<%BANNER%>

Eastern Mosquitofish as a Bioindicator of Pulp and Paper Mill Effluents


PAGE 1

EASTERN MOSQUITOFISH AS A BIOI NDICATOR OF PULP AND PAPER MILL EFFLUENTS By JESSICA JOY NOGGLE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Jessica Joy Noggle

PAGE 3

iii ACKNOWLEDGMENTS First of all, I thank my advisor, Dr. Timothy Gross, for accepting me as a young and nave graduate student. Dr. Gross taught me the finesse required to navigate the interdisciplinary realm of toxicology that stretches across academia, industries, and regulatory agencies. His encouragement to attend meetings, give presentations, and get involved with the regulatory implications of my research was inspiring. I am forever grateful for his gene rosity and patience. In addition, I would like to thank my other supervisory committee members for having high expectations that in spired me to work hard and strive for the objective ideals inherent in science. Dr. Seplveda served as a role model and example from the very beginning, helping me pick up the pulp and pape r mill fish research where she left off. Her friendship and support whenever I had a question or needed help will always be remembered and I will always admire her wo rk ethic and integrity. I thank Dr. Gallagher specifically for his strict a dherence to hypothesis testing a nd its central importance to experimental design – his sc ientific rigor has permanen tly shaped how I pursue and interpret research. Dr. Perciv al engaged me with stimulati ng discussions about the “Ph” half of my Ph.D. and his conceptual point of view enlightened me about the philosophical framework of science and its context in soci ety. Finally, Howard Je lks was my favorite special member! His ecological perspective wa s invaluable; as were his assistance with species identification and field work de sign and his reminder to enjoy the ride.

PAGE 4

iv I am forever indebted to the illustriou s crew in the Ecot oxicology Program at USGS: they put up with long days of field wo rk, processing frustratingly small fish, and gobs of mosquitofish samples in storage, all with jokes and laughter! Without the assistance of Shane Ruessler, Nikki Kerna ghan, Jon Wiebe, Beverly Arnold, Travis Smith, Jesse Grosso, and Janet Scarborough, these projects would never have been completed. Special thanks go to Carla Wies er for teaching me all the ins and outs of radioimmunoassay techniques and demonstrati ng how to run a very organized (and safe!) lab. Special thanks also go to Wendy Math is for taking care of my paychecks and ordering anything I needed for my projects, whenever I needed it – even cheerleading pompons. Fellow graduate students Brian Quinn, Heath Rauschenberger, Jennifer Muller, Eileen Monck, and Kevin Johnson were equally supportive in assistance with projects and swapping graduate student woes! These projects were primarily funded by the National Council for Air and Stream Improvement, Inc., and I thank Ken Bradle y and Dennis Borton for their congenial support and thoughtful discussions over data that furthered my development as a scientist. Ken aided with several field surv eys and much of the va lidation work on anal fin morphology, and provided insight into prac ticalities of experimental design. Without the initial financial and c ontinuous professional support of Stewart Holm of GeorgiaPacific Corporation, these projects would neve r have left the drawi ng board. In addition to thanking Stewart, I thank all the mill personnel who generously allowed access to retention ponds and mill grounds for sampling (specifically Myra Carpenter and Ted Kennedy of Georgia-Pacific; To m Deardorff and Joel Bolduc of International Paper; and Ray Andrews, Chet Thompson and Greg W ynn of Buckeye Technologies). Special

PAGE 5

v thanks go to Chet and Greg for assistance in the field—I would never have found those field sites (or unlocked my keys from the tr uck) without their ex tensive local knowledge! Sincere thanks go to my friends and fam ily who saw me through the rollercoaster life of a graduate student with love and suppor t. Special thanks go to my wonderful yoga family who supported me with vacations, massa ges, and the warmth and hospitality of their homes; and to my Ohio family who loved me through phone calls, snail mail, and happy family reunions. I extend deepest th anks to my parents, Denny and Donna Noggle, for their unconditional support of all my endeavours. I am eternally grateful for their tag team efforts through th e last semester of my doctora l program! Finally, heartfelt thanks go to my spiritual teachers for leading me with strength, compassion, and the perspective that all is sacred.

PAGE 6

vi TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................xi LIST OF FIGURES.........................................................................................................xiv ABSTRACT....................................................................................................................xvi i CHAPTER 1 MOSQUITOFISH EXPOSED TO PULP AND PAPER MILL EFFLUENTS: USE OF A POTENTIAL INDICATOR OF EXPOSURE AND EFFECTS................1 Background of Pulp and Paper Mills............................................................................1 Economic Importance of Pulp and Paper Industry in the United States...............1 Production Processes and Technologies................................................................2 Pulping...........................................................................................................3 Bleaching........................................................................................................5 Water Pollution and Regulation in the US............................................................6 Effects of Pulp and Paper Mill Effluents on Fish.........................................................9 Nonreproductive Effects......................................................................................13 Reproductive Effects...........................................................................................15 Masculinization and Femininization Effects.......................................................18 Effects of Pulp and Paper Mill Effluents on Mosquitofish.........................................22 Masculinization...................................................................................................22 Precocious maturation.........................................................................................26 Behavior..............................................................................................................27 Reproduction.......................................................................................................28 Mechanism of Action..........................................................................................29 Mosquitofish as a Model Species...............................................................................30 Occurrence and Availability in Effluent-Receiving Systems..............................30 Reproductive Characteristics...............................................................................31 Mosquitofish as a Bioindicator of Pulp and Paper Mill Effluent...............................32 Definitions: Bioindicator and Biomarker............................................................33 Bioindicator Criteria for Success.........................................................................35 Practicality....................................................................................................35 Variability.....................................................................................................35

PAGE 7

vii Predictability................................................................................................35 Contribution of My Study...........................................................................................37 Specific Aim 1.....................................................................................................38 Specific Aim 2.....................................................................................................39 2 VALIDATION OF MOSQUITOFISH ENDPOINTS USED TO ASSESS EFFECTS OF PULP AND PAPE R MILL EFFLUENT EXPOSURE......................42 Introduction.................................................................................................................43 Materials and Methods...............................................................................................44 Mill Characteristics and Field Collection............................................................44 Gender Identification Using the Urogenital Papilla............................................45 Anal Fin Morphology..........................................................................................45 Sex Steroids.........................................................................................................47 Statistics...............................................................................................................49 Results and Discussion...............................................................................................50 Water Quality......................................................................................................50 Validation of Gender Identification Using the Urogenital Papilla......................50 Morphology.........................................................................................................51 Validations...................................................................................................51 Body Size for Fall 2000 Collection..............................................................52 Influence of Body Size on Anal Fin Morphology........................................53 Seasonality...................................................................................................54 Sex Steroids.........................................................................................................55 Validations...................................................................................................56 Seasonality...................................................................................................57 Conclusions.................................................................................................................60 3 DIMINISHED EFFECTS OF PULP AND PAPER MILL EFFLUENT ON EASTERN MOSQUITOFISH BEFORE AND AFTER MAJOR PROCESS IMPROVEMENTS.....................................................................................................76 Introduction.................................................................................................................77 Materials and Methods...............................................................................................79 Mill Characteristics.............................................................................................79 Field Collections..................................................................................................80 Morphology.........................................................................................................81 Sex Steroids.........................................................................................................81 Statistics...............................................................................................................82 Results and Discussion...............................................................................................82 Water Quality......................................................................................................82 Body Size and Condition.....................................................................................83 Males............................................................................................................83 Females.........................................................................................................84 Anal Fin Morphology..........................................................................................85 Males............................................................................................................85 Females.........................................................................................................86

PAGE 8

viii Sex Steroids.........................................................................................................87 Males............................................................................................................88 Females.........................................................................................................89 Association Between Anal Fi n Morphology and Sex Steroids...........................90 Conclusions.................................................................................................................91 4 VARIABLE EFFECTS OF EFFLUENT ON EASTERN MOSQUITOFISH COLLECTED BELOW THREE FLOR IDA PULP AND PAPER MILLS.............105 Introduction...............................................................................................................106 Materials and Methods.............................................................................................108 Mill Characteristics...........................................................................................108 Water Samples...................................................................................................109 Fish Samples......................................................................................................109 Sex Steroids.......................................................................................................110 Statistics.............................................................................................................110 Results and Discussion.............................................................................................111 Water Quality....................................................................................................111 Water Chemistry................................................................................................112 Body Size and Condition...................................................................................113 Males..........................................................................................................113 Females.......................................................................................................114 Anal Fin Morphology........................................................................................114 Males..........................................................................................................115 Females.......................................................................................................116 Sex Steroids.......................................................................................................118 Males..........................................................................................................119 Females.......................................................................................................120 Anal Fin Elongation and Sex Steroids...............................................................122 Males..........................................................................................................122 Females.......................................................................................................123 Conclusions...............................................................................................................124 5 DIFFERENTIAL INDUCTION OF EFFECTS IN MOSQUITOFISH EXPOSED TO BLEACHED KRAFT MILL EFFLUENT......................................137 Introduction...............................................................................................................138 Materials and Methods.............................................................................................140 Mill Characteristics and Exposure Scenarios....................................................140 Water Samples...................................................................................................143 Morphological Endpoints..................................................................................143 Hormonal Endpoints..........................................................................................144 Statistics.............................................................................................................144 Results and Discussion.............................................................................................145 Water Quality....................................................................................................145 Water Chemistry................................................................................................146 Body Size and Condition...................................................................................147

PAGE 9

ix Males..........................................................................................................147 Females.......................................................................................................147 Anal Fin Morphology........................................................................................148 Males..........................................................................................................148 Females.......................................................................................................149 Sex Steroids.......................................................................................................150 Males..........................................................................................................150 Females.......................................................................................................153 Anal Fin Elongation and Sex Steroids...............................................................156 Conclusions...............................................................................................................156 6 INVESTIGATION OF REPRODUCT IVE SUCCESS IN MOSQUITOFISH LIVING IN PULP AND PAPER MILL EFFLUENT DOMINATED SYSTEMS.170 Introduction...............................................................................................................171 Materials and Methods.............................................................................................173 Mill Characteristics...........................................................................................173 Water Samples...................................................................................................173 Population Survey.............................................................................................174 Morphology................................................................................................175 Sex Steroids................................................................................................176 Fry Production...................................................................................................176 Statistics.............................................................................................................177 Results and Discussion.............................................................................................179 Water Quality....................................................................................................179 Water Chemistry................................................................................................180 Population Survey.............................................................................................181 Body Size...................................................................................................184 Anal fin morphology..................................................................................184 Sex steroids................................................................................................185 Anal Fin Elongation and Sex Steroids...............................................................186 Fry Production...................................................................................................187 Summer 2003.............................................................................................187 Summer 2004.............................................................................................189 Anal Fin Elongation and Fry Production...........................................................192 Conclusions...............................................................................................................193 7 EVALUATION OF MOSQUITOFISH AS A BIOINDICATOR OF PULP AND PAPER MILL EFFLUENT EXPOSURE................................................................216 Summary...................................................................................................................217 Specific Aims Revisited...........................................................................................219 Bioindicator Criteria Revisited.................................................................................221 Other Model Fish Species.........................................................................................223 Future Work..............................................................................................................225

PAGE 10

x APPENDIX A FIELD SITES...........................................................................................................229 B SEX STEROID RADIOIMM UNOASSAY PROTOCOLS.....................................237 C POSTER AND PLATFORM PRESENTATIONS OF DISSERTATION RESEARCH.............................................................................................................240 LIST OF REFERENCES.................................................................................................242 BIOGRAPHICAL SKETCH...........................................................................................259

PAGE 11

xi LIST OF TABLES Table page 1-1 Select characteristics of the mills in my study according to receiving stream.........40 2-1 Water quality parameters of Rice Creek field collection site s in winter 2000.........64 2-2 Correlation coefficients (r2) for morphological measurements made before and after preservation in formalin; between USGS and NCASI laboratories; and between manual and computer-aided measurement by the same observer..............64 2-3 Average coefficients of varia tion for manual and computer-aided measurements by observer and among observers....................................................64 2-4 Body size parameters (ave + se) for mo squitofish collected in winter 2000...........65 2-5 Digestion and extraction efficiencies, and coefficien ts of variation (CV), by exposure and reproductive status fo r mosquitofish whole body hormone analysis.....................................................................................................................66 3-1 Water quality parameters of field co llection sites before (2000) and after (2002) process changes at the Georgia-Pacific Palatka mill....................................95 3-2 Body size parameters (ave + se) and sa mple sizes for mosquitofish collected before (2000) and after (2002) process changes......................................................96 4-1 Water quality parameters of field collection sites associated with three effluent-receiving streams in Florida the summer of 2001....................................127 4-2 Concentration of selected effluent components in single grab water samples from field collection sites associated with three effluent-receiving streams in Florida the summer of 2001...................................................................................128 4-3 Body size parameters (ave + se) for mos quitofish collected from three effluentreceiving streams in Florida the summer of 2001..................................................129 5-1 Water quality parameters (ave + se) measured three times weekly (n = 13 total) during four week tank exposures of mos quitofish to bleached/unbleached kraft mill effluent in summer 2002.................................................................................159

PAGE 12

xii 5-2 Water quality parameters (ave + se) measured three times weekly (n = 12 total) during caged exposures of mosquitofish to field sites in Rice Creek during summer 2002..........................................................................................................159 5-3 Concentrations of selected effluent components in 100% final effluent sampled weekly midJanuary to midMay in 2002.................................................................159 5-4 Body size parameters (ave + se) for mosquitofish exposed to bleached/unbleached pulp mill effluents vi a whole effluent dilutions for four weeks in summer 2002...........................................................................................160 5-5 Body size parameters (ave + se) for mosquitofish caged in Rice Creek field sites for four weeks in summer 2002.....................................................................161 6-2 Water quality parameters (ave + se) measured three times weekly (n = 16 total) during laboratory fry production of female mosquitofish collected from field sites in Rice Creek during summer 2003...............................................................197 6-3 Water quality parameters (ave + se) at field sites where female mosquitofish were collected for fry production st udies over 4 months in summer 2004............198 6-4 Water quality parameters (ave + se) measured three times weekly (n = 16 total) during laboratory fry production of female mosquitofish collected from field sites in Rice Creek and Fenho lloway River during summer 2004.........................198 6-5 Concentrations of selected effluent components (ave + se) in single grab water samples from field sites where female mosquitofish were collected in Rice Creek and Fenholloway River during summer 2003.............................................199 6-6 Concentrations of selected effluent co mponents (ave + se) in single grab water samples from field sites where female mosquitofish were collected in Rice Creek and Fenholloway River during su mmer 2004 for fry production studies....200 6-7 Body size parameters (ave + se) fo r mosquitofish collected for population survey of Fenholloway River and Rice Creek in May 2003..................................201 6-8 Reproductive and morphological charact eristics of females collected from Fenholloway River and Rice Creek and monitored for fry production in 2003.....202 6-9 Reproductive and morphological characteristics of females collected for fry production from Fenholloway River in 2004.........................................................203 6-10 Reproductive and morphological characteristics of females collected for fry production from Rice Creek in 2004......................................................................204 A-1 Latitude, longitude, and descriptions fo r mosquitofish collection sites in Rice Creek......................................................................................................................232

PAGE 13

xiii A-2 Latitude, longitude, and descriptions for mosquitofish collection sites in Fenholloway River.................................................................................................233 A-3 Latitude, longitude, and descriptions for mosquitofish collection sites in Elevenmile Creek...................................................................................................234

PAGE 14

xiv LIST OF FIGURES Figure page 1-1 Categories of pulp and paper mill facilities.............................................................41 2-1 Maps of Rice Creek, a tributary of the Saint Johns River, FL, USA.......................67 2-2 Gender agreement between NCASI and USGS laboratories...................................69 2-3 Gender agreement within USGS laboratory.............................................................70 2-4 Female index of anal fin elongation for each site by 5 mm increments (winter 2000).........................................................................................................................7 1 2-5 Index of anal fin elongation for winter and summer months in 2000.....................72 2-6 Female whole body sex steroids from collections made in the summer and winter of 2000 (ave + se)..........................................................................................73 2-7 Percentage of female mosquitofish with masculine and feminine sex steroid ratios collected in 2000............................................................................................74 2-8 Male whole body sex steroids from colle ctions made in the summer and winter of 2000 (ave + se).....................................................................................................75 3-1 Maps of Rice Creek and Saint Johns River, USA....................................................97 3-2 Representative male gonopodia from th e upstream site collected before and after process changes................................................................................................98 3-3 Male index of anal fin elongation for each site by 0.1 mm increments...................99 3-4 Representative female anal fins from collections made before and after process changes...................................................................................................................100 3-5 Female index of anal fin elongati on for each site by 0.1 mm increments.............101 3-6 Male whole body sex steroids (ave + se ) from collections made before (2000) and after (2002) process changes...........................................................................102 3-7 Female whole body sex steroids (ave + se) from Rice Creek collections made before (2000) and after (2002) process changes....................................................103

PAGE 15

xv 3-8 Percentage of female mosquitofish with masculine and feminine sex steroid ratios collected after process changes in 2002.......................................................104 4-1 Maps of field sites..................................................................................................130 4-2 Index of anal fin elongation for mosq uitofish collected in summer 2001 from three effluent-receiving systems in Florida............................................................132 4-3 Whole body sex steroids (ave + se) for male mosquitofish collected from three effluent-receiving streams in Florida the summer of 2001....................................133 4-4 Whole body sex steroids (ave + se) fo r female mosquitofish collected from three effluent-receiving streams in Florida the summer of 2001...........................134 4-5 Percentage of female mosquitofish with masculine and feminine sex steroid ratios collected from three effluent-r eceiving streams in Florida the summer of 2001........................................................................................................................135 5-1 Diagram of tank facility for flow -through whole effluent exposure of mosquitofish in summer 2002 at Georgi a-Pacific’s Palatka, FL operation...........162 5-2 Map of cage locations for in situ field exposures at Rice Creek, FL, in 2002.......163 5-3 Concentrations of selected wood extract ives in 100% final effluent from the Rice Creek mill during tank and fiel d exposures of mosquitofish.........................164 5-4 Index of anal fin elongation (length ratio of Ray 4 to Ray 6) for mosquitofish exposed to bleached/unbleached pulp m ill effluents via whole effluent dilutions or onsite caged exposures for four weeks in summer 2002....................165 5-5 Whole body sex steroids (ave + se) for male mosquitofish exposed to bleached/unbleached pulp mill effluents vi a whole effluent dilutions or onsite caged exposures for four weeks in summer 2002..................................................166 5-6 Percentage of male mosquitofish wi th masculine and feminine sex steroid ratios exposed to bleached/unbleached pulp mill effluents via whole effluent dilutions or in situ field exposures for four weeks in summer 2002......................167 5-7 Whole body sex steroids (ave + se) for female mosquitofish exposed to bleached/unbleached pulp mill effluents vi a whole effluent dilutions or onsite caged exposures for four weeks in summer 2002..................................................168 5-8 Percentage of female mosquitofish with masculine and feminine sex steroid ratios exposed to bleached/unbleached pulp mill effluents via whole effluent dilutions or in situ field exposures for four weeks in summer 2002......................169 6-1 Maps of field sites..................................................................................................205

PAGE 16

xvi 6-2 Representative changes in resin ac id concentrations during summer 2003 at Fenholloway River and Rice Creek fiel d sites where mosquitofish were collected.................................................................................................................206 6-3 Representative changes in resin ac id concentrations during summer 2004 at Fenholloway River and Rice Creek fiel d sites were mosquitofish were collected at the same time......................................................................................207 6-4 Estimated relative abundances of mosquitofish for Fenholloway River and Rice Creek sites......................................................................................................208 6-5 Estimated age and sex structure of mosquitofish populations living near pulp and paper mill effluent discharge...........................................................................209 6-6 Index of anal fin elongation for mosq uitofish collected in summer 2003 from Fenholloway River and Rice Creek.......................................................................210 6-7 Whole body sex steroids (ave + se) for mosquitofish collected in summer 2003 from Fenholloway River and Rice Creek...............................................................211 6-8 Percentage of mosquitofish with ma sculine and feminine sex steroid ratios collected in summer 2003 from Fenhollo way River and Rice Creek (systems divided by solid black line)....................................................................................212 6-9 Viability of primary and secondary clutches produced by females collected from Fenholloway River and Rice Creek in 2003..................................................213 6-10 Fecundity and individual fry weight of primary and secondary clutches produced by female mosquitofish coll ected from Fenholloway River and Rice Creek in summer 2003...........................................................................................214 6-11 Adjusted fecundity of primary clut ches produced by female mosquitofish collected monthly in 2004......................................................................................215 A-1 Total monthly precipitation for Flor ida regions where mosquitofish were collected from pulp and paper mill effluent-receiving systems.............................235

PAGE 17

xvii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EASTERN MOSQUITOFISH AS A BIOI NDICATOR OF PULP AND PAPER MILL EFFLUENTS By Jessica Joy Noggle May 2005 Chair: Timothy S. Gross Major Department: Physiological Sciences A variety of sublethal physio logical effects have been re ported for fish exposed to pulp and paper mill effluents. As mill processing technologies improve, mounting evidence demonstrates fewer effects potentially linked to reduction in wood extractives. Repercussions of sublethal effects at hi gher levels of biological organization are important questions beginning to be explored. The goal of my study was to evaluate whether sublethal effects in mosquitofish can be used reliably to indicate adverse impact of pulp and paper mill effluents. Biomarke rs of anal fin morphology and whole body sex steroids were validated, then studied exte nsively in wild-caught mosquitofish from Florida effluent-receiving str eams that varied markedly in effluent composition. These biomarkers were also examined under shortterm controlled whole effluent exposures (caged in field and in tank s at 0, 10, 20, 40, and 80% diluti ons); and in relation to fry production and preliminary population survey s. Extent of female anal fin masculinization, or development of malelike secondary sex ch aracteristics, was

PAGE 18

xviii associated with increasing concentrations of wood extractives in wa ter samples from field sites. Implementation of EPA’s Cluster Ru le at one mill was followed by a significant reduction, but not elimination, of this response. However, masculin ization could not be reproduced under controlled e xposure, likely due to insufficient exposure duration. Alterations in sex steroids were manifested in both sexes for wild-caught and experimentally-exposed fish: males exhibite d feminized hormonal profiles will females displayed masculinized profiles. Differen tial responses among cage-exposed and tankexposed fish indicated additional environmen tal factors (such as bacterial communities hypothesized to degrade effluent compone nts into androgenic compounds) were influential in producing respons es. However, large natural variation at unexposed sites and indication of seasonality precluded definiti ve interpretation. The lack of association between these biomarkers demonstrated w hole body sex steroids cannot be used to predict morphological masculinizat ion; but they can be compared as biomarkers of recent versus past exposure. Finally, reproductiv e success studies implied mosquitofish may adapt different reproductive strategies in effluent-receiving streams. Since neither biomarker could be linked to differences in fry production or population structure, at this point mosquitofish may not be a suitable bi oindicator of adverse e ffects due to pulp and paper mill effluents.

PAGE 19

1 CHAPTER 1 MOSQUITOFISH EXPOSED TO PULP AND PAPER MILL EFFLUENTS: USE OF A POTENTIAL INDICATOR OF EXPOSURE AND EFFECTS Pollution from both point and nonpoint source s of human activity releases a variety of chemicals into the aquatic environment. While lethal effects have been addressed, sublethal effects of these chemicals on aquati c wildlife remain controversial. Whether observed sublethal effects lead to adverse impacts includ ing reproductive, population, or community level effects is arguable. This question has been strongly debated about pulp and paper mill effluents and sublethal effects in fish. The controversy becomes more complex as major processing improvements are implemented by the industry. The goal of my study was to evaluate wh ether sublethal effects in mo squitofish can be reliably used to indicate adverse impacts of pulp and paper mill effluents. Background of Pulp and Paper Mills Economic Importance of Pulp and Pa per Industry in the United States Pulp and paper products (such as writi ng and copy paper; sanitary tissues; cardboard; linerboard; and indirect products li ke pill capsules, diapers; and rayon) are significant economic commodities in the Unite d States (US). The US produces around 30% of the world’s paper and paperboard (US Environmental Protection Agency (EPA) 2002). In 2000, US pulp and paper mills produced 79 billion US dollars in shipments and employed 182,000 people, while Americans consume around 300 kg of paper-based products each year (EPA 2002). Industriali zation of nations incr eases demand for pulp

PAGE 20

2 and paper products (Smook 1999), resulting in hi gh capital investments and intensive use of forests, water, and energy. Furnish (tree species or sp ecifically cellulose fiber source) usually comes from harvested hardwood and softwood tree species ( Table 1-1 ). The fiber length varies tremendously between these two tree types (l onger in softwoods), so the final product determines which tree type and species is us ed. Alternative furnis hes include recycled paper (increasingly used, especially fo r products like corrugated board) and nonwood sources such as bagasse, bamboo, cereal straws cotton rags and linters, flax, hemp, and synthetic fibers (Smook 1999). Pulp and paper production consumes large amounts of forestry resources. Approximately 6 million acres of forest in the southeastern US are logged annually, mostly for paper production (W ear and Greis 2002). The process is also water and energy intensive. The pulp and pa per industry is classified as the largest industrial water consumer and the third largest in dustrial energy consumer in the US (US Department of Commerce 2000, US Department of Energy 2000). Efforts toward more sustainable forestry practices (e.g., the Fore st Stewardship Council certification program, http://www.fscus.org/ ), water use reduction, and increase d use of wood waste material for fuel are ongoing to reduce na tural resource consumption. Production Processes and Technologies The following summary is based on inform ation from Smook (1999) and US EPA (2002), unless cited otherwise. Final paper pr oducts are generated in two overall steps: pulp production and paper or paperboard manuf acture. Pulp and paper mills can be classified by whether they produce pulp, paper/paperboard, or both ( Figure 1-1 ). Most US mills are nonintegrative facilities and pr oduce paper products using pulp obtained off-

PAGE 21

3 site (54%). About one-third are integrative a nd produce pulp and final paper products (36%), Ten percent exclus ively produce market pulp. Depending on final product, 80% of market pulp is for paper and 20% for nonpaper ( Figure 1-1 ). Nonpaper pulps are either dissolvi ng pulp, fluff pulp, or specialty pulp. Dissolving pulp is a “chemical cellulose” that can be converted in to rayon, cellophane, cellulose acetates, cellulose ni trate, and carboxymethyl cellulo se via a modified kraft or sulfite pulping process. These chemicals are used in a variety of nonpaper products ranging from synthetic clothing to pill capsules to air filters. Fluff pulp is a very soft and absorbent form of pulp used in diapers, fe minine products, and hospital pads. Specialty pulp comprises the remaining nonpaper pulps that do not fit in the other two groups (final products include components of shoe soles and laminates). Facilities involved in pulp production (int egrative and market pulp facilities) face the biggest challenges of water pollution in th e industry. Pulp produc tion consists of five major steps: furnish preparation (debarking and chipping); pulping (breakdown of furnish into fibers); pulp refinement (removal of impurities, cleaning and thickening of pulp); bleaching (to whiten and brighten the pulp); and stock preparation (wet additives are integrated into pulp based upon desired end pr oduct). Two of these steps (pulping and bleaching) are considered the dominant sour ces of water pollution within pulp production (details next). Pulping Two major components of wood are cellulose (the fibers) and lignin (the glue holding fibers together). Broadly speaking, pulping unglues wood and reduces it to a fibrous mat. More specifically, the goal of pulping is to retain intact cellulose fibers while releasing all other wood components. These components include hemicellulose;

PAGE 22

4 lignin; and extractives such as resin acids, fatty acids, phytoste rols, turpenoids and alcohols. Pulping methodologies are generally classified as chemical, semichemical, or mechanical. Most North American pulping technologies (70%) involve chemical processes ( Table 1-1 ). Chemical pulping digests wood chips at high temperatures and pressures, usually in an alkaline solution (ca lled the kraft process), or historically in an acidic solution (the sulfite process). The kraft process (also know n as the sulfate process) dominates North American chemical pulping technologies ( 95%). Major advantages over the sulfite process are high strength pulp (“kraft” is Ge rman for strong) and recove ry and reuse of digestion chemicals. Lignin removal is high, allo wing for extensive bleaching without pulp degradation (via delignification). Additional advantages are the wide range of furnishes that can undergo the kraft process, and the to lerance for bark. However, pulp yield is relatively low (40-50% of furn ish) compared with mechan ical pulping. An additional disadvantage is pulp color: th e kraft process produces a da rk brown pulp that requires extensive bleaching, neutralizing the bleaching-related benefits of high lignin removal. Overall, the kraft process has proven to be the most cost-effective chemical pulping technique. Kraft pulping is cyclical, beginning and ending with white liquor. White liquor, composed of sodium hydroxide (NaOH) and sodium sulfide (Na2S) as the active ingredients, is the al kaline solution used to digest wood chips. Temperature and pressure is elevated using more conventional batch di gesters or less common c ontinuous digesters. Raw pulp and residual black liquor are produce d. The pulp is destined for refinement, bleaching, and stock preparation, while the blac k liquor is concentrated and burned into

PAGE 23

5 an inorganic smelt that is dissolved to produ ce green liquor. Some black liquor is washed away into effluent, and some carries over w ith the pulp. These black liquor losses impact effluent quality. Green liquor is then causticized to re generate white liquor. This efficient chemical recovery is integral to the success of the kraft process. Some kraft pulping by-products (turpentine and tall oil) are recovered and either reused as fuel sources or sold, depending on market prices. The rest of the noncellulose wood components are discharged (as part of th e final effluent) into an aquatic receiving environment. Bleaching Bleaching of pulp is often desirable because it produces a whiter, brighter, softer, and more absorbent end product. Roughly ha lf of all paper products in the US are bleached. Bleaching potential for a pulp depends on two factors. Inherent lignin content of furnish: highe r lignin content gives a darker color, and softwoods tend to have more lignin than hardwoods. Pulping process: sulfite chemical pulping pr oduces a relatively bright pulp with low residual lignin content, wh ile kraft chemical and semichemical pulping produces a darker pulp. The brightness of mechanic ally produced pulp is dependent on lignin content of furnish. In general, early stages of the bleaching sequence continue delignification begun during pulping, while later stages focus on oxida tion to remove any residual color. Modern bleaching uses a continuous sequen ce of alternating acidic and alkaline stages with washing between stages. Wash ing usually involves large amounts of water that is collected and discharged as part of the final effluent Several bleaching agents are available (shorthand used by the industry for bleaching sequences is given in parentheses). Hypochlorite (H)

PAGE 24

6 Elemental chlorine (C) Chlorine dioxide (D) Oxygen (O) Hydrogen peroxide (P) Ozone (Z) Sodium hydroxide (E) Historically, bleaching was accomplished using hypochlorite (Turoski 1998). Hypochlorite (as either calcium or sodium hypochlorite) was initia lly the only bleaching agent at the turn of the twen tieth century (H or HH bleachi ng sequence). The addition of elemental chlorine gas commercially in 1930 was a major advance that reduced the amount of hypochlorite needed, and became the standard first stage of bleaching followed by an extraction (E) stage (CEH). A decade late r, chlorine dioxide (e.g. CEHDED) and hydrogen peroxide (e.g. CEHD(Ep)D) began co mmercial use. Chlorine dioxide eventually replaced hypochlorite in the later stages of bleaching by the 1960s (e.g. CEDED) because of its powerful bright ening combined with high selectivity for lignin. As chlorine dioxide gained popularity, oxygen and ozone bleaching were initiated (e.g. OCEDED or OZED). These latter agents have been slow to gain acceptance by the industry because of complications with low se lectivity for lignin removal. However, as environmental and health concerns about ch lorine began forming in the 1970s, reducedchlorine and chlorine-free methods of bleaching (using 100% chlorine dioxide substitution, hydrogen peroxide, oxygen and oz one) have been e xpanded and refined. Table 1-1 compares bleaching strategies of the three mills in my study. Water Pollution and Regulation in the US As previously emphasized, effluent (dischar ge of liquid waste from a factory/plant) from integrative and market pulp facilities is of most concern to environmental and human health. Within pulp production, th e pulping and bleaching stages primarily

PAGE 25

7 contribute to water pollution. Four categories of water pollu tion are currently monitored: effluent solids, oxygen demand, color, and to xicity. In addition, major water quality characteristics of effluent-re ceiving waters (such as pH, temperature, dissolved oxygen, alkalinity, and conductivity) ar e expected to remain unchanged or minimally changed (and may be subject to regulation as well). Abatement efforts to control these variab les occur within plant processing systems and post-processing. More efficient use of raw materials, reuse of mill waters to create a more closed system, and reduced effluent vol ume are strategies within mill operations that are very effective at in creasing profits and at restri cting contaminants produced. Additionally, external or endof-pipe treatment of efflue nt helps reduce or remove contaminants. Primary external treatment en tails sedimentation in settling basins to remove suspended solids. Secondary treatment reduces biochemical oxygen demand using biological degradation/oxidation. Occasionally, mills also apply a tertiary treatment to reduce color (turbidi ty), but this step is costly. Since pulp production releases disproportiona te amounts of chemicals into air and water (compared to other industries releasing pr imarily to land), health-related concerns center on air emissions and aquatic toxicity. W ith the discovery of dioxins and furans in fish collected downstream of a pulp and paper mill in 1985 (Smook 1999) and related evidence for biological effect in fish by the Environment-Cellulose project of Sweden in the late 1980s (Lehtinen 2004), public attent ion focused on environmental impacts of pulp and paper mills. Dioxins and furans are a class of chlorinated organic compounds produced mainly by incomplete combustion of organic compounds. Natural sources are forest fires and volcanic erupt ions, while human sources incl ude waste incinerators, coal

PAGE 26

8 and oil-fired power plants, vehicle exhaust, chlorinate d pesticide and herbicide production, and chlorine bleaching during pulp production. Although pulp and paper mills represent a minor source of chlorinated organics, these compounds are lipophilic and persistent in the environment. Hence, they have the potential to biomagnify up the food chain through fish and potentially to hum ans, causing sublethal, chronic toxicity (which is not traditionally monito red). The most toxic congeners 2,3,7,8tetrachlorodibenzodioxin and 2,3,7,8-tetrach lorodibenzofuran (TCDD and TCDF respectively) are classified as pr obable human carcinogens by the EPA.1 Release and exposure predominately occurs as mixtures of chlorinated compounds (measured as adsorbable organic halides or AOX) includ ing dioxins and furans, which can enhance toxicity. So pollution prevention efforts have focused on reducing release of chlorinated compounds as a group. To reduce the toxic release of chlorina ted compounds to both air and water, EPA enacted a landmark regulation deemed the Cl uster Rule in April 1998. The rule set new baseline limits for toxic and nonconventi onal pollutant releases; and aims for approximately 60% reduction in air emissions and virtual elimination of chlorinated organic compounds in water (US EPA 1997). I ndividual mills are allo wed flexibility in tailoring pollution prevention technologies to their specif ic situations. A voluntary incentive program for technologies above a nd beyond the rule grants mills a variable compliance period (3 to 8 years). Paper-grade bleached kraft and su lfite mills are most affected by the Cluster Rule, requiring 100% chlorine dioxide substitution for elemental 1 The US Department of Health and Human Se rvices issues a Toxicological Profile for Chlorinated Dibenzop -Dioxins containing a detailed survey of human exposure and effects.

PAGE 27

9 chlorine in the bleaching sequence, rendering th ese mills Elemental Chlorine Free (ECF). In addition to this source control, spill c ontrol of black liquor is required. Beyond these two requirements, mills develop their own approved plan to meet the new limits, potentially including voluntary measures su ch as extended deligni fication, closed loop technologies, or Total Chlorine Free (TCF) bleaching. Table 1-1 shows different pollution prevention strategies adopted by th e mills in my study. Ultimately regulating both media (air and water) at the same time cr eates a synergistic re duction in pollution. Once fish living downstream of pulp and paper mills were discovered with measurable dioxins and furans in their tissu es, intensive research into exposure and effects on fish was initiated. Concerned with biomagnification to humans and carcinogenicity, regulatory agencies were also interested in potential adverse effects on aquatic life. As the next section shows, that research led to questions about reproductive impairment as an effect of pulp and paper mill effluent exposure. Effects of Pulp and Paper Mill Effluents on Fish The interaction between industry and governme nt regulatory agencies (primarily in North America and Scandinavia) produ ced a large body of knowledge (centered upon fish) concerning aquatic toxicity of pulp mill effluents. In general, effects have shifted from gross alterations in growth and acute, leth al toxicity; to more subtle sublethal effects influencing development, maturation and repr oduction. In Canada, regulation passed in the early 1990s (Environmen tal Effects Monitoring Progr am) produced a decade of consistent fish research at all Canadian mills (McMaster et al. 2003). Also, since 1991, five international conferences have provi ded a forum to discuss research on the environmental impacts of pulp and paper m ill effluents, all of which have published

PAGE 28

10 proceedings (Sodergren 1991, Servos et al. 1996, Ruoppa et al. 2000, Stuthridge et al. 2003, Borton et al. 2004). Many of these studies are field-based, pr ecluding an identification of causative agents in effluent. Converse ly, studies using controlled exposure to whole effluent dilutions preclude isolation of bioactive e ffluent components, yet retain environmental relevance. Several research efforts have addressed controlled exposure to specific effluent components that are not easily extrapolated to observe d effects in the field. Most recently, efforts to elucidate bioactive com pounds have used bioassay-based fractionation studies. While appealing in theory, van den Huevel (2004a) points out two major drawbacks of these studies. Isolation of bioactive com pounds within mill processes as opposed to in final effluent. Dependence on receptor-binding studies (some of which use human receptors as opposed to fish receptors) when a receptor-mediated mechanism has not been firmly established. Chlorinated organics were thought to be key com ponents causing toxicity. However, their virtual removal from effluent has not been associated with removal of chronic, sublethal effects (Lehtinen 2004). Importantly, implementation of the Cluster Rule has also led to large reductions in nonchlorinated, nonconventional pollutants such as wood extractives, which could be tied to reduction in effects. Despite the above caveats, bioassay-based fractiona tion studies have provided the most specific attempts at identifying which portions of the nonconventiona l pollutants may be causing effects. For example, in association with studies on redu ced steroidogenesis, lignin derivatives such as polyphenolics were identified as bioactiv e agents in condensates of black liquor

PAGE 29

11 (Hewitt et al. 2002). On the other hand, phytoste rols were not determined to be causing observed effects (Dube and MacLatchy 2001). Other mechanistic-based studies found liga nds for the estrogen receptor and sex steroid binding protein present in pulp mill effluents, implying an estrogenic cause for well-documented reproductive effects (Hewitt et al. 2000, Pryce-Hobby et al. 2003). In support of these findings, known (weakly) estrog enic compounds were recently identified in effluents such as genistein (Kiparississ et al. 2001) and industri al nonionic surfactants (nonylphenol ethoxylates) (Lee and Peart 1999). Additional studies showed significant estrogenic properties of the phytosterol -sitosterol in fish (MacLatchy and Van der Kraak 1995, Tremblay and Van der Kraak 1998). Androgenic properties of pulp and paper m ill effluents have also been reported (Svenson and Allard 2004). In vitro assays using the human androgen receptor showed effluents from softwood furn ish, but not hardwood, produced low levels of androgenic activity. Biological treatment of effluent had no effect on andr ogenicity. Uptake of these androgenic compounds by fish and conjugati on in bile was also demonstrated. However, researchers have had little success in associati ng androgenic compounds with masculinization effects. For inst ance, androstenedione and human androgen receptor binding was detected in water and sediment samples downstream from one of the mills in my study and associated with masc ulinized fish (Parks et al. 2001, Jenkins et al. 2003). Evidence for androstenedione as a bioactive effluent component was then refuted by more quantitative analysis show ing androstenedione was present only in fractions that did not induce human androgen receptor activit y and expression (Durhan et al. 2002). Phytosterol degradation analysis of effluents from these same waters did not

PAGE 30

12 reveal androstenedione metabolites but detect ed androsteneone (Quinn 2004). Separate studies also refuted androste nedione and testosterone as the active androgens causing masculinizing effects (Ellis et al. 2003), and showed in vitro fish receptor binding responses that do not correlate with in vivo effects. Follow up studies with this mill effluent (van den Huevel et al. 2004b) using more accurate measures of both masculinization and fish receptor binding, failed to produce any response. Although no major process changes occurred between st udies, treatment system maintenance was improved as indicated by gradual re duction in total suspended solids. In reality, the effluent components responsib le for causing subletha l toxicity in fish may never be determined, even though specifi c mechanisms of action may be narrowed down. As in other technology sectors, po llution prevention technology in the pulp and paper industry rapidly progresses. For this industry, the movement is toward a closed system with recycling and reuse of all ma terials (Lehtinen 2004). Implementation of these technologies is the lim iting factor, since the industry is capital-intensive; but voluntary incentive programs (such as the one associated with the Cluster Rule) help offset investment risks. As the monitored e ffects diminish and potentially disappear with improving pollution prevention technologies, id entification of the specific bioactive compounds may not be necessary. The following literature review of effects in fish exposed to pulp and paper mill effluent begins with nonreproduc tive effects, followed by gene ral reproductive effects. It ends with specific reproductiv e effects indicating masculini zation or femininization of fish.

PAGE 31

13 Nonreproductive Effects A variety of sublethal, nonr eproductive physiological effects have been reported in fish exposed to pulp and paper mill effluent s. Alteration of liver function (mainly induction of the detoxifying cytochrome P450 system as measured by ethoxyresorufin-odeethylase (EROD) activity) was the most cons istently reported nonr eproductive effect in fish. Additional work with conjugating det oxification systems, stress, hematology, and immunological responses received compara tively scant attention and effects are conflicting. As a more general measure of health, growth rates were also examined, although results are equally difficult to interpret. Significant EROD induction (typically a 2to 4-fold increase), usually accompanied by an increase in liver somatic index, is often considered a nonspecific marker of effluent exposure (Rogers et al 1989, McMaster et al. 1991, Servos et al. 1992, Gagn and Blaise 1993, Ahokas et al. 1994, Munkittrick et al. 1994, Gagnon et al. 1995, Soimasuo et al. 1995, Martel et al. 1996, Ma rtel and Kovacs 1997, Soimasuo et al. 1998, Seplveda et al. 2002, van den Huevel et al 2002). Yet the plethora of compounds known to induce this detoxification response (combined with the variable nature of effluent composition within and among mills) makes it difficult to link this biomarker to specific chemical compounds. Dioxins and fura ns are strong inducer s (Servizi et al. 1993), although EROD induction was also dete cted when these compounds were not present (Munkittrick et al. 1992). Unfortunate ly EROD activity could not be consistently tied to reproductive effects (Munkittrick et al. 1999). Induction of conjugating detoxification sy stems (as well as stress, hematological changes, and immunological res ponses) has also been addres sed in the literature, albeit with much less intensity compared to EROD act ivity. Results are often mixed; and most

PAGE 32

14 of this work has not been linked to advers e effects at higher levels of biological organization, so the consequences of thes e physiological changes remain unclear. For example, oxidative stress has been reported in fish due to induction of hepatic enzymatic (glutathione peroxidase, glutathione S -transferase, catalas e) and nonenzymatic (glutathione and metallothein) antioxidants, as well as lipid peroxidation in gill and kidney (Oikari et al. 1988, Stephensen et al 1998, Ahmad et al. 2000, Fatima et al. 2000). However, hepatic antioxida nts (mainly glutathione S -transferase and glutathione) were not induced by other studies (Mather-Miha ich and DiGuilio 1991, Bucher et al. 1992, Larsson et al. 2002). Similarly, effects on ac tivity of another conj ugating detoxification enzyme (uridine diphopsphate glucuronosyl transferase) range from induction to inhibition (Oikari et al. 1983, Frlin et al. 1985, Lindstrm-Seppa and Oikari 1988, Andersson et al. 1988b, Lindstrm-Seppa et al. 1989). These mixed results on conjugating detoxification systems are likely due to differences in effluent quality, experimental design, exposure c onditions and life stage at ti me of exposure. They are representative of results for the inter -related responses in stress, immune, and hematological functions (see Seplveda 2000 a nd van den Huevel 2004a for discussion of these latter parameters). Likewise, conflicting results ex ist for growth patterns of fish exposed to pulp and paper mill effluents. For instance, Warren et al. (1974) and Munk ittrick et al. (1991) detected reductions in growth of fish in laboratory and field co llections respectively. Other field and laboratory studies found no effects of pulp and paper mill effluents on growth rates in fish (Servizi et al. 1993, Sw anson et al. 1992). Additionally, some field collections documented increased growth at effluent-exposed sites (McLeay and Brown

PAGE 33

15 1974, Sandstrom et al. 1988). Explaining some of these discrepancies, Gagnon et al. (1995) found accelerated growth characteristic of downstream nutrient loading from both natural and anthropogenic sources, independent of effluent exposur e. Despite rapid growth, fish collected downstream of the bleached kraft mill did not have concomitant increases in reproductive effort; rather, th ey exhibited greater length at maturity, reduction in gonad size and highly variable fecundity compared to reference fish. Similar reproductive impairment was detected by Munkitt rick et al. (1991) in fish with reduced growth. Such findings focused researcher s toward evaluating reproductive impacts of pulp mill effluents on fish. Reproductive Effects Dominant reproductive effects of pulp a nd paper mill effluent exposure include depressed circulating sex steroi ds associated with alterati ons in steroidogenic capacity; reduced gonadal development; delayed sexua l maturation; and negative impacts on egg and fry quality. Effects on egg production and size have been debata ble, probably due to differences in the quality of effluent te sted. Although many of these reproductive parameters have improved with changing effl uent technologies, the virtual removal of chlorinated organics from effl uent has not eliminated responses. This finding leads researchers away from dioxins and furans as causative agents; and toward wood extractives such as phytosterols, which have been reduced but not eliminated by pollution prevention technologies. The most compelling evidence for physio logical reproductive alteration comes from a series of Canadian studies over the past 10 years (summarized by McMaster et al. 2003). Extensive work on white sucker ( Catostomus commersoni ) and many other fish species such as lake whitefish ( Coregonus clupeaformis ) and longnose sucker

PAGE 34

16 ( Catostomus catostomus ) showed inhibited gonadal development (primarily decreased gonadosomatic index) and depressed circ ulating sex steroids (primarily 17 -estradiol and 11-ketotestosterone) in both sexes (Munkittri ck et al. 1998). However, the steroid response was not 100% consistent, especially at more recent, modernized mills. Importantly, steroid effects not observed in the field at a modern mill (Servos et al. 1992) occurred in laboratory exposures of anothe r species at concentrations higher than observed in the receiving environment (R obinson 1994). Thus, bioactive compounds were still being produced, but differential spec ies sensitivity, effluent dilution, and/or conditions of the receiving environment pr otected wild fish. This emphasized the importance of field studies, despite inherent problems with identifyi ng causative effluent components. Along with the more persistent reduction in gonadal development of these fishes, a concerted effort was launched to identify mech anisms behind depression of sex steroids. Several sites along the pituitary-gonad axis appeared to be affected by exposure to bleached kraft mill effluents: pituitary f unction was decreased, ovarian biosynthetic capacity was reduced, and peri pheral steroids metabolism wa s inhibited (Van der Kraak et al. 1992, McMaster et al. 1995, 1996). C onflicting evidence was presented by Gagnon et al. (1994a), who conclude d that increased steroid me tabolism may reduce steroid levels. These discrepancies may have occurr ed from capture and handling stress (Jardine et al. 1996). Regardless of the contro versy over mechanism, improved processing technologies were followed by partial recove ry of reproductive function: steroids and potential mechanistic responses along the biosyn thetic pathway were either less impacted

PAGE 35

17 or no longer significant in both wild and cage-exposed fish, although gonad size was often still reduced (Mun kittrick et al. 1997, van den Huevel et al. 2004b). Whether these physiological effects tran slate into effects on the production, survival, and development of young has been mo re controversial. Early Scandinavian studies showed dramatic effects of pulp mill effluent on eggs and fry including reduced fecundity, smaller egg size, poorer fertilizati on of eggs, and decreased viability of fry (Vuorinen and Vuorinen 1985). More recent Sc andinavian experiments exposing fish to wood-derived phytoste rols (primarily -sitosterol) showed increas ed egg mortality, larval deformities, and maternal transfer of phytos terols to offspring (Lehtinen et al. 1999, Mattson et al. 2001). In contrast, exposure to another phytosterol (stigmastanol) unequivocally had no effect on egg, larval, or juvenile survival and quality (NCASI 1999). Similar to Canadian reports of reduced steroids at effluent concentrations higher than receiving stream concentrations, NCAS I (1996) documented reduced egg production in 18-100% v/v effluent from a bleached kraft mill. Unlike the Canadian steroid work, though, impacts in the field were not addresse d to verify a lack of observed laboratory effects. Initial Canadian reports on sex st eroids and gonadal devel opment also indicated reduced egg size and fecundity (Munkittrick et al. 1991). However, further research found fecundity to be quite va riable (Gagnon et al. 1994b) and fertility of eggs and sperm not affected at exposed sites (McMaster et al. 1992), despite the well-documented physiological responses. Research has been conducted on egg and fry characteristics of fish exposed to effluent from the primary mill investigated in our study, before implementation of Cluster Rule process changes. The NCASI (2000a) found egg production (but not hatchability)

PAGE 36

18 significantly reduced at 23% v/ v effluent, well within instre am concentrations (yearly averages aprroximately 60%).2 Additionally, exposure to 10 % or greater whole effluent dilutions did not result in e ffects on fecundity, egg size, or hatchability; but caused reduction in fry growth and survival (Seplve da et al. 2003). Parent fish also had reduced circulating sex steroi ds and gonad size at 20-40% effl uent dilution, similar to previous findings (Seplveda et al. 2001) and in support of Canadian research. These results are perhaps the most convincing li nk between physiological reproductive impact and more subtle influences on offspring, alt hough effects in fry could have originated from maternal transfer of bioactive efflue nt components instead of (or in addition to) direct impairment of pare ntal reproductive systems. Masculinization and Femininization Effects Studies of masculinization and femininizat ion of fish began with the desire to control sex ratios in the aquaculture industr y (Yamazaki 1983). A number of fish species have an innate capacity to regulate sex ratios in the population trigge red by subtle social and environmental contexts (Bar oiller et al. 1999). On this level, masculinization refers to complete sex reversal (females to male s) at the gonad level; and feminization, vice versa. Control of sex ratios is accomplishe d by careful exposure to sex steroids, altering ratios developmentally or in adult fish (t he latter sometimes resulting in sterility) (Pandian and Sheela 1995). Androgens (mainly the synthetic 17 -methyltestosterone) have been used to shift sex ratios to males. Natural estrogens (mainly 17 -estradiol) have been used to shift sex ratios to female s. However, male-biased sex ratios have also 2 Since implementation of process changes, NCASI has repeated its fish full life-cycle exposure and showed improvement in egg production (DL Borton, pers. comm.).

PAGE 37

19 been induced successfully, using aromatase i nhibitors (Jalabert et al. 2000, Kwon et al. 2000). (Aromatase converts androgens to es trogens in many tissues of both mammals and fish.) Effective doses vary drastically among species, with Po eciliids often requiring the largest doses compared to salmonids, c yprinids, cichlids and an abantids (Pandian and Sheela 1995). Environmental po llution has also been linked to unintentional shifts in sex ratio of fish (Jalabert et al. 2000). The terms masculinization and femininizati on have also been used to describe changes in secondary sex characte ristics. In this sense, masc ulinization refers to external appearance of male secondary sex characteristics in a fema le fish (i.e., she retains ovaries), and feminization vice versa. This phenomenon can o ccur naturally in the wild: arrhenoidy, or masculinization of older, re productively senescent females, has been documented at low levels in wild Poecil iid populations (Constanz 1989). Changes in secondary sex characteristics can be induced by administration of sex steroids (Turner 1941a, Turner 1942a, Turner 1942b, Hildemann 1954, Borg 1994). Androgens and estrogens are assumed to be key players, how ever progestins may also play a significant role (Jalabert et al. 2000). Changes in sec ondary sex characteristics have also been associated with human impacts on the environment (Jalabert et al. 2000). As endocrine disruption became a controvers ial issue in toxicology, the distinction between these levels of masculinization a nd femininization was important. Changes in appearance could behaviorally affect repr oduction and population size, or reproduction could be unaffected. On the other hand, a si gnificant shift in sex ratios could impact reproduction and population size more overtl y. The difficulty with sex ratios is determining how much of a sh ift is significant. Hence, for the purpose of my study,

PAGE 38

20 masculinization and femininization will re fer to alterations in secondary sex characteristics only. Shifts in sex ratios (imp lying alteration at the gonad level) will be referred to as such. Both of these endpoints have been associated with pulp and paper mill effluent exposure in fish. Effects on sex ratio varied by effluent exposure and species. Research on an indigenous species living near a TCF Swedish kraft mill demonstrated slight yet statistically significant male-biased sex ratios in embryos (55 to 58% male) compared to pooled reference sites (Larsson et al. 2000). Further, temporary mill shutdown allowed recovery of normal sex ratios, and the male bias reappeared after mill processes were restored (Larsson and Frlin 2002). Short-te rm (42 day) laboratory exposure of this effluent to a livebearing species did not reflect field results, failing to induce any change in sex ratios (Larsson et al. 2002). Full li fe-cycle exposure to bleached sulfite mill effluent demonstrated the opposite response in yet another species: sex ratios were female-biased at 30% effluent and greater (Parrott et al. 2004). In addition, egg production was reduced at 10% e ffluent and failed at 30% e ffluent or greater. Based upon these findings, egg production but not sex ratios may be impacted in the wild, since effluent concentrations vary from 1 to 15% by season and river flow. Finally, multigenerational laboratory exposure to enviro nmentally relevant levels of phytosterols, primarily -sitosterol, revealed male-biased sex ratios in the first offspring generation and female-biased sex ratios in the second (Nak ari and Erkomma 2003). Clearly, bias toward one sex or another cannot be generalized in response to pulp and paper mill effluent exposure, although changes in sex ratio may be a useful indicator of impacts on fish reproduction (Parrott et al. 2004).

PAGE 39

21 Changes in secondary sex characteristics of fish exposed to different types of pulp and paper mill effluent include precocious and delayed maturation, femininization, and masculinization. Among these alterations, ma sculinization is the most consistently reported effect across field and laboratory studies. Precocious maturation (or early development of secondary sex characteristics) has been reported in fish collected from a bleached kraft effluent receivi ng stream investigated in my study (Caruso and Suttkus 1988). Precocious maturation at 32% or grea ter effluent, masculinization (at 10% or greater) and femininization (at 32% or greater), were reported in fish exposed to dilutions of bleached sulfite effluent (Parrott and Wood 2002, Parrott et al. 2003, 2004). As with sex ratios and egg production reported by this group, the only environmentally significant response may be masculinization. Using the same model species, NCASI (2000b) found delayed maturation resulting from expos ure to bleached kraft mill effluent. Masculinization occurred in this species from exposure to a different bleached kraft mill effluent (Kovacs et al. 1995b) but not to effluent from a thermomechanical pulp mill (Kovacs et al. 1995a). In the study by Larrs on et al. (2002a) on a livebearing species, although sex ratios were not al tered, masculinization was weakly indicated by male-like coloration. Perhaps the strongest case fo r masculinization lies with effects on mosquitofish, initially documented by Howell et al. (1980). This sp ecies is elaborated upon in the next section. Examining these responses as a whole, in or der to be useful bi omarkers any of the suborganism level effects must be linked to exposure, as many studies included; and to effects on reproductive success, and if possibl e on populations. Full life cycle tests are very useful to this end, but these tests must be comparable to responses in wild fish

PAGE 40

22 actually living under exposure conditions. Henc e a two-pronged approach pairing field and laboratory exposures is ideal, especially if the same species can be used for both types of studies. The mosquitofish has poten tial for both types of exposures, as shown in the following section. The remainder of this chapter examines mosquitofish in relation to pulp and paper mill effluent, as a model species, and as a potential bioindicator of pulp and paper mill effluents. Effects of Pulp and Paper Mill Effluents on Mosquitofish The Eastern mosquitofish, Gambusia holbrooki was the first species recorded as masculinized by pulp and paper mill effluent exposure (Howell et al. 1980). Since then, improved analysis of field collections and laboratory exposure to degraded effluent components have supported the original obs ervational response. The degree of masculinization was highly variable with in a site and by season, yet considered comparable among three Florida mills (all of which were examined in my study). In contrast, controlled exposure to whole effl uent dilutions provided mixed evidence for masculinization in Western mosquitofish, Gambusia affinis (McCarthy et al. 2004). Beyond the masculinization response, precoci ous maturation, behavior, and aspects of reproduction (mainly brood size) were also addressed without si gnificant observable impacts. So far, attempts to isolate potential mechanism(s) of masculinization were inconclusive. Masculinization Sampling of Eastern mosquitofish in Elev enmile Creek, FL, USA revealed the first known occurrence of masculiniz ation associated with pulp and paper mill discharge (Howell et al. 1980, p. 676). Lacking quantific ation of data, the authors reported the following.

PAGE 41

23 All females within this stream are st rongly masculinized, possessing a male-like gonopodium and displaying male reproductiv e behavior. All males exhibit precocious secondary sex characte rs and reproductive behavior. Photomicrographs indicated elongation, se gmentation, and intermittent terminal differentiation of female anal fins similar to the male gonopodium (the copulatory organ used to inseminate females in this livebeari ng species). Apparently equivalent responses were detected in another e ffluent-receiving stream in Fl orida, the Fenholloway River (Bortone and Drysdale 1981). After significan t process changes at the Elevenmile Creek mill (including conversion to ECF bleaching and oxygen delignification), quantification of the response (anal fin length) and statis tical comparison to females from a reference stream showed masculinization rema ined (Cody and Bortone 1997). However photographs qualitatively indicated redu ced elongation and lack of terminal differentiation. Season (winter versus summer months, based upon seasonal drought conditions) also influenced an al fin length significantly: gr eater elongation occurred in summer months. As further evidence, Bortone and Cody (1999) detect ed a statistically significant increase in the ratio of anal fin length to fish standard length (finally accounting for the influence of body size on th is morphological feature), and inferred a distance/dose-dependent response downstream of pulp mill effluent discharge in Rice Creek, FL. In light of seasonal effects re ported previously (Cody and Bortone 1997), the inference was tenuous: they compared one upstream and three downstream sites collected in summer with a fourth, furthest downstream site collected twice; once in winter, and once several years previously by separate res earchers. Significant increase of anal fin elongation in females from the first two downstr eam sites was statistically comparable to Fenholloway River females collected several ye ars before. Variation was high (highest in the Fenholloway collection), even at the upstream (200 m above outfall) site. The

PAGE 42

24 authors speculated tidal influence may dr aw effluent above the discharge point, accounting for this unexpected result. Exposur e to pulp mill effluent, either potential or actual, was never documented in these collections. Based upon detailed observations of expos ure to androgens by Turner (1941a, 1942a, 1942b) and steroid production studies usin g bacterially degrad ed phytosterols by Marsheck et al. (1972), it was hypothesized that female mosquitofish may be masculinized when exposed to androgens fo rmed by degradation of phytosterols present in pulp and paper mill effluents. Subsequent ly, female mosquitofish were exposed to high concentrations of p hytosterols (approximately 0.1-0.5 g/L of stigmastanol and sitosterol) combined with a bacterium ( Mycobacterium smegmatis ) not common to effluent-receiving streams (Denton et al. 1985, Howell and Denton 1989). Although presence of androgens was not monitore d to verify androgen exposure, females developed male-like gonopodial structures with in two weeks. Stigmastanol produced a more potent effect. The male-like gonopodial st ructures did not elonga te to the length of normal male gonopodia, but developed terminal differentiations. Lacking quantification and statistics, nonetheless this mechanism wa s proposed to explain observed effects of masculinization. In support of these findings, Angus et al. (2001) detected rapid onset of anal fin elongation: 14 days at 60 g 11-keto testosterone/g food, deve lopment of terminal differentiations by 20 days in the high expos ure groups (80 and 100 g/g), and average of 40 days in the low exposure groups. Analysis of various morphological endpoint s revealed the most sensitive measures of masculinization. Bortone et al. (1989) determined an unranked suite of about ten morphological measures of body and fin sizes—including only one measure of the anal

PAGE 43

25 fin—statistically differentiated effluent-exposed fish from refe rence fish. These variables were proposed as a rapid bioassay to detect effects of pulp mill exposure, but were never fully developed. Howell and Denton (1989) developed five stages of increasing gonopodial development in females exposed to bacterially-degraded phytosterols. More recently, quantitative measures of anal fi n morphology were compared from wild-caught and androgen-exposed females (Angus et al. 200 1, Bradley et al. 2004). The length ratio of ray 4 to 6 was the most sensitive measure of masculinization of the anal fin. While the number of segments along rays 3 and 4 and the width ratio of ray 3 to 4 were also sensitive measures, they were more subject to variability. Controlled exposure to whole efflue nt dilutions produced inconsistent masculinization results. Initia lly in support of field collect ions, static renewal exposure of newborn mosquitofish to water collected 3.6 km downstream from Elevenmile Creek induced elongated anal fins (measured as anal fin length) in females upon maturity (Drysdale and Bortone 1989). While my study research was being conducted, researchers in Canada and New Zealand we re also studying mascul inization of adult female mosquitofish using controlled (mainly static renewal, one flow-through) exposures to 15%, 70% or 100% effluent (McCarthy et al. 2004 summarizes results across separate studies). Out of seven pulp mills and one sewage treatment facility tested, four of the pulp mill effluents and the sewage effluent3 induced masculinization (all static renewal exposures). Among th ese four pulp mill effluents, two induced masculinization relatively quickly (within 3 weeks) while the other two required 24 weeks of exposure. No association between induction and type of mill or concentration 3 In contrast, collection near other sewage discharges found effect on males, not females (Batty and Lim 1999, Angus et al. 2002). Bradley et al. (2004) also found no effect on females

PAGE 44

26 of -sitosterol could be established. Appa rently, duration of exposure required to produce effect varies widely. However, thr ee important caveats exist for these studies. Other than Elevenmile Creek, comparativ e field studies were not conducted to determine if effects existed in wild mo squitofish exposed to these effluents. All but one exposure required holding and transport of effluent back to the exposure system, with unknown consequences to effluent composition. Masculinization was measured qualitativel y, as presence/absence or staged using categories established by Howell and Denton (1989). One of these controlled expos ures detected differences in masculinization due to effluent treatment and filtration (Ellis et al. 2003). Secondary treatment of effluent at environmentally relevant concentrati on reduced gonopodial development by 25%, yet masculinization remained significantly greate r than controls. Filtration of treated effluent, removing organic extractives adso rbed to particulates, also removed the response. Exposure was repeated two years later, after treatment system maintenance was improved as indicated by gradual reduc tion in total suspended solids. Using the more specific ray 4 to 6 length ratio, masculin ization was not induced (van den Huevel et al. 2004b). Since exposure duration remained th e same (3 weeks), it is unknown if the effect was entirely removed or if time to ma nifestation was extended. Regardless, these experiments strongly correlate masculinizat ion with adsorbable organic effluent components, such as low mol ecular weight wood extractives. Precocious maturation Precocious (early) maturation of male mosquitofish exposed to pulp and paper mill effluent has been examined brie fly. Effects were associated with bleached kraft effluent from Elevenmile Creek before major proces s changes (Howell et al. 1980, Drysdale and Bortone 1989), but not with effluent from a thermomechanical/kraft/newsprint mill in

PAGE 45

27 Ontario (McCarthy et al. 2004). Males of this species grow steadily until maturation at which point growth plateaus, as opposed to females who grow steadily throughout life allowing body length to roughly approximate age (Snelson 1989). Howell et al. (1980) reported small males (12-13 mm standard leng th) began to develop gonopodia, and fullydifferentiated gonopodia occurred in male s measuring 13-18 mm standard length. Compared to males in unexposed sites with mature gonopodia at 18 mm or longer, they concluded males developed earlier due to efflue nt exposure. However, neither data nor statistics was provided to support this conclu sion. Static renewal exposure of newborn mosquitofish to water collected 3.6 km dow nstream from Elevenmile Creek provided more compelling, statistical ev idence of precocious matura tion in males (Drysdale and Bortone 1989). Exposed males began anal fin elongation approximately one month before unexposed males, with groups eveni ng out during late-stage gonopodial growth. Surprisingly, this endpoint was virtually i gnored by researchers until recently. In contrast, in the Ontario study, continuous fl ow-through exposure of mosquitofish to 15% and 100% effluent for 21 weeks did not aff ect male gonopodial length relative to body size (McCarthy et al. 2004). Behavior Investigation of male-like reproductive behavior was a logical step once male-like secondary sex characteristics were discovere d in female mosquitofish. As implied previously, changes in secondary sex characteristics could potentially lead to behavioral changes that keep females from copula ting and reproducing. Initial preliminary investigation, lacking statistical comparison, indicated both masculinized females and precocious males displayed more aggressive re productive behavior (Howell et al. 1980). Behavioral evaluation of masculinized fe males in the presence of normal females

PAGE 46

28 supported increased aggressive but not reproductive, behavior (Ellis et al. 2003). Analyzing the ability of a suite of reproducti ve behavioral changes to detect pulp mill effluent exposure, Bortone et al. (1989) determined behavior did not adequately discriminate effluent exposure from unexpos ed groups. In support of this conclusion, Krotzer (1990) performed a t horough reproductive behavioral analysis of mosquitofish females exposed to bacterially-degraded phytos terols. Masculinized females displayed aggressive male behaviors toward nonmas culinized females only in a noncopulatory fashion. Paired with males or other masculin ized females, they behaved normally. Thus, masculinization likely does not impact mosqu itofish populations in a behavioral sense. Reproduction Potential for reproduction does not appear impacted by pulp and paper mill effluent exposure in female mosquitofish. Excep t for histological evaluation of gonads, reproductive parameters were never directly co mpared with measures of masculinization. True to the distinction between masculini zation and actual sex reversal, normal ovaries lacking any testicular tissue were consistently reported in masculinized females (Howell et al. 1980, Hunsinger et al. 1988, Ellis et al. 2003, McCarthy et al. 2004). Fecundity (inferred by brood size or number of eyed em bryos plus mature eggs) relative to body length was depressed in females collected belo w effluent discharge in Elevenmile Creek (Rosa-Molinar and Williams 1984, p. 122). However, the authors state: “Estimated fecundities in re ference to length in the arrh enoid [masculinized] fishes were not found to be similar to th ose found in other studies…although the fecundity of the normal G. a. holbrooki [in the current study] was similar.” In contrast, static renewal exposure to se diments and waters of bleached and unbleached kraft effluent receiving streams for 56 days pr oduced no statistical differences in several measures of fecundity compar ed to reference sites (Felde r et al. 1998, D’Surney et al.

PAGE 47

29 2000). However, variation was high between re ference sites. One reference site was a research station, a well-documented habitat without pollution impact s but obviously very different from the receiving stream (e.g. very low water hardness contributing to increased skeletal abnormalities). Thus the se lection of unexposed sites is important, and should include an upstream site at the minimum and sites belonging to the same watershed. In support of this overall lack of histological and embryological impairment in effluent-exposed females, McCarthy et al. (2 004) did not detect al teration in sex ratios of mosquitofish reared in 100% effluent under laboratory conditions. The assumption must be made for all above data on fecundity and sex ratios that at least a portion of exposed females analyzed were masculinized as well. Mechanism of Action Since mosquitofish research has produced the only specific h ypothesis of causation linking one class of effluent components to potentially adverse effects, several researchers have attempted to isolate potential mechanism(s) of action. In support of the bacterial degradation hypothesis forming andr ogens, Jenkins et al (2001) detected low levels of androstenedione (0.14 nM) in th e Fenholloway River downstream of effluent discharge. However bioassay-based fracti onation studies, discussed previously under effects of pulp mill effluent on fish, have not supported androstenedione and testosterone as active androgens that bind the androgen receptor and masc ulinize females (Jenkins et al. 2001and2003, Parks et al. 2001, Durhan et al. 2002, Ellis et al. 2003, van den Huevel et al. 2004b). Orlando et al. (2002) investigated an altern ative mechanism used in the aquaculture industry, aromatase inhibition. Contrary to their hypothesis, aromatase activity was elevated in both brain and ovari an tissue of females co llected downstream of effluent discharge in the Fenholloway. Ma sculinization was i ndicated by increased

PAGE 48

30 segmentation of anal fin rays, but not increa sed anal fin length. The authors concluded aromatase inhibition was not a likely mechanism to account for masculinization. However, impaired activity could po tentially upregulate enzyme production. Measurement of endogenous steroid levels may provide more insight into this potential mechanism. Mosquitofish as a Model Species Several life history and ecological characte ristics of mosquitofish (the closely related Eastern and Western species ( Gambusia holbrooki and G. affinis ) of the family Poeciliidae), make these species an id eal model for both field and laboratory toxicological studies. Meffe and Snels on (1989) compiled the most comprehensive overview of Poeciliids, from which the fo llowing summary was derived unless noted otherwise. Occurrence and Availability in Effluent-Receiving Systems Mosquitofish are opportunistic, omnivorous feeders that can e xploit diverse foods ranging from planktonic invertebra tes and fish fry to detritus and algae; so food source should not limit their occurrence in effluent -receiving streams. Similarly, mosquitofish inhabit a diverse range of shallow habitats with the ability to occupy “fringe” habitats characterized by environmental extremes. Comb ined with their tolerance to high salinity (up to 50% seawater in the Western mosquitofish, G. affinis ), broad thermal range, and tolerance to low dissolved oxygen (mosquitofi sh gulp air at the su rface in response to hypoxia), mosquitofish should tolerate water qua lity of effluent-receiving streams. Their home range is small (several meters) maki ng chronic exposure likely. In addition, they readily colonize new populations via migra tion of a single gravid female, and have

PAGE 49

31 become ubiquitous around the world caused by deliberate introducti ons in attempt to control mosquitoes and associat ed mosquito-borne illnesses. In addition to their suitability as a field model, mosquitofish can be maintained under laboratory conditi ons with ease relative to other livebearing fish species. Also compared to other livebearers, much more is known about mosquitofish reproduction. Reproductive Characteristics As members of the livebearing fish family, Poeciliidae, mosquitofish develop eggs internally and appear to give bi rth to fry (ovoviviparity). This is in stark contrast to most fish that lay eggs (oviparity). Males inse minate females with packets of sperm called spermatozeugmata, and females can store sperm in ovarian folds and gonoduct for up to eight months/broods. Fertilization and embryol ogical development occur directly in the ovarian follicle, and ovulation is immediat ely proceeded by parturition. Estrogen (as opposed to prostaglandins in other fish species) stimulates postovulatory sexual receptivity of females. Livebearers exhibit a spectrum of matern al-embryo nutrient exchange, from more fish or reptilian like yolk loading before fertilization (lecithotrophy) to more mammalian like continuous provisioning throughout embryological development by the mother (matrotrophy). Mosquitofish represent the fo rmer group. In addition, mosquitofish carry one fertilized brood at a time, as opposed to many other livebearers that harbor several broods at different stages of development (superfetation). Brood size is dependent on female body size, with larger fish pr oducing larger broods. Reproduction is asynchronous and seasonal in temperate to subt ropical climates such as Florida. The reproductively active period is during sp ring and summer months followed by reproductive senescence in the fall and winter. Temperatur e and photoperiod are

PAGE 50

32 considered dominant environmental cues c ontrolling reproductive season in both sexes (Koya and Kamiya 2000, Koya and Iwase 2004). Mosquitofish begin life as hermaphrodite s, containing both ovar ian and testicular tissue (Koya et al. 2003). Hermaphroditisim (T eh et al. 2000) has also been detected in adult farm-reared albino mosqu itofish. Within 10 days, gonads differentiate into paired, fused testes or ovaries. Gonadal maturity is reached in approximately 3 months, with males maturing 2-3 weeks earlier than females. Mosquitofish, similar to many other livebea rers, are sexually dimorphic. Male and female mosquitofish have several gender-speci fic traits: females are larger and possess an anal/gravid spot and urogenital papilla, while males are smaller and possess a gonopodium. The gonopodium facilitates internal fertilization and is formed by the elongation of rays 3, 4, and 5 of the anal fi n. Formation of the gonopodium is controlled by androgens, and a fully-developed gonopodium (marked by terminal differentiations on the tips of rays 3, 4, and 5) signifies comp lete maturity (Turner 1941b). Turner (1941a, 1942a, 1942b) also documented formation of the fully mature gonopodium in females exposed to androgens such as ethynyl a nd methyl testosterone. Thus, female mosquitofish could be a useful model of exposure to environmental androgens. Mosquitofish as a Bioindicator of Pulp and Paper Mill Effluent Mosquitofish have been repeatedly propos ed as an indicator of environmental disturbance by pulp and paper mill effluents (Davis and Bortone 1992, Bortone and Davis 1994, Cody and Bortone 1997). At the state level, the Florida Department of Environmental Protection has explored this possibility (T.S. Gro ss pers. comm.) without implementation. Federally, the US EPA is de veloping the mosquitofish as an androgenic model of endocrine disruption (Angus et al. 1997). While mosquitofish have potential

PAGE 51

33 for use in regulatory testing and screening of pulp mill effluents, they have not been adequately assessed as an indicator species. Definitions: Bioindicator and Biomarker For every critical review of bioindicators and biomarkers there exists a slightly different definition. The term biomarker is more consistently defined as a measurable biological response to environmental polluti on observed below the organism level of biological organization (Foster et al. 1992, Peakall 1992, Jamil 2001). Biomarkers encompass changes at the molecular, biochemical, physiological, histological, morphological, or behavioral levels. The major premise for use of biomarkers in environmental regulation is the bridge they form between chemical exposure and adverse effect. However, biomarkers are usually clas sified as more indicative of either exposure or effect. Current challenges for biomarker us e include questions of natural variability, use in the field, and extrapol ation to higher levels of bi ological organization and to humans. Though the challenges appear formid able, researchers strive to meet these important demands of regulator y application (e.g. analysis of suites of biomarkers and increasing inclusion of biomarker analyses in population and co mmunity studies). Unfortunately, as a society we have little patience for the pace of science and often biomarkers are misjudged or overinterpreted. In contrast to biomarkers, the term bi oindicator usually implies changes at the organism level or above, includi ng individual reproduc tion, populations and communities. For example, US EPA (2004, website) states the following: Environmental scientists have determ ined that the presence, condition, and numbers of the types of fish, insects, algae, and plants can provide accurate information about the health of a specific river, stream, lake, wetland, or estuary. These types of plants and animal s are called biological indicators.

PAGE 52

34 Ecological bioindicators have been developed and used extensively by the US EPA to assess ecosystem health. Jamil (2001, p. 4) defines bioindicators for evaluating ecological health “as a species or groups of species (plants or animals) that, by their presence and/or abundance, play an importa nt role in the ecosystem to which they belong.” Further, he disti nguishes two classes of bioi ndicators used to evaluate environmental quality: bioaccumulator and sent inel species. Similar to separation of biomarkers into exposure and effect groups, bioaccumulators repres ent organisms that bioconcentrate toxicants from the surrounding media and biomagnify up the food chain, while sentinels indicate toxic effect that allows judgment of effects on human and/or environmental health. Sentinel species are further characterized as species with field application, either preexisting at sites of interest or capable of in situ exposure (such as caging onsite). Ideally, extens ive knowledge exists about no rmal states measured in sentinel species. A final condition is the surroga te nature for species at risk, i.e., sentinels substitute for sampling of already imperiled sp ecies that are difficult to study directly and could potentially be harmed by intensive research. For the purposes of my study, bioindicat or is defined as a species possessing measurable changes in biomarkers that co rrespond to impacts at higher levels of biological organization. Similar to applicab ility requirements for sentinel species, a bioindicator should be a versatil e subject in the field, not only in the laboratory. Thus the bioindicator retains environmental relevance while affording analysis of time and dosedependent responses.

PAGE 53

35 Bioindicator Criteria for Success Criteria, like definitions, abound for ra ting success or failure of potential bioindicator organisms (Peakall 1992, Jamil 2001) Regardless of variable definitions, several common criteria for success exist and overlap with criteria for biomarkers. Practicality Sufficient scientific knowledge of model species under normal, unexposed conditions. Ease of training and use by personnel. Costand labor-effective. Availability of model species fo r both field and laboratory study. Variability Intrinsic or natural variability of biom arkers such as seasonality and gender differences. Exposure variability esp ecially sensitivity and tolerance/acclimatization. Method variability such as observe r bias and instrumentation bias. Predictability Extrapolation to organism level or highe r adverse effects, i.e., reproductive or population impacts. Extrapolation to other species living in exposure conditions. Extrapolation to humans. Applying these criteria to th e existing research on mosquitofish exposed to pulp and paper mill effluents reveals practicality, but not variability and predictability, has been adequately addressed. Practicality was supported in the previous section about mosquitofish as a model species. The mosquito fish is one of the most intensely studied livebearing species because of its use in mos quito control. Mascul inization studies are not expensive, especially compared to molecular and biochemical research.

PAGE 54

36 Measurements require basic lab skills us ing dissecting scopes a nd ideally computer measurement software. The most expensive as pect for these studies is exposure facility construction, a common expense for any bioind icator. Similarly, exposures are the most labor-intensive aspect of masculinization studies. Perhaps the greatest strength for mosquitofish as a bioindicator is its globa l distribution (again because of mosquito control), making collection of large numbers in effluent-exposed and reference sites very easy. In terms of variability, natural fluctuati on has been indicated by season (in Florida) and the response is gender sp ecific, although precocious ma turation remains questionable in Florida streams. Exposure variability has been implied by the work in Canada and New Zealand, among mills and within m ills with improving technologies and maintenance of systems. Research in Florid a has not directly addr essed this type of variability, but a consistent re sponse to variable exposure is implicated. Since specific bioactive agents and mechanism of action have no t been isolated, sensitivity is difficult to address and can only be viewed from whole e ffluent exposures. Co mpared to binding of fish androgen receptor (Ellis et al. 2003), masculinization is a less sensitive, but potentially more relevant res ponse. Specificity of mascu linization for pulp and paper mill effluent versus sewage effluent appears high, with the abnormal exception to treated sewage effluent by McCarthy et al. (2004). Yet laboratory ex posures to other chemicals (such as the hypertensive drug spironolactone and the agricultu ral insecticide endosulfan) have also induced masculini zation (Howell et al. 1994, Park et. al 2004). In addition, Bradley et al. (2004) found masculinized female s living in the retention pond of an urban parking lot; therefore, nonpoint sources of pollution may confound results. So

PAGE 55

37 specificity, once considered high for pulp mill effluents, has become questionable. At the same time, exploration of how these alte rnative compounds masculinize may catalyze isolation of bioactive compounds Tolerance and acclimatization have not been studied. A handful of anecdotal reports (Davis and Bortone 1992, Bortone and Davis 1994, Cody and Bortone 1997) indicate masculinization is reversible when females are transferred to clean water; McCarthy et al (2004) did not observe reso rption of anal fin elongation under controlled exposure. Finally, method vari ability, other than determination of the more sensitive measures of anal fi n morphology, has not been addressed. The final major criterion for a successful bioindicator, predictability, has begun to be examined with studies of reproductive po tential. As previously discussed, gonad condition, fecundity, and sex ratios do not appe ar affected by effluent exposure. A major drawback to most of this wo rk is the lack of masculini zation measures. Actual fry production, quality and survival have not been examined either as a more accurate measure of reproductive success. Regardi ng extrapolation to the fish community, Bortone and Cody (1999) attempted to examine masculinization of other livebearing fish species in their Rice Creek collection, but failed to obtain adequa te fish numbers for analysis. The link to humans has yet to be determined as well. Obviously, further testing of mosquitofish is required to determine if this species would be a successful bioindicator of pulp and paper mill effluents. Practically speaking, mosquitofish are very promising, especially on a worldwide scale bu t perhaps less useful in nations with advanced processing technology. Contribution of My Study My study addresses two of th e three major bioindicator criteria for mosquitofish: variability and predicta bility. Industry and regulatory ag encies alike will thus have a

PAGE 56

38 better understanding of the potential for mosqu itofish as a bioindicator of pulp and paper mill effluent exposure. In addition to anal fin morphology as a biomarker, sex steroids were investigated for two reasons: 1) mascu linization as an androgenic model innately assumes alteration of steroid levels, either peripherally or systemi cally; 2) sex steroid levels are generally depressed in other fish species exposed to pulp and paper mill effluents. Variability by season, method, and exposure were addresse d directly. Method and seasonal variability, while not a stated objective of th e original research proposal, were necessary precursors in the developmen t of techniques and thus were included in my study results. Exposure variability studies focused upon the three Florida freshwater systems (Elevenmile Creek, Fenholloway River, and Rice Creek) for which masculinization has been reported and cons idered equivalent. Predictability was addressed by evaluating the relationship among biomarkers and reproductive success. Preliminary examination of population st ructure was also conducted during the reproductive experiments. Explicitly stated objectives for my study we re divided into two specific aims with associated hypotheses. Within each specifi c aim, three sub-aims were identified and studies developed for each. Specific Aim 1 Our first specific aim was to determine the effects of improved mill technology on masculinization of female mosquitofish. We hypothesized that redu ction in brown side effluent components (i.e., wood extractives su ch as phytosterols a nd resin acids) would reduce anal fin elongation and hormonal alteration in female mosquitofish. Aim 1A: Assess induction of ma sculinization in female mosquitofish under shortterm controlled exposure to effluent at one mill throughout process changes and at two mills using different processing techniques. Expected outcomes were:

PAGE 57

39 induction would be rapid; degree of response would reflect differences in concentration of wood extractives; and e ffect would be reduced following major process changes. Aim 1B: Compare and contrast an al fin morphology and sex hormone concentrations in female mosquitofish at one mill throughout process changes. The expected outcome was that process change s would reduce masculinization in wild fish to a similar degree as induction studies. Aim 1C: Compare and contra st anal fin morphology and sex hormones in female mosquitofish collected from systems exposed to different types of mill effluent. The expected outcome was degree of re sponse would reflect differences in concentration of wood extrac tives, similar to differences expected with induction studies. Specific Aim 2 Our second specific aim was to evalua te reproductive success of mosquitofish exposed to pulp and paper mill effluents. We hypothesized that exposure to pulp and paper mill effluents would not impair reproductive success of mosquitofish. Aim 2A: Determine population structure of mosquitofish living in Fenholloway River. The expected outcome was that population structures would reflect sex ratios and recruitment (juvenile s) found in unexposed references Aim 2B: Characterize organism level res ponses to effluent exposure associated with reproduction. The expected outcom e was that adult females would display normal reproductive status regardless of exposure and masculinization Aim 2C: Quantify offspring production of fe male mosquitofish exposed to pulp and paper mill effluents The expected outcome was that offspring production would not vary regardless of exposure or masculinization Studies to address these aims were designed with certain limitations and assumptions in mind. Home ranges do not overlap between sites sampled. However, the above implies field studies may be comparing genetically different populations of mosquitofish. Mosquitofish caught at field sites have lived there throughout life (birth, preand post-maturation).

PAGE 58

40 Mosquitofish ( G. affinis and holbrooki ) respond in the same manner. Masculinization of the anal fin is fin ite and permanent (as indicated by Turner 1942b), i.e., gonopodial development does not vary between acute and lifetime exposures. Fry production studies use wild-caught fi sh therefore reproductive history (i.e., fertilization, past broods) is not controlled and is unknown. Cannot expose to whole effluent by tran sfer back to the laboratory without potentially altering its composition. Composition of effluent changes drastica lly when mills cycle through different types of trees. Potential (as opposed to actual) exposure is sufficient to document toxicant exposure, since the masculinization hypothesis predicts actual expos ure is to as yet unknown androgens and not pare nt effluent components. Specific changes in mill technology are not always known or available to investigators. Table 1-1. Select characterist ics of the mills in my study according to receiving stream. Mill Characteristic Elevenmile Creek Fenholloway River Rice Creek Furnish 75% hardwood, 25% softwood 100% softwood 50% hardwood, 50% softwood Pulping Chemical/kraft Chemical/ dissolving kraft Chemical/kraft Bleaching ECFa (1995) Sodium hypochloriteb ECFa (May 2001) Product White copy paper, return postcards, market (paper) pulp High grain cellulose (dissolving nonpaper pulp) Paper towels, tissue paper, kraft bag, linerboard Effluent treatment Aeration with microbial degradation, chemical flocculation, oxygen delignification Aeration with microbial degradation Aeration with microbial degradation, activated sludge Effluent volume 21-26 mgdc 43 mgd 28 mgd a Elemental chlorine free; conversion date in parentheses bNote this mill was not subject to many Cluster Rule requirements, since it is not a papergrade bleached kraft or sulfite mill cmillion gallons per day

PAGE 59

41 Figure 1-1. Categories of pulp and paper mill f acilities. Mills associated with my study are abbreviated by receiving stream: EM = Elevenmile Creek, FH = Fenholloway River, RC = Rice Creek. dissolving specialty market pulp facilities nonintegrated facilities integrated facilities market pulp paper/paperboard paper pulp nonpaper pulp fluff newsprint linerboard writing/ copy tissues paper towels packaging/ wrapping corrugated EM EM EM EM EM FH FH FH FH RC RC RC RC RC RC

PAGE 60

42 CHAPTER 2 VALIDATION OF MOSQUITOFISH ENDPOINTS USED TO ASSESS EFFECTS OF PULP AND PAPER MILL EFFLUENT EXPOSURE Mosquitofish have been proposed as a bioi ndicator of pulp and paper mill effluents, although several aspects concerning variability in masculinization responses have not been addressed. Eastern mosqu itofish were collected in summ er and winter months from Rice Creek, the receiving stream for effluent from the Georgia-Pacific bleached/unbleached kraft mill in Palatka, FL. In this study, a series of validations were performed for morphological measurements and for a new biomarker in this species, whole body sex steroid concentrations. Seas onality in responses was also addressed. Gender identification using th e urogenital papilla was succes sfully validated against internal examination of gonads Morphological measurements validated adequately, and computer-aided measurement was a preferred alternative to manual measurement. Sex steroids also validated adequately consid ering the unavoidable limitations using whole body analysis. Greater masculinization was a ssociated with effluent-exposed sites for females, while males were not overtly affect ed. Females displayed seasonal effects on sex steroids, especially ster oid ratios, but not anal fi n morphology. Surprisingly, sex steroids were not altered in females collected from 100% final effluent before discharge, indicating a complex interplay of environm ental factor(s) may produce responses as opposed to effluent alone, such as different ial bacterial degradation of phytosterols.

PAGE 61

43 Introduction Concerns over release of chemicals into th e environment have shifted from lethal to sublethal effects on nontarget wildlife sp ecies (Lehtinen 2004). Reported sublethal effects of pulp and paper mill effluents on fish include induction of liver detoxification systems, alterations in sex steroids con centrations and production/metabolism, reduced gonadal development, decreased egg production and decreased fry survival (Van Der Kraak et al. 1992, Gagnon et al. 1994a, Munkittrick et al. 1999, NCASI 2000a, Sepulveda et al. 2003, Parrott et al. 2004, McMast er et al. 2003). Degree of these effects is often mill-specific with some effluents pr oducing no effect at all (Kovacs et al. 1995a, McCarthy et al. 2004). No clear pattern existed between eff ect and type or quality of effluent, other than reduced responses with improved mill technologies (Munkittrick et al. 1997, van den Huevel et al. 2004b). Fu rther complicating th e matter, bioactive effluent components have not been strictly identified yet and most likely different compounds and conditions influence response pathways. Several responses imply androgen-induced m echanisms in fish. For example, pulp mill effluents have been associated with male-biased sex ratios and development of male-like secondary sex characteristics in females or masculinization (Kovacs et al. 1995b, Larsson et al. 2000, Larsson and Frlin 2002, Larsson et al. 2002, Parrott and Wood 2002, Parrott et al 2003). The mascu linization response has been frequently reported in female mosquitofish as an el ongation of the anal fin into a male-like gonopodium, the copulatory organ in this live bearing species (Howell et al. 1980, Cody and Bortone 1997, Bortone and Cody 1999, Parks et al. 2001, Ellis et al. 2003, Bradley et al. 2004, McCarthy et al. 2004). Further, this species has been repeatedly proposed as a bioindicator of pulp and paper mill efflue nts (Davis and Bortone 1992, Bortone and

PAGE 62

44 Davis 1994, Cody and Bortone 1997). In a pract ical sense mosquitofish are a suitable model, yet several aspects require more thorough evaluation to warrant this broad application as a bioindicator of pulp and paper mill effluents. In collaboration with the National Co uncil for Air and Stream Improvement (NCASI), Southeastern Aquatic Biology Pr ogram, New Bern, NC, techniques to assess masculinization were validated and refine d. Method validation involved a new gender identification technique, anal fin morphol ogical measurements, and development of a novel assay for measurement of whole body se x steroids in mosquitofish. NCASI’s validation results have been reported in the proceedings of the 5th International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluents (Bradley et al. 2004). Materials and Methods Mill Characteristics and Field Collection Adult mosquitofish were collected along shallow vegetated banks in Rice Creek, the receiving stream for effluent discharg e from Georgia-Pacifi c’s Palatka, FL, USA operation ( Figure 2-1 ). Collections occurred in both summer (March and June 2000, July 2001: n = 174, n = 141, n = 368 respectively) and winter (November 2000, n = 899), corresponding to reproductive and nonreproductive periods respectively. All fish were measured for standard length (+ 0.01 mm) using digital calipers and weighed (+ 0.001 g) using a digital scale before preservation. Chap ter 3 gives details about the mill, and field collection techniques (site descriptions a nd latitude/longitude can be found in Appendix A).

PAGE 63

45 Gender Identification Using the Urogenital Papilla For all collections, gender was determ ined by external examination of the urogenital sinus. Most masculinization studies have used either the gonopodium (typically male-specific) and/or anal spot (female-specific) as external indicators of sex. Gender identification should not be based upon the gonopodium, since it is the primary measure of masculinization. This laboratory found the anal spot a difficult indicator as well, since many female fish have a partial an al spot and some females, mostly collected from effluent-exposed sites, have very light, brownish anal spots.4 Hence, a new gender identification technique other than examination of gonads, which is very labor-intensive, was developed. Females have a gender-specifi c urogenital papilla that protrudes from the urogenital sinus, and the urogenital opening is located on the tip of the papilla (Meffe and Snelson 1989). During copulation, males ho ld on to the papilla using terminal differentiations of the gonopodium. Gender identification via the urogenital pa pilla was validated using the winter 2000 and summer 2001 collections on several leve ls. First, variation among personnel was evaluated by having three tech nicians determine sex using th is new technique (n = 200 fish). Second, the new method was verified ag ainst gross internal examination of gonads by NCASI (n = 200 fish). Third, the new method was verified against histological identification of gonads within USGS usi ng a separate group of fish (n = 354). Anal Fin Morphology Several measures of anal fin elongation ha ve been employed by researchers, from qualitative scores of presence/absenc e and categorizing degree of gonopodial 4 Investigation did not reveal any dead embryos or ab normal ovaries within the body cavity, as initially suspected (pers. obs.). Cause(s) of brown anal spot remains unclear.

PAGE 64

46 development to quantitative measures such as total anal fin length, gonopodium/extension length, number of terminal differentiations, length ratio of Rays 4 and 6, thickness ratio of Rays 3 and 4, and number of segments on Ray 3. Measurement of anal fin and gonopodium/extension length has been shown to be dependent on body size (standard length) and thus must be account ed for during statistica l analysis (Bortone and Cody 1999, Bradley et al. 2004). For this reason, several studies restricted analysis of anal fin morphology to specific size classe s. However, length ratio of Rays 4 and 6 and width ratio of Rays 3 and 4 are independe nt of body size and such a restriction is not necessary (Angus et al. 2001). Ray 3 segment counts, surprisingly, di d not correlate with either of these ratios and was not suggested as an accurate measure of masculinization. Bradley et al. (2004) distinguished length ratio of Rays 4 and 6 as more sensitive than either width ratio of Rays 3 a nd 4 or segment number of Ray 3. Length ratio of Rays 4 to 6 was used to assess masculinization in our study. Body weight (+ 0.001 g) and standard length (+ 0.01 mm) were also measured before preservation using a digital scal e and a pair of digital calipers for all fish. For the winter 2000 collection, potential variati on in both sexes was addresse d due to preservation state (fresh versus fixed) and observer bias with in and among laboratories (n = 200 fish for each comparison). Manual measurements of the linear distance from base to tip of Rays 4 and 6 of the anal fin (+ 0.1 mm) were made using a di ssecting scope with ocular micrometer, before and after preservation in 10% neutral-buffered formalin. Preserved fish were additionally measured independently by two other technicians using the same equipment, and then shipped to NCASI for measurement. Influence of size class for females was also investigated in this colle ction, dividing females into four 5 mm groups

PAGE 65

47 (20–24.99 mm, 25–29.99 mm, 30–34.99 mm, 35–39.99 mm). The summer and winter 2000 collections were compared to address seasonal variation for both sexes (summer 2000 fish were measured after preservati on). Finally, the summer 2001 collection was used to test observer bias within the la boratory (n = 200 fish) using newly developed computer-aided measurements (+ 0.01 mm, SigmaScan Pro 5.0, SPSS, Inc.) of digital images taken of fish before preservation. Bradley et al. (2004) evaluated the use of computer-aided versus manual measurements and determined although they were comparable, computer-aided measurements were more accurate and useful for archiving data. Sex Steroids In addition to anal fin morphology as a bi omarker of masculinization, sex steroids were investigated as a second, physiological biomarker. Circulating sex steroids are generally depressed in othe r fish species exposed to pulp and paper mill effluents (McMaster et al. 2003, Sepulveda et al. 2001). Masculinization as an androgenic model innately assumes more specific alteration of steroid levels, either peripherally or systemically in favor of androgens. Orla ndo et al. (2002) hypothe sized inhibition of aromatase resulted in a masculinized hormone profile. Although they te sted the first half of this proposed mechanism and found aromat ase activity actually elevated, the second half, sex steroid profile, was not measured. Ex amination of sex steroids in mosquitofish would potentially shed light on hypothesized mechanisms. Primary sex steroids were analyzed using a modified radioimmunoassay (RIA) method originally developed for seru m and plasma samples of common carp, Cyprinus carpio (Goodbred et al 1997), and since adapted for use in a variety of other aquatic species and tissue media such as plasma of largemouth bass, Micropterus salmoides

PAGE 66

48 (Gross et al. 2001) and mantle of freshwat er invertebrates (Gross et al. 2000). Mosquitofish were initially analyzed for 17 -estradiol, 11-ketotestosterone and testosterone, considered the most active repr oductive hormones in fish. However, Borg (1994) demonstrated testosterone, and not 11-ketostosterone, as the only dominant androgenic hormone in poeciliid fishes. Anal ysis of 459 fish (both sexes) revealed nondetectable levels of 11-ketotestosterone confirming Borg’s finding. Therefore, testosterone is assumed to be the dominant androgen in mosquitofish and analysis of 11-ketotestosterone was discontinued. Ten fish of each sex from each site were analyzed for sex steroids in the summer 2000 collec tion; 42–50 females and 16–28 males per site were analyzed in the winter 2000 collection. RIA steroid analysis involves chemical digestion of an entire fish, followed by extraction, radiolabeling, and a competitive bi nding assay to quantify steroid levels (see Appendix B for laboratory prot ocols). Whole body chemical digestion is accomplished by boiling individual fish in pot assium hydroxide (30% w/v) at a volume three times the individual fish weight. Fi fty microliters of resultant homogenate are removed in duplicate for extraction. Diethyl ether added in excess (4 mL) is used to extract lipophilic compounds, including sex steroids, from th e digestion homogenate. Extraction and evaporation is performed twi ce to increase extraction effici ency. A reaction solution is prepared composed of evaporated extr act, tritiated hormone and a corresponding hormone-specific antibody. This solution incubates overnight to allow unlabeled hormone from the extract sample (at unknown concentration) and radiolabeled hormone (at a known concentration) to compete for an tibody binding sites. After incubation, the reaction solution is centrifuge d with charcoal dextran to remove any hormone not bound

PAGE 67

49 to antibody. Radioactivity is measured usi ng scintillation spectropho tometry. Standard curves (at 1, 5, 10, 25, 50, 100, 250, 500, and 1,000 pg hormone) are generated for each hormone using known concentrations of ra dioinert hormone in buffer. Steroid concentration in samples is then calculated by aligning values of an inhibition curve, generated from the competitive displacement of radiolabeled hormone in the sample, to the standard curve. Validation and char acterization of this procedure entailed determination of digestion a nd extraction efficiencies by re productive status and exposure (n = 5 each group for each hormone: males, gr avid and nongravid females, juveniles, and effluent-exposed males and females); minimu m detection limits on the standard curve; cross-reactivities of an tiserum with other steroids; and interand intra-assay variation. Statistics Body weight and standard length were used to calculate condition factor, K = weight / length3 x 100, as an indication of overal l health used by the aquaculture industry (values at least 1 are c onsidered healthy). The length ratio of anal fin Rays 4 and 6 was calculated as an index of anal fin elongation. Estrogen and testosterone concentrations were used to calculate a ratio indicating masculine hormone profile (E:T<1) or feminine hormone profile (E:T>1). Gender identification and anal fin mo rphological measurements used for validations were analyzed usi ng Pearson’s product moment co rrelations or calculation of coefficients of variance. An al fin morphology and sex steroid data were analyzed within sex using two-way analysis of covariance ( ANCOVA) to test for si gnificant variation by site and season for the summer and winter 2000 da ta, or by site and size class within the winter 2000 female anal fin da ta only. Size class was also analyzed by one-way ANOVA within site. Any data failing tests for normality and homogeneity of variance were

PAGE 68

50 transformed using log transformations. Angus et al. (2001) determined length ratio of Rays 4 to 6 is appropriately analyzed using parametric statistics after log transformation of the ratio data. Significant differe nces in the ANCOVA and ANOVA tests were analyzed for multiple comparisons using Tukey’ s HSD. Within site, differences between seasons were analyzed by t-test. St atistical significance was attained at <0.05 for all tests. All statistical analyses were conducte d using SAS version 9.0. Results and Discussion Water Quality As expected, conductivity, salin ity, turbidity and pH were elevated at effluentexposed sites compared to unexposed sites ( Table 2-1 ). Water temperature was highest at the reference site and predischarge pond, and increased along the length of Rice Creek. Reference and upstream sites were comparable, other than water temperature. Dissolved oxygen remained high enough to support fish at all sites (>4 mg/L). Validation of Gender Identificati on Using the Urogenital Papilla Agreement about gender was significant at al l three levels for which it was tested: among personnel within the USGS laboratory (r2 = 0.899); between USGS and NCASI laboratories (r2 = 0.93); and against hi stological evaluation (r2 = 0.99). Between laboratories, females were agreed upon slightly more than males and the error rate for the new technique was 3.5%. Disagreement occurred for 15 fish (7.5% of all fish examined, Figure 2-2 ): of these fish, half (7) were incorre ctly identified by the urogenital papilla, whereas the other half (8) could not be accura tely identified by inspection of gonads. Incorrectly identified fish were equally from both exposed [DIS] and unexposed [REF2, U(8)] sites, precluding bias ag ainst masculinized fish. Co mparison against histological identification of gonads was even more accura te with an error rate less than 1%.

PAGE 69

51 Disagreement occurred for 7 fish (2% of all fish examined Figure 2-3 ): of these fish, over half (5) could not be accurately identified by histological inspection, while the remaining two fish were incorrectly identified by th e urogenital papilla. Th erefore the urogenital papilla is a reliable indicator of internal sex. This new noninvasive external gender identification technique could be very us eful for a multitude of studies utilizing mosquitofish repeatedly sampled over time. Morphology Morphological measurements validated well, with computer-aided measurements preferable to manual measurements. Body size in winter 2000 was not impacted by effluent exposure relative to site; both sexes were in good general health judged by condition factor higher than 1. Increased gonopodial length was w eakly indicated for males at effluent-exposed sites. Precocious maturation in males was not apparent, since smallest males did not have fully developed gonopodia to afford statis tical analysis. In females, analysis by size class did not reveal specific size class(es) associated with anal fin elongation, although caution was implied for using appropriate sample sizes by females from the outfall site. Female anal fin elongation was significantly elevated at the discharge and first downstream site in both fa ll and winter. Seasonality was not evident for anal fin elongation in either sex. Validations Measurements on fish before and after pres ervation in formalin were significantly correlated ( Table 2-2 ), although measurements on formalin-preserved fish were consistently smaller than on fresh fish (data not shown). This bias was expected since preservation tends to dehydr ate and shrink specimens. Ob server bias between NCASI and USGS laboratories measured differences not only between two observers, but also

PAGE 70

52 differences between equipment. All meas urements were significantly correlated ( Table 2-2 ), although correlation coefficient for Ray 4 indicates somewhat inconsistently larger measurements by USGS. NCASI reports data to the nearest 0.01 mm for manual anal fin measurements (Bradley et al. 2004), while USGS reports data to the nearest 0.1 mm. Therefore, instrument bias is indicated. Si gnificant differences occurred between manual and computer-aided measurements ( Table 2-2 ). This discrepancy is likely due to the more accurate measurement by computer ( + 0.01 mm), similar to the nonsignificant differences between laboratories. Manual and computer-aided measurements of fish are thus not comparable when ex amining data across studies. Variation within one observer and among se veral observers was also addressed for both types of measurements. Coefficients of variation at 10% or less were considered acceptable. Repeated manual measurements by one observer were consistent, as were measurements among observers ( Table 2-3 ). Repeated computer-aided measurements by one observer were even more consistent than for manual measurements, but this was not quite the case among observers. This result means all computer-aided measurements should be (and were) made by the same observe r for masculinization studies. With the greater accuracy afforded by computer-aided measurements (the Ray can be traced along exact curvatures, as opposed to linear di stances with manual measurements), this technique is preferable. As Bradley et al. (2004) al so note, computer-aided measurements are ideal for archiving data, but are time-intensive with the extra step of photography involved. Body Size for Fall 2000 Collection Body size data is presented for winter 2000 in this chapter ( Table 2-4 ), while summer 2000 and 2001 data is presented in Chapters 3 and 4, respectively. Overall

PAGE 71

53 males were not impacted by exposure site. Weight decreased with increasing distance from outfall (i.e. decreasing exposure), and th e longest males were found at the upstream site even compared to the reference site [REF1]. Condition fa ctor, above one and indicating healthy males, significantly varied inversely with exposure similar to weight: highest for all sites at the out fall [DIS]; lower than outfall and unexposed sites but higher than furthest downstream site at the first dow nstream site [D(1)]; and lowest for all sites at the furthest downstream sites [D(3+6)]. However this index is not appropriately interpreted beyond the benchmark of above or below a value of one. Female body size was not affected by exposure site. The only significant difference was the presence of longest females in the lower half of Rice Cr eek. Condition factor was also above one for females at all sites, indicating adequate overall health. Influence of Body Size on Anal Fin Morphology Since body size (length) has been associated with anal fin length in masculinized females and precociously matured males, fish were divided into 5 mm size classes and analyzed as a covariate with si te. Size classes were also anal yzed within each site. When males were divided into size classes, th e majority fell into the 20–24.99 mm class, a minority fell into the 25–29.99 mm, and less th an 10 fell into the <20 mm class. This paucity of mature males in the smallest si ze class precluded ev aluation for precocious maturation. Precocious maturation may not be occurring in these males during the winter, since small males with fully differentiated gonopodia were not found in significant abundance. Precocious matura tion is addressed in the summer 2001 collections presented in Chapter 4. Standard length significantly correlated with total anal fin length (linear distance of Ray 4) for female mosquitofish in winter 2000 (r2 = 0.70, data not shown). However,

PAGE 72

54 index of anal fin elongation (Ray 4 to Ray 6 length) was statistic ally independent of standard length (r2 = 0.222, data not shown). Overall across size classes ( Figure 2-4 ), the first two exposure sites [DIS and D(1)] were significantly longer by th e index of anal fin elongation (Ray 4 to Ray 6 length) compared to nonexposed sites and the lower half of Rice Creek [REF1, U(8), D(3+6)] Size class di d not significantly covary with site. For size classes within each site, the only statistically significant difference was at the outfall [DIS]: females in the middle two size clas ses had significantly l onger anal fin elongation than the largest size class ( Figure 2-3 ). This result negates predicted responses of size classes due to drought: in light of the dr ought faced by females in 1999 and early 2000 (Appendix A), the older/larger females shoul d have greater elongation. Instead, this result may imply anal fin elongation is induced at a sensitive life stag e and/or represents a dynamic exposure to bioactive compounds. (Dynamic exposure refers to variable concentrations of effluent components over time, dependent on factors such as tree species for furnish, within plant processi ng spills, rainfall/dil ution, and bacterial degradation.) Without specific exposure data for these fish, these conclusions remain speculative. Seasonality Female anal fin elongation at the first tw o exposed sites was significantly elevated compared to unexposed sites for both seasons ( Figure 2-5A ). Elongations were more similar to a developing male gonopodium in le ngth and lacked termin al differentiations (data not shown, see Chapter 3 for photographs of anal fins of females collected in summer 2000). Site and season did not covary for any of these data. Season alone influenced the latter two downstream sites [D (1) and D(3+6)], with opposite trends: at the first downstream site [D(1)] elongation was larg er in winter than summer and vice versa

PAGE 73

55 for the lower half of Rice Creek [D(3+6)]. Therefore, season was not consistently influencing presence of anal fin elongation in females from Rice Creek. This result is in contrast to the seasonality study from El evenmile Creek (Cody and Bortone 1997), where winter months demonstrated a reduction in elongation compared to summer months. Granted, the Elevenmile Creek study was c onducted over the entire year, while the current study compares one month from each season. Variation within a season was not reported for Elevenmile Creek by Cody and Bortone (1997). Both reports agree the response is present regardless of season. Increased anal fin elongation, or greater gonopodial length, was not as apparent for males as for females ( Figure 2-5B ). The summer collection re vealed males with shorter gonopodia at the upstream site [U(8)] compared to the rest. The winter collection demonstrated significantly longe r gonopodia at the first downstream site compared to the reference [REF2] but not the upstream site [U(8)], and longer gonopodia further downstream [D(1) and D(3+6)] compared to both the upstream and reference sites. Hence selection of unexposed site(s) is of vital importance for in terpreting results. Season had no effect on gonopodial length as eith er a covariate with site or alone. Overall the evidence is weak for increased gonopodial length in males due to pulp mill effluent exposure. Sex Steroids Radioimmunoassay of primary sex steroids was validated for whole mosquitofish. To date, only one other study has reported this biomarker in mosquitofish for males only (Toft et al. 2003). Elevated testosterone was associated w ith instream effluent-exposed sites in females regardless of season. Howe ver seasonality, both alone and relation to site, was detected. Estrogen to testosterone ratios were masc ulinized (i.e., greater than 1)

PAGE 74

56 at impacted sites for the summer collec tion only. Importan tly, neither hormone concentrations nor their ratio were altered in 100% final effluent before discharge, indicating additional environmental factors interact to produce the response. For males, seasonality of sex steroids was indicated, and effluent exposure did not appear to alter concentrations and ratio. Validations Validations were completed for 17 -estradiol and testosterone. Since 11-ketotestosterone was not detected in thes e fish samples, only partial validation was possible and reported elsewh ere (Gross et al. 2001). Digestion and extraction effici encies were not influenced by reproductive status or exposure to pulp mill effluent (see Table 2-5 ), but greater efficiency was consistently achieved for 17 -estradiol than testosterone. For 17 -estradiol overall, digestion efficiency averaged 70 + 4.9% while extraction efficiency averaged 65 + 6.4%. For testosterone overall, diges tion efficiency averaged 63+ 3.9% while extraction efficiency averaged 51 + 12.5%. While these values would be considered low for plasma or serum samples, efficiencies at 60% or greater are high for whole body samples. Data was corrected for both digestion and extraction efficiencies. Minimum detection limits on standard curves were 6.4 pg/mL and 9.3 pg/mL for 17 -estradiol and testosterone, resp ectively. Cross-re activities of 17 -estradiol antiserum (produced and characterized by T. S. Gross, Un iversity of Florida) with other steroids were: 11.2% for estrone, 1.7% for estriol, less than 1% for 17 -estradiol and androstenedione, and less than 0.1% for all ot her steroids examined (ICN Biomedicals). Cross-reactivities of testosterone antiserum (produced and characterized by T. S. Gross,

PAGE 75

57 University of Florida) with other steroids were: 17.6% for dihydrotestosterone, 2.3% for androstenedione, 1.4% for 11-ketotestosterone <1.0% for androstenediol, and <0.1% for all other steroids examined (ICN Biomedi cals). A pooled sample (1.25 g of unexposed males and females randomly selected and approximately 135 pg 17 -estradiol/mL and 176 pg testosterone/mL added) was assayed se rially in 10, 20, 30, 40, and 50 L volumes (final volume of 50 L with boiled KOH). Re sulting inhibition curves were parallel to respective standard curve ba sed upon tests for homogeneity of regression indicating curves did not differ. Finally, average inter-a ssay and intra-assay coe fficients of variation were 7.8% and 9.4% for 17 -estradiol and 8.7% and 10.1% fo r testosterone. All values are reported as pg hormone per g body weight. Seasonality Both site and season significantly covaried for 17 -estradiol in female mosquitofish from the discharge [DIS] and first dow nstream [D(1)] sites in Rice Creek ( Figure 2-6A ). However, examination of site and season separa tely does not reveal a consistent pattern. 17 -estradiol appeared seasonally elevated in winter, although the opposite occurred for the lower half of Rice Creek. During the summer, 17 -estradiol was depressed in females from upstream and downstream sites [U (8), DIS, D(1)] compared to remaining sites. Notably, 17 -estradiol in females was not de pressed in 100% effluent before discharge [PRE-DIS]. Females collected in the winter revealed a different pattern: 17 estradiol was significantly elevated at the ups tream site [U(8)] compared to the discharge and first downstream sites [DIS and D(1)]. However, this hormone was not significantly different at effluent-exposed sites compared to reference fish [REF2], reiterating the influence of unexposed site sele ction. While it is ideal to collect upstream of effluent

PAGE 76

58 outfall, additional references pr ovide insight into natural va riability. Based upon inherent variation implied by these data within sites and between unexposed sites, 17 -estradiol alone does not appear influenced by effluent exposure but may be seasonally influenced. Site and season also significantly covaried for testosterone in female mosquitofish from the discharge [DIS] and first dow nstream [D(1)] sites in Rice Creek ( Figure 2-6B ). Patterns for this hormone are clearer. Test osterone was seasonally depressed in winter across all sites. For both seasons, testosterone was elevated at either the first downstream site (summer) or both the di scharge and first downstream sites (winter). Like 17 estradiol, testosterone was not impacted at the 100% final efflue nt before discharge [PRE-DIS]. Thus, elevated te stosterone concentrations were associated with instream effluent exposure but not at highest effluent concentr ations before discharge. Figure 2-7 illustrates percentage of females w ith normal, feminine sex steroid ratios (> 1) versus masculine steroid ratios biased toward testosterone (< 1). Estrogen to testosterone ratios were significantly masculinized for females in the summer (DIS and D(1) which averaged 0.8 and 0.3, respectively). In the winter masculine versus feminine sex steroid ratios were not significantly differe nt among sites. Thus, masculinized steroid profiles appear seasonally affected by efflue nt exposed sites. A low background level of masculine ratios existed in females from unexposed sites [U(8) and REF1 for summer, REF2 for winter], inferring this type of hormone profile can o ccur naturally in the population. In 100% effluent before discharge (summer collection), no females had masculinized estrogen to testosterone ra tios, emphasizing the difference in response between predischarge and instream effluent-exposed sites.

PAGE 77

59 Sex steroids in male mosquitofish covaried by site and season for 17 -estradiol but not for testosterone ( Figure 2-8 ). Within sites, seasonality was inferred at the upstream and outfall sites [U(8) and DIS] for 17 -estradiol and at the outfa ll and lower half of Rice Creek [DIS and D(3+6)] for testosterone. 17 -estradiol was elevated and testosterone was depressed in the winter for these sites, respectively. By site alone, in the winter 17 estradiol significantly peaked at the outfall and testosterone peaked at the upstream site. No differences were detected by site in th e summer. Testosterone data reveal large variation within site and between unexposed sites, similar to 17 -estradiol in females and again stressing importance of reference site se lection and an inherent natural variability (although mosquitofish have a reproductiv ely active season they are asynchronous breeders). Estrogen to testosterone ratios we re dominated by testosterone for all males across all sites for both seasons (significantl y less than one, approximately 0.01 to 0.001). Overall, sex steroids in ma les did not appear impacted by effluent exposed sites and seasonality was incons istently indicated. Recently Toft et al. (2003) examined whol e body sex steroids in male mosquitofish collected in lakes contaminated with agricultu ral pesticides. Techni que was similar to the radioimmunoassay used for our study, alt hough mechanical as opposed to chemical digestion of fish preceded extraction. St eroids were monitored December to May and seasonality was implied, although statisti cal relevance was unstated. Overall concentrations for both steroids were much hi gher than concentrations observed in this study: 17 -estradiol was five times higher and te stosterone was approximately 2 times higher. Extraction efficienci es were higher for the pestic ide lake study (83% and 111% for 17 -estradiol and testosterone respectively) and were not corrected, meaning actual

PAGE 78

60 concentrations were even higher for 17 -estradiol but lower for testosterone. Digestion efficiencies were not reporte d, nor were cross reactivities wi th other sex steroids. Thus direct comparison of results is difficult and laboratory differences, in addition to habitat and seasonal variations, ma y explain discrepancies. Conclusions Validations of gender iden tification, morphological and steroidal measurements support the use of mosquitofish as a bioindicator of pulp and pa per mill effluent. Site and seasonal differences were not readily appare nt in morphological measurements of male mosquitofish, especially in association to effluent-exposed sites. Precocious maturation could not be evaluated in the winter colle ction because small mature males were not captured in adequate numbers fo r statistical comparison. Theref ore, this aspect must be re-examined. For females, size class was not a complica ting factor for the index of anal fin elongation overall, although differential life stage exposure or dynamic exposure could be speculated at the outfall site. Masculinization at sites closest to effluent discharge was evident for both anal fin morphology and ster oids in females: anal fin elongation was independent of season, while effects on steroi ds were seasonal. This difference in seasonality may be the product of separa te mechanisms, exposure to bioactive compounds, and/or perhaps differential exposur e due to drought (see discussion below). However, both anal fin elongation and sex steroids were not measured in the same fish. The development and validation of computer-aided measurements allowed both endpoints to be measured in the same fresh fish for research presented in subsequent chapters.

PAGE 79

61 Importantly, neither hormone concentrations nor ratio were altered in 100% final effluent prior to discharge. Anal fin elongation at this site was not examined until 2001 and 2002, and data is presented in Chapters 4 and 3, respectively. This result points toward the presence of bioactive compounds in the receiving stream below discharge and not in the effluent by itself indicating additional environm ental factors interact to produce the response. Which environmental f actors are influential is pure speculation without chemistry, controlled exposure or l ong-term biomonitoring data associated with this study. With that caveat, there is one factor that is often overlooked yet crucial to the hypothesized mechanism of anal fin masculiniz ation via degraded phytosterols: bacteria. If the bacterial degradation hypothesis is corre ct, instream bacteria l communities may be more efficiently converting phytosterols to androgens than communities living in retention ponds. For example, aerobic microorganisms degraded 17 -estradiol 60 to 130 days faster and by first-orde r kinetics compared to anaerobic microorganisms (Quinn 2004). Natural variation in b acterial communities may also a lter degree of response for that matter. Only one type of bacteria ( Mycobacterium smegmatis ) has been examined in laboratory exposures to degraded phytoste rols (Denton et al. 1985, Howell and Denton 1989, McCarthy et al. 2004), and the Mycobacterium genus is more efficient at transforming phytosterols to androgens than other bacterial species (Marcheck et al. 1972). A potentially confounding factor in these studies was the variation in some responses at unexposed sites. Significant differences be tween reference and upstream sites existed for male gonopodial leng th and testosterone, and for 17 -estradiol in

PAGE 80

62 females. If collection had been restricted to one unexposed site, responses would have been either masked or abnormally inflated. Unfortunately the ideal reference site does not exist: often an upstream site is the best at representing environmental habitats at exposed sites, although habitat can vary subs tantially along a stream. Therefore multiple reference sites within the receiving system region are imperative to document natural variation in response and allow more releva nt interpretation. Overall, there was no difference in female anal fin elongation betw een unexposed sites, in agreement with the geographic study by Bradley et al. (2004). Their study collec ted at dozens of reference and polluted sites (not just pulp and paper m ill effluent), and they pooled reference data for analysis since reference values were not significantly different. Since our study investigated several other endpoints in addi tion to female anal fin elongation, multiple reference sites were ideal wh en funding and efforts allowed. Another confounding situation involved w eather during this study. Summer 2000 fish were living under drought conditions, while winter 2000 fish were collected after relief from drought with a s lightly higher than normal rainy season (Appendix A). According to Davis and Bortone (1992) and Cody and Bortone (1997), anecdotal observation of mosquitofish during times of drought corresponded to increased masculinization of the anal fin. The logical cause was a concentration of effluent, a plausible assumption for the low-flow streams characteristic of the Florida mills under investigation. This argument was used to support seasonal differences detected in anal fin length of females (Cody and Bortone 1997) since precipitation follows a seasonal pattern. However, since wild females surv ive 1 to 2 years, anal fin elongation among summer populations would still be present thro ugh the winter. According to McCarthy et

PAGE 81

63 al. (2004), gonopodial development in female mo squitofish does not regress once effluent exposure ceases. This controversial point, however, has not been well documented. Sex steroid levels, in contrast, are more labile a nd subject to recovery in effluent-exposed fish (McMaster et al. 2003), and th is study documented normal steroid ratios in the winter collection. Thus sex steroids may be a more sensitive marker of differential effluent exposure, while anal fin elongation is a more static and durable marker. Condition factor data cannot support or refute the potential influence of drought due to the experimental bias. A critical limitation of this drought hypot hesis is a lack of e xposure data to support concentration of effluent during times of drought. As an estimate of relative concentration of effluent, c onductivity was greater in the su mmer than winter, providing some support to the drought hypothesis. Re gular monthly field collections could shed more light on apparent trends. However, effects of drought should be considered when using mosquitofish as a bioindicator. Two aspects of previously defined bioindi cator criteria were addressed in this chapter: method and seasonal variation. Met hod variation was low, supporting the use of mosquitofish as bioindicators. Wild-caught females demonstrated greater effects than wild males. In females, seasonal variation related to exposure occurred for sex steroids but not anal fin elongation, i ndicating greater sensitivity of steroids than anal fin morphology. Ideally, both of these markers ar e measured for indivi dual fish to better evaluate the hypothesized li nk between anal fin elongation and altered sex steroid profiles. An expanded year long seasonality study of mosquitofish living in pulp mill effluents would allow better interpretation of absolute hormone concentrations, and changes in anal fin morphology, especially since the present results are in conflict with

PAGE 82

64 reported seasonal effects at Elevenmile Creek (anal fins) and reported concentrations of sex steroids in males. Table 2-1. Water quality parameters of Rice Creek field collection sites in winter 2000 Site REF2 U(8) PRE-DISDIS D(1) D(3+6) Winter 2000 Water temperature (C) 17.7 14.9 16.4 15.2 15.0 18.5 Conductivity (S) 167 173 2132 1112 1315 1336 Salinity (ppt) 0.1 0.1 1.1 0.6 1.3 0.7 Dissolved Oxygen (mg/L) 7.60 7.00 4.14 7.23 11.90 8.44 Turbidity (ntu) 1.90 3.34 32.5 18.9 18.4 13.5 pH 6.73 6.45 7.35 7.23 7.00 7.15 Table 2-2. Correlation coefficients (r2) for morphological measurements made before and after preservation in formalin; between USGS and NCASI laboratories; and between manual and computer-aided measurement by the same observer. Standard Length (SL) was only measured manually using digital caliper s. All correlations were statistically significant at p < 0.05 unless noted. Measurement Ray 4 Ray 6 SL Preservation 0.93 0.94 0.94 Between laboratories 0.54 0.80 0.97 Manual vs. computer-aided 0.27* 0.33* NAa a NA = not applicable *not significantly co rrelated (p > 0.05) Table 2-3. Average coefficients of va riation for manual and computer-aided measurements by observer and among obs ervers (three measurements per trait). Standard Length (SL) was only measured manually using digital calipers. Measurement Ray 4 Ray 6 SL Manual By observer 3.1% 3.6% 1.4% Among observers 8.8% 9.4% 5.6% Computer-aided By observer 1.1% 0.9% NAa Among observers 12.3% 10.1% NAa a NA = not applicable

PAGE 83

65 Table 2-4. Body size parameters (ave + se) for mosquitofish collected in winter 2000 Site REF2 U(8) DIS D(1) D(3+6) Winter 2000 Sample Size 45 (29,16)e 89 (61,28) 73 (73,16) 46 (29,17) 52 (32,20) Body Weight ) 0.255+ 0.010 0.289+ 0.008 0.255+ 0.013 0.221+ 0.006a 0.169+ 0.009a Standard Length (mm) 22.74+ 0.31 24.08+ 0.20b 22.24+ 0.31 22.95+ 0.25 23.78+ 0.33 Condition Factor (g/cm3) 2.13+ 0.03 2.03+ 0.02 2.24+ 0.02c 1.84+ 0.04c 1.22+ 0.03c Sample Size 141 (91,50)e 106 (64,42) 106 (53,53) 128 (99,29) 112 (65,47) Body Weight (g) 0.567+ 0.029 0.527+ 0.029 0.513+ 0.025 0.471+ 0.022 0.479+ 0.033 Standard Length (mm) 28.97+ 0.48 28.67+ 0.46 28.18+ 0.44 27.37+ 0.40 29.13+ 0.59d Condition Factor (g/cm3) 2.08+ 0.01 2.05+ 0.02 2.15+ 0.03 2.11+ 0.02 1.62+ 0.04 a D(1) statistically different than D(3+6); D(3+6) statistically different from all other sites bU(8) statistically different than REF2, DIS, and D(1) (p < 0.05) cDIS; D(1); and D(3+6) statistically different than rest (p < 0.05) dD(3+6) statistically different than rest (p < 0.05) esample sizes displayed as: total sample size (preserved anal fin measurements, hormone and computer-aided measurements)

PAGE 84

66 Table 2-5. Digestion and extrac tion efficiencies, and coeffici ents of variation (CV), by exposure and reproductive status fo r mosquitofish whole body hormone analysis. Digestion Extraction Efficiency CV Efficiency CV 17 -estradiol Exposed 70+ 4.7% 6.7% 66+ 6.6% 10% Unexposed 63+ 3.3% 5.2% 61+ 7.2% 11.7% Exposed 75+ 5.8% 7.7% 71.7+ 9.5% 13.2% Unexposed 76+ 7.6% 10% 60+ 5.8% 9.7% Adult 67+ 6.4% 9.6% 62+ 6.2% 10% Adult gravid 74+ 4.5% 8.9% 75+ 5.8% 12.7% Adult nongravid 72+ 4.1% 9.2% 60+ 6.3% 10.9% Juvenile 69+ 2.8% 5.7% 67+ 4.9% 10% Testosterone Exposed 64+ 2.5% 3.9% 41+ 4.4% 14% Unexposed 57+ 3.9% 7.0% 44+ 2.5% 9.9% Exposed 58+ 3.9% 6.8% 45+ 4.1% 9.1% Unexposed 50+ 2.9% 5.7% 46+ 6.1% 13% Adult 89+ 3.1% 3.1% 66+ 3.7% 14% Adult gravid 61+ 4.5% 7.3% 61+ 4.6% 15% Adult nongravid 67+ 3.8% 5.4% 66+ 5.3% 10% Juvenile 62+ 7.1% 11% 40+ 5.9% 15%

PAGE 85

67 Figure 2-1. Maps of Rice Cree k, a tributary of the Saint Johns River, FL, USA. A) Relative location in Florida. B) Summer 2000 field collection sites. C) Winter 2000 collection sites. D) Summer 2001 collection sites. Site symbols distinguish sites exposed to effluen t: circles = unexposed and triangles = exposed. Site abbreviations denote upstream (U) or downstream (D) of discharge, followed by approximate distance (km) from discharge in parentheses. PRE-DIS indicates site before discharge into the creek; DIS denotes site at discharge into creek; a nd REF indicates reference site, followed by identifying number. B A

PAGE 86

68 Figure 2-1. Continued D C

PAGE 87

69 Figure 2-2. Gender agreement between NCASI a nd USGS laboratories. Fish were sexed externally by USGS using the urogen ital papilla then gonads were grossly identified by NACSI. Question mark s corresponding to NCASI were unable to be reliably sexed by gross gonad examination. agree male agree female male/female* female/male* ?/male* ?/female*44.5% 48% 7.5% *NCASI/USGS

PAGE 88

70 Figure 2-3. Gender agreement within USGS la boratory. Fish were sexed externally using the urogenital papilla then gon ads were examined histologically. Question marks indicate gender could not be reliably identified by histological techniques. agree male agree female male/female* female/male* ?/female*49% 49% 2% *histology/urogenital papilla

PAGE 89

71 SITE REF2U(8)DISD(1)D(3+6) INDEX OF ANAL FIN ELONGATION 0.0 0.5 1.0 1.5 20-24.99 mm 25-29.99 mm 30-34.99 mm 35-39.99 mm u r r r u r u u r r r u u u r r u r r d d d* Figure 2-4. Female index of anal fin elongation (linear Ray 4 / Ray 6, manually measured on preserved fish) for each site by 5 mm increments (winter 2000). Significant differences (p < 0.05) are co lor coded within size class: u = different from upstream site [U(8)]; r = different from reference site [REF2]; and d = different from other exposed sites [DIS, D(1)]. Purple asterisks indicate significant differences to the la rgest size class within site (p < 0.05).

PAGE 90

72 REF1REF2U(8)DISD(1)D(3+6) INDEX OF ANAL FIN ELONGATION 0.0 0.5 1.0 1.5 2.0 2.5 SUMMER WINTER b b a SITE REF1REF2U(8)DISD(1)D(3+6) INDEX OF ANAL FIN ELONGATION 0.0 0.5 1.0 1.5 2.0 2.5 Aa c e d Bd Figure 2-5. Index of anal fin elongation ( linear Ray 4 / Ray 6, manually measured on preserved fish) for winter and summer months in 2000. A) Females. B) Males. Dashed lines separate sites not involved in ANCOVA analysis by site and season. Letters indicate significant differences among sites within season (p < 0.05): “a” denotes differences to nonlettered sites except D(1) was not different than REF1; “b” denotes differen ces to nonlettered si tes; “c” denotes differences to all but REF1; “d” denotes differences to REF1 and U(8); “d” denotes differences to DIS, D(1), and REF1. Season was significant within site (p < 0.05) for females at the tw o downstream sites only [D(1) and D(3+6), circled in yellow].

PAGE 91

73 SITE REF1REF2PRE-DISU(8)DISD(1)D(3+6) 17BETA-ESTRADIOL (pg/g) 0 200 400 600 800 1000 1200 SUMMER WINTER SITE REF1REF2PRE-DISU(8)DISD(1)D(3+6) TESTOSTERONE (pg/g) 0 100 200 300 400 500 600 700 a a a c d dA Bb* Figure 2-6. Female whole body se x steroids from collections made in the summer and winter of 2000 (ave + se). A) 17 -estradiol. B) Testos terone. Dashed lines separate sites not involv ed in ANCOVA analysis by site and season. Letters indicate significant differences by site within season (p < 0.05): “a” denotes differences to nonlettered sites; “b” de notes differences to DIS and D(1); “c” denotes differences to all but DIS; a nd “d” denotes differences to all other sites. Yellow circles demonstrate diffe rences between seasons (p < 0.05). Green asterisk show that both hormones covary by site and season (p < 0.05).

PAGE 92

74 % FEMALES 0102030405060708090100 SITE REF1 U(8) PRE-DIS DIS D(1) D(3) D(6) % FEMALES 0102030405060708090100 SITE REF2 U(8) DIS D(1) D(3+6) MASCULINE STEROID RATIO (E:T<1) FEMININE STEROID RATIO(E:T>1) A B Figure 2-7. Percentage of female mosquitofish with masculine and feminine sex steroid ratios collected in 2000. A) Summer ( 2 = 40.12, df = 6, p < 0.05). B) Winter ( 2 = 8.270, df = 4, p < 0.05).

PAGE 93

75 SITE REF1REF2PRE-DISU(8)DISD(1)D(3+6) 17BETA-ESTRADIOL (pg/g) 0 10 20 30 40 50 60 SUMMER WINTER A SITE REF1REF2PRE-DISU(8)DISD(1)D(3+6) TESTOSTERONE (pg/g) 0 200 400 600 800 1000 1200 a b* B Figure 2-8. Male whole body se x steroids from collections made in the summer and winter of 2000 (ave + se). A) 17 -estradiol. B) Testos terone. Dashed lines separate sites not involv ed in ANCOVA analysis by site and season. Letters indicate significant differences by site within season (p < 0.05): “a” denotes differences to all other sites; “b” deno tes differences to REF2, DIS and D(1). Yellow circles demonstrate differences between seasons (p < 0.05). Green asterisk signifies 17 -estradiol covaried by si te and season (p < 0.05).

PAGE 94

76 CHAPTER 3 DIMINISHED EFFECTS OF PULP AN D PAPER MILL EFFLUENT ON EASTERN MOSQUITOFISH BEFORE AND AFTER MAJOR PROCESS IMPROVEMENTS The US Environmental Protection Agency’s (EPA) Cluster Rule was enacted in 1998 to regulate both air and water pollu tion by the pulp and paper industry. Final implementation of this rule was carried out in 2001 by the Georgia-Pacific bleached kraft mill in Palatka, FL. Comparison of mosquitofish collections before and after these major process changes was conducted to determine if responses were reduced or eliminated. Males and females were evaluated for anal fin morphology and w hole body sex steroids, although male responses seemed more affect ed by potential seasonality as opposed to effluent exposure. Female anal fin elongation was cons istently reduced, but not eliminated, in effluent-exposed sites. Female sex steroid concentrati ons were difficult to interpret and may be seasonally a ffected; however the ratio of 17 -estradiol to testosterone indicated a masculnized horm one profile remained despite processing improvements. No association between anal fin elongation and sex steroid levels or ratios was apparent. Anal fin morphology likely portrays longer term, chronic exposure whereas sex steroids provide a snapshot of mo re recent exposure. Overall, reduced anal fin elongation associated with major process changes supports the use of mosquitofish as a bioindicator of effluent expos ure. Seasonal effects on sex st eroids need to be clarified before use of this biomarker.

PAGE 95

77 Introduction Pulp and paper mills release a complex mixture of chemical compounds into the aquatic environment that has been a source of environmental concern for the past two decades (Borton et al. 2004). Compounds such as chlorina ted organics, metals, and wood extractives (e.g. resin acids, fatty acids, phyt osterols, and lignin) have been shown to cause an array of effects in fish (EPA 2002 a nd Borton et al. 2004). Ac ute lethal toxicity concerns are no longer an issue, while chronic sublethal toxicity rema ins controversial. Linking actual adverse effects to effluent component(s) has been a challenge since effluents are complex by nature, and their variation in effluent composition among mills and even within a single mill. Within this effort, a wealth of information has been generated about effects in fish. Pulp and paper mill effluents have been shown to alter reproductive function in indigenous fish species. For example, white sucker ( Catostomus commersonii ) collected near Canadian mills had reduced ovarian steroid biosynthesis, delayed sexual maturity, decreased gonad size, and redu ced expression of secondary sex characteristics (McMaster et al. 1991, Munkittrick et al. 1991, McMaster et al. 1995, Van der Kraak et al. 1992). Largemouth bass ( Micropterus salmoides ) exposed to whole effluent dilutions at the same mill examined in our study exhibited depr essed sex steroids, vitellogenin, GSI, fry production and fry survival in largemouth ba ss (Sepulveda et al. 2001, Sepulveda et al. 2003). Several reported effects on fi sh reproduction imply an andr ogenic effect. A Finnish totally chlorine free (TCF) kraft mill was associated with male-biased sex ratios in wild eelpout ( Lycodes sp.) (Larsson et al. 2000). In a st udy of the same mill, increased male coloration was observed under laborato ry exposure of livebearing guppies ( Poecilia

PAGE 96

78 reticulate ) (Larsson et al. 2002). Full life cy cle testing using the fathead minnow ( Pimaphales promelas ) in Canada demonstrated both masculinization of females and feminization of males (Parrott and Wood 2002, Parro tt et al. 2003). Pulp mills in Florida have been associated with development of ma le secondary sex characteristics in female mosquitofish (Howell et al. 1980, Bortone and Drysdale 1981, Cody and Bortone 1997, Bortone and Cody 1999, Jenkins et al. 2001, Park s et al. 2001) and possibly precocious maturation in male Eastern mosquitofish ( Gambusia holbrooki ) (Howell et al. 1980, Drysdale and Bortone 1989). Unique among th e vast majority of these effects-based studies, the mosquitofish work has led to a proposition of bioactive compounds: androgens formed by bacterial degradati on of phytosterols. This hypothesis has stimulated controversial disc ussions about mosquitofish as a bioindicator or sentinel species for pulp and paper mill effluents. Eastern mosquitofish have been consider ed for use in regulatory testing and screening of pulp mill effluent toxicity at both the state (Florida Department of Protection, T. S. Gross, pers. comm.) a nd federal levels (Angus et al. 1997). Mosquitofish have been proposed since most other wild small fish species found in effluent-receiving streams cannot be collected in adequate nu mbers or maintained in the laboratory for use in research. As data on reproductive effects in fish has been generated, pulp and paper mills have been improving and refining process tec hnologies. Upgrades in pulp production processes have the poten tial to reduce if not abolish reporte d effects. For example, shortterm laboratory exposures of goldfish ( Carassius auratus ) revealed a rec overy of steroid function following unknown process changes (M cMaster et al. 1996). More recently,

PAGE 97

79 temporary shutdown of the TCF mill associated with male-biased sex ratios in eelpout allowed recovery of normal sex ratios in the exposed population (Larsson and Frlin 2002). With the implementation of US EPA’s Cl uster Rule in 2001, the Georgia-Pacific mill located in Palatka, FL, USA has modifi ed its pulping and bleaching processes. The objective of this study was to evaluate the in fluence of these process changes on a native small fish species, the Eastern mosquitofi sh, specifically examining effects on two endpoints: anal fin morp hology and sex steroids. Materials and Methods Mill Characteristics Georgia-Pacific’s mill in Palatka, Florida, USA is a paper grade bleached kraft mill established in 1947. It has two bleached (40% product) and one unbleached line (60% product). The bleaching lines manufacture pape r towels and tissue paper, whereas the unbleached line produces kraft bags and linerboa rd. Wood furnish for this mill typically consists of 50% softwood (slash, sand and loblolly pines) and 50% hardwood (gums, tupelo, magnolia and water oaks) cycled back and forth between the two types of furnish. Effluent receives secondary treatment consisting of anaerobic followed by aerobic degradation with a rete ntion time around 40 days. Effluent discharges into Rice Creek, near ly 6 km upstream of the confluence with the Saint Johns River (Figure 2-1). Rice Creek is a low-flow, tannic stream, so dilution factor for effluent is low until it reaches the Saint Johns River. Before process changes the average yearly instream effluent concen tration was approximately 60% until reaching the St Johns River, where concentration dr opped below 10%. In c ontrast, most North

PAGE 98

80 American mills average less than 5% year ly instream effluent concentration by discharging into larger and/or faster-flowing bodies of water. Before major process improvements, the Palatka mill released approximately 36 million gallons of effluent per day (mgd). Bleaching pre-process modifications used elemental chlorine and up to 10% chlorine di oxide substitution. The bleaching sequences were C90d10EopHDp and CEHD for the softwoods a nd hardwoods, respec tively. Process modifications in May 2001, to meet EPA the Cluster Rule, involved: 1) conversion to ECF bleaching via 100% chlorine dioxide subs titution; 2) reduction in black liquor losses; 3) added condensate stripping; 4) conversion of all rete ntion ponds to aerobic degradation; and 5) reduction in water use re sulting in release of approximately 28 mgd effluent. The current bleaching sequence is DEopD for both types of furnish. Field Collections Field collections of adult mosquitofish occurred during the reproductively active summer months for this species one year before (March and June 2000, n = 174 and n = 141 respectively) and one year after (April 2002, n = 363) pro cess modifications. Water quality parameters typically affected by pulp and paper mill effluents were measured before fish collection at each site: disso lved oxygen, temperature, pH, conductivity, salinity, and turbidity. Adult Eastern mos quitofish were collected along shallow vegetated banks using dip nets and a bac kpack electroshocker at several locations upstream and downstream of effluent discha rge in Rice Creek, and at reference sites lacking effluent exposure or any ot her known point sources of pollution ( Figure 3-1 and Appendix A). Fish were transported back to the laboratory in oxygenated bait buckets then euthanized with a termin al dose of buffered tricaine methanesulfonate (Tricaine-S, Western Chemical Inc ., Ferndale, WA, USA).

PAGE 99

81 Morphology Once euthanized, fish were examined unde r a dissecting scope to determine gender using the urogenital papilla (Chapter 2). Ge neral measurements of body size and selected measurements of anal fin morphology were taken for all adult fish collected. Body weight (+ 0.001 g) and standard length (+ 0.01 mm) were measured using a digital scale and a pair of digital calipers. In 2000, the March collection was preserved in 10% neutral-buffered formalin for anal fin measur ements while the June collection was frozen and stored at -80C for subsequent radioi mmunoassay (RIA) of sex steroids. Linear distance from base to tip of Ra ys 4 and 6 of the anal fin (+ 0.1 mm) were measured for formalin-preserved fish under a dissecting sc ope using an ocular micrometer. In 2002, a subset (n = 87) was preserved in 10% neut ral-buffered formalin for comparison to 2002 anal fin data. Remaining fish were photogra phed digitally before freezing for sex steroid analysis. Digital photographs of anal fins fo r these fish were measured using a computer software program (SigmaScan Pro 5.0, SPSS, Inc.), tracing along the lengths of Rays 4 and 6 (+ 0.01 mm). Chapter 2 gives validati on of these morphological measurements. Sex Steroids Whole body primary sex steroids (17 -estradiol and testoste rone for this species) were analyzed using a modified RIA method originally developed for serum and plasma samples of common carp, Cyprinus carpio (Goodbred et al. 1997), and since adapted for use in a variety of other aqua tic species and tissue media such as plasma of largemouth bass (Gross et al. 2001) and ma ntle of freshwater invert ebrates (Gross et al. 2000). Chapter 2 gives methods and validation of this assay.

PAGE 100

82 Statistics Body weight and standard length were us ed to calculate condition factor, K = weight / length3 x 100 (g/cm3), as an indication of overall health used by the aquaculture industry (values greater than 1 ar e considered healthy; less than one are considered poor). The length ratio of anal fin Rays 4 and 6 was calculated as an index of anal fin elongation. Estrogen and testoste rone concentrations were used to calculate a ratio indicating either a masculine hormone profile (E:T < 1) or a feminine hormone profile (E:T > 1). Any data failing tests for normality and homogeneity of variance were log transformed. Anal fin morphology and sex steroid concentrations were analyzed separately within sex using two-way analys is of covariance (ANCOVA) to test for significant variation by site and year. Site di fferences within year were also analyzed by one-way ANOVA. Significant differences in the ANCOVA and ANOVA were followed by multiple comparison tests using Tukey’s HSD. Within site, differences between years were analyzed by Student’s t-test. Fish measured for both anal fin morphology and sex steroids (2002 data only) were analyzed in two ways: first, by examining Pearson’s correlations of the index of anal fin elongation to sex ster oid concentrations and ratio, then by t-test for differences in index of anal fin elongation between females with masculine versus feminine E:T ratios. Statistical significance was set at < 0.05 for all tests. All statistical analyses were conducte d using SAS version 9.0. Results and Discussion Water Quality As expected, conductivity, sa linity and turbidity were higher at effluent-exposed sites compared to the upstream site ( Table 3-1 ). The reference site REF1 was more

PAGE 101

83 similar to effluent-exposed sites than the upstream site in terms of water quality parameters. Dissolved oxygen remained high enough to support fish at most sites (> 4 mg/L), except in 2002 at the predischarge si te [PRE-DIS] where levels were extremely low (< 1 mg/L). The fact that mosquitofish were present in this low oxygen environment demonstrates their tolerance to extreme environmental conditions. After process changes, conductivity was reduced at downstr eam sites sampled both years [DIS, D(1), D(3)] as a gross indication of reduced effl uent concentration and/or improved effluent quality. Body Size and Condition Effluent exposure, regardless of proce ss changes, was not associated with alterations in body size or condition in both male and female mosquitofish ( Table 3-2 ). Rather, differences in condition between year s were drastic (poor condition with CF < 1 in 2000 compared to good condition with CF > 1 in 2002), likely caused by fixation of 2000 fish before body size measurement (thereby decreasing weight of fish). For the 2000 collection, a statistically significant difference in these parameters occurred between populations in Rice Cr eek compared to Saint Johns River, therefore river sites were excluded from sampling in 2002. Males Before process changes, males at the first downstream site [D(1)] were statistically longer than males from the upstream site [(U8)] ( Table 3-2 ). Otherwise, males collected at effluent-exposed sites [DIS and D(3)] did not show any significant variation from unexposed sites. Males living in Saint Johns River [D(6) and REF1] were in the poorest relative condition of all sites collected, indicating these popula tions may be significantly different than creek populations. Condition fa ctor cannot be assessed in terms of general

PAGE 102

84 health status since this colle ction was preserved before meas urement of body size. After process changes, effluent exposure had no obvi ous effect on size or condition of males. Although length and weight of males at the discharge site [DIS] was significantly less than upstream males [U(8)] (length) or al l other sites (weight), condition factor was statistically equivalent and above one indicating good overall health. Females Before process changes, female body size and condition was not overtly affected by site when comparing exposed to nonexposed sites ( Table 3-2 ). Statistically significant variation occurred between females from une xposed sites [U(8) and REF1] for length, weight and condition factor. In addition, c ondition factor was st atistically different between the upstream site [U(8)] and the c onfluence of Rice Creek and Saint Johns River [D(6)]. Together with data on males, thes e data indicate varia tion between populations inhabiting the creek compared to the river. Thus, these river sites were excluded from collection in 2002. Similar to males, female body size and c ondition in 2002 were greater than what was observed in 2000 most likely caused by pres ervation state. Overall variability in length was reduced. Females collected from retention ponds before discharge into Rice Creek [PRE-DIS] were statistically larger in weight and length while condition factor was less, compared to the upstream site [U(8)] Although statistically significant differences in condition factor were dete cted (reduced in PRE-DIS and elevated in DIS), these differences are not appropriately interprete d beyond the benchmark of above or below a value of 1. All values were above one indicating good general health.

PAGE 103

85 Anal Fin Morphology Major process changes at the mill in 2001 we re associated with a reduction, but not elimination, in the masculinization response fo r female mosquitofish. Further, 2002 data implied a dose-dependent response. Le ngth of gonopodia in males was questionably affected before process changes, and definite ly not affected after improvements. Rather, an overall shift toward gr eater gonopodial length was evident after processing modifications, perhaps reflecting environmen tal stress on these animals in 2000 separate from effluent exposure. Vari ation in response at reference sites occurred for males but not females (2000 data only, additional reference site was not included in 2002 collection). Males Figure 3-2 shows the index of anal fin elonga tion for male gonopodia before (2000) and after (2002) major process modificatio ns. For both years, gonopodia (Rays 3, 4 and 5) extended at least twice as long as the rest of the fin. At the tips, the gonopodia were marked by terminal differentiations (hooks, serrae and blade; in the photographs, hooks are the most visible of these structures). These terminal structures signify complete maturation and the end of gonopodial development. Before process changes, male index of anal fin elongation was significantly different among sites ( Figure 3-3A ); however, the relationship to effluent exposure was unclear. Compared to males from the upstr eam site [U(8)], gonopodia were longer at the discharge [DIS] and at the first downstream s ite [D(1)]. Yet the reference site [REF1] also had males with significan tly longer fins in relation to upstream males. Therefore, significant differences were dependent on bot h unexposed and effluent-exposed sites.

PAGE 104

86 After process changes, males no longer had any statistically si gnificant differences in the index of anal fin elongation among sites ( Figure 3-3B ). Within each site, distribution of the index wa s normal, as opposed to the nonnormal distributions that characterized 2000 male data (compare Figure 3-3A and B ). Interaction of site and season was not statistically si gnificant for male gonopodial lengt h. For all sites collected both years [U(8), DIS, D(1)], male s had significantly longer gonopodia in 2002 regardless of effluent exposure. Since obser ver and experimental bi as was not an issue like the body size data (Chapter 2) these data may allude to environmental stress such as the 1999/2000 drought (Appendix A). Unfortunate ly, the potential st ress of a drought obscures any conclusions that could have b een made about precocious maturation, or the development of gonopodia in younger, smaller males. Under ideal environmental conditions, precocious maturation would be demonstrated by consistently longer gonopodia relative to standard length at e xposed sites versus unexposed sites and at exposed sites before process ch anges versus after modification. Females Figure 3-4 shows differences in female inde x of anal fin elongation among field collection sites before (2000) and after (2002) major process modifications. For both years, anal fin elongation resembled a de veloping male gonopodium, as opposed to a mature gonopodium, in both length of el ongation (averaging 1.2 in 2000 and 1.1 in 2002 for females versus 2.0 and 2.5 for males) and co mplete lack of terminal differentiations ( Figure 3-2 ). In 2000, female mosquitofish had significan t anal fin elongation at the discharge [DIS] and first downstream site [D(1)] compared to all ot her sites, as displayed in Figure 3-5A Further downstream, elongation of the anal fin was not significant [D(3) and

PAGE 105

87 D(6)]. In contrast to body length and wei ght measurements of females, there was no statistical difference in anal fin elongati on between the reference and upstream sites [REF1 and U(8)]. In 2002, anal fin elongation remained at effluent-exposed sites [PRE-DIS, DIS, D(1)] compared to the upstream site ( Figure 3-5B ). Dose-dependence was implied by decreasing elongation with increasing distan ce from the discharg e point. Anal fin elongation before discharge was not significantly different to elongation at the discharge site, while elongation was different between th e discharge and first downstream sites. Interestingly, anal fin elongation was presen t at the 100% final e ffluent site before discharge, despite no effects observed on sex steroid ratios in 2000 (Chapter 2 and Figure 3-7 ). Similar to males, the combination of site and year did not covary significantly. However, anal fin elongation in females was significantly reduced (average 8% reduction) after process change s for effluent-exposed sites [DIS, D(1), D(3)] but not the upstream site. Thus the masculinization re sponse appears reduced, but not eliminated, in regard to mill process improvements. Sex Steroids Absolute hormone concentrations between years for both males and females indicated seasonal changes as opposed to mill processing-related changes. The exception was elevated testosterone in females exposed to effluent in Rice Creek in 2000 but not 2002, implying a reduction in masculinized horm onal response. However, masculinized or testosterone-biased hormone profiles re mained dominant further downstream at effluent-exposed sites in 2002. For both year s, sex steroids in females living in 100% final effluent before discharg e were not altered similar to instream females, supporting the contribution of additional environmental f actors to produce an endocrine response or

PAGE 106

88 indicating a differential or dynamic exposure to bioactive compounds that allows steroid levels to recover. (Dynamic exposure refers to variable concentr ations of effluent components over time, dependent on factors such as tree species for furnish, within plant processing spills, rainfall/dilution, and bacteria l degradation.) Large individual variation within sites and significant differences betw een unexposed sites, indicate a naturally high variation in this biomarker. Males Male sex steroids before pr ocess changes were not affected by effluent exposure ( Figure 3-6 ). 17 -estradiol concentrations were betw een nondetectable le vels to 50 pg/g, while testosterone values ranged between 500-1200 pg/g and displayed large variation among individuals. Estrogen to testoster one ratios were extremely dominated by testosterone across all sites (less than one, approximatel y 0.01 to 0.001). After process changes in 2002, male sex steroids varied among effluent exposed sites but not to the upstream site. There was no statistical interact ion or covariance between site and year. Comparing years, a marked overall increase of 17 -estradiol in males characterized post-process changes, approximately five tim es higher than concentrations measured before improvements and within range of con centrations reported by Toft et al. (2003). Testosterone, on the other hand, remained within the same range and variation before and after process improvements. As fu rther evidence of the shift in 17 -estradiol, estrogen to testosterone ratios were mu ch closer to one (0.36 to 0.69), while still retaining a masculine profile in the majority of males. Since this change was i ndependent of effluent exposure, most likely this is further eviden ce for seasonality of sex steroids in male mosquitofish (compared to winter an d summer 2000 analyses in Chapter 2).

PAGE 107

89 In partial support of this evidence for seas onality, Toft et al. (2003) measured sex steroids in male mosquitofish from D ecember to May in reference and pesticidecontaminated lakes and found a decrease in 17 -estradiol concentrations from December (300 pg/g) to March (100 pg/g), with levels beginning to rise in April back to December values by May (regardless of exposure). Si nce 2000 and 2002 collections occurred in April and June, respectively, 17 -estradiol levels may reflect this normal rise during the beginning of reproductive season. Unfortunately th e other half of the cycle, from June to November, was not included in Toft et al. (2003), nor were hormones monitored continuously throughout one year for our study. Thus there is an incomplete inderstanding of the seasonality of hormone concentrations in males. Females Before process changes, 17 -estradiol was depressed in females from upstream and downstream sites [U(8), DIS, D(1) ] compared to remaining sites ( Figure 3-7A ). Notably, 17 -estradiol in females was not depressed in 100% effluent before discharge [PREDIS]. As explained in Chap ter 2, the differences between unexposed sites imply a natural variation for this hormone and do not consis tently indicate impacts from effluent exposure. Testosterone was elevated at th e first downstream site in 2000, but was not impacted in 100% final effluent before discharge [PRE-DIS] ( Figure 3-7B ). Thus, elevated testosterone concentrations were associated with initial instream effluent exposure but not at highest e ffluent concentrations before discharge. Estrogen to testosterone ratios were masculinized for most females (less than one, in favor of testosterone) at the discharge [DIS] and first downstream [D(1)] sites (average ratios of 0.8 and 0.3, respectively). Average ratios were normal for females (above one, in favor

PAGE 108

90 of estrogen) for most females from all ot her sites, although a low background level of masculine ratios existed at unexposed sites. In Chapter 2, Figure 2-6A depicts the percentage of females collected during the summer 2000 with normal estrogen to testosterone ratios (> 1) and those females with masculinized ratios in favor of testosterone. Sex steroids for females did not vary among sites for 2002, and were significantly reduced for both hormones at all but one site [17 -estradiol at DIS] compared to 2000 hormone data ( Figure 3-7 ). Consistent differences betw een years regardless of effluent exposure implied seasonality of a different natu re than males, although a decrease in both hormones is counter-intuitive to the onset of reproductive season. Lacking basic knowledge of year to year s easonality in unexposed females, these hormone data remain difficult to interpret. Estrogen to testosterone ratios, c onversely, clearly indicated hormonal masculinization (E:T > 1 for greater than 50% of females) at instr eam effluent exposed sites ( Figure 3-8 ). Average sex steroid ratios were te stosterone-biased at the two furthest downstream sites [0.7 at both D(1) and D(3)], but not before or at the discharge sites [1.3 at PRE-DIS and 1.0 at DIS]. This shift in masculinized hormone profile farther downstream of effluent outfall following proces sing improvements may also indicate that additional environmental factors influen ce hormonal response as predicted by the bacterial degradation hypothe sis for anal fin elongation. Association Between Anal Fin Morphology and Sex Steroids Females collected for sex steroid analys is in 2002 were also photographed for computer-aided measurement of anal fins (s eparate from manual measurements made for

PAGE 109

91 comparison to 2000 anal fin data). No correlation existed for the index of anal fin elongation to estrogen, testos terone, or the estrogen to testosterone ratio (r2 < 0.5 and p>0.05 for correlations). As further proof ag ainst relationship between these endpoints, Figure 3-8 gives average (+ se) index of anal fin elongati on for females with masculine and feminine hormone ratios. Across all si tes and within site, elongation did not differ between the two groups. Females with normal, feminine hormone profiles were just as likely to have masculinized anal fins as fe males with masculine hormone profiles. This does not rule out alterations in sex steroid ra tios contributing to elongation of the anal fin, since sex steroids were measured after ons et of elongation. Seasonal changes in sex steroids, but not anal fin elonga tion, support this point (Chapter 2). However, presence of the hormonal alteration cannot be used to pr edict occurrence of an al fin elongation in individual females living in effluent -receiving streams, based upon these data. Conclusions Major processing improvements implemented in 2001 by the Georgia-Pacific Palatka, FL mill to meet Cluster Rule requirements did not eliminate responses in wild female mosquitofish inhabiting Rice Creek, while males are likely not impacted. Female anal fin elongation was signifi cantly reduced at downstream sites (average 8% reduction), while masculinized sex steroid ratios remain ed with a majority bias shifted further downstream. No clear relationship exists between anal fin elongation and whole body primary sex steroid concentrations or rati os. Importantly, anal fin elongation at the predischarge 100% effluent site was equivalent to elongation detected in the creek, even though hormone profiles were not affected. The 2002 female anal fin data alludes to a dose-dependent response in the creek, however an increased effect is not observed at the site of highest effluent concentration

PAGE 110

92 (before discharge into the cr eek). Combined with the 2000 anal fin data, anal fin elongation appears to be a threshold response. Angus et al. (2001) concluded an all-ornone response of female anal fins to diet ary 11-ketotestosterone exposure (at 0, 20, 40, 60, 80, and 100 g/g). The highest concen tration produced elongation and terminal differentiation the fastest, while all other treatm ents were comparable in terms of rate of elongation (but not rate of di fferentiation which varied inco nsistently). Extent of elongation was less in the lowest dose, but similar among all other doses. Also, length ratio of Rays 4 and 6 (measured by com puter) ranged from 1.35 to 1.50. An actual threshold could not be calculated since all treatments responded, and can only be stated as less than 20 g/g feed. Comparing current data to these results, mosquitofish in Rice Creek are probably exposed below test c oncentrations of androgenic compounds but above the actual threshold. Average index of elongation at exposed sites was 1.2, and no terminal differentiations were recorded. This implies a partial or incomplete response at Rice Creek and exposure may be ne ar the threshol d concentration. Hormone data reveal within season differe nces in addition to effluent exposure likely influenced response. Without baseli ne knowledge of seasonality in hormones, interpretation of effects due to exposure is tentative. An acclimation of response in the highest exposure group (before discharge) may be possible, since both years revealed no effects of exposure on estrogen to testosterone ratios. Equally possibl e is the short-term recovery of normal steroid ratios duri ng periods of exposure below threshold. Concentration of bioactive compounds may be cycling above and below the threshold. Wood extractives present in final e ffluent can vary widely over short periods of time (Chapters 5 and 6). The Rice Creek mill cycles between hardwood and softwood

PAGE 111

93 tree species, and softwoods generally contain more wood extractives such as phytosterols (Smook 1999, Svenson and Allard 2004). In addition, there is th e “black box” of bacterial communities and how they change over time in the predischarge secondary treatment lagoons and the receiving stream. Even more exposure complexity may occur in these low-flow systems from precipitati on and periods of drought and flood (Chapter 2 and Appendix A). A scenario of dynamic exposure appears plausible. This concept would also explain the lack of relationship between anal fin elongation and sex steroids. Lister and Van der Kraak (2001) pointed out a similar lack of relationship between sex steroids and ot her measures of reproductive function (gonad size, age at maturity) among seve ral species exposed to pulp mill effluents in Ontario. As mentioned in Chapter 2, sex steroids are likely more sensitive and labile biomarkers than anal fin elongation which appears more static. Unfortunately, lacking exposure data specifica lly for these studies limits validity of this concept of dynamic effluent exposure and complete demonstration of improved effluent quality due to process changes. Re search on other fish species has also been conducted for the Rice Creek system using controlled effluent exposures, providing indirect information on exposure before and af ter process changes. Life cycle exposure to fathead minnows by NCASI occurred before (1998) and after (2002) mill process changes (NCASI 2000a and NCASI, unpublished data). Based upon chemical analyses of 100% whole effluent samples, before pr ocess changes this mill had some of the highest concentrations of organic compounds compared to other kraft mills studied by NCASI. For example, 615+ 369 g/L for three fatty acids, 4,008+ 1,675 g/L for nine resin acids, 174+ 33 g/L for three chlorinated resin acids and 380+ 169 g/L for four

PAGE 112

94 phytosterols. These concentr ations dropped substantially after process improvements: chlorinated resin acids dropped by 97% while resin acids, fatty acids and phytosterols dropped an average of 80%. The fact that we observed relatively modest reduction in effects in the face of these significant reducti ons to effluent com ponents further supports the concept of a threshold eff ect as opposed to a dose-response effect. Virtual removal of chlorinated compounds also implies they are not the bioactive agen ts causing anal fin elongation in mosquitofish. Additional data on resin acids and phytoste rols (campesterol mainly) in bile of largemouth bass exposed just before process changes in 2001 and after process changes in 2002 similarly indicates reduction in expos ure (Quinn 2004). Further, excretion of these compounds in bile dropped at 40% a nd 80% exposure indicating inhibition of detoxification pathways. Th is inhibition of excretion correlated to significant reproductive effects in these bass such as de pressed sex steroids and GSI (Noggle et al. 2004). Thus a threshold effect in mosquito fish, perhaps caused by overwhelming of the detoxification system, is supported by the largemouth bass work. This study focused upon bioindicator crit eria related to w ild exposure and, inadvertently, seasonal variability. Specifi cally, wild female mosquitofish appear sensitive to improved effluent quality in regards to anal fi n masculinization. Effect of upgrades on sex steroids was unclear because of: 1) large natural variation, and 2) unknown seasonal impacts hinder ing interpretation. While masculinized sex steroid ratios may be associated with effluent e xposure in the receiving stream, they are not predictive of anal fin responses so these biomarkers cannot be associated based upon these data. However the two biomarkers together may be useful to contrast static or long-

PAGE 113

95 term versus dynamic or short-term responses. This would be especially useful if the concept of dynamic exposure proves to be correc t. Exposure history is crucial to support the latter proposition. Finally, the unique response patter n in females living in the predischarge site may be key to de fining mechanisms in future studies. Table 3-1. Water quality parameters of fiel d collection sites before (2000) and after (2002) process changes at the Georgia-Pacific Palatka mill. Site REF1 U(8) PRE-DI S DIS D(1) D(3) D(6) 2000–before Temperature (C) 23.7 18.7 NAa 23.5 23.1 24.9 25.1 Conductivity (S) 955 240.4 NA 1909 1815 1091 684 Salinity (ppt) 0.5 0.1 NA 1.0 0.8 0.5 0.5 Dissolved Oxygen (mg/L) 6.84 6.31 NA 5.56 7.23 3.72 6.83 2002–after Temperature (C) NA 23 26.2 25.7 28.5 28.7 NA Conductivity (S) NA 230.6 2000.2 1814 1340 997 NA Salinity (ppt) NA 0.1 1.0 0.9 0.7 0.5 NA Dissolved Oxygen (mg/L) NA 6.58 0.74 10.1 8.27 4.60 NA Turbidity (ntu) NA 2.43 17.0 18.0 9.92 7.04 NA pH NA 7.41 7.85 7.45 7.32 7.18 NA aNA = not available

PAGE 114

96Table 3-2. Body size parameters (ave + se) and sample sizes for mosquitofish collected before (2000) and after (2002) process changes. Site REF1 U(8) PRE-DIS DIS D(1) D(3) D(6) 2000 – before Sample Sizeg 30 (20,10) 18 (8,10) NAa 29 (19, 10) 30 (20, 10) 26 (16,10) 21 (11,10) Body Weight (g) 0.119+ 0.010 0.114+ 0.001 NA 0.125+ 0.008 0.146+ 0.009 0.134+ 0.009 0.129+ 0.018 Standard Length (mm) 26.43+ 0.55 24.70+ 0.58 NA 26.23+ 0.43 28.24+ 0.51b 26.98+ 0.48 27.94+ 1.04b Condition Factor (g/cm3) 0.62+ 0.02b 0.74+ 0.02 NA 0.68+ 0.02 0.64+ 0.01 0.67+ 0.03 0.56+ 0.02b Sample Sizeg 27 (17,10) 17 (7,10) NA 29 (19,10) 30 (20,10) 22 (12,10) 14 (4,10) Body Weight (g) 0.217+ 0.019 0.584+ 0.072c NA 0.337+ 0.035 0.589+ 0.046c 0.419+ 0.085 0.550+ 0.292 Standard Length (mm) 30.85+ 0.76 40.14+ 1.28d NA 33.55+ 0.91 39.52+ 0.86d 34.81+ 2.40 39.09+ 5.54 Condition Factor (g/cm3) 0.71+ 0.01 0.88+ 0.04e NA 0.84+ 0.02e 0.91+ 0.02e 0.81+ 0.03e 0.70+ 0.06 2002 after Sample Sizeg NA 40 (20,20) 40 (20,20) 40 (20,20) 40 (20,20) 15(0,15) NA Body Weight (g) NA 0.186+ 0.010 0.193+ 0.008 0.147+ 0.007f 0.204+ 0.012 0.179+ 0.022 NA Standard Length (mm) NA 21.46+ 0.36 22.30+ 0.31 20.32+ 0.32b 22.05+ 0.36 21.40+ 0.85 NA Condition Factor (g/cm3) NA 1.82+ 0.03 1.71+ 0.03 1.71+ 0.03 1.82+ 0.03 1.71+ 0.04 NA Sample Sizeg NA 40 (20,20) 40 (20,20) 40 (20,20) 40 (20,20) 27 (17, 20) NA Body Weight (g) NA 0.557+ 0.036 0.851+ 0.053b 0.646+ 0.162 0.620+ 0.052 0.822+ 0.059b NA Standard Length (mm) NA 29.17+ 0.54 34.79+ 0.71b 29.69+ 0.85 30.27+ 0.66 33.09+ 0.81b NA Condition Factor (g/cm3) NA 2.25+ 0.03 2.02+ 0.02f 2.47+ 0.03f 2.24+ 0.02 2.27+ 0.04 NA *2000 fish were measured for length and weight after preservation. aNA = not available. bstatistically differs from U(8) (p < 0.05). cstatistically differs from REF1 (D(1) also from DIS) (p < 0.05). dstatistically differs from REF1 and DIS (p < 0.05). estatistically differs from REF1 and D(6) (p < 0.05). fstatistically differs from all other sites (p < 0.05). gsample sizes displayed as: total sample size (preserved anal fin measurements, hormone and com puter-aided meausurements)

PAGE 115

97 Figure 3-1. Maps of Rice Creek and Saint J ohns River, USA. A) Relative location in Florida. B) Field collec tion sites sampled before pro cess changes (2000). C) Field collection sites sampled after pr ocess changes (2002); asterisks indicate sites also collected in 2000. Site symbols distinguish sites exposed to effluent: circles = unexposed and tria ngles = exposed. Site symbols denote upstream (U) or downstream (D) of discharge, followed by approximate distance (km) from discharg e in parantheses. PRE-DI S indicates site before discharge into the creek; DIS denotes site at discharge into creek. A B

PAGE 116

98 Figure 3-1. Continued Figure 3-2. Representative male gonopodia from the upstream site collected before and after process changes. A) Before process changes (2000); preserved male photographed using 35 mm camera. B) After process changes (2002); fresh male photographed using digital camera Upstream site abbreviation is followed by that year’s average index of anal fin elongation represented by photograph. Notice terminal diffe rentiations on tips of gonopodia. U ( 8 ) – U ( 8 ) – A B 1 mm C

PAGE 117

99 SITE REF1U(8)PRE-DISDISD(1)D(3)D(6) # MALES IN 2000 0 2 4 6 8 10 12 1.4 mm 1.5 mm 1.6 mm 1.7 mm 1.8 mm 1.9 mm 2.0 mm 2.1 mm 2.2 mm 2.3 mm 2.4 mm SITE REF1U(8)PRE-DISDISD(1)D(3)D(6) # MALES IN 2002 0 2 4 6 8 10 12 2.1 mm 2.2 mm 2.3 mm 2.4 mm 2.5 mm 2.6 mm 2.7 mm 2.8 mm 2.9 mm 3.0 mm 3.1 mm A B * Figure 3-3. Male index of an al fin elongation (linear Ray 4 / Ray 6, manually measured on preserved fish) for each site by 0.1 mm increments. A) Before process changes (2000). B) After process cha nges (2002). Red asterisks indicate significant difference from upstream site [U(8)] within year (p < 0.05). Yellow boxes indicate significant differences between years.

PAGE 118

100 DIS –1.3 U(8) –1.1 D(3) –1.2 D(6) –1.1 D(1) –1.2 REF1 –1.1 1 mm U(8) –1.1 D(3) –1.1 PRE-DIS –1.2 D(1) –1.1 DIS –1.2 1 mm Figure 3-4. Representative female anal fins from collections made before and after process changes. A) Before pro cess changes (2000); preserved fish photographed using 35 mm camera. B) After process changes (2002); fresh fish photographed using digital camera. Fins listed by site abbreviation followed by average index of anal fi n elongation represented by photograph. A B

PAGE 119

101 SITE REF1U(8)PRE-DISDISD(1)D(3)D(6) # FEMALES IN 2000 0 2 4 6 8 10 12 14 16 18 20 1.0 mm 1.1 mm 1.2 mm 1.3 mm 1.4 mm 1.5 mm SITE REF1U(8)PRE-DISDISD(1)D(3)D(6) # FEMALES IN 2002 0 2 4 6 8 10 12 14 16 18 20 a b c d A B a Figure 3-5. Female index of anal fin elongation (linear Ray 4 / Ray 6, manually measured on preserved fish) for each site by 0.1 mm increments. A) Before process changes (2000). B) After pro cess changes (2002). Letters indicate statistically significant differences within year (p < 0.05): “a” denotes differences to nonlettered sites; “b” denotes diff erences to U(8) and D(3); “c” denotes differences to all sites except PRE-DIS; a nd “d” denotes differences to U(8). The yellow box surrounds sites with significan tly reduced anal fin elongation from 2000 to 2002.

PAGE 120

102 SITE REF1U(8)PRE-DISDISD(1)D(3)D(6) 17BETA-ESTRADIOL (pg/g) 0 100 200 300 400 500 600 2000 2002 SITE REF1U(8)PRE-DISDISD(1)D(3)D(6) TESTOSTERONE (pg/g) 0 200 400 600 800 1000 1200 WINTER 2000 a b A B c d Figure 3-6. Male whole body sex steroids (ave + se) from collections made before (2000) and after (2002) pro cess changes. A) 17 -estradiol. B) Teststosterone. Letters indicate significant differences by site within season (p < 0.05): “a” denotes differences to D(3); “b” denotes differences to PRE-DIS, D(1), D(3); “c” denotes differences to all other site s; “d” denotes differences to REF2, DIS and D(1). Yellow circles demonstrat e differences between years. Site and year did not covary for either hormone. Winter 2000 collection hormone values are given in grey for comparison.

PAGE 121

103 SITE REF1U(8)PRE-DISDISD(1)D(3)D(6) 17BETA-ESTRADIOL (pg/g) 0 200 400 600 800 1000 1200 2000 2002 SITE REF1U(8)PRE-DISDISD(1)D(3)D(6) TESTOSTERONE (pg/g) 0 100 200 300 400 500 600 700 WINTER 2000 a a a b A B Figure 3-7. Female whole body sex steroids (ave + se) from Rice Creek collections made before (2000) and after (2002) process changes. A) 17 -estradiol. B) Teststosterone. Letters i ndicate significant differen ces by site within season (p < 0.05): “a” denotes differences to nonlettered sites; “b” denotes differences to all but DIS. Yellow ci rcles demonstrate differences between years. Site and year did not cova ry for either hormone. Winter 2000 collection hormone values are given in grey for comparison.

PAGE 122

104 % FEMALES 0102030405060708090100 SITE U(8) PRE-DIS DIS D(1) D(3) MASCULINE STEROID RATIO (E:T<1) FEMININE STEROID RATIO(E:T>1) 1.13+ 0.02 1.19+ 0.01 1.27+ 0.03 1.32+ 0.01 1.34+ 0.04 1.33+ 0.04 1.12+ 0.04 1.20+ 0.02 1.19+ 0.02 1.20+ 0.02 Figure 3-8. Percentage of female mosquitofish with masculine and feminine sex steroid ratios collected after process changes in 2002 ( 2 = 8.270, df = 6, p < 0.05). Index of anal fin elongation (ave + se) is given for fish in each of these groups by site (no significant di fferences at p < 0.05).

PAGE 123

105 CHAPTER 4 VARIABLE EFFECTS OF EFFLUENT ON EASTERN MOSQUITOFISH COLLECTED BELOW THREE FLOR IDA PULP AND PAPER MILLS Masculinization, or development of male secondary sex characteristics, in female mosquitofish was considered equivalent am ong pulp and paper mills in Florida (FL) where this phenomenon was studied most exte nsively (Howell et al 1980, Drysdale and Bortone 1981, Cody and Bortone 1997, Bo rtone and Cody 1999, Jenkins et al. 2001, Parks et al. 2001). However, upgrades in pr ocessing technology have improved effluent characteristics, and reduction in response to upgrades at two of the three mills were documented (Cody and Bortone 1997, Chapter 3). In this study, field surveys within the same time period and using the same expe rimental design were conducted to assess differences in male and female mosquito fish responses among the three FL mills classically studied. Body size and anal fin morphology were measured, and sex steroids to provide insight into potent ial physiological disturbances. Female, and not male, anal fin morphology was significantly a ffected by effluent exposed s ite for all three systems. Despite high background and potential season al variation, hormone data indicated an androgenic or masculinized hormonal profil e in females and possible estrogenic or feminized hormonal profile in males. Degr ee of response in females mirrored relative concentrations of effluent components in wate r: greatest response and concentrations at Fenholloway River, intermediary responses a nd concentrations at Rice Creek, and lowest responses and concentrations at Elevenm ile Creek. These data support the use of

PAGE 124

106 mosquitofish as a bioindicator species to de tect extent of exposure and provided further characterization of seasonal changes in sex steroids. Introduction Sublethal effects of pulp and paper mill effl uent exposure on fish have been a major focus of aquatic environmental health concerns for over a decade (Sodergren 1991, Servos et al. 1996, Ruoppa et al. 2000, Stut hridge et al. 2003, Borton et al. 2004). Reported effects include inducti on of liver detoxification systems, alterations in sex steroid concentrations a nd production/metabolism, reduced gonadal development, decreased egg production, and decreased fry survival (Van der Kraak et al. 1992, Gagnon et al. 1994a, Munkittri ck et al. 1999, NCASI 2000a, Sepulve da et al. 2003, Parrott et al. 2004, McMaster et al. 2003). Whether or not these effects repres ent actual adverse effects in terms of reproductive success or population and community level impacts remains controversial. Equally debatable ar e potential mechanisms of action, since pulp mill effluents are a complex mixture and composition varies not only among mills but often within a mill. The discovery of carcinogeni c dioxins and furans, especi ally the polychlorinated congeners like TCDD and TCDF, in fish livi ng downstream of a pulp and paper mill in the 1980s initiated government re gulation of the industry to remove elemental chlorine and chlorinated compounds from effluent (Smook 1999). These chlorinated compounds were long considered the prim ary bioactive agents causing re ported sublethal effects in fish. Now that chlorine emissions were virt ually removed, a concomitant lack of effects in fish was not observed, although many eff ects were reduced (Lehtinen 2004, McMaster et al. 2003, and Chapter 3). Process change s implemented to remove chlorine also coincided with reduction in many other effluent components, especially wood extractives.

PAGE 125

107 Therefore, in recent years wood extractives such as resin acids, fatty acids, phytosterols, and lignin have become the focus as potential bioactive effluent components associated with noncarcinogenic sublet hal effects in fish. Phytosterol derivatives have been theoretically associated with fish effects for decades. In the laboratory, androgens we re formed by bacterial degradation of phytosterols and absorbed by mosquitofish (Denton et al. 1985, Howell and Denton 1989). One of the originally documented fish effects of pulp and paper mill effluents involved development of male secondary sex characteristics in female mosquitofish (Howell et al. 1980, Drysdale and Bortone 1981, Cody and Bortone 1997, Bortone and Cody 1999, Jenkins et al. 2001, Parks et al. 2001). Initial discovery was in fish from Elevenmile Creek, FL, a low flow minimal dilution stream (Howell et al. 1980), and reports of two other mills in FL that discharge into low flow streams reported equivalent masculinization (Drysdale and Bortone 1981, Cody and Bortone 1997, Bortone and Cody 1999). Yet these studies did not directly co mpare these systems in one comprehensive study and several processing improvements at all three mills over the past 20 years may have improved effluent quality and changed e quivalence of fish response. Other evidence supports a differential response among mills: r ecent whole-effluent exposure studies in Canada and New Zealand document large diffe rences in masculinization response time, from 3 wks to beyond 6 months (McCarthy et al. 2004, Van den Huev el et al. 2004b). Currently, variation in response at the mills in FL is likely but has not been specifically addressed. The objective of this study was to determ ine if FL mills still had equivalent masculinization responses in mosquitofish as previously reported. In light of processing

PAGE 126

108 upgrades and different processing technologies, masculinization was predicted to vary. Also, whole body sex steroids were analyzed by radioimmunoassay (RIA) to evaluate implications of the androgen-induced hypothe sis on steroid concentrations and ratios. Materials and Methods Mill Characteristics The three pulp mill effluent receiving systems in FL where masculinized female anal fins have been previously documented were surveyed within a three week period during the reproductively active seas on in July and August of 2001 ( Figure 4-1 Appendix A). Analogous site t ypes were sampled for each system: two reference sites; an upstream site; a site before discharge; th e discharge or outfall site; and at least one downstream site. Reference site s included sites with in the same basin as exposed sites, and sites within blackwater streams since all receiving streams were naturally tannic. Upstream sites were far enough upstream to prevent effects of backflushing from the effluent discharge, indicated by very low conductivity and low to no concentrations of effluent components ( Table 4-1 ). Predischarge sites comprised exposure to 100% effluent before discharge into the receiving stream, either in final retention ponds (Rice Creek and Elevenmile Creek) or the discharg e canal (Fenholloway River). Discharge sites were dominated by effluent (80 to 90 %), since dilution rates are low for all three systems. Downstream sites were at least one-third distance from discharge to mouth of the stream, where effluent was more d iluted (40 to 50% effluent maximum). The mills discharging into these systems are very different (Table 1-1, Chapter 1). In general, they differ in furnish, product, and secondary treatment. Two of the three mills are bleached kraft and were subject to EPA’s Cluster Rule: the Elevenmile Creek mill has been elemental chlorine free (ECF) since 1995, while the Rice Creek mill

PAGE 127

109 implemented major process improvements (i ncluding ECF technology) several months before fish collections in May. The Fenholloway River mill is much different than the other two, using a dissolving kraft pulping pr ocess to produce high grain cellulose. Thus it was not subject to many of the Cluster Rule requirements, such as ECF bleaching. Water Samples Before fish collection, water quality para meters typically affected by pulp and paper mill effluents were measured at each site: dissolved oxygen, temperature, pH, conductivity, salinity, and turbidity. Single grab wate r samples were also collected before fish collection to document potential exposure of fish to specific effluent components. Samples were preserved (buffe red), and sent to the National Council for Air and Stream Improvement, Inc. (NCASI) for chemical analysis. Water from all sites was analyzed for chlorinated phenolics (12 Cluster Rule compounds plus 16 others), 10 resin acids (including 3 chlorinated), 3 fatty acids, 4 phytosterols, total organic carbon, condensable tannins, and polyphenolics. Add itional effluent components were analyzed in 100% whole effluent: metals, nonmetals (suc h as chloride and fluoride), and neutral semivolatiles. Columbia Analytical Servi ces conducted the chlorophenolic analyses, CH2M Hill conducted the TOC analyses, a nd the NCASI West Coast Regional Center conducted all other analyses (NCASI 1986, 1997). Fish Samples Approximately 200 adult fish per site we re collected using dip nets and/or a backpack electroshocker. Fish designated for hormone analysis (20 to 30 each sex per site) were processed on-site. First, fish were euthanized with a terminal waterborne dose of buffered tricaine methanesulfonate (Trica ine-S, Western Chemical Inc., Ferndale, WA, USA), then weighed using a digital scale (+ 0.001 g) and measured for standard length (+

PAGE 128

110 0.01 mm) using a pair of digital calipers. U nder a dissecting scope, gender was identified using the presence (female) or absence (male) of a urogenital papi lla (see Chapter 2 for validation of this sexing technique). Each fish was photographed using a digital camera then placed on ice until tran sferred to a -80C freezer for subsequent radioimmunoassay (RIA) of sex steroids. Anal fin images of these fish were measured by computer (+ 0.01 mm) using trace mode in Sigma Scan Pro 5.0 from the base of Rays 4 and 6 along the curve of each ray to the tip. Remaining fi sh were euthanized and preserved in 10% neutral-buffered formalin for hist ological verification of gender. Sex Steroids Whole body primary sex steroids (17 -estradiol and testoste rone for this species) were analyzed using a modified RIA method originally developed for serum and plasma samples of common carp, Cyprinus carpio (Goodbred et al 1997), and since adapted for use in a variety of other aqua tic species and tissue media such as plasma of largemouth bass, Micropterus salmoides (Gross et al. 2001) and mantle of freshwater invertebrates (Gross et al. 2000). For methods and va lidation of this assay, see Chapter 2. Statistics Body weight and standard length were us ed to calculate condition factor, K = weight / length3 x 100 (g/cm3), as an indication of overall health used by the aquaculture industry (values at least 1 are considered healthy, Hile 1936). The length ratio of anal fin Rays 4 and 6 was calculated as an index of anal fin elongation. Estr ogen and testosterone concentrations were used to calculate a ratio indicating masculine hormone profile (E:T < 1) or feminine hormone profile (E:T > 1). Any data failing tests for normality and ho mogeneity of variance were transformed using log transformations. Originally arcsin e transformation was used for anal fin data

PAGE 129

111 (Noggle et al. 2004), but arcsine is most approp riate for ratio data ranging from 0 to 1 (Anderson and McLean 1974), and both anal fin elongation index and E:T ratio extend beyond 1.0. Also, log transformation was r ecommended by Angus et al. (2001) as appropriate for anal fin elongation index. Anal fin morphology and sex steroid data we re analyzed within sex using two-way analysis of covariance (ANCOVA) to test for significant variation by site and mill. Site differences within mill were also analy zed by one-way ANOVA, as were differences among mills by site type. Potential effects of size class was revisited and analyzed by one-way ANOVA within site. Significan t differences in ANC OVA and ANOVA were analyzed for multiple comparisons using Tukey’s HSD. Relationship between anal fin morphology and sex steroids were analyzed overall and by site in two ways: first, by examining Pearson’s correlations of the i ndex of anal fin elon gation to sex steroid concentrations and ratio, then by t-test fo r differences in index of anal fin elongation between females with masculine versus femini ne E:T ratios. Statis tical significance was set at < 0.05 for all tests. All statistical an alyses were conducte d using SAS version 9.0. Results and Discussion Water Quality As expected, conductivity, sa linity and turbidity were higher at effluent-exposed sites ( Table 4-1 ). Conductivity was highest in 10 0% effluent from the Fenholloway River (2,321 S) mill, followed by effluent from Rice Creek (1,916 S) and Elevenmile Creek (1,660 S). Temperature in effluent-dominated sites was also elevated compared to some, but not all, unexposed sites. Dissolved oxygen remained high enough to support fish at most sites in Rice Creek and Eleven mile Creek (> 4 mg/L), with the exception of

PAGE 130

112 the predischarge site for Rice Creek (2.14 mg/L). The Fenholloway River was poorly oxygenated along the entire leng th sampled regardless of effluent exposure: the predischarge site had more suitable dissolved oxygen levels to support fish. The fact that mosquitofish were present in this low oxyge n environment demonstrates their tolerance to extreme environmental conditions. Finally, pH was elevated in effluent-exposed sites compared to upstream sites; some references [REF2 for Rice Creek and REF1 for Fenholloway River] had similar pH. Notabl y, pH was very acidic at the upstream and second reference Fenholloway River sites [U(5 ) and REF2], likely reflecting the tannic, and blackwater nature of these systems. Water Chemistry Chemical analyses of single grab samples from the water column at fish collection sites distinguished the 100% final effluent before discharge site (highest chemical concentrations) from the discharge site (int ermediary concentrations) and unexposed sites (lowest concentrations) ( Table 4-2 ). Chlorinated compounds (chlorinated phenolics, chlorinated resin acids) were nondetectable or at the lower ca libration limits in all three systems (data not shown). Ions and heavy metals were within normal, acceptable ranges for pulp mill effluents: mg/L for ions (sodium calcium, fluoride, chloride, etc.) and g/L for metals (24 measured from aluminum to me rcury to zinc). Neut ral semivolatiles were at nondetectable concentrations in Rice Creek effluent samples; nondetectable concentrations in Elevenmile Creek effluent samples except 9.6 g/L dichlorodimethyl and 1.5 g/L 2,3,4,5-tetramethylcyclopentenoneu lfone; and 9 of 14 analytes (such as cyclopentophenones, acetophenone and camphor) averaged 7.6 g/L and ranged 1.6 to 31 g/L in Fenholloway River effluent samples. Essentially these latter data reveal the more

PAGE 131

113 odiferous nature of Fenholloway River efflue nt as opposed to an indication of increase effluent exposure. Reference and upstream sites had little to no measurable effluent components ( Table 4-2 ). Wood extractives (e.g. resin and fatty acids, phytosterols, polyphenolics (measuring tannin/lignin content), condens able tannins, and total organic carbon) decreased at downstream sites compared to predischarge sites for all systems ( Table 4-2 ), as expected with dilution. In accordance w ith conductivity measurements, most effluent components analyzed had an among mill trend of highest concentration in Fenholloway River effluent, intermediary concentration in Rice Creek effluent, and lowest levels in Elevenmile Creek effluent ( Table 4-2 ). Body Size and Condition No effect of effluent exposure on body size and condition, in terms of site, was detected for males or females across all three systems surveyed ( Table 4-3 ). Rather, variability in length and weight was associ ated with unexposed sites and demonstrated the importance of multiple reference streams for these more variable endpoints compared to female anal fin elongation (Chapters 2 and 3 and discussion below). Condition factor was sufficient for both sexes (> 1) and indi cated all fish were in good general health. Males Male body size was not affected by effluent -exposed site for any of the three systems, especially considering the va riation detected am ong unexposed sites ( Table 4-3 ). For the Rice Creek system, smaller males o ccurred at the second reference site [REF3] and the predischarge site [PRE-DIS]. However, male condition was > 1 indicating adequate general health for a ll sites (statistically greater at the upstream site although probably not biologically signifi cant). Males collected from the Elevenmile Creek and

PAGE 132

114 Fenholloway River systems were of equal body length across all sites, but body weight and condition factor were lowest at the s econd reference site [both REF2] compared to predischarge and discharge sites [PRE-DIS a nd DIS for both]. Condition factors were all above one and healthy for these males. Overall, male body size was not affected by effluent. Females Variation in body size related to reference site was detected in females among all three systems, as opposed to effect related to effluent exposed sites ( Table 4-3 ). In the Rice Creek system, the largest females (length and weight) were collected before discharge into Rice Creek [PRE-DIS]. The shortest females were collected from the outfall and first downstream sites [DIS and D( 1)]. However, condition factor was good (> 1) across all site s with the only statistical differe nces observed between the second reference site [REF3] and the first reference site [REF2] as well as the discharge site [DIS]. At the Fenholloway River system, si gnificant variation in body size occurred for the unexposed sites but no significa nt effects associated with exposed site were observed. Condition factor indicated good ge neral health (> 1) across all sites. Females from the Elevenmile Creek system were of equal le ngth and weight acro ss sites, and the only statistical difference in condition factor was observed at the second re ference site [REF2] compared to the other sites. As for the othe r systems, condition fact or was consistently < 1. Anal Fin Morphology Female, and not male, anal fin morphology was significantly affected by effluent exposed site for all three systems ( Figure 4-2 ). Precocious maturation in males was not supported, and natural variation as opposed to effluent-associated vari ation was detected.

PAGE 133

115 Female anal fin elongation was reduced compared to historical report s indicating terminal differentiations; overall elongation resembled a developing male gonopodium as opposed to a mature male gonopodium. Using water chemistry data as a mark of potential exposure ( Table 4-2 ), the greatest exte nt of elongation coincided with highest concentrations of wood extr actives (Fenholloway River); intermediary elongation with medium amounts of extractives (Rice Creek); and the smallest response with the lowest concentrations of extractives (Elevenmile Cr eek). Examining female anal fin elongation by size class did not consistently support the concept of a sensitive (adult) life stage; rather, differential or dynamic exposure was supported by these data. (Dynamic exposure refers to variable concentrations of effl uent components over time, dependent on factors such as tree species for furnish, within pl ant processing spills, rainfall/dilution, and bacterial degradation.) However genetic differences among these three mosquitofish populations can not be ruled out as an explanation for diffe rent responses by size class across mills. Males Males did not appear influenced by effl uent-exposed site in the 2001 collections ( Figure 4-2B ). No significant differences exis ted among males in the Rice Creek or Fenholloway River systems. At the Eleven mile Creek system, gonopodia were longer at 100% whole effluent and outfall sites [PRE-DIS and DIS] compared to one reference site [REF2] but not compared to the other unexposed sites [REF1 and U(1)]. Site and mill did not covary, but among mills Rice Creek males had significantly longer gonopodia at the discharge site [D] and the upstream site [U] compared to the other two mills. Variation was high and statisti cally significant among the thr ee mills’ reference sites as

PAGE 134

116 well. Again, this is evidence for a natural variability in this response or environmental stressors on males as opposed to abnormalities in effluent exposed males. Males were also analyzed by size class w ithin each system to address precocious maturation. Three size classes were identifie d (5 to 10 males per group): < 20 mm; 20 to 24.99 mm; and 25 to 29.99 mm. If precocious maturation was occurring, males in the smallest size class would exhibit abnorm ally long gonopodia similar in elongation to males in the larger size classes. Anal ysis by size class alone at Rice Creek and Fenholloway River field sites found significan tly smaller gonopodia in males from the < 20 mm class (2.37 + 0.05) compared to the 20 to 24.99 mm class (2.52 + 0.06), as expected under normal conditions. At Eleven mile Creek sites, the 25 to 29.99 mm class had a larger average (2.72 + 0.09) than the two smaller classes (2.07 + 0.10 and 2.02 + 0.11, respectively) which did not differ from each other. Direct analysis of standard length to gonopodial length (Ray 6) did not reveal any site -specific differences within mill (data not shown). Therefore, in this study there was no evidence for precocious maturation in males living in effluent-receiving streams. Females For all three systems, anal fin elongation in masculinized females never approached the male gonopodium in length or terminal di fferentiation. In term s of length, altered females from exposed sites averaged an index of 1.5, while normal males and normal females from unexposed sites averaged indice s of 2.5 and 1.1, respectively. No terminal structures (hooks, serrae, or blade) were observed on the ti p of any altered female, in contrast to historical collections of mosquito fish at these same sites (Howell et al. 1980, Drysdale and Bortone 1981, Cody and Bort one 1997). This lack of terminal

PAGE 135

117 differentiation indicated a reduction in respons e since process modifi cations have been implemented over the years. Anal fin elongation in female mosquitofi sh was detected at all three systems ( Figure 4-2A ). At Rice Creek, elongation was si gnificantly different between reference sites [R1 and R2] and effluent-exposed sites [P -D, D, D1 for R1 and only P-D to R2], but not between upstream [U] and exposed sites. Females from effluent-exposed sites along the Fenholloway River had significant anal fin elongation compared to nonexposed sites. Interestingly, elongation was great est at the farthest downstream site. Site fidelity in this species and the large distance between the disc harge and this site (~12 km) mean flushing of fish from upstream was not likely, thus le nding more support to an additional factor(s) in the receiving stream responsible for the observed response. Elevenmile Creek females also displayed anal fin elongation when data was log transformed as suggested by Angus et al. (2001). Females from the predischar ge and discharge site s [P-D and D] had significantly longer anal fin el ongation than females from the upstream site [U]. Interaction of site and mill significantly covaried. Among mills, Fenholloway River had the greatest degree of elongation for all effluent-e xposed sites, Rice Creek was in between the other two before and at the discharge [P-D and D], and Elevenmile Creek consistently had the least degree of elonga tion among exposed sites. This pattern mirrored that of wood extractives ( Table 4-2 ); therefore concentrations of these compounds may be useful in predicti ng degree of response (or vice versa). Effects of size class as an estimation of age was re-examined for these collections, since 2000 collections at Rice Cr eek (Chapter 2) indicated di fferences among size classes at the discharge site. In 2001, size class differe nces were again detected at the discharge

PAGE 136

118 site: the smallest size class (20 to 24.99 mm) had longer anal fin elongation than the other classes (25 to 29.99 mm and 30+ mm). This re sult was slightly different than what was observed in fall 2000 data, when females in the middle two size classes (25 to 29.99 mm and 30 to 34.99 mm) had significan tly longer anal fins compared to the largest size class (35 to 39.99 mm). Also different than 2000 da ta, in 2001 significantly longer elongations (0.1 mm longer) were detected in the smalle st size class (20 to 24.99 mm) versus the largest (30+ mm) at the first reference and upstream sites [REF2 and U(8)]. These data combined with the overall anal fin elongati on differences among sites described above for Rice Creek supports a background incidence of elongation in female mosquitofish and decreases the apparent specificity of this tr ait for use as a bioindicator of pulp and paper mill effluents. Size class analysis at the other two system s produced different results. Anal fin elongation did not vary by size class within sites at Fenholloway Ri ver, while size class was significantly influential at the predischarge site only at Elevenmile Creek. In this case, the largest class (30+ mm) had signifi cantly longer anal fin elongation compared to the smallest class (20 to 24.99 mm). Taken as a whole, these data on size class differences among mills do not support the idea of sensitive life stages among adult females; rather, they indicate a dynamic e xposure and/or the possibility of genetic differences among populations. Th e Rice Creek data also reveal ed a natural variation of the masculinization response at unexposed sites which had not been demonstrated so overtly in previous collec tions (Chapters 2 and 3). Sex Steroids Seasonal differences among systems were i ndicated by concentrations of individual steroids for both sexes. Additional envir onmental factors influencing steroid levels

PAGE 137

119 besides effluent exposure were also implied by different resp onse patterns among effluent exposed sites. 17 -estradiol was elevated in male s from Rice Creek and Fenholloway River, and the E:T ratio was estrogen-biased for a small number of these males, providing the first preliminary evidence for an estrogenic effect of pulp mill effluent on this species. Weak evidence of increased testosterone in females from Rice Creek and Fenholloway River was apparent, and boosted by an increased frequency of masculinized steroid ratios for these sites. Based upon steroid data among these mills, pulp mill effluents may result in estrogenic action on males and androgenic acti on on females at the physiological level. Perhaps more strongly, these data also stre ss the need for seasonal characterization of hormone levels in this species to allo w for a more conclusive interpretation. Males 17 -estradiol in males was significantly elev ated at the first downstream site for both Rice Creek and Fenholloway Rive r compared to all other sites ( Figure 4-3A ). Males living upstream of effluent discharge in El evenmile Creek also had significantly higher 17 -estradiol concentrations compared to all expo sed sites [P-D, D, D1]. In light of data from 2000 and 2002 for Rice Creek (Chapters 2 a nd 3), these contrasting differences may represent different seasonal stages among th ese fish populations. Values ranged from less than 100 pg/g similar to 2000 data, averaged around 500 pg/g similar to 2002 data, but ranged from 1000 to 2000 pg/g 17 -estradiol for Rice Cree k and Fenholloway River males (the highest recorded up to that point). Combined with the knowledge of (potential) exposure differences ( Table 4-2 ), one may speculate elevated estrogen in males as a potential effect of effluent expos ure in addition to the seasonal variation and dynamic exposure already implied.

PAGE 138

120 Similar to previous data on Rice Creek males (Chapters 2 and 3), individual testosterone concentrations vari ed greatly among nonexposed sites ( Figure 4-3B ). Statistically significant depres sion in testosterone at Fenhol loway River and Elevenmile Creek depended upon which unexposed site wa s used as reference, precluding strong conclusions about effects of e ffluent exposure on concentrations of testosterone in males. The vast majority of E:T ratios in male mosquitofish (95%) were normal and testosterone-biased with the exception of less than 5% of males that had estrogen-biased ratios. Half of these hormonally feminized males were from the downstream site on Fenholloway River [D1] where the average E:T ratio was 1.03+ 0.14. One to two were from the other two effluent-exposed sites of the Fenholloway system; one from an effluent-exposed site of the Elevenmile Creek system; two each from effluent-exposed sites; and one from a reference site of the Rice Creek system. Ratios other than at the downstream Fenholloway site were within th e tenths range (mean of 0.10 to 0.51 and standard error (se) of 0.01 to 0.09), similar to 2002 Rice Creek males as opposed to 2000 males. Considering the water chemistry data and the 17 -estradiol absolute concentrations it is possible the feminized ratio at downstream Fenholloway is an indication of an estrogenic effect on males. Since the ratio was normal before discharge and at the discharge of the Fenholloway River, additional environmental factors influencing this response (e .g. differential bacterial degr adation of phytosterols) are supported by these data. Females 17 -estradiol concentrations in female mosquitofish were not influenced by effluent exposure in summer 2001. Concentrati ons were within ranges reported for Rice

PAGE 139

121 Creek in 2000 and 2002 (Chapters 2 and 3), even the one significantly elevated peak at the upstream Fenholloway River site compared to all other sites for that system (see Figure 4-4A ). Site and mill did not significantly covary, i.e. there was no interaction between the two variables. 17 -estradiol varied among mills at two sites including the upstream site; therefore these differences are lik ely attributed to diffe rences in habitat and seasonality among regions as opposed to effluent-related differences. Weak evidence for elevated testosterone existed for females at Elevenmile Creek and Fenholloway River ( Figure 4-4B ). No significant differences in testosterone concentrations were detected in females from Rice Creek, despite the peak at the discharge site (which had the la rgest variation of all sites). Concentrations were within ranges reported for Rice Creek in 2000 and 2002 (Chapters 2 and 3). Site and mill did not covary for this hormone in females, although significant mill differences were detected for all site types. Testosterone was statistically elevated at Fenholloway River sites compared to the other two systems for both exposed and nonexposed sites. Similar to 17 -estradiol in females, these differences may reflect seasonality as opposed to effluent-related effects. Average E:T ratios were all above one and implied normal feminine hormonal profiles, although variation was high (average 10 + 2.8 standard error). Plotting the frequency of masculine versus feminine hor monal profiles for females revealed unique patterns for each system and a higher occu rrence of skewed pr ofiles than males ( Figure 4-5 ). Rice Creek females had signifi cantly different ratios among sites ( 2 = 27.95, df = 6, p < 0.05) with a skew toward masculinized profiles at the effluent-exposed sites. However, frequency was not as high as fo r 2002 females (Chapter 3), perhaps caused by

PAGE 140

122 frequent mill shutdowns as Cluster Rule pr ocess changes were being implemented that season (pers. obs.). Masculinized hormone profiles were more common in the Fenholloway River system at both effl uent exposed and unexposed sites ( 2 = 14.31, df = 5, p < 0.05), even the pristine Econfina River [R 1 and R2]. Skewed steroid ratios at the reference sites for this system are very impor tant and imply the E:T ratio can be altered by environmental factors completely separate from pulp and paper mill effluents. In contrast to the other two systems, masculini zed steroid ratios in females from Elevenmile Creek system occurred at very low levels (i n one or two females pe r site) and incidence was not significantly di fferent among sites ( 2 = 0.8064, df = 5, p > 0.05). Overall, frequency of masculinized hormone profiles in effluent exposed fish resembled the stronger trend among mills for anal fin elongation and was supported by relative concentrations of effluent components. Anal Fin Elongation and Sex Steroids Despite associations between these two biomarkers and effluent exposed sites in females, females with elongated anal fins were just as likely to have a masculinized hormonal profile as a normal female profile. La ck of demonstrated e ffect on anal fins in males and the low occurrence of feminized hormonal profiles essentially precluded comparison of these endpoints in males. Males Because of the low occurrence of feminized hormonal profiles for males (< 5%), statistical analysis was not robust when this group was compared to the dominant, normal masculinized profile. In the Fenholloway River and Elevenmile Creek systems, no correlation existed for the male index of anal fin elongation to estr ogen, testosterone, or the E:T ratio (r2 < 0.1 and p > 0.05 for correl ations). For Rice Cr eek males, statistical

PAGE 141

123 significance (p<0.05) was attain ed between the index and 17 -estradiol and testosterone separately but not as a ratio: correlation co efficients were low and indicated a weak negative correlation (r2 = -0.182 and r2 = -0.255, respectively). T-tests between masculinized and feminized hormone profile s of males were equally inconclusive because of small sample sizes for hormonally femi nized males. At this point, the lack of demonstrated effect on anal fin morphol ogy coupled with the low occurrence of feminized hormone profiles in males does not suggest a relationship between these biomarkers. Females Although the frequency of masc ulinized hormone profiles in effluent exposed fish resembled the stronger trend among mills for anal fin elongation, statistical comparison of these two biomarkers in these same fish di d not reveal any relations hip. This result is consistent with results obtained in the surv ey of mosquitofish before and after process changes at the Rice Creek system (Chapter 3) No correlation existed for the index of anal fin elongation to estrogen, te stosterone, or the E:T ratio (r2 < 0.5 and p > 0.05). Across all sites and within sites average el ongation did not differ between females with masculine and feminine hormone ratios ( Figure 4-5 ). Therefore, females with normal feminine hormone profiles were just as likely to have masculinized anal fins as females with masculine hormone profiles. This does not rule out alterations in sex steroid ratios contributing to elongation of the anal fin, since sex steroids were measured after onset of elongation. Seasonal changes in sex steroids, but not anal fin elongation, support this point (Chapter 2). However, presence of th e altered hormonal profile cannot be used to

PAGE 142

124 predict occurrence of anal fin elongation in individual females living in effluent-receiving streams, based upon these data. Conclusions Mosquitofish responded differentially to pulp and paper mill effluent exposure at three mills in FL, contrary to historical reports of similar responses in females at the mills investigated. With the adde d support of water chemistry da ta, anal fin and sex steroid biomarkers in females displayed a graded a ndrogenic response among mills: the greatest response at Fenholloway River, an intermedia ry response at Rice Cr eek, and the lowest response at Elevenmile Creek. Males do not appear affected by effluent exposure in terms of anal fin morphology, while a weak estrogenic response was indicated by changes in the E:T ratio. These gender-spe cific endocrine disruptive effects have also been reported for fathead minnows ( Pimaphales promelas ) under controlled exposure to Canadian pulp mill effluents (Parrott and Wood 2002, Parrott et al. 2003, 2004). Whether or not these effects lead to adverse impacts on reproductive success and concomitant impacts on higher levels of bi ological organization remains to be seen (Chapter 6). Several limitations to this data set make these conclusions tent ative and expose the need for basic biological data in toxicological studies. The inherent natural variation of these endpoints, especially sex steroids, within unexposed sites reduces the specificity of response for pulp mill effluents. Further co mplicating interpretation is the lack of baseline data on seasonality of steroid levels in this species. Without such a benchmark, a formal conclusion relating observed effects to pulp mill effluents remains elusive. A third major limitation was the single time poi nt water sampling for chemical analysis. Although an improvement over the studies repor ted in Chapters 2 and 3, these samples

PAGE 143

125 represent only a snapshot of continually changing effluent concentrations and composition. For example, one such factor that may change effluent concentrations over time is precipitation, especially for these low flow systems (Appendix A). Fish from Elevenmile Creek and Fenholloway River we re under drought conditions the year of collection, while Rice Creek fi sh experienced normal yearly rainfall. If a dynamic exposure scenario proves crucial to interpre tation of effects (see Chapter 3 for a more detailed discussion), documenting chemi cal exposure has to occur over time. Despite these limitations, the mosquitofish remains promising as a bioindicator species of pulp mill effluents. Differences in responses among mills coincided with effluent components meaning this species coul d potentially be used to document varying effluent quality. However, contribution of additional environmental factors is supported by the increased response detected further downs tream and not at sites of highest effluent concentration within a mill. A threshold res ponse is also supported by these data, and the apparent disparity between anal fin morphology and sex st eroid response may be quite useful to determine the threshold in future research. Controlled w hole effluent exposures are necessary to adequately document this po tential threshold (see Chapter 5). Recent work in Canada and New Zealand has shown wide variation in induction time for female anal fin elongation that does not correlate with phytosterols concentrations in effluent (McCarthy et al. 2004). This complexity s upports the contribution of additional factors within receiving streams which re quires further investigation. If the phytosterol bacterial degradation hypothesis is correct the additional factor may be a change in bacterial communities that transform phyt osterols into bioactive co mpounds at varying rates and efficiencies. While the mechanism behind observed responses requires more extensive

PAGE 144

126 and well-coordinated resear ch, effects-based monitori ng using species such as mosquitofish may be the best solution in the meantime.

PAGE 145

127 Table 4-1. Water quality parameters of fi eld collection sites associated with three effluent-receiving streams in Florida the summer of 2001. Site Typea R1 R2 U P-D D D1 D2 Rice Creek REF2 REF3 U(8) PRE-DIS DIS D(1) D(3) Temperature (C) 24.5 25.7 24.1 30.2 27.7 27.6 30.4 Conductivity (S) 143.3 330.7 227 1916 1580 1417 1185 Salinity (ppt) 0.1 0.2 0.1 1.0 0.8 0.7 0.6 Dissolved Oxygen (mg/L) 5.49 5.76 5.71 2.14 5.69 13.3 6.45 Turbidity (ntu) 2.02 4.29 15.3 13.5 20.4 12.7 5.76 pH 7.7 6.6 6.4 7.8 7.6 7.6 7.5 Fenholloway River REF1 REF2 U(5) PRE-DIS DIS D(12) NCb Temperature (C) 23.7 27.6 26.6 31.6 27.8 27.1 NAc Conductivity (S) 225 67.5 83.6 2321 1158 1336 NA Salinity (ppt) 0.1 0.0 0.0 1.0 0.3 0.7 NA Dissolved Oxygen (mg/L) 5.20 4.71 2.22 4.48 3.27 2.35 NA Turbidity (ntu) 5.73 2.92 1.1 39.4 19.8 15.3 NA pH 7.8 4.5 3.9 7.4 7.1 7.5 NA Elevenmile Creek REF1 REF2 U(1) PRE-DIS DIS D(5) NCb Temperature (C) 24.3 24.0 25.9 0.3 28.3 26.3 NAa Conductivity (S) 50.3 32.8 72.7 1660 1135 432.5 NA Salinity (ppt) 0.0 0.0 0.0 0.8 0.6 0.2 NA Dissolved Oxygen (mg/L) 5.46 7.71 7.74 4.09 5.09 6.13 NA Turbidity (ntu) 7.87 3.33 7.46 22.6 49.8 15.5 NA pH 5.9 6.2 6.2 8 7.8 7.1 NA aR# = first or second reference; U=upstream ; P-D=predischarge retention pond or canal; D = discharge or outfall; D#= first or second downstream bNC = not collected for this system

PAGE 146

128 Table 4-2. Concentration of selected effluent components in single grab water samples from field collection sites associated with three effluent-receiving streams in Florida the summer of 2001. Site Typea R1 R2 U P-D D Rice Creek REF2 REF3 U(8) PRE-DIS DIS Total RAFAa (g/L) 0 0 15 100 18 Campesterol (g/L) NDc ND ND 1.2 ND Stigmasterol (g/L) ND ND ND 4.6 1.7 Stigmastanol (g/L) ND ND ND 8.4 1.8 -sitosterol (g/L) ND ND ND 28.2 4.8 TOCb (mg/L) 4.1 10.4 18 69.7 24.9 polyphenolics (mg/L) 1.6 3.2 2.6 25 6.0 condensable tannins (mg/L) 0.6 1.0 0.7 3.5 1.2 Fenholloway River REF1 REF2 U(5) PRE-DIS DIS Total RAFAa (g/L) 4 5 6 326 102 Campesterol (g/L) ND ND ND 3.4 1.8 Stigmasterol (g/L) ND ND ND 10.1 7.9 Stigmastanol (g/L) ND ND ND 14.3 7.6 -sitosterol (g/L) ND ND ND 70.6 40.8 TOCb (mg/L) 73.2 55.9 107.7 164.7 83.5 polyphenolics (mg/L) 12 8.1 17 39 24 condensable tannins (mg/L) 5.1 4.0 2.5 9.5 5.1 Elevenmile Creek REF1 REF2 U(1) PRE-DIS DIS Total RAFAa (g/L) 3 9 3 5 2 Campesterol (g/L) ND ND ND ND ND Stigmasterol (g/L) ND ND ND 2.6 ND Stigmastanol (g/L) ND ND ND 4.2 1.5 -sitosterol (g/L) ND ND 0.8 2.5 ND TOCb (mg/L) 18 ND ND 31.8 24.9 polyphenolics (mg/L) 1.5 0.2 1.3 5.6 2.3 condensable tannins (mg/L) 0.7 0.6 0.7 5.6 0.9 aRAFA = resin acids and fatty acids bTOC = total organic carbon cND = nondetectable

PAGE 147

129Table 4-3. Body size parameters (ave + se) for mosquitofish collected from three effl uent-receiving streams in Florida the summer of 2001. Significant differences (p < 0.05) are noted by site within each system. Site typea R1 R2 U P-D D D1 D2 Rice Creek REF2 REF3 U(8) PRE-DIS DIS D(1) D(3) Sample Size 30 20 28 21 21 24 24 Body Weight (g) 0.235+0.021d 0.148+0.013 0.198+0.016 0.155+0.012 0.167+0.010 0.177+0.017 0.172+0.015 Standard Length (mm) 22.72+0.62 19.75+0.57e 22.49+0.61 19.68+0.54e 20.69+0.46 21.24+0.61 20.97+0.56 Condition Factor (g/cm3) 1.87+0.04 1.83+0.06 2.64+0.11f 1.97+0.04 1.87+0.06 1.73+0.03 1.76+0.03 Sample Size 53 55 55 50 34 51 55 Body Weight (g) 0.489+0.042 0.473+0.030 0.508+0.035 0.639+0.062g 0.336+0.028 0.444+0.045 0.531+0.048 Standard Length (mm) 27.65+0.68 26.90+0.47 28.00+0.52 29.45+0.90g 24.74+0.61h 26.58+0.63h 27.83+0.75 Condition Factor (g/cm3) 2.02+0.04 2.28+0.04i 2.13+0.04 2.19+0.04 2.09+0.11 2.13+0.09 2.10+0.03 Fenholloway River REF1 REF2 U(5) PRE-DIS DIS D(12) NCb Sample Size 20 17 20 24 23 20 NAc Body Weight (g) 0.152+0.012 0.127+0.012j 0.146+0.007 0.1.94+0.019 0.192+0.019 0.148+0.010 NA Standard Length (mm) 19.63+0.53 19.74+0.49 19.79+0.27 21.32+0.64 21.09+0.63 19.53+0.45 NA Condition Factor (g/cm3) 1.94+0. 07 1.59+0.06f 1.87+0.06 1. 88+0.04 1.93+0.08 1.93+0.04 NA Sample Size 44 46 36 55 46 55 NA Body Weight (g) 0.347+0.037 0.450+0.046 0.200+0.016f 0.445+0.032 0.364+0.022 0.441+0.024 NA Standard Length (mm) 24.18+0.55k 26.96+0.83 21.62+0.56l 26.51+0.57 25.29+0.51 26.66+0.48 NA Condition Factor (g/cm3) 2.32+0. 23 1.97+0.03m 1.89+0.05n 2.25+0.05 2.15+0.0 5 2.22+0.03 NA Elevenmile Creek REF1 REF2 U(1) PRE-DIS DIS D(5) NCa Sample Size 20 20 23 23 24 23 NA Body Weight (g) 0.150+0.008 0.144+0.008o 0.199+0.019 0.205+0.013 0.193+0.014 0.187+0.014 NA Standard Length (mm) 19.57+0.34 20.15+0.34f 20.76+0.59 21.16+0.50 21.18+0.50 21.06+0.47 NA Condition Factor (g/cm3) 1.98+0. 04 1.74+0.04 2.08+0.04 2.14 +0.05 1.97+0.04 1.93+0.05 NA Sample Size 55 42 55 54 55 55 NA Body Weight (g) 0.539+0.037 0.431+0.044 0.505+0.036 0.506+0.043 0.461+0.026 0.435+0.028 NA Standard Length (mm) 27.69+0.60 26.82+0.77 27.37+0.54 26.96+0.75 26.86+0.49 26.16+0.55 NA Condition Factor (g/cm3) 2.39+0. 09 1.98+0.04f 2.27+0.03 2. 28+0.03 2.25+0.02 2.27+0.03 NA aR# = first or second reference; U = upstream; P-D = predischarge retention pond or canal; D = discharge or outfa ll; D# = first or second downstream bNC = not collected cNA = not available statistically differs from: dREF1, PRE-DIS, DIS eREF2 and U(8) fall sites gDIS and D(1) hU(8), PRE-DIS, D(3) iREF2 and DIS jPRE-DIS and DIS kREF2 and D(5) lall sites except REF1 mPRE-DIS and D(12) nall sites except REF2 oDIS

PAGE 148

130 Figure 4-1. Maps of field sites. A) Relative location of the stream systems in Florida. B) Rice Creek sites. C) Fenholloway River sites. D) Elevenmile Creek sites. Symbols distinguish sites exposed to e ffluent: circles = unexposed sites and triangles = exposed sites. Site ab breviations denote upstream (U) or downstream (D) of discharge, followed by approximate distance (km) from discharge in parentheses. PRE-DIS indicates site be fore discharge into the creek; DIS denotes site at discharge into creek; a nd REF indicates reference site, followed by identifying number. A B B C D

PAGE 149

131 Figure 4-1. Continued C D

PAGE 150

132 SITE TYPE R1R2UP-DDD1D2 INDEX OF ANAL FIN ELONGATION 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Rice Creek Fenholloway River Elevenmile Creek SITE TYPE R1R2UP-DDD1D2 INDEX OF ANAL FIN ELONGATION 0.0 0.5 1.0 1.5 2.0 2.5 3.0 A B f a b c e c d* Figure 4-2. Index of anal fin elongation (t racings of Ray 4 / Ray 6, computer-aided measurement of fresh fish) for mosqu itofish collected in summer 2001 from three effluent-receiving systems in Florida. A) Females. B) Males. Letters indicate significant differences by si te within system (p<0.05): “a” denotes differences to P-D, D and D1 (Rice Creek system); “b” denotes differences to P-D (Rice Creek system); “c” denote s differences to nonlettered sites (Fenholloway River system); “d” denot es differences to all other sites (Fenholloway River system); “e” de notes differences to P-D and D (Elevenmile Creek system); “f” denotes differences to P-D and D (Elevenmile Creek system). Yellow circles signify statistically significant differences among mills (p < 0.05). Green asterisk marks interaction between site and mill for females only.

PAGE 151

133 SITE TYPE R1R2UP-DDD1D2 17BETA-ESTRADIOL (pg/g) 0 500 1000 1500 2000 2500 RICE CREEK FENHOLLOWAY RIVER ELEVENMILE CREEK SITE TYPE R1R2UP-DDD1D2 TESTOSTERONE (pg/g) 0 1000 2000 3000 4000 A B b a c e d f f Figure 4-3. Whole body sex steroids (ave + se ) for male mosquitofish collected from three effluent-receiving streams in Florida the summer of 2001. A) 17 estradiol. B) Testosterone. Letters indicate significant differences by site within mill (p < 0.05): “a” and “b” denote differences to all other sites; “c” denotes differences to P-D, D, D1; “d” de notes differences to R1 which is also different to U; “e” denotes differences to P-D; “f’ denotes differences to P-D (and R2 also different to D). Yellow ci rcle indicates significant differences among mills at that site. Site type an d mill did not covary for either hormone.

PAGE 152

134 SITE TYPE R1R2UP-DDD1D2 17BETA-ESTRADIOL (pg/g) 0 200 400 600 800 1000 1200 1400 RICE CREEK FENHOLLOWAY RIVER ELEVENMILE CREEK a SITE TYPE R1R2UP-DDD1D2 TESTOSTERONE (pg/g) 0 100 200 300 400 500 600 700 800 b c A B Figure 4-4. Whole body sex steroids (ave + se) for female mosquitofish collected from three effluent-receiving streams in Florida the summer of 2001. A) 17 estradiol. B) Testosterone. Letters indicate significant differences by site within mill (p < 0.05): “a” denotes diffe rences to nonlettered sites; “b” denotes differences to R2; “c” denotes differences to R1, D, and D1. Yellow circles indicate significant differences among mills at th at site. Site type and mill did not covary for either hormone.

PAGE 153

135 % FEMALES 0102030405060708090100 RICE CREEK SITE TYPE R1 R2 U P-D D D1 D2 MASCULINE STEROID RATIO (E:T<1) FEMININE STEROID RATIO(E:T>1) 1.43+ 0.02 1.20+ 0.03 / 1.16+ 0.02 1.07+ 0.01 / 1.12+ 0.01 1.16+ 0.05 / 1.17+ 0.03 1.23+ 0.04 / 1.19+ 0.02 1.20+ 0.05 / 1.27+ 0.05 1.15 / 1.13+ 0.02A % FEMALES 0102030405060708090100 FENHOLLOWAY RIVER SITE TYPES R1 R2 U P-D D D1 B1.53+ 0.07 / 1.52+ 0.03 1.37+ 0.03 / 1.37+ 0.02 1.41+ 0.03 / 1.40+ 0.03 1.16 / 1.19+ 0.04 1.17+ 0.03 / 1.14+ 0.02 1.18+ 0.02 / 1.14+ 0.02 Figure 4-5. Percentage of female mosquitofish with masculine and feminine sex steroid ratios collected from three effluent-r eceiving streams in Florida the summer of 2001. Index of anal fin elongation (ave + se) is given for fish in each of these groups by site (no significant differences at p < 0.05). A) Rice Creek ( 2 = 27.95, df = 6, p < 0.05). B) Fenholloway River ( 2 = 14.31, df = 5, p < 0.05). C) Elevenmile Creek ( 2 = 0.8064, df = 5, p < 0.05).

PAGE 154

136 % FEMALES 0102030405060708090100 ELEVENMILE CREEK SITE TYPES R1 R2 U P-D D D1 C1.17 / 1.12+ 0.12 1.08 / 1.21+ 0.08 1.05 / 1.21+ 0.08 1.07 / 1.08+ 0.01 1.30 / 1.18+ 0.014 1.32 / 1.22+ 0.02 Figure 4-5. Continued

PAGE 155

137 CHAPTER 5 DIFFERENTIAL INDUCTION OF EFFECT S IN MOSQUITOFISH EXPOSED TO BLEACHED KRAFT MILL EFFLUENT Masculinization, or development of male secondary sex charact eristics in female mosquitofish, has been detected in pulp and paper mill effluent-receiving streams for decades. While laboratory exposure to bacterially-degraded phytosterols, the hypothesized mechanism, has repl icated masculinization of females, controlled exposure to whole effluent dilutions has produced incons istent masculinization results. This study compared and contrasted short-term (4 week ) whole effluent expos ures in flow-through tanks (0, 10, 20, 40, and 80% effluent dilutions) against in situ exposure of caged fish at field sites (upstream, before discharge, and di scharge sites). While the anal fin Ray 4 to Ray 6 length ratio was not affected in e ither sex, whole body sex steroid alterations suggested either dynamic exposure and/or multiple modes of action along the reproductive-endocrine axis. Specifically, males expressed feminized estrogen to testosterone (E:T) ratios and females expressed masculinized ratios at midpoint sampling of field-exposed fish followed by partial or co mplete recovery of normal profiles at final sampling. Skewed steroid ratios did not ma nifest in tank-exposed fish until final sampling. Hormonal effects were detected at effluent dilutions with in yearly instream effluent concentrations (about 60%). Full ch aracterization of seasonality in hormones; intensive documentation of exposure; and th e study of additional environmental factors such as bacterial degradation of effluent components into bi oactive compounds would greatly improve strength of results.

PAGE 156

138 Introduction Masculinized mosquitofish collected in pulp and paper mill effluent receiving streams have been documented in Florida for over twenty year s (Howell et al. 1980, Drysdale and Bortone 1981, Cody and Bort one 1997, Bortone and Cody 1999, Jenkins et al. 2001, Parks et al. 2001, Chapters 2 through 4) Masculinization refers to development of male secondary sex characte ristics in females. Mosquito fish are sexually dimorphic: males develop a gender-specific copulatory organ, or gonopodium, as part of final maturation. The gonopodium is an extension an d differentiation of Rays 3, 4, and 5 of the anal fin (Turner 1941a). Based upon obser vations of female exposure to androgens (Turner 1941b, 1942a,b) and bacterial degrad ation of phytosterols into androgens (Marcheck et al. 1972), an androgen-medi ated mechanism has been hypothesized to explain this phenomenon. Laboratory exposure to bact erially-degraded phytosterol preparations confirmed the potential of th is hypothesis (Denton et al. 1985, Howell and Denton 1989), yet few studies ha ve demonstrated response under controlled exposure to whole effluents. Controlled exposure to whole effluent dilutions has produced inconsistent masculinization results. Initia lly in support of field collect ions, static renewal exposure of newborn mosquitofish to water collected 3.6 km downstream from Elevenmile Creek induced elongated anal fins (measured as anal fin length) in females upon maturity (Drysdale and Bortone 1989). While my study research was being conducted, researchers in Canada and New Zealand we re also studying mascul inization of adult female mosquitofish using controlled (mainly static renewal, one flow-through) exposures to 15%, 70% or 100% effluent (McCarthy et al. 2004). Four of seven pulp mill effluents induced masculinization (all static renewal). Among these four pulp mill

PAGE 157

139 effluents, two induced elongation relatively qui ckly (within 3 weeks) while the other two required 24 weeks of exposure. No associa tion between induction and type of mill or concentration of -sitosterol was established. Appare ntly, duration of exposure required to produce effect varies widel y. However, three important ca veats exist for these studies: Other than Elevenmile Creek, comparativ e field studies were not conducted to determine if effects existed in wild mo squitofish exposed to these effluents. All but one exposure required holding and transport of effluent back to the exposure system, with unknown consequences to effluent composition. Also noteworthy, masculinization was measured qualitatively, as either presence/absence or staged using cate gories established by Howell and Denton (1989). One of these controlled expos ures detected differences in masculinization due to effluent treatment and filtration (Ellis et al 2003). Secondary treatment of effluent reduced gonopodial development by 25%, yet masculinization remained significantly greater than controls. Filt ration of treated effluent, re moving many organic extractives adsorbing to particulates, also removed the response. This exposure was repeated two years late r, after treatment system maintenance was improved as indicated by gradual reduc tion in total suspended solids. Using the more specific Ray 4 to 6 length ratio, mascu linization was not induced (van den Huevel et al. 2004b). Since exposure duration remained the same (3 weeks), it was unknown if the effect was entirely removed or if time to manifestation was extended. Regardless, these experiments strongly correlate masculin ization with adsorbab le organic effluent components, such as low mol ecular weight wood extractives. The objectives of this study were: to assess the masculinization response (for secondary sex characters and steroids ) in mosquitofish under controlled and in situ field exposure to effluents from Rice Creek and Fenholloway River; to examine induction of

PAGE 158

140 responses under controlled conditions before and after major proces s changes at the Rice Creek mill; and to evaluate associations betw een concentrations of wood extractives in effluent and response in mosquitofish. Materials and Methods Controlled and in situ field exposure of mosquitofish to whole effluent dilutions was attempted on three separate occasions with limited success during the second attempt. Thus stated objectives were restrict ed to assessing the ma sculinization response, for secondary sex characters and steroids in mosquitofish under controlled and in situ field exposure to effluents from Rice Creek; and to eval uating associations between concentrations of wood extractiv es in effluent and responses in mosquitofish. Controlled tank exposures at the Rice Creek mill, using the facility described below, were tested in summer 2000. High fish mortality that was caused by several fact ors, including very high densities (over 1,000 fish per 1,500 L ta nk), precluded analysis. Tank exposures were modified and repeated concomitantly with field exposures at Ri ce Creek field sites in the summer of 2002. This study was complete d with partial fish loss and is the focus of this chapter. As part of the fry production study in 2003 (Chapter 6), in situ field exposures at Rice Creek and Fe nholloway River field sites were initiated but terminated prematurely because of several complications such as very acidic pH at the upstream Fenholloway site (averaging 3.5) and severe flooding that overflowed cages and made them inaccessible. Thus a large amount of e ffort yielded comparativ ely little data, and redesign of exposure scenarios is highly recommended. Mill Characteristics and Exposure Scenarios Mosquitofish were exposed to bleached/ unbleached pulp and paper effluent from the Georgia-Pacific mill in Palatka, Florid a, USA (Chapter 1, Table 1-1). The mill

PAGE 159

141 discharges into Rice Creek, a tributary of the St. Johns River. Rice Creek is a low-flow stream, so dilution factor for effluent is low until it reaches the Sain t Johns River. Major process changes to comply with US EPA’s Cluster Rule were implemented in May of 2001, and a reduction in effects on mosqu itofish was observed (Chapter 3). Controlled whole effluent exposure of mosquitofish was conducted using a unique onsite tank facility near the h ead of the discharge canal ( Figure 5-1 ). This facility was constructed primarily for flow-through exposur e of largemouth bass in 1998 (Seplveda 2000, Quinn 2004) and has also been used fo r exposure of freshwater mussels (T.S. Gross, unpublished data). Water from a near by well was pumped to a series of three 30 kL aboveground pools for filtration and remo val of sulfates and metals, and then pumped to a head tank. Simultaneously, 100% final effluent pulled midstream from the head of the canal was pumped to an adjacent head tank. Gravity was used to bring water from both lines down to aerated 1,500 L round tanks, and flow meters were used to control output from each line into the five se ts of tanks (duplicates were connected). Adjustment of flow meters (15 L per minute total per tank ) enabled effluent concentrations of 0, 10, 20, 40, and 80%. These concentrations were used for two reasons: 1) environmental relevance, since e ffluent accounts for an average of 60% flow upon discharge into Rice Creek and drops to < 10% at the St. Johns River; and 2) interspecies comparison, since largemout h bass were also exposed to these concentrations. Mosquitofish were also caged and submerged in situ at field sites surveyed regularly by our laboratory. Four sites were selected for field exposure, two each of exposed and unexposed sites ( Figure 5-2 ). Large 120 cm by 60 cm by 60 cm net cages

PAGE 160

142 (3 mm mesh) were framed with PVC pipe and fitted with PV C-reinforced screen lids. These same cages (without lids) housed fish in the tank exposures to provide equal densities of fish. Artificial cover (three to four green cheerleading pompons secured to top corners of cage) was provided to re duce aggression (a problem in the 2000 tank exposures). In situ cages were floated underneath shad e along vegetated banks (within 1 m from the bank). Targeted depths within the cage were approximately 30 to 40 cm, while distance between bottom of cages a nd stream sediments varied from 0 to approximately 60 cm (depending on tidal influenc e and rainfall). Cages were anchored to nearby trees and/or metal stakes. Exposures lasted for 4 weeks, based upon the length of time reported for mature gonopodial induction in females (2 to 3 week s) exposed to bacterially degraded phytosterols (Denton et al. 1985, Howell and Denton 1989). Approximately five thousand eastern mosquitofish were dona ted by Watts Aquatics (Lilith, FL) and transported to the US Geological Survey, Flor ida Integrated Science Center, Center for Aquatic Resource Studies, Ga inesville, FL, for acclimati on to captivity in 1500 L round tanks for one month. Artificial feed (Tetra min Flakes, Zeigler, Gardner, PA) used throughout the study was introduced at this time. Two to three hundred fish (estimated by weight at ~250 g/m3) were randomly assigned to each exposure group and transported to the mill in Palatka. Exposures commenced in midMarch 2002 on a staggered schedule over a ten day period, two to three treatment s or sites added every other day. Water quality parameters (dissolved oxygen, temp erature, pH, conductivity, salinity, and turbidity) and cages were monitored and fish we re fed three times a week in the mornings throughout exposure.

PAGE 161

143 Water Samples Single grab water samples were collected before fish sampling at weeks zero, two, and four to document potential exposure of fish to specific effluent components. Samples were preserved (buffered) a nd refrigerated for analysis. However, malfunction of the refrigerator caused spoilage and rendered the first two sets of samples entirely useless. Water samples for field exposures at week four were the only salvageable samples and were analyzed for phytosterol content by th e University of Flor ida Department of Environmental Engineering’s analytical chemistry core. Fortunately, NCASI was conducting fathead minnow ( Pimaphales promelas ) exposures at the same time and provided chem ical analysis data of 100% whole effluent measured for their study. Effluent was an alyzed for chlorinate d phenolics (12 Cluster Rule compounds plus 16 others), 10 resin ac ids (including 3 chlorinated), 3 fatty acids, 4 phytosterols, total organic carbon (TOC), condensable tannins, and polyphenolics. Additional effluent components were anal yzed in 100% whole effluent: metals, nonmetals (such as chloride and fluoride) and neutral semivolatiles. Columbia Analytical Services conducted the chlor ophenolic analyses, CH2M Hill conducted the TOC analyses, and the NCASI West Coast Re gional Center conducted all other analyses. Morphological Endpoints For each sampling week, 25 fish of each sex per treatment/site were targeted; however, low densities in the absence of obvious fish mortality at week two of sampling reduced this number to 15 fish (5 males a nd 10 females per treatment/site). First, fish were euthanized with a terminal dose of buffe red tricaine methanesulfonate (Tricaine-S, Western Chemical Inc., Ferndale, WA, USA) then weighed using a digital scale (+ 0.001 g) and measured for standard length (+ 0.01 mm) using a pair of di gital calipers. Under a

PAGE 162

144 dissecting scope, gender was identified using th e presence (female) or absence (male) of a urogenital papilla (Chapter 2). Each fish was photographed using a digital camera then placed on ice until transferred to -80C fr eezer for subsequent radioimmunoassay (RIA) of sex steroids. Anal fin images of these fish were measured by computer (+ 0.01 mm) using trace mode in Sigma Scan Pro 5.0 from the base of Rays 4 and 6 along the curve of each ray to the tip. Hormonal Endpoints Whole body primary sex steroids (17 -estradiol and testoste rone for this species) were analyzed using a modified RIA method originally developed for serum and plasma samples of common carp, Cyprinus carpio (Goodbred et al 1997), and since adapted for use in a variety of other aqua tic species and tissue media such as plasma of largemouth bass, Micropterus salmoides (Gross et al. 2001) and mantle of freshwater invertebrates (Gross et al. 2000). Chapter 2 provides deta iled methods and validation of this assay. Statistics Body weight and standard length were used to calculate condition factor, K = weight / length3 x 100 (g/cm3), as an indication of overall health used by the aquaculture industry (values of at least 1 are considered healthy, Hile 1936). The length ratio of anal fin Rays 4 and 6 was calculated as an index of anal fin elongation. Estrogen and testosterone concentrations were used to calculate a ratio indicating masculine hormone profile (E:T < 1) or feminine hormone profile (E:T > 1). Anal fin morphology and sex steroid data we re analyzed within sex using two-way analysis of covariance (ANCOVA) to test for significant variati on by treatment/site and week. Any data failing tests for norma lity and homogeneity of variance were log-transformed. Treatment or site differen ces within week were also analyzed by

PAGE 163

145 one-way ANOVA, as were differences by we ek. Significant di fferences in ANCOVA and ANOVA were analyzed for multiple co mparisons using Tukey’s HSD for field exposures and Dunnett’s for tank exposures. Relationship between anal fin morphology and sex steroids were analyzed overall a nd by treatment/site in two ways: first, by examining Pearson’s correlations of the i ndex of anal fin elon gation to sex steroid concentrations and ratio, then by t-test fo r differences in index of anal fin elongation between females with masculine versus femini ne E:T ratios. Statis tical significance was set at < 0.05 for all tests. All statistical an alyses were conducte d using SAS version 9.0. Results and Discussion All exposures continued to completion except for the field exposures at the reference site [REF2]. Just before the midpoint sampling for this site, cages were vandalized and destroyed. Water Quality Conductivity, salinity, turbidity and pH were significantly elevated in 40% and 80% effluent treatments ( Table 5-1 ). Dissolved oxygen was adequate for fish survival (> 4mg/L) and equivalent among treatments because of aeration of tanks. For field exposures, water temperature, conductivit y, salinity, dissolved oxygen, and turbidity significantly differe d among all sites ( Table 5-2 ). Temperature, conductivity, salinity, and turbidity were elevated at effluent e xposed sites, while dissolved oxygen was very low before discharge (PRE-DIS < 2 mg/L) a nd exceptionally high at the outfall (DIS > 10 mg/L). The pH differed significantly betw een exposed sites [PRE-DIS and DIS] but neither was different compared to the upstream site [U(8)].

PAGE 164

146 Water Chemistry Salvaged water samples from week four field exposures were analyzed for phytosterol content. Campesterol, stigma stanol, and stigmasterol were all below detection limits (< 12 g/L). -sitosterol was also below detection limit for the upstream site. -sitosterol was highest at the outfall site [DIS 45 g/L] and somewhat lower before discharge [PRE-DIS 32 g/L]. More sens itive analysis by NCASI confirmed higher concentrations of -sitosterol than the other phytost erols in 100% final effluent ( Table 5-3 ). Examining these compounds over the durati on of field and tank exposures provided compelling support for dynamic exposure of mosquitofish ( Figure 5-3 ). (Dynamic exposure refers to variable c oncentrations of effluent co mponents over time, dependent on factors such as tree species for furnish, w ithin plant processing sp ills, rainfall/dilution, and bacterial degradation.) In general e ffluent components decreased during exposures: the most dramatic drop was in resin acids while phytosterols were more stable ( Figure 5-3 and Table 5-3 ). Dilutions of these values coul d be roughly extrapolated to assess potential exposure of tank fish; however, additional environmental factors, such as rainfall and bacterial communities, may be influencing cage-exposed fish. These exposures began during a period of low ra infall in March, continued under average precipitation through April and ended at lo w rainfall in May (Appendix A). Therefore field exposure of caged fish may have been highest in March, lowest in April but then increased again in May, perhaps to concen trations intermediate between March and April. Without continuous monitoring, th ese concentrations are approximate and

PAGE 165

147 consider rainfall only. Additi onal factors (bacteri al activity) may have created even greater flux. Body Size and Condition Exposures had no impact on body size and condition for males. Females grew significantly longer under in situ field exposures (since fema les, and not males, grow continuously throughout life this result was not surprising). However, tank-exposed females did not grow suggesting increased f ood availability in fi eld cages, beyond the Tetramin flakes provided three times a wee k, may have caused growth in field-exposed females. Males There were no significant differences by week or among treatments in length, weight, or condition factor of males exposed to whole effluent dilutions at the tank facility ( Table 5-4 ). All males remained healthy (CF > 1) for the duration of exposure. Slight variations in body size occurred by week and by site in caged males exposed in situ at field sites, although these two factors did not covary ( Table 5-5 ). Standard length was slightly elevated for males at week tw o but not week four. Also, overall condition factor was higher for males from the discharge site [DIS] compared to the other two sites [U(8) and PRE-DIS], although they were al l above one indicating adequate general health. Females There were no significant differences by week or among treatments in length, weight, or condition factor of females exposed to whole effluent dilutions at the tank facility ( Table 5-4 ). All females remained healthy (CF > 1) for the duration of exposure. In contrast, females caged at field sites showed increased growth ( Table 5-5 ). Since

PAGE 166

148 females grow continuously throughout life, making body length approximate to age (Meffe and Snelson 1989), female growth by week was expected. However, given an absence of growth among tank exposed females, these data suggest field-exposed females may have had increased food consumption. Tank-exposed fish only received Tetramin flakes as a food source, while field-exposed fish potentially accessed any planktonic or plant particles passing through the cages. Anal Fin Morphology Neither sex responded to treatment in tanks or the field in terms of anal fin length. While this result was expected for males, it was unexpected for females. The only plausible conclusion for this lack of induc tion was the short duration of exposure which was based upon results published elsewhere af ter laboratory exposure of mosquitofish to degraded effluent components and not whol e effluent dilutions (Denton et al. 1985, Howell and Denton 1989). Males Male mosquitofish exposed to whole efflue nt dilutions in tanks did not vary in gonopodial length regardless of exposure dose or duration (Figure 5-4B ). Variation in gonopodial length was greater than for females. Similarly, caged males in field sites displayed no changes in gonopodial length, and the only si te-related difference was between the upstream and discharge sites [U (8) and DIS]. Gonopodia of males from the discharge site were shorter, although this was not significantly di fferent than control (week zero) males. Overall, effluent exposure did not affect male anal fin morphology in terms of length (2.5 + 0.08). These findings support prev iously reported surveys of male mosquitofish collected from effluent-expos ed Florida streams (C hapters 2 through 4).

PAGE 167

149 Females Similar to males, female anal fin elonga tion was not significantly different by week or treatment/site ( Figure 5-4A ). Average index of anal fin elongation (Ray 4 to Ray 6) was 1.14 + 0.02. These values are in accordance with field data for females at unexposed sites (Chapters 2 through 4). Also, reported Ray 4 to Ray 6 length ratios for unexposed females in the 11keto-testosterone study by Angus et al. (2001) averaged just above 1 (estimated 1.1 from Figure 5-4A ). In contrast to these exposures, wild female mosquitofish collected not only downstream of effluent discharge [DIS, D(1), D(3)] but also before discharge [PRE-DIS] displaye d masculinized anal fin elongations (1.25+ 0.03 Chapters 2 through 4). Thus induction of response was expected but not supported. Exposure research by McCarthy et al. (2004) indicates duration of exposure could explain this lack of induction. They found time to induction varied dramatically by mill effluent (from 3 to 24 weeks). Also po ssible but not probable, the mosquitofish population used for these exposures were someho w resistant or acclimated. To date, this has never been documented in eastern or wester n mosquitofish. Fish were farm-raised in manmade ponds separated from any known point sources of pollution, but these fish were not directly tested for prior chemical exposur e. Another alternativ e relates to sensitive life stage. For many other fish sp ecies, such as rainbow trout, Oncorhynchus mykiss (van den Huevel et al. 2002), increased sensitivity to effluent exposure occurs during critical windows of development, growth, and/or maturation. Laboratory exposures of mosquitofish to bacterially degraded phytosterols (Denton et al. 1985, Howell and Denton 1989), whole effluent exposures (Ell is et al. 2003, McCarthy et al. 2004, van den Huevel et al. 2004b), and lack of consistent size/age class related e ffects in wild-caught

PAGE 168

150 fish (Chapter 2) indicate mo squitofish females probably lack any particularly sensitive life stages relative to anal fin elongation. Sex Steroids In contrast to anal fin morphology, sex st eroids for both males and females were influenced by effluent exposure. Hormones were altered (mostly depressed except 17 -estradiol in males) by effluent expos ure and later appeared to recover among in situ field-exposed fish. Tank-exposed fish seemed to lag behind respons es in field-exposed fish, especially when E:T ratios were examined. Males seemed more sensitive than females regarding skewed sex hormone ratios Feminized hormone ratios (in males) and masculinized ratios (in females) were indu ced within two weeks in field-exposed fish with recovery or near recovery to normal ratio s by the end of four weeks. Skewed ratios were not significantly induced below the highe st effluent concentr ation in tank-exposed fish until four weeks. Very likely dynamic e xposure of mosquitofish to bioactive effluent components (or degradation products) explai ns these results, a lthough influence of toxicant(s) at different stages of steroid biosynthesis, tran sport, action, and degradation may also play a role. Males Before exposures, male sex steroids were normal and testosterone-biased at an E:T ratio of 3:1 ( Figure 5-5 ). Slight variation in 17 -estradiol levels wa s detected at 20% effluent ( Figure 5-5A ), but other than that no difference s in this hormone were detected from tank exposures by week and/or treatm ent. An exceptionally large peak in 17 -estradiol was detected in in situ field-exposed males at the predischarge site [PRE-DIS] at week two, with a pparent recovery at week four to levels similar to males

PAGE 169

151 from the other field sites. This recovery of normal steroid levels suggests either an acclimation response or a significant change in exposure. Unfortuna tely the latter cannot be directly deduced from available chemi cal data, although a decrease in exposure is indicated ( Figure 5-3 ). The rise in 17 -estradiol agrees with field collection data (Chapter 4), although significant elevation occurred at the di scharge site and not at the site before discharge. Testosterone concentrations did not vary among treatments at week two, but by week four they had dropped signifi cantly at 0% a nd 40% effluent ( Figure 5-5B ). Further, testosterone was significantly lower at 40% and 80% effluent compared to 0% for week four. The decrease at 0% over time implies a seasonal change in te stosterone, but does not explain the significant decreas es at the highest effluent co ncentrations. Testosterone in field-exposed males displayed a complete ly different pattern. Testosterone was depressed at the outfall for week two sa mples compared to males exposed upstream [U(8)]. By week four testos terone was significantly increased at both exposed field sites [PRE-DIS and DIS]. One possi ble explanation for this incr eased testosterone at week four is differential exposure to androg enic compounds between weeks and between whole effluent dilutions and instream fiel d dilutions. Tentative evidence for dynamic exposure in the sex steroid response of w ild-caught males (Chapters 3 and 4) lends credence to this conclusion. However, seve ral alternatives infl uencing the endocrine system also exist. Negative feedback loops are common in endocrine systems of fish and higher vertebrates (Van Der Kraak et al. 1998, Jala bert et al. 2000, Lister and Van Der Kraak 2001). Elevated estrogen or testosterone levels relay signals to the pituitary to decrease

PAGE 170

152 production of trophic hormones that are responsi ble for initiating biosynthesis. Elevated 17 -estradiol may be initiating a feedback l oop to bring these two hormones back into balance. Another point of regulation pot entially affecting these ster oid levels is aromatase, the monooxygenase enzyme responsible for converting androgens into estrogens. Orlando et al. (2002) pred icted inhibited activity of this enzyme in female mosquitofish collected from the Fenholloway River may cause elevated androgens that in turn produce masculinized anal fins in females. Contra ry to their hypothesis, aromatase activity was actually elevated in both br ain and ovarian tissue of fema les collected downstream of effluent discharge in the Fenholloway. Elev ated aromatase activity may explain results observed in males, leading to elevated 17 -estradiol. In support, E:T ratios were estrogen-biased by week two in males caged in both effluent-exposed field sites [PREDIS and DIS] and in 80% effluent in tanks ( Figure 5-6 ). By week four, field-exposed males demonstrated recovery back to normal testosterone-biased ratios, while even more tank-exposed males displayed feminized hormone ratios at 20%, 40% and 80%. Effects on ratios of tank-exposed males were due more to a drop in testosterone than a rise in 17 -estradiol, and this differen tial response within the same sex may allude to different chemicals affecting the hypothalamo-pituit ary-gonad axis or possibly different components of this axis being affected. Sin ce the difference in effect was between tankversus field-exposed fish, the hypothesis of dynamic exposure may be more plausible at this point.

PAGE 171

153 Females Significant elevation in 17 -estradiol occurred in c ontrol fish by week two (0% effluent) with recovery to preexposure levels by week four ( Figure 5-7A ). These changes are an important reminder about increasingly obvious seasonal effects on sex steroids in this species. Thr ough time, significant decreases in 17 -estradiol occurred in females from 40% and 80% exposure groups at week four, although the only significant difference among treatments at week four was at 40%. Similar to testosterone patterns betw een tankand field-exposed males, 17 -estradiol in females from field exposures demonstrated an opposite response. Large variation differences among sites were not de tected, but there was a significant increase in 17 -estradiol by the end of exposure. Pote ntially the females exposed in tanks may have lagged behind in response to depressed estrogen levels compared to field-exposed females. Such a “lag time” in response betw een the two groups may explain testosterone data in males as well, and suggests an e nvironmental factor in the natural system produced a more rapid response. It would be important a nd informative to extend tank exposures and increase sample sizes to determine if dominant hormones for each sex eventually catch up to observed responses in field-exposed fish. Testosterone concentrations in female s were not significan tly different among treatments in the tank exposures, nor wa s there a significant difference through time ( Figure 5-7B ). Testosterone in field-exposed females resembled the pattern for 17 -estradiol (and testosterone in field-exposed males), with a decrease in testosterone at the discharge site relative to upstream caged fish by week two followed by an increase at week four. These shifts become clearer when relative amounts of E:T are examined.

PAGE 172

154 Figure 5-8 depicts percentages of females with normal estrogen-biased sex steroid ratios and masculinized, te stosterone-biased ratios. In situ field-exposed females at week two had significantly different distribution of ratios than predicted by chance, with effluent-exposed females displaying a sligh tly greater tendency to have masculinized profiles. By the end of exposure (week f our), the difference disappeared and steroid ratios were mostly estrogen-biased and indi stinguishable from females exposed to the upstream unexposed site. Tank-exposed female s did not have signifi cantly masculinized hormone ratios until week four, and the res ponse was greater than that observed for field-exposed females at week two. What could be causing a delaye d reaction in tank exposed fish when they are most likely exposed to greater concentrations of effluent? And why do caged fish at the predischarge site res pond more similar to fish at the outfall site as opposed to highest concentrations in tanks? Because we used fish from the same aquaculture facility, genetic differences are an unlikely explanation for the differential response. Further, the fact that tank-exposed fish eventually re spond with similarly biased hormone profiles suggests a similar cause (although not conclusive). Assuming a similar cause, probable factors that vary between tank and field e xposures include chemical composition, food sources, and bacterial communities. There were actually three exposure syst ems that varied among each other: exposures in tanks to 100% w hole effluent dilutions; field ex posures before discharge in retention ponds; and field exposures at the di scharge point. All three were essentially flow-through systems in terms of water, but water parameters were literally and theoretically different. Fiel d cages hovered over and around the soft muck of sediments

PAGE 173

155 built up at field sites over time while tanks accumulated sediments from relatively recent effluent flow only. This difference, coupled wi th the probability of field-exposed fish to prey on wild food, indicates greater potent ial for bioaccumulation and biomagnification to influence responses in field-exposed fi sh compared to tank-exposed fish. The retention ponds before discharge [PRE-DIS ] were statistically different from the discharge site [DIS] in terms of water quali ty parameters such as pH and dissolved oxygen ( Table 5-3 ). Further, dissolved oxygen in tanks was in between values at field sites, and pH was highest in tanks. F actors such as pH and dissolved oxygen differentially affect bacteria l survival (Dick et al. 1998), thereby influencing which species thrive where. Conductivity, salin ity, and turbidity were comparable among exposures, and readings at bot h effluent-exposed field sites fell within 40–100% effluent for tank exposures (turbidity perhaps the most important for light penetration). The field sites also differed chemically. Phytosterol concentrations were estimated to be higher at the discharg e site (see water chemistry re sults above). Quinn (2004) detected an abundance of nonionic surfactan t degradation products (nonylphenol and octylphenol) at the outfall [DIS] where a liq uid oxygen injection syst em is also located (explaining extremely high dissolved oxygen le vels around 10 mg/L). Nonylphenol is considered a weak estrogen in terms of estrogen receptor binding and mRNA expression (Nimrod and Benson 1996), and has been impli cated as a bioactiv e agent in pulp and paper mill effluents (Lee and Peart 1999). Thus the environmental conditions among these three systems were likely very diffe rent and suggest variable exposure among exposure treatments and sites, in addition to the dynamic exposure ( Figure 5-3 and Table 5-3 ).

PAGE 174

156 Anal Fin Elongation and Sex Steroids Barring alteration in anal fin morphology fo r either sex, any relationship between anal fin elongation and sex ster oids could not be analyzed for this exposure study. Since changes in sex steroids were observed, the mo st that can be concluded is a sensitivity difference, alluded to by prev ious field work (Chapters 3 and 4): sex steroids were emphatically more sensitive than anal fins in terms of exposure. However, the usual caveats apply for these hormones as a biomarke r (normal seasonal fluctuations and actual adverse impacts of altered sex steroids are unknown). Conclusions In summary, controlled whole e ffluent exposure in tanks and in situ exposure in the field differentially affected anal fin morphol ogy and sex steroids in mosquitofish. Male anal fins were not affected by exposure, as expected based upon field surveys. However, anal fin elongation in females has been we ll-documented for this mill and exposures failed to induce an analogous response. This lack of induction may have been caused by insufficient exposure duration (4 weeks). On the other hand, sex st eroids for both sexes were more sensitive to effluent exposure, and males may be more sensitive for this endpoint based upon steroid ratios. Res ponse patterns differed between tankand field-exposed fish, and between fish from effluent-exposed fiel d sites. These differences indicated either variation in exposure and/ or multiple mechanisms affecting steroid production, transport, action and metabolism. Aspects of biosynthesis (aromatase) and direct action of hormones (androgenic compounds) were already discussed under steroi d hormone results in males as examples of the various endocrine pathways that may be affected by exposure to pulp and paper mill effluents. Several more have been documented both in vitro and in vivo for fish

PAGE 175

157 species. For example, fish from Jackfish Bay in Ontario displayed reduced ovarian steroidogenic capacity that was possibly lin ked to reduced cholesterol availability (cholesterol is the starting compound fo r lipophilic steroid hormone production of estrogens, progestins, androgens, and glucoc orticoids) (Van Der Kraak et al. 1998 and McMaster et al. 2003). Alternatively, pul p mill effluents were shown to bind sex hormone binding globulin (SHBG), a carrier prot ein for sex steroids in the bloodstream, potentially causing displacement and thus more rapid clearance of endogenous steroids (Hewitt et al. 2000). Degradation and excretion of sex steroids in and of itself may also be affected. For example, the conjuga ting detoxification enzyme UDP-glucoronosyl transferase has been both inhibited and i nduced by exposure to whole effluents or effluent components such as resin acids (Oik ari et al. 1983, Frlin et al. 1985, Andersson et al. 1988b, Lindstrom-Seppa and Oikari 1988, Lindstrom-Seppa and Oikari 1989). Increased metabolic clearance would depress endogenous steroid levels, while decreased clearance would build them up. None of these mechanistic studies have been performed on mosquitofish and would require characteri zation before evaluating which, if any, are responsible for observed effects on sex steroids. Although exposure of mosquitofish to bleached/unbleached kraft mill effluents failed to induce changes in anal fin morphol ogy, different patterns of the more sensitive response in sex steroids suppor t continued development of mosquitofish as a bioindicator species. The link between whole body steroid levels and anal fin morphology remains unclear. The action and effective concentration of steroids or steroid mimics causing anal fin elongation may be independe nt of whole body steroid leve ls. Indeed, systemic or peripheral action of sex ster oids on anal fin morphology has not been conclusively

PAGE 176

158 distinguished. At this point, without any ev idence of direct adverse effect of these biomarkers on mosquitofish (Chapter 6), th ey could be used as tiered responses indicating varying levels of exposure. Before such application, two points require further investigation: 1) typical seas onality of sex steroid concentr ations and ratios throughout the year (ideally this would be examined at more than one type of reference site and characterized in the laboratory ); and 2) thorough analysis of actual or direct chemical exposure to these fish, and documentation of potentially important e nvironmental factors such as bacterial communities and biomagni fication, should be executed in concert with repeat (and perhaps modified) exposures in th e field and in tank dilutions. Despite the limited amount of data returned for the la rge amount of effort expended, this exposure study was invaluable when contrasted against observed effects on th is species in wild field collections.

PAGE 177

159 Table 5-1. Water quality parameters (ave + se) measured three times weekly (n = 13 total) during four week tank e xposures of mosquitofish to bleached/unbleached kraft mill effluent in summer 2002. Effluent Concentration 0% 10% 20% 40% 80% Temperature (C) 21.7+ 0.4 22.6+ 1.0 21.8+ 0.4 22.6+ 0.5 23.3+ 0.4 Conductivity (S) 390.7+ 19.86 552.7+ 17.7 466.7+ 20.82 1380+ 120a 2228+ 33a Salinity (ppt) 0.2+ 0.008 0.3+ 0.01 0.2+ 0.01 0.7+ 0.06a 1.2+ 0.02a Dissolved Oxygen (mg/L) 8.59+ 0.49 8.10+ 0.45 8.30+ 0.46 7.42+ 0.43 6.97+ 0.38a Turbidity (ntu) 0.60+ 0.25 2.35+ 0.21 1.43+ 0.27 10.9+ 1.1a 18.1+ 0.7a pH 8.4+ 0.04 8.3+ 0.03 8.3+ 0.03 8.2+ 0.02a 8.2+ 0.03a asignificantly different from control (0% treatment) Table 5-2. Water quality parameters (ave + se) measured three times weekly (n = 12 total) during caged exposures of mosqu itofish to field sites in Rice Creek during summer 2002. Site U(8) PRE-DIS DIS Temperature (C)a 21.8+ 0.3 27.7+ 0.5 26.1+ 0.3 Conductivity (S)a 288.9+ 15.70 2245+ 14.05 1878+ 103.4 Salinity (ppt)a 0.1+ 0.01 1.1+ 0.01 1.0+ 0.03 Dissolved Oxygen (mg/L)a 6.29+ 0.16 1.74+ 0.20 10.15+ 0.77 Turbidity (ntu)a 3.96+ 0.51 33.8+ 14.9 13.2+ 0.53 pHb 7.4+ 0.1 8.0+ 0.03 7.8+ 0.02 asignificantly different among all three sites bsignificantly different between PRE-DIS and U(8) Table 5-3. Concentrations of selected e ffluent components in 100% final effluent sampled weekly midJanuary to midMay in 2002 (n = 10, data courtesy NCASI). MAX MEAN MIN SE Total RAFAa (g/L) 11.8 8.4 6.5 0.6 Campesterol (g/L) 12 5.9 NDc 1.4 Stigmasterol (g/L) 3.3 2.0 1.2 0.2 Stigmastanol (g/L) 7.8 5.6 4.3 0.4 -sitosterol (g/L) 71 47 33 3.8 TOCb (mg/L) 1767.2 804.6 96.3 180.7 polyphenolics (mg/L) 38.2 33.5 27.6 1.2 condensable tannins (mg/L) 6.6 5.1 2.5 0.4 aRAFA = resin acids and fatty acids bTOC = total organic carbon cND = nondetectable

PAGE 178

160 Table 5-4. Body size parameters (ave + se) for mosquitofish exposed to bleached/unbleached pulp mill effluents vi a whole effluent dilutions for four weeks in summer 2002. Effluent Concentration 0% 10% 20% 40% 80% Day 0 Sample Size 25 NAa NA NA NA Body Weight (g) 0.162+ 0.09 NA NA NA NA Standard Length (mm) 22.11+ 0.37 NA NA NA NA Condition Factor (g/cm3) 1.47+ 0.03 NA NA NA NA Sample Size 25 NA NA NA NA Body Weight (g) 0.460+ 0.039 NA NA NA NA Standard Length (mm) 28.62+ 0.716 NA NA NA NA Condition Factor (g/cm3) 1.83+ 0.05 NA NA NA NA Week 2 Sample Size 4 4 5 5 4 Body Weight (g) 0.213+ 0.021 0.253+ 0.029 0.200+ 0.020 0.235+ 0.006 0.201+ 0.030 Standard Length (mm) 24.23+ 0.82 24.57+ 0.95 24.56+ 0.75 24.88+ 0.46 24.60+ 1.36 Condition Factor (g/cm3) 1.49+ 0.10 1.68+ 0.06 1.34+ 0.10 1.54+ 0.09 1.32+ 0.03 Sample Size 11 10 10 10 10 Body Weight (g) 0.390+ 0.037 0.423+ 0.056 0.512+ 0.061 0.498+ 0.060 0.567+ 0.043 Standard Length (mm) 28.59+ 0.86 28.71+ 1.01 29.94+ 0.80 30.16+ 1.00 31.15+ 0.64 Condition Factor (g/cm3) 1.64+ 0.09 1.70+ 0.08 1.83+ 0.09 1.74+ 0.06 1.84+ 0.07 Week 4 Sample Size 10 10 10 10 10 Body Weight (g) 0.205+ 0.018 0.236+ 0.028 0.178+ 0.014 0.174+ 0.010 0.207+ 0.024 Standard Length (mm) 23.41+ 0.50 24.53+ 0.62 23.60+ 0.87 22.29+ 0.39 23.00+ 0.68 Condition Factor (g/cm3) 1.57+ 0.10 1.55+ 0.07 1.38+ 0.09 1.56+ 0.05 1.65+ 0.09 Sample Size 10 10 10 10 10 Body Weight (g) 0.642+ 0.108 0.557+ 0.044 0.503+ 0.101 0.505+ 0.020 0.554+ 0.045 Standard Length (mm) 31.88+ 1.25 30.74+ 0.76 29.52+ 1.26 30.72+ 0.40 31.35+ 0.68 Condition Factor (g/cm3) 1.84+ 0.11 1.89+ 0.06 1.78+ 0.10 1.74+ 0.04 1.77+ 0.06 aNA = not applicable

PAGE 179

161 Table 5-5. Body size parameters (ave + se) for mosquitofish caged in Rice Creek field sites for four weeks in summer 2002. Effluent Concentration Control U(8) PRE-DIS DIS Day 0 Sample Size 25 NAa NA NA Body Weight (g) 0.162+ 0.09 NA NA NA Standard Length (mm) 22.11+ 0.37 NA NA NA Condition Factor (g/cm3) 1.47+ 0.03 NA NA NA Sample Size 25 NA NA NA Body Weight (g) 0.460+ 0.039d NA NA NA Standard Length (mm) 28.62+ 0.716d NA NA NA Condition Factor (g/cm3) 1.83+ 0.05 NA NA NA Week 2 Sample Size NA 4 4 4 Body Weight (g) NA 0.144+ 0.018 0.180+ 0.028 0.207+ 0.040 Standard Length (mm) NA 22.83+ 0.50 23.15+ 1.17 24.16+ 1.07b Condition Factor (g/cm3) NA 1.19+ 0.06 1.42+ 0.04 1.43+ 0.09c Sample Size NA 10 10 10 Body Weight (g) NA 0.783+ 0.084 0.609+ 0.088 0.539+ 0.048 Standard Length (mm) NA 34.10+ 0.99 33.09+ 1.09 30.66+ 0.90 Condition Factor (g/cm3) NA 1.91+ 0.05 1.58+ 0.07 1.81+ 0.06 Week 4 Sample Size NA 10 10 9 Body Weight (g) NA 0.221+ 0.053 0.161+ 0.011 0.204+ 0.012 Standard Length (mm) NA 25.07+ 1.05 22.85+ 0.61 23.56+ 0.49 Condition Factor (g/cm3) NA 1.36+ 0.10 1.34+ 0.03 1.54+ 0.02c Sample Size NA 10 10 9 Body Weight (g) NA 0.625+ 0.076 0.541+ 0.04 0.616+ 0.034 Standard Length (mm) NA 32.85+ 0.84 31.99+ 0.66 33.55+ 0.74 Condition Factor (g/cm3) NA 1.70+ 0.12 1.63+ 0.06 1.63+ 0.06 aNA = not applicable bstatistically different than week 0 (p < 0.05) cstatistically different (over all 4 weeks) to other sites (p < 0.05) dstatistically different than week 2 and week 4 (p < 0.05)

PAGE 180

162 Figure 5-1. Diagram of tank facility for flow-through whole effluent exposure of mosquitofish in summer 2002 at Geor gia-Pacific’s Palatka, FL operation. retention well p onds 100% final effluent head tanks well water effluent 0 10 20 40 80 0 10 20 40 80 Rice Creek well water filtration pools

PAGE 181

163 Figure 5-2. Map of cage locations for in situ field exposures at Rice Creek, FL, in 2002 (red asterisks). Other field sites ro utinely surveyed for mosquitofish masculinization are also indicated. Site symbols distinguis h sites exposed to effluent: circles = unexposed and tria ngles = exposed. Site abbreviations denote upstream (U) or downstream (D) of discharge, followed by approximate distance (km) from discharg e in parentheses; PRE-DIS indicates site before discharge into the creek; DIS denotes site at discharge into creek; REF indicates reference site, fo llowed by identifying number. * *

PAGE 182

164 DATE 3-18-023-31-024-14-024-29-025-13-02 CONCENTRATION IN WHOLE EFFLUENT (ug/L) 0 25 50 75 100 125 150 175 12-Chlorodehydroabietic acid Pimaric acid Linoleic Acid Beta-sitosterol Figure 5-3. Concentrations of selected wood ex tractives in 100% final effluent from the Rice Creek mill during tank and fiel d exposures of mosquitofish. 12-chlorodehydroabietic acid is a chlorina ted resin acid, pimaric acid is a resin acid, linoleic acid is a fatty acid, an d beta-sitosterol is a phytosterol.

PAGE 183

165 SITE TYPE c o ntr o l 0% 1 0% 20% 40% 80% U(8) PRE-DIS DIS INDEX OF ANAL FIN ELONGATION 0.0 0.5 1.0 1.5 2.0 2.5 3.0 WK 0 WK 2 WK 4 SITE TYPE control 0% 10 % 20 % 40 % 80 % U (8) PRE-DIS D I S INDEX OF ANAL FIN ELONGATION 0.0 0.5 1.0 1.5 2.0 2.5 3.0 A B a Figure 5-4. Index of anal fin elongation (length ratio of Ray 4 to Ray 6) for mosquitofish exposed to bleached/unbleached pulp m ill effluents via whole effluent dilutions or onsite caged exposures for four weeks in summer 2002. A) Females. B) Males. Solid black line separates tank from in situ field exposures. Letter “a” indicates signif icant differences by site (p < 0.05).

PAGE 184

166 TREATMENT/SITE control 0 % 1 0 % 2 0 % 4 0 % 8 0 % U ( 8 ) P R E D I S DIS TESTOSTERONE (pg/g) 0 200 400 600 800 1000 1200 1400 1600 1800 TREATMENT/SITE c o n tr o l 0% 10% 20% 40% 80% U(8) PRE-DIS D I S 17BETA-ESTRADIOL (pg/g) 0 200 400 600 800 1000 1200 1400 1600 1800 2000 WK 0 WK 2 WK 4 A B a c b b Figure 5-5. Whole body se x steroids (ave + se) for male mosquitofish exposed to bleached/unbleached pulp mill effluents vi a whole effluent dilutions or onsite caged exposures for four weeks in summer 2002. A) 17 -estradiol. B) Testosterone. Solid black line separates tank from in situ field exposures. Letters indicate significa nt differences by treatment or site within week (p < 0.05): “a” denotes differences to PRE-DIS and DIS; “b” denotes differences to all other sites; “c” denote s differences to U(8). Yellow circles indicate significant differences by week. Treatment or site and week did not covary for either hormone.

PAGE 185

167 % MALES 0102030405060708090100 WK 0 CONTROL MASCULINE STEROID RATIO (E:T<1) FEMININE STEROID RATIO (E:T>1) % MALES 0102030405060708090100 WK 2 TREATMENT or SITE 0% 10% 20% 40% 80% U(8) PRE-DIS DIS % MALES 0102030405060708090100 WK 4 TREATMENT or SITE 0% 10% 20% 40% 80% U(8) PRE-DIS DIS A B C Figure 5-6. Percentage of ma le mosquitofish with mascu line and feminine sex steroid ratios exposed to bleached/unbleached pulp mill effluents via whole effluent dilutions or in situ field exposures for four weeks in summer 2002. A) Control males (week 0 sample). B) Midpoint sampling (week 2). C) Final sampling (week 4).

PAGE 186

168 TREATMENT/SITE c on tr ol 0 % 10 % 20% 40 % 80% U ( 8) P RE -D IS D IS 17BETA-ESTRADIOL (pg/g) 0 500 1000 1500 2000 WK 0 WK 2 WK 4 a b TREATMENT/SITE c on t rol 0 % 10 % 20% 40 % 80% U ( 8) P RE -D IS D IS TESTOSTERONE (pg/g) 0 200 400 600 800 1000 1200 1400 c A B Figure 5-7. Whole body sex steroids (ave + se) for female mosquitofish exposed to bleached/unbleached pulp mill effluents vi a whole effluent dilutions or onsite caged exposures for four weeks in summer 2002. A) 17 -estradiol. B) Testosterone. Solid black line separates tank from in situ field exposures. Letters indicate significa nt differences by treatment or site within week (p < 0.05): “a” denotes differences to 20% ; “b” denotes differences to 0%; “c” denotes differences to DIS. Site type and mill did not covary for either hormone.

PAGE 187

169 % FEMALES 0102030405060708090100 WK 0 CONTROL MASCULINE STEROID RATIO (E/T<1) FEMININE STEROID RATIO (E/T>1) % FEMALES 0102030405060708090100 WK 2 TREATMENT or SITE 0% 10% 20% 40% 80% U(8) PRE-DIS DIS % FEMALES 0102030405060708090100 WK 4 TREATMENT or SITE 0% 10% 20% 40% 80% U(8) PRE-DIS DIS Figure 5-8. Percentage of female mosquitofish with masculine and feminine sex steroid ratios exposed to bleached/unbleached pulp mill effluents via whole effluent dilutions or in situ field exposures for four weeks in summer 2002. A) Control females (week 0 sample). B) Midpoint sampling (week 2). C) Final sampling (week 4).

PAGE 188

170 CHAPTER 6 INVESTIGATION OF REPRODUCTIVE SU CCESS IN MOSQUITOFISH LIVING IN PULP AND PAPER MILL EFFLUENT DOMINATED SYSTEMS The impact of female mosquitofish ma sculinization or development of male secondary sex characteristics on reproductive success has not been directly addressed despite speculation of negativ e impacts. Histologically, masculinized females are normal, and limited brood size data (separ ate from masculinization studies) are conflicting but tend toward a lack of effect The current study evaluated fry production and masculinization in females collected fr om two effluent-receivi ng streams in Florida for two reproductive seasons. In additi on, population structure and abundance was tentatively evaluated the first season and masc ulinization and sex steroids were measured in adults. Potential exposure was document ed by analysis of wood extractives in water samples from field sites. Morphological ma sculinization was cons istent between years for one mill while the response was only detect ed the second year of study at the other mill. Sex steroid alterations measured the first year were weakly affected relative to measurements in previous studies (Chapters 2 through 5). Neither of these biomarkers could be associated with fry production or population structure differences. Fecundity appeared reduced in the first year’s single collection for bot h effluent-exposed sites but population structures implied mosquitofish at effluent-exposed sites may have started reproducing sooner than at unexposed sites. The second year of fry production studies over several months affirmed different repr oductive patterns in females among sites and through the season. Furthermore, overall fec undities were higher in females from one

PAGE 189

171 exposed site relative to respective refere nces. Rather than negatively impacting fecundity, pulp and paper mill effluent exposure may be stimulating modified reproductive strategies in mo squitofish influenced by ch anges in environmental and ecological factors as opposed to chemical exposure. Introduction Sublethal effects of pulp and paper mill effl uent exposure on fish have been a major focus of aquatic environmental health concerns for over a decade (Sodergren 1991, Servos et al. 1996, Ruoppa et al. 2000, Stut hridge et al. 2003, Borton et al. 2004). Reported effects include inducti on of liver detoxification systems, alterations in sex steroid concentrations a nd production/metabolism, reduced gonadal development, decreased egg production and decreased fry survival (Van der Kraak et al. 1992, Gagnon et al. 1994a, Munkittri ck et al. 1999, NCASI 2000a, McMast er et al. 2003, Seplveda et al. 2003, Parrott et al. 2004). Whether or not these effects represent actual adverse effects in terms of reproductive success or population and community level impacts remains controversial. Development of male-like secondary sex ch aracteristics in female mosquitofish, specifically masculinization of the anal fin into a gonopodial -like structure, has been reported in pulp mill effluent-receiving str eams for decades Howell et al. 1980, Drysdale and Bortone 1981, Cody and Bortone 1997, Bort one and Cody 1999, Jenkins et al. 2001, Parks et al. 2001, Chapters 2 through 4). A lthough impacts on reproduction were initially implied by this phenomenon, normal ovaries lacking any testic ular tissue were consistently reported in masculinized fema les (Howell et al. 1980, Hunsinger et al. 1988, Ellis et al. 2003, McCarthy et al. 2004). In a ddition, sex ratios of mosquitofish reared in 100% final effluent were not altered (McC arthy et al. 2004). Decreased potential

PAGE 190

172 fecundity, measured as brood size of de veloping embryos, was reported in early preliminary work (Rosa-Molinar and Williams 1984), but has not been detected in more recent studies (Felder et al. 1998, D’Surney et al. 2000). With the drastic improvements in processing technologies by the pulp a nd paper industry, masculinization has been reduced relative to initial reports (Howe ll et al. 1980, Drysdale and Bortone 1981, Cody and Bortone 1997, Chapters 3 and 4) and it is possible reproduction was previously impacted but may no longer be impaired. Mosquitofish, as members of the livebearing family Poeciliidae, develop eggs internally and ovulate im mediately before parturition of fry (Meffe and Snelson 1989). As nonsuperfetating lecithotrophe s, mosquitofish develop a single brood at a time and exhibit yolk-loading of eggs similar to egglaying (oviparous) speci es without maternal investment during embryological development (Turner 1937). Compared to egg-laying species, reproduction is asynchronous and re productive season occu rs through summer months with low to no reproduction in wint er months (Constanz 1989). Environmental cues control the beginning a nd end of the reproductive s eason: onset of reproductive season is triggered by a rise in water temperature while photoperiod (decreasing daylength) signals gonadal recrudescen ce (Koya and Kamiya 2000, Koya and Iwase 2004). A major drawback among existing studies of mosquitofish reproduction in pulp and paper mill effluents is the lack of corres ponding anal fin morphology data to evaluate potential association of masculinization a nd reproduction. Also, these studies represent potential effects on reproductive success as oppo sed to actual fecund ity or production of fry. No work has been done to assess potential impacts of effluent exposure on

PAGE 191

173 populations of mosquitofish in reference to reproductive level effects. The primary objective for the current study was to evalua te fry production in female mosquitofish collected from effluent receiving streams and relate this endpoint to masculinization of the anal fin. Preliminary investigation of population structures and abundance at effluent-exposed sites was a secondary objec tive to provide perspective on fry production studies. These collections were also evalua ted for anal fin morphology and sex steroids in adult fish. Materials and Methods Mill Characteristics Two pulp and paper mill effluent receiving systems in Florida where masculinized female anal fins have been previously documented were surveyed for mosquitofish reproduction studies in the summers of 2003 and 2004 ( Figure 6-1 and Appendix A). The Fenholloway River was the focus of an abbreviated population survey in 2003, while one instream exposed and one unexposed site from each of the two systems were surveyed to assess fry production in female s (once in 2003, and over a four month period in 2004). The Fenholloway River and Rice Creek mills are very different in furnish, processing and product (Chapter 1, Table 1-1). The Rice Creek mil is a bleached/unbleached kraft mill therefore subject to EPA’s Cluster Rule. The Fenholloway River mill uses a dissolving kr aft pulping process to produce high grain cellulose and is regulated under co mpletely different guidelines. Water Samples Before fish collection, water quality para meters typically affected by pulp and paper mill effluents were measured at each site: dissolved oxygen, temperature, pH,

PAGE 192

174 conductivity, salinity, and turbidity. Single gr ab, unfiltered surface water samples (1 L) were also collected before fish collection to document potential exposure of fish to specific effluent components. Samples were preserved (buffered) and sent to the National Council for Air and Stream Improveme nt, Inc. (NCASI) for chemical analysis by GC/MS. Water from all sites was an alyzed for 10 resin acids (including 3 chlorinated), 3 fatty acids, 4 phytosterols, total organic carbon (TOC), condensable tannins, and polyphenolics. Chlorinated phenol ics (12 Cluster Rule compounds plus 16 others) were analyzed in 100% final effluent for the firs t two field collections in 2003, and then discontinued since there were nondet ectable levels across sites. The exception to these findings was a small peak of 2-chlorosyringaldehyde (3.0 g/L) in the Fenholloway mill effluent at the first sampling. Columbia Analytical Services conducted the chlorophenolic analyses, CH2M Hill conducted the TOC analyses, and the NCASI West Coast Regional Center conducted all other analyses (NCASI 1986 1997). Population Survey As a preliminary investigation into populat ion structure and relative abundance of mosquitofish inhabiting efflue nt-receiving systems, field coll ection sites were surveyed systematically in May 2003 using dip nets al ong shallow vegetated banks. Three to five observers were spaced a minimum of 10 mete rs apart throughout sampling and number of sweeps, estimated time, and estimated area sampled were monitored and recorded by an independent observer. Sampling concluded when an estimated 75 to 100 adult female mosquitofish were collected fo r fry production studies. All mo squitofish were kept alive in aerated bait buckets an d transported back to th e laboratory for processing.

PAGE 193

175 Morphology Back at the laboratory, mosquitofish we re sorted into age-sex groups, giving preference to gravid females for preparation of fry production studies. Fish were handled as little as possible and w earing latex gloves to minimize stress on gravid females. Groups were divided as follows: gravid fe males (anal spot and swollen abdomen); nongravid females (lack of or partial anal sp ot and slim abdomen); adult males (fully differentiated gonopodium); developing males (elongated gonopodium lacking terminal differentiations); and juvenile s (< 20 mm standard length and lacking anal spot and gonopodium). Urogenital papillae were only used to distinguish gender in fish difficult to sex, as manipulation under a dissecting scope places more stress on gravid females. Fish designated for hormone analysis (20 to 30 adult males and females per site) were processed the day of collection. First, fish were euthanized with a terminal dose of buffered tricaine methanesulfonate (Tricain e-S, Western Chemical Inc., Ferndale, WA, USA), then weighed using a digital scale (+ 0.001 g) and measured for standard length (+ 0.01 mm) using a pair of digital caliper s. Under a dissecting scope, gender was reaffirmed using the presence (female) or absence (male) of a urogenital papilla (see Chapter 2 for validation of this sexing tec hnique). Each fish was photographed using a digital camera then placed on ice until tr ansferred to –80C freezer for subsequent radioimmunoassay (RIA) of sex steroids. Anal fin images of these fish were measured by computer (+ 0.01 mm) using trace mode in Sigma Scan Pro 5.0 from the base of Rays 4 and 6 along the curve of each ray to th e tip. Remaining fish were euthanized and preserved in 10% neutral-buffered formalin for determining population structure.

PAGE 194

176 Sex Steroids Whole body primary sex steroids (17 -estradiol and testoste rone) were analyzed using a modified RIA method originally de veloped for serum and plasma samples of common carp, Cyprinus carpio (Goodbred et al 1997), and since then adapted for use in a variety of other aquatic species and tissue media such as plasma of largemouth bass, Micropterus salmoides (Gross et al. 2001) and mantle of freshwater invertebrates (Gross et al. 2000). For methods and vali dation of this assay, see Chapter 2. Fry Production Fifty gravid females from one exposed and une xposed site per system (total 4 sites) were held for thirty days to monitor fry produc tion. Each female was placed individually in a modified plastic hatchery chamber purchased from Aquatic Ecosystems (Apopka, FL) that included hinged lids to prevent escape and 3” of artifici al green Cabomba to provide cover for females and fry. Upper portions of the hatchery chamber were available to females while the lower porti on was accessible only by fry. Newborn fry instinctively seek escape and protection from the mother and her cannibalistic instinct. Fish in hatchery chambers were initially held in 2’ round tanks with site water and acclimated to 50:50 pond:well water mix for 24 to 48 hours by gradual drip of lab water and graded pH shifts of 0.1 per hour and no more than 1 unit per day. After acclimation, females were transferred to two 4’ by 8’ by 6” shallow tanks receiving 50:50 filtered pond:well water mix from a head tank. Cham bers were randomized with respect to location in tanks, and 100 chambers filled each tank allowing for up to 200 chambers total. Full spectrum lighting was set on a 14:10 hour light:dark schedule to simulate increased photoperiod and keep females in reproductive mode.

PAGE 195

177 Chambers were monitored daily for fry production. Fry were removed immediately then euthanized, counted and preserved in 10% neutral buffered formalin for assessment of deformities and fry weight. Chambers were rinsed and females returned to the tank for observation of secondary brood pro duction. Initially, females were fed daily ad libitum with Tropical Prime flakes (Zeigl er Brothers, Gardners, PA, nutritional composition 45% min protein, 9% min fat, 4% max fiber). Feeding was later reduced to every other day because of a water mold infe station (Saproligniosis of Class Oomycetes) attributed to overfeeding. Water mold require d treatment with an overnight static salt immersion at 3 ppt. Water quality (disso lved oxygen, temperature, pH, conductivity, salinity; incident light and turbidity once a week) was measured three times a week and shallow tanks were cleaned. In 2004, this experiment was repeated m onthly from May to August to assess fry production throughout the reproductive season. A final collection in September was cancelled because of hurricane conditions. The only differences between years were: females were euthanized once fry were produced in 2004, and fry deformities and weights were not measured. Statistics As part of the population survey in 2 003, body weight and standard length were used to calculate condition factor, K = weight / length3 x 100 (g/cm3), as an indication of overall health used by the aquaculture industry (values at least 1 are considered healthy, Hile 1936). The length ratio of anal fin Rays 4 and 6 was calculated as an index of anal fin elongation. Estrogen and test osterone concentrations were used to calculate a ratio indicating masculine hormone profile (E:T < 1) or feminine hormone profile (E:T > 1). For fry production studies, body length has been shown to positively affect brood size

PAGE 196

178 (Krumholz 1948, Hughes 1985, Meffe and Snelson 1989), so the number of fry produced by each female was divided by her standard length for statistical analysis. Differences in population structur e among sites were analyzed by 2 test for independence. Relative abundance of mosqu itofish was estimated by calculating catch per unit effort (CPUE) as number of fish pe r sweep. Anal fin morphology and sex steroid data of adult fish from these collections were analyzed within sex using one-way analysis of variance (ANOVA) to test for significant variation by site (t -test for Rice Creek fish). Any data failing tests for normality and hom ogeneity of variance were transformed using log transformations. Significant differences in the ANOVA tests were analyzed for multiple comparisons using Tukey’s HSD. Relationship between anal fin morphology and sex steroids were analyzed by t-test for differences in index of anal fin elongation between females with masculine versus femini ne estrogen to testosterone (E:T) ratios. Since water parameters were measured repeatedly, water quality (2004 data) and chemistry (2003 and 2004) were analyzed us ing one-way ANOVA to test for significant variation by site (t-test for Rice Creek samples). Signifi cant differences in the ANOVA tests were analyzed for multiple comparisons using Tukey’s HSD. Fry production data were analyzed between si tes using t-tests within systems. Any data failing tests for normality and homogene ity of variance were transformed using log transformations. Differences in numbers of live versus dead fry and deformities were analyzed by 2 test for independence. For the 200 4 data, interaction by site and month was also analyzed using an analysis of c ovariance (ANCOVA), and influence of month within each site was analyzed by one -way ANOVA followed by Tukey’s HSD.

PAGE 197

179 Statistical significance was set at = 0.05 for all tests. All statistical analyses were conducted using SAS version 9.0. Results and Discussion Water Quality In general and as expected, most water quality parameters were higher at effluent-exposed sites compared to unexposed sites ( Table 6-1 for 2003 and Table 6-3 for 2004). Elevated temperatur e at effluent-exposed site s may be important, since reproduction is initiated by a rise in te mperature (Koya and Kamiya 2000). In 2004, dissolved oxygen was low at the effluent -exposed Fenholloway site (<1 mg/L), but elevated at the Rice Creek di scharge site (due to oxygen in jection system mentioned in Chapter 5). Conductivity, salinity and tu rbidity for fry producti on tanks were low and comparable to reference and upstream field sites ( Table 6-2 for 2003 and Table 6-4 for 2004). Temperatures were intermediate betw een exposed and unexposed sites, averaging closer to exposed sites. Di ssolved oxygen was adequate for fish survival (overall average 6.34 mg/L). Incident light varied from 60 to 120 Fc with greatest light directly beneath fixtures; light intensities may affect productivity of female mosquitofish when comparing indoor to outdoor lighting (W.K. Bradley pers. comm.), therefore this may have affected fry numbers. Randomization of hatchery ch ambers in the shallow tanks accounted for this bias. Overall pH was high (average 8.0): measur ements were most similar to the Rice Creek discharge site [DIS], and in May 2004 pH was significantly higher than other months. Although fish were acclimated to wa ter conditions in the laboratory, the large jump in pH during May 2004 (especially for une xposed fish) may have caused stress and

PAGE 198

180 subsequent higher mortality than observed for other months. In support, large shifts to acidic pH were observed to induce 100% mo rtality in caged fish during attempted exposures at the highly acidic upstream Fe nholloway River site [U(5)] regardless of acclimation to the low pH. Water Chemistry In general, effluent components were at higher concentrations in Fenholloway River compared to Rice Creek, although resi n acid concentrations were comparable across sites during 2003 ( Table 6-5 for 2003 and Table 6-6 2004). Water sampling coincided with fish sampling for the 2003 popul ation survey and 2004 female collections, and additional water samples were measured before and after the population survey and after female collection in 2003 (as part of the aborted caged exposures mentioned in Chapter 5). Total resin acids, 3 of 4 phytosterols, TOC, and polyphenolics (lignin content) statistically distinguished efflue nt-exposed from nonexposed sites. Of the phytosterols, only campesterol was different for Fenholloway River exposure sites in 2003; otherwise this compound was low to nondetect. Analyses of phytosterol concentrations were problematic for the Ri ce Creek discharge site both years and the Fenholloway River sites in 2004 in terms of low and variable su rrogate recoveries (failing quality control); theref ore these values must be considered estimates. Fatty acids and condensable tannins were not consis tently distinguishable among sites for 2003, while condensable tannins were hi gher in exposed sites for 2004. Examining concentrations over time reveal ed the dynamic (potential) exposure of mosquitofish to pulp and paper mill effluent components. Fish were collected for the 2003 population survey under relati vely stable concentrations in Fenholloway River, while concentrations in Rice Cr eek were dropping dramatically ( Figure 6-2 ). When

PAGE 199

181 females were collected for fry production in July, dehydroabietic acid (DHAA) was at its peak for Fenholloway River and was equivalent to May concentrations for Rice Creek. Comparison to monthly precipitation for each region (Appendix A) does not match perfectly (weekly preci pitation data would have been a better comparison). Population collections were made in May, during a month of average rainfall for both systems. At Rice Creek, June peaked with flooding, and fema les were collected for fry production in July when rainfall was low. At Fenholloway River, June began flood conditions of the rainy season, and females were collected for fry production studies in July when rainfall remained elevated. The apparent dispar ity between precipitation and effluent concentrations serves as another demonstra tion of multiple factors influencing effluent exposure, in this case perhaps including changes from mill output. DHAA concentrations over three of the four fish collections in 2004 revealed stable exposure in fish from Rice Creek and a la rge drop in exposure at Fenholloway River ( Figure 6-3 ). Concentrations of DHAA in June and July at Fenholloway were double concentrations in 2003. Again, monthly precip itation data did not match these patterns consistently (remembering rainfall data is a monthly average while chemistry data are single time points for these months). Rainfa ll was consistently high for the Rice Creek system over collections because of several intense storms, whereas precipitation at Fenholloway River was high in June, low in July, then high again in August. Population Survey Preliminary investigation of relative population abundance (CPUE) and age-sex structure showed mosquitofish populations at instream effluent-exposed sites were equally, if not more prolific than populations at unexposed sites. For example, median CPUE was highest at the first Fenholloway River downstream site [D(5)] ( Figure 6-4 ),

PAGE 200

182 and juvenile production was greatest instream and closest to discharge for both systems ( Figure 6-5 ). In contrast, relative abundance appe ared lower at the discharge point for Rice Creek [DIS] versus the upstream site [U (8)]. Abundances proba bly reflect ease of sampling related to water level and precipitati on just before or duri ng days of collection, rather than actual density of mosquitofi sh. The discharge site at Rice Creek was especially flooded the day of collection, alt hough monthly rainfall was average, and fish were spread out across shallow swamp backwa ters as opposed to concentrated near the banks of the creek. Intermittent rain show ers during fish collection at the Fenholloway River reference site [REF2] made it very diffi cult to see fish to even scoop. Therefore CPUE data may reflect an active sampling bi as instead of or in addition to actual differences in abundance. Further, these data sets were not large enough for formal statistical analysis such as nonparametric te sts on CPUE data or estimation and modeling. At the least, these data demonstrated mo squitofish populations were not severely impacted by effluent exposure which is not surprising given their tolerance for environmental extremes such as high salinity and low dissolved oxygen (Meffe and Snelson 1989, Nordlie 2000). A more salient point out of th ese preliminary data is the apparently different stages in reproductive season among sites ( Figure 6-5 ). Practically 100% of females from upstream sites were nongravid and most fe males from the Fenholloway River reference site were also not gravid. Internal gross examination of gonads of formalin-preserved fish confirmed these designations based upon anal spots and body shape (data not shown). In contrast, females closest to e ffluent outfall in both systems were mostly gravid; and approximately 50% of females collected further downstream and before

PAGE 201

183 discharge in Fenholloway and females collected before discharge into Fenholloway were gravid. Since gravid females had been collected ear lier in the season in previous years at unexposed sites (March and Apr il 2000), this lack of gravid females at unexposed sites was surprising and precluded c ontinuation into fry production studies. Juveniles were present at these sites, indicating females had already been pregnant earlier in the season, so either an environmental factor paused the reproductive season (which has not been documented) or these fish were more s ynchronous in reproductive strategy than previously believed. Evidence exists for tw o reproductively active female populations in mosquitofish: overwintering females a nd young-of-year females (Hughes 1985, Haynes and Cashner 1995, Fernandez-Delgado and Rossomano 1997). Overwintering females tend to produce larger clutches earlier in the season and eventually die out. Young-of-year females produce smaller clutches of fry, and these females may or may not live to become overwintering females the following year. Possibly the unexposed populations were in between major pr oduction by overwintering females and young-of-year initial fry production, wher eas exposed populations already had young-of-year females reproducing. Higher temper atures at effluent -exposed sites may trigger onset of reproduction ea rlier than at unexposed site s. This would explain the greater proportion of juveniles at instream effluent-exposed sites as well. Therefore, comparing fry production at the same time point, as performed in 2003, may not be an accurate reflection of reproductive success in the population. Instead, different stages of the reproductive season may have been compar ed. This is one of the main reasons

PAGE 202

184 females were examined on a monthly basis in 2004, to provide a more accurate picture of reproductive stages among sites. The other aspect examined by this portion of the study was adult sex ratios. Normal sex ratios are debatable but ra nge from approximately 1:1 to strongly female-biased (Krumholz 1948, Meffe and Snelson 1989). A variety of ecological factors such as predation a nd habitat preferences can alte r sex ratios (Casterlin and Reynolds 1977, Britton and Moser 1982) therefor e alterations in the field must be interpreted with caution. Normal sex ratios (1:1 to female-bia sed) were detected at both Rice Creek sites and at the upstream, predis charge, and furthest downstream sites in Fenholloway River ( Figure 6-5 ). However, male-dominated sex ratios (approximately 1:2) were observed at the Fenholloway refe rence site and the first downstream site. Overall, neither effluent-exposed site nor potential exposure documented by water chemistry can be associated with thes e apparent alterations in sex ratio. Body Size Body size was not impacted by effluent-e xposed site for both systems in fish collected for the population survey in summer 2003 ( Table 6-7 ). Condition factor was above one and indicated adequate general h ealth across all sites, regardless of any (erroneous) statistical differences by site within the Fenholloway River system. Compared to the upstream site at the Fenhollo way River [U(5)] but not the reference site [REF2], females from effluent-exposed sites we re larger. No significant differences were detected between upstream and discharg e sites in Rice Creek for either sex. Anal fin morphology Anal fin elongation in female mosquitofi sh was detected for the Fenholloway River system only in adult mosquitofish collect ed for the population survey in summer 2003

PAGE 203

185 ( Figure 6-6 ). No other statistically significant im pacts were observed for both systems. This collection represents the first sta tistically defined lack of morphological masculinization in Rice Creek female mosqu itofish, which was diminished since major process improvements in May 2001 (Chapter s 3 and 4). Anal fin elongation in Fenholloway River females was similar to elongation observed in 2001 (Chapter 4). These results may imply a threshold concen tration of bioactiv e component(s) lying somewhere between Fenholloway River and Rice Creek measurements in water samples ( Table 6-5 ). Alternatively, unknown contributions of environmental factors may have varied in the Rice Creek system for the 2003 collection. Sex steroids Sex steroids displayed high va riation, similar to previous field collections (Chapters 2 through 5). The strongest effect by efflue nt-exposed site occurred in males for the Fenholloway River system ( Figure 6-7 ). 17 -estradiol was low in females from the upstream Fenholloway River site [U(5)] compar ed to all other sites, and there was no difference between Rice Creek sites. Test osterone was elevated in females before discharge into Fenholloway River compared to the reference site [REF2] but not to the upstream site [U(5)]. Testosterone was al so elevated in effluent-exposed females compared to upstream females in Rice Creek. Males collected at the first downstream site in Fenholloway River [D (5)] had significantly higher 17 -estradiol concentrations compared to both unexposed sites and before discharge [REF2, U(5), PRE-DIS]. Males from Rice Creek had similar sex ster oid concentrations between sites. Sex steroid ratios were most altere d in males from Fenholloway River ( Figure 6-8 ): Twenty-two percent of males from the firs t downstream site [D(5)] and 56% of males

PAGE 204

186 from the second downstream site [D(12)] had fe minized ratios. Average ratios were 1.10 + 0.09 for D(5) and 0.93 + 0.28 for D(12). This response is similar to the response observed in 2001 collections (Chapter 4), when about half the male s (47%) at the second downstream Fenholloway River site [D (12)] had feminized ratios (1.03 + 0.14). In contrast to 2001 data for Fenholloway, skewed steroid ratios were nearly absent in females ( Figure 6-7A ). Sex steroid ratios were gene rally normal for females in both systems: 96% of females had normal femini zed ratios and the average ratio was above one for all sites. Among females with masculin ized ratios, most were from the discharge site in Rice Creek [DIS], at a frequency le ss than 2002 field data (Chapter 3) and in accordance with 2001 field data and 2002 caged da ta (Chapters 4 and 5). This variation in effect on ratios in females, all after ma jor process changes, pr ovides yet another piece of support for dynamic exposure of fish. Anal Fin Elongation and Sex Steroids Sex steroids and anal fin elongation were each altered in different groups of adult mosquitofish in the 2003 population survey. Females from Fenholloway River were the only group to display significant effect of effluent-exposed site on anal fin morphology ( Figure 6-6 ), while males from Fenholloway and fe males from Rice Creek were the only groups to display alterations in sex steroid ratios ( Figure 6-8 ). These results provided even stronger evidence against the predicti ve value of whole body sex steroids as a biomarker for anal fin elongation. Comparing the index of anal fi n elongation between hormonally masculine and feminine groups, th e only statistical di fference was observed at Rice Creek discharge [DIS] where normal fe minized females had a greater index than masculinized females. Of course, overall fema les from this site did not have significant anal fin elongation so the latter point is essentially moot. As discussed in Chapter 3, this

PAGE 205

187 does not rule out alterations in sex steroid ra tios contributing to elongation of the anal fin, since sex steroids were meas ured post-elongation (in Fenholloway females). However, presence of the physiological alteration cannot be used to predict occurrence of anal fin elongation in individual females living in effluent-receiving streams, based upon these data. Fry Production Fry production varied between years, with a general reduction in 2004. Initial 2003 production appeared consistently reduced in a ssociation with effluent-exposed sites for both systems, whereas 2004 production over 4 months was again reduced in Fenholloway River females but elevated in Rice Creek fema les relative to unexposed sites. However, fecundity between unexposed sites from each sy stem varied equally to fecundity within each system, implying different reproductive st rategies. Different patterns of anal fin elongation than those for fecundity suggested anal fin elongation was not predictive of observed effects on reproduction. Summer 2003 Despite mortality and potentia l stress from a water mold infestation, most females produced fry ( Table 6-8 ). In general, primary clutch sizes averaged around 15 to 20 fry and varied widely from one or two fry to seve ral dozen. These results were within ranges reported for mosquitofish (Krumholz 1948, Rosen and Bailey 1963, Hughes 1985, Meffe and Snelson 1989, Specziar 2004). Although mosquitofish are considered nonsuperfetating, a Bahaman mosquitofish species ( Gambusia hubbsi ) demonstrated superfetation as a possible reduction in reproductive costs (Downhower et al. 2002). S uperfetating species produce small broods (around 1 to 5 fry) more frequently than nonsuperf etating species, and a shift in this tactic

PAGE 206

188 could potentially bias compar ison to nonsuperfetating populati ons. Therefore this trait was examined for potential alteration by pulp and paper mill effluents by retaining females for the full 30 days even after primar y production. Fifteen to thirty percent of primary producing females also produced a seco nd clutch of comparable size (for all but Fenholloway River females producing a smaller secondary clutch) within the established 24 to 28 day interbrood interval for nonsuperfetation in this species ( Table 6-8 and Figure 6-10A ). Thus, effluent exposure did not alter this reproductive strategy. Further, sperm storage by female mosquitofish was reaffirmed since females were not exposed to males during the monitoring period. Females differed morphologically between systems and sites in 2003. Females from effluent-exposed sites were smalle r than females from unexposed sites ( Table 6-8 ), and body length correlated with ra w fecundity, or clutch size (r2 = 0.553, p < 0.05). Larger females tend to have larger clutches (Krumholz 1948, Hughes 1985, Meffe and Snelson 1989), therefore clut ch size was divided by standa rd length for each female before statistical analysis. The index of an al fin elongation (length of Ray 4 to Ray 6) was statistically greater at the Fenholloway effluent-exposed site but not at the Rice Creek discharge site, in agreement w ith data from the population survey ( Figure 6-6A ). Fry viability was high based upon few stillb orn/dead fry, low rates of deformity and little change in fry weight Most fry were born live ( Figure 6-9A ) and there were no significant differences in numbers of live vers us dead fry. Deformity rates were very low, usually less than 10% per clutch ( Figure 6-9B ). Relative to viviparous species that lay thousands of eggs (e.g. largemouth bass), mosquitofish invest much more energy into smaller clutches (Constanz 1989) and thus an extremely low rate of deformity was

PAGE 207

189 expected. Deformities were either edema, sk eletal abnormalities (lordosis and scoliosis) or premature abortion of embryos. The statistically significant peak in deformities at the Rice Creek discharge site was caused by one enti re clutch (n = 8) a borted prematurely; no other females displayed abortive behavior Individual fry weight was estimated by weighing clutches and dividing by clutch size. For the first ten clutches measured, individual fry were also we ighed to validate the estimati on. Individual weight was consistently underestimated by 0.1 mg and fina l weights were corrected as such. Fry weight was not affected by effluent-exposed site for primary production in Fenholloway females and secondary production in Rice Creek females, but was significantly depressed in secondary Fenholloway production and primary Rice Creek production ( Figure 6-10B ). This inconsistent result probably reflects weight differences due to clutch size as opposed to effluent-related effects: the few es pecially large clutches (n > 50) had much smaller fry as a constraint on body size of the female. Lower adjusted fecundities occurred at effluent-exposed sites for both stream systems ( Figure 6-10A ). (Since fecundity was analyzed in relation to standard length, data are presented as adjusted fecundity by multiplying each clut ch size to standard length ratio by the average standard length for th e female’s collection site.) However, the likelihood of different reproduc tive stages among sites ( Figure 6-5 ) precluded definitive conclusion based upon these data and dem onstrated the need for monitoring fry production throughout the reproductive season. Summer 2004 Fry production in 2004 was lower than the pr evious year in terms of number of females producing young and overall fecundities ( Table 6-9 and Table 6-10 Figure 6-11 ). Female mortality rates were simila r, and fungal growth was noted on several

PAGE 208

190 females in 2004 although not to the extent of outbreak or in festation requiring treatment in 2003. Females from unexposed sites were not as reproductively active as those from effluent-exposed sites in the May collection, indicated by lower rates of parturition (only two females from upstream Rice Creek produced clutches). Again, this suggests females living in effluent begin repr oducing earlier than females in unexposed sites. Standard length was more similar in females between sites than in 2003, a nd the pattern of body length changes throughout months reinforced the hypothesis of two distinct groups of females reproducing: larger, overwinteri ng females producing fry initially, then young-of-year females beginning to produce at smaller sizes. Anal fin elongation was present in female s from effluent-exposed sites for both systems and varied among months ( Table 6-9 and Table 6-10 Fenholloway River and Rice Creek, respectively). Statistically si gnificant monthly changes in elongation may support the concept of two reproductive classe s of mosquitofish females in populations, and/or differential exposures. The reappearan ce of masculinized anal fins in females from Rice Creek demonstrated this effect wa s not entirely eliminat ed, as implied by 2003 data. Potential exposure was different be tween years: resin and fatty acids were substantially lower in 2004, estimated phyt osterols were equivalent, and TOC, polyphenolics and condensable tannins were higher in 2004. Rainfall was higher and more stable in 2004 than 2003 (Appendix A). Without knowing specific bioactive compounds and fluctuation of effluent concentra tions, linking reappearance of this effect to exposure differences is difficult. Resi n and fatty acids may not be the bioactive components since they were actually higher in 2003 when anal fin elongation was absent, and the increase in TOC, lignin and tannin derivatives might support these compounds as

PAGE 209

191 bioactive agents. At the mi nimum, bioactive compounds remain present in Rice Creek effluent-exposed sites, perhaps teetering on th e threshold of the masc ulinization response. From the perspective of degraded ph ytosterols, the unknown contribution of variable bacterial communities may have b een influential since concentrations of phytosterols were similar. Yet quantification of these compounds was difficult and may not accurately reflect potential exposure. Regarding bacterial contributions, a separate study of -sitosterol degradation in April 2004, exam ining water samples from field sites just before fish collections began, determ ined microorganisms capable of degradation were present in both effluent-receiving sy stems and reference sites (Quinn 2004). Degradation half-life was fastest for Rice Creek water sampled at the discharge (22 to 24 days); slightly longer and mo re variable for Fenholloway River water (24 to 29 days); and slowest at reference sites (32 to 41 days ). This study provided strong evidence that bacterial communities differed between effluent-receiving systems, and that degradation can occur in unexposed systems. Thus the s upposition of variable bacterial degradation rates of similar phytosterol c oncentrations is plausible. Fry vitality was high again in 2004, base d upon percentages of fry born live (data not shown). Percentages were statistically lower at the Fenholloway downstream site June through August (Fisher’s Exact test), averaging 95% versus 99 to 100% at the reference site. Percentages were not differe nt between sites at Rice Creek and averaged 99% for the upstream site and 98% for the discharge site. Biological relevance of differences for Fenholloway River is probabl y low, since most females produced 100% live fry and dead fry across all sites occurred as single members of a clutch in most cases as opposed to the total death of a clutch.

PAGE 210

192 Since standard length positively correlated with raw fecundity overall (r2 = 0.434 and p < 0.050; females from upstream Ri ce Creek did not demonstrate a positive correlation dragging the overall correlation down slightly) fecundity was adjusted by length ( Figure 6-10 ). Site and month significantly c ovaried for the Fenholloway system but not for the Rice Creek system. Different patterns of fecundity were evident for Fenholloway, while Rice Creek females displaye d similar patterns across months. At the reference site for Fenholloway River, fecundity was relatively low for three of the four months with a large spike in July. May and August f ecundities were greater at the effluent-exposed downstream Fenholloway site across months and between sites, but overall fecundity was statistically reduced at th e effluent-exposed site In contrast, both Rice Creek sites peaked in fecundity later in the summer (August), and fecundity was increased relative to upstream in June at the effluent-exposed site. Overall fecundity at Rice Creek was higher in efflue nt-exposed fish. Thus the reduced fecundity observed in 2003 was reflected in 2004 sampling of th e Fenholloway River but the opposite was detected in Rice Creek, supporting the need for long-term observation of fry production over the reproductive season. In general thes e differences were related to reduced fecundities between years at both unexposed sites and the Fenholloway downstream site, while fecundity remained stable at the Rice Creek discharge site ( Figure 6-10 and Figure 6-11 ). Overall these data suggested diffe rences in reproductive output over the reproductive season and from year to year, as opposed to effluent-exposure, influenced fry production in female mosquitofish. Anal Fin Elongation and Fry Production For both years of fry production, overall fec undity (fry divided by female standard length) did not correlate with anal fin elongation (Ray 4 to Ray 6 length ratio);

PAGE 211

193 r2 = 0.0609 for 2003 and r2 = -0.0498 for 2004 with both p > 0.05. Examining fecundity by site for both years and by month for 2004 di d not reveal any stat istically significant correlations either, with Pearson's correla tion coefficient ranging from –0.344 to 0.204 and p>0.05. Furthermore, females with the greatest degree of anal fin elongation > 1.3 (all but 5 from effluent-exposed sites) produced fry in comparable numbers to all other females (ttest, p < 0.05). Therefore, anal fin elongation did not in fluence fecundity of female mosquitofish. Conclusions These studies represent the first examina tion of actual fry production as a measure of reproductive success in mosquitofish e xposed to pulp and paper mill effluents. Previous investigations of reproduction in this species under e ffluent exposure were confined to brood size of developing embr yos or histological evaluation of gonads, neither of which conclusively detected advers e impacts. Further, the current studies are the first to assess masculinization in addition to measures of reproduction. Taking these data as a whole, reproductiv e success of mosquitofish is likely not impacted by pulp and paper mill effluent exposure. Rather, seasonal differences in fecundities suggested mosquitofish may have adapted different site-specific reproductive strategies. Examining data by system, Fe nholloway River females displayed reduced fecundity and masculinized anal fins for both years. In contrast, Rice Creek females were less fecund in 2003 data, when masculinized anal fins were absent, but more fecund across months in 2004 when elongated anal fins returned. Combined with the overall lack of correlation between fecundity and an al fin elongation, anal fin elongation cannot be definitively tied to al terations in reproduction. Examining fecundity over time revealed site-specific patterns of produc tion, and age/sex struct ure in the population

PAGE 212

194 survey suggested effluent-exposed females may begin the reproductive season earlier than females living in nonexposed sites, pe rhaps caused by higher water temperatures. Increased temperature strongly triggers onset of the reproductive season (Koya and Kamiya 2000, Koya and Iwase 2004), and has been associated with an overall increase in reproductive output (Vondracek et al. 1988). Differences in seasonal reproductive pattern s have been described for mosquitofish populations living in unexposed conditions under the influence of different predation and food availability (Vondracek et al. 1988, Down hower et al. 2000). Further, Downhower et al. (2000) detected rapid phe notypic adjustment or plasticity in reproductive strategies for populations introduced to predator-free habi tats in less than 20 years. Therefore, ecological differences among sites caused by long-term effluent dominance could influence fecundity. For example, increased turbidity at effluent-dominated sites may decrease predation risk of mosquitofish; eu trophication of effluent-receiving systems may increase food availability as well. The co mbined effect of these types of ecological factors likely alters reproductive investment s and strategies and may explain observed variation in fecundities. Va riation in fecundity over the 2004 reproductive season within a site also supports the con cept of two separate reproduci ng populations, overwintering and young-of-year females (Hughes 1985, Haynes and Cashner 1995, Fernndez-Delgado and Rossomanno 1997), st ressing the importance of documenting reproduction throughout th e reproductive season. Studies evaluating effects of other contaminants on mosquitofish fry production suggested reproductive tolerance of this spec ies to exposure. Expe rimental exposure of adult mosquitofish to sublethal conc entrations of the nonionic surfactant Genapol

PAGE 213

195 OXD-080 (used as an agricultural pesticide) did not affect fry survival (Cabral et al. 1999). Field collection of mosquitofish living in coal ash settling basins did not have altered fecundities or fry viability despite el evated concentrations of metals in both females and fry (Staub et al. 2004). The latt er point suggested ma ternal transfer of contaminants from females to fry, which wa s also detected afte r dietary exposure of female mosquitofish to 4-nonylphenol (Thi baut et al. 2002), a weakly estrogenic microbial degradation product of industrial nonionic su rfactants that has been detected in Rice Creek (Nimrod and Benson 1996, Quinn 2004). In contrast to above reports, water-b orne exposure to 4-nonylphenol disrupted normal gonadal differentiation in maturing mosqu itofish at highest test concentrations (50 g/L) with a skew toward females, and lower test concentrations (0.5 and 5.0 g/L) were associated with partially-develope d gonopodium in fish with atrophied gonads (Drze et al. 2000). Gender of these latter fish was unclear, but this work insinuates estrogenic compounds detected in Rice Cr eek receiving waters could potentially contribute to effects on anal fin morphology and reproduction. Since a low level but persistent estrogen bias in male sex st eroid ratios was observed throughout our study, further examination of estrogenic effects of effluent components on mosquitofish is warranted. Before making a definitive conclusion about reproductive success based on fecundity, the picture needs to be widened ev en further to document onset and cessation of reproduction across sites. Ideally, populat ion-level studies inve stigating energetic investments in reproduction would be incl uded to address pote ntial variation in reproductive strategies at effluent-exposed si tes. It is possible an earlier onset of

PAGE 214

196 reproduction in effluent-exposed fish at Fenholloway River may counteract the lower fecundities observed in these studies. Pre liminary relative abundance data indicated greatly increased density at the downstr eam Fenholloway site in early summer 2003 (May) so an earlier onset of reproduction is tentatively supported. Also, the overall reduced fecundities between years suggested additional environmental factors were negatively influencing fry production. L ong term microcosm experiments studying population structure and recruitment dynamics using the tank facility at Rice Creek (Chapter 4) would be very us eful in determining these t ypes of differential reproductive strategies, and females could be subsampled periodically for fry production studies that could mimic experimental predation on populations. Bioindicator potential for mosquitofish was weakened by initial reproductive success studies that could not link anal fin elongation with al tered fecundity. Differences in fecundity may ultimately reflect adaptation of reproductive strategy in effluent-exposed fish, as opposed to negativ e impacts on reproductive success. Thus, any alterations in anal fin morphology and sex ster oids could serve as biomarkers indicating exposure rather than adverse effect.

PAGE 215

197 Table 6-1. Water quality parameters of Fenholloway River and Rice Creek population survey sites during summer 2003. Fenholloway River Rice Creek Site REF2 U(5) PRE-DIS D(5) D(12) U(8) DIS Water Temperature (C) 21.4 20.8 31.0 27.9 24.7 22.0 26.7 Conductivity (S) 278 80.5 2,229 1,820 1257 289 1,862 Salinity (ppt) 0.1 0.0 1.1 0.9 0.6 0.1 0.9 Turbidity (ntu) 0.62 1.32 12.0 8.18 5.79 1.03 26.5 pH 7.1 4.9 7.6 7.2 7.2 7.3 7.9 Table 6-2. Water quality parameters (ave + se) measured three times weekly (n = 16 total) during laboratory fry production of female mosquitofish collected from field sites in Rice Creek during summer 2003. Fry Production Tanks Temperature (C) 26.1+ 0.12 Conductivity (S) 310.7+ 1.8 Salinity (ppt) 0.1+ 0.02 Dissolved Oxygen (mg/L) 6.29+ 0.16 Turbidity (ntu)a 1.38+ 0.19 pH 8.0+ 0.04 Incident Light (Fc) 83.4+ 0.9 ameasured once a week

PAGE 216

198 Table 6-3. Water quality parameters (ave + se) at field sites where female mosquitofish were collected for fry production stud ies over 4 months in summer 2004. All parameters were statistically differe nt between exposed and unexposed sites within each system. Fenholloway River Rice Creek Site REF2 D(5) U(8) DIS Temperature (C) 23.1+ 0.7 28.8+ 1.1 23.2+ 1.5 27.1+ 1.5 Conductivity (S) 290.1+ 70.7 2,085+ 216.3 185.8+ 30.3 1,780+ 78.8 Salinity (ppt) 0.2+ 0.02 1.1+ 0.1 0.1+ 0.03 0.9+ 0.03 Dissolved Oxygen (mg/L) 4.50+ 0.27 0.92+ 0.7 7.20+ 0.49 8.59+ 1.03 Turbidity (ntu) 2.04+ 0.19 13.98+ 1.46 7.67+ 2.91 27.2+ 0.95 pHa 7.3+ 0.1 7.2+ 0.2 7.1+ 0.4 7.9+ 0.1 apH statistically differe nt at Rice Creek only Table 6-4. Water quality parameters (ave + se) measured three times weekly (n = 16 total) during laboratory fry production of female mosquitofish collected from field sites in Rice Creek and Fe nholloway River during summer 2004. Month May June July August Temperature (C) 25.3+ 0.3 25.4+ 0.4 24.7+ 0.1 24.4+ 0.3 Conductivity (S) 238.7+ 4.1a 260.9+ 12.5a 329.4+ 5.2 323.5+ 8.4 Salinity (ppt) 0.1+ 0.0a 0.1+ 0.01a 0.2+ 0.02 0.2+ 0.02 Dissolved Oxygen (mg/L) 6.76+ 0.17b 5.29+ 0.32 5.67+ 0.15 7.67+ 0.02 Turbidity (ntu) 1.01+ 0.20c 0.79+ 0.23 0.43+ 0.02 0.26+ 0.11 pH 8.6+ 0.06d 7.9+ 0.01c 7.7+ 0.04 7.7+ 0.02 asignificantly different from July and August (p < 0.05) bsignificantly different from June and July (p < 0.05) csignificantly different from August (p < 0.05) dsignificantly different from all other months (p < 0.05)

PAGE 217

199Table 6-5. Concentrations of sel ected effluent components (ave + se) in single grab water sample s from field sites where female mosquitofish were collected in Rice Creek and Fenholloway River during summer 2003. Statis tical significance (p < 0.05) is given by system. System Fenholloway River Rice Creek Site REF2 (n = 3) U(5) (n = 5) PRE-DIS (n = 3) D(5) (n = 7) D(12) (n = 3) U(8) (n = 7) DIS (n = 7) Total Resin Acids (g/L) NDj 1.29+ 1.29 125.46+ 6.94b 78.11+ 22.83b 57.34+ 4.21 0.20+ 0.13 98.01+ 24.44k Total Fatty Acids (g/L) 2.68+ 2.68 10.19+ 6.16 1.50+ 0.75 2.45+ 0.41 1.72+ 1.16 6.01+ 1.93 8.33+ 2.65 Campesterol (g/L) a ND ND 6.18+ 0.75c 3.12+ 0.86d 3.45+ 0.78b ND 0.68+ 0.18 Stigmasterol (g/L) a 0.27+ 0.27 1.39+ 0.15 13.26+ 2.44c 6.51+ 1.76d 7.26+ 1.58e ND 2.36+ 0.28* Stigmastanol (g/L) a ND ND 14.13+ 0.75f 7.19+ 1.83d 8.30+ 1.49b ND 2.72+ 0.32* -sitosterol (g/L) a 1.07+ 0.36 1.75+ 0.26 106.6+ 12.2c 54.58+ 14.15d 63.1+ 1.49b 0.19+ 0.0.13 13.97+ 2.30* Total Organic Carbon (mg/L) 41.0+ 8.03 109.1+ 14.8 154.7+ 8.6f 99.8+ 6.8g 85.2+ 0.8h 31.2+ 8.8 62.9+ 2.6* Polyphenolics (mg/L) 6.98+ 1.33 12.5+ 1.6 42.9+ 0.3.7f 25.9+ 2.6d 17.1+ 2.8g 4.9+ 1.2 21.4+ 2.9* Condensable Tannins (mg/L) 1.49+ 0.14 3.71+ 0.16 7.76+ 0.34i 5.02+ 0.37e 3.67+ 0.29 2.05+ 0.91 3.11+ 0.57 aphytosterol values for Rice Creek discharge site [DIS] failed quality control tests therefore are estimates bdifferent from REF2 and U(5) cdifferent from REF2, U(5), D(5) ddifferent from REF2, U(5), PRE-DIS edifferent from REF2 fdifferent from rest gdifferent from REF2, PRE-DIS hdifferent from U(5), PRE-DIS idifferent from REF2, U(5), D(12) jND = nondetectable kdifferent from U(8)

PAGE 218

200Table 6-6. Concentrations of sel ected effluent components (ave + se) in single grab water sample s from field sites where female mosquitofish were collected in Rice Creek and Fenhollo way River during summer 2004 fo r fry production studies. Statistical significance (p < 0.05) is given by system. System Fenholloway River Rice Creek Site REF2 (n = 3) D(5) (n = 3) U(8) (n = 3) DIS (n = 3) Total Resin Acids (g/L) 0.34+ 0.34 367.36+ 109.16d 0.30+ 0.30 44.97+ 6.79* Total Fatty Acids (g/L) 0.26+ 0.26 3.62+ 2.46 2.74+ 2.18 1.24+ 1.24 Campesterol (g/L)a NDb 2.07+ 1.11 ND ND Stigmasterol (g/L)a ND 5.75+ 1.85* ND 3.11+ 0.34* Stigmastanol (g/L)a ND 5.71+ 1.38* ND 3.56+ 0.39* -sitosterol (g/L)a 2.82+ 2.82 50.20+ 13.73* 0.46+ 0.46 14.53+ 0.79* Total organic carbon (mg/L)c 35.0+ 19.1 139.0+ 4.4* 44.8+ 15.1 91.0+ 15.7 Polyphenolics (mg/L) 4.20+ 2.12 35.83+ 5.38* 3.29+ 1.85 27.23+ 0.69* Condensable Tannins (mg/L) 1.36+ 0.77 8.03+ 0.25* 2.00+ 0.81 6.23+ 0.58* aphytosterol values for both Fenholloway River sites [REF2 and D(5)] and Rice Creek discharge site [DIS] failed quality control tests therefore are estimates bND = nondetectable cTOC values for August were above maximum temperature re quired for quality control a nd thus are only estimates ddifferent from refere nce [REF2 or U(8)]

PAGE 219

201Table 6-7. Body size parameters (ave + se ) for mosquitofish collected for population survey of Fenholloway River and Rice Cree k in May 2003. Significant differences (p < 0.05) are noted by site within each system. Fenholloway River Rice Creek Site REF2 U(5) PRE-DIS D(5) D(12) U(8) DIS Sample Size 10 10 10 10 10 10 10 Body Weight (g) 0.211+ 0.017 0.168+ 0.022 0.167+ 0.015 0.236+ 0.015 0.230+ 0.026 0.233+ 0.022 0.205+ 0.016 Standard Length (mm) 22.74+ 0.83 22.13+ 0.95 22.22+ 0.60 23.37+ 0.47 22.43+ 0.63 23.60+ 0.65 21.99+ 0.56 Condition Factor (g/cm3) 1.82+ 0.16 1.66+ 0.04 1.48+ 0.05a 1.85+ 0.10 1.99+ 0.16 1.73+ 0.05 1.90+ 0.07 Sample Size 23 24 24 25 25 25 13 Body Weight (g) 0.531+ 0.035 0.335+ 0.030b 0.589+ 0.057 0.754.+ 0.106 0.628+ 0.049 0.450+ 0.026 0.447+ 0.067 Standard Length (mm) 31.02+ 0.72 28.17+ 0.86 30.88+ 0.91 31.88+ 1.21c 30.45+ 0.69 28.42+ 0.56 27.78+ 1.36 Condition Factor (g/cm3) 1.74+ 0.04 1.45+ 0.04b 1.87+ 0.05c 2.12+ 0.04d 2.14+ 0.05d 1.92+ 0.03 1.91+ 0.04 astatistically differs from REF2 bstatistically differs from rest cstatistically differs from U(5) dstatistically differs from non“d” lettered sites

PAGE 220

202 Table 6-8. Reproductive and morphological ch aracteristics of females collected from Fenholloway River and Rice Creek and monitored for fry production in 2003. Fenholloway River Rice Creek Site REF3 D(5) U(8) DIS # Start 51 51 52 29 % Mortality 37% 27% 22% 10% # End 32 37 41 26 % Parturitiona 91% 92% 100% 85% Total Fry (1 production) 689 531 787 197 1 Clutch Size Range 4–62 1102 265 147 % with 2 Productionb 14% 30% 18% 14% Total Fry (2 production) 47 129 104 27 2 Clutch Size Range 420 629 825 515 Median Interbrood Interval (days) 24 24 25 27 Standard Length (mm)c 30.86+ 0.52 28.21+ 0.57d 31.80+ 0.61 28.25+ 0.99d Index of Anal Fin Elongation (Ray 4/Ray 6)c 1.17+ 0.01 1.54+ 0.06d 1.16+ 0.01 1.19+ 0.01 areferring to primary production of surviving females referring to secondary production of females that had primary production cave+ se dsignificantly different than unexposed site (p < 0.05)

PAGE 221

203Table 6-9. Reproductive and morphological ch aracteristics of females collected for fry production from Fenholloway River in 20 04. Significant differences in morphological variables are noted by site or month, and both covaried by site and month (p < 0.05). Fenholloway River Site REF2 D(5) May 2004 June 2004 July 2004 Aug 2004 May 2004 June 2004 July 2004 Aug 2004 # Start 50 50 50 42 50 50 50 49 % Mortality 32% 18% 4% 24% 6% 12% 20% 24% # End 34 41 48 32 47 44 40 37 % Parturitiona 38% 66% 80% 67% 74% 68% 66% 67% Total Fry (1 Production) 100 92 233 135 264 174 221 241 Clutch Size Range 18 7 16 13 20 15 28 20 Standard Length (mm)b 31.31+ 0.31d 26.74+ 0.46d 28.53+ 0.36 28.24+ 0.45 30.07+ 0.39ce 28.76+ 0.75e 24.70+ 0.60c 24.34+ 0.43c Index of Anal Fin Elongation (Ray 4/Ray 6)b 1.17+ 0.01d 1.11+ 0.01 1.12+ 0.01 1.12+ 0.01 1.51+ 0.03c 1.57+ 0.04c 1.51+ 0.04c 1.36+ 0.03cd areferring to primary production of all females bave + se csignificantly different than unexposed site within month dsignificantly different than rest of months within site esignificantly different than July and August within site

PAGE 222

204 Table 6-10. Reproductive and morphological ch aracteristics of females collected for fry production from Rice Creek in 2004. Significant differences are noted by site or month, and index of anal fin elonga tion only covaried by site and month (p < 0.05). Rice Creek Site U(8) DIS May 2004 June 2004 July 2004 Aug 2004 May 2004 June 2004 July 2004 Aug 2004 # Start 50 49 50 35 50 50 34 49 % Mortality 30% 12% 22% 37% 14% 4% 24% 29% # End 35 43 39 22 43 48 26 35 % Parturitiona 4% 67% 68% 57% 28% 74% 62% 71% Total Fry (1 Production) 20 155 627 217 74 141 141 541 Clutch Size Range 19 16 78 62 12 21 19 41 Standard Length (mm)b 28.55+ 0.43e 30.21+ 0.50e 30.00+ 0.72d 29.35+ 0.96d 28.35+ 0.42 28.57+ 0.51c 29.62+ 0.62c 31.89+ 0.54c,d Index of Anal Fin Elongation (Ray 4/Ray 6)b 1.16+ 0.01 1.16+ 0.01 1.13+ 0.01 1.14+ 0.02 1.24+ 0.05 1.21+ 0.01c 1.21+ 0.02c 1.27+ 0.01c areferring to primary production of all females bave + se csignificantly different than unexposed site within month dsignificantly different than rest of months within site esignificantly different than July and August within site

PAGE 223

205 Figure 6-1. Maps of field sites. A) Relative location of the stream systems in Florida. B) Fenholloway River sites. C) Rice Cr eek sites. Site symbols distinguish effluent exposure: circles = unexposed and triangles = exposed. Site abbreviations denote upstream (U) or dow nstream (D) of discharge, followed by approximate distance (km) from di scharge in parentheses. PRE-DIS indicates site before discharge into th e creek; DIS denotes site at discharge into creek; and REF indicates referen ce site, followed by identifying number. A B B C

PAGE 224

206 Figure 6-1. Continued WEEK OF MONTH M AY WK1 MA Y WK2 MAY W K 3 J ULY WK 4 AUG WK2 A UG W K 4 SEPT WK1 DEHYDROABIETIC ACID (ug/L) 0 10 20 30 40 50 REF2 U(5) PRE-DIS D(5) D(12) Rice Creek U(8) Rice Creek DIS Figure 6-2. Representative changes in resi n acid concentrations during summer 2003 at Fenholloway River and Rice Creek fiel d sites where mosquitofish were collected. Values at unexposed sites we re nondetect (at or below 2 ug/L) and may not be visible underneath Rice Cr eek U(8) symbols. Red underline indicates fish collection for populati on survey (May) and fry production (July). C

PAGE 225

207 MONTH JUNEJULYAUG DEHYDROABIETIC ACID (ug/L) 0 20 40 60 80 100 120 Fenholloway River REF2 Fenholloway River D(5) Rice Creek U(8) Rice Creek DIS Figure 6-3. Representative changes in resi n acid concentrations during summer 2004 at Fenholloway River and Rice Creek fiel d sites were mosquitofish were collected at the same time. Values at Fenholloway River reference site [REF2] were nondetect (at or belo w 1.0 ug/L) and may not be visible underneath Rice Creek U(8) symbols.

PAGE 226

208 SITE REF2U(5)PRE-DISD(5)D(12)RC U(8)RC DIS MEDIAN CPUE 0 10 20 30 40 50 60 Figure 6-4. Estimated relative abundances of mosquitofish for Fenholloway River and Rice Creek (RC) sites. Solid black line separates the two systems. Abundances were calculated as median catch per unit effort (CPUE) based upon number of fish per 10 sweeps by obser ver (3 to 5 observers per site).

PAGE 227

209 U(8)40% 19% 1% 40% 40% JUVENILE MALE GRAVID FEMALE NONGRAVID FEMALE DIS57% 23% 20% 12% 8% REF220% 50% 30% 24% 6% U(5)31% 14% 0% 55% 55% PRE-DIS23% 27% 50% 9% 41% D(5)60% 25% 15% 6% 9% D(12)47% 30% 23% 10% 13% Figure 6-5. Estimated age and sex structure of mosquitofish populations living near pulp and paper mill effluent discharge. A) Rice Creek. Distributions were significantly different ( 2=34.03, df=3, p<0.05 Fisher’s Exact Test). B) Fenholloway River. Distributions were significantly different ( 2=200.2, df=12, p<0.05 Chi Square Test for I ndependence). Site abbreviation descriptions are given in Figure 6-1 A B

PAGE 228

210 SITE R E F2 U(5) P RE -D IS D(5) D(12 ) R C U( 8 ) RC DIS INDEX OF ANAL FIN ELONGATION 0.0 0.5 1.0 1.5 2.0 2.5 3.0 SITE REF2 U( 5 ) PRE-DIS D ( 5) D (1 2) RC U(8) R C DIS INDEX OF ANAL FIN ELONGATION 0.0 0.5 1.0 1.5 2.0 2.5 3.0 a a b A B Figure 6-6. Index of anal fin elongation (t racings of Ray 4 / Ray 6, computer-aided measurement of fresh fish) for mosqu itofish collected in summer 2003 from Fenholloway River and Rice Creek (RC). Stream systems are divided by a solid black line. A) Females. B) Males. Letters indicate significant differences by site within system (p < 0.05): “a” denotes differences to all sites but PRE-DIS (Fenholloway River sy stem); “b” denotes differences to REF2 and U(5) (Fenholloway River system).

PAGE 229

211 SITE REF 2 U ( 5 ) P R ED IS D ( 5 ) D ( 1 2 ) U ( 8 ) DIS WHOLE BODY SEX STEROIDS (pg/g) 0 200 400 600 800 1000 1200 1400 1600 1800 2000 17beta-estradiol Testosterone SITE R EF2 U(5) PRED IS D(5) D(12) U(8) D I S WHOLE BODY SEX STEROIDS (pg/g) 0 200 400 600 800 1000 1200 B A a b c d Figure 6-7. Whole body sex steroids (ave + se) for mosquitofish collected in summer 2003 from Fenholloway River and Rice Cr eek. Stream systems divided by solid black line. A) Females. B) Males. Letters indicate significant differences by site within mill (p < 0.05) : “a” denotes differences to all other sites; “b” denotes differences to REF2; “c” denotes differences to U(8); “d” denotes differences to REF2, U(5), PRE-DIS.

PAGE 230

212 % FEMALES 0102030405060708090100 SITE REF2 U(5) PRE-DIS D(5) D(12) U(8) DIS MASCULINE STEROID RATIO (E/T<1) FEMININE STEROID RATIO (E/T>1) % MALES 0102030405060708090100 SITE REF2 U(5) PRE-DIS D(5) D(12) U(8) DIS A B 1.20/1.18+ 0.01 1.15+ 0.01/1.24+ 0.02 1.32/1.69+ 0.07 2.67+ 0.05/2.80+ 0.03 2.73+ 0.12/2.79+ 0.03* Figure 6-8. Percentage of mo squitofish with masculine and feminine sex steroid ratios collected in summer 2003 from Fenhollo way River and Rice Creek (systems divided by solid black line). A) Females. Frequency of ratios was statistically significant for Rice Creek (Fisher’s Exact test, p < 0.05) and Fenholloway River ( 2 = 20.20, df = 4, p < 0.05). B) Male s. Frequency of ratios was statistically significant for Fenholloway River ( 2 = 182.5, df = 4, p < 0.05). Index of anal fin elongation (ave + se) is given for fish with both masculine and feminine ratios by site. Red asteri sk indicates significa nt differences in elongation at p < 0.05 between ratios.

PAGE 231

213 SITE REF3D(5)U(8)DIS % FRY ALIVE 0 10 20 30 40 50 60 70 80 90 100 PRIMARY SECONDARY SITE REF3D(5)U(8)DIS % FRY DEFORMED 0 5 10 15 20 25 Figure 6-9. Viability of primary and seconda ry clutches produced by females collected from Fenholloway River and Rice Creek in 2003. A) Average percentages (+se) of live fry per female within one day of parturition. B) Average percentages (+se) of deformed fry per female. Field sites are divided by a solid black line (Fenholloway River sites on the left and Rice Creek on the right). Missing bars mean there were zer o deformed fry. Light blue asterisk indicates statistical difference from primary production at upstream Rice Creek site (for total fry deformed vers us normal within each site, Fisher’s Exact test, p < 0.05).

PAGE 232

214 SITE REF3D(5)U(8)DIS ADJUSTED FECUNDITY (# FRY / FEMALE) 0 5 10 15 20 25 PRIMARY SECONDARY A SITE REF3D(5)U(8)DIS INDIVIDUAL FRY WEIGHT (mg) 0 2 4 6 8 10 12 B * * * Figure 6-10. Fecundity and i ndividual fry weight of prim ary and secondary clutches produced by female mosquitofish coll ected from Fenholloway River and Rice Creek in summer 2003. A) Adjusted aver age fecundity (+se) or number of fry corrected for standard lengt h of individual females. B) Average individual fry weight (+se) calculated by wei ghing clutches and dividing by raw fecundity. Field sites are divided by a solid black line (Fenholloway River sites on the left and Rice Creek on the right ). Asterisks are color coded within primary and secondary production and i ndicate significant differences to unexposed site within each system (p < 0.05).

PAGE 233

215 MONTH MAYJUNEJULYAUGMAYJUNEJULYAUG ADJUSTED FECUNDITY (# FRY / FEMALE) 0 5 10 REF2 D(5)A *a*b* SITE MAYJUNEJULYAUGMAYJUNEJULYAUG ADJUSTED FECUNDITY (# FRY / FEMALE) 0 5 10 15 20 U(8) DIS* B7+ 16+ 1 5+ 0 3+ 0* *INDEX OF ANAL FIN ELONGATION 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 RAY 4 / RAY 6 INDEX OF ANAL FIN ELONGATION 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 b a a a Figure 6-11. Adjusted fecundity of primar y clutches produced by female mosquitofish collected monthly in 2004. A) Fenhollo way River. B) Rice Creek. Solid black line separates sites w ithin each system. Solid red circles indicate index of anal fin elongation (ave + se) for females producing fry. Site symbols are given in upper left-hand corners, while average (+ se) adjusted fecundity for sites are given in upper right-hand corner s. Significant differences by month within a site are denoted by letters: “ a” is different than May and June; and “b” is different than May. Blue aste risks indicate signi ficant differences between sites within a month (above bars ) or for the season (next to overall average) for each system. Green asterisk indicates significant interaction or covariance by site and month.

PAGE 234

216 CHAPTER 7 EVALUATION OF MOSQUITOFISH AS A BIOINDICATOR OF PULP AND PAPER MILL EFFLUENT EXPOSURE The goal of my study was to evaluate whethe r sublethal effects in mosquitofish can be reliably used to indicate adverse impacts of pulp and paper mill effluents. This chapter summarizes conclusions form previous research chapters, and then evaluates this research in light of specific aims and suitability of mosquitofish as a bioindicator species. In addition, mosquitofish are compared to ot her fish models that respond to pulp and paper mill effluents, and future studies ar e proposed that would resolve inconclusive aspects and address questions revealed by this work. My research has contributed several ke y points of knowledge about mosquitofish exposed to pulp and paper mill effluents. Anal fin morphology was only affected in females, not males. There was no appare nt precocious maturation among males exposed to mill effluents. Masculinization of the female anal fin is sensitive to differences in effluent composition: relative concentrations of wood extractives reflected degree of anal fin elongation. Also, the concept of a thre shold response was supported by an ephemeral occurrence of elongation in females from Rice Creek. Masculinization was not seasonally affected, as previously reporte d, indicating this endpoint may serve as a biomarker of past (chronic) exposure or at a po tentially sensitive life stage. In contrast, sex steroids may serve as a biomarker of current or recent exposure, especially considering whole effluent exposure resu lts. Both sexes responded hormonally, providing strong evidence for multiple endocri ne-mediated mechanisms not just an

PAGE 235

217 androgen-mediated response. Neither sex ster oids nor fry production were related to anal fin elongation in females, indicating mascu linization is not predictive of current physiology or ultimately reproductive output. Summary Gender identification techniques, anal fi n morphological measurements and whole body sex steroid analyses were suitably va lidated from a methodological viewpoint to support use of mosquitofish as a bioindicator of pulp and paper mill effluents. Seasonal changes in sex steroids, but not alterations in female an al fin morphology, suggested sex steroids may be a more sensitive and labile biomarker of differential effluent exposure, while female anal fin elongation may be a more static and historical biomarker. Across all studies, effects in males were restricted to changes in sex steroids, mainly elevated 17 -estradiol, accompanied by a shift to estrogen-biased steroid ratios. Precocious maturation of males was not supported by these studies. Female sex steroids generally responded with a masculinized steroid profile caused by increased testosterone. Analysis of female anal fin morphology by size class as an estimation of age did not consistently reveal a more sensitive adult life stage; however, juvenile versus adult exposure was never addressed and could potentially influence extent of elongation. Extensive field work in receiving stream s of one mill throughout process changes and among several mills indicated reductions in wood extractives in pulp mill effluents is associated with reduced, but not entirely elim inated, masculinization of the female anal fin. Masculinized hormone profiles in fema les and feminized hormone profiles in males were consistently detected for all but one mill where masculinization of the female anal fin was the least. Skewed steroid ratios at reference sites for the system with greatest

PAGE 236

218 degree of morphological masculinization was ve ry important and implied the estrogen to testosterone ratio can be alte red by environmental factors completely separate from pulp and paper mill effluents. Equally important, fish collected before discharge into receiving streams (i.e., retention ponds) did not always respond to the degree of fish collected in the receiving str eam. In general, reference site selection was an important aspect of determining statistically sign ificant responses, and comparison between unexposed sites demonstrated inherent natura l variability of biomarkers. No clear relationship existed between anal fin el ongation and whole body primary sex steroid concentrations or ratios, although this conclu sion does not mean development of anal fin elongation in females is independent of sex steroid concentrations. It does mean sex steroids are not predicti ve of anal fin elongation. Controlled whole effluent exposure in tanks and in situ field exposure in cages did not induce female anal fin elongation, while sex steroids in both genders were differentially altered between tanks versus cages. Masculin ization of the anal fin was likely not induced because of abbreviated e xposure; but the unexplored possibility of sensitive life stage(s) could also explain this lack of induction. Skewed sex steroid ratios were induced sooner in cages than tanks, and st eroids had recovered in caged fish by the end of the study. However, compared to field collections, steroid responses under controlled exposure were more pronounced. This study reinforced the idea of dynamic exposure dependent on environmental factor (s) in addition to effluent exposure. Finally, reproductive success was investigat ed over two years in conjunction with masculinization biomarkers. Fry production of wild-caught female mosquitofish was evaluated at one time point the first year and over several months the second year. A

PAGE 237

219 preliminary population survey was also c onducted the first year Morphological masculinization was consistent between years for one mill while the response was only detected the second year of st udy at the other mill. Sex steroid alterations were weakly affected relative to measurements in previous studies. Neither of these biomarkers could be associated with fry production or populati on structure differences. Initially, fry production appeared decreased at effluent-exposed sites but p opulation structures implied mosquitofish at effluent-exposed sites may ha ve started reproducing sooner than fish at unexposed sites. Fry production over severa l months the following year affirmed different reproductive patterns in females am ong sites and throughout th e year. Further, overall fecundities were higher in females from one exposed site and lower from the other relative to respective references. Rather than negatively impacting fecundity, pulp and paper mill effluent exposure may be stimul ating distinctive reproductive strategies in mosquitofish influenced by changes in envi ronmental and ecologica l factors as opposed to chemical exposure. Specific Aims Revisited Explicitly stated objectives for my study we re divided into two specific aims with associated hypotheses. Within each specifi c aim, three subaims were identified and studies developed for each. Broadly speaki ng, these aims were ambitious and reality dictated focus on one main aspect, variati on in masculinization response in the field. The first specific aim, to determine the effects of impr oved mill technology on masculinization of female mosquitofish, was sa tisfied for two of three subaims associated with this field work. Because of complicati ons with induction studies, field studies could neither be supported nor refuted by more cont rolled exposure. Th rough field studies, the hypothesis that reduction in br own side effluent components (i.e., wood extractives such

PAGE 238

220 as phytosterols and resin acids) will reduce an al fin elongation and hormonal alteration in female mosquitofish was supported. An im portant limitation to this work was the increasingly apparent seasonality of hormone concentrations that remained uncharacterized. Variations in response, potential exposure, and environmental factors (such as precipitation) suggested a scenario of dynamic exposure to pulp mill effluents. The second specific aim, to evaluate the reproductive success of mosquitofish exposed to pulp and paper mill effluents, and associated subaims were too broad for practical purposes and efforts were subseque ntly narrowed down toward fry production studies. Rather than provide a conclusive an swer to the hypothesis that exposure to pulp and paper mill effluents will not impair reproductive success of mosquitofish, this research posed questions about adaptation of reproductive strategies under effluent exposure. Similar to complications experienced with i nduction studies for masculinization in the first specific aim, controlled (caged) exposures designed to evaluate fry production were not completed. Thus these data require controlled exposures for more support. Nonethele ss, these studies did not demonstrate a relationship between fecund ity and masculinization. Overall, the proposed work was skewed to ward field work and away from whole effluent exposures. Controlled exposure to effluent components in the laboratory was avoided in favor of more relevant whole effluent exposures, since mechanistic-based questions were not directly addressed. However, difficulties with cages used for in situ field exposures created a bias in available data. This bias meant conclusions could be influenced by uncontrollable f actors inherent in field work such as poten tial population

PAGE 239

221 differences and unknown actual exposures. The ideal of paired field and laboratory work was reinforced by this inadvertent shortcoming. Bioindicator Criteria Revisited Based upon contributions of my study, mosqu itofish are not adversely affected by pulp and paper mill effluents. Therefore, at this point mosquitofish biomarkers may be used to indicate exposure but not effect. Wh ile responses were detected in both anal fin morphology and whole body sex steroids, a rela tionship between these parameters could not be established. Method variability was accep table but natural variability was high for sex steroids, therefore detection of change s due to effluent exposure were obscured especially in light of unchara cterized seasonal effects. Ma sculinization of the anal fin was sensitive to differences caused by changing effluent compositions, but the unique nature of the systems studied (low-flow, effluent dominated streams) questions applicability to more average effluent -receiving systems. Finally, impacts on reproduction were not consistent, and fecundity differences may actua lly reflect different reproductive strategies among sites. Whole effluent exposures by this laboratory and other researchers (Ellis et al. 2003, McCarthy et al. 2004, van den Huevel et al 2004b) have not c onsistently supported observations in the field relativ e to anal fin masculinization. This is another point that precludes implementation of mosquitofish as a bioindicator. Curre nt exposures were based on the premise that masculinization occurs in adult females as indicated by laboratory exposure to bacterially degraded effluent components (Denton et al. 1985, Howell and Denton 1989). Exposure throughout maturation may reveal juveniles are a more sensitive life stage.

PAGE 240

222 Pending such exposures and further analysis of seasonality in hormones, anal fin morphology and sex steroids could potentiall y be used on a mill-specific level as biomarkers of exposure. Candidate mills would include effluent-dominated receiving streams and receiving streams in devel oping countries lacking modern processing technologies. Sex steroids could serve as a marker of recent exposure, while anal fin morphology could serve as a more static ma rker of previous, longer-term exposure. A final point about utility of mosquitofish as a bioi ndicator is the biological significance of observed effects. Higher level impacts aside, what extent of anal fin elongation is biologically relevant ? Statistical differences con tinue to be detected, yet the degree of masculinization has lessened considerably from initial reports of fully formed gonopodia (Howell et al. 1980, Bortone and Dr ysdale 1989, Cody and Bortone 1992). Under dietary exposure to 11keto-testosterone female Ray 4 to Ray 6 length ratios ranged from 1.35 to 1.50 at 20 to 100 g/g f eed (Angus et al. 2001), which is at least 1 mm less than normal male ratios that averag ed 2.5. Therefore biologically significant differences based upon exposed females woul d be a smaller magnitude change than differences based upon normal male gonopodial le ngth. This debate ironically centers upon quantification of the mascu linization response, which was previously assumed to be a more appropriate and accurate measure of masculinization than categorically or nominally scored responses in previous st udies. Yet quantifying the response may have led to erroneous conclusions. Thus, biologi cal significance may be retained in future studies by blending the two types of measur ements and scoring females by absence and presence/extent of elongation, then measur ing ray lengths for each of these groups.

PAGE 241

223 My study addresses applicabilit y and relevance within the species, but applicability to other species remains to be studied. Mos quitofish presence and/or abundance may be negatively associated with pr esence/abundance of other nati ve small fish species and serve as a bioindicator of adverse effects on fish communities. Behavioral research on interactions among introduced mosquitofish and native small fish species in Australia and Spain concluded mosquitofish deleteriously compete with native species (Rincon et al. 2002, Warburton and Madden 2003). Possible m echanisms included predation on early life stages, increased aggressi on, and reduced feeding rates. Impacts of mosquitofish on fish communities in pulp and paper mill e ffluent receiving systems are unknown. Such research would have to carefully separate species interactions from environmental constraints that may negatively impact other fish living in effluent. Thus this last major bioindicator criterion, applicabil ity, requires further investigat ion before mosquitofish can be confidently accepted or rejected as a bioi ndicator of pulp and paper mill effluents. Other Model Fish Species What are the differences between mosquitofish responses to pulp and paper mill effluents and responses in other fish species? First of all, mo squitofish are unique in their reproductive mode. They are ovoviviparous/l ivebearers, while other species studied (such as fathead minnows, Pimephales promelas and largemouth bass, Micropterus salmoides ) are oviparous/egg-layers. Mosquitofi sh are better suited as models for maternal transfer of contaminants in this respect, since maternal nourishment (specifically facultative matrotrophy or condi tional maternal provisioning of developing embryos) was recently demonstrated for mosquitofish (Marsh-Matthews et al. 2001, Demarais 2003). Thus mosquitofish more accurately reflect placental nourishment of

PAGE 242

224 human fetuses than egg-laying fish speci es and may extrapolate to humans more effectively. Similarly, circulating sex ster oids in mosquitofish may be more similar to humans than other fishes. Testosterone, and not 11-ketotestosterone, is the dominant active androgen in mosquitofish (Borg 1994, Chap ter 2). Although progesterone was not measured in my study, research on ovovivipar ous and viviparous elasmobranches showed progesterone patterns were analogous to hum ans with a rise in progesterone at the periovulatory period that was sustai ned by pregnancy (Koob and Callard 1999). Research has been conducted on egg-laying fish species exposed to effluent from Rice Creek, affording more dir ect comparison of mosquitofish to other fish models. Largemouth bass, exposed for 56 days in the same tank facility used for mosquitofish, responded with reduced circulat ing sex steroids and gonad size in adults at 20 to 40% effluent dilution and decreased fry growth and survival at 10% dilution and greater (Seplveda 2000, Seplveda et al 2001, Seplveda et al. 2003) After EPA Cluster Rule process changes, effects on sex steroids and gonad size did not manifest until 40% effluent dilution or greater (Noggle et al. 2004b). (Fry pr oduction has not been assessed since process changes.) In contrast, my study showed effects on mosquitofish sex steroids at 10% or greater e ffluent dilution after process changes (Chapter 5), although anal fin elongation was not induced (Chapt er 5) and fry produc tion in wild-caught females was not affected (Chapter 6). Similar to largemouth bass, fathead minnow research conducted before process changes s howed reproductive-level effects (decreased egg production) at slightly higher effluent concentrations (23% or greater) (NCASI

PAGE 243

225 2000a). Masculinization in fathead minnows was not observed either before (NCASI 2000a) or after (DL Borton, pers. comm..) processing improvements. Relative to these data, mosquitofish are the most suitable model of the three fish species to examine masculini zation effects. Hormonally, th ey may be more sensitive than bass and minnows. However, bass may be the most sensitive in terms of adverse reproductive effects, followed by the minnows. Mosquitofish, on the other hand, may be responding with a shift in repr oductive strategy as opposed to direct adverse impact. If the mosquitofish responses can be adequately referenced to reproductive impacts in bass and minnows, then mosquitofish may indeed become a suitable bioi ndicator of adverse effect. Future Work Six aspects of mosquitofish responses to pulp and paper mill effluents require further investigation: seasonality; dynami c exposure; population-le vel responses; fish community assessments; mechanism of anal fin elongation; and adaptation to exposure. A major drawback to many toxicology studies is insufficient knowledge of background normal responses (Van Der Kraak et al. 1998). My study implied seasonality of mosquitofish hormones, and warrants a full characterization of hormone profiles at least monthly throughout the year in a reference site and under laboratory conditions. Progesterone should be included in this characterization to determine as well. Related characterization of fecundity (fry production) throughout th e entire reproductive season needs to be conducted as well for multiple re ference sites and effluent-exposed sites. Admittedly, such a project would be very labor intensive but would allow better interpretation of current reproductive success studies.

PAGE 244

226 A scenario of dynamic exposure was implied by my study. Concentrations of effluent components in water samples and precip itation levels in surveyed regions varied widely, but additional environmental factor s such as bacterial communities may be influential and crucial to actual exposures. These aspects need to be addressed then linked back to actual exposures and effects in fish. For example, bacterial communities could be surveyed over several weeks in th e field at receiving streams and before discharge in conjunction with chemical analysis of effluent, water, a nd sediment samples. Based upon these results, laborat ory studies could be designe d to evaluate degradation products of observed effluent components by different bacterial species. Finally, observed degradation products could be examined back in the field in water, sediment, and fish samples to provide insight into act ual exposures and effects. Since bacterial degradation products are a key component to the current hypothesis of mosquitofish masculinization, this type of study should be a high priority. Population-level work in microcosms woul d refine analysis of adverse impacts on mosquitofish. The flow-through tank facility at the Rice Creek mill would facilitate such studies, allowing controlled ex amination of changes in population structure and juvenile recruitment over time. This work could al so be compared to fry production studies conducted on females collected from field sites and provide further insi ght into variations in reproductive strategies in re sponse to effluent exposure. To address potential adverse impacts of mosquitofish on fish communities, community health assessments of receiving streams, such as the index of biologic integrity, could be performed (see Karr and C hu 1999 for an explanation of this metric and Adams et al. 1992 for a description of fish community assessment in a pulp and

PAGE 245

227 paper mill effluent-receiving stream). Coupled with laboratory studies of interactions among fish species from field sites, these data would provide the most insight into applicability and relevance of mosquitofish as a bioindicator of pulp and paper mill effluent exposure. Research on the mechanism of anal fin development may refine the search for bioactive effluent components by allowing iden tification of genes induced by exposure to specific components. Most research on fin growth and development has been conducted on the zebrafish, Danio rerio (Johnson and Bennett 1999). Mu ch of this work involved regeneration studies on adults (w hich is appropriate for invest igation of mosquitofish anal fin elongation), although the genetic control of anal fin regeneration specifically has not been fully characterized. Studies on livebeare rs themselves have identified at least two genes that may be involved in anal fin elonga tion in mosquitofish. Swordtail research showed development of male secondary sex characteristics in anal and caudal fins was linked to upregulation of msxC expression (Zauner et al. 2003). Further, this regulation was different than regulation of fin development in zebrafish. Most recently, research on gonopodial development of maturi ng western mosquitofish ( Gambusia affinis ) showed androgen-dependent fin elongation was associated with sonic hedgehog ( Shh ) expression in the distal ray epithelium (Ogino et al. 2004). Two isoforms of the androgen receptor, AR and AR were identified and predominantly expressed in distal regions of elongating anal fin rays. Ther efore, anal fin elongation appears locally stimulated by exogenous androgens at the fin itself. This la tter research on genetic control of anal fin elongation may prove invaluable in mosquito fish populations where the masculinization response is in question (i.e., Ri ce Creek and Elevenmile Creek).

PAGE 246

228 Finally, potential adaptation and tolerance of mosquitofish in pulp and paper mill effluents remains to be addressed. Anot her small fish species, the killifish or mummichog ( Fundulus heteroclitus ), was evaluated for sensitiv ity to laboratory exposure to potent dioxin-like compounds sp ecifically IUPAC PCB No. 126 and 3-methylcholanthrene (Nacci et al. 1999). Killifish from a Su perfund site heavily contaminated with PCBs were more tolerant to these exposures than fish from nearby reference populations: survival was greater and EROD activity was lower. Similarly, mosquitofish collected from sites with high c oncentrations of pestic ides showed inherited tolerance to organic contaminants (Andr eason 1985). Although survival and EROD activity may not be relevant to adaptation and tolerance of mosquitofi sh to pulp and paper mill effluents, my study implies reproductive strategies are adapted (Chapter 6). For example, Meffe and Snelson (1993) showed energy allocation during reproduction had large interindividual variation under laborator y study of mosquitofish collected from a site that was probably contaminated (on th e US Department of Energy’s Savannah River Site). These results also supported the potential for faculta tive matrotrophy in mosquitofish which was demonstrated more conclusively by direct experimental visualization of maternal transfer (Marsh -Matthews et al. 2001, Demarais and Oldis 2003). Thus, facultative matrotrophy could be one way mosquitofish adapt reproductive strategies to environmental conditions.

PAGE 247

229 APPENDIX A FIELD SITES This appendix contains information about locations for all field sites associated with my study (Tables A-1, A-2, and A-3). (Maps for these sites can be found in Chapters 2 through 6.) In a ddition, monthly rainfall data by region are represented in Figure A-1. These quality-controlled data were obtained from the National Oceanic Atmospheric Administration’s National Climat ic Data Center in Asheville, NC, USA (2005a,b,c). Since all three rece iving systems are low-flow st reams they are theoretically vulnerable to flooding with higher dilution of effluent. Conversely, periods of drought may concentrate effluent. Thus, precipitation adds yet another element of complexity to exposure of wild fish to pulp mill effluents. In Figure A-1A, Rice Creek fish were living under drought conditions in 1999 and up to initial collections in March and April 2000. (Females live 1 to 2 years, while males usually live no more than one year in the w ild (Meffe and Snelson 1989). Total yearly rainfall for 1999 was 40.90 inches versus a normal average of 50.42 inches, and by the end of 2000 the yearly total (49.28 inches) was back up to normal levels. In 2001, precipitation was similar to hi storical levels (52.93 inches total in 2001) with a large flood peak just after July and August co llections in September. Tank and in situ field exposures began during a period of low ra infall in March 2002 and continued under average precipitation in April. Flooding occurred after exposures in June; fish were sampled early June before any of the ma jor storms. Precipitation during 2003 was uncharacteristically cyclic with regular pe riods of flooding, the gr eatest in March.

PAGE 248

230 Population collections were made in May, dur ing a month of average rainfall. June peaked again with flooding, and then females were collected for fry production in July when rainfall was low. In situ field exposures occurred in August, and flooding during this month contributed to premature termina tion of exposures. Finally, the most recent collections for fry production studies were monthly from May to August 2004. Rainfall remained cyclic up to May, when monthly aver age was normal, but the rest of collections occurred in higher than norma l precipitation from several intense storms that released 2.3 to 3.0 inches per storm. Initial Fenholloway River collections we re conducted in August 2001. The region was under drought conditions, with yearly tota l precipitation of 41.37 inches compared to normal average of 58.15 inches. No collectio ns were made in 2002, when rainfall was slightly higher than normal at 62.12 inches. Intense floodi ng characterized the region in 2003, with abnormal highs in March and Octobe r and a higher than normal rainy season June to August. Total yearly precipitation was highest for all year s of collection at 81.73 inches. Population collections were made in May, during a month of average rainfall. June began flood conditions of the rainy s eason, and females were collected for fry production studies in July when rainfall remain ed elevated. Caged exposures occurred in August, and flooding during this month cont ributed to premature termination of exposures. Most recent collections for fry pr oduction studies were monthly from May to August 2004. Precipitation data is complete through September a nd indicates a normal nine month total at 50.30 inches compared to a historical nine month total of 48.59 inches. However, rainfall cycled between lower than average rainfall in May and July versus elevated rainfall in June and August.

PAGE 249

231 The third system, Elevenmile Creek, was surveyed in August 2001. Fish collection coincided with normal precipitation for the ra iny season. Preceding collection, rainfall was lower than normal with the exception of high rainfall in March. Total yearly precipitation indicates drought conditions at 47.53 inches versus a normal of 62.25 inches.

PAGE 250

232 Table A-1. Latitude, longitude, and descriptions for mosquitofish collection sites in Rice Creek. Rice Creeka Latitude Longitude Description REF1b N2936.422’ W08136.309’ Upst ream of Rice Creek at Blackbird Point in Saint Johns River REF2b N2943.020’ W08143.534’ Etonia Cree k (at Bardin Rd), tributary of Rice Creek REF3b N2952.461’ W08222.030’ Santa Fe River at Hwy 121 boat ramp near confluence of New River U(8)c N2941.233’ W08144.510’ Upstream of discharge at S.R. 100 in Rice Creek PRE-DISd N2941.155’ W08142.050’ Retention Pond 4, before discharge into Rice Creek DISe N2940.730’ W08141.648’ Discharg e point at first aerator into Rice Creek D(1)f N2941.324’ W08140.914’ Downstream of discharge at second aerator in Rice Creek D(6)f N2941.971’ W08139.960’ Downstream of discharge at SR 17 bridge in Rice Creek aMap of Elevenmile Creek in Chapter 4 bREF indicates reference site, followed by identifying number cU denotes upstream of discharge, followed by approximate distance (km) from discharge in parentheses dPRE-DIS indicates site before discharge into the creek eDIS denotes site at discharge into creek fD denotes downstream of discharge, follo wed by approximate distance (km) from discharge in parentheses

PAGE 251

233 Table A-2. Latitude, longitude, and descriptio ns for mosquitofish collection sites in Fenholloway River. Fenholloway Rivera Latitude Longitude Description REF1b N3015.079’ W08342.026’ Econfina River at US 19 crossing REF2b N3008.556’ W08351.944’ Econfina River at US 98 boat ramp U(5)c N3006.177’ W08326.369’ Upstream of discharge at US 27 in Fenholloway River PRE-DISd N3003.942’ W08333.275’ Canal leading from retention ponds to Fenholloway River DISe N3004.069’ W08333.326’ Discharge point at old railroad crossing in Fenholloway River D(12)f N3004.540’ W08339.769’ Downstream of discharge at Hampton Springs Bridge in Fenholloway River aMaps of Fenholloway River in Chapters 4 and 6 bREF indicates reference site, followed by identifying number cU denotes upstream of discharge, followed by approximate distance (km) from discharge in parentheses dPRE-DIS indicates site before discharge into the creek eDIS denotes site at discharge into creek fD denotes downstream of discharge, follo wed by approximate distance (km) from discharge in parentheses

PAGE 252

234 Table A-3. Latitude, longitude, and descriptio ns for mosquitofish collection sites in Elevenmile Creek. Elevenmile Creeka Latitude Longitude Description REF1b N3046.553’ W08720.326’ Pine Barren Creek, tributary of Escambia River REF2b N3029.605’ W08719.494’ Eight mile Creek, tributary of Elevenmile Creek U(1)c N3034.986’ W08719.705’ Upstream of discharge at US 297A in a headwater tributary of Elevenmile Creek PRE-DISd N3034.730’ W08719.223’ Retention pond before discharge into Elevenmile Creek DISe N3034.432’ W08719.329’ Discharge point at Kingsfield RD in Elevenmile Creek D(5)f N3032.075’ W08720.587’ Downstream of discharge at Ninemile RD in Elevenmile Creek aMap of Elevenmile Creek in Chapter 4 bREF indicates reference site, followed by identifying number cU denotes upstream of discharge, followed by approximate distance (km) from discharge in parentheses dPRE-DIS indicates site before discharge into the creek eDIS denotes site at discharge into creek fD denotes downstream of discharge, fo llowed by approximate distance (km) from discharge in parentheses

PAGE 253

235 MONTH J A N F EB MAR A P R MA Y J U N JL Y AUG SE P T O CT NO V DEC TOTAL PRECIPITATION (inches) 0 2 4 6 8 10 12 14 16 18 20 NORMAL 1999 2000 2001 2002 2003 2004 A MONTH JA N FEB MAR A PR MAY JUN JLY AU G SEP T O C T NOV DEC TOTAL PRECIPITATION (inches) 0 2 4 6 8 10 12 14 16 18 20 NORMAL 2001 2002 2003 2004 B Figure A-1. Total monthly precipitation for Florida regions where mosquitofish were collected from pulp and paper mill effluent-receiving systems (National Climactic Data Center, Asheville, NC ). A) Palatka region 1999 to 2004, Rice Creek. B) Perry region 2001 to 2004, Fe nholloway River. C) Pensacola region 2001, Elevenmile Creek. “Norma l” indicates historical average precipitation.

PAGE 254

236 MONTH JAN FE B MAR APR M A Y JUN J LY AUG SEPT OCT NOV D E C TOTAL PRECIPITATION (inches) 0 2 4 6 8 10 12 14 16 18 20 NORMAL 2001 C Figure A-1. Continued

PAGE 255

237 APPENDIX B SEX STEROID RADIOIMM UNOASSAY PROTOCOLS This appendix details the working protoc ols used to perform radioimmunoassay of whole body sex steroids in mosquitofish. Valid ation statistics for this assay are provided in Chapter 2. Digestion Add KOH (30%) to each fish at a rate of 3 times the body weight (volume) in individual vials. For example, add 1.2 mL 30% KOH to 0.4 g tissue. (Make sure the lid is on every cryobox before boiling; since caps will pop off of vials otherwise.) Boil samples in cryoboxes for 20 minutes in a shaking water ba th at 80 to 100C. Vortex cryoboxes for 1 minute. Chill cryoboxes for 5 minutes on ice. Remove 50 L from each vial (4 times fo r 2 hormones in duplicate) and aliquot in large glass tubes for ether extraction. 50 L samples in large glass tubes can be stored in a -10C freezer before extraction; any remaining samples shoul d be placed in a -80C freezer for longterm storage. Extraction Thaw 50 L samples. Turn on the vortex evaporator (Labonco); fi ll the ether trap with a thin layer of methanol and dry ice, and then seal the trap. Fill an aliquot bottle with diethyl ether, calibrate for 4 mL, and remove bubbles from aliquoter. Fill a shallow tray with methanol and dry ice. Once samples have thawed, squirt each larg e glass tube with 4 mL diethyl ether

PAGE 256

238 After filling a rack, vortex the rack for 1 minute. Place the rack in the dry ice/methanol tray and allow every sample to precipitate into a white pellet (up to 3 or 4 minutes ) Replenish the dry ice to keep the precipitation temperature low enough. Pour the supernatant into smaller glass tubes. Place small tubes containing supernatant in the evaporator. Turn on the vacuum and then the vort ex and evaporate samples for 10 to 15 minutes. (All the ether does not need to evaporate during th e first extraction.) Add another 4 mL diethyl ether, and th en precipitate and evaporate samples a second time. However, this time make sure all the ether has evaporated. Evaporated samples can be stored in a -10C freezer for up to a week. Radioimmunoassay Remove evaporated samples in sm all glass tubes from -10C freezer. Transfer small tubes to new white racks: 60 tubes to two racks for each hormone, leaving first and last columns for controls; 2 of each sample to a rack set, loaded top to bottom vertically; 11 new tubes at fi rst and last columns; leaving an opening at the second to bottom space. Four r acks can hold up to 100 samples at a time. Prepare the standards (8 exponential d ilutions from 1-1,000 pg per hormone). Vortex each standard upon first use. Pipette 50 L of each standard to corresponding new tubes at first and last columns of each rack set (3rd-10th tubes from top). Increase accuracy by using a new pipette tip for each sample. Load the weakest to strongest dilution from top to bottom. Save the standards until radioactivity has been added, then dispose them down the sink. Add 200 L PBSGA buffer to each dilution standard. Add 100 L PBSGA buffer to the total count tubes (TC) and nonspecifi c binding tubes (NSB) (the 4 corners of rack set). Add 250 L PBSGA buffer to NSB (to bring total volume of PBSGA to 530 L), maximum binding tubes (B0), and TC (first 2 and last tubes along first and last columns). Finally, add 250 L PB SGA buffer to all tissue samples. Prepare the antibody for each hormone (Ab) Vortex Ab upon first use. Add 100 L Ab to all tubes but TC and NSB (4 corners). Prepare radioactive-labeled hormones (mar ked with tritium). Carefully stir upon first use. Carefully add 100 L radioactive hormone to all tubes. Incubate samples in a refrigerator for 24 to 48 hours.

PAGE 257

239 Prepare charcoal dextran and vortex fo r several minutes upon first use. Add 250 L water to TC (at bottom 2 corners) and separate TC tubes from the rack set. Add 250 L charcoal dextran to all other tubes within 5 to 7 minutes. Shake the rack set a few times, and then load and balance sample tubes in the centrifuge (Beckman J-6). Set the centri fuge at 3 rpm and spin for 10 minutes. Mix 4 mL scintillation cocktail with 400 L of each sample and place mixture in a scintillation vial. Load vials into scintil lation racks, and then load racks into the liquid scintillation counter (Packard Tricarb 1600). Dispose of waste appropriately.

PAGE 258

240 APPENDIX C POSTER AND PLATFORM PRESENTATI ONS OF DISSERTATION RESEARCH Noggle JJ, Bradley WK, Smith JT, Gross TS Platform presentation, “Relationship between changing effluent quality and repr oductive effects in Ea stern gambusia.” 24th annual meeting of the Society of Environmental Toxicology & Chemistry, Austin, TX, November 9-13, 2003. Noggle JJ, Smith JT, Ruessler DS, Quinn BP, Holm SE, Sepulveda MS, Gross TS. Platform presentation, “Paper mill process modifications reduce biological effects on largemouth bass and Eastern gambusia.” 5th International Conference on Fate and Effects of Pulp and Paper Mill E ffluents, Seattle, WA, June 1-4, 2003. Noggle JJ, Bradley WK, Borton DL, Smith JT, Gross TS. Platform presentation, “Comparison of anal fin morphology & horm one status in gambusia among Florida pulp & paper mills.” 5th International Conference on Fa te and Effects of Pulp and Paper Mill Effluents, Seattle, WA, June 1-4, 2003. Noggle JJ, Bradley WK, Borton DL, Smith JT, Gross TS. Poster presentation, “Comparison of anal fin morphology in female gambusia among three Florida pulp and paper mills.” 23rd annual meeting of the Society of Environmental Toxicology & Chemistry, Salt Lake City, UT, November 16-20, 2002. Noggle JJ. Invited platform presentation, “Gambusia and pulp & paper mill effluents.” Georgia-Pacific Corp. 2002 Enviro nmental Conference, Atlanta, GA September 16 & 17, 2002. Noggle JJ, Ruessler DS, Sepul veda MS, Holm SE, Gross TS. Poster presentation, “Responses of Eastern mosquitofish to papermill effluent exposure.” 22nd annual meeting of the Society of Environmental Toxicology & Chemistry, Baltimore, MD, November 11-15, 2001. Noggle JJ, Bradley WK, Ruessler DS, Sepulve da MS, Gross TS. Po ster presentation, “Considerations in mosquitofish and papermill effluent studies: analysis of methodology.” 22nd annual meeting of the Society of Environmental Toxicology & Chemistry, Baltimore, MD, November 11-15, 2001. Noggle J, Ruessler D, Sepulveda M, Holm S, Gross T. Poster presentation, “Effects of papermill effluent on secondary sex characteristics in mosquitofish.” 40th annual meeting of the Society of Toxico logy, San Francisco, CA, March 25-29, 2001.

PAGE 259

241 Noggle JJ, Gross TS. Platform presenta tion, “Effects of paper mill effluents on secondary sex characteristics in mosquitofish.” 4th annual meeting of the Southern Conference of Researchers in Aquatic Di seases, Gainesville, FL, February 10-12, 2001. 2nd place Best Student Presentation.

PAGE 260

242 LIST OF REFERENCES Adams SM, Crumby WD, Greeley MS, Shugart LR, Saylor CF. 1992. Responses of fish populations and communities to pulp mill effluents: a holistic assessment. Ecotox Environ Safety 24:347-360. Ahmad I, Hamid T, Fatima M, Chand HS, Jain SK, Athar M, Raisuddin S. 2000. Induction of hepatic antioxidant s in freshwater catfish ( Channa punctatus Bloch) is a biomarker of paper mill effluent expos ure. Biochim Biophys Acta 1523:37-48. Ahokas JT, Holdway DA, Brennan SE, Goude y RW, Bibrowska HB. 1994. MFO activity in carp ( Cyprinus carpio ) exposed to treated pulp and paper mill effluent in Lake Coleman, Victoria, Australia, in relati on to AOX, EOX, and muscle PCDD/PCDF. Environ Toxicol Chem 13:41-50. Anderson VL, McLean RA. 1974. Volume 5, Design of experiments: a realistic approach. In: Owen DB, Lewis P, Mint on PD, Pratt JW, editors. Statistics: textbooks and monographs. New York: Marcel Dekker. 418 p. Andersson T, Frlin L, Hrdig J, La rsson . 1988. Biochemical and physiological disturbances in fish inhabiting coastal waters polluted with bleached kraft mill effluents. Mar Environ Res 24:233-236. Andreason JK. 1985. Insecticide resistance in mosquitofish of the lower Rio Grande Valley of Texas – an ecological hazard ? Arch Environ Contam Toxicol 14:573-57. Angus RA, Blanchard P, Howell WM Douglas WR. 1997. A short-term in vivo screening system for endocrine di srupters using mosquitofish ( Gambusia affinis and G. holbrooki ). US Environmental Protecti on Agency, National Center for Environmental Research, Grant R826130. http://cfpub.epa.gov/ncer_abstracts/index.c fm/fuseaction/display.abstractDetail/abs tract/167/report/0 Accessed 2005 February 5. Angus RA, McNatt HB, Howell WM, Peopl es SD. 2001. Gonopodium development in normal male and 11-ketotestosterone -treated female mosquitofish ( Gambusia afinis ): a quantitative study us ing computer image analys is. Gen Comp Endocrinol 123:222-234. Angus RA, Weaver SA, Gri zzle JM, Watson DR. 2002. Reproduc tive characteristics of male mosquitofish ( Gambusia affinis ) inhabiting a small Southeastern US river receiving treated domestic sewage e ffluent. Environ Toxicol Chem 21:1404-1409.

PAGE 261

243 Baroiller JF, Guiguen Y, Fostier A. 1999. Endocrine and environmental aspects of sex differentiation in fish. Cell Mol Life Sci 55:910-31. Batty J, Lim R. 1999. Morphological and re productive characte ristics of male mosquitofish ( Gambusia affinis holbrooki ) inhabiting sewage-contaminated waters in New South Wales, Australia. Arch Environ Contam Toxicol 36:301-307. Borg B. 1994. Androgens in teleost fishes. Comp Biochem Physiol C 109:219-45. Bortone SA, Cody RP. 1999. Morphological masc ulinization in poecili id females from a paper-mill effluent receiving tributary of the St. Johns River, Florida. Bull Environ Contam Toxicol 63:150-156. Bortone SA, Davis WP, Bundrick CM. 1989. Mor phological and behavioral characters in mosquitofish as potential bioindication of exposure to kraft mill effluent. Bull Environ Contam Toxicol 43:370-377. Bortone SA, Davis WP.1994. Fish intersexuality as indicator of environmental stress. BioScience 44:165-172. Bortone SA, Drysdale 1981. Additional evidence for environmentally induced intersexuality in poeciliid fishes. Assoc Southeastern Biol Bull 28:67. Borton DL, Hall TJ, Fisher RP, Thomas JE, editors. 2004. Pulp and paper mill effluent environmental fate and effects. 5th International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluent s; 2003 June 1-4; Seattle. Lancaster: DEStech Publications. 578p. Bradley WK, Borton DL, Noggle JJ, Gro ss TS. 2004. Sources of variability of mosquitofish ( Gambusia holbrooki ) anal fin morphology characteristics: measurements methods, geographic variabil ity, and exposure to pulp mill effluents. In: Borton DL, Hall TJ, Fisher RP, Thomas JE, editors. Pulp and paper mill effluent environmental fate and effects. 5th International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluent s; 2003 June 1-4; Seattle. Lancaster: DEStech Publications. p 25-38. Britton RH, Moser ME. 1982. Size specific preda tion by herons and its effect on the sexratio of natural populations of the mosquito fish Gambusia affinis (Baird and Girard). Oecologia 53:146-151. Bucher F, Hofer R, Salvenmoser W. 1992. E ffects of treated paper mill effluents on hepatic morphology in male bullhead ( Cottus gobio L.). Arch Environ Contam Toxicol 23:410-419. Cabral JA, Avila S, Marques JC. Acute and sublethal effects of a non-ionic surfactant, Genapol OXD-080 on mosquitofish Gambusia holbrooki (Girard). Ecotoxicology 8:245-252.

PAGE 262

244 Caruso JH, Suttkus RD. 1988. Abnormal expres sion of secondary sex characters in a population of Anguilla rostrata (Pisces: Anguillidae) from a dark colored Florida stream. Copeia 4:1077-1079. Casterlin ME, Reynolds WW. 1977. Aspects of habitat selection in the mosquitofish Gambusia affinis Hydrobiologia 55:125-127. Cody RP, Bortone SA. 1997. Masculinization of mo squitofish as an indicator of exposure to kraft mill effluent. Bull Environ Contam Toxicol 58:429-436. Constanz GD. 1989. Reproductive biology of Poeci liid fishes. In: Meffe GK, Snelson FF, editors. Ecology and evolution of liveb earing fishes (Poeciliidae). Englewood Cliffs: Prentice-Hall. p 33-50. Davis WP, Bortone SA. 1992. Effects of kraft m ill effluent on the sexuality of fishes: an environmental early warning? In: Colborn TC, Clement C, editors. Volume 21, Chemically-induced alterations in se xual and functional development: the wildlife/human connection. In: Mehlman MA, editor. Advances in modern environmental toxicology. Princeton: Pr inceton Scientific Publishing. p 113-127. DeMarais AA, Oldis DL. 2003. Visualization of mother-to-embryo transfer in livebearing fishes. Integr Comp Biol 43:1003. Denton TE, Howell WM, Allison JJ, McCollum J, Marks B. 1985. Masculinization of female mosquitofish by expos ure to plant sterols and Mycobacterium smegmatis. Bull Environ Contam Toxicol 35:627-632. Dick T, Heng Lee B, Murugasu-Oei B. 1998. Oxygen depletion induced dormancy in Mycobacterium smegmatis FEMS Microbiol Lett 163:159-164. Downhower JF, Brown LP, Matsui ML. 2000. Life history variation in female Gambusia hubbsi Environ Biol Fish 59:415-428. Downhower JF, Brown LP, Matsui ML. 2002. Litter overlap in Gambusia hubbsi : superfetation revisited. E nviron Biol Fish 65:423-430. Dreze V, Monod G, Cravedi JP, Biagianti-Risbourg S, Le Gac F. 2000. Effects of 4nonylphenol on sex differentiation an d puberty in mosquitofish ( Gambusia holbrooki ). Ecotoxicology 9:93-103. Drysdale DT, Bortone SA. 1989. Laboratory induction of inters exuality in the mosquitofish, Gambusia affinis using paper mill effluent. Bull Environ Contam Toxicol 43:611-617. D’Surney, SJ, Eddy LP, Felder DP, Rodgers JH, Deardroff TL. 2000. Assessment of the impact of a bleached kraft mill effluent on a south-central USA river. Environ Toxicol 15:28-39.

PAGE 263

245 Dube MG, MacLatchy DL. 2001. Identification and treatment of a waste stream at a bleached-kraft mill that depresses a sex steroid in the mummichog ( Fundulus heteroclitus ). Environ Toxicol Chem 20:985-995. Durhan E, Lambright C, Wilson V, Buterworth BC, Kuehl DW, Orlando EF, Guillette LJ, Gray LE,Ankley GT. 2002. Evaluation of androstenedione as an andreogenic component of river water downstream of a pulp and pa per mill effluent. Environ Toxicol Chem 21:1973-1976. Ellis RJ, Van Den Heuvel MR, Bandelj E, Smith MA, McCarthy LH, Stuthridge TR, Dietrich DR. 2003. In vivo and in vitro assessment of the andr ogenic potential of a pulp and paper mill effluent. Environ Toxicol Chem 22:1448-1456. Fatima M, Ahmad I, Sayeed I, Athar M, Raisuddin S. 2000. Pollutant-induced overactivation of phagocytes is concomitantly a ssociated with peroxidative damage in fish tissues. Aquat Toxicol 49:243-250. Felder DP, D’Surney, SJ, Rodgers JH Deardroff TL. 1998. A comprehensive environmental assessment of receiving aquatic system near an unbleached kraft mill. Ecotoxicology 7:313-324. Fernndez-Delgado C, Rossomano S. 1997. Repr oductive biology of the mosquitofish in a permanent natural lagoon in south-west Sp ain: two tactics for one species. J Fish Biol 51:80-92. Frlin L, Andersson T, Bengtsson BE, Hrdig J, Larsson . 1985. Effects of pulp bleached plant effluents on hepatic xenobiot ic biotransformation enzymes in fish: laboratory and field studies. Mar Environ Res 17:109-112. Foster ML, Versteeg DJ, McKee MJ, Fo lmar LC, Graney RL, McCume DC. 1992. Physiological and nonspecific biomarkers In: Huggett RJ, Kimerie RA, Mehrle PM, Bergman HL, editors. Biomarkers: bi ochemical, physiological and histological markers of anthropogenic stress. Boca Raton: Lewis Publishers. pp 5-86. Gagn F, Blaise C. 1993. Hepatic metallot hionen level and mixed function oxidase activity in fingerling trout ( Oncorhynchus mykiss ) after acute exposure to pulp and paper mill effluents. Wat Res 27:1669-1682. Gagnon MM, Bussieres D, Dodson JJ, Hodson PV. 1995. White sucker ( Catostomus commersoni ) growth and sexual maturation in pulp mill-contaminated and reference rivers. Environ Toxicol Chem 14:317-327. Gagnon MM, Dodson JJ, Hodson PV. 1994a. Ab ility of BKME (bleached kraft mill effluent) exposed white suckers ( Catostomus commersoni ) to synthesize steroid hormones. Comp Biochem Physiol C 107:265-273.

PAGE 264

246 Gagnon MM, Dodson JJ, Hodson PV, Van Der Kraak G, Carey JH. 1994b. Seasonal effects of bleached kraft mill effluent on reproductive parameters of white sucker ( Catostomus commersoni ) populations of the St. Maurice River, Quebec, Canada. Can J Aquat Sci 51:337-347. Goodbred SL, Gilliom RJ, Gross TS, Denslow NP, Bryant WL, Schoeb TR. US Geological Survey. 1997. Reconnaissance of 17-estradiol, 11-ketotestosterone, vitellogenin, and gonadal histopathology in common carp of United States streams: potential for contaminant-induced en docrine disruption. Sacramento: US Geological Survey. Open-File Report 96-627. 47 p. Gross TS, Kernaghan NJ, Ruessler DS, Holm SE. 2000. An evaluation of the potential effects of papermill effluents on freshwater mussels. In: Ruoppa M, Paasivrita J, Lehtinen KJ, Ruonala S, editors. Proceedings of the 4th International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluents; 2000 June 12-15; Helsinki. Helsinki: The Finnish Environment. Report 417. p 253-257. Gross TS, Seplveda MS, Wieser CM, Wieb e JJ, Schoeb TR, Denslow ND, Johnson WE. 2001. Characterization of annual reproducti ve cycles for pond-reared Florida largemouth bass ( Micropterus salmoides floridanus ). In: Proceedings of the Black Bass 2000 Symposium; 2000 November; St. Lo uis. American Fisheries Society Symposium 31. p 205-212. Haynes JL, Cashner RC. 1995. Life history and population dynamics of the western mosquitofish: a comparison of natural a nd introduced populati ons. J Fish Biol 46:1026-1041. Hewitt M, Parrott J, Wells K, Elliott K, Bi ddiscombe S, McMaster M, Munkittrick K, Van Der Kraak G. 2000. Accumulation of ligands for the Ah receptor, androgen receptor, estrogen receptor and sex steroid binding protein by white sucker exposed to bleached kraft mill efflue nt. In: Ruoppa M, Paasivrita J, Lehtinen KJ, Ruonala S, editors. Proceedings of the 4th International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluent s; 2000 June 12-15; Helsinki. Helsinki: Finnish Environment Institute. Report 417. p 159-164. Hewitt LM, Smyth SA, Dube MG, Gilman CI, MacLatchy DL. 2002. Isolation of compounds from bleached kraft mill chemi cal recovery condensates associated with reduced levels of te stosterone in mummichog ( Fundulus heteroclitus ). Envrion Toxicol Chem 21:1359-1367. Hildemann WH. 1954. Effects of sex hormones on the secondary sex characteristics of Lebistes reticulates J Exp Zool 126:1-15. Hile R. 1936. Age and growth of the cisco, Leucichthys artedi (Le Sueur), in the lakes of the northeastern highlands, Wisconsi n. Bull US Bureau Fish 19: 211-317.

PAGE 265

247 Howell WM, Black DA, Bortone SA. 1980. A bnormal expression of secondary sex characters in a populat ion of mosquitofish, Gambusia affinis holbrooki : evidence for environmentally induced masculinization. Copeia 4:676-681. Howell WM, Denton TE. 1989. Gonopodial mor phogenesis in female mosquitofish, Gambusia affinis holbrooki masculinized by exposure to degradation products from plant sterols. Environ Biol Fishes 24:43-51. Howell WM, Hunsinger RN, Blanchard PD. 1994. Paradoxical masculinization of female Western mosquitofish during exposure to spironolactone. Pr ogress Fish Cultur 56:51-55. Hughes AL. 1985. Seasonal changes in fecundity and size at first reproduction in an Indiana population of the mosquitofish Gambusia affinis American Midland Naturalist 114:30-36. Hunsinger RN, Byram BR, Howell WM. 1988. Unchanged gonadal morphology of mosquitofish masculinized by exposure to degraded plant ster ols. J Fish Biol 32:795-796. Jalabert B, Baroiller JF, Breton B, Fos tier A, Le Gac F, Guiguen Y, Monod G. 2000. Main neuro-endocrine, endocrine, and para crine regulations of fish reproduction, and vulnerability to xenobi otics. Ecotoxicology 9:25-40. Jamil K. 2001. Bioindicators and biomarke rs of Environmental pollution and risk assessment. Enfield: Science Publishers. 204 p. Jardine JJ, Van Der Kraak GJ, Munkittrick KR. 1996. Capture and confinement stress in white sucker exposed to bleached kraft mill effluent. Ecotoxicol Environ Safety 33:287-298. Jenkins R, Angus RA, McNatt H, Howell WM Kemppainen JA, Kirk M, Wilson EM. 2001. Identification of androstenedione in a river containing paper mill effluent. Environ Toxicol Chem 20:1325-1331. Jenkins RL, Wilson EM, Angus RA, Howell WM, Kirk M. 2003. Androstenedione and progesterone in the sediment of a river receiving paper mill effluent. Toxicol Sci 73:53-59. Johnson SL, Bennett P. 1999. Growth contro l in the ontogenetic and regenerating zebrafish fin. In: Detrich HW, Westerfi eld M, Zon LI, editors. Volume 59. The zebrafish: biology. In: American Society fo r Cell Biology. Methods in cell biology. San Diego: Academic Press. p301-311. Karr JR, Chu EW. 1999. Restoring life in runni ng waters: better biological monitoring. Washington, DC: Island Press. 206 p.

PAGE 266

248 Kiparississ Hughes R, Metcalfe C. 2001. Iden tification of the isoflavonoid genistein in bleached kraft mill effluent. Environ Sci Technol 35:2423-2427. Koob TJ, Callard IP. 1999. Reproductive endo crinology of female elasmobranches: lessons from the little skate ( Raja erinacea ) and spiny dogfish ( Squalus acanthias ). J Exp Zool 284:557-574. Kovacs TG, Gibbons JS, Martel PH, O’Connor BI, Voss RH. 1995a. Effects of a secondary-treated thermomechanical pulp mill effluent on aquatic organisms as assessed by shortand long-term laboratory tests. J Toxicol Environ Health 44:485502. Kovacs TG, Gibbons JS, Tremblay LA, O’Connor BI, Martel PH, Voss RH. 1995b. Effects of a secondary-treated bleached kr aft mill effluent on aquatic organisms as assessed by shortand long-term laborato ry tests. Ecotox E nviron Safety 31:7-22. Koya Y, Fujita A, Niki F, Ishihara E. 2003. Sex differentiation and pubertal development of gonads in the viviparous mosquitofish, Gambusia affinis Zool Sci 20:12311242. Koya Y, Iwase A. 2004. Annual reproductive cycl e and rate of spermatogenic process in male mosquitofish, Gambusia affinis Ichthyol Res 51:131-136. Koya Y, Kamiya E. 2000. Environmental regul ation of annual reproductive cycle in the mosquitofish, Gambusia affinis J Exp Zool 286:204-211. Krotzer MJ. 1990. The effects of induced masc ulinization on reproductiveand aggressive behaviors of the female mosquitofish, Gambusia affinis affinis Environ Biol Fishes 29:127-134. Krumholz LA. 1948. Reproduction in the western mosquitofish, Gambusia affinis affinis (Baird and Girard) and its use in mo squito control. Ecol Monogr 18:1-43. Kwon JY, Haghpanah V, Kogson-Hurtado LM, McAndrew BJ, Penman DJ. 2000. Masculinization of genetic female Nile tilapia ( Oreochromis niloticus ) by dietary administration of an aromatase inhibito r during sexual differentiation. J Exp Zool 287:46-53. Larsson DGJ, Frlin L. 2002. Male-biased se x ratios of fish embryos near a pulp mill:temporary recovery after a short-term shutdown. Environ Health Perspect 110:739-742. Larsson DGJ, Hallman H, Frlin L. 2000. More male fish embryos near a pulp mill. Environ Toxicol Chem 19:2911-2917. Larsson DGJ, Kinnberg K, Sturve J, Stephensen E, Skon M, Frlin L. 2002. Studies of masculinization, detoxification, and oxi dative stress responses in guppies ( Poecilia reticulata ) exposed to effluent from a pulp mill. Ecotox Environ Safety 52:13-20.

PAGE 267

249 Lee HB, Peart TE. 1999. Supercri tical carbon dioxide extraction of resi n and fatty acids from sediments at pulp mill sites. J Chromatogr 594:309-315. Lehtinen KJ. 2004. Relationship of the technical development of pulping and bleaching to effluent quality and aquatic toxicology. In: Bort on DL, Hall TJ, Fisher RP, Thomas JE, editors. Pulp and paper mill effluent environmental fate and effects. 5th International Conference on Environmental Fa te and Effects of Pulp and Paper Mill Effluents; 2003 June 1-4; Seattle. Lan caster: DEStech Publications. p 273-284. Lehtinen KJ, Mattsson K, Tana J, Engstrom C, Lerche O, Hemming J. 1999. Effects of wood-related sterols on the reproduction, egg survival, and offspring of brown trout ( Salmo trutta lacustris L.). Ecotox Environ Safety 42:40-49. Lindstrm-Sepp P, Oikari A. 1988. Hepatic biot ransformation in fishes exposed to pulp mill effluents. Wat Sci Technol 20:167-170. Lindstrm-Sepp P, Oikari A. 1989. Biotrans formation and other phys iological responses in whitefish caged in a lake receiving pulp and paper mill effluents. Ecotox Environ Safety 18:191-203. Lister AL, Van Der Kraak GJ. 2001. Endocrine disruption: why is it so complicated? Wat Qual Res J Can 36:175-190. MacLatchy D, Van Der Kraak GJ. 1995. The phytoestrogen -sitosterol alters the reproductive endocrine status of goldf ish. Toxicol Appl Pharmacol 134:305-312. Marsh-Matthews E, Skierrkowski P, De Marais A. 2001. Direct evidence for mother-to-embryo transfer of nutri ents in the livebearing fish Gambusia geiseri Copeia 1:1-6. Marsheck WJ, Kraychy S, Muir RD. 1972. Mi crobial degradation of sterols. Appl Microbiol 23:72-77. Martel P, Kovacs T. 1997. A comparison of the potential of primary and secondary treated pulp and paper mill effluents to induce mixed function oxidase (MFO) activity in fish. Wat Res 31:1482-1488. Martel PH, Kovacs TG, Voss RH. 1996. Efflue nts from Canadian pulp and paper mills: a recent investigation of their potential to induce mixed function oxygenase activity in fish. In: Servos MR, Munkittrick KR, Carey JH, Van Der Kraak GJ, editors. Environmental Fate and Effects of Pulp and Paper Mill Effluents. 2nd International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluents; 1994 November 6-10; Vancouver. DelRay Beach: St. Lucie Press. pp 401-412. Mather-Mihaich E, DiGuilio RT. 1991. Oxidant, mixed-function oxidase and peroxisomal responses in channel catfish e xposed to a bleached kraft mill effluent. Arch Environ Contam Toxicol 20:391-397.

PAGE 268

250 Mattsson K, Tana J, Engstrom C, Hemming J, Lehtinen KJ. 2001. Effects of woodrelated sterols on the offspri ng of the viviparous blenny, Zoarces viviparus L. Ecotox Environ Safety 49:122-130. McCarthy LH, Bostan IV, Choi W. 2004. Co mparison of some studies assessing the androgenic potential of compounds in pulp mill and municipal effluents using the mosquitofish Gambusia affinis In: Borton DL, Hall TJ, Fisher RP, Thomas JE, editors. Pulp and paper mill effluent environmental fate and effects. 5th International Conference on Environmental Fa te and Effects of Pulp and Paper Mill Effluents; 2003 June 1-4; Seattle. Lan caster: DEStech Publications. p 361-373. McMaster ME, Munkittrick KR, Van Der Kraak GJ, Flett PA, Servos MR. 1996. Detection of steroid hormone disruptions as sociated with pulp mill effluent using artificial exposures of goldfish. In: Se rvos MR, Munkittrick KR, Carey JH, Van Der Kraak GJ, editors. Environmental Fate and Effects of Pulp and Paper Mill Effluents. 2nd International Conferen ce on Environmental Fate and Effects of Pulp and Paper Mill Effluents; 1994 November 6-10; Vancouver. DelRay Beach: St. Lucie Press. pp 425-437. McMaster ME, Parrott JL, Hewitt LM. 2003. A decade of research on the environmental impacts of pulp and paper mill effluent s in Canada (1992-2002). National Water Research Institute, Burlington, Ontario. NW RI Scientific Assessment Report Series no. 4. 84 p. McMaster ME, Portt CB, Munkittrick KR Dixon DG. 1992. Milt characteristics, reproductive performance, and larval surv ival and development of white sucker exposed to bleached kraft mill effluent. Ecotox Environ Safety 23:103-117. McMaster ME, Van Der Kraak GJ, Munkittrick KR. 1995. Exposure to bleached kraft pulp mill effluent reduces the steroid bios ynthetic capacity of white sucker ovarian follicles. Comp Biochem Physiol C 112:169-178. McMaster ME, Van Der Kraak GJ, Portt CB, Munkittrick KR, Sibley PK, Smith IR, Dixon DG. 1991. Changes in hepatic mi xed funxtion oxygenase (MFO) activity, plasma steroid levels and age at maturity of a white sucker ( Catostomus commersoni ) population exposed to bleached kraft pulp mill effluent. Aquat Toxicol 21:199-218. Meffe GK, Snelson FF, editors. 1989. Ecol ogy and evolution of livebearing fishes (Poeciliidae). Englewood Cliffs: Prentice-Hall. 453 p. Meffe GK, Snelson FF. 1993. Lipid dynamics during reproduction in two livebearing fishes, Gambusia holbrooki and Poecilia latipinna Can J Fish Aquat Sci 50:2185-2191. Munkittrick KR, McMaster ME, McCarthy LH, Servos MR, Van Der Kraak GJ. 1998. An overview of recent studies on the poten tial of pulp-mill effluents to alter reproductive parameters in fish. J Toxi col Environ Health B Part B 1:347-371.

PAGE 269

251 Munkittrick KR, Portt CB, Van Der Kraak GJ Smith IR, Rokosh DA. 1991. Impact of bleached kraft mill effluent on population ch aracteristics, liver MFO activity, and serum steroid levels of a Lake Superior white sucker ( Catostomus commersoni ) population. Can J Aquat Sci 48:1371-1380. Munkittrick KR, Servos MR, Carey JH, Va n Der Kraak GJ. 1997. Environmental impacts of pulp and paper wastewater: evidence fo r a reduction in environmental effects since 1992. Wat Sci Technol 35:329-338. Munkittrick KR, Van Der Kraak GJ, McMast er ME, Portt CB. 1992. Response of hepatic MFO activity and plasma sex steroids to s econdary treatment of bleached kraft pulp mill effluent and mill shutdown. Environ Toxicol Chem 11:1427-1439. Munkittrick KR, Van Der Kraak GJ, McMast er ME, Portt CB, van den Huevel MR, Servos MR. 1994. Survey of receiving-wat er environmental impacts associated with discharges from pulp-mills: 2) gona d size, liver size, hepatic EROD activity, and plasma sex steroid-levels in wh ite sucker. Environ Toxicol Chem 13:10891101. Munkittrick KR, Servos MR, Gorman K, Bl unt B, McMaster ME. 1999. Characteristics of EROD induction associated with exposur e to pulp and paper mill effluent. In: Rao SS, editor. Impact Assessment of H azardous Aquatic Contaminants: Concepts and Approaches. Lewis Publishers. p 79-77. Nacci D, Coiro L, Champlin D, Jayaraman S, McKinney R, Gleason TR, Munns WR, Specker JL, Cooper KR. 1999. Adaptations of wild populations of the estuarine fish Fundulus heteroclitus to persistent environmenta l contaminants. Mar Biol 134:917. Nakari T, Erkomaa K. 2003. Effects of phytosterols on zebrafish reproduction in multigeneration test. Environ Poll 123:267-273. National Climactic Data Center. 2005a. Annual Climatological Summary; 1999-2004; Palatka, Florida. Asheville: US Depa rtment of Commerce/National Oceanic and Atmospheric Administration. Station ID 086753. National Climactic Data Center. 2005b. Annual Climatological Summary; 2001; Pensacola Regional Airport, Florida. Asheville: US Department of Commerce/National Oceanic and Atmos pheric Administrati on. Station ID 086997. National Climactic Data Center. 2005c. Annual Climatological Summary; 2001-2004; Perry, Florida. Asheville: US Depart ment of Commerce/National Oceanic and Atmospheric Administration. Station ID 087025. National Council for Air and Stream Improve ment. 1981. Experience with the analysis of pulp mill effluents for chlorinate d phenols using acetic acid anhydride derivitization proce dure. Research Triangle Par k. Technical Bulletin no. 347.

PAGE 270

252 National Council for Air and Stream Improveme nt. 1986. Procedures for the analysis of resin and fatty acids in pulp mill effluents. Research Triangle Park. Technical Bulletin no. 501. National Council for Air and Stream Improveme nt. 1996. Effects of biologically treated bleached kraft mill effluent on the early life stage and life cycle of Pimephales promelas (fathead minnow) and Ceriodaphnia dubia ; a comparison before and after conversion to oxygen delignificati on and ECF bleaching. Research Triangle Park. Technical Bulletin no. 722. National Council for Air and Stream Improvement. 1997. Development and evaluation of a method for the determination of phytoste rols in pulp and paper mill effluents. Research Triangle Park. Technical Bulletin no. 746. National Council for Air and Stream Improve ment, Inc. 1999. Effects of the phytosterol stigmastanol on early life-stages a nd life-cycles of fathead minnows ( Pimephales promelas ) and Ceriodaphnia dubia Research Triangle Par k. Technical Bulletin no. 788. National Council for Air and Stream Improveme nt. 2000a. Effects of biologically treated chlorine-bleached/unbleached kraft pulp mill effluent on early life stages and life cycles of the fathead minnow ( Pimephales promelas ) and Ceriodaphnia dubia before process modifications to meet cl uster rules. Research Triangle Park. Technical Bulletin no. 813. National Council for Air and Stream Impr ovement. 2000b. Effects of Western hemlock wood leachate on the early life stage a nd life cycle of the fathead minnow ( Pimephales promelas ). Research Triangle Park. Technical Bulletin no. 816. Nimrod AC, Benson WH. 1996. Environmenta l estrogenic effects of alkylphenol ethoxylates. Crit Rev Toxicol 26:335-364. Noggle JJ, Bradley WK, Borton DL, Smith JT Gross TS. 2004a. Comparison of anal fin morphology and testosterone le vels in female gambusia among three Florida pulp and paper mills. In: Borton DL, Hall TJ, Fisher RP, Thomas JE, editors. Pulp and paper mill effluent environmental fate and effects. 5th International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluents; 2003 June 1-4; Seattle. Lancaster: DESt ech Publications. p 344-353. Noggle JJ, Smith JT, Ruessler DS, Quinn BP, Holm SE, Sepulveda MS, Gross TS. 2004b. Paper mill process modifications re duce biological effects on largemouth bass and Eastern gambusia. In: Borton DL, Hall TJ, Fisher RP, Thomas JE, editors. Pulp and paper mill effluent environmental fate and effects. 5th International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluents; 2003 June 1-4; Seattle. Lancaste r: DEStech Publications. p 344-353. Nordlie FG. 2000. Patterns of reproduction and de velopment of selected resident teleosts of Florida salt marshes. Hydrobiologia 434:165-182.

PAGE 271

253 Ogino Y, Katoh H, Yamada G. 2004. Androge n dependent development of a modified anal fin, gonopodium, as a model to unders tand the mechanism of secondary sexual character expression in vert ebrates. FEBS Lett 575:119-126. Oikari A, Lonn BE, Castren M, Nakari T. 1983. Toxicological effect s of dehydroabietic acid (DHAA) on the trout, Salmo gairdneri Richardson, in fresh water. Water Res 17:81-89. Oikari A, Lindstrom-Seppa P, Kukkonen J. 1988. Subchronic metabolic effects and toxicity of a simulated pulp mill effluent on juvenile lake trout, Salmo trutta m. lacustris Ecotox Envrion Safety 16:202-218. Orlando EF, Davis WP, Guillette LJ. 2002. Aromat ase activity in the ovary and brain of the eastern mosquitofish ( Gambusia holbrooki ) exposed to paper mill effluent. Environ Health Perspect 110:429-433. Pandian TG, Sheela SG. 1995. Hormonal induction of sex reversal in fish. Aquaculture 138:1-22. Park D, Minor MD, Propper CR. 2004. Toxic response of endosulfan to breeding and nonbreeding female mosquitofis h. J Environ Biol 25:119-124. Parks LG, Lambright CS, Orlando EF, Guille tte LJ, Ankley GT, Gray LE. 2001. Masculinization of female mosquitofish in kraft mill effluent-contaminated Fenholloway River water is associated w ith androgen receptor agonist activity. Toxicol Sci 62:257-267. Parrott JL, Wood CS. 2002. Fathead minnow lif ecycle tests for detection of endocrinedisrupting substances in efflue nts. Wat Qual Res J Can 37:651-667. Parrott JL, Wood CS, Boutot P, Dunn S. 2003. Changes in growth and secondary sex characteristics of fathead minnows exposed to bleached sulfite mill effluent. Ecotox Environ Chem 22:2908-2915. Parrott JL, Wood CS, Boutot P, Dunn S. 2004. Changes in growth, secondary sex characteristics, and reproduction of fath ead minnows exposed for a life cycle to bleached sulfite mill effluent. J T oxicol Environ Health A 67:1755-1764. Peakall D. 1992. Ecotoxicology Series 1, Animal biomarkers as pollution indicators. New York: Chapman and Hall. 291 p. Pryce-Hobby AC, McMaster ME, Hewitt LM Van Der Kraak G. 2003. The effects of pulp mill effluent on the sex steroid binding protein in white sucker ( Catostomus commersoni ) and longnose sucker ( C. catostomus ). Comp Biochem Physiol C 134:241-250. Quinn B. 2000. Determination of selected resin acids in paper mill effluent and bile from exposed largemouth bass. Gainesville (F L): University of Florida. 74 p.

PAGE 272

254 Quinn B. 2004. Analysis of selected natural compounds and their degradation products in pulp and paper mill effluent: exploration of possible endocrine disruption. Gainesville (FL): University of Florida. 88 p. Rincon PA, Correas AM, Morcillo F, Risu eno P, Lobon-Cervia J. 2002. Interaction between the introduced eastern mosqu itofish and two autochthonous Spanish toothcarps. J Fish Biol 61:1560-1585. Robinson RD. 1994. Evaluation and development of laboratory protocols for estimating reproductive impacts of pulp mill effluent on fish. Guelph (ON): University of Guelph. 175p. Rogers IH, Levings CD, Lockhart WL Norstrom RJ. 1989. Observations on overwintering juvenile chinook salmon ( Oncorhynchus tshawytscha ) exposed to bleached kraft mill eflluent in the upper Fraser River, British Columbia. Chemosphere 19:1853-1868. Rosa-Molinar E, Williams CS. 1984. Notes on fecundity of an arrhenoid population of mosquitofish, Gambusia affinis holbrooki Northeast Gulf Sci 7:121-125. Rosen DE, Bailey RM. 1963. The poeciliid fish es (Cyprinodontiformes), their structure, zoogeography and systematics. Bull Amer Mus Natur Hist 126:1-176. Ruoppa M, Paasivrita J, Lehtinen KJ, Ruona la S, editors. 2003. Proceedings of the 4th International Conference on Environmental Fa te and Effects of Pulp and Paper Mill Effluents; 2000 June 12-15; Helsinki. Helsinki: Finnish Environment Institute. Report 417. Sandstrom O, Neuman E, Karas P. 1988. E ffects of a bleached pulp mill effluent on growth and gonad function in baltic co astal fish. Wat Sci Technol 20:107-118. Seplveda MS. 2000. Effects of paper mill e ffluents on health and reproductive success of largemouth bass ( Micropterus salmoides ): field and laboratory studies. Gainesville (FL): University of Florida. 265 p. Seplveda MS, Johnson WE, Higman JC, De nslow ND, Schoeb TR, Gross TS. 2002. An evaluation of biomarkers of reproduc tive function and potential contaminant effects in Florida largemouth bass ( Micropterus salmoides floridanus ) sampled from the St. Johns River. Sci Total Environ 289:133-144. Seplveda MS, Quinn BP, Denslow ND, Holm SE, Schoeb TR, Gross TS. 2003. Effects of pulp and paper mill effluents on reproductive success of largemouth bass. Environ Toxicol Chem 22:205-213. Seplveda MS, Ruessler DS, Denslow ND, Holm SE, Schoeb TR, Gross TS. 2001. Assessment of reproductive effects in largemouth bass ( Micropterus salmoides ) exposed to bleached/unbleached kraft mill effluents. Arch Environ Contam Toxicol 41:475-482.

PAGE 273

255 Servizi JA, Gordon RW, Martens DW, Lockha rt WL, Metner DA, Rogers LH, McBride JR, Norstrom RJ. 1993. Effect of biotre ated bleached kraft mill effluent on fingerling Chinook salmon ( Onchorhynchus tshawytscha ). Can J Aquat Sci 50:846-857. Servos MR, Carey JH, Ferguson ML, Van Der Kraak G, Ferguson H, Parrott J, Gorman K, Cowling R. 1992. Impact of a moderniz ed bleached kraft mill on white sucker populations in the Spanish River, Ontario. Wat Poll Res J Can 3:423-437. Servos MR, Munkittrick KR, Carey JH, Van Der Kraak GJ, editors. 1996. Environmental Fate and Effects of Pulp and Paper Mill Effluents. 2nd International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluents; 1994 November 6-10; Vancouver. DelRay Beach: St. Lucie Press. 703 p. Smook GA. 1999. Handbook for pulp and paper technologists. 2nd ed. Vancouver: Angus Wilde Publications. 419 p. Snelson FF. 1989. Social and environmental cont rol of life history traits in Poeciliid fishes. In: Meffe GK, Snelson FF, editors Ecology and evolution of livebearing fishes (Poeciliidae). Englewood C liffs: Prentice-Hall. pp 149-161. Sodergren A, editor. 1991. Environmental fate and effects of bleached pulp mill effluents. 1st International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluents; 1991 November 19-21; Saltsjobaden. Sweden: SEPA Conference. Report 4031. Soimasuo MR, Aaltonen T, Nikinmaa M, Pellinen J, Ristola T, Oikari AO. 1995. Physiological toxicity of low-chlorine bleached pulp and paper mill effluent on whitefish ( Coregonus lavaretus L. s.l. ): a laboratory exposure simulating lake pollution. Ecotox Environ Safety 31:228-237. Soimasuo MR, Karels AE, Leppanen H, Santti R, Oikari AO. J. 1998. Biomarker repsonses in whitefish ( Coregonus lavaretus L. s.l.) experimentally exposed in a large lake receiving effluent s from pulp and paper indus try. Arch Environ Contam Toxicol 34:69-80. Specziar A. 2004. Life history pattern and feeding ecology of the introduced eastern mosquitofish, Gambusia holbrooki in a thermal spa under temperate climate, of Lake Heviz, Germany. Hydrobiologia 522:249-260. Staub BP, Hopkins WA, Novak J, Congdon JD. 2004. Respiratory and reproductive characteristics of eastern mosquitofish ( Gambusia holbrooki ) inhabiting a coal ash settling basin. Arch Environ Contam Toxicol 46:96-101. Stephensen E, Adolfsson-Erici M, Frlin L. 1998. Biochemical effects in primary cultured hepatocytes from rainbow trout afte r treatment with extracts of effluent waters from pulp mills. Mar Environ Res 46:395.

PAGE 274

256 Stuthridge T, van den Huevel M, Marvin N, Slade A, Gifford J, editors. 2003. Environmental impacts of pulp and paper waste streams. 3rd International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluents; 1997; Rotorua. Pensacola: SETAC Press. CD. Svenson A, Allard AS. 2004. In vitro androgenicity in pulp and paper mill effluents. Environ Toxicol 19: 510-517. Swanson S, Shelast R, Schryder R, Kloepper-S ams P, Marchant T, Kroeber K, Bernstein J, Owens JW. 1992. Fish populations and biomarker responses at a Canadian bleached kraft mill site. TAPPI 75:139-149. Teh SJ, Miller CE, Hinton DE. 2000. Herma phroditism in laboratory-cultured albino Western mosquitofish. J Aqua t Animal Health 12:78-80. Thibaut R, Monod G, Cravedi JP. 2002. Residues of 14 C-4n-nonylphenol in mosquitofish ( Gambusia holbrooki ) oocytes and embryos during dietary exposure of mature females to this xenohormone. Mar Environ Res 54:685-689. Toft G, Edwards TM, Baatrup E, Guillette LJ. 2003. Disturbed sexual characteristics in male mosquitofish ( Gambusia holbrooki ) from a lake contaminated with endocrine disrupters. Environ Health Perspect 111:695-701. Tremblay L, Van der Kraak G. 1998. Use of a series of homologous in vitro and in vivo assays to evaluate the endoc rine modulating actions of -sitosterol in rainbow trout. Aquat Toxicol 43:149-162. Turner CL. 1937. Reproductive cycles and superfe tation in poeciliid fishes. Biol Bull 72:145-164. Turner CL. 1941a. Gonopodial characteristics pr oduced in the anal fins of females of Gambusia affinis affinis by treatment with ethinyl te stosterone. Biol Bull 30:371383. Turner CL. 1941b. Morphogenesis of the gonpodium in Gambusia affinis affinis J Morph 69:161-185. Turner CL. 1942a. Morphogenesis of the gonopodial suspensorium in Gambusia affinis affinis and the induction of male suspenso rium characters in the female by androgenic hormones. J Exp Zool 91:167-193. Turner CL. 1942b. A quantitative study of the e ffects of different concentrations of ethinyl testosterone and methyl testosterone in the production of gonopodia in females of Gambusia affinis affinis Physiol Zool 15:263-280. Turoski V. 1998. Chlorine and chlorine co mpounds in the paper industry. 1998. Sleeping Bear Press, Chelsea, MI.

PAGE 275

257 US Department of Commerce. 2000. US Industry and Trade Outlook 2000. Washington DC: US Department of Commerce/Intern ational Trade Administration. C-61-34-2. CD. US Department of Energy. 2000. Forest Pr oducts Project Fact Sheet: Closed Cycle Bleached Kraft Production. Office of Industrial Technologies. US Environmental Protection Agency. 1997. EP A’s final pulp, paper, and paperboard “Cluste Rule” – overview fact sheet. Of fice of Water. EPA-821-F-97-010. 2 p. US Environmental Protection Agency. 2002. Profile of the pulp and paper mill industry. 2nd ed. Washington DC: Office of Complia nce. Sector Notebook Project. EPA-310R-02-002. 135 p. US Environmental Protection Agenc y. 2004. About biological indicators. http://www.epa.gov/bioindicators/html/about.html Accessed 2004 December 7. van den Huevel MR. 2004a. Overview of fi sh health studies—progress since 2000. In: Borton DL, Hall TJ, Fisher RP, Thomas JE, editors. Pulp and paper mill effluent environmental fate and effects. 5th International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluent s; 2003 June 1-4; Seattle. Lancaster: DEStech Publications. p 3-13. van den Huevel MR, Ellis RJ, Smith MA, Fi nley M, Stuthridge TR. 2004b. Review of reproductive-endocrine effects of a New Z ealand pulp and paper mill effluent. In: Borton DL, Hall TJ, Fisher RP, Thomas JE, editors. Pulp and paper mill effluent environmental fate and effects. 5th International Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluent s; 2003 June 1-4; Seattle. Lancaster: DEStech Publications. p 55-66. van den Huevel MR, Ellis RJ, Tremblay LA, Stuthridge TR. 2002. Exposure of reproductively maturing rainbow trout to a New Zealand pulp and paper mill effluent. Ecotox Environ Safety 51:65-75. Van Der Kraak GJ, Munkittrick KR, McMa ster ME, Portt CB, Chang JP. 1992. Exposure to bleached kraft pulp mill effluent disrupts the pituitary-gonadal axis of white sucker at multiple sites. Toxicol Appl Pharmacol 115:224-265. Van Der Kraak GJ, Zacharewski T, Janz DM, Sanders BM, Gooch JW. 1998. Comparative endocrinology and mechanisms of endocrine modulation in fish and wildlife. IN: Principles and processes for evaluating endocrine disruption in wildlife. Kendall R, Dickerson R, Giesey J, Suk W editors. SETAC Press, Pensacola, FL, 97-119. Vondracek B, Wurtsbaugh WA, Cech JJ. 1988. Growth and reproduction of the mosquitofish, Gambusia affinis in relation to temper ature and ration level: consequences fro life history. Environ Biol Fishes 21:45-57.

PAGE 276

258 Vuorinen PK, Vuorinen M. 1985. Effects of bleached kraft mill e ffluent on reproduction of brown trout ( Salmo trutta L.) on a restricted diet. Finnish Fish Res 6:92-105. Warburton K, Madden C. 2003. Behavioural responses of two native Australian fish species ( Melanotaenia duboulayi and Pseudomugil signifer ) to introduced Poeciliids ( Gambusia holbrooki and Xiphophorus helleri ) in controlled conditions. Proc Linn Soc New South Wales 124:115-123. Warren CE, Seim WK, Blosser R, Caron OG, Owens EL. 1974. Effects of kraft effluent on the growth and production of salmonid fishes. TAPPI 57:127-132. Wear DN, Greis JG, editors. 2002. Southern forest resource assessment. Ashville: US Department of Agriculture, Forest Servi ce, Southern Research Station. General Technical Report SRS-53. 635 p. White R, Jobling S, Hoare SA, Sumpter JP, Parker MG. 1994. Environmentally persistent alkylphenolic compounds are es trogenic. Endocrinology 135:175-182. Yamazaki F. 1983. Sex control and manipul ation in fish. Aquaculture 33:329-354. Zauner H, Begemann G, Mar-Beffa M, Meyer A. 2003. Differential regulation of msx genes in the development of the gonopodi um, an intromittent organ, and of the “sword,” a sexually selected trait of swordtail fishes ( Xiphophorus ). Evol Devel 5:466-477.

PAGE 277

259 BIOGRAPHICAL SKETCH Jessica Joy Noggle was born north of Cincinnati in Middletown, Ohio, on September 5th, 1978. Her primary education was at public schools in the Lakota Local School District, also north of Cincinnati, and she graduated in the Top 25 from Lakota High School with an Honors Diploma in 1996. Jessica came to the University of Florida that same year on a National Merit Scholarship with an interest in wildlife veterinary medicine. Graduating Summ a Cum Laude in December 1999 with her Bachelor of Science in Zoology and a minor in Wildlif e Ecology and Conservation, her focus had shifted to interdisciplinary research. In January 2000, she began working as a research technician in the Ecotoxicology Program at the United States Geological Survey, Florida Integrated Science Center, Center for Aquatic Resource Studies in Gainesville, FL. Her main projects evaluated effects of pulp and paper mill effluents on aquatic wildlife, including mosquitofish. Within the year, Jess ica began graduate studi es at the University of Florida, becoming a Ph.D. candidate in the Department of Physiological Sciences, College of Veterinary Medicine in 2002. She received her Ph.D. with a specialization in Toxicology in the spring of 2005.


Permanent Link: http://ufdc.ufl.edu/UFE0009720/00001

Material Information

Title: Eastern Mosquitofish as a Bioindicator of Pulp and Paper Mill Effluents
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0009720:00001

Permanent Link: http://ufdc.ufl.edu/UFE0009720/00001

Material Information

Title: Eastern Mosquitofish as a Bioindicator of Pulp and Paper Mill Effluents
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0009720:00001


This item has the following downloads:


Full Text












EASTERN MOSQUITOFISH AS A BIOINDICATOR OF PULP AND PAPER MILL
EFFLUENTS















By

JESSICA JOY NOGGLE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Jessica Joy Noggle















ACKNOWLEDGMENTS

First of all, I thank my advisor, Dr. Timothy Gross, for accepting me as a young

and naive graduate student. Dr. Gross taught me the finesse required to navigate the

interdisciplinary realm of toxicology that stretches across academia, industries, and

regulatory agencies. His encouragement to attend meetings, give presentations, and get

involved with the regulatory implications of my research was inspiring. I am forever

grateful for his generosity and patience.

In addition, I would like to thank my other supervisory committee members for

having high expectations that inspired me to work hard and strive for the objective ideals

inherent in science. Dr. Sepulveda served as a role model and example from the very

beginning, helping me pick up the pulp and paper mill fish research where she left off.

Her friendship and support whenever I had a question or needed help will always be

remembered and I will always admire her work ethic and integrity. I thank Dr. Gallagher

specifically for his strict adherence to hypothesis testing and its central importance to

experimental design his scientific rigor has permanently shaped how I pursue and

interpret research. Dr. Percival engaged me with stimulating discussions about the "Ph"

half of my Ph.D. and his conceptual point of view enlightened me about the philosophical

framework of science and its context in society. Finally, Howard Jelks was my favorite

special member! His ecological perspective was invaluable; as were his assistance with

species identification and field work design and his reminder to enjoy the ride.









I am forever indebted to the illustrious crew in the Ecotoxicology Program at

USGS: they put up with long days of field work, processing frustratingly small fish, and

gobs of mosquitofish samples in storage, all with jokes and laughter! Without the

assistance of Shane Ruessler, Nikki Kernaghan, Jon Wiebe, Beverly Arnold, Travis

Smith, Jesse Grosso, and Janet Scarborough, these projects would never have been

completed. Special thanks go to Carla Wieser for teaching me all the ins and outs of

radioimmunoassay techniques and demonstrating how to run a very organized (and safe!)

lab. Special thanks also go to Wendy Mathis for taking care of my paychecks and

ordering anything I needed for my projects, whenever I needed it even cheerleading

pompons. Fellow graduate students Brian Quinn, Heath Rauschenberger, Jennifer

Muller, Eileen Monck, and Kevin Johnson were equally supportive in assistance with

projects and swapping graduate student woes!

These projects were primarily funded by the National Council for Air and Stream

Improvement, Inc., and I thank Ken Bradley and Dennis Borton for their congenial

support and thoughtful discussions over data that furthered my development as a

scientist. Ken aided with several field surveys and much of the validation work on anal

fin morphology, and provided insight into practicalities of experimental design. Without

the initial financial and continuous professional support of Stewart Holm of Georgia-

Pacific Corporation, these projects would never have left the drawing board. In addition

to thanking Stewart, I thank all the mill personnel who generously allowed access to

retention ponds and mill grounds for sampling (specifically Myra Carpenter and Ted

Kennedy of Georgia-Pacific; Tom Deardorff and Joel Bolduc of International Paper; and

Ray Andrews, Chet Thompson and Greg Wynn of Buckeye Technologies). Special









thanks go to Chet and Greg for assistance in the field-I would never have found those

field sites (or unlocked my keys from the truck) without their extensive local knowledge!

Sincere thanks go to my friends and family who saw me through the rollercoaster

life of a graduate student with love and support. Special thanks go to my wonderful yoga

family who supported me with vacations, massages, and the warmth and hospitality of

their homes; and to my Ohio family who loved me through phone calls, snail mail, and

happy family reunions. I extend deepest thanks to my parents, Denny and Donna

Noggle, for their unconditional support of all my endeavours. I am eternally grateful for

their tag team efforts through the last semester of my doctoral program! Finally, heartfelt

thanks go to my spiritual teachers for leading me with strength, compassion, and the

perspective that all is sacred.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ................................................................................................. iii

LIST OF TABLES .............. ................. ........... ................... ........ xi

LIST OF FIGURES ......... ....................... .......... ....... ............ xiv

ABSTRACT ........ .......................... .. ...... .......... .......... xvii

CHAPTER

1 MOSQUITOFISH EXPOSED TO PULP AND PAPER MILL EFFLUENTS:
USE OF A POTENTIAL INDICATOR OF EXPOSURE AND EFFECTS................1

B background of Pulp and Paper M ills.................................................... .................. 1
Economic Importance of Pulp and Paper Industry in the United States ...............1
Production Processes and Technologies..................................... .....................2
P u lp in g ............................................... .......................... 3
B leaching ..................................... ............................................ 5
W ater Pollution and R regulation in the U S ........................................ .................6
Effects of Pulp and Paper Mill Effluents on Fish.................................... ..................9
N onreproductive E ffects.......................................................... ............... 13
R productive E effects ................... ............................................ ....... ...... 15
Masculinization and Femininization Effects............. ...................18
Effects of Pulp and Paper Mill Effluents on Mosquitofish.................. ...............22
M ascu lin ization .................................................. ................ 2 2
P recociou s m atu ration .............................................................. .....................2 6
B behavior ........................................... ........................... 27
R ep ro du ctio n .................................................... ................ 2 8
M echanism of A action ................................................ .............................. 29
M osquitofish as a M odel Species ...................... .................. ............... .... 30
Occurrence and Availability in Effluent-Receiving Systems............................30
Reproductive Characteristics................. ..... ............................ ...............31
Mosquitofish as a Bioindicator of Pulp and Paper Mill Effluent ..............................32
Definitions: Bioindicator and Biomarker ...............................................33
B ioindicator Criteria for Success...................................... ........ ............... 35
Practicality ............. .. ........... ......... ................... ............ 35
V ariab ility ............................................... ................ 3 5









P red ictab ility ............................................................3 5
C contribution of M y Study ............................................................................ ...... 37
S p e c ifi c A im 1 ............................................................................................... 3 8
S p e cifi c A im 2 ............................................................................................... 3 9

2 VALIDATION OF MOSQUITOFISH ENDPOINTS USED TO ASSESS
EFFECTS OF PULP AND PAPER MILL EFFLUENT EXPOSURE.....................42

Introdu action ...................................... ................................................. 4 3
M materials and M methods ............................................................. ............................44
M ill Characteristics and Field Collection................................. ............... 44
Gender Identification Using the Urogenital Papilla................. ............. .....45
A nal Fin M orphology .............................................. .............................. 45
Sex Steroids .................................... ................................ ........47
Statistics ...................................... ............................... .......... ...... 4 9
R results and D discussion ........................... ...... ..... ...... .. .............. 50
W after Q quality ........................................50
Validation of Gender Identification Using the Urogenital Papilla.....................50
M o rp h o lo g y ................................................................. ................................5 1
Validations .............................................................. 51
Body Size for Fall 2000 Collection................................... ............... 52
Influence of Body Size on Anal Fin Morphology......................................53
S e a so n a lity ............................................................................................. 5 4
Sex Steroids ............................................................................................... 55
V a lid atio n s ............................................................................................. 5 6
S e a so n a lity ............................................................................................. 5 7
C o n c lu sio n s........................................................................................................... 6 0

3 DIMINISHED EFFECTS OF PULP AND PAPER MILL EFFLUENT ON
EASTERN MOSQUITOFISH BEFORE AND AFTER MAJOR PROCESS
IMPROVEMENTS ................................................................................... ........ .... ........76

Introdu action ............. ......................................................................................77
M materials an d M eth od s ......................................................................................... 79
M ill C h aracteristics ....................................................................................... 7 9
Field C ollections............................................. 80
Morphology ........... 8........... .. ........ ......... 81
Sex Steroids ............ ......... .. ......... ......... 81
S statistics ...................... .. ............. ..................................................... 82
Results and Discussion ....................... ............ .. ............... 82
W after Q u ality ...............................................................82
Body Size and Condition.............................................................83
M a le s ................................................................8 3
F e m a le s ................................................................................................... 8 4
A nal Fin M orphology ...................... ..................... .. .. ...........................85
M a le s ................................................................ 8 5
F em ale s....................................................8 6









S ex S te ro id s .................................................................................................... 8 7
M a le s ................................................................ 8 8
F em ales ............................. .... .. ... .. .. ... .................. 8 9
Association Between Anal Fin Morphology and Sex Steroids ..........................90
C o n c lu sio n s..................................................... ................ 9 1

4 VARIABLE EFFECTS OF EFFLUENT ON EASTERN MOSQUITOFISH
COLLECTED BELOW THREE FLORIDA PULP AND PAPER MILLS.............. 105

Introduction ..................................................... .. ........ ................. 106
M materials and M methods ............................................ ...................................... 108
M ill C characteristics ........................ .. .............................. .... ............108
W ater Sam ples ............................................................... 109
F ish S am p les............................................................................. ............... 10 9
Sex Steroids ............ ...... .......... ................... ...........110
Statistics ........... ........ ........ .. ............... ........................ 110
R results and D discussion ............. ....................... ........ ..... ...... .......... .... 111
W after Quality .............. .. ....... .... .................. .................. 111
W after C h em istry ............................................................. ........ ..... .. ... 112
B ody Size and C condition ...................... .. ............. ................... ............... 113
M ales ....................... ...... ............. .......... ..... ......... 113
F e m a le s ................................................ ................ 1 14
A nal Fin M orphology ............... ............................ ...............................114
M ales ............. ......... ... .. ............. .... ...... ......... .115
F e m a le s ............................................ ....... .............. 1 1 6
Sex Steroids ............. .. ........... ............................. ... .... 118
M ales ............. ......... ... .. ............. .... ...... ......... .119
F em ales ............................................ .............................. 12 0
Anal Fin Elongation and Sex Steroids...................................... ............... 122
M ales ....................... ...... ............. .. ....................... 122
F e m a le s ................................................ ................ 12 3
C o n clu sio n s.................................................... ................ 12 4

5 DIFFERENTIAL INDUCTION OF EFFECTS IN MOSQUITOFISH
EXPOSED TO BLEACHED KRAFT MILL EFFLUENT.................................... 137

In tro du ctio n .................................................................................................. .... 13 8
M materials and M ethods ............................................ ...................................... 140
M ill Characteristics and Exposure Scenarios...................................................140
W ater Sam ples ............................................................... 14 3
M orphological Endpoints ......................... ..... ............ ................. .. 143
H orm onal E ndpoints.................................................. .............................. 144
S statistic s .................................................... .............. ................ 14 4
R results and D iscu ssion ............................. .................................................. 14 5
W ate r Q u ality ............................................................................................... 14 5
W after C h em istry ............................................................. ................... ... 14 6
B ody Size and C condition .......................................................... ............... 147


viii









M ales ......................................... .................. .. .... ........ 147
F e m a le s ................................................................................................. 1 4 7
A nal Fin M orphology .......................................................... ............... 148
M ales ......................................... .................. .. .... ........ 148
F e m a le s ....................................................................................................... 1 4 9
Sex Steroids ......................................................................................... ....... 150
M a le s ................................................................. ................................. 1 5 0
F em ales.................... ..... ...................153
Anal Fin Elongation and Sex Steroids............................................... 156
C o n clu sio n s......................................................................................15 6

6 INVESTIGATION OF REPRODUCTIVE SUCCESS IN MOSQUITOFISH
LIVING IN PULP AND PAPER MILL EFFLUENT DOMINATED SYSTEMS .170

Introdu action ...................................................................................................17 1
M materials and M methods ........................................... ....................................... 173
M ill C characteristics ............................................ .. .. .. ...... ...............173
W ate r S a m p le s ............................................................................................. 17 3
P population Survey ..................... .. ........................ .. .. .. ...............174
Morphology ...... ...................... ......... ........175
Sex Steroids.............................................. 176
Fry Production ........................ .......... ......... 176
S ta tistic s ....................................................................................................... 1 7 7
Results and Discussion ................................. ........................ .. ...... ........ 179
W ate r Q u ality .............................................................................................. 17 9
W after C h em istry .......................................................................................... 18 0
P population Survey ...........................................................18 1
B o d y S iz e ..............................................................18 4
A nal fin m orphology ............................................................. ............ 184
Sex steroids ........................................185
Anal Fin Elongation and Sex Steroids............................................... 186
Fry Production ................................................................. .. ......... 187
Sum m er 2003 ....................................... ................ .... .... 187
Sum m er 2004 ...................................... ... .. ... ... ............ 189
Anal Fin Elongation and Fry Production.............................. ............... 192
Conclusions........................... .............. ......... ... ......193

7 EVALUATION OF MOSQUITOFISH AS A BIOINDICATOR OF PULP AND
PAPER MILL EFFLUENT EXPOSURE ........................................ ....216

Su m m ary .................................................................. ................. 2 17
Specific Aims Revisited ...........................................................219
B ioindicator Criteria R visited ........................................... .................. ........ 221
Other M odel Fish Species ................ ........ .. ............... ............ 223
F future W ork .......................................................................................... 22 5










APPENDIX

A F IE L D S IT E S ....................................................................................... ...... .. 2 2 9

B SEX STEROID RADIOIMMUNOASSAY PROTOCOLS .................................... 237

C POSTER AND PLATFORM PRESENTATIONS OF DISSERTATION
R E SE A R C H ........................................... .......... ................. 240

L IST O F R EFER EN CE S ........................................................................... ..............242

B IO G R A PH IC A L SK E T C H ........................................... ...........................................259












































x















LIST OF TABLES


Table pge

1-1 Select characteristics of the mills in my study according to receiving stream.........40

2-1 Water quality parameters of Rice Creek field collection sites in winter 2000.........64

2-2 Correlation coefficients (r2) for morphological measurements made before and
after preservation in formalin; between USGS and NCASI laboratories; and
between manual and computer-aided measurement by the same observer.............64

2-3 Average coefficients of variation for manual and computer-aided
measurements by observer and among observers ..........................................64

2-4 Body size parameters (ave + se) for mosquitofish collected in winter 2000 ..........65

2-5 Digestion and extraction efficiencies, and coefficients of variation (CV), by
exposure and reproductive status for mosquitofish whole body hormone
analy sis. .............................................................................66

3-1 Water quality parameters of field collection sites before (2000) and after
(2002) process changes at the Georgia-Pacific Palatka mill.................................95

3-2 Body size parameters (ave + se) and sample sizes for mosquitofish collected
before (2000) and after (2002) process changes. ............................................ 96

4-1 Water quality parameters of field collection sites associated with three
effluent-receiving streams in Florida the summer of 2001 ..............................127

4-2 Concentration of selected effluent components in single grab water samples
from field collection sites associated with three effluent-receiving streams in
Florida the sum m er of 2001 ......... .................. ................... ....................... 128

4-3 Body size parameters (ave + se) for mosquitofish collected from three effluent-
receiving streams in Florida the summer of 2001 ................................................ 129

5-1 Water quality parameters (ave + se) measured three times weekly (n = 13 total)
during four week tank exposures of mosquitofish to bleached/unbleached kraft
m ill effluent in sum m er 2002. ..........................................................................159









5-2 Water quality parameters (ave + se) measured three times weekly (n = 12 total)
during caged exposures of mosquitofish to field sites in Rice Creek during
sum m er 2002. .......................................................................159

5-3 Concentrations of selected effluent components in 100% final effluent sampled
weekly midJanuary to midM ay in 2002. ...................................... ............... 159

5-4 Body size parameters (ave + se) for mosquitofish exposed to
bleached/unbleached pulp mill effluents via whole effluent dilutions for four
w eeks in sum m er 2002. ..... ........................... ......................................... 160

5-5 Body size parameters (ave + se) for mosquitofish caged in Rice Creek field
sites for four weeks in summer 2002. ....................................... ............... 161

6-2 Water quality parameters (ave + se) measured three times weekly (n = 16 total)
during laboratory fry production of female mosquitofish collected from field
sites in Rice Creek during sum m er 2003 ...............................................................197

6-3 Water quality parameters (ave + se) at field sites where female mosquitofish
were collected for fry production studies over 4 months in summer 2004............198

6-4 Water quality parameters (ave + se) measured three times weekly (n = 16 total)
during laboratory fry production of female mosquitofish collected from field
sites in Rice Creek and Fenholloway River during summer 2004.......................198

6-5 Concentrations of selected effluent components (ave + se) in single grab water
samples from field sites where female mosquitofish were collected in Rice
Creek and Fenholloway River during summer 2003 ..........................................199

6-6 Concentrations of selected effluent components (ave + se) in single grab water
samples from field sites where female mosquitofish were collected in Rice
Creek and Fenholloway River during summer 2004 for fry production studies....200

6-7 Body size parameters (ave + se) for mosquitofish collected for population
survey of Fenholloway River and Rice Creek in May 2003 ...............................201

6-8 Reproductive and morphological characteristics of females collected from
Fenholloway River and Rice Creek and monitored for fry production in 2003.....202

6-9 Reproductive and morphological characteristics of females collected for fry
production from Fenholloway River in 2004............... ............... ..................203

6-10 Reproductive and morphological characteristics of females collected for fry
production from Rice Creek in 2004 .................................... ......... ............... 204

A-1 Latitude, longitude, and descriptions for mosquitofish collection sites in Rice
C reek .............................................................................232









A-2 Latitude, longitude, and descriptions for mosquitofish collection sites in
Fenhollow ay R iver. ...................... .... .............. ............................ 233

A-3 Latitude, longitude, and descriptions for mosquitofish collection sites in
Elevenm ile Creek. ..................................... .............. .. ..............234
















LIST OF FIGURES


Figure p

1-1 Categories of pulp and paper m ill facilities ...................................... .....................41

2-1 Maps of Rice Creek, a tributary of the Saint Johns River, FL, USA.....................67

2-2 Gender agreement between NCASI and USGS laboratories .................................69

2-3 Gender agreement within USGS laboratory .........................................................70

2-4 Female index of anal fin elongation for each site by 5 mm increments (winter
2 0 0 0 ) .......................................................................... 7 1

2-5 Index of anal fin elongation for winter and summer months in 2000 .................72

2-6 Female whole body sex steroids from collections made in the summer and
w inter of 2000 (ave + se)...................... ..... ......... .............................. .. ......... ...... 73

2-7 Percentage of female mosquitofish with masculine and feminine sex steroid
ratios collected in 2000. ......................... .................. ... ...... .. .... ........... 74

2-8 Male whole body sex steroids from collections made in the summer and winter
of 2000 (ave + se) ............ .... .......................... ...... ..... ....... 75

3-1 Maps of Rice Creek and Saint Johns River, USA............ .............................97

3-2 Representative male gonopodia from the upstream site collected before and
after process changes.......... ............................................................ ..... ... ... 98

3-3 Male index of anal fin elongation for each site by 0.1 mm increments ...................99

3-4 Representative female anal fins from collections made before and after process
changes. .............................................................................100

3-5 Female index of anal fin elongation for each site by 0.1 mm increments ...........101

3-6 Male whole body sex steroids (ave + se) from collections made before (2000)
and after (2002) process changes ................................................ ............... 102

3-7 Female whole body sex steroids (ave + se) from Rice Creek collections made
before (2000) and after (2002) process changes ......................................... 103









3-8 Percentage of female mosquitofish with masculine and feminine sex steroid
ratios collected after process changes in 2002 .............. ......................... ......... 104

4-1 M aps of field sites .......................... ........................ .. .... .. .. ..... ........ 130

4-2 Index of anal fin elongation for mosquitofish collected in summer 2001 from
three effluent-receiving system s in Florida................................. ...... ............ ...132

4-3 Whole body sex steroids (ave + se) for male mosquitofish collected from three
effluent-receiving streams in Florida the summer of 2001 ...............................133

4-4 Whole body sex steroids (ave + se) for female mosquitofish collected from
three effluent-receiving streams in Florida the summer of 2001 ........................134

4-5 Percentage of female mosquitofish with masculine and feminine sex steroid
ratios collected from three effluent-receiving streams in Florida the summer of
2 0 0 1 ........................................................................................... 1 3 5

5-1 Diagram of tank facility for flow-through whole effluent exposure of
mosquitofish in summer 2002 at Georgia-Pacific's Palatka, FL operation. ..........162

5-2 Map of cage locations for in situ field exposures at Rice Creek, FL, in 2002.......163

5-3 Concentrations of selected wood extractives in 100% final effluent from the
Rice Creek mill during tank and field exposures of mosquitofish.........................164

5-4 Index of anal fin elongation (length ratio of Ray 4 to Ray 6) for mosquitofish
exposed to bleached/unbleached pulp mill effluents via whole effluent
dilutions or onsite caged exposures for four weeks in summer 2002 ..................165

5-5 Whole body sex steroids (ave + se) for male mosquitofish exposed to
bleached/unbleached pulp mill effluents via whole effluent dilutions or onsite
caged exposures for four weeks in summer 2002 ..............................................166

5-6 Percentage of male mosquitofish with masculine and feminine sex steroid
ratios exposed to bleached/unbleached pulp mill effluents via whole effluent
dilutions or in situ field exposures for four weeks in summer 2002....................167

5-7 Whole body sex steroids (ave + se) for female mosquitofish exposed to
bleached/unbleached pulp mill effluents via whole effluent dilutions or onsite
caged exposures for four weeks in summer 2002. ............................................168

5-8 Percentage of female mosquitofish with masculine and feminine sex steroid
ratios exposed to bleached/unbleached pulp mill effluents via whole effluent
dilutions or in situ field exposures for four weeks in summer 2002......................169

6-1 M aps of field sites .......................... ........................ .. .... .. .. ..... ..... .. 205









6-2 Representative changes in resin acid concentrations during summer 2003 at
Fenholloway River and Rice Creek field sites where mosquitofish were
collected ...........................................................................2 0 6

6-3 Representative changes in resin acid concentrations during summer 2004 at
Fenholloway River and Rice Creek field sites were mosquitofish were
collected at the sam e tim e ............................................. ............................. 207

6-4 Estimated relative abundances of mosquitofish for Fenholloway River and
R ice C reek sites .......................................................................208

6-5 Estimated age and sex structure of mosquitofish populations living near pulp
and paper m ill effluent discharge .................................. ...................................... 209

6-6 Index of anal fin elongation for mosquitofish collected in summer 2003 from
Fenhollow ay River and Rice Creek ............................................ ............... 210

6-7 Whole body sex steroids (ave + se) for mosquitofish collected in summer 2003
from Fenholloway River and Rice Creek .............................. .......... ...............211

6-8 Percentage of mosquitofish with masculine and feminine sex steroid ratios
collected in summer 2003 from Fenholloway River and Rice Creek (systems
divided by solid black line). ..... .......................................................................212

6-9 Viability of primary and secondary clutches produced by females collected
from Fenholloway River and Rice Creek in 2003...............................213

6-10 Fecundity and individual fry weight of primary and secondary clutches
produced by female mosquitofish collected from Fenholloway River and Rice
Creek in sum m er 2003 .................................................................. .....................214

6-11 Adjusted fecundity of primary clutches produced by female mosquitofish
collected m monthly in 2004 .............................................. ............................ 215

A-1 Total monthly precipitation for Florida regions where mosquitofish were
collected from pulp and paper mill effluent-receiving systems...........................235















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

EASTERN MOSQUITOFISH AS A BIOINDICATOR OF PULP AND PAPER MILL
EFFLUENTS

By

Jessica Joy Noggle

May 2005

Chair: Timothy S. Gross
Major Department: Physiological Sciences

A variety of sublethal physiological effects have been reported for fish exposed to

pulp and paper mill effluents. As mill processing technologies improve, mounting

evidence demonstrates fewer effects potentially linked to reduction in wood extractives.

Repercussions of sublethal effects at higher levels of biological organization are

important questions beginning to be explored. The goal of my study was to evaluate

whether sublethal effects in mosquitofish can be used reliably to indicate adverse impact

of pulp and paper mill effluents. Biomarkers of anal fin morphology and whole body sex

steroids were validated, then studied extensively in wild-caught mosquitofish from

Florida effluent-receiving streams that varied markedly in effluent composition. These

biomarkers were also examined under short-term controlled whole effluent exposures

(caged in field and in tanks at 0, 10, 20, 40, and 80% dilutions); and in relation to fry

production and preliminary population surveys. Extent of female anal fin

masculinization, or development of male-like secondary sex characteristics, was









associated with increasing concentrations of wood extractives in water samples from field

sites. Implementation of EPA's Cluster Rule at one mill was followed by a significant

reduction, but not elimination, of this response. However, masculinization could not be

reproduced under controlled exposure, likely due to insufficient exposure duration.

Alterations in sex steroids were manifested in both sexes for wild-caught and

experimentally-exposed fish: males exhibited feminized hormonal profiles will females

displayed masculinized profiles. Differential responses among cage-exposed and tank-

exposed fish indicated additional environmental factors (such as bacterial communities

hypothesized to degrade effluent components into androgenic compounds) were

influential in producing responses. However, large natural variation at unexposed sites

and indication of seasonality precluded definitive interpretation. The lack of association

between these biomarkers demonstrated whole body sex steroids cannot be used to

predict morphological masculinization; but they can be compared as biomarkers of recent

versus past exposure. Finally, reproductive success studies implied mosquitofish may

adapt different reproductive strategies in effluent-receiving streams. Since neither

biomarker could be linked to differences in fry production or population structure, at this

point mosquitofish may not be a suitable bioindicator of adverse effects due to pulp and

paper mill effluents.


xviii














CHAPTER 1
MOSQUITOFISH EXPOSED TO PULP AND PAPER MILL EFFLUENTS: USE OF A
POTENTIAL INDICATOR OF EXPOSURE AND EFFECTS

Pollution from both point and nonpoint sources of human activity releases a variety

of chemicals into the aquatic environment. While lethal effects have been addressed,

sublethal effects of these chemicals on aquatic wildlife remain controversial. Whether

observed sublethal effects lead to adverse impacts including reproductive, population, or

community level effects is arguable. This question has been strongly debated about pulp

and paper mill effluents and sublethal effects in fish. The controversy becomes more

complex as major processing improvements are implemented by the industry. The goal

of my study was to evaluate whether sublethal effects in mosquitofish can be reliably

used to indicate adverse impacts of pulp and paper mill effluents.

Background of Pulp and Paper Mills

Economic Importance of Pulp and Paper Industry in the United States

Pulp and paper products (such as writing and copy paper; sanitary tissues;

cardboard; linerboard; and indirect products like pill capsules, diapers; and rayon) are

significant economic commodities in the United States (US). The US produces around

30% of the world's paper and paperboard (US Environmental Protection Agency (EPA)

2002). In 2000, US pulp and paper mills produced 79 billion US dollars in shipments and

employed 182,000 people, while Americans consume around 300 kg of paper-based

products each year (EPA 2002). Industrialization of nations increases demand for pulp









and paper products (Smook 1999), resulting in high capital investments and intensive use

of forests, water, and energy.

Furnish (tree species or specifically cellulose fiber source) usually comes from

harvested hardwood and softwood tree species (Table 1-1). The fiber length varies

tremendously between these two tree types (longer in softwoods), so the final product

determines which tree type and species is used. Alternative furnishes include recycled

paper (increasingly used, especially for products like corrugated board) and nonwood

sources such as bagasse, bamboo, cereal straws, cotton rags and linters, flax, hemp, and

synthetic fibers (Smook 1999). Pulp and paper production consumes large amounts of

forestry resources. Approximately 6 million acres of forest in the southeastern US are

logged annually, mostly for paper production (Wear and Greis 2002). The process is also

water and energy intensive. The pulp and paper industry is classified as the largest

industrial water consumer and the third largest industrial energy consumer in the US (US

Department of Commerce 2000, US Department of Energy 2000). Efforts toward more

sustainable forestry practices (e.g., the Forest Stewardship Council certification program,

http://www.fscus.org/), water use reduction, and increased use of wood waste material for

fuel are ongoing to reduce natural resource consumption.

Production Processes and Technologies

The following summary is based on information from Smook (1999) and US EPA

(2002), unless cited otherwise. Final paper products are generated in two overall steps:

pulp production and paper or paperboard manufacture. Pulp and paper mills can be

classified by whether they produce pulp, paper/paperboard, or both (Figure 1-1). Most

US mills are nonintegrative facilities and produce paper products using pulp obtained off-









site (54%). About one-third are integrative and produce pulp and final paper products

(36%), Ten percent exclusively produce market pulp.

Depending on final product, 80% of market pulp is for paper and 20% for nonpaper

(Figure 1-1). Nonpaper pulps are either dissolving pulp, fluff pulp, or specialty pulp.

Dissolving pulp is a "chemical cellulose" that can be converted into rayon, cellophane,

cellulose acetates, cellulose nitrate, and carboxymethyl cellulose via a modified kraft or

sulfite pulping process. These chemicals are used in a variety of nonpaper products

ranging from synthetic clothing to pill capsules to air filters. Fluff pulp is a very soft and

absorbent form of pulp used in diapers, feminine products, and hospital pads. Specialty

pulp comprises the remaining nonpaper pulps that do not fit in the other two groups (final

products include components of shoe soles and laminates).

Facilities involved in pulp production (integrative and market pulp facilities) face

the biggest challenges of water pollution in the industry. Pulp production consists of five

major steps: furnish preparation (debarking and chipping); pulping (breakdown of furnish

into fibers); pulp refinement (removal of impurities, cleaning and thickening of pulp);

bleaching (to whiten and brighten the pulp); and stock preparation (wet additives are

integrated into pulp based upon desired end product). Two of these steps (pulping and

bleaching) are considered the dominant sources of water pollution within pulp production

(details next).

Pulping

Two major components of wood are cellulose (the fibers) and lignin (the glue

holding fibers together). Broadly speaking, pulping unglues wood and reduces it to a

fibrous mat. More specifically, the goal of pulping is to retain intact cellulose fibers

while releasing all other wood components. These components include hemicellulose;









lignin; and extractives such as resin acids, fatty acids, phytosterols, turpenoids and

alcohols. Pulping methodologies are generally classified as chemical, semichemical, or

mechanical. Most North American pulping technologies (70%) involve chemical

processes (Table 1-1). Chemical pulping digests wood chips at high temperatures and

pressures, usually in an alkaline solution (called the kraft process), or historically in an

acidic solution (the sulfite process).

The kraft process (also known as the sulfate process) dominates North American

chemical pulping technologies (95%). Major advantages over the sulfite process are high

strength pulp ("kraft" is German for strong) and recovery and reuse of digestion

chemicals. Lignin removal is high, allowing for extensive bleaching without pulp

degradation (via delignification). Additional advantages are the wide range of furnishes

that can undergo the kraft process, and the tolerance for bark. However, pulp yield is

relatively low (40-50% of furnish) compared with mechanical pulping. An additional

disadvantage is pulp color: the kraft process produces a dark brown pulp that requires

extensive bleaching, neutralizing the bleaching-related benefits of high lignin removal.

Overall, the kraft process has proven to be the most cost-effective chemical pulping

technique.

Kraft pulping is cyclical, beginning and ending with white liquor. White liquor,

composed of sodium hydroxide (NaOH) and sodium sulfide (Na2S) as the active

ingredients, is the alkaline solution used to digest wood chips. Temperature and pressure

is elevated using more conventional batch digesters or less common continuous digesters.

Raw pulp and residual black liquor are produced. The pulp is destined for refinement,

bleaching, and stock preparation, while the black liquor is concentrated and burned into









an inorganic smelt that is dissolved to produce green liquor. Some black liquor is washed

away into effluent, and some carries over with the pulp. These black liquor losses impact

effluent quality. Green liquor is then causticized to regenerate white liquor. This

efficient chemical recovery is integral to the success of the kraft process.

Some kraft pulping by-products (turpentine and tall oil) are recovered and either

reused as fuel sources or sold, depending on market prices. The rest of the noncellulose

wood components are discharged (as part of the final effluent) into an aquatic receiving

environment.

Bleaching

Bleaching of pulp is often desirable because it produces a whiter, brighter, softer,

and more absorbent end product. Roughly half of all paper products in the US are

bleached. Bleaching potential for a pulp depends on two factors.

* Inherent lignin content of furnish: higher lignin content gives a darker color, and
softwoods tend to have more lignin than hardwoods.

* Pulping process: sulfite chemical pulping produces a relatively bright pulp with low
residual lignin content, while kraft chemical and semichemical pulping produces a
darker pulp. The brightness of mechanically produced pulp is dependent on lignin
content of furnish.

In general, early stages of the bleaching sequence continue delignification begun during

pulping, while later stages focus on oxidation to remove any residual color.

Modem bleaching uses a continuous sequence of alternating acidic and alkaline

stages with washing between stages. Washing usually involves large amounts of water

that is collected and discharged as part of the final effluent. Several bleaching agents are

available (shorthand used by the industry for bleaching sequences is given in

parentheses).

* Hypochlorite (H)









* Elemental chlorine (C)
* Chlorine dioxide (D)
* Oxygen (0)
* Hydrogen peroxide (P)
* Ozone (Z)
* Sodium hydroxide (E)

Historically, bleaching was accomplished using hypochlorite (Turoski 1998).

Hypochlorite (as either calcium or sodium hypochlorite) was initially the only bleaching

agent at the turn of the twentieth century (H or HH bleaching sequence). The addition of

elemental chlorine gas commercially in 1930 was a major advance that reduced the

amount of hypochlorite needed, and became the standard first stage of bleaching

followed by an extraction (E) stage (CEH). A decade later, chlorine dioxide (e.g.

CEHDED) and hydrogen peroxide (e.g. CEHD(Ep)D) began commercial use. Chlorine

dioxide eventually replaced hypochlorite in the later stages of bleaching by the 1960s

(e.g. CEDED) because of its powerful brightening combined with high selectivity for

lignin. As chlorine dioxide gained popularity, oxygen and ozone bleaching were initiated

(e.g. OCEDED or OZED). These latter agents have been slow to gain acceptance by the

industry because of complications with low selectivity for lignin removal. However, as

environmental and health concerns about chlorine began forming in the 1970s, reduced-

chlorine and chlorine-free methods of bleaching (using 100% chlorine dioxide

substitution, hydrogen peroxide, oxygen and ozone) have been expanded and refined.

Table 1-1 compares bleaching strategies of the three mills in my study.

Water Pollution and Regulation in the US

As previously emphasized, effluent (discharge of liquid waste from a factory/plant)

from integrative and market pulp facilities is of most concern to environmental and

human health. Within pulp production, the pulping and bleaching stages primarily









contribute to water pollution. Four categories of water pollution are currently monitored:

effluent solids, oxygen demand, color, and toxicity. In addition, major water quality

characteristics of effluent-receiving waters (such as pH, temperature, dissolved oxygen,

alkalinity, and conductivity) are expected to remain unchanged or minimally changed

(and may be subject to regulation as well).

Abatement efforts to control these variables occur within plant processing systems

and post-processing. More efficient use of raw materials, reuse of mill waters to create a

more closed system, and reduced effluent volume are strategies within mill operations

that are very effective at increasing profits and at restricting contaminants produced.

Additionally, external or end-of-pipe treatment of effluent helps reduce or remove

contaminants. Primary external treatment entails sedimentation in settling basins to

remove suspended solids. Secondary treatment reduces biochemical oxygen demand

using biological degradation/oxidation. Occasionally, mills also apply a tertiary

treatment to reduce color (turbidity), but this step is costly.

Since pulp production releases disproportionate amounts of chemicals into air and

water (compared to other industries releasing primarily to land), health-related concerns

center on air emissions and aquatic toxicity. With the discovery of dioxins and furans in

fish collected downstream of a pulp and paper mill in 1985 (Smook 1999) and related

evidence for biological effect in fish by the Environment-Cellulose project of Sweden in

the late 1980s (Lehtinen 2004), public attention focused on environmental impacts of

pulp and paper mills. Dioxins and furans are a class of chlorinated organic compounds

produced mainly by incomplete combustion of organic compounds. Natural sources are

forest fires and volcanic eruptions, while human sources include waste incinerators, coal









and oil-fired power plants, vehicle exhaust, chlorinated pesticide and herbicide

production, and chlorine bleaching during pulp production. Although pulp and paper

mills represent a minor source of chlorinated organic, these compounds are lipophilic

and persistent in the environment. Hence, they have the potential to biomagnify up the

food chain through fish and potentially to humans, causing sublethal, chronic toxicity

(which is not traditionally monitored). The most toxic congeners 2,3,7,8-

tetrachlorodibenzodioxin and 2,3,7,8-tetrachlorodibenzofuran (TCDD and TCDF

respectively) are classified as probable human carcinogens by the EPA.1 Release and

exposure predominately occurs as mixtures of chlorinated compounds (measured as

adsorbable organic halides or AOX) including dioxins and furans, which can enhance

toxicity. So pollution prevention efforts have focused on reducing release of chlorinated

compounds as a group.

To reduce the toxic release of chlorinated compounds to both air and water, EPA

enacted a landmark regulation deemed the Cluster Rule in April 1998. The rule set new

baseline limits for toxic and nonconventional pollutant releases; and aims for

approximately 60% reduction in air emissions, and virtual elimination of chlorinated

organic compounds in water (US EPA 1997). Individual mills are allowed flexibility in

tailoring pollution prevention technologies to their specific situations. A voluntary

incentive program for technologies above and beyond the rule grants mills a variable

compliance period (3 to 8 years). Paper-grade bleached kraft and sulfite mills are most

affected by the Cluster Rule, requiring 100% chlorine dioxide substitution for elemental


1 The US Department of Health and Human Services issues a Toxicological Profile for
Chlorinated Dibenzo-p-Dioxins containing a detailed survey of human exposure and
effects.









chlorine in the bleaching sequence, rendering these mills Elemental Chlorine Free (ECF).

In addition to this source control, spill control of black liquor is required. Beyond these

two requirements, mills develop their own approved plan to meet the new limits,

potentially including voluntary measures such as extended delignification, closed loop

technologies, or Total Chlorine Free (TCF) bleaching. Table 1-1 shows different

pollution prevention strategies adopted by the mills in my study. Ultimately regulating

both media (air and water) at the same time creates a synergistic reduction in pollution.

Once fish living downstream of pulp and paper mills were discovered with

measurable dioxins and furans in their tissues, intensive research into exposure and

effects on fish was initiated. Concerned with biomagnification to humans and

carcinogenicity, regulatory agencies were also interested in potential adverse effects on

aquatic life. As the next section shows, that research led to questions about reproductive

impairment as an effect of pulp and paper mill effluent exposure.

Effects of Pulp and Paper Mill Effluents on Fish

The interaction between industry and government regulatory agencies (primarily in

North America and Scandinavia) produced a large body of knowledge (centered upon

fish) concerning aquatic toxicity of pulp mill effluents. In general, effects have shifted

from gross alterations in growth and acute, lethal toxicity; to more subtle sublethal effects

influencing development, maturation and reproduction. In Canada, regulation passed in

the early 1990s (Environmental Effects Monitoring Program) produced a decade of

consistent fish research at all Canadian mills (McMaster et al. 2003). Also, since 1991,

five international conferences have provided a forum to discuss research on the

environmental impacts of pulp and paper mill effluents, all of which have published









proceedings (Sodergren 1991, Servos et al. 1996, Ruoppa et al. 2000, Stuthridge et al.

2003, Borton et al. 2004).

Many of these studies are field-based, precluding an identification of causative

agents in effluent. Conversely, studies using controlled exposure to whole effluent

dilutions preclude isolation of bioactive effluent components, yet retain environmental

relevance. Several research efforts have addressed controlled exposure to specific

effluent components that are not easily extrapolated to observed effects in the field. Most

recently, efforts to elucidate bioactive compounds have used bioassay-based fractionation

studies. While appealing in theory, van den Huevel (2004a) points out two major

drawbacks of these studies.

* Isolation ofbioactive compounds within mill processes as opposed to in final
effluent.

* Dependence on receptor-binding studies (some of which use human receptors as
opposed to fish receptors) when a receptor-mediated mechanism has not been
firmly established.

Chlorinated organic were thought to be key components causing toxicity.

However, their virtual removal from effluent has not been associated with removal of

chronic, sublethal effects (Lehtinen 2004). Importantly, implementation of the Cluster

Rule has also led to large reductions in nonchlorinated, nonconventional pollutants such

as wood extractives, which could be tied to reduction in effects. Despite the above

caveats, bioassay-based fractionation studies have provided the most specific attempts at

identifying which portions of the nonconventional pollutants may be causing effects. For

example, in association with studies on reduced steroidogenesis, lignin derivatives such

as polyphenolics were identified as bioactive agents in condensates of black liquor









(Hewitt et al. 2002). On the other hand, phytosterols were not determined to be causing

observed effects (Dube and MacLatchy 2001).

Other mechanistic-based studies found ligands for the estrogen receptor and sex

steroid binding protein present in pulp mill effluents, implying an estrogenic cause for

well-documented reproductive effects (Hewitt et al. 2000, Pryce-Hobby et al. 2003). In

support of these findings, known (weakly) estrogenic compounds were recently identified

in effluents such as genistein (Kiparississ et al. 2001) and industrial nonionic surfactants

(nonylphenol ethoxylates) (Lee and Peart 1999). Additional studies showed significant

estrogenic properties of the phytosterol P-sitosterol in fish (MacLatchy and Van der

Kraak 1995, Tremblay and Van der Kraak 1998).

Androgenic properties of pulp and paper mill effluents have also been reported

(Svenson and Allard 2004). In vitro assays using the human androgen receptor showed

effluents from softwood furnish, but not hardwood, produced low levels of androgenic

activity. Biological treatment of effluent had no effect on androgenicity. Uptake of these

androgenic compounds by fish and conjugation in bile was also demonstrated.

However, researchers have had little success in associating androgenic compounds

with masculinization effects. For instance, androstenedione and human androgen

receptor binding was detected in water and sediment samples downstream from one of

the mills in my study and associated with masculinized fish (Parks et al. 2001, Jenkins et

al. 2003). Evidence for androstenedione as a bioactive effluent component was then

refuted by more quantitative analysis showing androstenedione was present only in

fractions that did not induce human androgen receptor activity and expression (Durhan et

al. 2002). Phytosterol degradation analysis of effluents from these same waters did not









reveal androstenedione metabolites but detected androsteneone (Quinn 2004). Separate

studies also refuted androstenedione and testosterone as the active androgens causing

masculinizing effects (Ellis et al. 2003), and showed in vitro fish receptor binding

responses that do not correlate with in vivo effects. Follow up studies with this mill

effluent (van den Huevel et al. 2004b), using more accurate measures of both

masculinization and fish receptor binding, failed to produce any response. Although no

major process changes occurred between studies, treatment system maintenance was

improved as indicated by gradual reduction in total suspended solids.

In reality, the effluent components responsible for causing sublethal toxicity in fish

may never be determined, even though specific mechanisms of action may be narrowed

down. As in other technology sectors, pollution prevention technology in the pulp and

paper industry rapidly progresses. For this industry, the movement is toward a closed

system with recycling and reuse of all materials (Lehtinen 2004). Implementation of

these technologies is the limiting factor, since the industry is capital-intensive; but

voluntary incentive programs (such as the one associated with the Cluster Rule) help

offset investment risks. As the monitored effects diminish and potentially disappear with

improving pollution prevention technologies, identification of the specific bioactive

compounds may not be necessary.

The following literature review of effects in fish exposed to pulp and paper mill

effluent begins with nonreproductive effects, followed by general reproductive effects. It

ends with specific reproductive effects indicating masculinization or femininization of

fish.









Nonreproductive Effects

A variety of sublethal, nonreproductive physiological effects have been reported in

fish exposed to pulp and paper mill effluents. Alteration of liver function (mainly

induction of the detoxifying cytochrome P450 system as measured by ethoxyresorufin-o-

deethylase (EROD) activity) was the most consistently reported nonreproductive effect in

fish. Additional work with conjugating detoxification systems, stress, hematology, and

immunological responses received comparatively scant attention and effects are

conflicting. As a more general measure of health, growth rates were also examined,

although results are equally difficult to interpret.

Significant EROD induction (typically a 2- to 4-fold increase), usually

accompanied by an increase in liver somatic index, is often considered a nonspecific

marker of effluent exposure (Rogers et al. 1989, McMaster et al. 1991, Servos et al. 1992,

Gagne and Blaise 1993, Ahokas et al. 1994, Munkittrick et al. 1994, Gagnon et al. 1995,

Soimasuo et al. 1995, Martel et al. 1996, Martel and Kovacs 1997, Soimasuo et al. 1998,

Sepulveda et al. 2002, van den Huevel et al. 2002). Yet the plethora of compounds

known to induce this detoxification response (combined with the variable nature of

effluent composition within and among mills) makes it difficult to link this biomarker to

specific chemical compounds. Dioxins and furans are strong inducers (Servizi et al.

1993), although EROD induction was also detected when these compounds were not

present (Munkittrick et al. 1992). Unfortunately EROD activity could not be consistently

tied to reproductive effects (Munkittrick et al. 1999).

Induction of conjugating detoxification systems (as well as stress, hematological

changes, and immunological responses) has also been addressed in the literature, albeit

with much less intensity compared to EROD activity. Results are often mixed; and most









of this work has not been linked to adverse effects at higher levels of biological

organization, so the consequences of these physiological changes remain unclear. For

example, oxidative stress has been reported in fish due to induction of hepatic enzymatic

(glutathione peroxidase, glutathione S-transferase, catalase) and nonenzymatic

(glutathione and metallothein) antioxidants, as well as lipid peroxidation in gill and

kidney (Oikari et al. 1988, Stephensen et al. 1998, Ahmad et al. 2000, Fatima et al. 2000).

However, hepatic antioxidants (mainly glutathione S-transferase and glutathione) were

not induced by other studies (Mather-Mihaich and DiGuilio 1991, Bucher et al. 1992,

Larsson et al. 2002). Similarly, effects on activity of another conjugating detoxification

enzyme (uridine diphopsphate glucuronosyltransferase) range from induction to

inhibition (Oikari et al. 1983, Forlin et al. 1985, Lindstrom-Seppa and Oikari 1988,

Andersson et al. 1988b, Lindstrom-Seppa et al. 1989). These mixed results on

conjugating detoxification systems are likely due to differences in effluent quality,

experimental design, exposure conditions and life stage at time of exposure. They are

representative of results for the inter-related responses in stress, immune, and

hematological functions (see Sepulveda 2000 and van den Huevel 2004a for discussion of

these latter parameters).

Likewise, conflicting results exist for growth patterns of fish exposed to pulp and

paper mill effluents. For instance, Warren et al. (1974) and Munkittrick et al. (1991)

detected reductions in growth of fish in laboratory and field collections respectively.

Other field and laboratory studies found no effects of pulp and paper mill effluents on

growth rates in fish (Servizi et al. 1993, Swanson et al. 1992). Additionally, some field

collections documented increased growth at effluent-exposed sites (McLeay and Brown









1974, Sandstrom et al. 1988). Explaining some of these discrepancies, Gagnon et al.

(1995) found accelerated growth characteristic of downstream nutrient loading from both

natural and anthropogenic sources, independent of effluent exposure. Despite rapid

growth, fish collected downstream of the bleached kraft mill did not have concomitant

increases in reproductive effort; rather, they exhibited greater length at maturity,

reduction in gonad size and highly variable fecundity compared to reference fish. Similar

reproductive impairment was detected by Munkittrick et al. (1991) in fish with reduced

growth. Such findings focused researchers toward evaluating reproductive impacts of

pulp mill effluents on fish.

Reproductive Effects

Dominant reproductive effects of pulp and paper mill effluent exposure include

depressed circulating sex steroids associated with alterations in steroidogenic capacity;

reduced gonadal development; delayed sexual maturation; and negative impacts on egg

and fry quality. Effects on egg production and size have been debatable, probably due to

differences in the quality of effluent tested. Although many of these reproductive

parameters have improved with changing effluent technologies, the virtual removal of

chlorinated organic from effluent has not eliminated responses. This finding leads

researchers away from dioxins and furans as causative agents; and toward wood

extractives such as phytosterols, which have been reduced but not eliminated by pollution

prevention technologies.

The most compelling evidence for physiological reproductive alteration comes

from a series of Canadian studies over the past 10 years (summarized by McMaster et al.

2003). Extensive work on white sucker (Catostomus commersoni) and many other fish

species such as lake whitefish (Coregonus clupeaformis) and longnose sucker









(Catostomus catostomus) showed inhibited gonadal development (primarily decreased

gonadosomatic index) and depressed circulating sex steroids (primarily 173-estradiol and

11-ketotestosterone) in both sexes (Munkittrick et al. 1998). However, the steroid

response was not 100% consistent, especially at more recent, modernized mills.

Importantly, steroid effects not observed in the field at a modem mill (Servos et al. 1992)

occurred in laboratory exposures of another species at concentrations higher than

observed in the receiving environment (Robinson 1994). Thus, bioactive compounds

were still being produced, but differential species sensitivity, effluent dilution, and/or

conditions of the receiving environment protected wild fish. This emphasized the

importance of field studies, despite inherent problems with identifying causative effluent

components.

Along with the more persistent reduction in gonadal development of these fishes, a

concerted effort was launched to identify mechanisms behind depression of sex steroids.

Several sites along the pituitary-gonad axis appeared to be affected by exposure to

bleached kraft mill effluents: pituitary function was decreased, ovarian biosynthetic

capacity was reduced, and peripheral steroids metabolism was inhibited (Van der Kraak

et al. 1992, McMaster et al. 1995, 1996). Conflicting evidence was presented by Gagnon

et al. (1994a), who concluded that increased steroid metabolism may reduce steroid

levels. These discrepancies may have occurred from capture and handling stress (Jardine

et al. 1996). Regardless of the controversy over mechanism, improved processing

technologies were followed by partial recovery of reproductive function: steroids and

potential mechanistic responses along the biosynthetic pathway were either less impacted









or no longer significant in both wild and cage-exposed fish, although gonad size was

often still reduced (Munkittrick et al. 1997, van den Huevel et al. 2004b).

Whether these physiological effects translate into effects on the production,

survival, and development of young has been more controversial. Early Scandinavian

studies showed dramatic effects of pulp mill effluent on eggs and fry including reduced

fecundity, smaller egg size, poorer fertilization of eggs, and decreased viability of fry

(Vuorinen and Vuorinen 1985). More recent Scandinavian experiments exposing fish to

wood-derived phytosterols (primarily 3-sitosterol) showed increased egg mortality, larval

deformities, and maternal transfer of phytosterols to offspring (Lehtinen et al. 1999,

Mattson et al. 2001). In contrast, exposure to another phytosterol (stigmastanol)

unequivocally had no effect on egg, larval, or juvenile survival and quality (NCASI

1999). Similar to Canadian reports of reduced steroids at effluent concentrations higher

than receiving stream concentrations, NCASI (1996) documented reduced egg production

in 18-100% v/v effluent from a bleached kraft mill. Unlike the Canadian steroid work,

though, impacts in the field were not addressed to verify a lack of observed laboratory

effects. Initial Canadian reports on sex steroids and gonadal development also indicated

reduced egg size and fecundity (Munkittrick et al. 1991). However, further research

found fecundity to be quite variable (Gagnon et al. 1994b) and fertility of eggs and sperm

not affected at exposed sites (McMaster et al. 1992), despite the well-documented

physiological responses.

Research has been conducted on egg and fry characteristics offish exposed to

effluent from the primary mill investigated in our study, before implementation of Cluster

Rule process changes. The NCASI (2000a) found egg production (but not hatchability)









significantly reduced at 23% v/v effluent, well within instream concentrations (yearly

averages approximately 60%).2 Additionally, exposure to 10% or greater whole effluent

dilutions did not result in effects on fecundity, egg size, or hatchability; but caused

reduction in fry growth and survival (Sepulveda et al. 2003). Parent fish also had

reduced circulating sex steroids and gonad size at 20-40% effluent dilution, similar to

previous findings (Sepulveda et al. 2001) and in support of Canadian research. These

results are perhaps the most convincing link between physiological reproductive impact

and more subtle influences on offspring, although effects in fry could have originated

from maternal transfer of bioactive effluent components instead of (or in addition to)

direct impairment of parental reproductive systems.

Masculinization and Femininization Effects

Studies of masculinization and femininization of fish began with the desire to

control sex ratios in the aquaculture industry (Yamazaki 1983). A number of fish species

have an innate capacity to regulate sex ratios in the population triggered by subtle social

and environmental contexts (Baroiller et al. 1999). On this level, masculinization refers

to complete sex reversal (females to males) at the gonad level; and feminization, vice

versa. Control of sex ratios is accomplished by careful exposure to sex steroids, altering

ratios developmentally or in adult fish (the latter sometimes resulting in sterility)

(Pandian and Sheela 1995). Androgens (mainly the synthetic 17a-methyltestosterone)

have been used to shift sex ratios to males. Natural estrogens (mainly 173-estradiol)

have been used to shift sex ratios to females. However, male-biased sex ratios have also


2Since implementation of process changes, NCASI has repeated its fish full life-cycle exposure and
showed improvement in egg production (DL Borton, pers. comm.).









been induced successfully, using aromatase inhibitors (Jalabert et al. 2000, Kwon et al.

2000). (Aromatase converts androgens to estrogens in many tissues of both mammals

and fish.) Effective doses vary drastically among species, with Poeciliids often requiring

the largest doses compared to salmonids, cyprinids, cichlids and anabantids (Pandian and

Sheela 1995). Environmental pollution has also been linked to unintentional shifts in sex

ratio of fish (Jalabert et al. 2000).

The terms masculinization and femininization have also been used to describe

changes in secondary sex characteristics. In this sense, masculinization refers to external

appearance of male secondary sex characteristics in a female fish (i.e., she retains

ovaries), and feminization vice versa. This phenomenon can occur naturally in the wild:

arrhenoidy, or masculinization of older, reproductively senescent females, has been

documented at low levels in wild Poeciliid populations (Constanz 1989). Changes in

secondary sex characteristics can be induced by administration of sex steroids (Turner

1941a, Turner 1942a, Turner 1942b, Hildemann 1954, Borg 1994). Androgens and

estrogens are assumed to be key players, however progestins may also play a significant

role (Jalabert et al. 2000). Changes in secondary sex characteristics have also been

associated with human impacts on the environment (Jalabert et al. 2000).

As endocrine disruption became a controversial issue in toxicology, the distinction

between these levels of masculinization and femininization was important. Changes in

appearance could behaviorally affect reproduction and population size, or reproduction

could be unaffected. On the other hand, a significant shift in sex ratios could impact

reproduction and population size more overtly. The difficulty with sex ratios is

determining how much of a shift is significant. Hence, for the purpose of my study,









masculinization and femininization will refer to alterations in secondary sex

characteristics only. Shifts in sex ratios (implying alteration at the gonad level) will be

referred to as such. Both of these endpoints have been associated with pulp and paper

mill effluent exposure in fish.

Effects on sex ratio varied by effluent exposure and species. Research on an

indigenous species living near a TCF Swedish kraft mill demonstrated slight yet

statistically significant male-biased sex ratios in embryos (55 to 58% male) compared to

pooled reference sites (Larsson et al. 2000). Further, temporary mill shutdown allowed

recovery of normal sex ratios, and the male bias reappeared after mill processes were

restored (Larsson and Forlin 2002). Short-term (42 day) laboratory exposure of this

effluent to a livebearing species did not reflect field results, failing to induce any change

in sex ratios (Larsson et al. 2002). Full life-cycle exposure to bleached sulfite mill

effluent demonstrated the opposite response in yet another species: sex ratios were

female-biased at 30% effluent and greater (Parrott et al. 2004). In addition, egg

production was reduced at 10% effluent and failed at 30% effluent or greater. Based

upon these findings, egg production but not sex ratios may be impacted in the wild, since

effluent concentrations vary from 1 to 15% by season and river flow. Finally,

multigenerational laboratory exposure to environmentally relevant levels of phytosterols,

primarily P-sitosterol, revealed male-biased sex ratios in the first offspring generation and

female-biased sex ratios in the second (Nakari and Erkomma 2003). Clearly, bias toward

one sex or another cannot be generalized in response to pulp and paper mill effluent

exposure, although changes in sex ratio may be a useful indicator of impacts on fish

reproduction (Parrott et al. 2004).









Changes in secondary sex characteristics of fish exposed to different types of pulp

and paper mill effluent include precocious and delayed maturation, femininization, and

masculinization. Among these alterations, masculinization is the most consistently

reported effect across field and laboratory studies. Precocious maturation (or early

development of secondary sex characteristics) has been reported in fish collected from a

bleached kraft effluent receiving stream investigated in my study (Caruso and Suttkus

1988). Precocious maturation at 32% or greater effluent, masculinization (at 10% or

greater) and femininization (at 32% or greater), were reported in fish exposed to dilutions

of bleached sulfite effluent (Parrott and Wood 2002, Parrott et al. 2003, 2004). As with

sex ratios and egg production reported by this group, the only environmentally significant

response may be masculinization. Using the same model species, NCASI (2000b) found

delayed maturation resulting from exposure to bleached kraft mill effluent.

Masculinization occurred in this species from exposure to a different bleached kraft mill

effluent (Kovacs et al. 1995b) but not to effluent from a thermomechanical pulp mill

(Kovacs et al. 1995a). In the study by Larrson et al. (2002a) on a livebearing species,

although sex ratios were not altered, masculinization was weakly indicated by male-like

coloration. Perhaps the strongest case for masculinization lies with effects on

mosquitofish, initially documented by Howell et al. (1980). This species is elaborated

upon in the next section.

Examining these responses as a whole, in order to be useful biomarkers any of the

suborganism level effects must be linked to exposure, as many studies included; and to

effects on reproductive success, and if possible on populations. Full life cycle tests are

very useful to this end, but these tests must be comparable to responses in wild fish









actually living under exposure conditions. Hence a two-pronged approach pairing field

and laboratory exposures is ideal, especially if the same species can be used for both

types of studies. The mosquitofish has potential for both types of exposures, as shown in

the following section. The remainder of this chapter examines mosquitofish in relation to

pulp and paper mill effluent, as a model species, and as a potential bioindicator of pulp

and paper mill effluents.

Effects of Pulp and Paper Mill Effluents on Mosquitofish

The Eastern mosquitofish, Gambusia holbrooki, was the first species recorded as

masculinized by pulp and paper mill effluent exposure (Howell et al. 1980). Since then,

improved analysis of field collections and laboratory exposure to degraded effluent

components have supported the original observational response. The degree of

masculinization was highly variable within a site and by season, yet considered

comparable among three Florida mills (all of which were examined in my study). In

contrast, controlled exposure to whole effluent dilutions provided mixed evidence for

masculinization in Western mosquitofish, Gambusia affinis (McCarthy et al. 2004).

Beyond the masculinization response, precocious maturation, behavior, and aspects of

reproduction (mainly brood size) were also addressed without significant observable

impacts. So far, attempts to isolate potential mechanisms) of masculinization were

inconclusive.

Masculinization

Sampling of Eastern mosquitofish in Elevenmile Creek, FL, USA revealed the first

known occurrence of masculinization associated with pulp and paper mill discharge

(Howell et al. 1980, p. 676). Lacking quantification of data, the authors reported the

following.









All females within this stream are strongly masculinized, possessing a male-like
gonopodium and displaying male reproductive behavior. All males exhibit
precocious secondary sex characters and reproductive behavior.

Photomicrographs indicated elongation, segmentation, and intermittent terminal

differentiation of female anal fins similar to the male gonopodium (the copulatory organ

used to inseminate females in this livebearing species). Apparently equivalent responses

were detected in another effluent-receiving stream in Florida, the Fenholloway River

(Bortone and Drysdale 1981). After significant process changes at the Elevenmile Creek

mill (including conversion to ECF bleaching and oxygen delignification), quantification

of the response (anal fin length) and statistical comparison to females from a reference

stream showed masculinization remained (Cody and Bortone 1997). However

photographs qualitatively indicated reduced elongation and lack of terminal

differentiation. Season (winter versus summer months, based upon seasonal drought

conditions) also influenced anal fin length significantly: greater elongation occurred in

summer months. As further evidence, Bortone and Cody (1999) detected a statistically

significant increase in the ratio of anal fin length to fish standard length (finally

accounting for the influence of body size on this morphological feature), and inferred a

distance/dose-dependent response downstream of pulp mill effluent discharge in Rice

Creek, FL. In light of seasonal effects reported previously (Cody and Bortone 1997), the

inference was tenuous: they compared one upstream and three downstream sites collected

in summer with a fourth, furthest downstream site collected twice; once in winter, and

once several years previously by separate researchers. Significant increase of anal fin

elongation in females from the first two downstream sites was statistically comparable to

Fenholloway River females collected several years before. Variation was high (highest

in the Fenholloway collection), even at the upstream (200 m above outfall) site. The









authors speculated tidal influence may draw effluent above the discharge point,

accounting for this unexpected result. Exposure to pulp mill effluent, either potential or

actual, was never documented in these collections.

Based upon detailed observations of exposure to androgens by Turner (1941a,

1942a, 1942b) and steroid production studies using bacterially degraded phytosterols by

Marsheck et al. (1972), it was hypothesized that female mosquitofish may be

masculinized when exposed to androgens formed by degradation of phytosterols present

in pulp and paper mill effluents. Subsequently, female mosquitofish were exposed to

high concentrations of phytosterols (approximately 0.1-0.5 g/L of stigmastanol and 3-

sitosterol) combined with a bacterium (Mycobacterium smegmatis) not common to

effluent-receiving streams (Denton et al. 1985, Howell and Denton 1989). Although

presence of androgens was not monitored to verify androgen exposure, females

developed male-like gonopodial structures within two weeks. Stigmastanol produced a

more potent effect. The male-like gonopodial structures did not elongate to the length of

normal male gonopodia, but developed terminal differentiations. Lacking quantification

and statistics, nonetheless this mechanism was proposed to explain observed effects of

masculinization. In support of these findings, Angus et al. (2001) detected rapid onset of

anal fin elongation: 14 days at 60 [g 11-ketotestosterone/g food, development of terminal

differentiations by 20 days in the high exposure groups (80 and 100 [g/g), and average of

40 days in the low exposure groups.

Analysis of various morphological endpoints revealed the most sensitive measures

of masculinization. Bortone et al. (1989) determined an unranked suite of about ten

morphological measures of body and fin sizes-including only one measure of the anal









fin-statistically differentiated effluent-exposed fish from reference fish. These variables

were proposed as a rapid bioassay to detect effects of pulp mill exposure, but were never

fully developed. Howell and Denton (1989) developed five stages of increasing

gonopodial development in females exposed to bacterially-degraded phytosterols. More

recently, quantitative measures of anal fin morphology were compared from wild-caught

and androgen-exposed females (Angus et al. 2001, Bradley et al. 2004). The length ratio

of ray 4 to 6 was the most sensitive measure of masculinization of the anal fin. While the

number of segments along rays 3 and 4 and the width ratio of ray 3 to 4 were also

sensitive measures, they were more subject to variability.

Controlled exposure to whole effluent dilutions produced inconsistent

masculinization results. Initially in support of field collections, static renewal exposure

of newborn mosquitofish to water collected 3.6 km downstream from Elevenmile Creek

induced elongated anal fins (measured as anal fin length) in females upon maturity

(Drysdale and Bortone 1989). While my study research was being conducted,

researchers in Canada and New Zealand were also studying masculinization of adult

female mosquitofish using controlled (mainly static renewal, one flow-through)

exposures to 15%, 70% or 100% effluent (McCarthy et al. 2004 summarizes results

across separate studies). Out of seven pulp mills and one sewage treatment facility

tested, four of the pulp mill effluents and the sewage effluent3 induced masculinization

(all static renewal exposures). Among these four pulp mill effluents, two induced

masculinization relatively quickly (within 3 weeks) while the other two required 24

weeks of exposure. No association between induction and type of mill or concentration


3 In contrast, collection near other sewage discharges found effect on males, not females (Batty and Lim
1999, Angus et al. 2002). Bradley et al. (I 2 14) also found no effect on females









of P-sitosterol could be established. Apparently, duration of exposure required to

produce effect varies widely. However, three important caveats exist for these studies.

* Other than Elevenmile Creek, comparative field studies were not conducted to
determine if effects existed in wild mosquitofish exposed to these effluents.

* All but one exposure required holding and transport of effluent back to the
exposure system, with unknown consequences to effluent composition.

* Masculinization was measured qualitatively, as presence/absence or staged using
categories established by Howell and Denton (1989).

One of these controlled exposures detected differences in masculinization due to

effluent treatment and filtration (Ellis et al. 2003). Secondary treatment of effluent at

environmentally relevant concentration reduced gonopodial development by 25%, yet

masculinization remained significantly greater than controls. Filtration of treated

effluent, removing organic extractives adsorbed to particulates, also removed the

response. Exposure was repeated two years later, after treatment system maintenance

was improved as indicated by gradual reduction in total suspended solids. Using the

more specific ray 4 to 6 length ratio, masculinization was not induced (van den Huevel et

al. 2004b). Since exposure duration remained the same (3 weeks), it is unknown if the

effect was entirely removed or if time to manifestation was extended. Regardless, these

experiments strongly correlate masculinization with adsorbable organic effluent

components, such as low molecular weight wood extractives.

Precocious maturation

Precocious (early) maturation of male mosquitofish exposed to pulp and paper mill

effluent has been examined briefly. Effects were associated with bleached kraft effluent

from Elevenmile Creek before major process changes (Howell et al. 1980, Drysdale and

Bortone 1989), but not with effluent from a thermomechanical/kraft/newsprint mill in









Ontario (McCarthy et al. 2004). Males of this species grow steadily until maturation at

which point growth plateaus, as opposed to females who grow steadily throughout life

allowing body length to roughly approximate age (Snelson 1989). Howell et al. (1980)

reported small males (12-13 mm standard length) began to develop gonopodia, and fully-

differentiated gonopodia occurred in males measuring 13-18 mm standard length.

Compared to males in unexposed sites with mature gonopodia at 18 mm or longer, they

concluded males developed earlier due to effluent exposure. However, neither data nor

statistics was provided to support this conclusion. Static renewal exposure of newborn

mosquitofish to water collected 3.6 km downstream from Elevenmile Creek provided

more compelling, statistical evidence of precocious maturation in males (Drysdale and

Bortone 1989). Exposed males began anal fin elongation approximately one month

before unexposed males, with groups evening out during late-stage gonopodial growth.

Surprisingly, this endpoint was virtually ignored by researchers until recently. In

contrast, in the Ontario study, continuous flow-through exposure of mosquitofish to 15%

and 100% effluent for 21 weeks did not affect male gonopodial length relative to body

size (McCarthy et al. 2004).

Behavior

Investigation of male-like reproductive behavior was a logical step once male-like

secondary sex characteristics were discovered in female mosquitofish. As implied

previously, changes in secondary sex characteristics could potentially lead to behavioral

changes that keep females from copulating and reproducing. Initial preliminary

investigation, lacking statistical comparison, indicated both masculinized females and

precocious males displayed more aggressive reproductive behavior (Howell et al. 1980).

Behavioral evaluation of masculinized females in the presence of normal females









supported increased aggressive, but not reproductive, behavior (Ellis et al. 2003).

Analyzing the ability of a suite of reproductive behavioral changes to detect pulp mill

effluent exposure, Bortone et al. (1989) determined behavior did not adequately

discriminate effluent exposure from unexposed groups. In support of this conclusion,

Krotzer (1990) performed a thorough reproductive behavioral analysis of mosquitofish

females exposed to bacterially-degraded phytosterols. Masculinized females displayed

aggressive male behaviors toward nonmasculinized females only in a noncopulatory

fashion. Paired with males or other masculinized females, they behaved normally. Thus,

masculinization likely does not impact mosquitofish populations in a behavioral sense.

Reproduction

Potential for reproduction does not appear impacted by pulp and paper mill effluent

exposure in female mosquitofish. Except for histological evaluation of gonads,

reproductive parameters were never directly compared with measures of masculinization.

True to the distinction between masculinization and actual sex reversal, normal ovaries

lacking any testicular tissue were consistently reported in masculinized females (Howell

et al. 1980, Hunsinger et al. 1988, Ellis et al. 2003, McCarthy et al. 2004). Fecundity

(inferred by brood size or number of eyed embryos plus mature eggs) relative to body

length was depressed in females collected below effluent discharge in Elevenmile Creek

(Rosa-Molinar and Williams 1984, p. 122). However, the authors state:

"Estimated fecundities in reference to length in the arrhenoid masculinizedd] fishes
were not found to be similar to those found in other studies... although the
fecundity of the normal G. a. holbrooki [in the current study] was similar."

In contrast, static renewal exposure to sediments and waters of bleached and unbleached

kraft effluent receiving streams for 56 days produced no statistical differences in several

measures of fecundity compared to reference sites (Felder et al. 1998, D'Surney et al.









2000). However, variation was high between reference sites. One reference site was a

research station, a well-documented habitat without pollution impacts but obviously very

different from the receiving stream (e.g. very low water hardness contributing to

increased skeletal abnormalities). Thus the selection of unexposed sites is important, and

should include an upstream site at the minimum and sites belonging to the same

watershed. In support of this overall lack of histological and embryological impairment

in effluent-exposed females, McCarthy et al. (2004) did not detect alteration in sex ratios

of mosquitofish reared in 100% effluent under laboratory conditions. The assumption

must be made for all above data on fecundity and sex ratios that at least a portion of

exposed females analyzed were masculinized as well.

Mechanism of Action

Since mosquitofish research has produced the only specific hypothesis of causation

linking one class of effluent components to potentially adverse effects, several

researchers have attempted to isolate potential mechanisms) of action. In support of the

bacterial degradation hypothesis forming androgens, Jenkins et al (2001) detected low

levels of androstenedione (0.14 nM) in the Fenholloway River downstream of effluent

discharge. However bioassay-based fractionation studies, discussed previously under

effects of pulp mill effluent on fish, have not supported androstenedione and testosterone

as active androgens that bind the androgen receptor and masculinize females (Jenkins et

al. 2001and2003, Parks et al. 2001, Durhan et al. 2002, Ellis et al. 2003, van den Huevel

et al. 2004b). Orlando et al. (2002) investigated an alternative mechanism used in the

aquaculture industry, aromatase inhibition. Contrary to their hypothesis, aromatase

activity was elevated in both brain and ovarian tissue of females collected downstream of

effluent discharge in the Fenholloway. Masculinization was indicated by increased









segmentation of anal fin rays, but not increased anal fin length. The authors concluded

aromatase inhibition was not a likely mechanism to account for masculinization.

However, impaired activity could potentially upregulate enzyme production.

Measurement of endogenous steroid levels may provide more insight into this potential

mechanism.

Mosquitofish as a Model Species

Several life history and ecological characteristics of mosquitofish (the closely

related Eastern and Western species (Gambusia holbrooki and G. affinis) of the family

Poeciliidae), make these species an ideal model for both field and laboratory

toxicological studies. Meffe and Snelson (1989) compiled the most comprehensive

overview of Poeciliids, from which the following summary was derived unless noted

otherwise.

Occurrence and Availability in Effluent-Receiving Systems

Mosquitofish are opportunistic, omnivorous feeders that can exploit diverse foods

ranging from planktonic invertebrates and fish fry to detritus and algae; so food source

should not limit their occurrence in effluent-receiving streams. Similarly, mosquitofish

inhabit a diverse range of shallow habitats with the ability to occupy "fringe" habitats

characterized by environmental extremes. Combined with their tolerance to high salinity

(up to 50% seawater in the Western mosquitofish, G. affinis), broad thermal range, and

tolerance to low dissolved oxygen (mosquitofish gulp air at the surface in response to

hypoxia), mosquitofish should tolerate water quality of effluent-receiving streams. Their

home range is small (several meters) making chronic exposure likely. In addition, they

readily colonize new populations via migration of a single gravid female, and have









become ubiquitous around the world caused by deliberate introductions in attempt to

control mosquitoes and associated mosquito-borne illnesses.

In addition to their suitability as a field model, mosquitofish can be maintained

under laboratory conditions with ease relative to other livebearing fish species. Also

compared to other livebearers, much more is known about mosquitofish reproduction.

Reproductive Characteristics

As members of the livebearing fish family, Poeciliidae, mosquitofish develop eggs

internally and appear to give birth to fry (ovoviviparity). This is in stark contrast to most

fish that lay eggs (oviparity). Males inseminate females with packets of sperm called

spermatozeugmata, and females can store sperm in ovarian folds and gonoduct for up to

eight months/broods. Fertilization and embryological development occur directly in the

ovarian follicle, and ovulation is immediately proceeded by parturition. Estrogen (as

opposed to prostaglandins in other fish species) stimulates postovulatory sexual

receptivity of females.

Livebearers exhibit a spectrum of maternal-embryo nutrient exchange, from more

fish or reptilian like yolk loading before fertilization (lecithotrophy) to more mammalian

like continuous provisioning throughout embryological development by the mother

(matrotrophy). Mosquitofish represent the former group. In addition, mosquitofish carry

one fertilized brood at a time, as opposed to many other livebearers that harbor several

broods at different stages of development (superfetation). Brood size is dependent on

female body size, with larger fish producing larger broods. Reproduction is

asynchronous and seasonal in temperate to subtropical climates such as Florida. The

reproductively active period is during spring and summer months followed by

reproductive senescence in the fall and winter. Temperature and photoperiod are









considered dominant environmental cues controlling reproductive season in both sexes

(Koya and Kamiya 2000, Koya and Iwase 2004).

Mosquitofish begin life as hermaphrodites, containing both ovarian and testicular

tissue (Koya et al. 2003). Hermaphroditisim (Teh et al. 2000) has also been detected in

adult farm-reared albino mosquitofish. Within 10 days, gonads differentiate into paired,

fused testes or ovaries. Gonadal maturity is reached in approximately 3 months, with

males maturing 2-3 weeks earlier than females.

Mosquitofish, similar to many other livebearers, are sexually dimorphic. Male and

female mosquitofish have several gender-specific traits: females are larger and possess an

anal/gravid spot and urogenital papilla, while males are smaller and possess a

gonopodium. The gonopodium facilitates internal fertilization and is formed by the

elongation of rays 3, 4, and 5 of the anal fin. Formation of the gonopodium is controlled

by androgens, and a fully-developed gonopodium (marked by terminal differentiations on

the tips of rays 3, 4, and 5) signifies complete maturity (Turner 1941b). Turner (1941a,

1942a, 1942b) also documented formation of the fully mature gonopodium in females

exposed to androgens such as ethynyl and methyl testosterone. Thus, female

mosquitofish could be a useful model of exposure to environmental androgens.

Mosquitofish as a Bioindicator of Pulp and Paper Mill Effluent

Mosquitofish have been repeatedly proposed as an indicator of environmental

disturbance by pulp and paper mill effluents (Davis and Bortone 1992, Bortone and Davis

1994, Cody and Bortone 1997). At the state level, the Florida Department of

Environmental Protection has explored this possibility (T.S. Gross pers. comm.) without

implementation. Federally, the US EPA is developing the mosquitofish as an androgenic

model of endocrine disruption (Angus et al. 1997). While mosquitofish have potential









for use in regulatory testing and screening of pulp mill effluents, they have not been

adequately assessed as an indicator species.

Definitions: Bioindicator and Biomarker

For every critical review of bioindicators and biomarkers there exists a slightly

different definition. The term biomarker is more consistently defined as a measurable

biological response to environmental pollution observed below the organism level of

biological organization (Foster et al. 1992, Peakall 1992, Jamil 2001). Biomarkers

encompass changes at the molecular, biochemical, physiological, histological,

morphological, or behavioral levels. The major premise for use of biomarkers in

environmental regulation is the bridge they form between chemical exposure and adverse

effect. However, biomarkers are usually classified as more indicative of either exposure

or effect. Current challenges for biomarker use include questions of natural variability,

use in the field, and extrapolation to higher levels of biological organization and to

humans. Though the challenges appear formidable, researchers strive to meet these

important demands of regulatory application (e.g. analysis of suites of biomarkers and

increasing inclusion of biomarker analyses in population and community studies).

Unfortunately, as a society we have little patience for the pace of science and often

biomarkers are misjudged or overinterpreted.

In contrast to biomarkers, the term bioindicator usually implies changes at the

organism level or above, including individual reproduction, populations and

communities. For example, US EPA (2004, website) states the following:

Environmental scientists have determined that the presence, condition, and
numbers of the types of fish, insects, algae, and plants can provide accurate
information about the health of a specific river, stream, lake, wetland, or estuary.
These types of plants and animals are called biological indicators.









Ecological bioindicators have been developed and used extensively by the US EPA to

assess ecosystem health. Jamil (2001, p. 4) defines bioindicators for evaluating

ecological health "as a species or groups of species (plants or animals) that, by their

presence and/or abundance, play an important role in the ecosystem to which they

belong." Further, he distinguishes two classes of bioindicators used to evaluate

environmental quality: bioaccumulator and sentinel species. Similar to separation of

biomarkers into exposure and effect groups, bioaccumulators represent organisms that

bioconcentrate toxicants from the surrounding media and biomagnify up the food chain,

while sentinels indicate toxic effect that allows judgment of effects on human and/or

environmental health. Sentinel species are further characterized as species with field

application, either preexisting at sites of interest or capable of in situ exposure (such as

caging onsite). Ideally, extensive knowledge exists about normal states measured in

sentinel species. A final condition is the surrogate nature for species at risk, i.e., sentinels

substitute for sampling of already imperiled species that are difficult to study directly and

could potentially be harmed by intensive research.

For the purposes of my study, bioindicator is defined as a species possessing

measurable changes in biomarkers that correspond to impacts at higher levels of

biological organization. Similar to applicability requirements for sentinel species, a

bioindicator should be a versatile subject in the field, not only in the laboratory. Thus the

bioindicator retains environmental relevance while affording analysis of time and dose-

dependent responses.









Bioindicator Criteria for Success

Criteria, like definitions, abound for rating success or failure of potential

bioindicator organisms (Peakall 1992, Jamil 2001). Regardless of variable definitions,

several common criteria for success exist and overlap with criteria for biomarkers.

Practicality

* Sufficient scientific knowledge of model species under normal, unexposed
conditions.

* Ease of training and use by personnel.

* Cost- and labor-effective.

* Availability of model species for both field and laboratory study.

Variability

* Intrinsic or natural variability of biomarkers such as seasonality and gender
differences.

* Exposure variability especially sensitivity and tolerance/acclimatization.

* Method variability such as observer bias and instrumentation bias.

Predictability

* Extrapolation to organism level or higher adverse effects, i.e., reproductive or
population impacts.

* Extrapolation to other species living in exposure conditions.

* Extrapolation to humans.

Applying these criteria to the existing research on mosquitofish exposed to pulp

and paper mill effluents reveals practicality, but not variability and predictability, has

been adequately addressed. Practicality was supported in the previous section about

mosquitofish as a model species. The mosquitofish is one of the most intensely studied

livebearing species because of its use in mosquito control. Masculinization studies are

not expensive, especially compared to molecular and biochemical research.









Measurements require basic lab skills using dissecting scopes and ideally computer

measurement software. The most expensive aspect for these studies is exposure facility

construction, a common expense for any bioindicator. Similarly, exposures are the most

labor-intensive aspect of masculinization studies. Perhaps the greatest strength for

mosquitofish as a bioindicator is its global distribution (again because of mosquito

control), making collection of large numbers in effluent-exposed and reference sites very

easy.

In terms of variability, natural fluctuation has been indicated by season (in Florida)

and the response is gender specific, although precocious maturation remains questionable

in Florida streams. Exposure variability has been implied by the work in Canada and

New Zealand, among mills and within mills with improving technologies and

maintenance of systems. Research in Florida has not directly addressed this type of

variability, but a consistent response to variable exposure is implicated. Since specific

bioactive agents and mechanism of action have not been isolated, sensitivity is difficult to

address and can only be viewed from whole effluent exposures. Compared to binding of

fish androgen receptor (Ellis et al. 2003), masculinization is a less sensitive, but

potentially more relevant response. Specificity of masculinization for pulp and paper

mill effluent versus sewage effluent appears high, with the abnormal exception to treated

sewage effluent by McCarthy et al. (2004). Yet laboratory exposures to other chemicals

(such as the hypertensive drug spironolactone and the agricultural insecticide endosulfan)

have also induced masculinization (Howell et al. 1994, Park et. al 2004). In addition,

Bradley et al. (2004) found masculinized females living in the retention pond of an urban

parking lot; therefore, nonpoint sources of pollution may confound results. So









specificity, once considered high for pulp mill effluents, has become questionable. At the

same time, exploration of how these alternative compounds masculinize may catalyze

isolation of bioactive compounds. Tolerance and acclimatization have not been studied.

A handful of anecdotal reports (Davis and Bortone 1992, Bortone and Davis 1994, Cody

and Bortone 1997) indicate masculinization is reversible when females are transferred to

clean water; McCarthy et al. (2004) did not observe resorption of anal fin elongation

under controlled exposure. Finally, method variability, other than determination of the

more sensitive measures of anal fin morphology, has not been addressed.

The final major criterion for a successful bioindicator, predictability, has begun to

be examined with studies of reproductive potential. As previously discussed, gonad

condition, fecundity, and sex ratios do not appear affected by effluent exposure. A major

drawback to most of this work is the lack of masculinization measures. Actual fry

production, quality and survival have not been examined either as a more accurate

measure of reproductive success. Regarding extrapolation to the fish community,

Bortone and Cody (1999) attempted to examine masculinization of other livebearing fish

species in their Rice Creek collection, but failed to obtain adequate fish numbers for

analysis. The link to humans has yet to be determined as well.

Obviously, further testing of mosquitofish is required to determine if this species

would be a successful bioindicator of pulp and paper mill effluents. Practically speaking,

mosquitofish are very promising, especially on a worldwide scale but perhaps less useful

in nations with advanced processing technology.

Contribution of My Study

My study addresses two of the three major bioindicator criteria for mosquitofish:

variability and predictability. Industry and regulatory agencies alike will thus have a









better understanding of the potential for mosquitofish as a bioindicator of pulp and paper

mill effluent exposure. In addition to anal fin morphology as a biomarker, sex steroids

were investigated for two reasons: 1) masculinization as an androgenic model innately

assumes alteration of steroid levels, either peripherally or systemically; 2) sex steroid

levels are generally depressed in other fish species exposed to pulp and paper mill

effluents. Variability by season, method, and exposure were addressed directly. Method

and seasonal variability, while not a stated objective of the original research proposal,

were necessary precursors in the development of techniques and thus were included in

my study results. Exposure variability studies focused upon the three Florida freshwater

systems (Elevenmile Creek, Fenholloway River, and Rice Creek) for which

masculinization has been reported and considered equivalent. Predictability was

addressed by evaluating the relationship among biomarkers and reproductive success.

Preliminary examination of population structure was also conducted during the

reproductive experiments.

Explicitly stated objectives for my study were divided into two specific aims with

associated hypotheses. Within each specific aim, three sub-aims were identified and

studies developed for each.

Specific Aim 1

Our first specific aim was to determine the effects of improved mill technology on

masculinization of female mosquitofish. We hypothesized that reduction in brown side

effluent components (i.e., wood extractives such as phytosterols and resin acids) would

reduce anal fin elongation and hormonal alteration in female mosquitofish.

* Aim 1A: Assess induction of masculinization in female mosquitofish under short-
term controlled exposure to effluent at one mill throughout process changes and at
two mills using different processing techniques. Expected outcomes were:









induction would be rapid; degree of response would reflect differences in
concentration of wood extractives; and effect would be reduced following major
process changes.

* Aim 1B: Compare and contrast anal fin morphology and sex hormone
concentrations in female mosquitofish at one mill throughout process changes. The
expected outcome was that process changes would reduce masculinization in wild
fish to a similar degree as induction studies.

* Aim 1C: Compare and contrast anal fin morphology and sex hormones in female
mosquitofish collected from systems exposed to different types of mill effluent.
The expected outcome was degree of response would reflect differences in
concentration of wood extractives, similar to differences expected with induction
studies.

Specific Aim 2

Our second specific aim was to evaluate reproductive success of mosquitofish

exposed to pulp and paper mill effluents. We hypothesized that exposure to pulp and

paper mill effluents would not impair reproductive success of mosquitofish.

* Aim 2A: Determine population structure of mosquitofish living in Fenholloway
River. The expected outcome was that population structures would reflect sex
ratios and recruitment (juveniles) found in unexposed references.

* Aim 2B: Characterize organism level responses to effluent exposure associated
with reproduction. The expected outcome was that adult females would display
normal reproductive status regardless of exposure and masculinization.

* Aim 2C: Quantify offspring production of female mosquitofish exposed to pulp and
paper mill effluents. The expected outcome was that offspring production would
not vary regardless of exposure or masculinization.

Studies to address these aims were designed with certain limitations and

assumptions in mind.

* Home ranges do not overlap between sites sampled.

* However, the above implies field studies may be comparing genetically different
populations of mosquitofish.

* Mosquitofish caught at field sites have lived there throughout life (birth, pre- and
post-maturation).










* Mosquitofish (G. affinis and holbrooki) respond in the same manner.

* Masculinization of the anal fin is finite and permanent (as indicated by Turner
1942b), i.e., gonopodial development does not vary between acute and lifetime
exposures.

* Fry production studies use wild-caught fish therefore reproductive history (i.e.,
fertilization, past broods) is not controlled and is unknown.

* Cannot expose to whole effluent by transfer back to the laboratory without
potentially altering its composition.

* Composition of effluent changes drastically when mills cycle through different
types of trees.

* Potential (as opposed to actual) exposure is sufficient to document toxicant
exposure, since the masculinization hypothesis predicts actual exposure is to as yet
unknown androgens and not parent effluent components.

* Specific changes in mill technology are not always known or available to
investigators.

Table 1-1. Select characteristics of the mills in my study according to receiving stream.
Mill Characteristic Elevenmile Creek Fenholloway River Rice Creek
Furnish 75% hardwood, 100% softwood 50% hardwood,
25% softwood 50% softwood
Pulping Chemical/kraft Chemical/ dissolving kraft Chemical/kraft
Bleaching ECFa (1995) Sodium hypochloriteb ECFa (May 2001)
Product White copy paper, High grain cellulose Paper towels, tissue paper,
return postcards, (dissolving nonpaper pulp) kraft bag, linerboard
market (paper) pulp
Effluent treatment Aeration with microbial Aeration with microbial Aeration with microbial
degradation, chemical degradation degradation, activated
flocculation, oxygen sludge
delignification
Effluent volume 21-26 mgd" 43 mgd 28 mgd
aElemental chlorine free; conversion date in parentheses
bNote this mill was not subject to many Cluster Rule requirements, since it is not a papergrade bleached
kraft or sulfite mill
"million gallons per day











FH market pulp EM integrated RC nonintegrated
facilities facilities facilities


FH market pulp EM 0 RC
V W paper/paperboard
FH nonpaper pulp paper pulp EM m EM


fluff /writing/ tissues P "' orrugate
copy aPPIn
FH\ \EM / RC /RC RC RC
paper
dissolving specialty \/ /newsprint \/ / towels inerboar


Figure 1-1. Categories of pulp and paper mill facilities. Mills associated with my study
are abbreviated by receiving stream: EM= Elevenmile Creek, FH =
Fenholloway River, RC = Rice Creek.














CHAPTER 2
VALIDATION OF MOSQUITOFISH ENDPOINTS USED TO ASSESS EFFECTS OF
PULP AND PAPER MILL EFFLUENT EXPOSURE

Mosquitofish have been proposed as a bioindicator of pulp and paper mill effluents,

although several aspects concerning variability in masculinization responses have not

been addressed. Eastern mosquitofish were collected in summer and winter months from

Rice Creek, the receiving stream for effluent from the Georgia-Pacific

bleached/unbleached kraft mill in Palatka, FL. In this study, a series of validations were

performed for morphological measurements and for a new biomarker in this species,

whole body sex steroid concentrations. Seasonality in responses was also addressed.

Gender identification using the urogenital papilla was successfully validated against

internal examination of gonads. Morphological measurements validated adequately, and

computer-aided measurement was a preferred alternative to manual measurement. Sex

steroids also validated adequately considering the unavoidable limitations using whole

body analysis. Greater masculinization was associated with effluent-exposed sites for

females, while males were not overtly affected. Females displayed seasonal effects on

sex steroids, especially steroid ratios, but not anal fin morphology. Surprisingly, sex

steroids were not altered in females collected from 100% final effluent before discharge,

indicating a complex interplay of environmental factors) may produce responses as

opposed to effluent alone, such as differential bacterial degradation of phytosterols.









Introduction

Concerns over release of chemicals into the environment have shifted from lethal to

sublethal effects on nontarget wildlife species (Lehtinen 2004). Reported sublethal

effects of pulp and paper mill effluents on fish include induction of liver detoxification

systems, alterations in sex steroids concentrations and production/metabolism, reduced

gonadal development, decreased egg production and decreased fry survival (Van Der

Kraak et al. 1992, Gagnon et al. 1994a, Munkittrick et al. 1999, NCASI 2000a,

Sepulveda et al. 2003, Parrott et al. 2004, McMaster et al. 2003). Degree of these effects

is often mill-specific with some effluents producing no effect at all (Kovacs et al. 1995a,

McCarthy et al. 2004). No clear pattern existed between effect and type or quality of

effluent, other than reduced responses with improved mill technologies (Munkittrick et

al. 1997, van den Huevel et al. 2004b). Further complicating the matter, bioactive

effluent components have not been strictly identified yet and most likely different

compounds and conditions influence response pathways.

Several responses imply androgen-induced mechanisms in fish. For example, pulp

mill effluents have been associated with male-biased sex ratios and development of

male-like secondary sex characteristics in females or masculinization (Kovacs et al.

1995b, Larsson et al. 2000, Larsson and Forlin 2002, Larsson et al. 2002, Parrott and

Wood 2002, Parrott et al 2003). The masculinization response has been frequently

reported in female mosquitofish as an elongation of the anal fin into a male-like

gonopodium, the copulatory organ in this livebearing species (Howell et al. 1980, Cody

and Bortone 1997, Bortone and Cody 1999, Parks et al. 2001, Ellis et al. 2003, Bradley et

al. 2004, McCarthy et al. 2004). Further, this species has been repeatedly proposed as a

bioindicator of pulp and paper mill effluents (Davis and Bortone 1992, Bortone and









Davis 1994, Cody and Bortone 1997). In a practical sense mosquitofish are a suitable

model, yet several aspects require more thorough evaluation to warrant this broad

application as a bioindicator of pulp and paper mill effluents.

In collaboration with the National Council for Air and Stream Improvement

(NCASI), Southeastern Aquatic Biology Program, New Bern, NC, techniques to assess

masculinization were validated and refined. Method validation involved a new gender

identification technique, anal fin morphological measurements, and development of a

novel assay for measurement of whole body sex steroids in mosquitofish. NCASI's

validation results have been reported in the proceedings of the 5th International

Conference on Environmental Fate and Effects of Pulp and Paper Mill Effluents (Bradley

et al. 2004).

Materials and Methods

Mill Characteristics and Field Collection

Adult mosquitofish were collected along shallow vegetated banks in Rice Creek,

the receiving stream for effluent discharge from Georgia-Pacific's Palatka, FL, USA

operation (Figure 2-1). Collections occurred in both summer (March and June 2000, July

2001: n = 174, n = 141, n = 368 respectively) and winter (November 2000, n = 899),

corresponding to reproductive and nonreproductive periods respectively. All fish were

measured for standard length (+ 0.01 mm) using digital calipers and weighed (+ 0.001 g)

using a digital scale before preservation. Chapter 3 gives details about the mill, and field

collection techniques (site descriptions and latitude/longitude can be found in Appendix

A).









Gender Identification Using the Urogenital Papilla

For all collections, gender was determined by external examination of the

urogenital sinus. Most masculinization studies have used either the gonopodium

(typically male-specific) and/or anal spot (female-specific) as external indicators of sex.

Gender identification should not be based upon the gonopodium, since it is the primary

measure of masculinization. This laboratory found the anal spot a difficult indicator as

well, since many female fish have a partial anal spot and some females, mostly collected

from effluent-exposed sites, have very light, brownish anal spots.4 Hence, a new gender

identification technique other than examination of gonads, which is very labor-intensive,

was developed. Females have a gender-specific urogenital papilla that protrudes from the

urogenital sinus, and the urogenital opening is located on the tip of the papilla (Meffe and

Snelson 1989). During copulation, males hold on to the papilla using terminal

differentiations of the gonopodium.

Gender identification via the urogenital papilla was validated using the winter 2000

and summer 2001 collections on several levels. First, variation among personnel was

evaluated by having three technicians determine sex using this new technique (n = 200

fish). Second, the new method was verified against gross internal examination of gonads

by NCASI (n = 200 fish). Third, the new method was verified against histological

identification of gonads within USGS using a separate group of fish (n = 354).

Anal Fin Morphology

Several measures of anal fin elongation have been employed by researchers, from

qualitative scores of presence/absence and categorizing degree of gonopodial


4 Investigation did not reveal any dead embryos or abnormal ovaries within the body cavity, as initially
suspected (pers. obs.). Cause(s) of brown anal spot remains unclear.









development to quantitative measures such as total anal fin length,

gonopodium/extension length, number of terminal differentiations, length ratio of Rays 4

and 6, thickness ratio of Rays 3 and 4, and number of segments on Ray 3. Measurement

of anal fin and gonopodium/extension length has been shown to be dependent on body

size (standard length) and thus must be accounted for during statistical analysis (Bortone

and Cody 1999, Bradley et al. 2004). For this reason, several studies restricted analysis

of anal fin morphology to specific size classes. However, length ratio of Rays 4 and 6

and width ratio of Rays 3 and 4 are independent of body size and such a restriction is not

necessary (Angus et al. 2001). Ray 3 segment counts, surprisingly, did not correlate with

either of these ratios and was not suggested as an accurate measure of masculinization.

Bradley et al. (2004) distinguished length ratio of Rays 4 and 6 as more sensitive than

either width ratio of Rays 3 and 4 or segment number of Ray 3.

Length ratio of Rays 4 to 6 was used to assess masculinization in our study. Body

weight (+ 0.001 g) and standard length (+ 0.01 mm) were also measured before

preservation using a digital scale and a pair of digital calipers for all fish. For the winter

2000 collection, potential variation in both sexes was addressed due to preservation state

(fresh versus fixed) and observer bias within and among laboratories (n = 200 fish for

each comparison). Manual measurements of the linear distance from base to tip of Rays

4 and 6 of the anal fin (+ 0.1 mm) were made using a dissecting scope with ocular

micrometer, before and after preservation in 10% neutral-buffered formalin. Preserved

fish were additionally measured independently by two other technicians using the same

equipment, and then shipped to NCASI for measurement. Influence of size class for

females was also investigated in this collection, dividing females into four 5 mm groups









(20-24.99 mm, 25-29.99 mm, 30-34.99 mm, 35-39.99 mm). The summer and winter

2000 collections were compared to address seasonal variation for both sexes (summer

2000 fish were measured after preservation). Finally, the summer 2001 collection was

used to test observer bias within the laboratory (n = 200 fish) using newly developed

computer-aided measurements (+ 0.01 mm, SigmaScan Pro 5.0, SPSS, Inc.) of digital

images taken of fish before preservation. Bradley et al. (2004) evaluated the use of

computer-aided versus manual measurements and determined although they were

comparable, computer-aided measurements were more accurate and useful for archiving

data.

Sex Steroids

In addition to anal fin morphology as a biomarker of masculinization, sex steroids

were investigated as a second, physiological biomarker. Circulating sex steroids are

generally depressed in other fish species exposed to pulp and paper mill effluents

(McMaster et al. 2003, Sepulveda et al. 2001). Masculinization as an androgenic model

innately assumes more specific alteration of steroid levels, either peripherally or

systemically in favor of androgens. Orlando et al. (2002) hypothesized inhibition of

aromatase resulted in a masculinized hormone profile. Although they tested the first half

of this proposed mechanism and found aromatase activity actually elevated, the second

half, sex steroid profile, was not measured. Examination of sex steroids in mosquitofish

would potentially shed light on hypothesized mechanisms.

Primary sex steroids were analyzed using a modified radioimmunoassay (RIA)

method originally developed for serum and plasma samples of common carp, Cyprinus

carpio (Goodbred et al 1997), and since adapted for use in a variety of other aquatic

species and tissue media such as plasma of largemouth bass, Micropterus salmoides









(Gross et al. 2001) and mantle of freshwater invertebrates (Gross et al. 2000).

Mosquitofish were initially analyzed for 173-estradiol, 11-ketotestosterone and

testosterone, considered the most active reproductive hormones in fish. However, Borg

(1994) demonstrated testosterone, and not 11-ketostosterone, as the only dominant

androgenic hormone in poeciliid fishes. Analysis of 459 fish (both sexes) revealed

nondetectable levels of 11-ketotestosterone, confirming Borg's finding. Therefore,

testosterone is assumed to be the dominant androgen in mosquitofish and analysis of

11-ketotestosterone was discontinued. Ten fish of each sex from each site were analyzed

for sex steroids in the summer 2000 collection; 42-50 females and 16-28 males per site

were analyzed in the winter 2000 collection.

RIA steroid analysis involves chemical digestion of an entire fish, followed by

extraction, radiolabeling, and a competitive binding assay to quantify steroid levels (see

Appendix B for laboratory protocols). Whole body chemical digestion is accomplished

by boiling individual fish in potassium hydroxide (30% w/v) at a volume three times the

individual fish weight. Fifty microliters of resultant homogenate are removed in

duplicate for extraction. Diethyl ether added in excess (4 mL) is used to extract lipophilic

compounds, including sex steroids, from the digestion homogenate. Extraction and

evaporation is performed twice to increase extraction efficiency. A reaction solution is

prepared composed of evaporated extract, tritiated hormone and a corresponding

hormone-specific antibody. This solution incubates overnight to allow unlabeled

hormone from the extract sample (at unknown concentration) and radiolabeled hormone

(at a known concentration) to compete for antibody binding sites. After incubation, the

reaction solution is centrifuged with charcoal dextran to remove any hormone not bound









to antibody. Radioactivity is measured using scintillation spectrophotometry. Standard

curves (at 1, 5, 10, 25, 50, 100, 250, 500, and 1,000 pg hormone) are generated for each

hormone using known concentrations of radioinert hormone in buffer. Steroid

concentration in samples is then calculated by aligning values of an inhibition curve,

generated from the competitive displacement of radiolabeled hormone in the sample, to

the standard curve. Validation and characterization of this procedure entailed

determination of digestion and extraction efficiencies by reproductive status and exposure

(n = 5 each group for each hormone: males, gravid and nongravid females, juveniles, and

effluent-exposed males and females); minimum detection limits on the standard curve;

cross-reactivities of antiserum with other steroids; and inter- and intra-assay variation.

Statistics

Body weight and standard length were used to calculate condition factor,

K = weight / length3 x 100, as an indication of overall health used by the aquaculture

industry (values at least 1 are considered healthy). The length ratio of anal fin Rays 4 and

6 was calculated as an index of anal fin elongation. Estrogen and testosterone

concentrations were used to calculate a ratio indicating masculine hormone profile

(E:T<1) or feminine hormone profile (E:T>1).

Gender identification and anal fin morphological measurements used for

validations were analyzed using Pearson's product moment correlations or calculation of

coefficients of variance. Anal fin morphology and sex steroid data were analyzed within

sex using two-way analysis of covariance (ANCOVA) to test for significant variation by

site and season for the summer and winter 2000 data, or by site and size class within the

winter 2000 female anal fin data only. Size class was also analyzed by one-way ANOVA

within site. Any data failing tests for normality and homogeneity of variance were









transformed using log transformations. Angus et al. (2001) determined length ratio of

Rays 4 to 6 is appropriately analyzed using parametric statistics after log transformation

of the ratio data. Significant differences in the ANCOVA and ANOVA tests were

analyzed for multiple comparisons using Tukey's HSD. Within site, differences between

seasons were analyzed by t-test. Statistical significance was attained at a<0.05 for all

tests. All statistical analyses were conducted using SAS version 9.0.

Results and Discussion

Water Quality

As expected, conductivity, salinity, turbidity and pH were elevated at effluent-

exposed sites compared to unexposed sites (Table 2-1). Water temperature was highest at

the reference site and predischarge pond, and increased along the length of Rice Creek.

Reference and upstream sites were comparable, other than water temperature. Dissolved

oxygen remained high enough to support fish at all sites (>4 mg/L).

Validation of Gender Identification Using the Urogenital Papilla

Agreement about gender was significant at all three levels for which it was tested:

among personnel within the USGS laboratory (r2 = 0.899); between USGS and NCASI

laboratories (r2 = 0.93); and against histological evaluation (r2 = 0.99). Between

laboratories, females were agreed upon slightly more than males and the error rate for the

new technique was 3.5%. Disagreement occurred for 15 fish (7.5% of all fish examined,

Figure 2-2): of these fish, half (7) were incorrectly identified by the urogenital papilla,

whereas the other half (8) could not be accurately identified by inspection of gonads.

Incorrectly identified fish were equally from both exposed [DIS] and unexposed [REF2,

U(8)] sites, precluding bias against masculinized fish. Comparison against histological

identification of gonads was even more accurate with an error rate less than 1%.









Disagreement occurred for 7 fish (2% of all fish examined Figure 2-3): of these fish, over

half (5) could not be accurately identified by histological inspection, while the remaining

two fish were incorrectly identified by the urogenital papilla. Therefore the urogenital

papilla is a reliable indicator of internal sex. This new noninvasive external gender

identification technique could be very useful for a multitude of studies utilizing

mosquitofish repeatedly sampled over time.

Morphology

Morphological measurements validated well, with computer-aided measurements

preferable to manual measurements. Body size in winter 2000 was not impacted by

effluent exposure relative to site; both sexes were in good general health judged by

condition factor higher than 1. Increased gonopodial length was weakly indicated for

males at effluent-exposed sites. Precocious maturation in males was not apparent, since

smallest males did not have fully developed gonopodia to afford statistical analysis. In

females, analysis by size class did not reveal specific size classes) associated with anal

fin elongation, although caution was implied for using appropriate sample sizes by

females from the outfall site. Female anal fin elongation was significantly elevated at the

discharge and first downstream site in both fall and winter. Seasonality was not evident

for anal fin elongation in either sex.

Validations

Measurements on fish before and after preservation in formalin were significantly

correlated (Table 2-2), although measurements on formalin-preserved fish were

consistently smaller than on fresh fish (data not shown). This bias was expected since

preservation tends to dehydrate and shrink specimens. Observer bias between NCASI

and USGS laboratories measured differences not only between two observers, but also









differences between equipment. All measurements were significantly correlated (Table

2-2), although correlation coefficient for Ray 4 indicates somewhat inconsistently larger

measurements by USGS. NCASI reports data to the nearest 0.01 mm for manual anal fin

measurements (Bradley et al. 2004), while USGS reports data to the nearest 0.1 mm.

Therefore, instrument bias is indicated. Significant differences occurred between manual

and computer-aided measurements (Table 2-2). This discrepancy is likely due to the

more accurate measurement by computer (+ 0.01 mm), similar to the nonsignificant

differences between laboratories. Manual and computer-aided measurements of fish are

thus not comparable when examining data across studies.

Variation within one observer and among several observers was also addressed for

both types of measurements. Coefficients of variation at 10% or less were considered

acceptable. Repeated manual measurements by one observer were consistent, as were

measurements among observers (Table 2-3). Repeated computer-aided measurements by

one observer were even more consistent than for manual measurements, but this was not

quite the case among observers. This result means all computer-aided measurements

should be (and were) made by the same observer for masculinization studies. With the

greater accuracy afforded by computer-aided measurements (the Ray can be traced along

exact curvatures, as opposed to linear distances with manual measurements), this

technique is preferable. As Bradley et al. (2004) also note, computer-aided

measurements are ideal for archiving data, but are time-intensive with the extra step of

photography involved.

Body Size for Fall 2000 Collection

Body size data is presented for winter 2000 in this chapter (Table 2-4), while

summer 2000 and 2001 data is presented in Chapters 3 and 4, respectively. Overall









males were not impacted by exposure site. Weight decreased with increasing distance

from outfall (i.e. decreasing exposure), and the longest males were found at the upstream

site even compared to the reference site [REF1]. Condition factor, above one and

indicating healthy males, significantly varied inversely with exposure similar to weight:

highest for all sites at the outfall [DIS]; lower than outfall and unexposed sites but higher

than furthest downstream site at the first downstream site [D(1)]; and lowest for all sites

at the furthest downstream sites [D(3+6)]. However this index is not appropriately

interpreted beyond the benchmark of above or below a value of one. Female body size

was not affected by exposure site. The only significant difference was the presence of

longest females in the lower half of Rice Creek. Condition factor was also above one for

females at all sites, indicating adequate overall health.

Influence of Body Size on Anal Fin Morphology

Since body size (length) has been associated with anal fin length in masculinized

females and precociously matured males, fish were divided into 5 mm size classes and

analyzed as a covariate with site. Size classes were also analyzed within each site. When

males were divided into size classes, the majority fell into the 20-24.99 mm class, a

minority fell into the 25-29.99 mm, and less than 10 fell into the <20 mm class. This

paucity of mature males in the smallest size class precluded evaluation for precocious

maturation. Precocious maturation may not be occurring in these males during the

winter, since small males with fully differentiated gonopodia were not found in

significant abundance. Precocious maturation is addressed in the summer 2001

collections presented in Chapter 4.

Standard length significantly correlated with total anal fin length (linear distance of

Ray 4) for female mosquitofish in winter 2000 (r2 = 0.70, data not shown). However,









index of anal fin elongation (Ray 4 to Ray 6 length) was statistically independent of

standard length (r2 = 0.222, data not shown). Overall across size classes (Figure 2-4), the

first two exposure sites [DIS and D(1)] were significantly longer by the index of anal fin

elongation (Ray 4 to Ray 6 length) compared to nonexposed sites and the lower half of

Rice Creek [REF1, U(8), D(3+6)] Size class did not significantly covary with site. For

size classes within each site, the only statistically significant difference was at the outfall

[DIS]: females in the middle two size classes had significantly longer anal fin elongation

than the largest size class (Figure 2-3). This result negates predicted responses of size

classes due to drought: in light of the drought faced by females in 1999 and early 2000

(Appendix A), the older/larger females should have greater elongation. Instead, this

result may imply anal fin elongation is induced at a sensitive life stage and/or represents a

dynamic exposure to bioactive compounds. (Dynamic exposure refers to variable

concentrations of effluent components over time, dependent on factors such as tree

species for furnish, within plant processing spills, rainfall/dilution, and bacterial

degradation.) Without specific exposure data for these fish, these conclusions remain

speculative.

Seasonality

Female anal fin elongation at the first two exposed sites was significantly elevated

compared to unexposed sites for both seasons (Figure 2-5A). Elongations were more

similar to a developing male gonopodium in length and lacked terminal differentiations

(data not shown, see Chapter 3 for photographs of anal fins of females collected in

summer 2000). Site and season did not covary for any of these data. Season alone

influenced the latter two downstream sites [D(1) and D(3+6)], with opposite trends: at the

first downstream site [D(1)] elongation was larger in winter than summer and vice versa









for the lower half of Rice Creek [D(3+6)]. Therefore, season was not consistently

influencing presence of anal fin elongation in females from Rice Creek. This result is in

contrast to the seasonality study from Elevenmile Creek (Cody and Bortone 1997), where

winter months demonstrated a reduction in elongation compared to summer months.

Granted, the Elevenmile Creek study was conducted over the entire year, while the

current study compares one month from each season. Variation within a season was not

reported for Elevenmile Creek by Cody and Bortone (1997). Both reports agree the

response is present regardless of season.

Increased anal fin elongation, or greater gonopodial length, was not as apparent for

males as for females (Figure 2-5B). The summer collection revealed males with shorter

gonopodia at the upstream site [U(8)] compared to the rest. The winter collection

demonstrated significantly longer gonopodia at the first downstream site compared to the

reference [REF2] but not the upstream site [U(8)], and longer gonopodia further

downstream [D(1) and D(3+6)] compared to both the upstream and reference sites.

Hence selection of unexposed site(s) is of vital importance for interpreting results.

Season had no effect on gonopodial length as either a covariate with site or alone.

Overall the evidence is weak for increased gonopodial length in males due to pulp mill

effluent exposure.

Sex Steroids

Radioimmunoassay of primary sex steroids was validated for whole mosquitofish.

To date, only one other study has reported this biomarker in mosquitofish for males only

(Toft et al. 2003). Elevated testosterone was associated with instream effluent-exposed

sites in females regardless of season. However seasonality, both alone and relation to

site, was detected. Estrogen to testosterone ratios were masculinized (i.e., greater than 1)









at impacted sites for the summer collection only. Importantly, neither hormone

concentrations nor their ratio were altered in 100% final effluent before discharge,

indicating additional environmental factors interact to produce the response. For males,

seasonality of sex steroids was indicated, and effluent exposure did not appear to alter

concentrations and ratio.

Validations

Validations were completed for 17p-estradiol and testosterone. Since

11-ketotestosterone was not detected in these fish samples, only partial validation was

possible and reported elsewhere (Gross et al. 2001).

Digestion and extraction efficiencies were not influenced by reproductive status or

exposure to pulp mill effluent (see Table 2-5), but greater efficiency was consistently

achieved for 173-estradiol than testosterone. For 173-estradiol overall, digestion

efficiency averaged 70 + 4.9% while extraction efficiency averaged 65 + 6.4%. For

testosterone overall, digestion efficiency averaged 63+ 3.9% while extraction efficiency

averaged 51 + 12.5%. While these values would be considered low for plasma or serum

samples, efficiencies at 60% or greater are high for whole body samples. Data was

corrected for both digestion and extraction efficiencies.

Minimum detection limits on standard curves were 6.4 pg/mL and 9.3 pg/mL for

17p-estradiol and testosterone, respectively. Cross-reactivities of 173-estradiol antiserum

(produced and characterized by T. S. Gross, University of Florida) with other steroids

were: 11.2% for estrone, 1.7% for estriol, less than 1% for 17a-estradiol and

androstenedione, and less than 0.1% for all other steroids examined (ICN Biomedicals).

Cross-reactivities of testosterone antiserum (produced and characterized by T. S. Gross,









University of Florida) with other steroids were: 17.6% for dihydrotestosterone, 2.3% for

androstenedione, 1.4% for 11-ketotestosterone, <1.0% for androstenediol, and <0.1% for

all other steroids examined (ICN Biomedicals). A pooled sample (1.25 g of unexposed

males and females randomly selected and approximately 135 pg 173-estradiol/mL and

176 pg testosterone/mL added) was assayed serially in 10, 20, 30, 40, and 50 [tL volumes

(final volume of 50 [iL with boiled KOH). Resulting inhibition curves were parallel to

respective standard curve based upon tests for homogeneity of regression indicating

curves did not differ. Finally, average inter-assay and intra-assay coefficients of variation

were 7.8% and 9.4% for 173-estradiol and 8.7% and 10.1% for testosterone. All values

are reported as pg hormone per g body weight.

Seasonality

Both site and season significantly covaried for 173-estradiol in female mosquitofish

from the discharge [DIS] and first downstream [D(1)] sites in Rice Creek (Figure 2-6A).

However, examination of site and season separately does not reveal a consistent pattern.

17p-estradiol appeared seasonally elevated in winter, although the opposite occurred for

the lower half of Rice Creek. During the summer, 173-estradiol was depressed in

females from upstream and downstream sites [U(8), DIS, D(1)] compared to remaining

sites. Notably, 173-estradiol in females was not depressed in 100% effluent before

discharge [PRE-DIS]. Females collected in the winter revealed a different pattern: 173-

estradiol was significantly elevated at the upstream site [U(8)] compared to the discharge

and first downstream sites [DIS and D(1)]. However, this hormone was not significantly

different at effluent-exposed sites compared to reference fish [REF2], reiterating the

influence of unexposed site selection. While it is ideal to collect upstream of effluent









outfall, additional references provide insight into natural variability. Based upon inherent

variation implied by these data within sites and between unexposed sites, 173-estradiol

alone does not appear influenced by effluent exposure but may be seasonally influenced.

Site and season also significantly covaried for testosterone in female mosquitofish

from the discharge [DIS] and first downstream [D(1)] sites in Rice Creek (Figure 2-6B).

Patterns for this hormone are clearer. Testosterone was seasonally depressed in winter

across all sites. For both seasons, testosterone was elevated at either the first downstream

site (summer) or both the discharge and first downstream sites (winter). Like 173-

estradiol, testosterone was not impacted at the 100% final effluent before discharge

[PRE-DIS]. Thus, elevated testosterone concentrations were associated with instream

effluent exposure but not at highest effluent concentrations before discharge.

Figure 2-7 illustrates percentage of females with normal, feminine sex steroid ratios

(> 1) versus masculine steroid ratios biased toward testosterone (< 1). Estrogen to

testosterone ratios were significantly masculinized for females in the summer (DIS and

D(1) which averaged 0.8 and 0.3, respectively). In the winter masculine versus feminine

sex steroid ratios were not significantly different among sites. Thus, masculinized steroid

profiles appear seasonally affected by effluent exposed sites. A low background level of

masculine ratios existed in females from unexposed sites [U(8) and REF 1 for summer,

REF2 for winter], inferring this type of hormone profile can occur naturally in the

population. In 100% effluent before discharge (summer collection), no females had

masculinized estrogen to testosterone ratios, emphasizing the difference in response

between predischarge and instream effluent-exposed sites.









Sex steroids in male mosquitofish covaried by site and season for 173-estradiol but

not for testosterone (Figure 2-8). Within sites, seasonality was inferred at the upstream

and outfall sites [U(8) and DIS] for 173-estradiol and at the outfall and lower half of Rice

Creek [DIS and D(3+6)] for testosterone. 173-estradiol was elevated and testosterone

was depressed in the winter for these sites, respectively. By site alone, in the winter 173-

estradiol significantly peaked at the outfall and testosterone peaked at the upstream site.

No differences were detected by site in the summer. Testosterone data reveal large

variation within site and between unexposed sites, similar to 173-estradiol in females and

again stressing importance of reference site selection and an inherent natural variability

(although mosquitofish have a reproductively active season they are asynchronous

breeders). Estrogen to testosterone ratios were dominated by testosterone for all males

across all sites for both seasons (significantly less than one, approximately 0.01 to 0.001).

Overall, sex steroids in males did not appear impacted by effluent exposed sites and

seasonality was inconsistently indicated.

Recently Toft et al. (2003) examined whole body sex steroids in male mosquitofish

collected in lakes contaminated with agricultural pesticides. Technique was similar to the

radioimmunoassay used for our study, although mechanical as opposed to chemical

digestion of fish preceded extraction. Steroids were monitored December to May and

seasonality was implied, although statistical relevance was unstated. Overall

concentrations for both steroids were much higher than concentrations observed in this

study: 17p-estradiol was five times higher and testosterone was approximately 2 times

higher. Extraction efficiencies were higher for the pesticide lake study (83% and 111%

for 173-estradiol and testosterone, respectively) and were not corrected, meaning actual









concentrations were even higher for 173-estradiol but lower for testosterone. Digestion

efficiencies were not reported, nor were cross reactivities with other sex steroids. Thus

direct comparison of results is difficult and laboratory differences, in addition to habitat

and seasonal variations, may explain discrepancies.

Conclusions

Validations of gender identification, morphological and steroidal measurements

support the use of mosquitofish as a bioindicator of pulp and paper mill effluent. Site and

seasonal differences were not readily apparent in morphological measurements of male

mosquitofish, especially in association to effluent-exposed sites. Precocious maturation

could not be evaluated in the winter collection because small mature males were not

captured in adequate numbers for statistical comparison. Therefore, this aspect must be

re-examined.

For females, size class was not a complicating factor for the index of anal fin

elongation overall, although differential life stage exposure or dynamic exposure could be

speculated at the outfall site. Masculinization at sites closest to effluent discharge was

evident for both anal fin morphology and steroids in females: anal fin elongation was

independent of season, while effects on steroids were seasonal. This difference in

seasonality may be the product of separate mechanisms, exposure to bioactive

compounds, and/or perhaps differential exposure due to drought (see discussion below).

However, both anal fin elongation and sex steroids were not measured in the same fish.

The development and validation of computer-aided measurements allowed both

endpoints to be measured in the same fresh fish for research presented in subsequent

chapters.









Importantly, neither hormone concentrations nor ratio were altered in 100% final

effluent prior to discharge. Anal fin elongation at this site was not examined until 2001

and 2002, and data is presented in Chapters 4 and 3, respectively. This result points

toward the presence of bioactive compounds in the receiving stream below discharge and

not in the effluent by itself, indicating additional environmental factors interact to

produce the response. Which environmental factors are influential is pure speculation

without chemistry, controlled exposure or long-term biomonitoring data associated with

this study.

With that caveat, there is one factor that is often overlooked yet crucial to the

hypothesized mechanism of anal fin masculinization via degraded phytosterols: bacteria.

If the bacterial degradation hypothesis is correct, instream bacterial communities may be

more efficiently converting phytosterols to androgens than communities living in

retention ponds. For example, aerobic microorganisms degraded 173-estradiol 60 to 130

days faster and by first-order kinetics compared to anaerobic microorganisms (Quinn

2004). Natural variation in bacterial communities may also alter degree of response for

that matter. Only one type of bacteria (Mycobacterium smegmatis) has been examined in

laboratory exposures to degraded phytosterols (Denton et al. 1985, Howell and Denton

1989, McCarthy et al. 2004), and the Mycobacterium genus is more efficient at

transforming phytosterols to androgens than other bacterial species (Marcheck et al.

1972).

A potentially confounding factor in these studies was the variation in some

responses at unexposed sites. Significant differences between reference and upstream

sites existed for male gonopodial length and testosterone, and for 173-estradiol in









females. If collection had been restricted to one unexposed site, responses would have

been either masked or abnormally inflated. Unfortunately the ideal reference site does

not exist: often an upstream site is the best at representing environmental habitats at

exposed sites, although habitat can vary substantially along a stream. Therefore multiple

reference sites within the receiving system region are imperative to document natural

variation in response and allow more relevant interpretation. Overall, there was no

difference in female anal fin elongation between unexposed sites, in agreement with the

geographic study by Bradley et al. (2004). Their study collected at dozens of reference

and polluted sites (not just pulp and paper mill effluent), and they pooled reference data

for analysis since reference values were not significantly different. Since our study

investigated several other endpoints in addition to female anal fin elongation, multiple

reference sites were ideal when funding and efforts allowed.

Another confounding situation involved weather during this study. Summer 2000

fish were living under drought conditions, while winter 2000 fish were collected after

relief from drought with a slightly higher than normal rainy season (Appendix A).

According to Davis and Bortone (1992) and Cody and Bortone (1997), anecdotal

observation of mosquitofish during times of drought corresponded to increased

masculinization of the anal fin. The logical cause was a concentration of effluent, a

plausible assumption for the low-flow streams characteristic of the Florida mills under

investigation. This argument was used to support seasonal differences detected in anal

fin length of females (Cody and Bortone 1997) since precipitation follows a seasonal

pattern. However, since wild females survive 1 to 2 years, anal fin elongation among

summer populations would still be present through the winter. According to McCarthy et









al. (2004), gonopodial development in female mosquitofish does not regress once effluent

exposure ceases. This controversial point, however, has not been well documented. Sex

steroid levels, in contrast, are more labile and subject to recovery in effluent-exposed fish

(McMaster et al. 2003), and this study documented normal steroid ratios in the winter

collection. Thus sex steroids may be a more sensitive marker of differential effluent

exposure, while anal fin elongation is a more static and durable marker. Condition factor

data cannot support or refute the potential influence of drought due to the experimental

bias. A critical limitation of this drought hypothesis is a lack of exposure data to support

concentration of effluent during times of drought. As an estimate of relative

concentration of effluent, conductivity was greater in the summer than winter, providing

some support to the drought hypothesis. Regular monthly field collections could shed

more light on apparent trends. However, effects of drought should be considered when

using mosquitofish as a bioindicator.

Two aspects of previously defined bioindicator criteria were addressed in this

chapter: method and seasonal variation. Method variation was low, supporting the use of

mosquitofish as bioindicators. Wild-caught females demonstrated greater effects than

wild males. In females, seasonal variation related to exposure occurred for sex steroids

but not anal fin elongation, indicating greater sensitivity of steroids than anal fin

morphology. Ideally, both of these markers are measured for individual fish to better

evaluate the hypothesized link between anal fin elongation and altered sex steroid

profiles. An expanded year long seasonality study of mosquitofish living in pulp mill

effluents would allow better interpretation of absolute hormone concentrations, and

changes in anal fin morphology, especially since the present results are in conflict with









reported seasonal effects at Elevenmile Creek (anal fins) and reported concentrations of

sex steroids in males.

Table 2-1. Water quality parameters of Rice Creek field collection sites in winter 2000
Site REF2 U(8) PRE-DIS DIS D(1) D(3+6)
Winter 2000
Water temperature (C) 17.7 14.9 16.4 15.2 15.0 18.5
Conductivity (iS) 167 173 2132 1112 1315 1336
Salinity (ppt) 0.1 0.1 1.1 0.6 1.3 0.7
Dissolved Oxygen (mg/L) 7.60 7.00 4.14 7.23 11.90 8.44
Turbidity (ntu) 1.90 3.34 32.5 18.9 18.4 13.5
pH 6.73 6.45 7.35 7.23 7.00 7.15

Table 2-2. Correlation coefficients (r2) for morphological measurements made before
and after preservation in formalin; between USGS and NCASI laboratories; and between
manual and computer-aided measurement by the same observer. Standard Length (SL)
was only measured manually using digital calipers. All correlations were statistically
significant at p < 0.05 unless noted.
Measurement Ray 4 Ray 6 SL
Preservation 0.93 0.94 0.94
Between laboratories 0.54 0.80 0.97
Manual vs. computer-aided 0.27* 0.33* NAa
aNA = not applicable
*not significantly correlated (p > 0.05)

Table 2-3. Average coefficients of variation for manual and computer-aided
measurements by observer and among observers (three measurements per
trait). Standard Length (SL) was only measured manually using digital
calipers.
Measurement Ray 4 Ray 6 SL
Manual
By observer 3.1% 3.6% 1.4%
Among observers 8.8% 9.4% 5.6%
Computer-aided
By observer 1.1% 0.9% NAa
Among observers 12.3% 10.1% NAa
aNA = not applicable










Table 2-4. Body size parameters (ave + se) for mosquitofish collected in winter 2000
Site REF2 U(8) DIS D(1) D(3+6)
Winter 2000
Sample Size 45 (29,16)e 89 (61,28) 73 (73,16) 46 (29,17) 52 (32,20)
SBody Weight 0.255+0.010 0.289+0.008 0.255+0.013 0.221+0.006a 0.169+0.009a

SStandard Length 22.74+0.31 24.08+0.20b 22.24+0.31 22.95+0.25 23.78+0.33
(mm)
S Condition Factor 2.13+0.03 2.03+0.02 2.24+0.02c 1.84+0.04c 1.22+0.03c
(g/cm3)
Sample Size 141 (91,50)e 106 (64,42) 106 (53,53) 128 (99,29) 112 (65,47)
SBody Weight 0.567+0.029 0.527+0.029 0.513+0.025 0.471+0.022 0.479+0.033
(g)
SStandard Length 28.97+0.48 28.67+0.46 28.18+0.44 27.37+0.40 29.13+0.59d
(mm)
Condition Factor 2.08+0.01 2.05+0.02 2.15+0.03 2.11+0.02 1.62+0.04
(g/cm3)
aD(1) statistically different than D(3+6); D(3+6) statistically different from all other sites
bU(8) statistically different than REF2, DIS, and D(1) (p < 0.05)
cDIS; D(1); and D(3+6) statistically different than rest (p < 0.05)
dD(3+6) statistically different than rest (p < 0.05)
sample sizes displayed as: total sample size (preserved anal fin measurements, hormone and
computer-aided measurements)











Table 2-5. Digestion and extraction efficiencies, and coefficients of variation (CV), by
exposure and reproductive status for mosquitofish whole body hormone
analysis.


17p-estradiol
Exposed 0
Unexposed 0
Exposed Y
Unexposed
Adult g
Adult gravid
Adult nongravid
Juvenile
Testosterone
Exposed 0
Unexposed 0
Exposed Y
Unexposed
Adult g
Adult gravid
Adult nongravid
Juvenile


Efficiency

70+4.7%
63+3.3%
75+5.8%
76+7.6%
67+6.4%
74+4.5%
72+4.1%
69+2.8%

64+2.5%
57+3.9%
58+3.9%
50+2.9%
89+3.1%
61+4.5%
67+3.8%
62+7.1%


Digestion
CV


6.7%
5.2%
7.7%
10%
9.6%
8.9%
9.2%
5.7%

3.9%
7.0%
6.8%
5.7%
3.1%
7.3%
5.4%
11%


Extraction
Efficiency CV


66+6.6%
61+7.2%
71.7+9.5%
60+5.8%
62+6.2%
75+5.8%
60+6.3%
67+4.9%

41+4.4%
44+2.5%
45+4.1%
46+6.1%
66+3.7%
61+4.6%
66+5.3%
40+5.9%


10%
11.7%
13.2%
9.7%
10%
12.7%
10.9%
10%

14%
9.9%
9.1%
13%
14%
15%
10%
15%













A

















B





DUS
St Johns Ri r



EtoniD(6)





0 Palatka



0051 4
Z-USGS Putnam County

Figure 2-1. Maps of Rice Creek, a tributary of the Saint Johns River, FL, USA. A)
Relative location in Florida. B) Summer 2000 field collection sites. C)
Winter 2000 collection sites. D) Summer 2001 collection sites. Site symbols
distinguish sites exposed to effluent: circles = unexposed and triangles =
exposed. Site abbreviations denote upstream (U) or downstream (D) of
discharge, followed by approximate distance (km) from discharge in
parentheses. PRE-DIS indicates site before discharge into the creek; DIS
denotes site at discharge into creek; and REF indicates reference site, followed
by identifying number.























































!.. Putnam County
Figure 2-1. Continued










7.5%
48% 44.5%









agree male U agree female

D male/female* female/male*

?/male* U ?/female*

*NCASI/USGS

Figure 2-2. Gender agreement between NCASI and USGS laboratories. Fish were sexed
externally by USGS using the urogenital papilla then gonads were grossly
identified by NACSI. Question marks corresponding to NCASI were unable
to be reliably sexed by gross gonad examination.











2%
49% 49%









agree male U agree female

E male/female* female/male*

?/female*

*histology/urogenital papilla
Figure 2-3. Gender agreement within USGS laboratory. Fish were sexed externally
using the urogenital papilla then gonads were examined histologically.
Question marks indicate gender could not be reliably identified by histological
techniques.






















- 20-24.99 mm
25-29.99 mm
30-34.99 mm
- 35-39.99mm


REF2 U(8) DIS D(1) D(3+6)
SITE
Figure 2-4. Female index of anal fin elongation (linear Ray 4 / Ray 6, manually
measured on preserved fish) for each site by 5 mm increments (winter 2000).
Significant differences (p < 0.05) are color coded within size class: u =
different from upstream site [U(8)]; r = different from reference site [REF2];
and d = different from other exposed sites [DIS, D(1)]. Purple asterisks
indicate significant differences to the largest size class within site (p < 0.05).


1.5
0



0
1.0





O 0.5





0.0






























REF1 REF2 U(8) DIS D(1) D(3+6)


I SUMMER
I WINTER


REF1 REF2 U(8) DIS D(1) D(3+6)
SITE


Figure 2-5. Index of anal fin elongation (linear Ray 4 / Ray 6, manually measured on
preserved fish) for winter and summer months in 2000. A) Females. B)
Males. Dashed lines separate sites not involved in ANCOVA analysis by site
and season. Letters indicate significant differences among sites within season
(p < 0.05): "a" denotes differences to nonlettered sites except D(1) was not
different than REF 1; "b" denotes differences to nonlettered sites; "c" denotes
differences to all but REF 1; "d" denotes differences to REF1 and U(8); "d"
denotes differences to DIS, D(1), and REF 1. Season was significant within
site (p < 0.05) for females at the two downstream sites only [D(1) and D(3+6),
circled in yellow].


2.5

0
< 2.0
0


W 1.5


S1.0


0
X 0.5














1000


800


600


400


200


REF1 REF2 PRE-DIS U(8) DIS D(1) D(3+6)
SITE


* SUMMER
* WINTER


600
~ o-
So 500
-

I 400
-
H 300
C)-

( 200-
H
100

0


REF1 REF2 PRE-DIS U(8)
SITE


DIS D(1)


D(3+6)


Figure 2-6. Female whole body sex steroids from collections made in the summer and
winter of 2000 (ave + se). A) 173-estradiol. B) Testosterone. Dashed lines
separate sites not involved in ANCOVA analysis by site and season. Letters
indicate significant differences by site within season (p < 0.05): "a" denotes
differences to nonlettered sites; "b" denotes differences to DIS and D(1); "c"
denotes differences to all but DIS; and "d" denotes differences to all other
sites. Yellow circles demonstrate differences between seasons (p < 0.05).
Green asterisk show that both hormones covary by site and season (p < 0.05).


A





0
A i


II

i


II


7nn







74



V MASCULINE STEROID RATIO (E:T<1)
A FEMININE STEROID RATIO(E:T>1)


D(6)

D(3)

D(1)

DIS


PRE-DIS

U(8)

REF1


0 10 20 30 40 50 60 70 80 90 100


% FEMALES


D(3+6)


D(1)


DIS


U(8)


REF2


0 10 20 30 40 50 60 70 80 90 100

% FEMALES

Figure 2-7. Percentage of female mosquitofish with masculine and feminine sex steroid
ratios collected in 2000. A) Summer (X2 = 40.12, df = 6, p < 0.05). B)
Winter (X2 = 8.270, df= 4, p < 0.05).


I










A i

oa












REFi REF2 PRE-DIS U(8) DIS D(1) D(3+6)
SITE

B
I
















- I






\ ^ [I


REF1 REF2 PRE-DIS U(8)


* SUMMER
* WINTER


DIS D(1) D(3+6)


SITE
Figure 2-8. Male whole body sex steroids from collections made in the summer and
winter of 2000 (ave + se). A) 173-estradiol. B) Testosterone. Dashed lines
separate sites not involved in ANCOVA analysis by site and season. Letters
indicate significant differences by site within season (p < 0.05): "a" denotes
differences to all other sites; "b" denotes differences to REF2, DIS and D(1).
Yellow circles demonstrate differences between seasons (p < 0.05). Green
asterisk signifies 17p-estradiol covaried by site and season (p < 0.05).


1200

1000














CHAPTER 3
DIMINISHED EFFECTS OF PULP AND PAPER MILL EFFLUENT ON EASTERN
MOSQUITOFISH BEFORE AND AFTER MAJOR PROCESS IMPROVEMENTS

The US Environmental Protection Agency's (EPA) Cluster Rule was enacted in

1998 to regulate both air and water pollution by the pulp and paper industry. Final

implementation of this rule was carried out in 2001 by the Georgia-Pacific bleached kraft

mill in Palatka, FL. Comparison of mosquitofish collections before and after these major

process changes was conducted to determine if responses were reduced or eliminated.

Males and females were evaluated for anal fin morphology and whole body sex steroids,

although male responses seemed more affected by potential seasonality as opposed to

effluent exposure. Female anal fin elongation was consistently reduced, but not

eliminated, in effluent-exposed sites. Female sex steroid concentrations were difficult to

interpret and may be seasonally affected; however the ratio of 173-estradiol to

testosterone indicated a masculnized hormone profile remained despite processing

improvements. No association between anal fin elongation and sex steroid levels or

ratios was apparent. Anal fin morphology likely portrays longer term, chronic exposure

whereas sex steroids provide a snapshot of more recent exposure. Overall, reduced anal

fin elongation associated with major process changes supports the use of mosquitofish as

a bioindicator of effluent exposure. Seasonal effects on sex steroids need to be clarified

before use of this biomarker.









Introduction

Pulp and paper mills release a complex mixture of chemical compounds into the

aquatic environment that has been a source of environmental concern for the past two

decades (Borton et al. 2004). Compounds such as chlorinated organic, metals, and wood

extractives (e.g. resin acids, fatty acids, phytosterols, and lignin) have been shown to

cause an array of effects in fish (EPA 2002 and Borton et al. 2004). Acute lethal toxicity

concerns are no longer an issue, while chronic sublethal toxicity remains controversial.

Linking actual adverse effects to effluent components) has been a challenge since

effluents are complex by nature, and their variation in effluent composition among mills

and even within a single mill. Within this effort, a wealth of information has been

generated about effects in fish.

Pulp and paper mill effluents have been shown to alter reproductive function in

indigenous fish species. For example, white sucker (Catostomus commersonii) collected

near Canadian mills had reduced ovarian steroid biosynthesis, delayed sexual maturity,

decreased gonad size, and reduced expression of secondary sex characteristics (McMaster

et al. 1991, Munkittrick et al. 1991, McMaster et al. 1995, Van der Kraak et al. 1992).

Largemouth bass (Micropterus salmoides) exposed to whole effluent dilutions at the

same mill examined in our study exhibited depressed sex steroids, vitellogenin, GSI, fry

production and fry survival in largemouth bass (Sepulveda et al. 2001, Sepulveda et al.

2003).

Several reported effects on fish reproduction imply an androgenic effect. A Finnish

totally chlorine free (TCF) kraft mill was associated with male-biased sex ratios in wild

eelpout (Lycodes sp.) (Larsson et al. 2000). In a study of the same mill, increased male

coloration was observed under laboratory exposure of livebearing guppies (Poecilia









reticulate) (Larsson et al. 2002). Full life cycle testing using the fathead minnow

(Pimaphalespromelas) in Canada demonstrated both masculinization of females and

feminization of males (Parrott and Wood 2002, Parrott et al. 2003). Pulp mills in Florida

have been associated with development of male secondary sex characteristics in female

mosquitofish (Howell et al. 1980, Bortone and Drysdale 1981, Cody and Bortone 1997,

Bortone and Cody 1999, Jenkins et al. 2001, Parks et al. 2001) and possibly precocious

maturation in male Eastern mosquitofish (Gambusia holbrooki) (Howell et al. 1980,

Drysdale and Bortone 1989). Unique among the vast majority of these effects-based

studies, the mosquitofish work has led to a proposition of bioactive compounds:

androgens formed by bacterial degradation of phytosterols. This hypothesis has

stimulated controversial discussions about mosquitofish as a bioindicator or sentinel

species for pulp and paper mill effluents.

Eastern mosquitofish have been considered for use in regulatory testing and

screening of pulp mill effluent toxicity at both the state (Florida Department of

Protection, T. S. Gross, pers. comm.) and federal levels (Angus et al. 1997).

Mosquitofish have been proposed since most other wild small fish species found in

effluent-receiving streams cannot be collected in adequate numbers or maintained in the

laboratory for use in research.

As data on reproductive effects in fish has been generated, pulp and paper mills

have been improving and refining process technologies. Upgrades in pulp production

processes have the potential to reduce if not abolish reported effects. For example, short-

term laboratory exposures of goldfish (Carassius auratus) revealed a recovery of steroid

function following unknown process changes (McMaster et al. 1996). More recently,









temporary shutdown of the TCF mill associated with male-biased sex ratios in eelpout

allowed recovery of normal sex ratios in the exposed population (Larsson and Forlin

2002).

With the implementation of US EPA's Cluster Rule in 2001, the Georgia-Pacific

mill located in Palatka, FL, USA has modified its pulping and bleaching processes. The

objective of this study was to evaluate the influence of these process changes on a native

small fish species, the Eastern mosquitofish, specifically examining effects on two

endpoints: anal fin morphology and sex steroids.

Materials and Methods

Mill Characteristics

Georgia-Pacific's mill in Palatka, Florida, USA is a paper grade bleached kraft mill

established in 1947. It has two bleached (40% product) and one unbleached line (60%

product). The bleaching lines manufacture paper towels and tissue paper, whereas the

unbleached line produces kraft bags and linerboard. Wood furnish for this mill typically

consists of 50% softwood (slash, sand and loblolly pines) and 50% hardwood (gums,

tupelo, magnolia and water oaks) cycled back and forth between the two types of furnish.

Effluent receives secondary treatment consisting of anaerobic followed by aerobic

degradation with a retention time around 40 days.

Effluent discharges into Rice Creek, nearly 6 km upstream of the confluence with

the Saint Johns River (Figure 2-1). Rice Creek is a low-flow, tannic stream, so dilution

factor for effluent is low until it reaches the Saint Johns River. Before process changes

the average yearly instream effluent concentration was approximately 60% until reaching

the St Johns River, where concentration dropped below 10%. In contrast, most North









American mills average less than 5% yearly instream effluent concentration by

discharging into larger and/or faster-flowing bodies of water.

Before major process improvements, the Palatka mill released approximately 36

million gallons of effluent per day (mgd). Bleaching pre-process modifications used

elemental chlorine and up to 10% chlorine dioxide substitution. The bleaching sequences

were C90dioEopHDp and CEHD for the softwoods and hardwoods, respectively. Process

modifications in May 2001, to meet EPA the Cluster Rule, involved: 1) conversion to

ECF bleaching via 100% chlorine dioxide substitution; 2) reduction in black liquor

losses; 3) added condensate stripping; 4) conversion of all retention ponds to aerobic

degradation; and 5) reduction in water use resulting in release of approximately 28 mgd

effluent. The current bleaching sequence is DEopD for both types of furnish.

Field Collections

Field collections of adult mosquitofish occurred during the reproductively active

summer months for this species one year before (March and June 2000, n = 174 and n =

141 respectively) and one year after (April 2002, n = 363) process modifications. Water

quality parameters typically affected by pulp and paper mill effluents were measured

before fish collection at each site: dissolved oxygen, temperature, pH, conductivity,

salinity, and turbidity. Adult Eastern mosquitofish were collected along shallow

vegetated banks using dip nets and a backpack electroshocker at several locations

upstream and downstream of effluent discharge in Rice Creek, and at reference sites

lacking effluent exposure or any other known point sources of pollution (Figure 3-1 and

Appendix A). Fish were transported back to the laboratory in oxygenated bait buckets

then euthanized with a terminal dose of buffered tricaine methanesulfonate (Tricaine-S,

Western Chemical Inc., Ferndale, WA, USA).









Morphology

Once euthanized, fish were examined under a dissecting scope to determine gender

using the urogenital papilla (Chapter 2). General measurements of body size and selected

measurements of anal fin morphology were taken for all adult fish collected. Body

weight (+ 0.001 g) and standard length (+ 0.01 mm) were measured using a digital scale

and a pair of digital calipers. In 2000, the March collection was preserved in 10%

neutral-buffered formalin for anal fin measurements while the June collection was frozen

and stored at -800C for subsequent radioimmunoassay (RIA) of sex steroids. Linear

distance from base to tip of Rays 4 and 6 of the anal fin (+ 0.1 mm) were measured for

formalin-preserved fish under a dissecting scope using an ocular micrometer. In 2002, a

subset (n = 87) was preserved in 10% neutral-buffered formalin for comparison to 2002

anal fin data. Remaining fish were photographed digitally before freezing for sex steroid

analysis. Digital photographs of anal fins for these fish were measured using a computer

software program (SigmaScan Pro 5.0, SPSS, Inc.), tracing along the lengths of Rays 4

and 6 (+ 0.01 mm). Chapter 2 gives validation of these morphological measurements.

Sex Steroids

Whole body primary sex steroids (173-estradiol and testosterone for this species)

were analyzed using a modified RIA method originally developed for serum and plasma

samples of common carp, Cyprinus carpio, (Goodbred et al. 1997), and since adapted for

use in a variety of other aquatic species and tissue media such as plasma of largemouth

bass (Gross et al. 2001) and mantle of freshwater invertebrates (Gross et al. 2000).

Chapter 2 gives methods and validation of this assay.









Statistics

Body weight and standard length were used to calculate condition factor, K =

weight / length3 x 100 (g/cm3), as an indication of overall health used by the aquaculture

industry (values greater than 1 are considered healthy; less than one are considered poor).

The length ratio of anal fin Rays 4 and 6 was calculated as an index of anal fin

elongation. Estrogen and testosterone concentrations were used to calculate a ratio

indicating either a masculine hormone profile (E:T < 1) or a feminine hormone profile

(E:T> 1).

Any data failing tests for normality and homogeneity of variance were log

transformed. Anal fin morphology and sex steroid concentrations were analyzed

separately within sex using two-way analysis of covariance (ANCOVA) to test for

significant variation by site and year. Site differences within year were also analyzed by

one-way ANOVA. Significant differences in the ANCOVA and ANOVA were followed

by multiple comparison tests using Tukey's HSD. Within site, differences between years

were analyzed by Student's t-test. Fish measured for both anal fin morphology and sex

steroids (2002 data only) were analyzed in two ways: first, by examining Pearson's

correlations of the index of anal fin elongation to sex steroid concentrations and ratio,

then by t-test for differences in index of anal fin elongation between females with

masculine versus feminine E:T ratios. Statistical significance was set at a < 0.05 for all

tests. All statistical analyses were conducted using SAS version 9.0.

Results and Discussion

Water Quality

As expected, conductivity, salinity and turbidity were higher at effluent-exposed

sites compared to the upstream site (Table 3-1). The reference site REF1 was more