<%BANNER%>

Beavers of the Fish World

Permanent Link: http://ufdc.ufl.edu/UFE0023785/00001

Material Information

Title: Beavers of the Fish World Can Wood-Eating Catfishes Actually Digest Wood? A Nutritional Physiology Approach
Physical Description: 1 online resource (154 p.)
Language: english
Creator: German, Donovan
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: alimentary, degradation, detritivore, detritus, digest, digestive, enzyme, fermentation, fiber, gut, isotope, microbial, stable, xylivory, xylophagy
Zoology -- Dissertations, Academic -- UF
Genre: Zoology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Because wood is composed almost entirely of structural polysaccharides (primarily cellulose) few animals can digest it, let alone thrive on it. Most animals that do subsist on wood require endosymbiotic microorganisms in their digestive tracts to aid in the digestion of structural polysaccharides. Several species of catfishes (family Loricariidae) from the Amazonian basin (South America) were recently described as being 'wood-eating,' or 'xylivorous.' However, beyond some cursory analyses of gut contents and digestive enzyme activities, little is known about these animals and whether they can digest wood. In this dissertation I explored the structure and function of the digestive tract in four species of loricariid catfishes: three xylivorous species (Panaque cf. nigrolineatus ?Mara?on?, P. nocturnus, and Hypostomus pyrineusi) and one detritivorous species (Pterygoplichthys disjunctivus) that represents the common feeding mode of the family. Thus, I was able to examine whether the xylivorous species had specializations of the digestive tract affording them the ability to digest wood in comparison to the non-xylivorous, detritivorous species. I measured several aspects of the fishes' gut morphology, including intestinal folding patterns, microvilli surface area, pH and redox potentials, and microbial diversity in different regions of the gut. I also measured the activity levels of 14 digestive enzymes in the guts of the fish, and determined the sources of these enzymes: endogenous (produced by the fish) vs. exogenous (produced by microbes). These data were then compared to concentrations of microbial fermentative end products, called short chain fatty acids (SCFAs), and soluble carbohydrate profiles in the intestinal fluids of the fish to determine where microbes might be most active, and where nutrients were being hydrolyzed and absorbed. And finally, I measured the ability of captive fish to digest wood, to grow on it, and how quickly the fish pass wood through their digestive tracts. The results of these analyses suggested that, unlike termites, the alleged xylivorous catfishes of the Amazon cannot digest the fibrous components of wood in their digestive tracts. I found no evidence that they harbored endosymbionts in their guts capable of digesting the structural polysaccharides of wood. And, the laboratory feeding trials showed that the fish could not assimilate significant amounts of fiber or energy from wood, resulting in the fish losing weight on a wood diet. The fishes' entire feeding strategy, ranging from intake, to gut passage rates, digestive enzyme activities, intestinal morphology, soluble carbohydrate profiles, and levels of SCFAs throughout the gut suggest that the fish eat as much as they can, pass it through the digestive tract quickly, and assimilate the soluble, non-fibrous components available to them. Unlike many other wood-eating animals (e.g., termites, beavers), these fishes consume decaying wood in aquatic systems; decaying wood is in the process of being degraded by microbes, which produce soluble degradation products (e.g., beta-glucosides) that the fish can actually digest and assimilate. Thus, rather than harboring endosymbiotic microorganisms to digest wood fiber within their guts, the fish rely on microbial decomposition occurring in the environment. In this vein, the wood-eating catfishes are actually detritivores like so many other loricariid catfishes.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Donovan German.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Evans, David H.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0023785:00001

Permanent Link: http://ufdc.ufl.edu/UFE0023785/00001

Material Information

Title: Beavers of the Fish World Can Wood-Eating Catfishes Actually Digest Wood? A Nutritional Physiology Approach
Physical Description: 1 online resource (154 p.)
Language: english
Creator: German, Donovan
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: alimentary, degradation, detritivore, detritus, digest, digestive, enzyme, fermentation, fiber, gut, isotope, microbial, stable, xylivory, xylophagy
Zoology -- Dissertations, Academic -- UF
Genre: Zoology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Because wood is composed almost entirely of structural polysaccharides (primarily cellulose) few animals can digest it, let alone thrive on it. Most animals that do subsist on wood require endosymbiotic microorganisms in their digestive tracts to aid in the digestion of structural polysaccharides. Several species of catfishes (family Loricariidae) from the Amazonian basin (South America) were recently described as being 'wood-eating,' or 'xylivorous.' However, beyond some cursory analyses of gut contents and digestive enzyme activities, little is known about these animals and whether they can digest wood. In this dissertation I explored the structure and function of the digestive tract in four species of loricariid catfishes: three xylivorous species (Panaque cf. nigrolineatus ?Mara?on?, P. nocturnus, and Hypostomus pyrineusi) and one detritivorous species (Pterygoplichthys disjunctivus) that represents the common feeding mode of the family. Thus, I was able to examine whether the xylivorous species had specializations of the digestive tract affording them the ability to digest wood in comparison to the non-xylivorous, detritivorous species. I measured several aspects of the fishes' gut morphology, including intestinal folding patterns, microvilli surface area, pH and redox potentials, and microbial diversity in different regions of the gut. I also measured the activity levels of 14 digestive enzymes in the guts of the fish, and determined the sources of these enzymes: endogenous (produced by the fish) vs. exogenous (produced by microbes). These data were then compared to concentrations of microbial fermentative end products, called short chain fatty acids (SCFAs), and soluble carbohydrate profiles in the intestinal fluids of the fish to determine where microbes might be most active, and where nutrients were being hydrolyzed and absorbed. And finally, I measured the ability of captive fish to digest wood, to grow on it, and how quickly the fish pass wood through their digestive tracts. The results of these analyses suggested that, unlike termites, the alleged xylivorous catfishes of the Amazon cannot digest the fibrous components of wood in their digestive tracts. I found no evidence that they harbored endosymbionts in their guts capable of digesting the structural polysaccharides of wood. And, the laboratory feeding trials showed that the fish could not assimilate significant amounts of fiber or energy from wood, resulting in the fish losing weight on a wood diet. The fishes' entire feeding strategy, ranging from intake, to gut passage rates, digestive enzyme activities, intestinal morphology, soluble carbohydrate profiles, and levels of SCFAs throughout the gut suggest that the fish eat as much as they can, pass it through the digestive tract quickly, and assimilate the soluble, non-fibrous components available to them. Unlike many other wood-eating animals (e.g., termites, beavers), these fishes consume decaying wood in aquatic systems; decaying wood is in the process of being degraded by microbes, which produce soluble degradation products (e.g., beta-glucosides) that the fish can actually digest and assimilate. Thus, rather than harboring endosymbiotic microorganisms to digest wood fiber within their guts, the fish rely on microbial decomposition occurring in the environment. In this vein, the wood-eating catfishes are actually detritivores like so many other loricariid catfishes.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Donovan German.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Evans, David H.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0023785:00001


This item has the following downloads:


Full Text

PAGE 1

1 BEAVERS OF THE FISH WORLD: C AN WOOD-EATING CATFISHES ACTUALLY DIGEST WOOD? A NUTRITIONAL PHYSIOLOGY APPROACH By DONOVAN PARKS GERMAN 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 2008

PAGE 2

2 2008 Donovan Parks German

PAGE 3

3 To my parents, John C. and Gillian R. Germa n, for their unending support for whatever it is I choose to do in life.

PAGE 4

4 ACKNOWLEDGMENTS I thank my advisor, David H. Evans, as well as m y committee members, Karen A. Bjorndal, Douglas J. Levey, Larry M. Page, and Richard D. Miles for their invaluable support and guidance throughout this study. The academic and financial support given to me by the committee simply could not be offered by any othe r biology program in the country, and in that regard, the University of Florid a is second to none. I am extremely indebted to the following undergraduate research a ssistants whose hard work and wil lingness to do almost anything made this project possible: Jennette Villeda, Ana Ru iz, Daniel Neuberger, Ankita Patel, Meaghan Callahan, Rosalie Bittong, Norma Lizardo, Robyn M onckton, and Joseph Taylor. Additionally, I am grateful to Alfred Thomson, Dieldrich Berm udez, Samantha Hilber, Nathan Lujan, Krista Capps, Don Taphorn, Alex Flecker, David Weineke, and Jeremy Wright who assisted with the collection and dissectio n of fishes. A special thank you is offered to the Barker-Emmerson family in Orlando, FL, for giving me access to their land and allowing me to collect fish from their private spring (Starbuck Spri ng). Pratap Pullammanappallil Jason Curtis, Andrew Zimmerman, Lou Guillette, David Julian, and Stev e Phelps graciously allowed me to conduct analyses in their laboratories or store samples in their freezers at the University of Florida. Karen Kelly, Kim Backer-Kelly, and Lynda Schneider provided indispen sible support in the Electron Microscope Core Laborator y in the ICBR at the University of Florida. Brian Silliman offered his unending financial and lo gistic support, without which I would not have been able to function outside the university. I also thank the graduate students, post docs, and faculty in the De partment of Zoology for providing intellectual and emoti onal support throughout my graduate career. In particular, I would like to acknowledge Jada White, Ryan Mc Cleary, Greg Pryor, Larisa Grawe-DeSantis, Derek DeSantis, Alex Jahn, Silvia Lomascol o, Brandon Moore, Thea Edwards, James Nifong,

PAGE 5

5 Schuyler Van Montfrans, Benjamin Predmore, and Kim Reich. I would like to thank my wonderful lab mates Kelly Hynd man, Keith Choe, Leslie Babonis, and Justin Havird for methodological and emotional support. I am extr emely grateful to my wife, Lisa, and our adorable son, Merrick, who have literally kept me alive through this pr ocess. And finally, I thank my parents, John C. and Gillian R. German for supporting me in whatever I choose to do in life. Funding for this work came from the Univ ersity of Florida Mentoring Opportunity Program (2), SPICE (NSF GK-12) Fellowships (2), American Museum of Natural History Theodore Roosevelt Memorial Fund, American Society of Ichthyologists and Herpetologists Raney Awards (2), Brian Reiwald Memorial Schol arships (2), The Archie Carr Center for Sea Turtle Research, NSF All Catfish Species I nventory (PI: L.M. Page), Hartz-Mountain Corporation, and the deep pockets of my advi sor, David H. Evans, and my committee member Richard D. Miles. All work described in this dissertation was approved by the Institutional Animal Care and Use Committee at the Univer sity of Florida (protocols D995 and E822).

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES.......................................................................................................................10 ABSTRACT...................................................................................................................................12 CHAPTER 1 INTRODUCTION: WHO EATS WO OD AND HOW DO THEY DO IT?.......................... 14 2 BEAVERS OF THE FISH WORLD: CAN WOOD-EATING CA TFISHES ACTUALLY DIGEST WOOD?............................................................................................ 18 Introduction................................................................................................................... ..........18 Materials and Methods...........................................................................................................23 Fish Collection.................................................................................................................23 Gut Morphology and Length...........................................................................................23 Gut pH and Redox Measurements................................................................................... 24 Histological and TEM Analyses...................................................................................... 25 Tissue Preparation for Digestive Enzyme Analyses....................................................... 26 Assays of Digestive Enzyme Activity............................................................................. 28 Gut Fluid Preparation, Gastrointestinal Fermentati on, and Lum inal Carbohydrate Profiles.........................................................................................................................33 Fiber Digestibility............................................................................................................ 34 Transit Time of Wood in the Digestive Tract................................................................. 36 Statistical Analyses.......................................................................................................... 37 Results.....................................................................................................................................38 Gut Length, pH and Redox..............................................................................................38 Histology and TEM Analyses......................................................................................... 39 Digestive Enzyme Activities........................................................................................... 39 Gastrointestinal Fermentation a nd Lumi nal Carbohydrate Profiles................................ 46 Fiber Digestibility and Gut Transit.................................................................................. 47 Discussion...............................................................................................................................49 3 CAN WOOD-EATING CATFISHES ASSI MILATE NUTRIENTS AND ENERGY FROM WOOD? INSIGHTS FR OM STABLE ISOTOPES IN THE LABORATORY AND IN THE FIELD............................................................................................................. 87 Introduction................................................................................................................... ..........87 Methods..................................................................................................................................92 Fish Collection and Main tenance in Laboratory............................................................. 92 Tissues Used for Stable Isotope Analysis....................................................................... 93

PAGE 7

7 Stable Isotope Trial 1: Initial Turnover........................................................................... 94 Stable Isotope Trial 2: Wood A ssimilation and Negative C ontrol.................................. 94 Stable Isotopic Profiles of W ild-Caught Fish and Resources......................................... 97 Sample Preparation for Mass Spectrometry.................................................................... 98 Statistical Analyses.......................................................................................................... 99 Results...................................................................................................................................102 Trial One: Initial Turnover............................................................................................102 Trial Two: Wood Assimilation and Negative Control.................................................. 103 Wild-Caught Fish and Resources.................................................................................. 105 Discussion.............................................................................................................................106 4 CONCLUSIONS.................................................................................................................. 135 LIST OF REFERENCES.............................................................................................................139 BIOGRAPHICAL SKETCH.......................................................................................................154

PAGE 8

8 LIST OF TABLES Table page 2-1 Digestive enzymes assayed in this study........................................................................... 64 2-2 Interspecific comparisons of body mass, relative gut length, gut length as a function of snout-vent length, and Zihlers index in three species of lo ricariid catfishes............... 65 2-3 pH and redox conditions in four regions of the gut of Panaque nocturnus P. cf. nigrolineatus Maraon, and Pterygoplichthys disjunctivus ..........................................66 2-4 Amylase, laminarinase, cellulase, an d xylanase activities in the intestinal fluid and microbial extracts of Panaque cf. nigrolineatus Maraon, P. nocturnus Pterygoplichthys disjunctivus and Hypostomus pyrineusi ................................................67 2-4 (continued).........................................................................................................................68 2-5 Summary of ANOVA and t -test st atistics for interspeci fic comparisons of digestive enzyme activities.............................................................................................................. ..69 2-6 Michaelis-Menten constants of disaccharidases in the gut walls and microbial extracts of the proximal intestines of Panaque cf. nigrolineatus Maraon, P. nocturnus, Pterygoplichthys disjunctivus, and Hypostomus pyrineusi ..............................70 2-7 -mannosidase and aminopeptidase activities in the gut wall and m icrobial extracts of Panaque cf. nigrolineatus Maraon, P. nocturnus, Pterygoplichthys disjunctivus and Hypostomus pyrineusi ............................................................................71 2-8 Total short chain fatty acid concentrations (mM) in three gut regions of Hypostomus pyrineusi Pterygoplichthys disjunctivus Panaque cf. nigrolineatus Maraon, and Panaque nocturnus............................................................................................................72 2-9 Nutritional compos ition of water oak ( Quercus nigra ) wood consumed by Panaque nigrolineatus and Pterygoplichthys disjunctivus in laboratory feeding trials.................... 73 2-10 Digestibilities (%) of various fractions of wood consumed by Panaque nigrolineatus and Pterygoplichthys disjunctivus in laboratory feeding trials.......................................... 74 2-11 Particle sizes of intestinal contents pr esented as the percen t of total contents for each of the proximal, mid, an d distal intestine of Panaque nigrolineatus .................................75 3-1 Overall isotopic signatur es of the pelleted al gae and artificial w ood-detritus diets fed to Ptergoplichthys d isjunctivus........................................................................................120 3-2 Resources and animals collected from the Ro Maraon, Per, and their respective stable isotopic signatures................................................................................................. 121

PAGE 9

9 3-3 Taxa collected from Wekiva Springs, Fl orida and their stable isotopic signatures......... 122 3-4 The isotopic incorporation of carbon from an algal diet (trial one) into tissues of Pterygoplichthys disjunctivus ..........................................................................................123 3-5 The isotopic incorporati on of nitrogen from an algal di et (trial one) into tissues of Pterygoplichthys disjunctivus ..........................................................................................124 3-6 Isotopic signatures ( 13C and 15N) of three tissues of wild-caught Pterygoplichthys disjunctivus collected from Wekiva Springs, Florida...................................................... 125 3-7 The isotopic incorporation of carbon from an artificial wood-detr itus diet (trial two) into tissues of Pterygoplichthys disjunctivus ...................................................................126 3-8 The isotopic incorporation of nitrog en from an artificial wood-detritus diet (trial two) into tissues of Pterygoplichthys disjunctivus ...................................................................127

PAGE 10

10 LIST OF FIGURES Figure page 2-1 Partial phylogenetic hypothe sis for three tribes in the ca tfish fam ily Loricariidae........... 76 2-2 Photographs of the digestive trac t of Pterygoplichthys disjunctivus .................................77 2-3 Histological images and transm ission elec tron microscope (TEM) micrographs of the proximal, mid, and distal intestine of Pterygoplichthys disjunctivus Panaque cf. nigrolineatus Maraon, P. nocturnus, and Hypostomus pyrineusi ................................78 2-4 Total activities (intestinal fluid + microbial extract) of amylase, lam inarinase, cellulase, and xylanase in three regions of the intestine of Panaque cf. nigrolineatus Maraon, P. nocturnus Pterygoplichthys disjunctivus and Hypostomus pyrineusi .....80 2-5 Maltase, -g lucosidase, and N-acetyl-D-glucosaminidase activities in the gut walls and microbial extracts of the proximal intestine, mid intestine, and distal intestine of Panaque cf. nigrolineatus Maraon and P. nocturnus...................................................82 2-6 Maltase, -g lucosidase, and N-acetyl-D-glucosaminidase activities in the gut walls and microbial extracts of the proximal intestine, mid intestine, and distal intestine of Pterygoplichthys disjunctivus and Hypostomus pyrineusi .................................................84 2-7 Total activities (intestinal fluid + microbial extract) of trypsin in three regions of the intestine of Panaque cf. nigrolineatus Maraon, P. nocturnus, Pterygoplichthys disjunctivus and Hypostomus pyrineusi ............................................................................85 2-8 Position of stained wood in the intes tine of Panaque nigrolineatus at different time intervals following its consumption...................................................................................86 3-1 Changes in 13C and 15N in Pterygoplichthys disjunctivus 0-203 days after a diet switch...............................................................................................................................128 3-2 Changes in 13C and 15N in Pterygoplichthys disjunctivus across 155 days while consuming an artificial wood-detritus diet or an algal diet..............................................129 3-3 Changes in 13C and 15N in plasma solutes of Pterygoplichthys disjunctivus across 155 days while consuming an artificial w ood-detritus diet or deprived of food............. 130 3-4 Changes in 13C and 15N in red blood cells of Pterygoplichthys disjunctivus that were deprived of food across 155 days............................................................................ 131 3-5 Hepato-somatic indices a nd %C:%N ratios in the livers of Pterygoplichthys disjunctivus fed an artificial wood-detritus diet, an algal diet, or those that were deprived of food...............................................................................................................132

PAGE 11

11 3-6 Carbon and nitrogen dual-isotope plots of animals and resources collected in the upper Ro Maraon, Per ................................................................................................. 133 3-7 Carbon and nitrogen dual-isotope plots of animals and resources collected Wekiva Springs, FL, USA .............................................................................................................134

PAGE 12

12 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 BEAVERS OF THE FISH WORLD: C AN WOOD-EATING CATFISHES ACTUALLY DIGEST WOOD? A NUTRITIONAL PHYSIOLOGY APPROACH By Donovan Parks German December 2008 Chair: David H. Evans Major: Zoology Because wood is composed almost entirely of structural polysaccharides (primarily cellulose) few animals can digest it, let alone th rive on it. Most animals that do subsist on wood require endosymbiotic microorganisms in their di gestive tracts to aid in the digestion of structural polysaccharides. Several species of catfishes (family Loricariidae) from the Amazonian basin (South America) were recen tly described as being wood-eating, or xylivorous. However, beyond some cursory anal yses of gut contents and digestive enzyme activities, little is known about these animals and whether they can digest wood. In this dissertation I explored the structur e and function of the digestive tract in four species of loricariid catfishes: three xylivorous species ( Panaque cf. nigrolineatus Maraon, P. nocturnus, and Hypostomus pyrineusi ) and one detritivorous species ( Pterygoplichthys disjunctivus ) that represents the common feeding mode of the fam ily. Thus, I was able to examine whether the xylivorous species had specializations of the diges tive tract affording them the ability to digest wood in comparison to the non-xylivor ous, detritivorous species. I measured several aspects of the fishes gut morphology, including intestinal folding patterns, microvilli surface area, pH and redox potentials, and microbial diversity in different regions of the gut. I also measur ed the activity levels of 14 digestive enzyme s in the guts of the

PAGE 13

13 fish, and determined the sources of thes e enzymes: endogenous (produced by the fish) vs. exogenous (produced by microbes). These data we re then compared to concentrations of microbial fermentative end products, called shor t chain fatty acids (SCFAs), and soluble carbohydrate profiles in the intestinal fluids of the fish to determine where microbes might be most active, and where nutrient s were being hydrolyzed and absorbed. And finally, I measured the ability of captive fish to digest wood, to grow on it, and how quickly the fish pass wood through their diges tive tracts. The results of these analyses suggested th at, unlike termites, the alleged xylivorous catfishes of the Amazon cannot digest the fibrous components of wood in their digestive tracts. I found no evidence that they harbored endosymbion ts in their guts capab le of digesting the structural polysaccharides of wood. And, the labora tory feeding trials showed that the fish could not assimilate significant amount s of fiber or energy from wood, resulting in the fish losing weight on a wood diet. The fish es entire feeding st rategy, ranging from intake, to gut passage rates, digestive enzyme activities, intestinal morphology, soluble car bohydrate profiles, and levels of SCFAs throughout the gut suggest that the fish eat as much as they can, pass it through the digestive tract quickl y, and assimilate the soluble, non-fibr ous components available to them. Unlike many other wood-eating animals (e.g., term ites, beavers), these fishes consume decaying wood in aquatic systems; decaying wood is in th e process of being degraded by microbes, which produce soluble degradation products (e.g., -glucosides) that the fish can actually digest and assimilate. Thus, rather than harboring endos ymbiotic microorganisms to digest wood fiber within their guts, the fish rely on microbial deco mposition occurring in the environment. In this vein, the wood-eating catfishes are actually detritivores like so ma ny other loricariid catfishes.

PAGE 14

14 CHAPTER 1 INTRODUCTION: WHO EATS WO OD AND HOW DO THEY DO IT? W ith 1012 metric tons produced annually, cellulose is the most abundant organic molecule in the biosphere (Wilson and Irwin 1999; Karas ov and Martnez del Rio 2007). Cellulose is the major structural polysaccharide in the cell wa lls of most photosynthetic organisms and in the sheaths of tunicates, and is found in nearly all habitats. Despite its ubiquitous distribution and overall abundance, relatively few animals can di gest it in their alimentary tracts. The -1,4linkage between adjacent glucose molecules makes cellulose particularly resistant to hydrolysis; thus, it is considered refractory or recalcitrant to digestion. In fact, digestion of cellulose requires a specialized set of en zymes, collectively called cellulases, to be degraded. Although many microorganisms (encompassing the phylogenetically disparate groups of bacteria, fungi, and protists) possess genes for thes e cellulase enzymes, relatively few animals have these genes (Watanabe and Tokuda 2001; Lo et al. 2003). Thus, many herb ivorous animals do not digest cellulose in plants they consume, and those that do mostly require the aid of symbiotic microorganisms to extract energy from ce llulose (Karasov and Martnez del Rio 2007). Of all the photosynthetic organisms on the planet, woody plants pr oduce more cellulose on a proportional basis than any gro up of herbaceous plants or alga e; 90% of a trees biomass is made of cell wall components, primarily cellulo se (Karasov and Martnez del Rio 2007). Given the recalcitrance of cellulose to di gestion, wood is, therefore, a diffi cult resource for an animal to rely on to meet its daily energetic needs. Thus, few animals target wood as their primary food source and are considered xylivorous. The re liance on wood as a food resource is limited to a few families of insects, one family of molluscs, two lineages of mammals, and two genera of fishes. Among insects, silverfish, cockroaches woodroaches, lower termites, higher termites, beetles, and wood wasps are known to be able to digest wood and s ubsist on it (Prins and

PAGE 15

15 Kreulen 1991). Shipworms (actually a family of bivalves) represent the only molluscan family to be xylivorous (Xu and Distel 2004), and b eavers and porcupines, both of which are large rodents, are the only mammals that are known to be able to extract sufficient amounts of energy from wood (Vispo and Hume 1995; Felicetti et al. 2000). Recently, several new species of Amazonian catfishes (genus Panaque family Loricariidae) were described as xylivorous, with wood composing the only macroscopic material in the fishes intestines (Schaefer and Stewart 1993). The mechanisms of digestion in most xylivorous animals have been described, at least on a basic level; most use symbiotic microorganisms to digest cell ulose (Prins and Kreulen 1991; Vispo and Hume 1995; Felicetti et al. 2000; Xu and Distel 2004). However, beyond the measurement of cellulase activity in the guts of Panaque maccus (Nelson et al. 1999), little is known about the digestive tracts of the alleged wood-eating catfishes. Ve rtebrate animals do not possess cellulase genes (Lo et al. 2003) a nd are, therefore, ab solutely reliant upon microorganisms to aid in the digestion of cellu lose (Stevens and Hume 1998). For example, beavers and porcupines possess microorganisms in their intestines that ferment cellulose and other refractory polysaccharides (Vispo and Hume 1995), producing byproducts called short chain fatty acids (SCFAs), which the animals then absorb and use as an energy source (Bergman 1990; Stevens and Hume 1998). Thus, the expectation would be that xylivorous catfishes function similarly to beavers and porcupines with a reliance on endosymbiotic fermentation to digest and aassimilate wood. The family to which the wood-eating catfishes belong, the Loricariidae is diverse, with over 680 described species in 80 genera, and is Neot ropical in its distribu tion (Armbruster 2004). Most species appear to be herbivorous or de tritivorous (Nelson et al. 1999; Delariva and

PAGE 16

16 Agostinho 2001; Pouilly et al. 2003 ), although xylivory evolved twice in the family, as species in the genus Hypostomus (Armbruster 2003) and Panaque (Schaefer and Stewart 1993) are considered xylivorous. However, given that ma ny loricariid catfishes co nsume low-quality food rich in cellulose (i.e., detritus), not too dissimilar from decaying wood, how different are the wood-eating catfishes from other de tritivorous loricariid catfish es? Do wood-eating catfishes possess specializations similar to beavers and por cupines allowing them to consume and digest wood, or are these wood-eating fishes simply detri tivores that specialize on a form of detritus (i.e., wood) that is ubiquitous in th e forested Amazonian basin? The focus of this dissertation was to investig ate the structure and f unction of the digestive tracts of wood-eating catf ishes to determine whether they can actually digest the cellulose in wood, or whether they are detritivores, like so ma ny other loricariid catfis hes, and rely more on souble components of their food. In chapter two I describe a detailed suite of analyses of the gut structure and function of three wood-eating species, representing both xylivorous genera within the Loricariidae and a generalized detritivore that represents the most common feeding mode of the family. Specifically, I examined the structure of the fishes digestive tracts by looking at the gross gut morphology, the folding patterns of the intestine with histolog ical staining, and the surface of the cells lining the gut with transmi ssion electron microscopy (TEM). To determine the conditions of the gut milieu I measured the pH and redox conditions of the fishes intestines to determine if any region of the gut was hos pitable for microbes to reside in and ferment cellulose. Accordingly, I investigated whether the fish used endosymbiotic microorganisms to digest cellulose by measuring con centrations of SCFAs along the fish es digestive tracts. I also used the TEM micrographs to search for conglomer ations of microbes in di fferent regions of the gut. To determine what compounds the fishes were capable of digesting I measured the activity

PAGE 17

17 levels of 14 digestive enzymes, including cel lulases, and examined where food was being digested and absorbed in their guts. And, I performed detailed feeding trials to determine whether xylivorous and detritivorous catfish could digest wood and grow on a wood diet. Thus, this study was designed to investigate, on multip le levels, the capabilit ies of the wood-eating catfishes to harbor endosymbionts in their guts and digest wood. In chapter three I explored the temporal dyna mics of stable isotopic incorporation in Pterygoplichthys disjunctivus to determine whether non-invasi vely sampled tissues plasma solutes, red blood cells, and fin ti ssue could be used to isotopi cally track the diet of wildcaught fishes. Furthermore, I took advantage of naturally occurring stab le isotopic ratios of different plant types to discern whether this detritivorous fish could assimilate carbon from wood cellulose. I then took the lessons learned from la boratory experiments and applied them to data gathered in the field. The application of this resear ch lies in the potential for biofuel production. With the current interest in cellulosic-et hanol, and the isolation of a f ungus from the guts of a wood-eating beetle capable of producing ethanol from wood (Nigam 2001), there is great potential for the discovery of new and novel microorganisms from the guts of the xylivorous and detritivorous catfishes. However, until we understand more about the structure and function of the guts of these animals, this remains a potential rather than a resource. Thus, the po int of this dissertation is to reveal whether xylivorous and detritivorous catfishes rely on endogenous or exogenous digestive mechanisms to gain nutrition from their food.

PAGE 18

18 CHAPTER 2 BEAVERS OF THE FISH WORLD: C AN WOOD-EATING CA TFISHES ACTUALLY DIGEST WOOD? Introduction The consumption of wood for food is rare am ong animals. Unlike th e greener portions of plants, woody tissues are m ade of cells that are dead at functional ma turity, and hence, are lacking the nutritional cell conten ts on which many herbivorous animals thrive. Because wood is comprised almost entirely of structural polysacch arides (e.g., lignocellulose), it is considered nutrient poor (Karasov and Mart nez del Rio 2007). Thus, ma ny wood-eating, or xylivorous, animals (e.g., lower termites, b eavers) require the aid of symb iotic microorganisms in their alimentary tracts to digest cellulose and make the energy in this compound available to the host (Prins and Kreulen 1991; Vispo and Hume 1995) Indeed, a common theme among xylivorous animals is that they possess an expanded hindgut or caecum in which microbes reside and produce cellulolytic enzymes to aid in the digest ion of woody material (Prins and Kreulen 1991; Vispo and Hume 1995; Mo et al. 2004). Because the conditions in this expanded hindgut are typically anaerobic, microbial endosymbionts operate under fermentative pathways, reducing glucose (and other monomers) to by-products called short chain fatty acids (SCFAs; e.g., acetate), which are then absorbed by the host anim al and used to generate ATP (Bergman 1990; Karasov and Martnez del Rio 2007). In 1993, Schaefer and Stewart described seve ral new species as part of a lineage of Neotropical catfishes, genus Panaque family Loricariidae, that appeared to be xylivorous. The enlarged teeth these animals use to scrape wood from the surface of fallen trees in the river, and the presence of wood as the onl y macroscopic material in the fi shes digestive tracts intrigued the authors (Schaefer and Stewart 1993). What s more, xylivory evolved twice in loricariid catfishes, as a clade in the genus Hypostomus (formerly Cochliodon ; Armbruster 2003) is

PAGE 19

19 recognized as wood-eating in addition to the Panaque (Figure 2-1). Both xylivorous clades are derived within the larger phylogeny of the fam ily. Although there is some knowledge of the phylogenetic history of loricariid catfishes (Armbruster 2004), in cluding the wood-eating lineages (Schaefer and Stewart 1993, Armbrust er 2003), little is know n of the digestive physiology in these fishes, and whether th ey can digest cellu lose from wood. The catfish family to which these alleged xyliv ores belong, the Loricariidae, is incredibly diverse, with 680 described species in 80 genera, and is entirely Neotropi cal in its distribution (Armbruster 2004). Although some authors have co mmented that all lorica riids are herbivorous or detritivorous (Nelson et al. 1999), the diets of relatively fe w species are known (e.g., Delariva and Agostinho 2001; Pouilly et al. 2003), and app ear to include animal, plant, and detrital material from the benthos. It is clear, however that these fishes have undergone evolutionary rearrangements of jaw structure, allowing fo r diversity in feeding modes and trophic specialization (Schaefer and Lauder 1986). Furthermore, lori cariids have extremely long, coiled intestines (Delariva and Agostinho 2001), which suggests they have high levels of intake of lowquality food (Sibly and Calow 1986; Horn a nd Messer 1992; Karasov and Martnez del Rio 2007), such as detritus (Ara ujo-Lima et al. 1986). In the only investigation of digestive physiol ogy in xylivorous catfishes and in loricariids in general, Nelson et al. (1999) examined digestive enzyme activ ities and cultured microbes from the digestive tracts of Panaque maccus and an undescribed species of Pterygoplichthys (formerly Liposarcus ; Armbruster 2004), both of which they obtained via the aquarium trade. Nelson and colleagues were able to isolate microbe s with cellulolytic capabilities from the guts of the two species, and they were able to meas ure cellulase activities in the fishes guts. However, the finding of cellulolytic microbes in the guts of the fish does not mean that those

PAGE 20

20 microorganisms are endosymbionts digesting wood. For example, grass carp, which eat aquatic macrophytes rich in cellulose, ha ve cellulase activities in thei r guts (Das and Tripathy 1991) and an active microbial population (Trust et al. 1979 ; Lesel et al. 1986), ye t poorly digest the cellulose component of their plan t diet (Van Dyke and Sutton 1977). This is likely due to rapid gut transit and low levels of mi crobial fermentation in the gra ss carp guts (Stevens and Hume 1998). What is clearly needed is an understanding of the gut st ructure and function of the woodeating catfishes, including a tradit ional fiber digestibility investig ation, to determine whether the xylivorous catfishes can digest cellulose from wood and subsis t on it. Moreover, such an investigation should be conducted in a theoreti cal context to better understand exactly how the guts of these fishes function. That is, do their guts act more like those of a terrestrial hindgut fermenter (Penry and Jumars 1987; Breznak and Brune 1994; Vispo and Hu me 1995; Felicetti et al. 2000), with some mechanism for slowing th e flow of digesta and allowing microbes to ferment structural polysaccharides (Clements and Raubenheimer 2006; Karasov and Martnez del Rio 2007), or are their guts more similar to those of other detritivorous fishes (Horn and Messer 1992; Crossman et al. 2005 ; German 2008), with rapid gut transit and little digestion of structural polysaccharides? These divergent digestive strategies not only fe ature differences in digesta transit rate and gut morphology, but also produce completely different profiles of digestive enzyme activities and SCFA concentrations along the gut (Horn and Messer 1992; Jumars 2000; German 2008). For example, an animal with hindgut fermentation would be expected to show retention of small particles in the hindgut fermentative region (P arra 1978; Vispo and Hume 1995), to have high concentrations of SCFAs in that region (Vispo and Hume 1995; Mountfort et al. 2002; Pryor and Bjorndal 2005), and to have high activities of microbially produced digestive enzymes in the

PAGE 21

21 hindgut (e.g., carrageenase, Skea et al. 2005; cell ulase, Potts and Hewitt 1973; Nakashima et al. 2002; Mo et al. 2004). Thus, if xylivorous catfishes are reliant upon an endosymbiotic community to digest wood in their digestive tract s, they should display these patterns. I examined the gut structure and function of xylivorous and detrivorous loricariid catfishes to determine what traits of the digestive tract, if any, these animals have for digesting a diet rich in refractory polysaccharides. I traveled to the Ro Maraon in northern Per, where xylivorous catfishes are most diverse and abundant (Sch aefer and Stewart 1993), and collected animals directly from their natural habitat. In all, I collected two species from the genus Panaque ( P. nocturnus Schaeffer and Stewart 1993, and an undesc ribed species I am tentatively calling P. cf. nigrolineatus Maraon), representing the two clades of this genus, and one species of Hypostomus ( H. pyrineusi Ribeiro 1920) representing the ot her clade of xylivorous catfishes (Figure 2-1). All of these taxa are sympatric in the Ro Maraon. Additionally, I made use of an introduced population of a detritivorous loricariid, Pterygoplichthys disjunctivus Weber 1991, which has been living in Florida for nearly two decades (Nico 2005). This study, therefore, included both clades of xylivorous catfishes, and a less-derived detritivore from the same family (Figure 2-1). Thus, I was able to examine gut st ructure and function in th ese fishes in dietary and phylogenetic contexts. This study had five main components. I first ex amined the fishes gut structure, including the morphology of the digestive tract, to look for the presence of any kinks, valves, or caeca that might serve as a refuge for microbial endosym bionts (Pryor and Bjor ndal 2005; Pryor et al. 2006). I also measured the length of the gut (German and Horn 2006), examined the gut ultrstructure with histological and electron microscopical technique s, and qualitatively examined the surface area of the intestine. Second, I measured the pH and redox conditions along the

PAGE 22

22 digestive tract to determine whether any portion of the gut would be hospitable for microbes to operate under anaerobic conditions. Third, I measured the biochemical activity levels of 14 digestive enzymes acting in the gut lumen or alo ng the brush border of the intestine that reflect the ability of the fish to hydrolyze substrates commonly encountered in wood or detritus (Table 2-1). Following the methodology of Skea et al. ( 2005), I measured enzyme activities relative to location along the gut and determined whether the sources of these enzyme activities were endogenous (host-produced) or exogenous (pr oduced by microorganisms). This was done by collecting three fractions from the gut sections: gut wall ti ssue (endogenous), gut fluid (enzymes secreted either by the fish or microorganisms) a nd microbial extract (exogenous). If the fishes were relying on endosymbionts to digest cellulose I would expect cellula se activities to be highest where the microbes are most densely pop ulated in the gut. Fourth, luminal carbohydrate profiles and SCFA concentrations were measured along the diges tive tract to determine where nutrients were being hydrolyzed and absorbed, and where microbes might be most concentrated. Transmission electron micrographs were also used to examine mi crobial diversity in different regions of the gut. And fifth, two types of feed ing trials were performed to determine whether the fish could actually digest wood fiber, how quickly wood passes th rough their guts, and whether there is selective reten tion of particles (large or small) anywhere along the digestive tract. Overall, this study was designed to exam ine, on multiple levels, the capabilities of these fishes to harbor endosymbionts and digest wood. Given the dearth of information available on loricariid digestion, I adopted a null hypothesis for all queries, and expected to find little evidence of microbial digestion and elaborations of the fish digestive tract for harboring endosymbionts.

PAGE 23

23 Materials and Methods Fish Collection Twenty five adult individuals each of Panaque cf. nigrolineatus Maraon and P. nocturnus, and seven adult individuals of Hypostomus pyrineusi were captured by seine and a backpack electroshocker from the upper Ro Maraon in northern Peru (4.957 S, 77.283 W) in August 2006. Thirty four individuals of Pterygoplichthys disjunctivus were captured by hand while snorkeling from the Wekiva Springs complex in north central Florida (28.321 N, 81 .464 W) in March 2006. Upon capture, fishes were placed in coolers of aerated river water and held until euthanized (l ess than two hours). Fishes were euthanized in buffered water containing MS-222 (1 g l-1), measured [standard length (SL) 1 mm], and dissected on a chilled (~4oC) cutting board. Guts were removed by cu tting at the esophagus a nd at the anus and processed in a manner appropriate for specific anal yses (see below). All handling of fish from capture to euthanasia was conducted under approved protocol D995 of th e Institutional Animal Care and Use Committee of the University of Florida. Gut Morphology and Length Gu ts from 11 P. cf. n. Maraon, 13 P. nocturnus, and 13 P. disjunctivus were removed from the fish, uncoiled, and measured. These gut s were then immediately frozen and saved for gut content analyses. Gut content analyses were performed following the techniques described by German et al. (2008). However, to save space, the results of this part of the study will not be shown. Briefly, the gut cont ents of the species of Panaque contained wood (55%), amorphous detritus (40%), and di atoms (5%), whereas Pt. disjunctivus guts contained amorphous detritus (45%), algae (30%; green filamentous and blue-gr een algae), diatoms (15%), unidentified insect parts (5%), and sediment (5%).

PAGE 24

24 The measured guts were used to calculate th ree digestive somatic indices commonly used in studies of fish feeding ecology (German and Horn 2006): relative gut length (RGL = gut length/standard length); gut length as a function of snout-vent length (GL/SVL = gut length/snout-vent length), where snout vent length is the meas urement on the ventral surface from the tip of head to the anus; and Zihler s index [Zihler 1982; ZI = gut length/(10 x body mass1/3)], which relates gut length to body mass. These indices allow for the comparison of gut length among fishes with different diets while controlling for differences in body size (Kramer and Bryant 1995; German and Horn 2006). Gut pH and Redox Measurements Upon dissection, the complete digestive tracts of f our individuals each of P. cf. n. Maraon, P. nocturnus and Pt. disjunctivus were placed on a sterilized, stainless-steel dissection tray at ambient temperature (22-25 C) and gently uncoiled without tearing or stretching. The pH and redox conditions of th e digestive tracts were measured following Clements et al. (1994) with calibrated pH and redox microelectrodes (models PHR-146S and ORP-146, respectively; Lazar Laborat ories Inc., Los Angeles, CA, USA) connected to a portable pH-redox meter (model 601A, Jenco Inc., San Die go, CA, USA). Incisions large enough to allow penetration of the microelectrode tip (~0.25 mm) into the gut fluid were made in the stomach and intestinal wall and the pH and re dox conditions were measured immediately after each incision was made. Overall, pH and redox cond itions were measured in five sections of the stomach and 10 sections each of the proximal, mid, and distal intestine of each individual fish. The mean pH and redox conditions were then determ ined for each region of the digestive tract in an individual fish, and mean values determined for each gut region for each species.

PAGE 25

25 Histological and TEM Analyses Upon remo val from the body, the digestive tracts of two individuals of each species were immediately placed in ice-cold Trumps fixati ve [4% formaldehyde, 1% glutaraldehyde, in 10 mM sodium phosphate (monobasic) and 6.75 mM sodium hydroxide; McDowell and Trump 1976], pH 7.5, to prevent any degradation of the gut ultrastructure. The guts were gently uncoiled while submerged in the fixative, and the length of the intestine (IL 0.05 mm) was measured with calipers. Six 1-mm sections were excised from each of the proximal, mid, and distal intestine and placed in their own indivi dual vials containing fres h Trumps fixative, and kept cool (~4-10C) for transport back to the University of Florida. Three of the sections were designated for analysis with tr ansmission electron microscopy (TEM), whereas the other three were designated for use in histol ogical analyses and light microscopy. Upon arrival at the laboratory (approximately three weeks), tissues were removed from the fixative and rinsed in 0.1 M phosphate buffered saline (PBS), pH 7.5, for 3 x 20 min, and a final rinse overnight at 4C. Following rinsing in PBS, the tissues designated for histological analyses were rinsed for 40 min in running deio nized water and dehydrated in a graded ethanol series. The samples were then impregnated in two changes of Citrisolve for 20 min each, and infiltrated in four changes of paraffin (TissuePrep 2, Fisher Scientific, Fair Lawn, NJ, USA) for 30, 45, 60, and 60 min in a vacuum oven at 60C. The tissues were then embedded in paraffin at 57C and the blocks stored at room temperature before use. In testinal tissues were serially sectioned at 7 m, stained in a modified Massons trichrome (Presnell and Schreibman 1997), and photographed at 40X, 100X, and 400X with a H itachi KP-D50 digital camera attached to an Olympus BX60 bright-field light microscope. Imag es (n=5 per intestinal region, per individual fish; 30 images per species) were used to qualita tively examine the gut stru cture of the fish.

PAGE 26

26 Following the rinsing in PBS, those tissues desi gnated for TEM were then postfixed in 1% osmium tetroxide for 12 h at 4C, and rinsed in running deionized water for 40 min. The tissues were then dehydrated in a graded ethanol series, followed by a graded acetone series at room temp (22C), and embedded in Spurrs resin (T ed Pella Inc., Redding, CA USA). Resin blocks were cut into 1-mm thick sections using a Reic hert-Jung Ultracut-E microtome (Jena, Germany). The sections were stained with 1% toluidine blue, and examined under a bright-field light microscope (Olympus BX60) to find sections with appropriate intestinal folds (Horn et al. 2006). Ultrathin sections (70 nm) were then cut with a diamond blade from the same central part of each selected mucosal fold, mounted on honeycomb copper grids (Pelco 8GC 180 or 270, Ted Pella) and stained with 1% uranyl aceta te and 2% lead citrat e. Cross sections of 5 enterocytes with undistorted (i.e., cylindrical) microvilli were photographed us ing a transmission electron microscope (H-7000, Hitachi, Japan). Images ( n = 15 per intestinal region per individual fish; 90 images per species) were used to qualitativel y assess how the surface area of the intestine changes from the proximal to the distal ends. Tissue Preparation for Digestive Enzyme Analyses For fishes designated for digestive enzyme an alyses, guts were diss ected out, placed on a sterilized, chilled (~4C) cu tting board, and uncoiled. The stomachs were excised, and the intestines divided into three sections of equal length representing the pr oximal, mid, and distal intestine. The gut contents were gently squeezed from each of the three intestinal regions with forceps and the blunt side of a r azorblade into sterile centrifuge vials. These vials (with their contents) were then centrifuged at 10,000 x g for 5 min (Skea et al. 2005) in an Eppendorf 5415R desktop centrifuge powered by a 12V car batt ery via a power inverter. Following centrifugation, the supernatants (her etofore called intestinal fluid ) were gently pipetted into a separate sterile centrifuge vials, and the pelleted gut contents and intestinal fluid were frozen in

PAGE 27

27 liquid nitrogen. Gut wall sections were collected from each intestinal region of each specimen by excising an approximately 30 mm piece each of th e proximal, mid, and distal intestine. These intestinal pieces were then cut longitudinally, an d rinsed with ice-cold 0.05 M Tris-HCl buffer, pH 7.5, to remove any trace of intestinal contents. The entire liver and hepatopancreas were also excised from each animal. The gut wall sections livers, and hepatopancreas were placed in sterile centrifuge vials and frozen in liquid nitrogen. All of the samples were then transported on dry ice back to the University of Florida where they were stored at oC until analyzed. The intestinal fluids and pelleted gut conten ts were homogenized on ice following Skea et al. (2005). Intestinal fluids were defrosted, diluted 5-10 vol umes in 0.05 M Tris-HCl, pH 7.5, and gently homogenized using a Polytron homog enizer (Brinkmann Instruments, Westbury, NY) with a 7-mm generator at a setti ng of 1100 rpm for 30 s. The intest inal fluid samples were then stored at C in small aliquots (100-200 L) until use. To ensure the rupture of microbial cells and the complete release of enzymes from the gut contents, the pelleted gut contents were defrosted, diluted 3-5 volumes in 0.05 M Tris-HCl pH 7.5, sonicated at 5 W output for 3 x 20 s, with 40-s intervals between pulses, and hom ogenized with the Polytron homogenizer at 3000 rpm for 3 x 30 s. The homogenized pelleted gut contents were then centrifuged at 12,000 x g for 10 min at 4C, and the resulting supernatant designated microbial extract. Gut wall, liver, and hepatopancreas samples were homogenized according to German et al. (2004). Gut wall sections were defrosted, dilu ted in 5-100 volumes of 0.3 M mannitol in 0.001 M Hepes/NaOH (Martnez del Rio et al. 1995; Leve y et al. 1999), pH 7.0, homogenized with the Polytron homogenizer at 3000 rpm for 3 x 30 s, and centrifuged at 9,400 x g for 2 min at 4C. The liver and hepatopancreas samples were treate d in the same manner, with the exception that they were diluted 3-10 volumes with 0.05 M Tris-HCl, pH 7.5. Following centrifugation, the

PAGE 28

28 supernatants from the pelleted gut contents (microbial extract), the gut wall sections, liver, and hepatopancreas samples were collected and stored in small aliquots (100-200 l) at oC until just before use in spectrophotometric assays of activities of digestive enzymes. The protein content of the homogenates was measured usi ng bicinchoninic acid (S mith et al. 1985) as detailed by German (2008). Assays of Digestive Enzyme Activity All assays we re carried out at 25oC, consistent with meas ured temperatures (24-26oC) of the Ro Maraon, in triplicate using the BioRad Benchmark Plus microplate spectrophotomer and Falcon flat-bottom 96-well microplates (Fisher Sc ientific). All pH values listed for buffers were measured at room temperature (22oC), and all reagents were purchased from Sigma-Aldrich Chemical (St. Louis). All reactions were run at saturating substrate concentrations as determined for each enzyme with gut tissues from the four species. Each enzyme activity (Table 2-1) was measured in each gut region of each individual fi sh, and blanks consisting of substrate only and homogenate only (in buffer) were conducted simultaneously to account for endogenous substrate and/or product in the tissue homogenates and subs trate solutions (Skea et al. 2005; German et al. 2008). Polysaccharidase activities (i.e., activities against starch, laminarin, cellulose, mannan, and xylan) were measured in the intestinal fluid, pelleted gut contents, liver, and hepatopancreas according to the Somogyi-Nelson method (N elson 1944; Somogyi 1952). Polysaccharide substrate was dissolved [starch (2%), laminari n (0.5%), carboxymethyl cellulose (0.5%), or mannan (0.5%)] or suspended (xylan, 0.5%) in 0.8 M sodium citrate buffer, pH 7.5. In a microcentrifuge vial, 50 l of polysaccharide solution was combined with 50 l of a mixture of sodium citrate buffer and intestinal fluid, tissue, or microbial extract homogenate. Homogenate

PAGE 29

29 volumes ranged from 1-30 l depending on the enzyme concentr ation in the homogenates. The incubation period varied with subs trate the assays were carried out for 10 min for starch, two hours for laminarin, each in a water bath, and 24 hours for each of carboxymethyl cellulose, mannan, and xylan, under constant shaking on a rotary shaker in an incubator. The incubation was stopped by adding 20 l of 1 M NaOH and 200 l of Somogyi-Nelson re agent A. SomogyiNelson reagent B was added after the assay solu tion was boiled for 10 min (see German et al. 2004 for reagent recipes). The resulting solution was diluted in water and centrifuged at 6,000 x g for 5 min. The reducing sugar content of the solution was then determined spectrophotometrically at 650 nm, and polysaccharidase activity was determined from a standard curve constructed with the respective monome r (i.e., glucose for starch, laminarin, and carboxymethyl cellulose; mannose for mannan; and xylose for xylan). Enzyme activities are expressed in U (1 mol reducing sugar liberated per minute) per gram wet weight of fluid, tissue, or content. Maltase activity was measured in gut wall tissues and pelleted gut contents following Dahlqvist (1968) as descri bed by German (2008). In a microcentrifuge tube, 10 L of 56 mM maltose dissolved in 100 mM maleate buffer, pH 7.0, was combined with 10 L of regional gut wall or pellet homogenate. After 10 min, th e reaction was stopped by the addition of 300 L of assay reagent (Sigma GAGO20) dissolved in 1 M tris-HCl, pH 7.0. The reaction mixture was incubated for 30 min at 37 oC, and was stopped by the addition of 300 L of 12 N H2SO4. The amount of glucose in the solution was then dete rmined spectrophotometrically at 540 nm. The maltase activity was determined from a gluc ose standard curve and expressed in U (1 mol glucose liberated per minute) per gram wet wei ght of gut tissue or pe lleted contents. The

PAGE 30

30 Michaelis-Menten constant (Km) for maltase was determined for gut wall and pelleted gut content samples with substrate concen trations ranging from 0.56 mM to 112 mM. Tris is known to be an inhibitor of malta se activity (Dahlquist 1968), but in higher concentrations (e.g., 1 M; Levey et al. 1999) than those used in our homogenate buffer (0.05 M). Nevertheless, to confirm that th e different buffers used for th e gut wall (Hepes-Mannitol) and microbial extract (Tris-HCl) homogenates did not directly affect the Km or activity for maltase, the gut walls and pelleted gut contents of the proximal intestine of five additional P. disjunctivus were homogenized in the opposite buffers gut walls in Tris-HCl and pell eted gut contents in Hepes-Mannitol. For maltase, the different buffers did not produce different Km (Tris-HCl: 7.72 1.91 mM; Hepes-Mannitol: 7.97 0.99 mM; t =0.10, p=0.92, d.f.=10) or activity (Tris-HCl: 20.74 4.76 U g tissue-1; Hepes-Mannitol: 12.62 1.66 U g tissue-1; t =1.38, p=0.20, d.f.=10) values in the microbial extract, or Km (Tris-HCl: 4.98 0.72 mM; Hepes-Mannitol: 3.87 0.58 mM; t =1.20, p=0.26, d.f.=10) or activity (Tris-HCl: 2.05 0.41 U g tissue-1; Hepes-Mannitol: 2.44 0.37 U g tissue-1; t =0.70, p=0.50, d.f.=10) values in the gut wall homogenates. The lowconcentration Tris-HCl was observed to have li ttle effect on maltase activity in two previous investigations (German et al. 2004, German 2008) in which the gut tissues were homogenized in 0.05 M Tris-HCl buffer. The different buffers also did not affect the Km and activity levels of the other disaccharidases measured in this study (see below), and thus I can be confident that any differences in Km and enzyme activity among the gut wall and pelleted gut content homogenates are not due to the different buffers used in their homogenization. The activities of the disaccharidases -glucosidase, -mannosidase, -xylosidase, and Nacetyl-D-glucosaminidase (NAG) were measured in gut wall tissues and pelleted gut contents using p-nitrophenol conjugated subs trates (Nelson et al. 1999; Xie et al. 2007) dissolved in 0.1M

PAGE 31

31 sodium citrate, pH 7.0. In a microplate well, 90 L of 11.1 mM substrate (1.33 mM for NAG) was combined with 10 L of gut wall or pelleted gut conten t homogenate and the reaction was read kinetically at 405 nm for 15 min. The disacc haridase activities were determined from a pnitrophenol standard curv e and expressed in U (1 mol p-nitrophenol liberated per minute) per gram wet weight of gut tissue or pelleted contents. The Km was determined for gut wall and pelleted gut content samples for -glucosidase and NAG. The s ubstrate concentrations ranged from 0.002 mM to 10 mM for -glucosidase and 0.04 to 1.2 mM for NAG Chitinase activities we re measured following German et al. (2008), but no activity was detected in the four species used in this study. In all assays, the background levels of N-acetylglucosamine detected in the blanks (>1 mM) matched what was measurable in the assay mixtures, making activity determinations impossi ble. However, the measurable N-acetylglucosamine in the gut in addition to measurable NAG activities makes it likely that the fish can utilize chitin as a nutrient source. Trypsin activity was assayed in the intestin al fluid and pelleted gut contents using a modified version of the method designed by Erlange r et al. (1961) as described by Gawlicka et al. (2000). The substrate, 2 mM N -benzoyl-L-arginine-p-nitroanilide hydrochloride (BAPNA), was dissolved in 100 mM tris-HCl buffer (pH 7.5) by heating to 95oC (Preiser et al. 1975; German et al. 2004). In a microplate, 95 l of BAPNA was combined with 5 l of homogenate, and the increase in absorbance was read continu ously at 410 nm for 15 min. Trypsin was also assayed in the liver and hepat opancreas, but tissues homogenates from these organs were first incubated with enterokinase for 15 min to activate trypsinogen prio r to combining the homogenates with substrate (German et al. 2004). Trypsin activity was determined with a p-

PAGE 32

32 nitroaniline standard curv e, and expressed in U (1 mol p-nitroaniline liberated per minute) per gram wet weight of tissue, gut fluid, or pelleted gut contents. Aminopeptidase activity was measured in gut wall tissues and pe lleted gut contents according to Roncari and Zuber (1969) as described by German et al. (2004). In a microplate, 90 L of 2.04 mM L-alanine-p-nitroanilide HCl dissolved in 200 mM sodium phosphate buffer (pH 7.5) was combined with 10 L of homogenate. The increase in absorbance was read continuously at 410 nm for 15 min and activity determ ined with a p-nitroaniline standard curve. Aminopeptidase activity was expressed in U (1 mol p-nitroaniline liberated per minute) per gram wet weight of gut tissu e or pelleted gut contents. Lipase (nonspecific bile-salt activated E.C. 3.1.1.-) activ ities were assayed in the intestinal fluids and pelleted gut contents us ing a modified version of the method designed by Iijima et al. (1998). In a microplate, 86 l of 5.2 mM sodium cholate dissolved in 250 mM trisHCl (pH 7.5) was combined with 6 l of homogenate and 2.5 l of 10 mM 2-methoxyethanol and incubated at room temperature for 15 min to allow for lipase activation by bile salts. The substrate p-nitrophenyl myristate (5.5 l of 20 mM p-nitrophenyl my ristate dissolved in 100% ethanol) was then added and the increase in abso rbance was read continuously at 405 nm for 15 min. Lipase activity was determined with a p-n itrophenol standard curve, and expressed in U (1 mol p-nitrophenol liberated per minute) per gram wet weight of gut tissue. The activity of each enzyme was regre ssed against the protein content of the homogenates to confirm that there were no signif icant correlations betwee n the two variables. Because no significant correlations were observed, th e data are not reported as U per mg protein.

PAGE 33

33 Gut Fluid Preparation, Gastrointestinal Ferm entation, and Luminal Carbohydr ate Profiles Measurements of symbiotic fermentation activ ity were based on the methods of Pryor and Bjomdal (2005). Fermentation activity was indica ted by relative concentrations of short-chain fatty acids (SCFA) in the fluid contents of the guts of the fishes at the time of death. As homogenates were prepared from the intes tinal fluid samples (see above under Tissue preparation for digestive enzyme analyses), 30 L of undiluted intestinal fluid was pipetted into a sterile centrifuge vial equipped with a 0.22 m cellulose acetate filter (Costar Spin-X gamma sterilized centrifuge tube filters, Coming, NY) and centrifuged under refrigeration at 13,000 x g for 15 min to remove particles from the fluid (including bacterial cells). The filtrates were collected and frozen until they were analy zed for SCFA and nutrient concentrations. Concentrations of SCFA in the intestinal fluid samples from each gut region in each species were measured using gas chromatograp hy as described by Pr yor et al. (2006) and German et al. (2008). Glucose co ncentrations were analyzed in 2 L of gut fluid using the same glucose content assay described for the maltase a ssay above. The only depa rture being that there was no pre-incubation with maltose; the gut fluid was immediately combined with the assay reagent and incubated at 37 oC for 30 min, the reaction stopped with 12N H2SO4, and the resulting mixture read in a sp ectrophotometer at 540 nm against a glucose standard curve. To examine the presence of reduc ing sugars of various sizes in the intestinal fluids of the fish, 1 L of filtered intestinal fluid was spotted on to pre-coated silica gel plates (Whatman, PE SIL G) together with standards of glucose, maltose, and trito penta-oligosa ccharides of glucose. The thin layer chromatogram (TLC) was develo ped with ascending solvent (isopropanol/acetic acid/water, 7:2:1 (v/v)) and stained with thymol reagent (A dachi 1965; Skea et al. 2005).

PAGE 34

34 Fiber Digestibility To evaluate whether Pt. disjunctivus and a wood-eating loricariid catfish, Panaque nigrolineatus, could digest wood fiber, I performe d a tr aditional fiber diges tibility feeding trial using a total collection method (Galetto and Bellwood 1994). Five individuals of Pt. disjunctivus (mean S.D.; 221.88 9.83 mm SL; 214.44 23.09 g BM) were collected from Wekiva Springs, FL, in September 2007 and brought back to the University of Florida. Seven individuals of P. nigrolineatus (176.43 20.73 mm SL; 227.76 78.71 g BM) were obtained from an aquarium wholesaler (5-D Tropical, Ta mpa, FL) as the fish arrived from Venezuela, where they were captured from their native habitat. Individuals of both species were individually assigned to 75.6-L a quaria equipped with a 2.5 cm plas tic mesh at the bottom, which allowed feces and uneaten food (orts) to fa ll through and be undisturbed by the fish. The intake tube of the mechanical filter was covered with 250m mesh screen to prevent uneaten orts and feces from being sucked into the filter. Every two days the fish were given a new piece (~100 g, wet weight) of waterlogged, degraded water oak (Quercus nigra) wood on which to graze. The fish were allowed to acclimate to these conditions for at least one month prior to the beginning of the experiment. Once the experiments were started, they were executed in two-week intervals. Two twoweek feeding trials (i.e., a total of four weeks) were performed for P. nigrolineatus, and three two-week trials were performed for Pt. disjunctivus to obtain enough fecal material for all of the analyses. Each afternoon (~15:00) the fish we re gently removed from their tanks into an individual bucket containing aquarium water. The plastic mesh was removed from the bottom of the tank, and the feces, which were completely di stinguishable from the orts, were siphoned off with a 25 mL bulb pipette into a weigh boat. Or t debris was siphoned off with a piece of vinyl

PAGE 35

35 tubing onto a 250m mesh screen, from which it was scrape d with a razorblade into weigh boats for each tank. The feces and debris were dried at 60C for 24 hours, weighed, and stored in sealed glass vials. On a daily basis the wood pieces provided to each fish were blotted with a paper towel, weighed (wet mass), and a small s ubsample was scraped off with a razorblade. The wood was then weighed again before being returned to the tank. Daily diet samples, as well as daily feces and ort samples for an individual fis h, were combined to obtain a composite sample for each of the diet, feces, and orts across th e feeding trial (Bouchard and Bjorndal 2006). Wood and ort samples were ground to pass thro ugh a 1-mm screen in a coffee grinder, and fecal samples were ground with a mortar and pestle. Fecal, ort, and diet samples were analyzed for dry matter, organic matter, neutra l detergent fiber (NDF), acid detergent fiber (ADF), lignin+cutin, and %N (Bouchard and Bj orndal 2006). Ort analysis allowed us to examine whether the fishes fed selectively on more nutritious portions of the wood. Dry matter and ash (mineral) content were de termined by drying subsamples overnight at 105C and then combusting them at 550C for 4 h. The difference between these two measures represents the organic matter component of the sample. NDF and ADF were determined by sequentially refluxing samples with neutral detergent and acid de tergent solutions (Goering and Van Soest 1970) in an Ankom200 Fiber Analyzer according to the guidelines supplied with the equipment (Ankom Technology 1998; Ankom Technology 1999). NDF repres ents the cell-wall component of the wood (cellulose, hemicellulo se, and lignin), and ADF represents the lignocellulose component. The actual lignin+cutin portion was determin ed by refluxing the samples in 72% sulfuric acid for 3 hours at room temp (24C; Moran and Bj orndal 2007). The N content of the samples was determined using a Carlo Erba elemental analyzer.

PAGE 36

36 Daily consumption rates, on a dry matter basis, were calculated from a regression of wet weight vs. dry weight of water oak wood pieces mi nus the mass of orts. Thus, we were able to calculate digestibility coefficients based on total intake and fecal output. Digestibility was determined using the equation (int ake-feces)/intake, where intake is total grams of dry matter or organic matter consumed during the trial and fece s is grams of dry matter or organic matter in the feces produced during the trial. Transit Time of Wood in the Digestive Tract One of the most important elem ents in studies of digestion is how long food is held in the digestive tract, as this provides key information into the strategy an animal takes to digest a meal (Karasov and Martnez del Rio 2007). At the conc lusion of the fiber dige stibility experiment, P. nigrolineatus and Pt. disjunctivus were fed water oak wood that had been stained red with carmine dye (Fris and Horn 1993). The wood pi eces were submerged in a 2% carmine dye solution for at least one month. They were then pre-steeped under agitation (3x10 min) in aquarium water to remove loose dye particle s and ensure that only fully stained wood was offered to the fish. Wood pieces (~100 g) were placed in tanks as described above under Fiber digestibility only the tanks were without th e plastic mesh at the bottom. Because P. nigrolineatus and Pt. disjunctivus are nocturnal, the fish were given the stained wood an hour before the lights turned off in the aquari um laboratory, and the fish usually took 0.5-1 h to adjust to the darkness and start feeding (DPG, pers. obs.). Four hours after the commencement of feeding, two individuals were taken and euthanized in MS-222 as described above. After this initial four hour feeding peri od, the stained wood was replaced with non-stained wood in the tanks of the remaining fish to allow for the tr acking of the stained w ood (i.e., the pulse) through the gut. Two additional fish were taken and euthan ized at six and eight hours, and the remaining

PAGE 37

37 fish at 18 hours post feeding. At each sampling interval, the presence or absence of red-stained feces in the aquaria was noted. At each sampling interval the fish were dissected on a sterile cutting board, and the gut was removed, uncoiled, measured, and photographed. The red dye on the wood was visible in the gut contents of P. nigrolineatus, but not in Pt. disjunctivus mainly because the latter species produces an inordinate amount of b ile that is nearly black in colo r. Thus, I was unable to address particle retention in Pt. disjunctivus and only have the approximate amount of time taken for red feces to appear in the aqua ria for this species. In P. nigrolineatus, the intestine was divided into three sections of equal length and the proximal-mo st location of the red stained wood in the gut was noted at each sampling interval. The intestinal contents were squeezed from each intestinal section with forceps and the blunt si de of a razor blade into sterile centrifuge vials and frozen at 80C until analyzed (~one mont h). When analyzed, the samples were defrosted, and to determine whether specific particle sizes of the digesta were bei ng selectively retained along the intestine, the contents were wet sieved (Vis po and Hume 1995) using mesh sizes ranging from 0.25 to 1.5 mm. Following the sieving, the contents were dried at 60C for 24 hours, and the samples of the various particle size classes were weighed. The masses of all of the particle size classes were then added together, and the proporti on of each particle size class was determined. This design also allowed me to ascertain whether particles (stain ed or non-stained) of any size were held longer in one region of the digestive tract than another. This experiment was repeated one month later with an additional seven P. nigrolineatus obtained from the same aquarium wholesaler. Statistical Analyses Prior to all significance tests, a Levenes test for equal vari ance was performed and residual versus fits plots were ex amined to ensure the appropriateness of the data for parametric

PAGE 38

38 analyses. All tests were run using SPSS (ver sion 11) and Minitab (ver sion 12) statistical software packages. The various digestive-soma tic indices were compared among species with ANCOVA, using body mass as a covariate, followed by a Tukeys HSD with a family error rate of P = 0.05. The activities of amylase, lamina rinase, cellulase, and xylanase were compared between the intestinal fluid and microbial extract fractions of each gut region in each species with t -test, using a Bonferroni correction. Intersp ecific and intraspecific (i.e., among gut regions in a single species) comparisons of total enzymatic activities (intestinal fluid + microbial extract) and total SCFA concentrations were made with ANOVA followed by a Tukeys HSD with a family error rate of P = 0 .05. The activities of maltase, -glucosidase, N-acetyl-Dglucosaminidase, -mannosidase, and aminopeptidase were compared between the gut wall and microbial extract fractions of each gut region in each species with t -test, using a Bonferroni correction. Similarly, the Km values of maltase, -glucosidase, and N-acetyl-Dglucosaminidase from the proximal intestine of the fish were compared between the gut wall and microbial extract fractions of each species with t -test. Results Gut Length, pH and Redox The digestive tracts of P. cf. n. Maraon, P nocturnus, Pterygoplichthys disjunctivus, and Hypostomus pyrineusi were very similar in morphology. All four species have extrem ely long, thin-walled intestines that exceed 10x the body length of the animal (Table 2-2; Figure 22), but Pt. disjunctivus possesses a significantly longe r gut than either of the Panaque species (Table 2-2). I was only able to measur e the gut length of two individuals of H. pyrinuesi and thus, they were not included in the statistical an alyses. But, the digestive somatic indices in this species (RGL = 11.85; GL/SVL = 20.13; ZI = 40.03) were more similar to the species of

PAGE 39

39 Panaque than to Pt. disjunctivus The pH of the digestive tracts of P. cf. n. Maraon, P. nocturnus, and Pt. disjunctivus were all neutral, whereas th e redox conditions of the stomach were positive ( Pt. disjunctivus ) or less negative (P. cf. n. Maraonand P. nocturnus ), and the redox conditions of the intestines of all three spec ies were extremely negative (Table 2-3). Thus, the guts of the three species were aerobic or slightly anaerobic in the stomach region, and extremely anaerobic along the intestine. Histology and TEM Analyses The digestive tra cts of the four loricariid catfish species were very similar on the histological and transmission elect ron micrograph levels (Figure 2-3). All four species showed the same pattern of decreasing height of intest inal folds, and decreasing microvilli surface area moving distally along the intes tine. Furthermore, no conglom erations of microbes were observed anywhere along the in testines of the four species (Figure 2-3). Digestive Enzyme Activities No differences were observed in amylase, lam i narinase, or cellulase activities between the intestinal fluid and the microbial extracts of any species (Table 2-4). However, xylanase activity was significantly greater in the microbial extracts of the proximal and mid intestine of P. nocturnus than in the intestinal fluids of these regions. Total amylase activity was significantly greater in the proximal intestine than in the distal intestine of all four species (Table 2-5, Figure 2-4). Panaque nocturnus possessed significantly lower amylase activity in its proximal intestine than P. cf. n. Maraon and H. pyrineusi but not significantly lower than Pt. disjunctivus. In the mid intestine, P. nocturnus possessed lower amylase than Pt. disjunctivus and H. pyrineusi but not lower than the mid intestine of P. cf. n. Maraon (Table 25, Figure 2-4). No differences were detected among the species fo r amylase activity in the distal intestine.

PAGE 40

40 Laminarinase activity was significantly higher in the proximal intestine of all four species than in their mid or distal intestines (Table 2-5, Figure 2-4). No la minarinase activity was detected in the distal intestines of P. nocturnus and H. pyrineusi Pterygoplichthys disjunctivus possessed significantly higher lamina rinase activity in its proximal intestine than in this gut region of P. cf. n. Maraon and H. pyrinuesi but not P. nocturnus (Table 2-5, Figure 2-4). In turn, P. nocturnus and Pt. disjunctivus possessed greater laminarina se activity in their mid intestines than P. cf. n. Maraon and H. pyrineusi Pterygoplichthys disjunctivus and H. pyrineusi exhibited significantly higher cellulase activity in their proximal intestines than in their mid or distal intestines ( H. pyrineusi lacked detectable cellulase activity in its distal intestine), whereas the two species of Panaque showed no difference in cellulase activity along the gut (Table 2-5, Figure 24). Individuals of Pt. disjunctivus possessed significantly greater cellulase ac tivity in their proximal intestines than individuals of P. nocturnus but not greater than P. cf. n. Maraon or H. pyrineusi And, Pt. disjunctivus exhibited higher cellulase activit ies in its mid intestine than H. pyrineusi but not higher than the species of Panaque (Table 2-5, Figure 2-4). Individuals of P. cf. n. Maraon, Pt. disjunctivus, and H. pyrineusi possessed significantly greater xylanase activity in their prox imal intestines than in their mid or distal intestines (like cellulase, H. pyrineusi lacked detectable xylanase activ ity in its distal intestine). Panaque nocturnus, on the other hand, showed a slight increase in xylanase activitiy moving distally along the intestine, albeit not a sign ificant increase (Table 2-5, Figure 2-4). No significant differences were observed in the activ ity levels of xylanase among the species for any gut region. No mannanase activity was detect ed in any gut region of any species.

PAGE 41

41 The maltase activity in the microbial extrac t was significantly higher than the activity of this enzyme in the gut wall of the proximal intestin es of all four species (Figures 2-5 and 2-6). No significant differences were observed in the mid intestine. The maltase activity in the gut walls of the distal intestines of the wood-eating taxa was higher than the maltase activity of the microbial extract, whereas the opposite was true for the detritivorous Pt. disjunctivus (Figures 25 and 2-6). All four species showed decreasi ng maltase activities in the microbial extract moving distally along the in testine, whereas all f our taxa showed slight increases in gut wall maltase activity in the mid intestine in comparison to the proximal intestine (Figures 2-5 and 26). Hypostomus pyrineusi possessed significantly greater maltase activity in its gut wall fractions (PI: ANOVA F3,22 = 21.84, P < 0.001; MI: ANOVA F3,22 = 55.34, P < 0.001; DI: ANOVA F3,22 = 11.20, P < 0.001, Tukeys P < 0.011 for all) than the other species, which did not differ from one another. Similarly, significant differences were detected among the microbial extract maltase activities of the proximal intestines of the species (ANOVA F3,26 = 32.55, P < 0.001; Tukeys P < 0.011); H. pyrineusi possessed significantly greater ma ltase activity in its proximal intestine microbial extract than all of the other species, and Pt. disjunctivus exhibited greater maltase activity than the species of Panaque which did not differ from one another. A similar pattern was found for the mid intestine microbial extract (ANOVA F3,26 = 20.30, P < 0.001; Tukeys P < 0.011), with H. pyrineusi and Pt. disjunctivus possessing significantly greater maltase than the two species of Panaque There were no differences among the species in the microbial extract maltase activities of the distal intestine. The -glucosidase activities in the microbial extracts of the proximal intestines of P. cf. n. Maraon, P. nocturnus, and Pt. disjunctivus were all significantly higher than the activities of this enzyme in the gut wall frac tions, however the opposite was true for H. pyrineusi (Figures

PAGE 42

42 2-5 and 2-6). Only Pt. disjunctivus showed significant differences in -glucosidase activity in their mid and distal intestines, with the gut wall activity being significantly higher in the mid intestine, and the activity in the microbial extract being higher in the distal intestine. All four species showed decreasing -glucosidase activity in the microbial extracts of their distal intestines (Figures 2-5 and 2-6). However, th ere were several different patterns for gut wall glucosidase activity: P. nocturnus and H. pyrineusi showed decreasing activ ity in their distal intestine, P. cf. n. Maraon showed increasing activity towards their distal intestine, and Pt. disjunctivus showed a spike in activity in the mid intes tine, followed by a decrease in the distal intestine. Like maltase, H. pyrineusi possessed significantly greater -glucosidase in the gut wall fraction of their proximal intestines than in th e other species (ANOVA F3,22 = 49.68, P < 0.001; Tukeys P < 0.011), which did not differ from one another. Pterygoplichthys disjunctivus and H. pyrineusi exhibited significantly greater -glucosidase activity in the gut wall fractions of their mid intestines than P. nocturnus, but not P. cf. n. Maraon, which in turn was not different from P. nocturnus Individuals of P. cf. n. Maraon possessed significantly greater glucosidase activity in the gut wall of their distal intestine than in the distal intestine gut wall of Pt. disjunctivus ( t = 2.62, P = 0.026, df = 10); P. nocturnus and H. pyrineusi lacked glucosidase activity in the gut wall fractions of their distal intestines. No significant differences were detected in the -glucosidase activities of the proximal and mid intestine microbial extracts of any species. However, P. cf. n Maraon and Pt. disjunctivus exhibited significantly greater -glucosidase activity in the microbial extracts of their distal intestines than in P. nocturnus (ANOVA F3,26 = 4.35, P = 0.014; Tukeys P < 0.011) but not greater than in H. pyrineusi which did not differ from P. nocturnus

PAGE 43

43 Panaque nocturnus exhibited significantly greater N-acetyl-D-glucosaminidase (NAG) activity in the gut wall of its proximal intestine than in the microbial extract, whereas none of the other species showed differences in NAG activity between these two fracti ons in their proximal intestines (Figures 2-5 and 2-6). The wood-ea ting taxa all exhibited significantly higher NAG activity in the gut walls of their mid intestines than in the mi crobial extracts from this gut region, whereas Pt. disjunctivus showed no differences between the two fractions. However, P. nocturnus and Pt. disjunctivus had significantly greater NAG activ ity in the gut walls of their distal intestine gut walls than in their microbi al extracts, whereas the other species showed no differences between the two frac tions (Figures 2-5 and 2-6). Panaque cf. n. Maraon, P. nocturnus, and Pt. disjunctivus showed increases in their gut wall NAG activities moving distally along the intestine, whereas H. pyrineusi showed a decrease. The NAG activities of the microbial extracts were variable and didnt follow one pattern (increase or decrease) along the guts of any of the four species (Figures 2-5 and 2-6). Pterygoplichthys disjunctivus exhibited significantly greater NAG activity in its proximal intestine gut wall than in P. nocturnus (ANOVA F3,22 = 4.97, P = 0.010; Tukeys P < 0.011), but no t greater than the other species, which did not differ from one another. The mid intestine gut wall NAG activities were not different among the species, but Pt. disjunctivus displayed significantly greater distal intestine gut wall NAG than the other species (ANOVA F3,22 = 15.39, P < 0.001; Tukeys P < 0.011), which did not vary from one another. No diffe rences were detected among the species for NAG activity in the proximal intestine microbial extracts, but Pt. disjunctivus had significantly higher NAG activity in its mid intestine microbial ex tract than in the other species (ANOVA F3,26 = 26.37, P < 0.001; Tukeys P < 0.011), which did not differ from one another. Pterygoplichthys disjunctivus showed significantly greater distal in testine microbial extract NAG activity than P.

PAGE 44

44 cf. n. Maraon and P. nocturnus (ANOVA F3,26 = 5.25, P = 0.007; Tukeys P < 0.011), but not greater than H. pyrineusi The maltase Michaelis-Menten constants (Km) from the wall of the proximal intestines of the fish were generally lower, alt hough not significantly so, than the Km values of the microbial extracts from the proximal intestines (Table 2-6). However, the Km values of -glucosidase were all significantly lower in the fish gut walls than in the microbial extracts, and the same is generally true for NAG, except for in P. nocturnus (Table 2-6). Pterygoplichthys disjunctivus had a significantly higher Km for maltase in the gut wall fraction than any of the xylivorous species (ANOVA F3,22: 9.61, P <0.001; Tukeys P <0.011), wh ich did not differ from one another. No differences in the maltase Km values of the microbial extracts were detected among the species (ANOVA F3,26: 1.80, P = 0.176; Tukeys P <0.011). The two species of Panaque possessed significantly lower Km values for -glucosidase in their gut walls than in Pt. disjunctivus or H. pyrineusi (ANOVA F3,22: 10.28, P <0.001; Tukeys P <0.011), which did not differ from one another. And, P. cf. n. Maraon possessed significantly lower -glucosidase Km in its microbial extract than Pt. disjunctivus and H. pyrineusi (ANOVA F3,26: 7.28, P = 0.001; Tukeys P <0.011), but not P. nocturnus, and the microbial extract -glucosidase Km of the three remaining species were not statisti cally different from each other. Panaque cf. n. Maraon displayed significantly lower NAG Km in its gut wall than any other species (ANOVA F3,22: 6.60, P = 0.003; Tukeys P <0.011), which were not different from one another. Both species of Panaque exhibited significantly lower Km values for their microbial extract NAG than that of Pt. disjunctivus (ANOVA F3,26: 4.27, P = 0.015), but not of H. pyrineusi All four species generally possessed significantly greater -mannosidase and activities in their gut walls than in the microbial ex tracts (Table 2-7). The activity of -mannosidase

PAGE 45

45 increased in activity moving di stally along the intestine of P. cf. n. Maraon, decreased in activity moving distally along the intestines of P. nocturnus and Pt. disjunctivus and spiked in the mid intestine of H. pyrineusi Panaque nocturnus possessed significantly greater mannosidase activity in the gut wall of its proxi mal intestine than in the other species (ANOVA F3,22: 24.72, P <0.001; Tukeys P <0.011), which didnt differ from one another. Panaque nocturnus also showed greater -mannosidase activity in the gut wa ll of their mid intestine than P. cf. n. Maraon and Pt. disjunctivus (ANOVA F3,22: 8.36, P = 0.001; Tukeys P <0.011), and H. pyrineusi exhibited greater activity than Pt. disjunctivus, which did not differ from P. cf. n.Maraon. However, P. cf. n. Maraon displayed significantly higher -mannosidase activity in the gut wall of its distal intestine than in Pt. disjunctivus and H. pyrineusi but not P. nocturnus (ANOVA F3,22: 4.97, P = 0.010; Tukeys P <0.011). The -mannosidase activity in the gut wall of the distal intestine of the other taxa did not differ from one another. All four species generally po ssessed significantly greater ami nopeptidase activities in their gut walls than in the microbial extracts (Table 2-7). Aminopeptidase activities increased in activity moving distally along the intestine in the wood-eating taxa, and spiked in the mid intestine of Pt. disjunctivus (Table 2-7). No significant differe nces were observed in the gut wall aminopeptidase activities of the proximal and mid intestine among the species. However, the three wood-eating species possess ed significantly higher aminopep tidase activities in the gut walls of their distal intestin es than in this region of Pt. disjunctivus (ANOVA F3,22: 5.17, P = 0.009; Tukeys P <0.011). -xylosidase activity (not shown) was low and only observed in the microbial extracts of the four taxa. The activity of this enzyme, like xylanase, genera lly decreased in activity moving distally along the intestine.

PAGE 46

46 Trypsin activities significantly decreased moving distally alo ng the intestines of all four species (Figure 2-7). Panaque nocturnus possessed significantly greater trypsin activity in their proximal intestine than all of the other species, whereas the tw o remaining wood-eating species ( P. cf. n. Maraon and H. pyrineusi ) possessed similar and significan tly greater trypsin activity in their proximal intestines than in Pt. disjunctivus (Figure 2-7). No diffe rences were detected among the species in their mid or distal intestines. Lipase activities (not shown) followed similar patterns to trypsin and decreas ed moving distally along the in testines of the fish. The trypsin:lipase activity ratios for all fish exceeded 400:1. Enzymatic activities of the hepa topancreas and liver (not show n) varied by enzyme. No cellulase or xylanase activities were detected in the hepatpancreas or liver of any species, whereas amylase, laminarinase, tryp sin, and lipase were all detected in the hepatopancreas of the fish. Only amylase and lipase were de tectable in the liver. Gastrointestinal Fermentation and Luminal Carbohydrate Profiles Overall, the concentrations of SCFAs were low in the digestive tracts of all four species of catfish (Table 2-8). Only H. pyrineusi showed any significant change in SCFA concentration along the gut, with significantly higher SCFA concentrations in the mid intestine than in the proximal intestine. The trends of SCFA concentrations varied among species, with Pt. disjunctivus showing an increasing concentr ation of SCFAs along the gut, and P. cf. n. Maraon showing a decrease, albeit no significan t change (Table 2-8). The TLC plates (not shown) revealed that all four species had soluble oligo-, dia nd monosaccharides in the proximal intestine, and that these concentrations decrease d until there were no soluble sugars remaining in the distal intestine. Similarly, measurable glucose was observed in the fluid of the proximal intestine of P. cf. n. Maraon (2.70 0.29 mM) and P. nocturnus (2.86 0.38 mM), but these concentrations disappeared in the mid and distal intestine. Only H. pyrineusi showed measurable

PAGE 47

47 glucose in all regions of the intestine and these concentratio ns decreased, significantly so (ANOVA: F2,14 = 84.75, P < 0.001), from the proximal (4.98 0.43 mM) to the mid (0.93 0.09 mM) to the distal (0.73 0.03 mM) intestine. No glucose was detected in the fluid of any gut region of P. disjunctivus Fiber Digestibility and Gut Transit The wood I offered to the fish in the laborat ory wa s almost entirely organic matter (97%), and moderately rich in lingocellulose (~60% ; Table 2-9). Because I observed no feeding selectivity on specific ty pes of wood in the wild in Per, degraded wood of any riparian tree seemed appropriate for this part of the study. Neither P. nigrolineatus nor Pt. disjunctivus readily assimilated large proportion s of the dry or organic matter, or of the fiber types of the water oak (Table 2-10). However, because th e organic matter digestibilities are apparent digestibilities (because the fish contribute orga nic waste, like sloughed in testinal cells, to the feces) and the fiber digestibilitie s represent true digestibilities (because the fish contribute no fiber to the feces) it is difficult to match the calculations for the digestibilities for the two fractions to each other. Nevertheless, the digestibilties for NDF and ADF each contributed approximately the same amount of overall organi c matter digestion (Table 2-10). For example, 22% digestibility of NDF that composes 86% of the total organic matter equals a total digestibility of 20%. Similarly, 24% digestibility of ADF that composes 61% of the total equals a total digestibility of 15%. Thus, most of the fiber dige stibility can be accounted for via ADF and lignin digestion. Furthermore, the lignin:cell ulose ratio (Abril and Bu cher 2002) of the feces (0.90 0.16) was not significantly greater than the lignin:cellulose ra tio of the wood (0.81 0.08; t = 0.49, P = 0.63, d.f. = 20), suggesting that the fish were unable to assimilate significant amounts of cellulose from the wood diet.

PAGE 48

48 The digestibility data are further confounded by the presence of proportionally more ash in the feces than in the wood (12% ash in the feces vs. 3% ash in the wood, overall). On a daily basis two to five times more ash was found in the feces than was consumed with the wood in the first place. This addition of inorganic material to the feces is mysterious and may have artificially inflated the levels of fiber digestibility observe d in this study. The fish consumed 2-5% of their body mass ( on a wet weight basis) in wood per day, but were not thriving on it, as P. nigrolineatus lost 1.8 0.15 % of thei r body mass over the course of the experiment, and Pt. disjunctivus lost 8.4 0.81 % of their body mass. This stands in contrast to a 41% mass gain by Pt. disjunctivus on an algal diet in the laboratory (DPG, unpubl. data). Furthermore, P. nigrolineatus and Pt. disjunctivus excreted more nitrogen in their feces than they consumed in the wood (Table 2-10), an d this excretion was significantly greater for Pt. disjunctivus than for P. nigrolineatus. Wood traversed the digestive tr acts of the two species extrem ely quickly, with red stained wood appearing in feces less than four hours after its consumption. Furthermore, there appeared to be no retention of the stained wood along the gut of P. nigrolineatus (Figure 2-8). The proportion of particles (stain ed or non-stained) <250 m in diameter in each region of the intestine at each time interval were as follows (mean SEM; n=3-4): 4 hours PI: 47.59 1.66, MI: 25.44 1.50, DI 22.53 2.77; 6 hours PI: 44.09 3.61, MI: 22.33 2.33, DI 17.26 2.66; 8 hours PI: 45.21 7.74, MI: 28.60 4.28, DI 23.31 3.63; 18 hours DI: 19.05 1.08 (digesta was only present in the DI at the 18 hour interval). Thus, no sele ctive retention of small particles was observed in this study. Additionally, the overall analysis of particle size in the guts of P. nigrolineatus revealed that small, more digestible particles were not retained anywhere along the gut, with particles >350 m making up more of the total moving distally along the

PAGE 49

49 intestine (Table 2-11). However, the results coul d be interpreted to show selective retention of larger particles (>350 m in diameter) in the mid and distal intestine. Discussion The data gathered in this study overwhe lming s upport the null hypothesis that the xylivorous loricariid catfishes are actually detritivores and do not digest the fibrous components of wood in their alimentary tracts. E ach of the analyses provided evidence that the fish do not exhibit specialized gut anatomy for harboring endosy mbionts: no kinks, valves, or caeca are present anywhere along their long, narr ow intestines; the mi crovilli surface area decreases moving distally along the intestine, in dicating that most absorption takes place in the proximal and mid intestine, which was corr oborated by the luminal carbohydrate profiles; no conglomerations of microbes were observed in the TEM micrographs; th e fish clearly have enzymatic activities geared for th e assimilation of soluble compone nts of their diet; they lack significant amounts of gastro intestinal fermentation anywhere w ithin their digestive tracts; they pass wood through the gut too quickly (< four hours) for microbial digestion of cellulose; and the fish do not retain small particles anywhere along their digestive tract. Furthermore, the catfish were unable to digest wood and thrive on it in the laboratory. Each of these components would be expected to be the opposite in an animal that digests wood via an endosymbiotic community of microbes living in their guts. However, loricari id catfishes certainly have interesting digestive tracts (the longest among all fishes measured to date; Horn 1989; Kram er and Bryant 1995) and subsist on detritus in the wild, which they do appear suited to digest (Bow en et al. 1995; German 2008; German et al. 2008). There has been some debate over the last 15 years as to whether xylivorous catfishes can digest wood. Schaefer and Stewart (1993) suggested that species in the genus Panaque could

PAGE 50

50 be capable of extracting energy from wood and this assertion has been assu med to be true ever since, especially on the internet and among aquari um fish enthusiasts. However, the one study published to date examining digestion in species of Panaque and Pterygoplichthys provided only inferential evidence of cellulose di gestion (Nelson et al. 1999). Th e data gathered in the current study systematically refute that wood-eating species in the genera Panaque or Hypostomus nor the detritivorous Pt. disjunctivus have the capability to digest and subsist on wood. Two of the key factors c ontributing to this inability are th e lack of specialized gut anatomy and rapid gut transit. The lori cariid catfishes examined in this study clearly have long, narrow, anatomically unspecialized intestines, and no conglomerations of microbes were observed in the fishes guts. However, unlike terrestrial herbiv ores and xylivores, a specialized gut anatomy is not a prerequisite for fishes to harbor an endosymbiont community in their guts. Many herbivorous fishes with active endosymbiotic communities and high levels of SCFA production and assimilation (Mountfort et al. 2002) have anatomically unspecialized digestive tracts (e.g., Odax pullus and O. cyanomelas ; Clements and Choat 1995; Clem ents and Raubenheimer 2006). Odax pullus and O. cyanomelas do, however, possess voluminous guts, and O. pullus has relatively long retention of digesta in the alimentary tract (12-20 hours), mainly because of low gut contractility (Clements and Rees 1998). This low contractility is due to the very thin musculature surrounding the intestine of this sp ecies (Clements and Rees 1998), and also may result in significant amounts of axial mixing of digesta in the gut, an important component of microbially-mediated digestion (Horn 1989; Ho rn and Messer 1992; Karasov and Martnez del Rio 2007). Xylivorous termites have extremely long digesta transit cons idering their size (~24 hours; Breznak and Brune 1994), po rcupines consuming wood have mean retention times of food in the gut exceeding 34 hours (Felicetti et al. 2000), and beavers selectively retain small particles

PAGE 51

51 and fluid in their hindgut caeca (Vispo and Hume 1995). Both long retention time of digesta in the gut, and selective retention of small particles allow resident microbes to remain in the gut and digest cellulose in thes e taxa. The species of Panaque Pterygoplichthys, and Hypostomus investigated in this study ha ve relatively thick musculature surrounding their long intestines (Figure 2-3; especially when compared to O. pullus ; Clements and Rees 1998), have rapid gut transit, and no selective retention of small particles, suggest ing more of a unidirectional, plugflow movement of digesta with little axial mixing along the inte stine (Penry and Jumars 1987; Horn and Messer 1992; Jumars 2000). Furthermore, the peristaltic contrac tions of the intestine of P. nocturnus are strong and continue after the death of the animal (view video at: http://www.zoology.ufl.edu/dgerman/images/Peristalsis1.AVI ). Thus, the combination of long, narrow intes tines, rapid gut transit, and strong contractility do not support an active microbial community in the catfishes dige stive tracts. The decreasing inte stinal surface area observed in the catfishes is also consistent with other de tritivorous fishes (Frierson and Foltz 1992), which typically rely more on endogenous digestive mechanisms than on microbial endosymbionts (Bowen 1984; Smith et al. 1996; Smoot and Findla y 2000; Crossman et al. 2005; German 2008). Small particles from wood may not always be h igher quality than larg er particles. For example, in ruminant mammals, small, more indigestible particles escape the rumen following digestion, whereas larger, more digestible partic les are retained in the rumen (Van Soest 1994). Hence, the reduction of the proportion of small partic les in the mid and distal intestine of the fish in this study could indicate a ra pid movement of small indigestib le particles (perhaps rich in lignin) through their guts, and retention of larger particles in the mid and distal intestine. However, the predominance of small particles in the proximal intestine, which seemingly disappear in the mid and distal intestine, is confounded by endogenous inputs of bile and

PAGE 52

52 digestive enzymes in the proximal intestine in comparison to other gut regions. Endogenous enzyme activities, like those of amylase, lamina rinase, and trypsin are all higher in the proximal intestine (Figures 2-4 and 2-7), and these digestive secretions, along with bile, are all part of the small particle fraction (<250 m diameter) of the proximal inte stine. The disappearance of this fraction in the mid intestine may simply reflect the reabsorption of bile and endogenous digestive enzymes, which all decrease in ac tivity moving towards the distal intestine. Additionally, the disappearance of smaller particles from the pr oximal intestine could indicate digestion and assimilation of nutrients from smaller wood degradation products from the degraded wood. The low wood fiber assimilation efficiencies in the catfishes are hi ghly indicative that they cannot subsist on a wood only diet. Other xylivorous animals that have an active microbial community in their guts are capable of digesting the fibrous cell wall fraction of wood (Breznak and Brune 1994; Felicetti et al. 2000; Karas ov and Martnez del Rio 2007). For example, porcupines assimilate about 70% of NDF from wood with similar biochemical composition to that offered to the fish in this study (Felicetti et al. 2000). Panaque nigrolineatus and Pt. disjunctivus did clearly assimilate some cellulose from wood [given the small increase in the lignin:cellulose ratio (Abril and Bucher 2002) in the feces compared to the wood], and potentially some lignin, although th e latter may be an artifact of the detergent fiber analysis system (Jung 1997), which was designed for grasses (Goering and Van Soest 1970), not wood. The small change in the fecal lignin:cellulose ratio in the fish pales in comparison to the change in this ratio observed in termite feces (at least a 50% increase in the ratio in feces relative to wood; Karasov and Martnez del Rio 2007). Furthe rmore, the low DM and OM digestibilities, and net loss of N combined w ith a loss of weight while eating wood further show that P.

PAGE 53

53 nigrolineatus and Pt. disjunctivus cannot thrive on a wood diet. Individuals of Pt. disjunctivus likely did worse from weight and fecal nitr ogen loss perspectives than individuals of P. nigrolineatus beause the former lack the spoon-shaped teeth (Nelson et al. 1999) necessary to gouge wood in significant quantities (i.e., they had lower daily intake rates of wood). This is also supported by the observation that it took six weeks for Pt. disjunctivus to produce amounts of feces that P. nigrolineatus produced in four weeks. To my knowledge, no studies have evaluated the digestibility of a diet in an animal that lost weight over the course of the experiment, leaving me with no basis for comparison to the current study. Although nitrogen loss in feces has been observed for rats on a high-fiber, lowprotein diet (Jrgensen et al. 2003), there are no observations of mineral (ash) loss in feces. However, there are three possible ways for more ash to be recovered from fishes feces than from the food they consumed. First, the fish actually deposited minerals into the feces, thereby increasing the ash content. S econd, a significant proportion of the fecal organic matter was consumed by microbes over the 12 hours the feces sat in the aquaria. And third, I grossly underestimated the intake of wood by fish in th is study. In terms of th e first scenario, it is possible that the fish excreted minerals into their feces. Gonzales and Brown (2007) observed that Nile tilapia ( Oreochromis niloticus ) lost body mineral content on low-quality foods that insufficiently met the fishes daily energetic need s. Perhaps the same pattern was occurring in this study. The loricariid catfishes are character ized as having dermal plates composed of calcium phosphate (bone) on their skin (Armbrus ter 2004). Just as turtles are known to mobilize the calcium carbonate from their shells to buffe r lactate accumulation du ring periods of anoxia (Jackson 2002; Reese et al. 2004), perhaps the fish in this study, which were in negative energy balance, mobilized the phosphate from their boney plates for ATP production, and excreted the

PAGE 54

54 excess calcium in their feces. The second scenar io listed above is not likely, as the amount of ash in the feces is two to five times greater than that of the wood the fish consumed, and a microbial population could not cons ume that much organic matter in just 12 hours. For example, Galetto and Bellwood (1994) observed li ttle change in fecal organic ma tter after fish feces sat in an aquarium for 24 hours at 25C. The third scen ario is also not likely because the fish would have produced more orts on a daily basis if they had consumed more wood; the loricariids produce a large proportion of debris as they graze on wood (i.e., they are messy eaters), and if intake was higher, then there would have been mo re debris in the aquari a on a daily basis than was observed. Thus, scenario one above needs to be further investigated to determine whether the fish added inorganic material to their feces. This can be done by determining the concentration of calcium in th e feces of the fish relative to the wood they consumed. Furthermore, the mass and calcium content of th e dermal plates, as well as circulating blood calcium concentrations can be compared in fed an d starved fish to assess whether more calcium was being mobilized in fish in negative energy balance. The pH levels and redox potentials in the ca tfish intestines indicated the possibility of supporting a population of anaerobic microbes, but only in the in testine. Many loricariids breathe air and have modified stomachs that qualify as air breathing organs (ABOs; Graham and Baird 1982; Armbruster 1998), which explains why the redox potentials of the stomachs of the fish in this study were positive (Pt. disjunctivus ) or only slightly negative ( P. cf. n. Maraon and P. nocturnus). However, the loricariid stomach is not involved in digestion. For example, the stomach of Pt. disjunctivus is usually filled with air, is al kaline, and ingesta are not held in the stomach for any length of time; even individuals of this species killed minutes after consuming food had already passed the ingesta into the proximal inte stine, bypassing the

PAGE 55

55 stomach via a small groove at its base (DPG pers. obs.). Redox potentials measured 1-mm beyond the pyloric sphincter in th is study were already -600 mV, i ndicating that even the most proximal region of the intestine is not oxygenated by the fishs breathing activity. Similar to some detritivorous termites (Kappler and Brune 2002), detritivorous fish es potentially consume large proportions of humic acids, which, in additi on to other components of the intestinal fluid (e.g., bile salts, and ingested metals; Kappler and Brune 2002) can increase the reductive potential, thus producing negativ e redox conditions. Either way, the redox potentials measured in wild-caught fish in this study suggest that the intest inal environment is highly reductive. The digestion of lignin, which composes roughly 18-35% of woody material (Petterson 1984), requires oxidative (positive redox potential) c onditions (Zimmer and Brune 2005). Seeing that the stomach is the only oxidative site along the catfish gut, and food is not digested there, I do not see how it is possible for these fishes, via mi crobial endosymbionts, to digest lignin. That is, unless, similar to termites (Ebert and Brune 1997) and terrestrial isopods (Zimmer and Brune 2005), the catfish possess a radial oxygen gradient in the intestinal lumen that I was unable to detect. However, termites and isopods are extremely small, terrestrial organisms with high surface area:volume ratios in compar ison to the loricariid catfishes so I do not think this is likely. The low concentrations of SCFAs observed in the fishes intestines further challenge the hypothesis that wood-eating catfis hes use an endosymbiotic comm unity to ferment recalcitrant polysaccharides. Choat and Clements (1998) sugges ted that fish with less than 20 mM total SCFAs in the peak fermentative region of the in testine had low fermentation potential. The highest peak concentrations observed in this study were in the mid intestine of H. pyrineusi (3.20 0.79 mM), and were far below 20 mM. Furthermore, H. pyrineusi was the only species

PAGE 56

56 to show any significant difference in SCFA c oncentrations along its gut. The low SCFA concentrations coupled with the lack of any localization of SCFAs in one gut region are consistent with other herbivorous and detritivorous fishes that do not rely on gastrointestinal fermentation to digest refractory polysaccharides (Smith et al. 1996; Crossman et al. 2005; German 2008; German et al. 2008). Additionally, the loricariid catfishes are consuming detrital material that is already being degraded by mi crobes in the environmen t (Sinsabaugh et al. 1991b; Sinsabaugh et al. 1992; Tank et al 1998; Hendel and Marxsen 2000). Some of this detritus may already be in the process of fermentative digest ion when consumed by the fish (i.e., the food is prefermented), and thus, SCFAs themselves are consumed with the detritus producing the low and unchanging SCFA concentrations along the gut (German 2008). Perhaps the most informative biochemical data gathered in this study suggesting a lack of microbially-mediated digestion in the catfishes comes from the patterns of digestive enzyme activities along the fishes intestines. A co mmon pattern in lower termites, which digest cellulose in their hindgut via an endosymbiotic microbial community, is increasing cellulase activities in the hindgut region (Nakashima et al. 2002; Mo et al. 2004). Similarly, marine herbivorous fishes with active hindgut microbi al populations have increasing exogenously produced enzyme activities (e.g., carrageenase) in the microbial extracts of their hindguts (Skea et al. 2005). However, none of the catfish sp ecies showed increasing ce llulase activity in the distal intestine, and instead, showed no pattern (no increase or decrease; P. cf. n. Maraon and P. nocturnus ) or decreasing activity ( Pt. disjunctivus and H. pyrineusi ) towards the distal intestine. Moreover, the cellulase activities in the catfish guts were five orders of magnitude lower than amylase activities, and one to two orders lower than laminarinase activities. Thus, the

PAGE 57

57 fish clearly digest soluble polysac charides, like starch and laminari n, more rapidly than structural polysaccharides, especially given the rapid transit time of food through the gut. Decaying wood in an aquatic environment will likely have more nutritious dietary items collecting on the surface of the wood, and in sp aces among fibers, than the wood it self. The epilithic algal complex (EAC), which is a l oose assemblage of bacteria, cyanobacteria, filamentous green algae, diatoms, and detritus th at grows on hard substrates in aquatic systems (Hoagland et al. 1982; van Dam et al. 2002; Kloc k et al. 2007; German et al. 2008) contains soluble polysaccharides in the algae (Painter 1983) and in exopolymeric substances produced by microbes (Wotton 2004; Klock et al. 2007). Th ese soluble polysaccharides are likely an important energy source, not only to grazing species like Pt. disjunctivus but also to the xylivorous species digging into the decaying wood. Other EAC consuming fishes (e.g., species in the genus Campostoma ) have very similar patterns and magnitudes of amylase and laminarinase activities to the catfishes (German 2008; German et al. 2008), suggesting that they are targeting similar suites of nutrients from their foods. The xylanase activities in P. nocturnus were the only luminal enzyme activities to be different between the intestinal fluid and the microbial extract, and the activities of this enzyme slightly increased in activity, albeit not significantly so, towards the distal intestine of this species. Xylan is a component of hemicellulo se (Petterson 1984; Brezn ak and Brune 1994), but mammals (and probably vertebrates in gene ral) are not known to possess an endogenous xylanase, nor to be able to meta bolize the monomer of xylan, xylose, without the aid of intestinal microorganisms (Johnson et al. 2006a; Johnson et al. 2006b). Additionall y, all of the catfish species lacked a -xylosidase in their gut walls and had low activities of this enzyme in their intestinal contents that decreased moving distally al ong the digestive tract.

PAGE 58

58 Given the low and variable cellulase and xylanas e activities observed in the catfish, and the lack of any consistent pattern of activity along th e guts of the fish, it is likely that these enzymes are ingested (and produced by microbes ingested ) with detritus rather than produced by a resident endosymbiotic community, per se. This is especially true in Pt. disjunctivus and H. pyrineusi which showed decreasing ce llulase and xylanase activities moving distally along their guts. Furthermore, cellulase and xylanase activ ities were not higher in the xylivorous catfish species. For example, detritivorous Pt. disjunctivus possessed the highest cellulase activity in its proximal intestine, and xylivorous P. nocturnus the lowest. The cellulase activities measured in this st udy are three orders of magnitude lower than those reported for Panaque maccus and Pterygoplichthys sp. by Nelson et al. (1999). However, there are several methodological differences be tween this study and that performed by Nelson and colleagues. First, they used assay conditions designed for ruminant mammals (pH 5, 40C), which differ wildly from any conditions occurrin g in the fishes guts. I designed my assay conditions to reflect the fishes gut pH (pH = 7.5; Table 2-3) and ambient temperatures of their environment (25C). Second, Nelson et al. (1999) did not specify in which region of the gut they measured the enzyme activities. Third, when performing a general reducing sugar assay for polysaccharidase activity that includes intes tinal contents (as was done by Nelson and colleagues, and in this study), it is essential to perform the appropriate blanks to account not only for background reducing sugars in the gut, but also for additional substrate that may be a source of other reducing sugars released during the a ssay (Skea et al. 2005; German et al. 2008). Not doing so will result in an over-e stimation of activity levels, an d Nelson and colleagues did not perform this type of blank w ith their assays (J.A. Nelson, pe rs. comm.). And fourth, the activities were likely calculated differently be tween the two studies. All of these reasons (and

PAGE 59

59 more) have led several authors (Peres et al. 1998; Logothetis et al. 2001; Chan et al. 2004; German et al. 2004; Horn et al. 2006) to cau tion against making comparisons of digestive enzyme activities among different studies. T hus, I will not do so here, referring only to similarities in patterns of dige stive enzyme activities along the in testines of different animals. The most striking digestive enzyme activity data suggesting that the catfishes digest mainly soluble components from their detrital diet comes from the disaccharidase activities. The Michaelis-Menten constants (Km) for -glucosidase in the gut walls of the fishes were an order of magnitude lower than thos e of the microbial extracts (Table 2-6). Although the -glucosidase activities were higher in the microbial extracts than in the gut walls of the proximal intestines of P. cf. n. Maraon, P. nocturnus, and Pt. disjunctivus, this may be outweighed by the more efficient (lower Km) gut wall -glucosidase of the fish. Hypostomus pyrineusi had the double effect of lower Km and higher activity of -glucosidase in its proximal intestine gut wall. These results are important because microbes degrading the cellulose of wood in the river excrete enzymes extracellularly (Sinsabaugh et al. 1992 ; Hendel and Marxsen 2000) and depend on diand monosaccharides [such as cellobiose (a -glucoside) and glucose, respectively] to diffuse back to them that they can then further digest and assimilate (Allison and Jastrow 2006). The fishes are consuming wood detritus that is in this process of degradation, and thus, there are likely many soluble components, like cellobios e, on the decaying wood (Sinsabaugh et al. 1992; Hendel and Marxsen 2000). Because the fishes -glucosidases are more efficient than those produced by the microbes degrading the wood, th e fish quickly digest and assimilate the cellobiose in their detrital di et. Additionally, because microbe s in the environment secrete digestive enzymes extracellularly, the enzymes themselves are also likely on the detritus (Sinsabaugh et al. 1992; Hendel and Marxsen 2000) as occurs in soils (Allison 2006; Allison

PAGE 60

60 and Jastrow 2006), and are thus digested within th e guts of the fish. This may explain why the microbial extract enzyme activit ies, almost without exception (T ables 2-4 and 2-7; Figures 2-5 and 2-6), tend to decrease moving distally along the in testines of the fish. This is especially true for -glucosidase and stands in stark contra st with patterns in lower termites, as -glucosidase activities increase in th e hindguts of these taxa (McEwen et al. 1980). However, decreasing glucosidase activity moving distally along the gut ha s been observed in other detritivorous fishes (Smoot and Findlay 2000). The more efficient and higher N-acetyl-D-glucosaminidase (NAG) activity in the fish gut walls may indicate that chitin, and its degr adation products (i.e., ch itobiose), are important energy and nitrogen sources to the fish. Fungi, wh ich make cell walls of chitin, are some of the most active microorganisms in wood degradation and digestion (Swift et al. 1979; Breznak and Brune 1994; Hendel and Marxsen 2000), and are likel y consumed by the fish with wood detritus. I attempted to measure chitinase activity in the guts of the catfishes, but there was so much background N-acetyl-glucosamine which is the monomer of chitin and the endpoint of chitin digestion in the fishes guts (>1 mM), that the determination of chitinase activity was impossible using colorimetric methods. This stands in contrast to other fishes that consume chitinous arthropods, in which ch itinase activity was readily m easured (Gutowska et al. 2004; German et al. 2008). N-acetyl-glucosamine is a usable energy source for vertebrate animals (Gutowska et al. 2004), and th e presence of such large amounts of this compound in the intestines of the fish suggests that chitin digestion is proceeding rapidly, and thus, fungi may be an important dietary item of the fishes. Indeed, microbes in general may be an important nutrient source to the catfishes, which would make ly sozyme an important enzyme for nutrient acquisition in these animals (K rogdahl et al. 2005; Karasov and Martnez del Rio 2007).

PAGE 61

61 Lysozyme is important not only in bacterial cell wall degradation, but also for the degradation of chitin (Marsh et al. 2001; Krogdahl et al. 2005) in fungal cell walls. T hus, future studies of digestion in loricariid catfishes should take lysozyme activity into account, and should explore non-colorimetric methods for the determina tion of chitinase activity (e.g., release of 14C; Marsh et al. 2001). Some differences in disaccharidase activities were observed among the catfish species. Hypostomus pyrineusi possessed higher microbial extract and gut wall maltase activities than the other species. Because I did not collect enough individuals of this species on which to conduct a thorough analysis of gut contents, it is difficult speculate why they have higher maltase activities than the other species. However, cursory obser vations of their gut c ontents did reveal the presence of a large proportion of wood. Pterygoplichthys disjunctivus increases its maltase activity when consuming a diet rich in soluble polysaccharides in the laboratory (DPG, unpubl.), which might suggest that H. pyrineusi consumes wood with more EAC, and thus, more soluble polysaccharides on it than the other wood-eating speci es. The presence of measurable glucose in the intestinal fluids in all gut regions in this sp ecies, and glucose concentr ations in the fluid of the proximal intestine that were double those of the ot her xylivores further suggest that they consume more soluble polysaccharides than the othe r taxa. This is offered with the caveat that all four species had abundant soluble oligoto disaccharides in the fluid of their proximal intestines (as determined with TLC plates) that disappeared by the distal intestine. Most of the catfish specie s showed significantly higher -mannosidase activities in their gut walls than in the microbial extracts in all regi ons of the intestine. I am not sure what these activities mean for the fish, as I am aware of only one study that examined -mannosidase activity in other fish taxa: Nelson et al. (1999) found detectable activity of this enzyme in P.

PAGE 62

62 maccus and Pterygoplichthys sp. Most analyses of mannan, the products of mannan degradation, and the enzymes involved in mannan digestion, have been aimed at bacteria and fungi (Valaskova and Baldrian 2006; Moreira and F ilho 2008). Mannan, and the monomer, mannose, are major components of hemicelluloses in w ood (Petterson 1984; More ira and Filho 2008). Thus, an animal digesting wood, or woody detritu s, should possess the enzymes to digest this compound. None of the catfish species posse ssed any activity against the polysaccharide mannan. However, the detectable activity of -mannosidase in the gut walls of the fish suggest that, like -glucosidase and cellobiose, the fish may be able to digest the soluble component of mannan degradation, namely, -mannosides. This may provide yet another example of how the fish are geared for assimilating the more soluble compone nts of their detrital diet. Many animals, including herbivorous and detr itivorous fishes, feed to meet protein requirements (Bowen et al. 1995; Raubenheimer and Simpson 1998; Raubenheimer et al. 2005). The increasing aminopeptidase activiti es in the distal intestines of the catfishes likely reflect increased efforts by the fish to absorb whatever pr otein is available in their detrital diet (Fraisse et al. 1981; Harpaz and Uni 1999; German 2008), especially given the decreasing microvilli surface area of the distal intestine. Furthermore, the trypsin activ ities in the loricariid catfishes are the highest I have measured in a number of fish taxa usin g identical methodology (German et al. 2004; Horn et al. 2006; German 2008; German et al. 2008). And, as one final piece of information supporting detritivory in these fishes, the extremely high trypsin:lipase ratios (exceeding 400:1) are consistent with other detri tivorous animals in marine (Mayer et al. 1997) and freshwater (Smoot and Findlay 2000) habitats. In conclusion, loricariid catfishes in the genera Panaque and Hypostomus appear to be little more than detritivores that specialize on a rather ubiquitous form of co arse detritus in their

PAGE 63

63 environment, namely, wood. The digestive tracts of these fishes, and of a closely related nonwood-eating detritivore, Pt. disjunctivus, are clearly geared for the consumption of large amounts of low-quality food, and rapid transit of this food thr ough the gut. Enzyme activities in their alimentary tracts hydrolyze soluble components of detritus more efficiently than structural polysaccharides, and the majority of this hydrolysis takes place in the proximal and mid intestine. These patterns match well with the higher microvilli surface area in these regions of the gut. Additionally, even though the guts of th ese animals are highly reductive and hospitable for anaerobic microbes, the low SCFA concentrations throughout the fishes intestines show that they are not relying on microbial symbionts to digest structural polysaccharides via fermentative pathways. Further investigations in these fishes should emphasize th eir ability to digest bacteria and fungi, as these organisms may be likely food s ources for the fish. Nonetheless, loricariid catfishes are impressive organi sms that are highly abundant in the Amazonian basin, and the wood-eating species likely contribute heavily to nutr ient cycling in these habitats by reducing the particle size of wood from coar se debris to particles on th e scale of 1-mm in diameter.

PAGE 64

64 Table 2-1. Digestive enzymes assayed in this study of gut stru cture and function in wood-eating loricariid catfishes. Enzyme Location1 Substrate Substrate source Fractions assayed2 Amylase Lum., cont. Starch Algae, detritus Fluid, contents, HP, Liver Laminarinase Lum., cont. Laminarin Diatoms Fluid, contents, HP, Liver Cellulase Lum., cont. Cellulose Wood, algae, detritus Fluid, contents, HP, Liver Xylanase Lum., cont. Xylan Wood, detritus Fluid, contents, HP, Liver Mannanase Lum., cont. Mannan Wood, detritus Fluid, contents, HP, Liver Chitinase Lum., cont. Chitin Fungi, insects, detritus Fluid, contents, HP, Liver Trypsin Lum., cont. Protein Algae, detritus, animals Fluid, contents, HP, Liver Lipase Lum., cont. Lipid Algae, detritus, animals Fluid, contents, HP, Liver Maltase BB, cont. Maltose Algae, detritus Contents, gut wall -glucosidase BB, cont. -glucosides Algae, wood, detritus Contents, gut wall -xylosidase BB, cont. -xylosides Wood, detritus Contents, gut wall -mannosidase BB, cont. -mannosides Wood, detritus Contents, gut wall N-acetyl-Dglucosaminidase BB, cont. N-acetyl-Dglucoaminides Fungi, insects, detritus Contents, gut wall Aminopeptidase BB, cont. Dipeptides Algae, detritus, animals Contents, gut wall Notes. 1 Indicates where the enzyme is active. Lum = lumen of the intestine; cont. = contents (ingesta) of the intestine; BB = brushborder of the intestine. 2 The portions of gut content or intestinal tissue in which I assayed the activity of the enzyme.

PAGE 65

65Table 2-2. Interspecific comparisons of body mass (BM), relative gut length (RGL), gut length as a functi on of snout-vent leng th (GL/SVL), and Zihlers index (ZI) in three species of loricariid catfishes. Species (sample size) DietSL BM RGL GL/SVL ZI P. cf. nigrolineatus Maraon (11) W 87.65 9.99 29.59 8.83a 11.56 0.50a 18.97 0.96a 36.71 1.58a P. nocturnus (17) W 102.24 3.62 32.69 3.79a 11.47 0.41a 18.44 0.62a 37.53 1.28a Pt. disjunctivus (17) D 203.94 8.23 196.83 23.01b 17.24 0.55b 27.19 0.92b 61.93 1.90b Species F2,44 = 28.03 F2,44 = 55.58 F2,44 = 12.58 F2,44 = 28.01 P < 0.001 P < 0.001 P < 0.001 P < 0.001 Body Mass -F1,41 = 1.42 F1,41 = 0.72 F1,41 = 1.21 -P = 0.240 P = 0.401 P = 0.279 Note: Values are mean ( SEM). Abbreviations for diet are as follows: W = wood; D = detritus. Interspecific comparisons of BM were made with ANOVA followed by a Tukeys HSD with a family er ror rate of P=0.05. Interspecifi c comparisons of gut dimension parameters were analyzed with ANCOVA (using body mass as a cova riate) and Tukeys HSD with a family error rate of P=0.05. Values for a parameter that share a supers cript letter are not significantly different.

PAGE 66

66 Table 2-3. pH and redox conditions in four regions of the gut of Panaque nocturnus P. cf. nigrolineatus Maraon, and Pterygoplichthys disjunctivus Species Stomach Proximal Intestine Mid Intestine Distal Intestine pH P. cf. n. Maraon 7.42 0.04 7.51 0.11 7.48 0.03 7.50 0.04 P. nocturnus 8.31 0.17 8.29 0.14 8.28 0.14 8.28 0.14 Pt. disjunctivus 7.12 0.15 7.48 0.15 7.28 0.09 7.47 0.12 Redox (mV) P. cf. n. Maraon -59.38 6.33 -605.47 2.87 -611.52 5.54 -616.74 6.51 P. nocturnus -45.72 5.81 -605.81 8.91 -616.73 9.11 -608.96 8.91 Pt. disjunctivus 105.90 67.80 -595.40 4.16 -596.23 4.04 -599.63 6.39 Note: Values are mean ( SEM).

PAGE 67

67Table 2-4. Amylase, laminarinase, cellulase, and xylanase activities (U g-1) in the intestinal fluid and microbial extracts of Panaque cf. nigrolineatus Maraon ( Pm ), P. nocturnus ( Pn ), Pterygoplichthys disjunctivus ( Ptd ), and Hypostomus pyrineusi ( Hp ). Amylase Laminarinase Species (n) Proximal intestine Mid intestine Distal intestine Proximal intestine Mid intestine Distal intestine Pm (6) Intestinal Fluid 884.68 114.63 196.94 74.76 214.03 123.06 0.052 0.015 0.015 0.006 0.013 0.006 Microbial Extract 859.09 70.77 451.96 129.06 78.20 26.02 0.125 0.035 0.022 0.009 0.008 0.002 t 0.21 1.71 1.08 1.96 0.66 1.16 P 0.84 0.12 0.31 0.08 0.53 0.27 Pn (6) Intestinal Fluid 323.38 40.14 146.59 30.84 88.46 27.17 0.262 0.040 0.093 0.037 n.d. Microbial Extract 308.31 41.29 142.86 29.53 36.53 8.80 0.222 0.036 0.137 0.027 n.d. t 0.26 0.09 2.18 0.73 0.96 N/A P 0.80 0.93 0.06 0.48 0.38 Ptd (10) Intestinal Fluid 581.64 148.54 468.74 68.70 90.59 15.79 0.276 0.050 0.123 0.018 0.018 0.006 Microbial Extract 615.83 150.35 473.46 54.66 90.78 18.80 0.340 0.091 0.128 0.019 0.023 0.010 t 0.16 0.05 0.01 0.62 0.19 0.44 P 0.87 0.96 0.99 0.55 0.85 0.67 Hp (5) Intestinal Fluid 804.53 .12 522.44 122.85 105.41 33.79 0.107 0.049 0.006 0.002 n.d. Microbial Extract 1012.67 .21 533.74 98.99 85.54 18.85 0.083 0.047 0.005 0.001 n.d. t 0.95 0.07 0.51 0.34 0.68 N/A P 0.37 0.95 0.62 0.74 0.51

PAGE 68

68Table 2-4 (continued) Cellulase Xylanase Species (n) Proximal intestine Mid intestine Distal intestine Proximal intestine Mid intestine Distal intestine Pm (6) Intestinal Fluid 0.0034 0.0020 0.0013 0.0003 0.0042 0.0029 0.0051 0.0007 0.0010 0.0002 n.d. Microbial Extract 0.0034 0.0015 0.0040 0.0014 0.0016 0.0005 0.0071 0.0023 0.0031 0.0014 0.0008 0.0001 t 0.04 1.86 0.89 0.81 1.43 N/A P 0.97 0.12 0.40 0.44 0.18 Pn (6) Intestinal Fluid 0.0013 0.0003 0.0028 0.0001 0.0022 0.0016 0.0001 0.0001 0.0005 0.0002 0.0024 0.0013 Microbial Extract 0.0010 0.0001 0.0012 0.0002 0.0006 0.0002 0.0004 0.0001 0.0015 0.0003 0.0007 0.0003 t 0.85 1.56 1.22 3.33 3.32 1.61 P 0.41 0.15 0.25 <0.01 <0.01 0.14 Ptd (10) Intestinal Fluid 0.0055 0.0010 0.0041 0.0009 0.0002 0.0001 0.0064 0.0025 0.0062 0.0025 n.d. Microbial Extract 0.0091 0.0023 0.0039 0.0011 0.0005 0.0001 0.0088 0.0035 0.0052 0.0018 0.0005 0.0001 t 1.45 0.07 2.13 0.54 0.34 N/A P 0.17 0.94 0.05 0.60 0.74 Hp (5) Intestinal Fluid 0.0043 0.0012 0.0007 0.0004 n.d. 0.0037 0.0007 n.d. n.d. Microbial Extract 0.0048 0.0009 0.0003 0.0001 n.d. 0.0024 0.0014 0.0003 0.0001 n.d. t 0.31 0.98 N/A 0.81 N/A N/A P 0.76 0.36 0.44 Note: Values are mean ( SEM). Comparisons were made of the activities of each enzyme between the intestinal fluid and microbi al extract of each gut region in each species with t -test. Following a Bonferroni correction for each enzyme and species, differences are considered significant at P=0.017.

PAGE 69

69Table 2-5. Summ ary of ANOVA and t -test* statistics for interspecific comparisons of digestive enzyme activities for each of the proximal intestine (PI), mid intestine (MI), and distal intes tine (DI), and intraspecific comparisons of digestive enzyme activities among gut regions. Enzyme Interspecific comparisons by gut region Intras pecific comparisons (among gut regions) by species PI MI DI P. cf. n. Maraon P. nocturnus Pt. disjunctivus H. pyrineusi Amylase F3,26 = 4.87 F3,26 = 7.12 F3,26 = 0.93 F2,17 = 28.30 F2,17 = 23.34 F2,29 = 8.56 F2,14 = 51.68 P = 0.009 P = 0.001 P = 0.444 P < 0.001 P < 0.001 P = 0.001 P < 0.001 Laminarinase F3,26 = 3.75 F3,26 = 12.61 t = 0.93* F2,17 = 17.66 t = 2.68* F2,29 = 13.02 t = 1.96* P = 0.025 P < 0.001 P = 0.368 P < 0.001 P = 0.023 P < 0.001 P = 0.086 Cellulase F3,26 = 4.37 F3,26 = 4.08 F2,21 = 3.06 F2,17 = 0.20 F2,17 = 0.74 F2,29 = 11.11 t = 3.86* P = 0.015 P = 0.019 P = 0.072 P = 0.818 P = 0.492 P < 0.001 P = 0.018 Xylanase F3,26 = 0.44 F3,26 = 0.92 F2,21 = 1.43 F2,17 = 6.56 F2,17 = 0.81 F2,29 = 3.04 t = 3.18* P = 0.726 P = 0.445 P = 0.265 P = 0.009 P = 0.463 P = 0.065 P = 0.013 Trypsin F3,26 = 42.40 F3,26 = 2.45 F3,26 = 1.17 F2,17 = 23.59 F2,17 = 208.28 F2,29 = 9.03 F2,14 = 36.21 P < 0.001 P = 0.090 P = 0.343 P = 0.009 P = 0.463 P = 0.001 P < 0.001 Note: If only two values were compared, t -test was used instead of ANOVA. For example, P. nocturnus and H. pyrineusi lacked laminarinase activity in their distal in testines, and thus, for the DI, laminari nase activities were only compared among P. cf. n. Maraon and Pt. disjunctivus with t -test. Similarly, intraspecific compar isons of laminarinase activities in P. nocturnus and H. pyrineusi were only made among the PI and MI with t -test. Sample sizes: P. cf. nigrolineatus Maraon n=6; P. nocturnus n=6; Pt. disjunctivus n=10; H. pyrineusi n=5.

PAGE 70

70Table 2-6. Michaelis-Menten constants (Km) of disaccharidases in the gut walls and micr obial extracts of the proximal intestines of Panaque cf. nigrolineatus Maraon ( Pm ), P. nocturnus ( Pn ), Pterygoplichthys disjunctivus ( Ptd ), and Hypostomus pyrineusi ( Hp ). Maltase -glucosidase N-acetyl-D-glucosaminidase Species Gut Wall Microbial Extract Gut Wall Microbial Extract Gut Wall Microbial Extract Pm 1.84 0.24 2.66 0.39 t = 1.83 P = 0.097 0.041 0.005 0.708 0.042 t = 15.94 P < 0.001 0.075 0.004 0.317 0.063 t = 3.86 P = 0.003 Pn 2.85 0.60 4.33 0.31 t = 2.20 P = 0.053 0.026 0.004 0.976 0.069 t = 13.67 P < 0.001 0.146 0.018 0.187 0.015 t = 1.76 P = 0.108 Ptd 4.35 0.25 5.47 1.53 t = 0.56 P = 0.587 0.121 0.018 1.175 0.113 t = 7.12 P < 0.001 0.172 0.012 0.979 0.237 t = 3.40 P = 0.008 Hp 2.07 0.19 2.09 0.13 t = 0.10 P = 0.923 0.103 0.025 1.391 0.101 t = 12.36 P < 0.001 0.141 0.027 0.470 0.095 t = 3.34 P = 0.010 Note: values are mean ( SEM), and concentrations are in mM. Gut wall and microbial extract constants were compared with t -test for each species and enzyme, and after a B onferroni correction, are considered signif icantly different at P = 0.013. Samples sizes were Pm : n=6; Pn : n=6; Ptd : n=6 (gut wall), n=10 (microbial extract); Hp : n=5.

PAGE 71

71Table 2-7. -mannosidase and aminopeptidase activities (U g-1) in the gut wall and microbial extracts of Panaque cf. nigrolineatus Maraon ( Pm ), P. nocturnus ( Pn ), Pterygoplichthys disjunctivus ( Ptd ), and Hypostomus pyrineusi ( Hp ). -mannosidase Aminopeptidase Species (n) Proximal intestine Mid intestine Distal intestine Proximal intestine Mid intestine Distal intestine Pm (6) Gut Wall 0.418 0.073 0.399 0.129 1.133 0.403 0.223 0.027 0.421 0.123 1.309 0.361 Microbial Extract 0.078 0.015 0.115 0.045 0.141 0.0560.045 0.015 0.069 0.011 0.208 0.053 t 4.57 2.08 2.44 5.75 2.86 3.02 P 0.006 0.065 0.035 <0.001 0.017 0.013 Pn (6) Gut Wall 1.816 0.209 1.453 0.294 0.455 0.105 0.319 0.057 0.923 0.194 1.294 0.236 Microbial Extract 0.041 0.011 0.102 0.019 0.024 0.006 0.038 0.003 0.078 0.006 0.210 0.025 t 8.48 4.57 4.10 4.93 4.53 4.56 P <0.001 0.006 0.009 0.001 0.001 0.001 Ptd Gut Wall (6) 0.229 0.101 0.240 0.056 0.059 0.019 0.358 0.029 0.712 0.089 0.217 0.052 Microbial Extract (10) 0.162 0.027 0.043 0.008 n.d. 0.254 0.045 0.262 0.030 0.237 0.053 t 0.79 3.49 N/A 1.64 5.78 0.25 P 0.441 0.018 0.123 <0.001 0.805 Hp (5) Gut Wall 0.630 0.168 0.931 0.213 0.157 0.058 0.364 0.059 0.973 0.148 1.066 0.318 Microbial Extract 0.134 0.024 0.071 0.021 0.236 0.1260.111 0.0010 0.134 0.029 0.173 0.052 t 2.92 4.01 0.57 4.21 5.56 2.77 P 0.019 0.004 0.586 0.003 0.001 0.024 Note: Values are mean ( SEM). Comparisons were made between the activities of each enzyme from the gut wall and microbial extract of each gut region in each species with t -test. Following a Bonferroni correction for each enzyme and species, differences are considered significant at P=0.017.

PAGE 72

72 Table 2-8. Total short chain fatty acid con centrations (mM) in three gut regions of Hypostomus pyrineusi Pterygoplichthys disjunctivus Panaque cf. nigrolineatus Maraon, and Panaque nocturnus. Gut Region P. cf. n. Maraon P. nocturnus Pt. disjunctivus H. pyrineusi Proximal 2.95 0.65 1.50 0.23 2.44 0.41 1.00 0.16a Middle 2.85 1.40 1.94 0.39 2.40 0.44 3.20 0.79b Distal 2.10 0.33 1.65 0.32 3.50 0.68 2.01 0.40ab F2,17 = 0.26 P = 0.77 F2,17 = 0.48 P = 0.63 F2,17 = 1.28 P = 0.31 F2,14 = 4.55 P = 0.03 Total 7.90 1.95 5.10 0.54 8.44 1.17 6.21 1.00 F3,22 = 1.45 P = 0.26 Note. Values are mean ( SEM). Comparis ons of SCFA concentrations among gut regions within a species, and for total SCFA concentr ation between species, we re made with ANOVA, with differences considered significant at P = 0.0 5. If significant differen ces were detected with ANOVA, this was followed by a Tukeys HSD multiple comparison test with a family error rate of P = 0.05. Those values sharing a superscript letter are not significantly different. Samples sizes were as follows: P. cf. n. Maraon, n=6; P. nocturnus n=6; Pt. disjunctivus, n=10; H. pyrineusi n=5. Acetate:Propionate:B utyrate ratios for total SCFAs were as follows: P. cf. n. Maraon = 62:23:15; P. nocturnus = 44:31:25; Pt. disjunctivus = 70:16:14; H. pyrineusi = 52:28:20.

PAGE 73

73 Table 2-9. Nutritional co mposition of water oak ( Quercus nigra ) wood consumed by Panaque nigrolineatus and Pterygoplichthys disjunctivus in laboratory feeding trials. Component Percent of total Dry Matter (DM) 92.16 0.20 Organic Matter (OM) 97.19 0.19 Nitrogen 0.13 0.01 NDF (total fiber) 86.87 3.88 ADF (lingocellulose) 61.16 2.96 Acid detergent lignin 26.65 1.57 Note: Values are mean ( SEM). n = 11.

PAGE 74

74 Table 2-10. Digestibilities (%) of various fractions of wood consumed by Panaque nigrolineatus and Pterygoplichthys disjunctivus in laboratory feeding trials. Panaque nigrolineatus Pterygoplichthys disjunctivus Dry Matter (DM) 3.35 0.39b 1.72 0.41a Organic Matter (OM) 11.37 0.89 12.76 1.52 Nitrogen -114.93 22.90b -397.78 86.50a NDF (total fiber) 22.95 2.40 31.54 3.56 ADF (lignocellulose) 23.69 2.55 32.08 3.04 Acid detergent lignin 36.87 2.62 32.92 3.51 Note: Values are mean ( SEM). DM, OM, and nitrogen digestib ilities are apparent digestibilities because of endogenous inputs fr om the fish, whereas digestibilities for NDF, ADF, and lignin are true digestibil ities because the fish do not ex crete any substances that are considered fibrous. Values were compared between the species for each digestibility coefficient with t -test (d.f. = 9), and following a Bonferroni correction, values are considered significantly different at P = 0.008. Values for a particular digestibility coeffi cient with different superscript letters are signifi cantly different. P. nigrolineatus, n = 7; Pt. disjunctivus n = 4.

PAGE 75

75Table 2-11. Particle sizes of intestinal contents presented as the percent of tota l contents for each of the proximal, mid, an d distal intestine of Panaque nigrolineatus Region of Intestine <250 m 250-350 m 351-700 m 701-1000 m 1001-1500 m > 1501 m Proximal 44.59 4.97 10.73 0.69 18.10 1.76 9.26 1.19 8.76 1.58 8.56 1.47 Mid 25.42 3.37 13.47 1.75 24.20 2.69 10.17 2.63 13.23 1.50 13.51 1.93 Distal 23.82 2.64 8.56 0.93 30.06 3.51 10.21 1.04 13.09 2.23 14.25 3.90 Note: Values are mean SEM. N=12 for proximal and mid intestine, N=14 for distal intestine.

PAGE 76

76 Figure 2-1. Partial phylogenetic hy pothesis for three tribes in the catfish fam ily Loricariidae. Phylogeny based on parsimony and summarized from Armbrust er (2004). Genera in bold include wood-eating species, and the asterisks (*) indicate genera from which species were invest igated in this study. Numbers in parenthe ses indicate approximate number of taxa not shown. Panaque (small; 5) Panaque (large; 3) *Pterygoplichthini Pter yg o p lichth ys ( 14 ) *Hemiancistrus ( 20 ) H yp ostomu s ( 2 ) H yp ostomu s ( 5 ) H yp ostomu s ( 6 ) H yp ostomu s round snout ( 1 ) H yp ostomu s ( 6 ) Hypostomus cochliodon group (7)* Hemiancistrus sp. (1) Pekoltia (2) Hypancistrus/Parancistrus (2) Six genera (7) Hemiancistrus/Pekoltia (3) Chaetostoma and others (30)Hypostomini Ancistrini

PAGE 77

77 Figure 2-2. Photographs of the digestive tract of Pterygoplichthys disjunctivus : coiled within th e body cavity (A; scale bar = 75 mm), uncoiled beneath the body of the fish (B; scale bar = 150 mm ), and freshly removed from the body, coiled (C; scale bar = 75 mm). A C B

PAGE 78

78 Figure 2-3. Histological images and transmission electron microscope (TEM) micrographs of the proximal, mid, and distal intestine of Pterygoplichthys disjunctivus ( Ptd ), Panaque cf. nigrolineatus Maraon ( Pm ), P. nocturnus ( Pn ), and Hypostomus pyrineusi ( Hp ). Scale bar for histology = 65 m, scale bar for TEM = 1 m. Pm Pn Hp Ptd Proximal Mid Distal Ptd Intestinal region Species

PAGE 79

79 0 0.005 0.01 0.015 0.02 0.025Xylanase(U .g-1) n.d Pm Pn Ptd Hp Proximal Intestine Mid Intestine Distal Intestine Amylase (U .g-1)a a b b b a ab b b a a a a 0 400 800 1200 1600 2000 2400 Laminarinase(U .g-1) 0.0 0.2 0.4 0.6 0.8 1.0a a ab b a a b b 0 0.004 0.008 0.012 0.016 0.02Cellulase(U .g-1)b a a b a b a b a b b a n.d n.d n.d

PAGE 80

80 Figure 2-4. Total activities (int estinal fluid + microbial extrac t) of amylase, laminarinase, cellulase, and xylanase in three regions of the intestine of Panaque cf. nigrolineatus Maraon ( Pm ), P. nocturnus ( Pn ), Pterygoplichthys disjunctivus ( Ptd ), and Hypostomus pyrineusi ( Hp ). Values are means and error bars represent SEM. Interspecific comparisons of each enzyme activity in each gut region were made with ANOVA followed by a Tukeys HSD with a family error rate of P = 0.05. Bars of a specific color and for a specific enzyme shari ng a letter are not significantly different. Intraspecific comparisons of each enzyme among gut regions were made with ANOVA followed by a Tukeys HSD with a family error rate of P = 0.05. Lines of a specific elevation passing over two or more bars indicate a significant difference in enzyme activity (P < 0.01) among those gut regions.

PAGE 81

81 0 1 2 3 4 5 6 0 1 2 3 4 5 6* -glucosidase(U .g tissue-1) 0 1 2 3 4 5 6 7 8* *Panaque nocturnus 0 1 2 3 4 5 6 7 8 Gut Wall Microbial ExtractPanaque cf. nigrolineatus Maraon *Maltase (U .g tissue-1) 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10PI MI DI PI MI DI *NAG (U .g tissue-1)

PAGE 82

82Figure 2-5. Maltase, -glucosidase, and N-acetyl-D-glucosaminidase (NAG) activities in the gut walls and microbial extracts of the proximal intestine (PI), mid intestine (MI), and distal intestine (DI) of Panaque cf. nigrolineatus Maraon (left column) and P. nocturnus (right column). Comparisons were made of the activities of each enzyme between the gut walls and microbial extracts of each gut region with t -test. Following a Bonferroni corre ction for each enzyme and species, differences are considered significant at P=0.013 [indicated with an asterisk (*)].

PAGE 83

83 0 2 4 6 8 10 0 2 4 6 8 10* -glucosidase(U .g tissue-1) 0 10 20 30 40 50 60 Gut Wall Microbial Extract 0 10 20 30 40 50 60* *Maltase (U .g tissue-1)Pterygoplichthys disjunctivus Hypostomus pyrineusi 0 5 10 15 20 25 0 5 10 15 20 25PI MI DI PI MI DI *NAG (U .g tissue-1)

PAGE 84

84Figure 2-6. Maltase, -glucosidase, and N-acetyl-D-glucosaminidase activities in the gut walls and microbial extracts of the proximal intestine (PI), mid intestine (MI), and distal intestine (DI) of Pterygoplichthys disjunctivus (left column) and Hypostomus pyrineusi (right column). Comparisons were made of the activities of each enzyme between the gut walls and microbial extracts of each gut region in each species with t -test. Following a Bonferroni correction for each enzyme and species, differences are consid ered significant at P=0.013 [indicat ed with an asterisk (*)].

PAGE 85

85 Figure 2-7. Total activities (intestin al fluid + microbial extract) of trypsin in three regions of the intestine of Panaque cf. nigrolineatus Maraon ( Pm ), P. nocturnus ( Pn ), Pterygoplichthys disjunctivus ( Ptd ), and Hypostomus pyrineusi ( Hp ). Values are means and error bars represent SEM. Comparative data as in Figure 2-4. Proximal Intestine Mid Intestine Distal Intestine 0 40 80 120 160 200a c b b Pm Pn Ptd Hp Trypsin(U .g-1)

PAGE 86

86 Figure 2-8. Position of staine d wood in the intestine of Panaque nigrolineatus at different time intervals following its consumption. Fish were allowed to graze on the wood for four hours (pulse), at which time the stained wood was taken away from the fish and replaced with non-stained wood so the fish c ould continue to fee d. Four individual fish were sacrificed and th e position of stained wood in the gut observed at four, six, eight, and 18 hours following commencement of feeding on the stained wood. No stained wood remained in the intestine at 18 hours post feeding. Stained wood was already present in feces at the four-hour interval, indicating that transit time is less than four hours. PI MI DI 4 6 8 18 Hours after consuming stained wood Position of stained wood in the intestine

PAGE 87

87 CHAPTER 3 CAN WOOD-EATING CATFISHES ASSIMI LATE NUTRIENTS AND ENE RGY FROM WOOD? INSIGHTS FROM STABLE ISOTOPES IN THE LABORATORY AND IN THE FIELD Introduction Plants and animals are m ade primarily of carbon, nitrogen, hydroge n, and oxygen. Each of these elements exists in different forms that vary in mass because of different numbers of neutrons in their nuclei (Fry 2007) These different forms of the atoms with different masses are called isotopes, and are denoted by their mass (e.g., 13C has an atomic mass of 13, whereas 12C had an atomic mass of 12). In the atmosphere, the proportional amounts of heavy and light isotopes of carbon, nitrogen, and oxygen are relativ ely constant. However, plants and animals use these isotopes in different wa ys resulting in the organisms ha ving varying ratios of heavy and light isotopes relative to one another. For example, many mo nocot plants, like grasses, use the C4 photosynthetic pathway, which results in tissu es with more of the heavy isotope of carbon (i.e., 13C) than dicot plants, which use the C3 photosynthetic pathway (Fry 2007). From an isotopic perspective, an animal truly is what it eats because the atoms they use to synthesize new tissues come directly from their diet. T hus, the ratios of heavy and light isotopes of different elements in an animals diet (e.g., 13C/12C and 14N/15N) can be used to trace what foods these animals actually digest, assimila te, and use to make new tissue. When attempting to discern an animals dietary habits using stable isotopes it is critical to recognize that the animals isotopi c signature may be similar to th at of their diet (DeNiro and Epstein 1978; DeNiro and Epstein 1981), plus or minus some difference, commonly referred to as an isotopic discrimination factor ( Xtissue-diet; Reich et al. 2008). The isotopic discrimination factor between an animals tissue and their diet is typically caused by three mechanisms: first, isotopic memory, which is the observation that the isotopic signature of an animals tissues do

PAGE 88

88 not immediately match that of the new diet fo llowing a dietary switch, and instead follow some temporal dynamics of isotopic incorporation (F ry and Arnold 1982; Phillips and Eldridge 2006); second, metabolic fractionation, which is the isot opic difference between reactants and products in biochemical reactions (Fry 2007); and third, isotopic routing, which is the shuttling of nutrients into pools by nutrient class (e.g., absorbed amino acids enter the amino acid pool, absorbed fatty acids enter the fatty acid pool, etc.; Martnez del Rio and Wolf 2005), as opposed to homogenous mixing of atoms. Each of thes e factors can produce differences in the isotopic signatures among the nutrient fractions of a single tissue (e.g., lip ids tend to be depleted in 13C; Fry 2007), and produce isotopic diffe rences among tissues in a si ngle animal, especially when compared to that animals diet. Thus, when usi ng stable isotopes to analyze the dietary history of an animal in the wild, expected discrimina tion factors for targeted tissue types should be determined in the laboratory for that animal beforehand (Gannes et al. 19 97; Reich et al. 2008). Of course, the advantage of using stable isotopi c analyses is that they provide an important dietary reference point integrated over some time period, unlike gut content analyses, which provide only a snapshot view of diet (Fry 2007). However, the length of time it takes for a tissue to turn over and equilibrate with the isotop ic signature of food is a necessary piece of information in stable isotopic investigations. The turnover time varies with tissue type (Tieszen et al. 1983; Hobson and Clark 1992; Martnez de l Rio and Wolf 2005), with dietary quality (Gaye-Siessegger et al. 2003; Gaye-S iessegger et al. 2004), and with growth rate (Hesslein et al. 1993; MacAvoy et al. 2001; Sakano et al. 2005; Trueman et al. 2005; Reich et al. 2008). Therefore, without knowledge of expected discrimination factors and turnover times, it is difficult to discern the dietary hi story of an animal with stab le isotopes and to know the time frame at which you are operating (Parker et al. 20 08; Reich et al. 2008). That is, does a tissue

PAGE 89

89 take days, weeks, months, or years to turn ove r and match the isotopic co mposition of a diet, and just how large is the discriminati on when turnover is reached? Although the number of studies of fish eco logy using stable isotopic analyses has expanded in recent years, few have attempted to accurately determine the turnover times and discrimination factors of tissues in the labor atory (Hesslein et al. 1993; MacAvoy et al. 2001; Jardine et al. 2004; Sakano et al. 2005; Guelin ckx et al. 2007) before gathering tissues and dietary items from the field (e.g., Ho et al. 2007). The results of several studies with fish suggest that discrimination factors commonly used in the literature to reflect tr ophic shifts (e.g., 1 for 13C, 3.4 for 15N; Fry and Sherr 1984; Focken 2004; Fry 2007) are not correct for many species, or dietary items (Hesslein et al. 1993; Jardine et al. 2004; Trueman et al. 2005; Barnes et al. 2007; Guelinckx et al. 2007; Mi ll et al. 2007; Gamboa-Delgado et al. 2008). Therefore, it is clear that the variabil ity in turnover times and discrimi nation factors among species obligates laboratory investigations in each species under study, or at least in closely related taxa that are morphologically and physiologically similar to the target species (MacAvoy et al. 2001). Laboratory isotopic turnover experiments provi de valuable information, not only for the better understanding of field isotop ic signatures, but also for observing the assimilation of specific nutrients, especially if a component of a diet is indi gestible (Gamboa-Delgado et al. 2008). For example, wood is made primarily of cellulose and hemicellulose, but is also composed of 18-35% lignin (Petterson 1984), which is absolutely indigestible by vertebrates, or by endosymbiotic microorganisms inhabiting their digestive tracts (Van Soest 1994; Stevens and Hume 1995; Karasov and Martnez del Rio 2007). Furthermore, different fiber types of wood (e.g., cellulose vs. lignin) have different isotopic signatures, with cellulo se being ~2 enriched in 13C compared to bulk wood (Gaudinski et al. 2005; Bowling et al. 2008), and lignin being

PAGE 90

90 ~4 depleted (Bowling et al. 2008). Thus, one would expect an animal consuming wood to have discrimination factors skewed towards the digestible, cellulosic portions of their diet rather than the bulk wood. However, few studies have attempted to trace th e assimilation of wood carbon using stable isotopes in animals (e.g. Taya su et al. 1997; Nonogaki et al. 2007), and I know of none that have examined whether a woodeating animal can specifically assimilate the isotopes from the cellulosic portion of their diet, rather than the overall isotopic signature of the bulk wood (including lignin). In this study I combined laboratory and fiel d stable isotopic inves tigations to discern whether wood-eating catfishes (fam ily Loricariidae) can assimila te wood and thrive on a woody diet. I performed turnover studies with captive individuals of the de tritivorous loricariid, Pterygoplichthys disjunctivus and used several tissues that c ould be sampled non-invasively (red blood cells, plasma, and fin tissue) over time. I determined turnover rates and discrimination factors for 13C and 15N for these tissues, and used this information to guide stable isotopic analyses of this and related speci es gathered in the field. Although Pt. disjunctivus is not a true wood-eating catfish, they were appropriate laboratory surrogates because they are closely related to the wood-eating catfishes (i.e., in the genera Panaque and Hypostomus ; Armbruster 2004), and because they have similar gut morphology, digestive physiology (Chapter 2), metabolic rates (Nelson 2002), and low-quality diet to their w ood-eating brethren. Furthermore, the wood-eating catfishes are native to rugged regions of the Amazonian basin a nd are difficult to obtain in sufficient numbers for use in th e laboratory (Schaefer and Stew art 1993; DPG, pers. obs.), even via the aquarium trade. A lthough native to South America, Pt. disjunctivus has been introduced to waterways in Florida, where it has been es tablished for nearly two decades (Nico 2005).

PAGE 91

91 Thus, I had access to an abundant lo cal population of this species wi thin transport distance of the laboratory. I performed two stable isotopic turnover experime nts in this investigation. In trial one, I established turnover rates and di scrimination factors for the diffe rent tissues (red blood cells, plasma, and fin tissue) of Pt. disjunctivus on an algal diet. In trial two I determined whether this species was capable of assimilating the fibrous po rtions of wood I fed the fish an artificial wood-detritus diet with fibrous ( 13C = -26.36; 15N = 2.13) and soluble components ( 13C = -11.82; 15N = 3.39) of different isotopic signatures to observe whether the fish were assimilating the wood fiber, the soluble components, or a mixture of the two. I also observed whether the fish were growing on a wood-domina ted diet. Wood-detritu s is not completely made of fibrous material, as degraded wood is typically composed of 20% soluble components (Chapter 2), most likely repres enting microbes, digestive enzy mes of microbial origin, and soluble wood-degradation products (Sinsabaugh et al. 1991b; Sinsabaugh et al. 1992; Tank et al. 1998; Hendel and Marxsen 2000). Thus, I attemp ted to mimic the biochemical composition of natural wood-detritus with my ar tificial version for trial two. I also extracted the cellulosic components of wood used in the laboratory, and fr om that grazed upon in the field to determine whether fishes were specifically assimilating th e carbon isotopes of cellulose and hemicellulose as opposed to other elements of bulk wood, such as lignin. Although Nonogaki et al. (2007) suggested that lorcariid ca tfishes can assimilate wood carbon (based on bulk wood 13C), I provided physiological ev idence in Chapter 2 of this dissertation showing that wood-eat ing loricariid catfishes, and Pt. disjunctivus, poorly digest the fibrous components of wood in their intestines. These fishes clearly digest the more soluble components of their diet. Thus, I predicted that Pt. disjunctivus could not assimilate the carbon

PAGE 92

92 isotopes of wood fiber in the laboratory. Howe ver, because many of the soluble components of natural wood-de tritus (e.g., -glucosides) do come from cellulo se degradation (Sinsabaugh et al. 1992; Tank et al. 1998; Hendel and Marxsen 2000), I predicted th at wild-caught wood-eating catfishes would have carbon isot opic signatures reflecting the cellulosic component of their wood diet, and not that of bulk wood or lignin. Furthermore, I hypothesized that the nitrogen signatures of wild-caught wood-eating catfishes would be much greater (i.e., 15N > 2; Tayasu et al. 1997) than those of wood-eating termites, which are reliant upon endosymbiotic nitrogen fixation to meet their nitrogen requirements (Sla ytor and Chappell 1994). Symbiotic nitrogenfixing bacteria fix atmospheric nitrogen into usable compounds for the host organism (Karasov and Martnez del Rio 2007). However, from a st able isotopic perspective, this results in low 15N values for the host (near zero) becau se the standard used to determine 15N is atmospheric nitrogen. Instead, I pr edicted the catfishes would have enriched 15N signatures reflecting additional nitrogen sources in their food; spec ifically, from detritus, microbes (fungi in particular) that are common in degrading wood in aquatic systems (Sinsabaugh et al. 1991a; Hendel and Marxsen 2000), and poten tially from aquatic insects. Methods Fish Collection and Maintenance in Laboratory Twenty-two adult Pterygoplichthys disjunctivus were captured from the Wekiva Springs com plex in north central Florida (28.321 N, 81.464 W) in May 2005. The fish were similar in length [mean SD; 128.77 2.12 mm standard length (SL)] and mass (51.29 2.19 g). Upon capture, fishes were plac ed in 128-L coolers of aerated river water and transported alive back to the University of Florida. Upon arriva l, fish were randomly assigned, in pairs, to 75.6-L aquaria equipped with mechanical filtration, cont aining naturally dechlor onated, aged tap water,

PAGE 93

93 and under a 12L:12D light cycle (Cha pter 2). The thermostat in the aquarium laboratory was set at 25C for the duration of the experiment and the temperature of each tank was monitored daily to confirm that the temperature did not vary by mo re than 1C. The tanks were scrubbed, debris and feces siphoned out, and 95% of the water changed every five to seven days to limit algal and microbial growth in the tanks as possible confounding food sources. Tissues Used for Stable Isotope Analysis I analyzed the isotopic composition of red blood cells, plasm a solutes, and pelvic and caudal fin tissue. I chose thes e tissue types be cause I could sample them in a non-invasive manner. Indeed, one of the goals of this study was to investigate isotopic turnover and discrimination in blood and skin tiss ue (i.e., fin clips) for use in st able isotopic studies of fishes, as has been done for other ectothermic (Kelly et al. 2006; Reich et al 2008) and endothermic (Bearhop et al. 2002) vertebra tes. Approximately 150 L of blood was drawn with a 23-gauge needle from the haemal arch just posterior to the anal fin of the fish, transferred to unheparinized capillary vials, and immediately centrifuged at 13,000 x g for five min to separate the red blood cells, white blood cells, and plasma. Following the separation, the red blood cells (RBCs) and plasma were placed in separate, sterile centrifuge vials, and the white blood cells were discarded. Fin clips (approximately 0.5 g wet weight) we re taken with steril e scissors by cutting membranous tissue and soft rays (no thicker, harden ed rays) from left or right pelvic fin, or the dorsal or ventral caudal fin, depe nding on the individual fish. Because the specific fin taken was also used to identify individual fish in their respective aquaria, the same fin was sampled at each sampling interval from each individual fish. Th e fin tissue completely regenerated within the 30 days between sampling intervals (a common observa tion in fishes; Wills et al. 2008), and thus, the isotopic turnover examined for the fin tissue was not for steady-st ate fin tissue, but rather, for

PAGE 94

94 regenerated fin tissue. The isotopic turnover for this regenerated tissue may differ from nonregenerated tissue, and therefore, may not re flect fin tissue isotopic turnover in nature. Stable Isotope Trial 1: Initial Turnover Upon arrival in the laboratory, tissues were taken from four i ndividual fish to provide an initial stable isotopic reading. Thereafter, fish were fed a commercial algal diet (W ardley Premium Algae Discs, Hartz-Mountain Corporation, Secaucus, NJ) for 203 days (Table 3-1). These algae discs contained primarily the green alga Spirulina as an ingredient, but also contained a variety of grains, le gumes, and plant proteins, in addi tion to vitamins and minerals. Fish were offered seven discs per night, which they consumed, equating to approximately 6% of their body mass, on a wet mass basis, per day. For the first 100 days of the feeding trial, tissues were sampled from three to four fish (from different tanks) every five days, whereas six to seven individuals were sampled every 10-11 days for th e final 103 days of the experiment. I designed my sampling regime so that indi vidual fish were sampled ever y 30 days, giving them ample time to recover from the blood draws and fin tissue biopsies. Three additional fish were captured from the wild, and blood and fin clips take n for isotopic analyses on days 100 and 210 to evaluate how the isotopic signatur e of the fish changed in nature over this same time period. The blood and finclip samples from days 25 to 55 of the experiment were lost due to non-demonic intrusion (Hurlbert 1984), which, in this case, was a careless resear ch assistant. However, these missing samples did not necessarily affect the over all analyses because the rates of isotopic incorporation could st ill be determined (see below). Stable Isotope Trial 2: Wood A ssimilation and Negative Control After 203 days on the algae diet, the f ish were given a refractory period of 40 days, during which they were fed the algae di scs but not handled. After this intermission, fish were divided into two groups: one group (n=6) remained on th e algae diet (positive controls), and one group

PAGE 95

95 (n=16) was switched to a wood di et (experimental group). An a dditional six fish were captured from the wild in February 2006 to act as negativ e controls and were not fed for the duration of the experiment (see below). The goal of this experiment was to examine whether P. disjunctivus could assimilate carbon and nitrogen from the structural compone nts of wood (cellulose, hemicellulose, lignin), or if they simply digested and assimilated soluble degradation produc ts found on wood as it is decomposed. Thus, I designed a diet that featur ed wood (and its inherent fiber types) of one isotopic signature, and soluble components of another isotopic signature (Table 3-1). The wood was that of decomposed riparian water oak (Quercus nigra ), collected from the Sampson River, FL (29.37 N, 82.16 W), and the soluble components were composed of corn gluten meal, corn meal, vitamins, and minerals (Table 3-1). The wood was chopped into smaller pieces with a hatchet, and ground to partic le sizes of 0.25 mm in a coffee grinder (Krups GX 4100). The wood was then autoclaved to ensure it was ster ile, and dried at 60C for 24 hours. The corn gluten meal, corn meal, lysine, vitamins, and minerals (gifts from Hartz-Mountain Corporation, Secaucus, NJ, and some of the same ingredients in the algal diet described above), intended specifically for use in fish food, were then added to the dried wood, followed by xanthan gum (6%; Now Foods, Bloomingdale, IN) and a small amount of water (20 mL/100 g dry mass; Table 3-1). The xantham gum and water, combined, acte d as binding agents to keep the ingredients bound when submerged in water. The mixture was then homogenized vigorously by hand with a stirring rod, and pressed into 4x2 mm (0.45 g) circular pellets in a hand press (Parr Instruments, Moline, IL). Because dried wood tends to float and the fish pr imarily feed on the benthos, the pellets were then adhered to a piece of PVC pipe (n=10 pellets per PVC piece) with a small drop of superglue (Loctite, Avon, Ohio), and sunk in the aquaria where the fish actively fed on the

PAGE 96

96 pellets during the evening hours (i.e., in the dark ). Because some of each pellet was permanently polymerized in the superglue, a small portion (~ 0.05g) of each pellet was inedible by the fish. The fish were offered, and consumed, 10 pellets pe r night, equating to approximately 8% of their body mass, on a wet mass basis, per day. Given that Panaque nigrolineatus (a true wood-eating loricariid catfish) was observed to consume 2-5% of the their body mass in wood per day in the laboratory (Chapter 2), I considered 8% of the fishes body mass per evening in food to be ad libitum The fish were fed this diet for 155 days, and, as with the initial turnover experiment, tissues were taken every five da ys for the first 50 days of the experiment, and every 10 days for the remaining 105 days. Fish remaining on the algae diet (positive controls) were fed as described above for trial 1. Because previous observations suggested that P. disjunctivus did not gain weight on a wood-only diet in the laboratory (Chapter 2), I include d a group of fish that acted as negative controls. Six individuals were deprived of food for the duration of the wood-feeding experiment. These fish were also weighed and tissues take n from individual fish every 30 days. One individual died during the course of the food-deprivation, so I had a final sample size of five for the negative controls. At the end of trial 2 (155 days) the fish we re euthanized in buffered water containing MS222 (1 g l-1), measured [standard length (SL) 1 mm], and weighed [body mass (BM) 0.5 g ]. Epaxial white muscle samples were taken fr om each fish for endpoint stable isotopic measurements. The livers of all of the fish were weighed, the hepato-soma tic index [HSI = liver mass (g)/body mass (g)] determined, and a subsampl e of the liver tissue was dried at 60C and analyzed for %C and %N in a Carlo-Erba elem ental analyzer. The HSI and %C:%N ratio were used to assess the health of the fish on the diffe rent diets (or deprived of food). The HSI is a

PAGE 97

97 body condition index (lower HSI = lower body cond ition, poor health; Lloret and Planes 2003) and the %C:%N ratio indicated whether the fish were mobilizing nitrogen from their livers (higher %C:%N ratio = less nitr ogen; Karasov and Martnez del Rio 2007), which would indicate that the fish were in negative nitrogen balance. All handling of fish from capture to euthanasia was conducted under approved protocols D995 and E 822 of the Institutional Animal Care and Use Committee of the University of Florida. Stable Isotopic Profiles of Wild-Caught Fish and Resources Wood, biofilm, algae, seston, abunda nt invertebrate taxa and fi sh were collected from the field sites for stab le isotopic measurements (T ables 3-2 and 3-3). I was not exhaustive in my sample collection, but I attemp ted to characterize a small por tion of the foodwebs in two locations: first, in the upper Ro Maraon in northern Peru (4.957 S, 77.283 W; Chapter 2), and second, in Wekiva Springs, FL (see abov e under Fish collection and maintenance in the laboratory). Wood, algae, and invertebrates we re rinsed with deionized water and frozen in liquid nitrogen in the field. Biofilm was colle cted by gently brushing the surface debris from wood and collecting it in a ziploc k bag. The bag was sealed and the contents of the bag were mixed by shaking the bag. I then pipetted 1 mL of the slurry into 1.5-mL centrifuge vials and centrifuged them at 10,000 x g for five min in an Eppendorf 5415R desktop centrifuge powered by a 12V car battery via a power inverter. The s upernatant was discarded and the pelleted debris was frozen in liquid nitrogen. Seston was collected by filtering river water through a 0.25 m glass fiber filter. Blood and fin clip samples were taken from fish as described above under Tissues used for stable isotope analysis. I also collected epaxial white muscle (Peru) or fin clip (Florida) samples from several fish specie s that were too small from which to draw an adequate blood sample (Tables 3-2 and 3-3). All fish tissues were frozen in liquid nitrogen in the

PAGE 98

98 field. All samples were shipped back to the University of Florida on dry ice, dried at 60C, and stored in sealed containers at room temperature until analyzed. Because wood is composed of compounds i ndigestible by the fish (e.g., lignin), it was necessary to extract the more d igestible portions (i.e., cellulose hemicellulose) of the wood for stable isotopic analyses. Holo-cellulose (Leav itt and Danzer 1993; Gaudins ki et al. 2005), which comprises the cellulosic and hemicellulosic compounds of wood, was extracted following a slightly modified Jayme-Wise method (Leavitt and Danzer 1993), essen tially as described by Gaudinski et al. (2005). Wood from Per and Florida was dr ied at 60C, ground in a coffee grinder to pass through a 1-mm screen, and 200-500 g samples of the ground wood were placed in polyester solvent bags (Ankom Technology, M acedon, New York). The samples were then extracted in 2:1 Toluene:Ethanol in a Dionex Accelerated Solvent Extractor (ASE), followed by a second extraction in 100% ethanol. The samples were then boiled for 4 hours in deionized water to remove soluble components, and a llowed to dry in a drying chamber at room temperature. The bags (containing the samples) were then soaked in 1-L of an aqueous bleach solution containing 10 g of sodium chlorite and 6 mL of glacial ace tic acid at 70C for five days. The bleach/acetic acid solution was changed every 12 hours. Following the bleaching, the samples were thoroughly rinsed in running deioni zed water for 3 hours, and dried at 60C. The remaining white fibrous material was holo-cellulo se as determined by Gaudinski et al. (2005). Sample Preparation for Mass Spectrometry Blood (red blood cells and plasma solutes), fin clip, and muscle samples from fish, and, wood, holo-cellulose, algae, invertebrate, biofilm, and seston (on a gla ss-fiber filter) samples from the environment, were dried to a constant weight for 24-48 hours at 60C (Reich et al. 2008). Lipids were extracted from all animal a nd plant material (except the red blood cells and

PAGE 99

99 plasma solutes, which were too small) with petroleum ether in a Dionex Accelerated Solvent Extractor (Dodds et al. 2004; Reic h et al. 2008). Approximately 500 g of animal, plant, or detrital tissue were load ed into pre-cleaned tin capsules combusted in a COSTECH ECS 4010 elemental analyzer interfaced via a FinniganMAT ConFlow III device (Finnigan MAT, Breman, Germany) to a Finnigan-MAT DeltaPlus XL (Breman, Germany) isotope ratio mass spectrometer in the light stable isotope lab at th e University of Florida, Gainesville, FL, USA. Stable isotope abundances are expressed in delta ( ), defined as parts per thousand () relative to the standard as follows: = [(Rsample/Rstandard) 1] (1000) (1) where Rsample and Rstandard are the corresponding ratios of heavy to light isotopes (13C/12C and 15N/14N) in the sample and standard, respectively. Rstandard for 13C was Vienna Pee Dee Belemnite (VPDB) limestone formation international standard. Rstandard for 15N was atmospheric N2. IAEA CH-6 ( 13C = -10.4) and IAEA N1 Ammonium Sulfate ( 15N = +0.4), calibrated monthly to VPDP and atmospheric N2, respectively, were inserted in all runs at regular intervals to calibrate the system and assess drift over time. The analyti cal accuracy of our measurements, measured as the SD of replicates of standards, was 0.14 for 13C and 0.11 for 15N (N = 120). Statistical Analyses I estimated growth rates (g in days-1) of Pt. disjunctivus using an exponential model (y = aebt; Reich et al. 2008). The fractiona l rate of isotopi c incorporation, was estimated (in days-1) with a non-linear fitting procedure using the equation X( t ) = X( ) + [ X(0) X( )]et, (2)

PAGE 100

100 where X(t) is the isotopic composition at time t, X( ) is the asymptotic, equilibrium isotopic composition, X(0) is the initial isotopic composition, and is the fractional rate of isotope incorporation in a tissue (Martnez del Rio and Wolf 2005; Reich et al. 2008). X( ) and X(0) were estimated using the same non-linear procedure. can be defined as the sum of tissue net growth (kgt) and tissue catabolic turnover (kdt); thus, = kgt + kdt. (Hesslein et al. 1993; Reich et al. 2008). If a tissue is growing exponentially, then I can measur e the growth and determine the contribution of growth and ti ssue catabolic turnover to (Reich et al. 2008). Following Reich et al. (2008), I assumed that the fractional growth rate of a tissue was equal to the fractional rate of growth of the whole animal (kg). I then compared to kg using t -tests, to determine whether the contribution of tissue catabolic turnover to isotopic incorporation was significant. A large difference between and kg indicates a large contribution of ti ssue catabolic tur nover to isotopic incorporation as opposed to new tissue accreti on. Because I used adult animals in our experiments, I anticipated a large difference between and kg, and thus, a large contribution of catabolic turnover, unlik e growing juvenile anim als (Trueman et al. 2005; Gamboa-Delgado et al. 2008; Reich et al. 2008). Isotopic discrimination factors ( Xtissue-diet) were calculated as X( )tissue Xdiet. Turnover times (average residence times) of C a nd N molecules in trial one were calculated as 1/ (Reich et al. 2008). In tr ial two, I used the following con centration-dependent linear mixing model incorporating digestibility estimates to pred ict the final isotopic signatures of the fish if they were assimilating the diet: n j jjx iiixpe xpej1 xip (3)

PAGE 101

101 where pxi is the expected isotopic signatu re assimilated of isotope X, ei is the digestibility coefficient for a dietary ingredient (e.g., wood), pi is the proportion of the diet composed of that ingredient, xi is the concentration (%) of the atom in question (e.g., C) in the ingredient, and the denominator is the summed totals of all of the ingredients in the diet (Martnez del Rio and Wolf 2005). Digestibility estimates for wood and cellulose were garnered (for Pt. disjunctivus specifically) from Chapter 2 and di gestibilities for corn gluten m eal (Guimares et al. 2008) and corn meal (Krogdahl et al. 2005) were taken from the literature. Xanthan gum was considered an indigestible non-starch polysaccharide (Leenhouwers et al. 2006), and the vitamins and minerals were considered 100% absorbable. Because 20 % of the bulk wood mass is made of soluble degradation components (Chapter 2) this was also added into the mixing model. However, the isotopic signature and actual digestibility of this fraction is unknown. Thus, I assumed that most of these degradation products are of cellulosic orig in (Sinsabaugh et al. 1992), and would, therefore, have an isotopic signature identical to that of cellulose ( 13C = -25.39). Pterygoplichthys disjunctivus has digestive enzyme activities in its gut indicative of efficient digestion of soluble wood de gradation products (i.e., high -glucosidase, -glucosidase, and mannosidase activities), so I assumed digestibilie s of 90% for this soluble component. With these ingredients and digestibiltie s, the artificial wood-detritus diet had the following predicted digestible isotopic signature: 13C = -21.79, and 15N = 3.25. Hepato-somatic indices and liver %C:%N ratios were compared among the experimental (wood-detritus diet), control (a lgae diet), and food-deprived groups with ANOVA, followed by a Tukeys HSD with a family error rate of P = 0.05. The same ANOVA procedure was also used to compare field isot opic signatures in Pt. disjunctivus among days 0, 100, and 210 of the experiment. Prior to all t -tests and ANOVA, a Levenes test for equal variance was performed to

PAGE 102

102 ensure the appropriateness of the data for parametric analyses All modeling and tests were run with SPSS statistical so ftware (version 12). Results Trial One: Initial Turnover The exponential growth rate for algae-fe d fish was (mean SD) 0.0017 0.0006 day-1, and the fish gained 40.86 0.17% of their body mass during the experiment. The exponential model (y = aebt) described the growth rates of the fish reasonably well (r2 = 0.82-0.99). Equation 2 described the changes in 13C and 15N through time adequately well in all tissues (r2 ranged from 0.73 to 0.92; Figure 3-1). All tissues s howed isotopic incorporation rates that were significantly greater than those expected by growth alone (0.0017 day-1), indicating a significant contribution of tissue catabolic turnover to isot opic incorporation (Tables 3-4 and 3-5). Tissue catabolic turnover contributed between 72 and 97% of carbon isotopic incorporation, and between 95 and 99% of nitrogen incorporation. 13Ctissue-diet and 15Ntissue-diet varied by tissue, with 13Ctissue-diet ranging from -0.13 to 1.75, and 15Ntissue-diet ranging from 4.08 to 5.17 (Tables 3-4 and 3-5). The turnover times of C varied by tissue, with fin clips showing the shortest residence time of 20 days, and red blood cells the longest at 167 days (Table 3-4). Similarly, nitrogen turnover times varied by tissue, as red blood cells showed residence times of approximately 4.5 days, and fin clips exhibited re sidence times of 33.2 days (Table 3-5). Wildcaught fishes captured on days 100 and 210 of the experiment had 13C signatures that varied from the first day of the experiment (significantly so for plasma solutes and fin tissue), whereas 15N was less variable over time (Table 3-6). Nonetheless, the 13C of wild fish became more depleted after 100 days, not more enriched as I ob served in the algae-fed fish in the laboratory (Table 3-6).

PAGE 103

103 Trial Two: Wood Assimilation and Negative Control All of the fish lost weight on the artificial wood-detritus diet, with the mean loss being (mean SD) 7.87 2.74% of their body mass ove r 155 days, whereas the control fish on the algae diet gained 31.23 8.10% of their body mass over the same time period. The fooddeprived fish lost 3.67 1.49% of their body mass over 150 days. The fish on the artificial wood-detritus diet lost significantly more weight than t hose that were food-deprived ( t = 10.61, P < 0.001, d.f. = 9), likely because the former were more active on a daily basis than the latter (DPG, pers. obs.). Equation 2 adequately described the changes in 13C over time for plasma solutes, but less so for red blood cells (Table 3-7; Figure 3-2). A linear procedure was necessary to describe the incorporation of 13C for the fin clip samples. Equation 2 appropriately described the incorporation of 15N over time in plasma solutes and fin clips, but not for red blood cells, which required a linear procedure (Table 3-8; Figure 3-2). It is clear that the 13C of all tissues became enriched over the course of the experiment, moving toward the predicted signature, but not perfectly so, especially for the fin tissue (Table 3-7). The 13C values of control fish (i.e., those consuming the algal diet) were unchanging ove r the course of the expe riment (Figure 3-2). The isotopic incorporation for 13C in plasma solutes in trial one was similar to that observed in trial two for the fish consuming the ar tificial wood-detritu s (0.026 and 0.023 day-1, respectively; Tables 3-4 and 3-7, respectively). Comparable resu lts were observed for the fin clips, as the rate of 13C incorporation in this tissue was similar be tween trial one and trial two (0.031 and 0.021 day-1, respectively; Tables 3-4 and 3-7 respectively). Different patterns of plasma 13C incorporation were observed between the fishes consuming the artificial wood-de tritus and those that were depr ived of food (Figure 3-3). Whereas the 13C incorporation in plasma solutes was described by equation 2 for the fish

PAGE 104

104 consuming the artificial wooddetritus (Table 3-7), a polynomi al distribution described the pattern of incorporation in the food-deprived fish (Figure 3-3). These differences are likely the result of the assimilation of th e soluble components from the artif icial wood-detritus diet. The patterns of red blood cell 13C incorporation were similar between the fish consuming the artificial wood-detritus, those that were food-de prived, and the control fi sh consuming the algae diet, as all three were relativ ely unchanging over the course of the experiment and not welldescribed by equation 2 or linear procedures (Figures 3-2 and 3-4). The patterns of plasma 15N incorporation were similar between the fish consuming the artificial wood-detritus and those that were deprived of food (Fi gure 3-3) both were described by equation 2 (Table 3-8), and the equation for the food-deprived fish was 11.26 2.28e-0.005(time) (r2 = 0.98). Remarkably, the fractional ra te of nitrogen inco rporation (0.005 day-1) was identical in the two groups. However, th e overall change in plasma 15N from the beginning to the end of the experiment was greater for the fish consuming the artificial wood-detritus (mean SD; 1.83 0.66) than for the food-de prived fish (1.24 0.23; t = 1.74, P = 0.10, D.F. = 17). Nevertheless, the loss of wei ght and the pattern of plasma 15N incorporation in the fish consuming the artificial wood-detritus indicate that these fish were at le ast partially starving and in negative nitrogen balance. This is further corroborated by th e hepato-somatic indices of the fish, which were significantly lower in the f ood-deprived fish and in those consuming the artificial wood-detritus than in those consuming algae (Figure 3-5). Protein reserves were definitely mobilized in the food-deprived fish, as the %C:%N ratio in thei r liver was significantly greater than in the algae-fed fish (Figure 3-5). However, the %C:%N ratios in the livers of the fish consuming the artificial w ood-detritus were indistinguishable from either group, suggesting that these fish were assimilating at least some nitrogen from the corn gluten meal and lysine

PAGE 105

105 (Table 3-1) in their diet. Nevertheless, the rate of 15N incorporation in the plasma solutes of the wood-fed and food deprived fish was identical ( = 0.005 d-1; Table 3-8), and this stands in starck contrast to the rate of 15N incorporation in plasma solutes from trial 1 on the algae diet ( = 0.213 d-1; Table 3-5). Thus, from an isotopic pers pective, the fish consuming the artificial wood-detritus appeared to be in negative nitrogen balance. The muscle tissue of the fish consuming the ar tificial wood-detritus ha d the following final isotopic signatures (mean SEM): 13C = -23.19 0.24, 15N = 8.40 0.21, 13Ctissue-diet = 1.40, 15Ntissue-diet = 5.02. The fish consuming the alg ae diet had the following final isotopic signatures in their muscle tissue: 13C = -22.03 0.21, 15N = 7.39 0.16, 13Ctissue-diet = 0.01, 15Ntissue-diet = 5.13. Wild-Caught Fish and Resources The wood-eating catfishes from the upper Ro Maraon, Per, had 13C signatures consistent with the assimilation of cellulosic ca rbon, and were isotopically different from the other loricariid species in the genera Lamontichthys and Spatuloricaria (Figure 3-6). The 13C of plasma solutes, red blood cells, and muscle tissue of Panaque cf. nigrolineatus Maraon, P. nocturnus, and Hypostomus pyrineusi were all consistent with that of cellulose. Furthermore, assuming a 13Ctissue-diet of 1.75 for fin tissue, the 13C of the fins is also consistent with a diet of cellulose carbon (~26.5; Figure 3-6). The w ood-eating catfish are cl early enriched in 13C compared to biofilm and bulk wood, and are deplet ed compared to seston (Figure 3-6). All three wood-eating catfish species had stable isotopic signatures that varied from one another, especially for 15N. The two species of Panaque had similar 13C signatures, but P. cf. n. Maraon had enriched 15N signatures in comparison to P. nocturnus and H. pyrineusi which in turn were different from each another (Figure 3-6). When including other animal taxa in the

PAGE 106

106 carbon-nitrogen dual-isotope plot, all of the wood-eating catfishes (including P. albomaculatus and P. gnomus in addition to those listed a bove) cluster together and have 15N signatures that are likely too enriched to reflect solely the assimilation of wood or of nitrogen fi xation occurring in their digestive tracts (assuming 15Ntissue-diet 4-5). Pterygoplichthys disjunctivus is not likely utilizing cellu losic carbon as a food source in Floridian spring habitats (Fi gure 3-7). This fish has 13C signatures that are more depleted than periphyton or coarse benthic orga nic matter, although, by assuming a 15Ntissue-diet of 4-5, these fish may be getting some nitr ogen from periphyton sources. Because I was not exhaustive in my collection of potential resour ces it is difficult to discern the exact carbon sources for Pt. disjunctivus but we can be certain it is not from wood or wood cellu lose. When including other animal taxa in the carbon-nitrogen du al-isotope plot, it is clear that Pt. disjunctivus is feeding at a lower trophic level than all of the common fish species observed in the spring habitat (Figure 37). However, the Wekiva Springs complex in Or lando, Florida is an urba n site and is clearly enriched in 15N in comparison to the Peruvian rive rine habitat (Figures 3-6 and 3-7). Discussion To my knowledge, this study is the first to report on the isotopic incorporation and discrimination factors for plasma solutes, red bl ood cells, and fin tissue in a fish species. In comparison to mammals and birds, fishes (and ect otherms in general; Reich et al. 2008) are less well-studied in this regard. Most of my predicti ons were supported by the results of this study. First, Pt. disjunctivus clearly cannot assimilate carbon from the fibrous portions of wood, and probably doesnt use wood as a re source in the wild. Second, w ild-caught wood-eating catfishes from Per had 13C signatures consistent with the assimila tion of cellulosic carbon. And third, the 15N signatures of these same wild-caught wood-eat ing catfishes were too enriched to reflect

PAGE 107

107 solely the assimilation of wood, or of nitrogen fixed by endosymbi onts in their digestive tracts. Overall, the data presented here support my la rger hypothesis that woodeating catfishes in the family Loricariidae are not capable of digesting w ood in their digestive tract s, and instead rely on soluble wood-degradation products produced by microbial digestion of wood in the environment (Chapter 2). This study is one of very few to determine the effects of growth versus catabolic tissue turnover to isotopic incorporati on in fishes. Hesslein et al (1993) observed very little contribution of catabolic tissue turnover to isotopic incorporation (~0.002 d-1) in liver and muscle of whitefish ( Coregonus nasus ), and Trueman et al. (2005) observed isotopic incorporation rates due to catabolic tissue tur nover of approximately 0.005 d-1 in muscle and liver of growing Atlantic salmon ( Salmo salar ). However, the C. nasus (0.05 d-1) and the S. salar (0.02 d-1) in those studies were growi ng at faster rates than Pt. disjunctivus in this study (0.0017 d-1). Obviously, growth will play a more significant role in isotopic incorporation in more rapidly growing animals, especially in juveniles (Gamboa-Delgado et al 2008; Reich et al. 2008). The Pt. disjunctivus used in this study were all adults, as ripe ovaries and testes were observed upon dissection of many of the wild fi sh, and some fish at the end of the experiment, and this may explain why they did not grow as rapidly as animals in previous studies that focused on juveniles. The rates of isotopic incorporation varied by tissue in Pt. disjunctivus with plasma solutes and fin tissue turning over more quickly than red blood cells for 13C, and plasma solutes and red blood cells turning over mo re quickly than fin tissue for 15N. The half-lives for 13C in these tissues (calculated as ln(2)/ ) ranged from 14 to 116 days, and from 3 to 22 days for 15N. Differences in turnover rates among tissues ar e not uncommon in fish es and other animals

PAGE 108

108 though. For example, MacAvoy et al. (2001) obser ved large half-life values of approximately 170 days for 13C and 15N in channel catfish ( Ictalurus punctatus ) muscle and whole blood, whereas Sakano (2005) observed half-liv es ranging from 24 to 65 days for 15N in sockeye salmon (Oncorhynchus nerka) muscle. McIntyre and Flecker (2006) estimated half-lives of 17 and 12 days for 15N in muscle and fin, respectively, in the loricariid catfish Ancistrus triradiatus. Sand goby ( Pomatoschistus minutus ) muscle and liver exhibited half-lives of 25 and 9 days, respectively, for 13C, and 28 and 3 days, respectively, for 15N (Guelinckx et al. 2007). Thus, the turnover rates observed in Pt. disjunctivus fall within the range reported for other fish tissues. Given the slow turnover of 167 days for 13C in red blood cells, it is difficult to imagine how the 15N turnover rate of less than five da ys was possible for this tissue in Pt. disjunctivus The few previous investigations of whole-blood (RBC + plasma) 15N turnover in fish (>250 days, MacAvoy et al. 2001; 25 days, McIntyre and Flecker 2006) varied widely, but were still much slower than five days. Furthermore, my observation of a turnover rate of 167 days for 13C in red blood cells is consistent with slow blood isotopic turnove r in channel catfish (MacAvoy et al. 2001). Thus, somethi ng appears to be wrong with the 15N turnover observed for red blood cells in this study. Several possi ble methodological issues can explain this rapid turnover, and all of them involve an err oneously high starting point for red blood cell 15N. First, the isotope ratio mass spectrom eter can occasionally produce incorrect 15N values, and if unnoticed, may have resulted in inflated starting 15N values. Second, there may have been incomplete separation of plasma and red blood ce lls during the centrifugation process, resulting in plasma proteins contamina ting the red blood cell samples; this may have produced unusually high starting 15N values and rapid turnover. And third, there is some unknown contaminant in

PAGE 109

109 the red blood cell samples skewing the results. The separation of the red blood cell and plasma fractions of the blood was very clear, so I do not think that option two is likely. Therefore, option one or three appear to be the probable po ssibilities. Both of these can be addressed by simply running the red blood cell samples through the isotope ratio mass spectrometer a second time to observe whether they are erroneously hi gh. A change of just 1 in the starting 15N of the red blood cells can result in a reduction of the turnover rate from five days to 25 days. Thus, the implications are large and should be addressed. However, they will not be addressed for this dissertation. If the 15N turnover rate is found to be corr ectly rapid for red blood cells in Pt. disjunctivus the mechanism for this turnover, and the complete independence from 13C turnover need to be investigated. The 13C of the fin tissue turned over significantly more quickly than the plasma solutes (20 days vs. 36 days; t = 3.71, P = 0.021, d.f. = 4), but the opposite was true for 15N (32 days vs. five days; t = 3.46, P = 0.026, d.f. = 4). I consiste ntly sampled the same fin tissue from each individual fish every 30 days, a nd thus, the fin tissue used in th is study was regenerated tissue as opposed to undamaged, steady-state fin tissue (Wills et al. 2008). The regeneration process which involves several different cell types, cell migration, and substantial new tissue synthesis (Wills et al. 2008) may have caused the differe nces in isotopic incorporation among the fin tissue and the plasma solutes. In the only other study of fish fin isotopic turnover, McIntyre and Flecker (2006) sampled undamaged fi n tissue from different individual Ancistrus triradiatus for each of their sampling intervals. In terestingly, these authors observed 15N turnover rates of about 18 days for undamaged fin tissue, a rate that is nearly half of wh at I observed. This suggests one of two scenarios: either the catfish studied by Mc Intyre and Flecker (2006) were growing at a faster rate than Pt. disjunctivus in this study, resulting in a larger contribution of

PAGE 110

110 growth to isotopic turnover; or the fin regene ration process in this study slowed the isotopic incorporation of nitrogen from the fishs diet into the newly generated fin tissue. McIntyre and Flecker (2006) did not measur e fin carbon turnover, and, to my knowledge, no studies have measured carbon isotopic turnover in fish fin tissue. Thus, I am left with no explanation for how the 13C turnover over rate is faster in fin (regenerated or non-re generated) than in plasma solutes. Nevertheless, the obs ervation that regenerated fin 15N turned over more slowly than non-regenerated fin, suggests that my findings are more conservative than those of McIntyre and Flecker (2006), and thus, may still apply to turnover rates of fin tissue in wild fish. Additionally, differences in turnover rates of 13C and 15N in a single tissue are not uncommon in fishes. For example, Guelinckx et al. (2007) observed 15N turnover rates that were 3X faster than 13C turnover rates in sandgoby liver. Thus, the obs ervation that plasma solutes had faster 15N than 13C turnover is not unknown, especially in rapidly cycling tissues like plasma and liver. Differences in turnover rates among tissues are not consistent and can disappear in rapidly growing ectotherms. For example, Reich et al. (2008) observed half-lives ranging from 27-35 days in five different tissues of rapidly growi ng loggerhead turtle hatchlings, illustrating that turnover rate can become somewhat homogenous if the animal is growing sufficiently fast. Obviously, this can change the interpretation of field isotopic data if the animal is growing quickly and argues that isotopic da ta gathered in the laboratory s hould not only target the species of interest, but also should attempt to mimic growth conditions experienced by the animals in nature. Because I did not measure growth of the catfish in the wild, it is difficult to discern whether Pt. disjunctivus were growing at similar rates in captivity and in nature. Growth rate varies as an allometric functi on of body mass (West et al. 2001), and thus, the rate of isotopic incorporation in an animals tissues should also vary with body size. This

PAGE 111

111 hypothesis was supported in house sparrows ( Passer domesticus ), which exhibited rates of isotopic incorporation that varied as an allometric function of body mass-0.25 (Carleton and Martnez del Rio 2005). In ectothermic vertebra tes, which are indeterminate growers, it is, therefore, essential to study animals in the labora tory that are of a similar mass to those studied in the field. I was able to do so, as I only studied adult animals, and the average mass of Pt. disjunctivus used in the laboratory (51.29 g) was si milar to the wild-caught fish in Orlando (61.93 g) and Per (31.55 g). Another caveat in the consideration of growth is that some fish tissues only reflect the is otopic signature acquired during periods of growth. For example, Perga and Gerdeaux (2005) i llustrated that the 13C and 15N of whitefish muscle were relatively unchanging during the non-growi ng season, and reflected the isotopic signature of food consumed during the growing season. However, liver 13C and 15N changed with diet across season regardless of growth. The same could be observed in Pt. disjunctivus in this study by comparing the patterns of 13C incorporation between red blood cells and plasma solutes. In trial two, when the fish were not growing, the 13C of red blood cells changed very little, whereas the plasma solute 13C changed considerably. Thus, when a ttempting to use stable isotopes to discern the diets of fishes that grow seasonally, it may help to use multiple tissues, some that turnover slowly, reflecting the longer-term signal, and some that turnove r quickly, to address more recent dietary changes (Reich et al. 2008). In this regard, blood makes the perfect tissue to sample non-invasively, as it provides two tissues that turn over at different rates plasma solutes, which are primarily synthesized in the liver (Turner and Hulme 1970; Adkins et al. 2002) and turnover quickly, and red blood cells, wh ich turnover slowly. Fin tissue may also provide an additional tissue that turns over at a different rate than plasma solutes.

PAGE 112

112 Correctly observing seasonal changes in diet is relevant to loricariid catfishes, as they experience wet and dry seasons each year in the Amazonian basin, which translates into drastic changes in water level, habitat, and food ava ilability (Fink and Fink 1979). Nonogaki et al. (2007) measured the 13C in otoliths of grazing and wood-eat ing catfishes from the Brazilian Amazon. Because otoliths are laid down in annua l growth increments, the isotopic signature of each otolith annulus can be used to monitor changes in 13C that occur over time. Indeed, Nonogaki and colleagues illust rated that the wood-eating Panaque nigrolineatus had 13C signatures that were relatively unchanging over time, whereas a gr azing loricariid with a more variable diet, Hypostomus regani showed temporal changes in otolith annulus 13C. In agreement with this, all of the wild-ca ught wood-eating catfishes in this study ( P. cf. nigrolineatus Maraon, P. nocturnus, and H. pyrineusi ) had red blood cell and plasma solute isotopic signatures that were similar to one another for each species, showing similarity in shortterm and long-term dietary history. This may not be surprising, however, as wood-detritus is available in flooded forests duri ng the wet season and in main river channels during the dry season. The 13Ctissue-diet values observed in trial one of this study varied by tissue, but were relatively consistent with discrimi nation factors of -1.5 to 3.4 reported for different tissues in the literature (Reich et al. 2008, and references therein). Plasma solutes and red blood cells had 13Ctissue-diet values indistinguishable fr om the diet, whereas fin tissue was 1.75 enriched over lipid-extracted diet. The 15Ntissue-diet values observed in Pt. disjunctivus tissues (Table 3-5) were higher than the widely used 3.4 for this is otope (Fry and Sherr 1984; Focken 2004; Fry 2007), but not inconsistent with 15Ntissue-diet in herbivorous animals (Robbins et al. 2005). The major contributing factors to 15Ntissue-diet are growth rate and dietar y protein quality (Karasov and

PAGE 113

113 Martnez del Rio 2007), with the latter defined as biological value, or how well the amino acid profile of the dietary protein m eets the needs of the animal (R obbins et al. 2005). Herbivorous mammals generally consume protein that is cons idered low-quality, and hence have larger 15Ntissue-diet than carnivores (Robbins et al. 2005). M ill et al. (2007) extended this argument for fishes showing that herbivores with high-in take of a low-protein diet have larger 15Ntissue-diet (> 4) because of the low quality of the protein in the food. Undoubtedly, there were differences in the amino acid profile of the algal diet fed to Pt. disjunctivus in this study and the amino acid requirements of the fish, as the 15Ntissue-diet varied from 4.08 in fin tissue to 5.13 in muscle. I do not know the quality of the protein available to the fish in the wild, but am confident assuming it is equal to or lower in quality than the protein in the algal diet offered to the fish in the lab. Therefore, I feel safe assuming the 15Ntissue-diet values observed in the lab translate into the field. In this regard, it appears possi ble to predict the tr ophic standing of the wild-caught woodeating catfishes based on their is otopic signatures. According to Figure 3-6, all of these species consume carbon sources consistent with the isot opic signature of cellulose. However, the 15N signatures indicate that they ar e clearly enriched >6 over the 15N of biofilm or bulk wood, suggesting one of two things: first, the protei n quality in the wild is so low that the 15Ntissue-diet is larger than that observed in the lab; or second, they are getting their nitrogen from some intermediate source between the wood and the fish such as wood-degrading microbes or insects. Bostrm et al. (2008) and Kohzu et al. (1999) showed that wood-degrading fungi can have 15N values ranging from -2 to 4, depending on the w ood being digested. If the fungi degrading the wood in the upper Ro Maraon have 15N values ranging from 2 to 4, and I assume the observed 15Ntissue-diet of 4 to 5 for the fish, then the predicted 15N for the wood-eating

PAGE 114

114 catfishes would be 6-9, exactly the range in which the fish fall in the wild. However, additional sampling to observe the 15N of the wood degrading fungi would be necessary to confirm or refute this assertion. Furthermore, species of Spatuloricaria are known to consume insects (de Melo et al. 2004), and th e wood-eating catfishes have simlar 15N signatures to a sympatric species of Spatuloricaria collected from the Ro Maraon (Figure 3-6). This suggests that the alleged wood-eating catfishes may supple ment their diet with protein from animal sources. True wood-eating termites that rely upon nitrog en-fixing bacteria in their digestive tracts (Slaytor and Chappell 1994) te nd to have more depleted 15N values because of the depleted 15N signature of fixed nitrogen (Tayasu et al. 1997). Not only do they have lower 15N signatures (2-4), true wood-eating termites also have 15Ntissue-diet values near zero over that of wood, whereas detritivorous termites have larger 15Ntissue-diet values (>3), and 15N values near 8 (Tayasu et al. 1997), consistent with the wood-eating catfishes in this study. Thus, it does not appear likely that w ood-eating catfishes rely upon nitroge n fixation in their digestive tracts, despite the presence of species of Spirochetes in their guts, as has been suggested (J.A. Nelson, pers. comm.). Furthermore, I did not observe any conglomerations of microbes anywhere along the digestive tracts of the wood-eating catfishes or of Pt. disjunctivus (Chapter 2), suggesting that any microbes isolated from th eir guts (Nelson et al. 199 9) are ingested with their detrital diet as opposed to being endosymbiotic. Wood-degrading fungi in streams ar e approximately 2 enriched in 13C in comparison to bulk wood (Kohzu et al. 1999; Bost rm et al. 2008), consistent with cellulose digestion. Therefore, digestion of fungi may provide an ad ditional avenue for the wood-eating catfishes to have 13C signatures on par with that of wood-cellulose. This is further supported by the wood-

PAGE 115

115 eating catfishes having elevated N-acetyl-D-glucosaminidase activities in their digestive tracts, and >1 mM N-acetyl-glucosamine in their intestin al fluid, both of which suggest digestion of fungal cell walls (Chapter 2; German et al. 2008). Pterygoplichthys disjunctivus is clearly feeding at a lower trophic level than other fishes in Florida spring habitats (Figure 3-7). Gut content analyses of th is species suggest they consume 45% green algae and diatoms and 45% detritus (Cha pter 2). Green algae (Evans-White et al. 2001) and diatoms (Nichols and Garling 2000) ca n individually be more depleted in 13C (< 30) than the signature observed for periphyton (-29.26) in this study, which may explain the depleted 13C (-30.3) observed in Pt. disjunctivus Furthermore, I do not know the isotopic signature of fine benthic organic ma tter (detritus) in the spring habitats. However, what is clear is that despite all of the similarities between Pt. disjunctivus and the wood-eating catfishes in terms of gut morphology and digestive physiology (Chapter 2), the former does not rely upon wood-carbon in the wild, at leas t in Florida. The enlarged spoon-shaped teeth of the woodeating catfishes allow them to gouge into wood much more efficiently than the villiform teeth of Pt. disjunctivus (Nelson et al. 1999). To my knowledge, this study is only the second to use a diet with ingredients of varying isotopic signatures to monitor a ssimilation of specific nutrients in a fish (Gamboa-Delgado et al. 2008). Gamboa-Delgado et al. (2008) determined the contribution of a readily assimable food item to growth in Senegalese Sole ( Solea senegalensis ). These authors designed a diet that featured one portion composed of Artemia of one isotopic signat ure, and another portion composed of an inert ingredient of a different isotopic signature. They then traced growth and changes in 13C over time, ultimately showing that the in ert ingredient contributed very little to the growth of the fish. I attempted something very similar, but with results that were more

PAGE 116

116 ambiguous. Pterygoplichthys disjunctivus lost weight on the artificial wood-detritus diet and was clearly in negative nitrogen balance. But, the 13C signatures of the fish suggest that they were assimilating at least some carbon from the food; the fish on the artifici al wood-detritus diet clearly had different patterns of carbon isotopic incorporation than those that were deprived of food, and rates of carbon isotopic incorporation on par with the fish fr om the initial turnover experiment with the algal diet. Th is portion of the study suggested that Pt. disjunctivus could not assimilate the refractory polysaccharides from wood in any meaningful quanti ties, a result that is not surprising. In chapter 2, I showed that neither Pt. disjunctivus nor Panaque nigrolineatus (a true wood-eating catfish) could assimilate si gnificant amounts of organic matter (1.72% and 3.35%, respectively) or cellulose (10% and 9%, respectively) from a strictly wood diet in the laboratory, and both species lost weight when cons uming wood. This inability to digest cellulose in their digestive tracts comes from rapid gut transit and low cellulase activities, the latter of which appear to be ingested with their detrital diet rather than produced by endosymbionts in their guts. However, the fish are quite effici ent at digesting soluble components of detritus (starch-like polysaccharides, -glucosides), and disaccharid es from wood degradation ( glucosides, -mannosides; Chapter 2), and thus, this supports my claim that they assimilated some soluble components from the wood in the arti ficial wood-detritus. The similarity between the predicted digestible 13C of the artificial wood-detritus and the 13C observed in the fish tissues further supports this contention (Table 3-7). The fish consuming the artificial wood-detr itus were definitely in negative nitrogen balance despite the presence of corn gluten meal, which is 70% protein and known to be a highly-digestible protein source for detr itivorous fishes (e.g., Nile Tilapia, Oreochromis niloticus ; Guimares et al. 2008), in their food. Fu rthermore, we supplemented the artificial

PAGE 117

117 wood-detritus with L-lysine to account for a deficiency of this amino acid in corn gluten meal (Guimares et al. 2008). However, there are at least two, non-mutually exclusive reasons why the fish eating the artific ial wood-detritus were in negative nitrogen balance. First, the diet was so dilute that I did not offer them enough food on a daily basis to meet their energetic or protein requirements (Raubenheimer and Simpson 1998). And second, the xanthan gum I used as a binding agent compromised the digestibility of th e artificial wood-detritu s via an increase in digesta viscosity (Leenhouwers et al. 2006; Lentle and Janssen 2008). Although I offered the fish 8% of their body mass per day of the artificial wood-detritus, ex ceeding what the fish consume when they naturally regulate intake on a wood di et in the laboratory (2-5% per day; Chapter 2), the food may have been lower in quality than I intended. Nonstarch polysaccharides, such as guar gum and xanthan gum, are commonly used as binding agents in fish feeds. And, at 6% of the total mass of the diet, the chosen concentr ation of xanthan gum was the minimum at which the artificial wood-detritu s pellets remained bound in water. However, this concentration of xanthan gum is also high enough to increase digesta viscosity a nd compromise the digestibility of energy and protein for fish es (Leenhouwers et al. 2006). For example, Leenhouwers and colleagues showed that clariid catfishes consuming diets containing 8% guar gum had decreased energy and protein digestibility in comparison to fish eating a guar gum-free diet, and as a result, the fish consuming th e 8% guar gum food increased intake (up to 3% of their body ma ss per day) to meet their nitr ogen requirements. However, the diet consumed by those fishes was extremely high in protein (~50%), and thus, the fish were still able to meet their energetic and protein demands by increasing intake. Because the artificial wood-detritus offered to the fish in this experime nt started off low in pr otein (7.75%, indicative of detritus; Bowen et al. 1995), a ny compromise of protein digestibility may have resulted in the

PAGE 118

118 fish simply not being able to eat enough to meet their protein requirements. Detritivorous fishes are known to target detrital aggregates that are highest in protein (relatively speaking), or to supplement their low-quality diet with animal mate rial (Bowen et al. 1995). Indeed, the natural diet of Pt. disjunctivus includes about 5% insect s (Chapter 2). Thus, ev en though I attempted to provide adequate intake to the fi sh on the artificial wood-detritus, it simply may have not been enough for them to meet their nitrogen requireme nts given the increased viscosity of the food. However, it is difficult to imagine a different wa y of feeding a benthic fish the artificial wooddetritus short of force-feeding them, which has its own shortcomings (e.g., daily anesthesia, not allowing them to regulate intake, physical damage to the esophagus, etc.). In conclusion, this study is one of the fi rst to integrate laborat ory and field isotopic investigations in fishes, and to show that bl ood and fin tissue can be used as non-invasively sampled tissues for stable isotopic analyses (Kel ly et al. 2006; McIntyre and Flecker 2006). The results clearly show the importance of estimating turnover times and discri mination factors in the laboratory before gathering data in the field. If I assumed the commonly used discrimination factors of 1 for 13C and 3.4 for 15N I might draw very different conclusions from the field isotopic data. The isolation of the more dig estible cellulosic portion of wood provided insight into the carbon sources for the wild-caught wood-eating catfishes that would have been missed by assuming a bulk wood signal for 13C, as was done by Nonogaki et al. (2007). By using the proper positive and negative control groups in the laboratory feeding trial I was able to show that Pt. disjunctivus consuming the artificial wood-detritus simply werent starving, but were definitely in negative nitr ogen balance. The observation that plasma solute 15N values increased in the food-deprived fish argues fo r further evaluation of the effects of food deprivation on 15N signatures in ectothermic animals, wh ich is not always consistent with

PAGE 119

119 observations in mammals and birds (McCue and Pollock 2008). Overall, this study, in addition to my digestive investigations (Chapter 2), s uggests that the wood-eating catfishes are not true xylivores like lower termites and beavers, a nd instead should be called detritivores, which specialize on degraded wood. Within tropical freshwater habitats, wood is important to fish communities, as it provides structure, and a hard surface from which to feed (Wright and Flecker 2004). In this regard, wood in tropical freshwater systems is like the coral reef in a tropical marine system. Parrotfish are reef eroders in tropical marine systems, literally taking bites from the reef in search of detritus (Crossman et al. 2005). Thus, in the Amazon, the wood-eating catfishes are the reef eroders, and can literally be considered the Parrotfish of the Amazon.

PAGE 120

120 Table 3-1. Overall isotopic sign atures of the pelleted algae and artificial wood-detritus diets fed to Ptergoplichthys disjunctivus. Proportions of ingredient s composing the artificial wood-detritus and the isotopic signatures ( 13C and 15N) of each are also shown. Diet component g/100 g 13C 15N Algae diet Total isotopic signature -22.97 0.02 2.43 0.25 Lipid extracted signature -22.02 0.19 2.26 0.16 Wood diet Bulk Wooda 80.0 -26.36 0.25 2.13 0.20 Wood celluloseb (33.6) -25.39 0.12 N/A Soluble wood componentsc (16.0) ? ? Corn gluten meald 9.0 -12.98 0.11 3.26 0.25 Xanthan gume 6.0 -9.64 0.05 1.46 0.01 Corn meald 2.1 -10.65 0.06 4.34 0.31 L-Lysined 1.0 -14.63 0.00 0.01 0.00 Vitamin premixd 1.0 -27.59 1.12 2.63 0.31 Trace mineral mixd 0.5 N/A N/A Water stable vitamin Cd 0.4 -18.90 0.01 N/A Total isotopic signature -23.42 0.18 2.30 0.20 Values are mean ( SEM). a bulk-wood from decomposed water oak ( Quercus nigra ) b cellulose isolated from wood following Gaudinski et al. (2005); proport ion determined on a dry matter basis, and mass presented represents 42 % (Chapter 2) of 80g of bulk wood; does not constitute an additional ingredient. c 20% of water oak wood is composed of solubl e degradation products (C hapter 2), the stable isotopic signatures of this fraction is unknown; does not constitute an additional ingredient d gifts from Hartz-Mountain Corp. and in tended specifically for use in fish food e indigestible non-starch polysaccharide added as a binding agent to keep pellets bound in water

PAGE 121

121 Table 3-2. Resources and animals collected fr om the Ro Maraon, Per, and their respective stable isotopic signatures. Taxa Diet N 13C () 15N () Plant/detrital material Grazed wood 4 -28.76 0.39 1.31 0.08 Cellulose (extracted from wood) 4 -26.90 0.63 N/A Epixylic biofilm 4 -29.25 0.05 1.19 0.06 Seston 4 -21.83 0.36 3.13 0.53 Crustacea* Shrimp Detritus? 3 -25.08 0.18 9.01 0.26 Crabs Detritus? 3 -27.14 0.68 4.47 0.30 Catfishes* Lamontichthys filamentosus Algae/Biofilm?6 -29.09 0.38 8.20 0.17 Spatuloricaria sp. Insects/Seeds2 3 -26.32 0.06 8.72 0.09 Panaque nocturnus Wood/detritus3 6 -26.95 0.09 7.84 0.15 P. cf. nigrolineatus Maraon1 Wood/detritus3 6 -26.15 0.36 8.10 0.20 P. albomaculatus Wood4 6 -26.89 0.22 7.55 0.16 P. gnomus Wood4 6 -27.22 0.52 7.89 0.39 Hypostomus pyrineusi Wood/detritus3 6 -26.79 0.21 7.01 0.16 Isotopes measured in lipid extracted mu scle tissue of crustaceans and catfishes. 1 This is an undescribed species of Panaque belonging to the P. nigrolineatus clade. 2 de Melo et al. (2004) 3 Chapter 2 4 Schaefer and Stewart (1993)

PAGE 122

122 Table 3-3. Taxa collected from Wekiva Springs Florida and their stable isotopic signatures. Taxa Diet N 13C () 15N () Plant/detrital material Duckweed 4 -30.45 0.22 18.91 0.21 Lily pads 4 -26.51 0.10 13.24 0.13 Periphyton 4 -29.26 0.25 7.81 0.31 Coarse Benthic Organic Matter 4 -29.23 0.43 3.43 0.29 Wood 4 -25.24 0.11 2.01 0.20 Cellulose (extracted from wood) 4 -24.09 0.23 N/A Mollusca Snail* Grazer? 3 -28.83 0.15 20.70 0.05 Crustacea Crayfish* Shredder 3 -29.46 0.21 19.58 0.35 Fishes* Lepisosteus platyrhinchus (Florida Gar) Fish 3 -29.22 0.17 15.00 0.31 Micropterus salmoides (Bass) Fish 3 -27.27 0.82 20.31 0.40 Lepomis punctatus (Spotted sunfish) Invertebrates 3 -26.72 0.48 18.56 1.75 L. macrochirus (Bluegill sunfish) Invertebrates 3 -25.96 0.92 16.46 1.77 Gambusia holbrooki (Mosquito fish) Invertebrates 3 -25.07 0.29 24.24 0.38 Lucania goodei (Bluefin killifish) Invertebra tes 3 -30.36 1.51 22.13 1.02 Heterandria formosa (Least killifish) Invertebrate s 3 -29.02 0.48 21.51 0.33 Pterygoplichthys disjunctivus Algae/Detritus 9 -28.62 0.65 11.51 0.42 Isotopes measured in lipid extracted muscle ti ssue of molluscs, crustaceans, and small fishes ( G. holbrooki, Lu. goodei H. Formosa ), and lipid extracted fin clips from larger fishes.

PAGE 123

123Table 3-4. The isotopic incor poration of carbon from an algal di et (trial one) into tissues of Pterygoplichthys disjunctivus using the equation: 13C(t) = 13C ( ) + [ 13C (0) 13C ( )]et. Tissue Equation vs. kgt t -test 13Ctissue-diet Average residence time (days) Plasma solutes -22.57 6.99e-0.026(time) 13.54** 0.12 0.08 38.46 9.62 Red blood cells -20.74 8.05e-0.006(time) 4.15* 0.13 0.30 166.67 10.56 Fin clips -20.57 5.66e-0.050(time) 9.82** 1.75 0.15** 20.00 2.80 Note: and ** indicate significant differen ces (P < 0.05, and P < 0.001, respectively) between the fractional rate of isotopic incorporation ( ) and the growth rate (kgt = 0.0017 day-1) for all tissue types. 13C is the mean ( SE) diet-tissue discrimination factor, and ** indicates significant differen ce from 0 with 1-sample t-test (P < 0.001); no difference from 0 was detected for 13Ctissuediet for plasma solutes or red blood cells. Because the fin clip tissue was lipid extract ed prior to analyses, the 13Ctissue-diet for fin clips were calculated against lipid extracted diet, whereas the 13Ctissue-diet for plasma solutes and red blood cells were calculated against non-lipid extracted di et. Average residence ( SE) time was estimated as 1/

PAGE 124

124Table 3-5. The isotopic incorpor ation of nitrogen from an algal diet (trial one) into tissues of Pterygoplichthys disjunctivus using the equation: 15N(t) = 15N ( ) + [ 15N (0) 15N ( )]et. Tissue Equation vs. kgt t -test 15Ntissue-diet Average residence time (days) Plasma solutes 6.97 + 3.57e-0.213(time) 33.58** 4.39 0.05* 4.69 0.24 Red blood cells 8.42 + 5.83e-0.216(time) 31.66** 5.17 0.13* 4.63 1.94 Fin clips 6.51 + 4.48e-0.031(time) 7.20** 4.08 0.14* 32.26 6.24 ** indicates significant difference (P < 0.001) between the fractional ra te of isotopic incorporation ( ) and the growth rate (kgt) for all tissue types. 15Ntissue-diet is the mean ( SE) diet-tissue discrimination factor and indicates significan t difference from 0 with 1sample t-test (P < 0.001). Average re sidence time ( SE) was estimated as 1/

PAGE 125

125 Table 3-6. Isot opic signatures (13C and 15N) of three tissues of wild-caught Pterygoplichthys disjunctivus collected from Wekiva Springs, Florida during the course of the laboratory feeding experiment. Day 13C Plasma 13C RBC 13C Fin 15N Plasma 15N RBC 15N Fin 1 -29.21 0.42a -29.26 0.44 -25.85 0.82a 12.82 0.14 13.12 0.45 11.09 0.26 100 -31.78 0.79b -31.40 0.81 -29.68 0.40b 11.14 1.18 11.54 1.49 12.22 0.54 210 -31.55 0.51ab -31.26 0.44 -29.60 1.14b 12.82 0.14 13.12 0.46 10.73 1.17 F2,8 = 5.76 F2,8 = 4.16 F2,10 = 8.41 F2,8 = 1.96 F2,8 = 0.96 F2,10 = 1.33 P = 0.040 P = 0.074 P = 0.011 P = 0.221 P = 0.436 P = 0.317 Values are mean ( SEM). Isotopic values were compared across days for each tissue with ANOVA followed by a Tukeys HSD with a family error rate of P = 0.05. Values for a particular isotope and tissue that share a superscript letter are not significantly different. RBC = red blood cells.

PAGE 126

126Table 3-7. The isotopic incorporation of carbon from an artificial wood-detritu s diet (trial two) into tissues of Pterygoplichthys disjunctivus Tissue EquationA r2 13Ctissue (final) 13Cdiet (final)predicted B Plasma solutes -21.16 1.83e-0.023(time) 0.71 0.41 Red blood cells -20.17 2.52e-0.002(time) 0.48 -0.24 Fin clips 0.014(time) -20.72 0.87 2.18 A The non-linear procedure using equation 13C(t) = 13C ( ) + [ 13C (0) 13C ( )]et described the isotopic incorporations for plasma and red blood cells reasonably well, but a linear procedure was necessary for the fin clip samples. B This value represents the difference between the final 13C observed for the tissues at the end of the experiment minus that predicted by a linear, concentration-dependent mixing model (Martnez del Rio and Wolf 2005) fo r the artificial wood-detritus ingredients plus the expected 13Ctissue-diet (i.e., the 13Ctissue-diet observed in trial one) for each tissue; 13C (final)predicted = -21.79.

PAGE 127

127Table 3-8. The isotopic incorporation of nitrogen from an artificial wood-detr itus diet (trial two) into tissues of Pterygoplichthys disjunctivus Tissue EquationA r2 15Ntissue (final) 15Ndiet (final)predicted B Plasma solutes 10.12 3.03e-0.005(time) 0.83 0.79 Red blood cells 0.011(time) + 7.21 0.82 -0.64 Fin clips 8.57 2.15e-0.021(time) 0.94 1.29 A The non-linear procedure using equation 15N(t) = 15N ( ) + [ 15N (0) 15N ( )]et described the isotopic incorporations for plasma and fin clips reasonably well, but a linear pr ocedure was necessary for the red blood cell samples. B This value represents the difference between the final 15N observed for the tissues at the end of the experiment minus that predicted by a linear, concentration-dependent mixing model (Martnez del Rio and Wolf 2005) fo r the artificial wood-detritus ingredients plus the expected 15Ntissue-diet (i.e., the 15Ntissue-diet observed in trial one) for each tissue; 15N (final)predicted = 3.25.

PAGE 128

128 Figure 3-1. Changes in 13C and 15N in Pterygoplichthys disjunctivus 0-203 days after a diet switch. Values are mean ( SEM). Curv es were fit by a nonlinear routine with the equation X( t ) = X( ) + [ X(0) X( )]et (see Statistical Analyses). -32 -30 -28 -26 -24 -22 -2004080120160200 -32 -30 -28 -26 -24 -22 -20 -32 -30 -28 -26 -24 -22 -20 5 7 9 11 13 15 5 7 9 11 13 15 5 7 9 11 13 1504080120160200a) Plasma 13C b) Plasma 15N c) Red blood cells 13C d) Red blood cells 15N e) Fin 13C f) Fin 15N Time (days)

PAGE 129

129 Figure 3-2. Changes in 13C and 15N in Pterygoplichthys disjunctivus across 155 days while consuming an artificial wood-detritus diet or an algal diet (positive control). Values are mean ( SEM). The solid lines represent the 13C and 15N signatures of the algal diet, whereas the dashed lines represent the 13C and 15N signatures of the bulk wood in the artificial wooddetritus diet. In the car bon plots, the dash-dot line represents the 13C of cellulose isolated from the wood. In the nitrogen plots, the dotted line represents the 15N signature of the corn products representing the soluble component of the artificial wood-detritus diet. The 13C of the corn products (11.82) is off the scale of the carbon plots. Equations for all relationships presented in Tables 7 and 8. 2 4 6 8 10 b) Plasma 15N 2 4 6 8 10 d) Red blood cells 15N 2 4 6 8 10020406080100120140160 f) Fin 15N -27 -25 -23 -21 -19 -17 c) Red blood cells 13C e) Fin 13C -27 -25 -23 -21 -19 -17020406080100120140160 -27 -25 -23 -21 -19 -17a) Plasma 13C Wood diet Algae diet (Control) Time (days)

PAGE 130

130 Figure 3-3. Changes in 13C and 15N in plasma solutes of Pterygoplichthys disjunctivus across 155 days while consuming an artificial wood-detritus diet or deprived of food (negative control). Values are mean ( SEM). Curves for the fish consuming the artificial wood-detritus, and the 15N of the food-deprived fish were fit with the equation X( t ) = X( ) + [ X(0) X( )]et, whereas the 13C of the food-deprived fish was best fit by a polynomial distribution (y = 0.00009x2 + 0.009x 24.09; r2 = 0.97). -26 -24 -22 -20Time (days) 6 8 10 12 -26 -24 -22 -20 0 40 80120160 6 8 10 120 40 80120160a) Plasma 13C wood diet b) Plasma 15N wood diet c) Plasma 13C food-deprived d) Plasma 15N food-deprived

PAGE 131

131 Figure 3-4. Changes in 13C and 15N in red blood cells of Pterygoplichthys disjunctivus that were deprived of food across 155 days. Symbols represent measurements in an individual fish, and lines are provided to trace how individual fish changed over the course of the experiment. 8 10 12 14 160255075100125150Time (days) -32 -30 -28 -26 -24 -2213C15N

PAGE 132

132 Figure 3-5. Hepato-somatic indices and %C:%N ratios in the livers of Pterygoplichthys disjunctivus fed an artificial wood-detritus diet, an algal diet, or those that were deprived of food. Values are mean ( SEM). Values compared among groups for an index with ANOVA followed by Tukeys HSD with a family error rate of P = 0.05. Bars that share a letter ar e not significantly different. 0 2 4 6 8 0.000 0.002 0.004 0.006 0.008 0.010 0.012Wood diet Algae diet Food-de p rived%C:%NHepato-somatic indexa a b ab a b

PAGE 133

133 Figure 3-6. Carbon and nitrogen dual-isotope plots of animals and resources collected in the upper Ro Maraon, Per a) shows the different tissues of only wood-eating catfishes and potential resour ces, b) shows wood-eating catf ishes in addition to other animals. Only lipid-extracted muscle tissue was used to analyze the isotopic signatures of the animals in plot b. Pm = Panaque cf. nigrolineatus Maraon, Pn = P. nocturnus Pg = P. gnomus, Pa = P. albomaculatus Hp = Hypostomus pyrineusi 0 1 2 3 4 5 6 7 8 9 10 -30-29-28-27-26-25-2413C15Na) Wood-eating catfishes Pm Pn Hp Biofilm Wood (bulk) Cellulose Fin Muscle Red blood cells Plasma 0 1 2 3 4 5 6 7 8 9 10 -30-28-26-24-22-2013C15Nb) Wood-eating catfishes, ot her animals, and resources wood-eating catfishes Lamontichthys Spatuloricaria Shrimp Crab Biofilm Wood (bulk) Cellulose Seston Pm Pn Hp Pg Pa Fin tissue

PAGE 134

134 Figure 3-7. Carbon and nitrogen dual-isotope plots of animals and resources collected Wekiva Springs, FL, USA a) shows the different tissues of only Pterygoplichthys disjunctivus and potential resources, b) shows Pt. disjunctivus in addition to other animals. Only lipid-extracted fin tissue (large fish) and muscle tissue (invertebrates and small fish, see Table 3) was used to analyze the isotopic signatures of the animals in plot b. CBOM = coarse benthic organic matter. 0 4 8 12 16 20 -32-30-28-26-24a) Pterygoplichthys disjunctivus Fin Muscle Red blood cells Plasma CBOM Periphyton Wood (bulk) Cellulose Lilypad Duckweed 13C15N 0 4 8 12 16 20 24 28 -32-30-28-26-24b) Pt. disjunctivus other animals, and resources Wood (bulk) Cellulose CBOM Periphyton Pt. disjunctivus Gar Bass Spotted sunfish Bluegill sunfish Lilypad Duckweed Cra y fish Snail Blue-fin killifish Least killifish Mos q uito fish 13C15N

PAGE 135

135 CHAPTER 4 CONCLUSIONS The purpose of this dissertation was to e xplore th e capabilitie s of xylivorous and detr itivorous catfishes to digest and assimilate wood, and to determine the extent these animals rely on endosymbiotic microorganisms in this process. The data presented in chapter two unequivocally support the null hypothes is that these fishes do not and cannot digest wood in their digestive tracts in contrast to some of the be tter known xylivores (e.g., termites, beavers). The following key points all suggest that the alleged xy livorous catfishes are actu ally just detritivores and digest primarily soluble components of thei r diet rather than cellulose and other cell wall polysaccharides: They have extremely long, narrow digestive tracts (11-18X their body lengths) with no kinks, valves, or caeca to slow th e flow of digesta through the gut. The intestinal folding patterns and microvilli surface area decrease moving distally along the intestine suggesting that most absorption o ccurs in the proximal and mid intestine. This is corroborated by soluble carbohydrates be ing detectable exclusively in the proximal and mid intestines of the fish and disappearing in the distal intestine, indicating that most sugars are actually absorbed in the proximal and mid intestine. The pH conditions in the intestine were alka line, and the redox potentials were clearly negative, signifying that the fishes intestines are anaerobic. However, the absence of any conglomerations of microbes in the TEM imag es indicates that th e fish do not harbor endosymbionts in their guts. The extremely low and unchanging SCFA concentrations along the fishes intestines suggest that, despite the anaerobic conditions, fe rmentation of cellulose is not occurring at a rapid pace, and is not likely a mechanis m for cellulose digestion in these fishes. The low and variable cellulase and xylanase activities, and the fact that these activities did not increase towards the fishes distal intestin es signify that these enzymes are ingested with the food (i.e., decaying wood) rather than produced by endosymbionts. In fact, almost without exception, enzy me activities in the microbial extracts decreased moving distally along the intestine, supporting this supposition. Activity levels of enzymes that digest soluble polysaccharides (i.e., amylase, laminarinase) were one to five orders of magnitude greater than the cellulase a nd xylanase activities, indicating that the fishes prefer entially digest soluble polysaccha rides rather than refractory ones.

PAGE 136

136 The soluble components of wood degradation (i.e., -glucosides, -mannosides) were efficiently digested and assimilated by th e fish. This is especially true for -glucosidase, as the Km values of this enzyme in the gut walls of the fish were an order of magnitude lower than the Km values in the microbial extracts. Chitinous compounds may be important energy and nitrogen sources to the fish as significant amounts (>1 mM) of N-acetyl-gluco samine (the monomer of chitin) were detected in the intestinal fluids of the fish. Furthermore, the fishes N-acetyl-Dglucosaminidase (NAG) activities were elevated compared to the microbial extracts. The elevated protease activities (trypsin a nd aminopeptidase), and the heavily skewed trypsin:lipase ratios (400:1) in the fishes al imentary tracts are consistent with other detritivorous animals rather than xylivorous ones. The fish passed wood through the gut too quickly (< 4 hours) to allow symbionts to digest cellulose, and there was no retention of small particles along the gut. Consequently, the digestibilities of cell wa ll compounds (NDF, ADF, lignin) were low in the loricariid catfishes, and th e fish lost weight on a wood di et, showing that they cannot digest and thrive on a wood diet. In chapter three, I attempted to discern, us ing stable isotopes, wh ether the wood-eating and detritivorous catfishes could assimilate carbon from wood and to what extent they could subsist on a wood diet. The results from this part of the study also clearly s uggest that loricariid catfishes primarily digest the soluble components of their diet. The key findings of this chapter were: From an isotopic ( 13C and 15N) standpoint, that the non-inva sively sampled tissues of plasma solutes, red blood cells, and fin tissu e turnover and match the isotopic signature of the diet on sufficiently different time scales to be used to track the diets of fishes in the wild. Laboratory investigations of wood assimilation using stable isotope s indicate that the detritivorous fish, Pterygoplichthys disjunctivus is incapable of assimilating structural polysaccharides from wood, and that they can only assimilate the more soluble components of wood detritus. Wild-caught wood-eating catfishes from Per do show that they are assimilating carbon ( 13C) with a baseline consistent with that of cellulose, but that their nitrogen ( 15N) signature is too enriched to re flect a digestive strategy consis tent with true wood digestion via endosymbionts (like termites and beavers).

PAGE 137

137 Wild-caught Pt. disjunctivus from Orlando, FL, are feeding at a lower trophic level than other native Floridian fishes, and appear to be using detritus, periphyton, and diatoms as a resource as opposed to wood like the true wood-eating catfishes. The bottom line of this dissertation is that each and every finding in this study says the same thing: the structure and function of the di gestive tract of the wood-eating catfishes is not different from detritivorous ones and neither th e wood-eating nor the detr itivorous species are capable of digesting wood in any significant am ount. However, these fish do consume food, detritus, which is in the process of being degrad ed by microbes in the environment. The fishes take advantage of this degradation by si phoning off soluble components of environmental microbial wood degradation (e.g., -glucosides, -mannosides), and probably by digesting environmental microbes themselves. Swift et al. (1979) called this kind of feeding pattern the external rumen, making reference to the specializ ed forestomach of a cow. However, rather than harboring microbial endosymbionts in a sp ecialized region of the gut and reaping the benefits internally, the fish allow the microbial action to occur outside of their bodies. Then, they consume the detritus, and take what they are equipped to, the soluble components, and excrete the rest in their feces. Thus, these fishes should not be referred to as xylivorous, but rather as wood-eating detritivores. The way in which these fishes feed is likely very important ecologica lly in the context of nutrient cycling. Wood-eating detritivorous cat fishes take a coarse form of detritus (woody debris) and reduce it to particles generally less than 1-mm in diameter. Furthermore, they add nitrogen to the excreta. Thus, by increasing th e surface area of the w ood particles and by adding nitrogen, they are creating a perfect milieu, in their feces, for further microbial degradation of wood in the environment. Given that these fi shes inhabit rivers and streams in tropical rainforests, the amount of wood that falls in the waterways is enorm ous. Without the grazing

PAGE 138

138 activities of these fishes on coarse woody debris the waterways of tropical South America might be choked with natural logjams. Furthermore, with the increase in deforestation in many rainforest habitats, more and more wood is findi ng its way into Amazonian waterways. At the same time, local human populations in some of these regions (e.g., northwestern Per) are becoming more reliant on fishes as a source of protein because they have nearly hunted all mammalian and bird taxa to extinction (DPG, pers. obs.). Many of the wood-eating detritivorous fishes have lower fecundity than some of the ot her fishes (DPG, pers. obs.), and are, therefore, susceptible to overfishing. So, coupled with pollution, the increasing removal of fish from Amazonian waterways, including the wood-eatin g detritivorous species, may spell serious trouble in terms of wood degradation and nutrien t cycling in these habitats. The people are reliant upon the rivers for water, food, and for tran sportation. There simply are too few roads in the outlying areas of the Amazonian rainforest on wh ich to travel by car. Thus, if continued, the uncontrolled removal of these fishes from the ri vers may lead to transportation issues, and unforeseen ecosystem-wide consequences.

PAGE 139

139 LIST OF REFERENCES Abril A. and E. Bucher. 2002. Evidence that th e fungus cultu red by leaf-c utting ants does not metabolize cellulose. Ec ology Letters 5: 325-328. Adachi S. 1965. Thin-layer chromatography of car bohydrates in the presence of bisulfite. Journal of Chromatography 17: 295-299. Adkins J.N., S.M. Varnum, K.J. Auberry, K.J. Moore, N.H. Angell, R.D. Smith, L.D. Springer and J.G. Pounds. 2002. Toward a human blood serum proteome: analysis by multidimensional separation coupled with ma ss spectrometry. Molecular and Cellular Proteomics 1: 947-955. Allison S. 2006. Soil minerals and humic acids alte r enzyme stability: implications for ecosystem processes. Biogeochemistry 81: 361-373. Allison S. and J. Jastrow. 2006. Activities of extracellular enzymes in physically isolated fractions of restored grassland soils. Soil Biology and Biochemistry 38: 3245-3256. Ankom Technology. 1998. Method for determini ng neutral detergent fiber (aNDF). Ankom Technical Manual, Fairport, NY. Ankom Technology. 1999. Method for determini ng acid detergent fiber. Ankom Technical Manual, Fairport, NY. Araujo-Lima C., B. Forsberg, R. Victoria and L. Martinelli. 1986. Energy sources for detritivorous fishes in the Amazon. Science 234: 1256-1258. Armbruster J. 1998. Modifications of the dige stive tract for holding air in loricariid and scoloplacid catfishes. Copeia 1998: 663-675. Armbruster J. 2003. The species of the Hypostomus cochliodon group (Siluriformes: Loricariidae) Z ootaxa 249. Magnolia Press, Auckland, New Zealand. Armbruster J. 2004. Phylogeneti c relationships of the suckermouth armoured catfishes (Loricariidae) with emphasi s on the Hypostominae and th e Ancistrinae. Zoological Journal of The Linnean Society 141: 1-80. Barnes C., C. Sweeting, S. Jennings, J. Barry a nd N.V.C. Polunin. 2007. Effect of temperature and ration size on carbon and nitrogen stable isotope trophic fract ionation. Functional Ecology 21: 356-362. Bearhop S., S. Waldron, S. Votier and R. Furness. 2002. Factors that influence assimilation rates and fractionation of nitrogen and carbon stab le isotopes in avian blood and feathers. Physiological and Biochemical Zoology 75: 451-458.

PAGE 140

140 Bergman E. 1990. Energy contributions of volatile fatty acids from the gastrointestinal tract in various species. Physiolo gical Reviews 70: 567-590. Bostrm B., D. Comstedt and A. Ekblad. 2008. Can isotopic fractionation during respiration explain the 13C-enriched sporocarps of ectomycor rhizal and saprotrophic fungi? New Phytologist 177: 1012-1019. Bouchard S. and K. Bjorndal. 2006. Nonadditive interactions between animal and plant diet items in an omnivorous freshwater turtle Trachemys scripta. Comparative Biochemistry and Physiology Part B 144: 77-85. Bowen S.H. 1984. Microorganisms and detritus in the diet of a typical neotropical riverine detritivore, Prochilodus platensis (Pisces: Prochilodontidae). Limnology and Oceanography 29: 1120-1122. Bowen S.H., E.V. Lutz and M.O. Ahlgren. 1995. Dietary protein and energy as determinants of food quality: trophic strategies compared. Ecology 76: 899-907. Bowling D., D. Pataki and J. Randerson. 2008. Carbon isotopes in terrestrial ecosystem pools and CO2 fluxes. New Phytologist 178: 24-40. Breznak J. and A. Brune. 1994. Role of microorganisms in the digestion of lignocellulose by termites. Annual Reviews of Entomology 39: 453-487. Carleton S. and C. Martnez del Rio. 2005. The effect of cold-i nduced increased metabolic rate on the rate of 13C and 15N incorporation in house sparrows ( Passer domesticus ). Oecologia 144: 226-232. Chan A.S., M.H. Horn, K.A. Dickson and A. Gawlicka. 2004. Digestive enzyme activity in carnivores and herbivores: comparisons among four closely related prickleback fishes (Teleostei: Stichaeidae) from a California roc ky intertidal habitat. Journal of Fish Biology 65: 848-858. Choat J.H. and K.D. Clements. 1998. Vertebra te herbivores in ma rine and terrestrial environments: A nutritional ecology perspe ctive. Annual Review of Ecology and Systematics 29: 375-403. Clements K.D. and J.H. Choat. 1995. Fermentation in tropical marine herbivorous fishes. Physiological and Biochemical Zoology 68: 355-378. Clements K.D., V. Gleeson and M. Slaytor. 1994. Short-chain fatty acid metabolism in temperate marine herbivorous fish. Journal of Comparative Physiology B 164: 372-377. Clements K.D. and D. Raubenheimer. 2006. Feeding and nutrition. Pp. 47-82 in D.H. Evans (ed) The physiology of fishes. CRC Press, Boca Raton, FL.

PAGE 141

141 Clements K.D. and D. Rees. 1998. Preservation of i nherrent contractility in isolated gut segments of herbivorous and carnivorous marine fish. Journal of Comparative Physiology B 168: 61-72. Crossman D.J., J.H. Choat and K.D. Clements. 2005. Nutritional ecology of nominally herbivorous fishes on coral reefs. Mari ne Ecology Progress Series 296: 129-142. Dahlqvist A. 1968. Assay of inte stinal disacharidases. Analyt ical Biochemistry 22: 99-107. Das K.M. and S.D. Tripathy. 1991. Studies on the digestive enzy mes of grass carp, Ctenopharyngodon idella (Val.). Aquaculture 92: 21-32. Delariva R. and A. Agostinho. 2001. Relatio nship between morphology and diets of six neotropical loricariids. Jour nal of Fish Biology 58: 832-847. de Melo C.E., F. de Arruda Machado and V. Pint o-Silva. 2004. Feeding habits of fish from a stream in the savanna of central Brazil, Ar aguaia Basin. Neotropi cal Ichthyology 2: 3744. DeNiro M.J. and S. Epstein. 1978. Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta 42: 495-506. DeNiro M.J. and S. Epstein. 1981. Influence of diet on the distribution of nitrogen isotopes in animals. Geochimica et Cosmochimica Acta 45: 341-351. Dodds E.D., M.R. McCoy, A. Geldenhuys, L.D. Rea and J.M. Kennish. 2004. Microscale recovery of total lipids from fish tissue by accelerated solvent extraction. Journal of the American Oil Chemists' Society 81: 835-840. Ebert A. and A. Brune. 1997. Hydrogen concentra tion profiles at the ox ic-anoxic interface: a microsensor study of the hindgut of the wood-feeding lower termite Reticulitermes flavipes (Kollar). Applied and Environm ental Microbiology 63: 4039-4046. Erlanger B.F., N. Kokowsky and W. Cohen. 1961. The preparation and properties of two new chromogenic substrates of trypsin. Archiv es of Biochemistry and Biophysics 95: 271278. Evans-White M., W.K. Dodds, L.J. Gray and K.M. Fritz. 2001. A comparison of the trophic ecology of the crayfishes ( Orconectes nais (Faxon) and Orconectes neglectus (Faxon)) and the central stoneroller minnow ( Campostoma anomalum (Radinesque)): omnivory in a tallgrass prarie stream. Hydrobiologia 462: 131-144. Felicetti L., L. Shipley, G. Witmer and C.T. Ro bbins. 2000. Digestibility, nitrogen excretion, and mean retention time by North American porcupines ( Erethizon dorsatum ) consuming natural forages. Physiological and Biochemical Zoology 73: 772-780.

PAGE 142

142 Fink W.L. and S.V. Fink. 1979. Central Amazonia and its fishes. Comparative Biochemistry and Physiology A 62: 13-29. Focken U. 2004. Feeding fish with diets of different ratios of C3-and C4-plant-derived ingredients: a laboratory analys is with implications for the back-calculation of diet from stable isotope data. Rapid Communica tions in Mass Spectrometry 18: 2087-2092. Fraisse M., N.Y.S. Woo, J. Noaillac-Depeyre and J.C. Murat. 1981. Distribution pattern of digestive enzyme activities in intestine of the catfish ( Ameiurus nebulosus L.) and of the carp ( Cyprinus carpio L.). Comparative Biochemistry and Physiology 70A: 443-446. Frierson E. and J. Foltz. 1992. Comparison and estim ation of absorptive intestinal surface area in two species of cichlid fish. Transactions of the American Fisheries Society 121: 517-523. Fris M.B. and M.H. Horn. 1993. Effects of diets of different pr otein content on food consumption, gut retention, protein conversion, and growth of Cebidichthys violaceus (Girard), an herbivorous fish of temperate zone marine waters. Jour nal of Experimental Marine Biology and Ecology 166: 185-202. Fry B. 2007. Stable Isotope Ec ology. Springer, New York, NY. Fry B. and C. Arnold. 1982. Rapid 13C/12C turnover during growth of brown shrimp (Penaeus aztecus). Oecologia 54: 200-204. Fry B. and E. Sherr. 1984. d13C measurements as indicators of carbon flow in marine and freshwater ecosystems. Contribtions in Marine Science 27: 13-47. Galetto M.J. and D.R. Bellwood. 1994. Digestion of algae by Stegastes nigicans amd Amphiprion akindynos (Pisces: Pomacentridae) with an evaluation of methods used in digestibility studies. Journal of Fish Biology 44: 415-428. Gamboa-Delgado J., J.P. Caavate, R. Zerolo and L. Le Vay. 2008. Natural carbon stable isotope ratios as indicators of the rela tive contribution of live and iner t diets to growth in larval Senegalese sole ( Solea senegalensis). Aquaculture 280: 190-197. Gannes L.Z., D.M. O'Brien and C. Martnez del Rio. 1997. Stable isotopes in animal ecology: assumptions, caveats, and a call for more laboratory experiments. Ecology 78: 12711276. Gaudinski J.B., T.E. Dawson, S. Quideau, E.A. G. Schuur, J.S. Roden, S.E. Trumbore, D.R. Sandquist, S.W. Oh and R.E. Wasylishen. 2005. Comparative analysis of cellulose preparation techniques for use with 13C, 14C, and 18O isotopic measurements. Analytical Chemistry 77: 7212-7224.

PAGE 143

143 Gawlicka A., B. Parent, M.H. Horn, N. Ross, I. Opstad and O.J. Torrissen. 2000. Activity of digestive enzymes in yolk-sac larvae of Atlantic halibut ( Hippoglossus hippoglossus ): indication of readiness for first feeding. Aquacult ure 184: 303-314. Gaye-Siessegger J., U. Focken, H. Abel and K. Becker. 2003. Feeding level and diet quality influence trophic shift of C and N isotopes in Nile tilapia ( Oreochromis niloticus (L.)). Isotopes in Environmental a nd Health Studies 39: 125-134. Gaye-Siessegger J., U. Focken, S. Muetzel, H. Abel and K. Becker. 2004. Feeding level and individual metabolic rate affect d13C and d15N values in carp: implications for food web studies. Oecologia 138: 175-183. German D.P. 2008. Do herbivorous minnows ha ve "plug-flow reactor" guts? Evidence from digestive enzyme activities, gastrointestinal fermentation, and luminal nutrient concentrations. Journal of Comparative Physiology B (Accepted, in revision): German D.P. and M.H. Horn. 2006. Gut lengt h and mass in herbivor ous and carnivorous prickleback fishes (Teleostei: Stichaeidae): ont ogenetic, dietary, and phylogenetic effects. Marine Biology 148: 1123-1134. German D.P., M.H. Horn and A. Gawlicka. 2004. Digestive enzyme activities in herbivorous and carnivorous prickleback fishes (Teleostei : Stichaeidae): ontogenetic, dietary, and phylogenetic effects. Physiological and Biochemical Zoology 77: 789-804. German D.P., B.C. Nagle, J.M. Villeda, A.M. Ruiz, A.W. Thomson, S. Contreras-Balderas and D.H. Evans. 2008. Evolution of herbivory in a carnivorous clade of minnows (Teleostei: Cyprinidae): effects on gut structure and function. Physiological and Biochemical Zoology (Accepted: in revision). Goering H.K. and P. Van Soest. 1970. Forage fi ber analyses (apparatus reagents, procedures and some applications). United States Depart ment of Agriculture, Washington, D.C. Gonzales J.M.J. and P.B. Brown. 2007. Nutrie nt retention capabilities of Nile tilapia ( Oreochromis niloticus ) fed bioregenerative life suppor t system (BLSS) waste residues. Advances in Space Research 40: 1725-1734. Graham J. and T. Baird. 1982. The transition to ai r breathing in fishes. 1. Environmental effects on the facultative air breathing of Ancistrus chagresi and Hypostomus plecostomus (Loricariidae). Journal of E xperimental Biology 96: 53-67. Guelinckx J., J. Maes, P. Van Den Driessche, B. Geysen, F. Dehairs and F. Ollevier. 2007. Changes in d13C and d15N in different tissues of juve nile sand goby Pomatoschistus minutus: a laboratory diet-switch experiment Marine Ecology Progress Series 341: 205215.

PAGE 144

144 Guimares I.G., L.E. Pezzato and M.M. Barros. 2008. Amino acid availa bility and protein digestibility of several protei n sources for Nile Tilapia, Oreochromis niloticus. Aquaculture Nutrition 14: 396-404. Gutowska M., J. Drazen and B. Robison. 2004. Digestive chitinolytic activity in marine fishes of Monterey Bay, California. Comparative Bi ochemistry and Physiology Part A 139: 351358. Harpaz S. and Z. Uni. 1999. Activity of intestin al mucosal brush border membrane enzymes in relation to the feeding habits of three aquacu lture fish species. Comparative Biochemistry and Physiology Part A 124: 155-160. Hendel B. and J. Marxsen. 2000. Extracellular en zyme activity associated with degradation of beech wood in a central European stream. In ternational Review of Hydrobiology 85: 95105. Hesslein R., K. Hallard and P. Ramlal. 1993. Re placement of sulfur, carbon, and nitrogen in tissues of growing broad whitefish (Coregonus nasus) in response to a change in diet traced by d34S, d13C, d15N. Canadian Journa l of Fisheries Aquatic Sciences 50: 20712076. Ho C.T., S.J. Kao, C.F. Dai, H.L. Hsieh, F.K. Shiah and R.Q. Jan. 2007. Dietary separation between two blennies and the Pacific gregor y in northern Taiwan: evidence from stomach content and stable isotope anal yses. Marine Biology 151: 729-736. Hoagland K., S. Roemer and J. Rosowski. 1982. Colonization and community structure of two periphyton assemblages, with emphasis on the diatoms (Bacillariophyceae). American Journal of Botany 69: 188-213. Hobson K.A. and R. Clark. 1992. Assessing avian diets using stable isot opes I: turnover of 13C in tissues. Condor 94: 181-188. Horn M.H. 1989. Biology of Marine Herbivor ous Fishes. Oceanography and Marine Biology Annual Review 27: 167-272. Horn M.H., A. Gawlicka, D.P. German, E.A. Lo gothetis, J.W. Cavanagh and K.S. Boyle. 2006. Structure and function of the stomachless dige stive system in three related species of New World silverside fishes (Atherinopsidae) representing herbivory, omnivory, and carnivory. Marine Biology 149: 1237-1245. Horn M.H. and K.S. Messer. 1992. Fish guts as chemical reactors: a model for the alimentary canals of marine herbivorous fi shes. Marine Biology 113: 527-535. Hurlbert S.H. 1984. Pseudoreplication and the desi gn of ecological field e xperiments. Ecological Monographs 54: 187-211.

PAGE 145

145 Iijima N., S. Tanaka and Y. Ota. 1998. Purification and characterization of bile salt-activated lipase from the hepatopanc reas of red seabream, Pagrus major Fish Physiology and Biochemistry 18: 59-69. Jackson D.C. 2002. Hibernating without oxygen: phys iological adaptations of the painted turtle. Journal of Physiology 543: 731-737. Jardine T., D. MacLatchy, W. Fairchild, R. C unjak and S. Brown. 2004. Rapid carbon turnover during growth of Atlantic Salmon (Salmo salar ) smolts in sea water, and evidence for reduced food consumption by growth -stunts. Hydrobiologia 527: 63-75. Johnson S., S. Jackson, V. Abratt, V. Wolfaar dt, R. Cordero-Otero and S. Nicolson. 2006a. Xylose utilization and short-chain fatty ac id production by selected components of the intestinal microflora of a rodent pollinator ( Aethomys namaquensis ). Journal of Comparative Physiology B 176: 631-641. Johnson S., S. Nicolson and S. Jackson. 2006b. Nect ar xylose metabolism in a rodent pollinator ( Aethomys namaquensis ): defining the role of gastro intestinal microflora using 14Clabeled xylose. Physiological a nd Biochemical Zoology 79: 159-168. Jrgensen H., X.Q. Zhao, P.K. Theil, V.M. Gabert and K.E. Bach Knudsen. 2003. Energy metabolism and protein balance in growing rats fed different levels of dietary fibre and protein. Archives of Animal Nutrition 57: 83-98. Jumars P.A. 2000. Animal guts as ideal chemical reactors: maximi zing absorption rates. American Naturalist 155: 527-543. Jung H.J.G. 1997. Analysis of Forage Fiber and Cell Walls in Ruminant Nutrition. Journal of Nutrition 127 (supplement): 810S-813S. Kappler A. and A. Brune. 2002. Dynamics of redox potential and changes in redox state of iron and humic acids during gut passa ge in soil feeding termites ( Cubitermes spp.). Soil Biology and Biochemistry 34: 221-227. Karasov W.H. and C. Martnez del Rio. 2007. Physiological ecology: how animals process energy, nutrients, and toxins. Princeton Un iversity Press, Princeton, NJ USA. Kelly M.H., W.G. Hagar, T.D. Jardine and R. A. Cunjak. 2006. Nonlethal sampling of sunfish and slimy sculpin for stable isotope analys is: how scale and fin tissue compare with muscle tissue. North American Journal of Fisheries Management 26: 921-925. Klock J.H., A. Wieland, R. Seifert and W. Mi chaelis. 2007. Extracellular polymeric substances (EPS) from cyanobacterial mats: characteris ation and isolation method optimisation. Marine Biology 152: 1077-1085.

PAGE 146

146 Kohzu A., T. Yoshioka, T. Ando, M. Takahashi, K. Koba and E. Wada. 1999. Natural 13C and 15N abundance of field-collected fungi a nd their ecological implications. New Phytologist 144: 323-330. Kramer D.L. and M.J. Bryant. 1995. Intestine leng th in the fishes of a tropical stream: 2. Relationships to diet the long and the s hort of a convoluted issue. Environmental Biology of Fishes 42: 129-141. Krogdahl ., G.I. Hemre and T. Mommsen. 2005. Carbohydrates in fish nut rition: digestion and absorption in postlarval stages. Aquaculture Nutr ition 11: 103-122. Leavitt S.W. and D.R. Danzer. 1993. Method for batch processing small wood samples to holocellulose for stable-carbon isotope an alysis. Analytical Chemistry 65: 87-89. Leenhouwers J.I., D. Adjei-Boateng, J.A.J. Verre th and J.W. Schrama. 2006. Digesta viscosity, nutrient digestibility and organ weights in African catfish ( Clarias gariepinus) fed diets supplemented with different levels of a soluble non-starch polysaccharide. Aquaculture Nutrition 12: 111-116. Lentle R.G. and P.W.M. Janssen. 2008. Physical characteristics of digest a and their influence on flow and mixing in the mammalian intestin e: a review. Journal of Comparative Physiology B 178: 673-690. Lesel R., C. Fromageot and M. Lesel. 1986. Cellulose digestibility in grass carp, Ctenopharyngodon idella and in goldfish, Carassius auratus Aquaculture 54: 11-17. Levey D.J., A.R. Place, P.J. Rey and C. Martnez del Rio. 1999. An experimental test of dietary enzyme modulation in pine warblers Dendroica pinus Physiological and Biochemical Zoology 72: 576-587. Lloret J. and S. Planes. 2003. Condition, feedi ng, and reproductive potential of white seabream Diplodus sargus as indicators of habitat quality and the effect of reserve protection in the northwest Mediterranean. Marine Ec ology Progress Series 248: 197-208. Lo N., H. Watanabe and M. Sugimura. 2003. Eviden ce for the presence of a cellulase gene in the last common ancestor of bilaterian animals. Proceedings of the Royal Society of London B (Suppl.) 270: S69-S72. Logothetis E.A., M.H. Horn and K.A. Di ckson. 2001. Gut morphology and function in Atherinops affinis (Teleostei: Atherinopsidae), a stomachless omnivore feeding on macroalgae. Journal of Fish Biology 59: 1298-1312. MacAvoy S., S. Macko and G. Garman. 2001. Isotopi c turnover in aquatic predators: quantifying the exploitation of migratory prey. Canadian Journal of Fisheries Aquatic Sciences 58: 923-932.

PAGE 147

147 Marsh R.S., C. Moeb, R.B. Lomneth, J.D. Fawcetta and A.R. Place. 2001. Characterization of gastrointestinal chitinase in the lizard Sceloporus undulatus garmani (Reptilia: Phrynosomatidae). Comparative Biochemistry and Physiology Part B 128: 675-682. Martnez del Rio C., K.E. Brugger, J.L. Rios M.E. Vergara and M.C. Witmer. 1995. An experimental and comparative study of dietary modulation of intestinal enzymes in the European starling (Sturnus vulgaris ). Physiological Zoology 68: 490-511. Martnez del Rio C. and B. Wolf. 2005. Mass balance models for animal isotopic ecology: linking diets stoichiometry and physiological processes with br oad scale ecological patterns. Pp. 141-174 in J. Starck and T. Wang (eds) Physiological and ecological adaptations to feeding in vertebra tes. Science Publishers, Berlin. Mayer L., L. Schick, R. Self, P.A. Jumars, R. Fi ndlay, Z. Chen and S. Sampson. 1997. Digestive environments of benthic macroinvertebrate guts: enzymes, surfactants and dissolved organic matter. Journal of Ma rine Research 55: 785-812. McCue M.D. and E. Pollock. 2008. Stable isot opes may provide evidence for starvation in reptiles. Rapid Communications in Mass Spectrometry 22: 2307-2314. McDowell E. and B. Trump. 1976. Histological fixatives for diagnostic light and electron microscopy. Archives of Pathology a nd Laboratory Medici ne 100: 405-414. McEwen S., M. Slaytor and R. O'Brien. 1980. Cell obiase activity in three species of Austrailian termites. Insect Biochemistry 10: 563-567. McIntyre P. and A. Flecker. 2006. Rapid turnove r of tissue nitrogen of primary consumers in tropical freshwaters. Oecologia 148: 12-21. Mill A., J.K. Pinnegar and N.V.C. Polunin. 2007. Explaining isotope trophic-step fractionation: why herbivorous fish are differe nt. Functional Ecology 21: 1137-1145. Mo J., T. Yang, X. Song and J. Chang. 2004. Ce llulase activity in five species of important termites in China. Applied Entomology and Zoology 39: 635-641. Moran K.L. and K. Bjorndal. 2007. Simulated green turtle grazing aV ects nutrient composition of the seagrass Thalassia testud inum. Marine Biology 150: 1083-1092. Moreira L. and E. Filho. 2008. An overview of mannan structure and mannan-degrading enzyme systems. Applied Microbiology and Biotechnology 79: 165-178. Mountfort D., J. Campbell and K.D. Clements. 2002. Hindgut fermentation in three species of marine herbivorous fish. Applied and Environmental Microbiology 68: 1374-1380.

PAGE 148

148 Nakashima A., H. Watanabe, H. Saitoh, G. Tokuda and J.I. Azuma. 2002. Dual cellulosedigesting system of th e wood-feeding termite, Coptotermes formosanus Shiraki. Insect Biochemistry and Molecular Biology 32: 777-784. Nelson J.A. 2002. Metabolism of three species of herbivorous loricariid ca tfishes: influence of size and diet. Journal of Fish Biology 61: 1586-1599. Nelson J.A., D. Wubah, M. Whitmer, E. Johnson and D. Stewart. 1999. Wood-eating catfishes of the genus Panaque : gut microflora and cellulolytic enzyme activities. Journal of Fish Biology 54: 1069-1082. Nelson N. 1944. A photometric adaptation of th e Somogyi method for the determination of glucose. Journal of Biological Chemistry 153: 375-380. Nichols S.J. and D. Garling. 2000. Food-web dynamics and trophic-level interactions in a multispecies community of freshwater unioni ds. Canadian Journal of Zoology 78: 871882. Nico L. 2005. Changes in the fish fauna of th e Kissimmee River basin, penninsular Florida: nonnative additions. Pp. 523-556 in J.N. Rinne R.M. Hughes and B. Calamusso (eds) Historical changes in large river fish asse mblages of the Americas. American Fisheries Society, Bethesda, MD. Nigam J.N. 2001. Ethanol production from wh eat straw hemicellulose hydrolysate by Pichia stipitis Journal of Biotechnology 87: 17-27. Nonogaki H., J.A. Nelson and W.P. Patterson. 2007. Dietary histories of herbivorous loricariid catfishes: evidence from 13C values of otoliths. Experime ntal Biology of Fishes 78: 1321. Painter T.J. 1983. Algal Polysaccharides Pp. 196-285 in G.O. Aspinall (ed) The Polysaccharides, Volume 2. Academic Press Inc., New York, NY. Parker J., J. Montoya and M. Hay. 2008. A specialist detritivore links Spartina alterniflora to salt marsh food webs. Marine Ecology Progress Series 364: 87-95. Parra R. 1978. Comparison of foregut and hindgu t fermentation in herb ivores. Pp. 205-229 in G.G. Montgomery (ed) The ecology of ar boreal folivores. Smithsonian Institution, Washington D.C. Penry D.L. and P.A. Jumars. 1987. Modeling anim al guts as chemical reactors. The American Naturalist 129: 69-96. Peres A., J.L. Zambonino Infante and C. Cahu. 1998. Dietary regulation of activities and mRNA levels of trypsin and amylase in sea bass ( Dicentrarchus labrax ) larvae. Fish Physiology and Biochemistry 19: 145-152.

PAGE 149

149 Perga M.E. and D. Gerdeaux. 2005. 'Are fish wh at they eat' all year round? Oecologia 144: 598606. Petterson R. 1984. The chemical composition of wood. Pp. 58-126 in R. Rowell (ed) The chemistry of solid wood, advan ces in chemistry series 207. American Chemical Society, Washington D.C. Phillips D.L. and P. Eldridge. 2006. Estimating the timing of diet shifts using stable isotopes. Oecologia 147: 195-203. Potts R.C. and P.H. Hewitt. 1973. The distribution of intestinal bacteria a nd cellulase activity in the harvester termite Trinervitermes trinervoides (Nasutitermitinae). Insectes Sociaux 20: 215-220. Pouilly M., F. Lino, J. Bretenoux and C. Rosale s. 2003. Dietary-morphological relationships in a fish assemblage of the Bolivian Amazonian floodplain. Journal of Fish Biology 62: 11371158. Preiser H., J. Schmitz, D. Maestracci and R.K. Crane. 1975. Modification of an assay for trypsin and its application for the es timation of enteropeptidase. Clinica Chimica Acta 59: 169175. Presnell J.K. and M.P. Schreibman. 1997. Humasons Animal Tissue Techniques. The Johns Hopkins University Press, Baltimore, MD. Prins R.A. and D.A. Kreulen. 1991. Comparative as pects of plant cell wall digestion in insects. Animal Feed Science and Technology 32: 101-118. Pryor G.S. and K. Bjorndal. 2005. Symbiotic ferm entation, digesta passage, and gastrointestinal morphology in bullfrog tadpoles ( Rana catesbeiana). Physiological and Biochemical Zoology 78: 201-215. Pryor G.S., D.P. German and K. Bjorndal. 2006. Ga strointestinal Fermentation in Greater Sirens ( Siren lacertina). Journal of Herpetology 40: 112-117. Raubenheimer D. and S. Simpson. 1998. Nutrient transfer functions: th e site of integration between feeding behaviour and nutritional physiology. Chemoecology 8: 61-68. Raubenheimer D., W.L. Zemke-White, R.J. Phillips and K.D. Clements. 2005. Algal macronutrients and food selectivity by the omnivorous marine fish Girella tricuspidata Ecology 86: 2601-2610. Reese S.A., G.R. Ultsch and D.C. Jackson. 2004. Lactate accumulation, glycogen depletion, and shell composition of hatchling turtles during simulated aquati c hibernation. Journal of Experimental Biology 207: 2889-2895.

PAGE 150

150 Reich K., K. Bjorndal and C. Martnez del Rio. 2008. Effects of growth and tissue type on the kinetics of 13C and 15N incorporation in a rapidly gr owing ectotherm. Oecologia 155: 651-663. Robbins C.T., L.A. Felicetti and M. Sponheimer. 2005. The effect of dietary protein quality on nitrogen isotope discrimination in mammals and birds. Oecologia 144: 534-540. Roncari G. and H. Zuber. 1969. Thermophilic aminopeptidases from Bacillus stearothermophilus. I. Isolation, specificity, and general properties of the thermostable aminopeptidase I. Intern ational Journal of Protein Research 1: 45-61. Sakano H., E. Fujiwara, S. Nohara and H. Ue da. 2005. Estimation of nitrogen stable isotope turnover rate of Oncorhynchus nerka. Environmental Biology of Fishes 72: 13-18. Schaefer S. and G.V. Lauder. 1986. Historical transformation of functional design: evolutionary morphology of feeding mechanisms in lorica rioid catfishes. Systematic Zoology 35: 489508. Schaefer S. and D. Stewart. 1993. Systematics of the Panaque dentex species group (Siluriformes: Loricariidae), wood-eating ar mored catfishes from Tropical South America. Ichthyological Exploratio ns of Freshwaters 4: 309-342. Sibly R.M. and P. Calow. 1986. Physiological Ec ology of Animals, an Evolutionary Approach. Blackwell Scientific Publications, Oxford, England. Sinsabaugh R.L., R.K. Antibus, A. E. Linkins, C.A. McClaugherty, L. Rayburn, D. Repert and T. Weiland. 1992. Wood decomposition over a first order watershed: mass loss as a function of lignocellulase activity. Soil Biol ogy and Biochemistry 24: 743-749. Sinsabaugh R.L., S.W. Golladay and A.E. Linki ns. 1991a. Comparison of epilithic and epixylic biofilm development in a boreal ri ver. Freshwater Biology 25: 179-187. Sinsabaugh R.L., D. Repert, T. Weiland, S.W. Golladay and A.E. Linkins. 1991b. Exoenzyme accumulation in epilithic biofilms. Hydrobiologia 222: 29-37. Skea G., D. Mountfort and K.D. Clements. 2005. Gut carbohydrases fr om the New Zealand marine herbivorous fishes Kyphosus sydneyanus (Kyphosidae), Aplodactylus arctidens (Aplodactylidae), and Odax pullus (Labridae). Comparative Biochemistry and Physiology Part B 140: 259-269. Slaytor M. and D.J. Chappell. 1994. Nitrogen meta bolism in termites. Comparative Biochemistry and Physiology Part B 107: 1-10.

PAGE 151

151 Smith P., R. Krohn, G. Hermanson, A. Mallia, F. Gartner, M. Provenzano, E. Fujimoto, N. Goeke, B. Olson and D. Klenk. 1985. Measur ement of protein using bicinchoninic acid. Analytical Biochemistry 150: 76-85. Smith T., D. Wahl and R. Mackie. 1996. Volatile fatty acids and anaerobic fermentation in temperate piscivorous and omnivorous freshwater fish. Journal of Fish Biology 48: 829841. Smoot J.C. and R.H. Findlay. 2000. Digestive en zyme and gut surfcant ac tivity of detrivorous gizzard shad ( Dorosoma cepedianum ). Canadian Journal of Fisheries and Aquatic Sciences 57: 1113-1119. Somogyi M. 1952. Notes on Sugar Determination. Journal of Biological Chemistry 195: 19-23. Stevens C.E. and I.D. Hume. 1995. Comparative phys iology of the vertebrate digestive system. Press Syndicate of the University of Cambridge, Melbourne, Australia. Stevens C.E. and I.D. Hume. 1998. Contributions of Microbes in Vertebra te Gastrointestinal Tract to Production and Conser vation of Nutrients. Physiological Reviews 78: 393-427. Swift M.J., O.W. Heal and J.M. Anderson. 1979. Decomposition in terrestrial ecosystems. University of California Press, Berkeley, CA, USA. Tank J.L., J.R. Webster, E.F. Benfield and R.L. Sinsabaugh. 1998. Effect of leaf litter exclusion on microbial enzyme activity a ssociated with wood biofilms in streams. Journal of the North American Bethological Society 17: 95-103. Tayasu I., T. Abe, P. Eggleton and D.E. Bignell. 1997. Nitrogen and carbon isotope ratios in termites: an indicator of trophic habit along the gradient from wood-feeding to soilfeeding. Ecological Entomology 22: 343-351. Tieszen L., T. Boutton, K. Tesdahl and N. Slade. 1983. Fractionation and turnover of stable carbon isotopes in animal tissu es: implications for delta 13C analysis of diet. Oecologia 57: 32-37. Trueman C., R. McGill and P. Guyard. 2005. The effect of growth rate on tissue-diet isotopic spacing in rapidly growing animals. An experimental study with Atlantic salmon ( Salmo salar). Rapid Communications in Mass Spectrometry 19: 3239-3247. Trust T., L. Bull, B. Currie and J. Buckley. 1979. Obligate anaerobi c bacteria in the gastrointestinal microflo ra of the grass carp ( Ctenopharyngodon idella ), Goldfish ( Carassius auratus ), and rainbow trout (Salmo gairdneri ). Journal of the Fisheries Research Board of Canada 36: 1174-1179.

PAGE 152

152 Turner M.H. and B. Hulme. 1970. The plasma proteins: an introducti on. Pitman Medical & Scientific Publishing Company, London, U.K. Valaskova V. and P. Baldrian. 2006. Degradation of cellulose and hemicelluloses by the brown rot fungus Piptoporus betulinus production of extr acellular enzymes and characterization of the major cellu lases. Microbiology-SGM 152: 3613-3622. van Dam A., M. Beveridge, M. Azim and M. Verdegem. 2002. The potential of fish production based on periphyton. Reviews in Fish Biology and Fisheries 12: 1-31. Van Dyke J.M. and D.L. Sutton. 1977. Digestion of duckweed ( Lemna spp.) by the grass carp ( Ctenopharyngodon idella ). Journal of Fish Biology 11: 273-278. Van Soest P. 1994. Nutrional Ecology of the Rumi nant. Cornell University Press, Ithica, NY. Vispo C. and I.D. Hume. 1995. The digestive trac t and digestive function in the North American porcupine and beaver. Canadian Journal of Zoology 73: 967-974. Watanabe H. and G. Tokuda. 2001. Review: Anim al cellulases. Cellular and Molecular Life Sciences 58: 1167-1178. West G.B., J.H. Brown and B.J. Enquist. 2001. A general model for ontogenetic growth. Nature 413: 628-631. Wills A.A., A.R.I. Kidd, A. Lepilina and K.D. Poss. 2008. Fgfs control homeostatic regeneration in adult zebrafish fins. Development 135: 3063-3070. Wilson D.B. and D.C. Irwin. 1999. Genetics and Properties of Cellulases. Pp. 1-21 in G.T. Tsao and T. Schepes (eds) Recent Progress in Bioc onversion of Lignocellulosics. Advances in Biochemical Engineering/Biotechnol ogy. Springer Verlag, New York. Wotton R.S. 2004. The ubiquity and many roles of exopolymers (EPS) in aquatic systems. Scientia Marina 68 (suppl. 1): 13-21. Xie X.L., J. Du, Q.S. Huang, Y. Shi and Q.X. Chen. 2007. Inhibitory kinetics of bromacetic acid on -N-acetyl-D-glucosaminidase from prawn (P enaeus vannamei). International Journal of Biological Macromolecules 41: 308-313. Xu P.E. and D.L. Distel. 2004. Purification and characterization of an endo-1, 4--D glucanase from the cellulolytic system of the wood-boring marine mollusk Lyrodus pedicellatus (Bivalvia: Teredinidae). Marine Biology 144: 947-953.

PAGE 153

153 Zihler F. 1982. Gross morphology and configuration of digestive tr acts of Cichlidae (Teleostei: Perciformes): phylogenetic and functional significance. Netherlands Journal of Zoology 32: 544-571. Zimmer M. and A. Brune. 2005. Physiological prope rties of the gut lumen of terrestrial isopods (Isopoda: Oniscidea): adaptive to digesting lignocellulose ? Journal of Comparative Physiology B 175: 275-283.

PAGE 154

BIOGRAPHICAL SKETCH Donovan Parks German was born in Orange, CA, in 1975. He received a Bachelor of Arts degree in M arine Sc ience from the University of San Diego where he was a student athlete playing football and lacrosse. During his undergraduate career, Donovan participated in numerous research projects, incl uding one in Bermuda, and another that resulted in a publication (Sturz, A., D.P. German and D. Putnam (1998) Salton Sea Geothermal Area Mud Pots. In Geology and Geothermal Resources of the Imperial and Mexicali Valleys L. Lindsey and W. Hample (eds). San Diego Asso ciation of Geologists publicati on 98-1, Pg. 109-128). He then attended California State University Fullerton, wh ere he received a Master of Science degree. For his thesis research he studied the digestiv e physiology of herbivorou s and carnivorous fishes under the stewardship of Dr. Michael H. Horn. While at CSU Fullerton he was heavily involved in community outreach through the world-peace Buddhist organization, SGI-USA, for which he was awarded the Presidents Associates Outstandi ng Graduate Student Award. He also received the Best Thesis award in the Depa rtment of Biological Science at CSU Fullerton. He then began his doctoral training at the Univer sity of Florida under the tutela ge of Dr. David H. Evans, to whom he is gratefully endebted. While at UF, Donovan was part of the NSF SPICE program, which aimed at increasing scien tific literacy and gene ral interest in science in under-resourced middle schools in east Gainesville. He also taught Functional Vertebrate Anatomy Lab, and the graduate course Integrative Prin ciples of Zoology at the Universi ty. In January 2009, he began a postdoctoral research position in the Department of Ecology and Evolutionary Biology at University of California Irvine under Dr. Steven D. Allison.