This item is only available as the following downloads:
1 MORPHOLOGY, ECOPHYSIOLOGY, AND IMPACTS OF NONINDIGENOUS POMACEA IN FLORIDA By JENNIFER L. BERNATIS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FO R THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014
2 2014 Jennifer L. Bernatis
3 T his is for all of the grad students who th ink the process will never end And to Tabitha and Snuffy, you are t ruly missed but will never be forgotten.
4 ACKNOWLEDGMENTS I would like to thank my committee Dr. Steve Johnson (Chair), Dr. Mark Brenner (Co chair) Dr. Ken Lang e land, Dr. Jim Williams, and Dr. Tom Frazer for their support and patience through this pr ocess. I am also very th ankful to Dr. Chad Cross, Environmental and Occupational Health, University of Nevada Las Vegas, and Dr. Erin Leone, Florida Fish and Wildlife Conservation Commission, for invaluable statistical help and support. In addition, I woul d like to thank Mr. Gary Warren, Florida Fish and Wildlife Conservation Commission, and Dr. Iain McGaw, Memorial University, for their continued support and encouragement throughout this project. I would also like to thank Mr. Bill Coleman, Florida Fish an d Wildlife Conservation Commission, for providing the resou rces for much of this project. And finally, I would like to thank my parents, Robert and Terry Bernatis, for not minding the fact they have a professional student for a daughter
5 TABLE OF CONTENT S page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 Background ................................ ................................ ................................ ............. 12 Morphological Characteristics ................................ ................................ ................. 13 Biology ................................ ................................ ................................ .................... 15 Habitat and Environmental Tolerances ................................ ............................ 15 Feeding ................................ ................................ ................................ ............ 16 Rep roduction ................................ ................................ ................................ .... 17 Age and Growth ................................ ................................ ............................... 19 Pathways of Introduction ................................ ................................ ......................... 20 Effects of Int roduction ................................ ................................ ............................. 21 Summary and Research Goals ................................ ................................ ............... 24 2 MORPHOLOGY ................................ ................................ ................................ ...... 32 Background ................................ ................................ ................................ ............. 32 Methods ................................ ................................ ................................ .................. 35 Collection and Holding Procedures ................................ ................................ .. 35 Measuring Procedures ................................ ................................ ..................... 36 Statistical Analysis ................................ ................................ ............................ 37 Results ................................ ................................ ................................ .................... 38 Nonparametric MANOVA ................................ ................................ ................. 38 Preliminary Canonical Discriminant (CDF) and Discriminant Function Analyses ................................ ................................ ................................ ........ 40 Forward Stepwise Analysis ................................ ................................ .............. 43 Follow up Canonical Discriminant and Discriminant Function Analyses .......... 44 Pre and Post Cluster Frequencies ................................ ................................ .... 46 Discuss ion ................................ ................................ ................................ .............. 47 3 PHYSIOLOGICAL TOLERANCES ................................ ................................ ......... 67 Background ................................ ................................ ................................ ............. 67 Methods ................................ ................................ ................................ .................. 71 Collection and Holding ................................ ................................ ..................... 71
6 General Procedures ................................ ................................ ......................... 72 Starvation ................................ ................................ ................................ ......... 73 Salinity ................................ ................................ ................................ .............. 73 pH ................................ ................................ ................................ ..................... 74 Desiccation ................................ ................................ ................................ ....... 74 S tatistical Analyses ................................ ................................ .......................... 76 Results ................................ ................................ ................................ .................... 76 Starvation ................................ ................................ ................................ ......... 76 Salinity ................................ ................................ ................................ .............. 77 pH ................................ ................................ ................................ ..................... 78 Desiccation ................................ ................................ ................................ ....... 79 Discussion ................................ ................................ ................................ .............. 81 4 FEEDING RATES AND PREFERENCES ................................ ............................... 93 Background ................................ ................................ ................................ ............. 93 Methods ................................ ................................ ................................ .................. 96 Gen eral Procedures ................................ ................................ ......................... 96 Feeding Rates ................................ ................................ ................................ .. 97 Feeding Preferences ................................ ................................ ........................ 98 Single age gro up ................................ ................................ ........................ 98 Mixed age group ................................ ................................ ........................ 99 Statistical Analysis ................................ ................................ ............................ 99 Results ................................ ................................ ................................ .................. 100 Feeding Rates ................................ ................................ ................................ 100 Feeding rates of adult P. canaliculata ................................ ...................... 100 Feeding rates of adult P maculata ................................ .......................... 101 Feeding rates of juvenile P. maculata ................................ ...................... 102 Comparison of feeding rates of adult snails ................................ ............. 103 Comparisons of feeding rates of P. maculata age groups ....................... 103 Feeding Preferences ................................ ................................ ...................... 104 Six week feeding pr eference study ................................ .......................... 104 Two week feeding preference trials ................................ ......................... 106 Discussion ................................ ................................ ................................ ............ 107 5 EFFECTIVENESS OF MANUAL REMOVAL FOR CONTROL OF NONINDIGENOUS APPLE SNAILS ................................ ................................ ..... 121 Background ................................ ................................ ................................ ........... 121 Methods ................................ ................................ ................................ ................ 125 Removal Protocols ................................ ................................ ......................... 125 Dive Survey ................................ ................................ ................................ .... 127 Results ................................ ................................ ................................ .................. 127 Primary Location ................................ ................................ ............................ 127 Reference Ponds ................................ ................................ ............................ 129 Discussion ................................ ................................ ................................ ............ 129
7 6 CONC LUSIONS ................................ ................................ ................................ ... 137 APPENDIX A P VALUE COMPARISONS OF RAW CONSUMPTION FEEDING RATES .......... 145 B P VALUE COMPARISONS OF STANDARDIZED CONSUMPTION FEEDI NG RATES ................................ ................................ ................................ .................. 147 LIST OF REFERENCES ................................ ................................ ............................. 149 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 163
8 LIST OF TABLES Table page 2 1 General egg characteristics of Pomacea spp. in North America ......................... 53 2 2 Results of the 45 ratio measurements ................................ ................................ 55 2 3 Results of the 45 ratio measurements for juveniles.. ................................ .......... 58 2 4 Results of the initial 45 ratio model and the reduced follow up model for the canonical discrimi nant function analysis ................................ ............................. 61 3 1 Comparison of survival curves ................................ ................................ ............ 88 4 1 Plant species used in the three feeding trials ................................ ................... 114 4 2 Consumption ( X C) and body weight adjusted (standardized) X consumption ( X SC), in grams, for the 72 hour feeding trials. ................................ ............... 115 4 3 Results of the 6 week feeding preference for adults and juveniles.. ................. 116 4 4 P value comparisons of foraging rates in the 6 week preference trial. ............. 117 4 5 Results of the presence/absence of plant species for the 6 week feeding preference trial. ................................ ................................ ................................ 118 A 1 P value comparisons of RCR feeding rates. ................................ ..................... 145 B 1 P value comparisons of standardized (g consumed/ gram of snail weight) feeding rates. ................................ ................................ ................................ .... 147
9 LIST OF FIGURES Figure page 1 1 States wit h Pomacea in North America. ................................ ............................. 27 1 2 Nonindigenous Pomacea populations throughout Florida as of 2013. ................ 28 1 3 Egg masses of Pomacea spp. in Florida. ................................ ........................... 29 1 4 Holding tank observation ................................ ................................ .................... 30 1 5 Measurement of added shell ................................ ................................ .............. 30 1 6 Pomacea maculata egg masses in various stages of development and hatching. ................................ ................................ ................................ ............. 31 1 7 Two day old hatchling apple snails, on the left P. paludosa and on the right P. canaliculata ................................ ................................ ................................ ... 31 2 1 Shell measurements for adult and juvenile snails ................................ ............... 54 2 2 Results of the canonical discriminant function analysis ................................ ...... 62 2 3 Results of the combined adult size class ................................ ............................ 63 2 4 Results of the juvenile, small adult, and large adult canonical ............................ 64 2 5 Results of the within species canonical discriminant function analyses ............. 65 2 6 Results of the prelimi nary cluster analysis ................................ .......................... 66 3 1 Salinity survival curves ................................ ................................ ....................... 90 3 2 pH survival curves ................................ ................................ .............................. 91 3 3 Wet/dry survival curves ................................ ................................ ...................... 92 4 1 Representative tanks at the end of the six week trial ................................ ........ 119 4 2 Total consumption during the 2 week trial ................................ ........................ 120 5 1 l l lustration of primary study site in Duval County, FL. ................................ ...... 134 5 2 The number of snails collected in each year ................................ ................... 135 5 3 The number of egg masses collected on each collection ................................ 136
10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MORPHOLOGY, ECOPHYSIOLOGY, AND IMPACTS OF NONINDIGENOUS POMACEA IN FLORIDA By Jennifer L. Bernatis May 2014 Chair: Steve Johnson Major: Interdisciplinary Ecology Nonindigenous a pple snails (Ampullariidae : Pomacea ) are freshwate r gastropods n ative to South and Central America and have been introduced in North Am erica and Asia. Two species, Pomacea canaliculata and Pomacea maculata are known to damage wetland crops and because of their indiscriminate and voracious appetites have been implicated in decimation of large areas of aquatic vegetation. Florida ha s nonindigenous Pomacea populations in at least 29 watersheds in 38 counties Although there are many questions to be examined regarding P. canaliculata and P. maculata three g eneral questions are addressed in this study : 1) what are the physiological tolerances affect ing habitat use of nonindigenous Pomacea in Florida, 2) is there a means to distinguish between t he species using morphology and 3) what are the potential impacts of nonindigenous Pomacea on aquatic flora? Results of this research provide a framework for understanding the differences between species as well as potential impacts. Differences in shell morphology were observed in the large adult snails (> 50.00 mm len gth) Most notably, P. maculata is generally wider than it is long ( X ratio < 0.97 ) and the opposite is true for P. canaliculata ( X ratio > 1.05) In smaller adults ( 30.00 49.99 mm ), this guideline should be adjusted to
11 1.02. Physiolo gical testing showed that overall snails can tolerate salinit y levels up to 8 ppt and pH 5.5 9.5 for at least 28 days, and desiccation for up to one year (adults) and six months (juveniles). The potential for damage to rooted aquatic macrophytes is great er in young plants; snails appear to prefer new stem and shoot growth of emergent vegetation. Plants with herbivore defense mechanisms may be able to withstand the foraging pressure of the snails. In some systems control and even eradication may be possib le as was confirmed in a manual removal project A substantial reduction in snail numbers (> 90%) was noted by the end of the first year and control was effectively achieved after three years. The biology of these nonindigenous Pomacea makes them a thr e a t to some aquatic eco systems, but the ir long term impacts o n large natural systems remain unclear.
12 CHAPTER 1 INTRODUCTION Background Apple snails (Ampullariidae) are freshwater gastropods and are classified into nine extant genera w ith an estimated 150 species (Hayes et al 2009). Nonindigenous Pomacea are native to South and Central America and have been introd uced in North America and Asia One species, P omacea paludosa (Say 1829), is native to North America. In the last decade, sn ails in the genus Pomacea have received much attention because of the ir ability to damage crops through foraging. The genus is divided into groups and two of these groups, Pomacea canaliculata (Lamarck 1819) and Pomacea bridgesii (Reeve 1856), account for most of the diversity of introduced apple snails. In introduced locations, the most abundant P. canaliculata group members are P. canaliculata and Pomacea maculata ( Perry 18 10 ; previously described as Pomacea insularum Hayes et al. 2009 20 12 ). Pomacea canaliculata is listed as Lowe et al. 2000 ). Commonly introduced species from the P. bridgesii group include Pomacea sp. ( previously described as Pomacea haustrum Reeve 1856), Pomacea diffusa (Blume 1957), and Pomacea scalaris In the U.S., nonindigenous apple snails are found in at least 1 0 states (Figure 1 1). Florida has more apple snai l species than any other state. These include the native, Pomacea paludosa (Say 1829 ; Florida app le snail ) and at least five nonindigenous species: P. canaliculata (Channeled apple snail ) P. maculata (Island apple snail) Pomacea sp. P. diffusa (Spike top apple snail) and Marisa cornuarietis (Linnaeus 1758 ; Giant ramshorn snail ). As of 201 3 Flori da ha d nonindigenous Pomacea populations in at least 29 watersheds in 38 of 67 counties (Figure 1 2).
13 Introductions of these species have resulted in negative agricultural impact and have led to a small body of research on reproductive behavior, feedi ng preferences, and control methods (Estebenet 1995, Albrecht et al 1999, Yusa and Wada 1999, Estoy et al 2002a, 2002b, Carlsson et al 2004, Carlsson and Lacoursiere 2005 Yusa et al. 2006). Results from some of the research, however, are inconsistent w ith respect to the same response variable and this prompted a genetic analysis of the populations to determine if all of the snails belonged to the same species. The genetic analysis demonstrated that many of the papers reporting to have P. canaliculata sn ails actually had P. maculata snails or both species (Dr. Timothy Collins, Florida International University, personal communication 2006 ). This raised two important questions: 1) which species occur in Florida and 2) what are their potential impacts on Flo aquatic ecosystems. Morphological Characteristics Originally d escribed in the early 1800s, analysis of shell morphology and limited study of soft tissues has led to members of the P. canaliculata complex being added, rem oved, or considered synonymou s Morphological characteristics are easily influenced by food quality and quantity, habitat characteristics (e.g. water flow) and snail density (Alderson 1925 ; Cazzaniga 1990 ; Mart in and Estebenet 2002 ). S hells within the species complex vary in size an d often in color and pattern. Selective breeding for the aquarium trade (e.g. color traits) has added to the variation within certain species, further confounding morphological identification. R ecent studies on internal anatomy found variation in male rep roductive structures between species ( Hayes et al. 2012 ).
14 Generally, the shells are spherical and have five to six whorls that are separated by a deep indentat ion, hence the common apple snails for one species. The characteristics of the channel can be altered by habitat and snail density. T rophic changes (e.g. reduced food availability from grazing pressure) allow P. canaliculata males to reach sexual maturity earlier and this results in a less channeled and more spike top a ppearance, similar to P. diffusa ( Dr. Pablo Martin, Universidad Nacional del Sur, personal communication 2008). P omacea maculata is the largest of the known species in Florida, with shell sizes up to 120 mm; P. canaliculata may reach 80 mm wh ereas the rem aining species in Florida range in size up to 80 mm but are generally smaller than 60 mm The shell aperture is l arge and may be oval to round. Snails in naturally occurring (wild) populations range in shell color from yellow to brownish black, and may co ntain stripes or dimpling. Because these snails have been cultivated for the aquarium trade, selective breeding has led to distinct color variations and patterns; snails may have bright shell colors (e.g., y ellow/orange, green, or copper) and the tissues o f the foot may have pigment variations ranging from yellow to orange t o gray (personal observation). These extremes in shell patterns may result from sperm competition. Accord ing to Yusa (2004) females are able to accept sperm from multiple males and there is evidence to suggest that sperm with certain color and pattern genes can out compete other sperm. Multiple e nvironmental factors that affect growth patterns, such as population density, age of the snail, and food quality, often make field identification s difficult (Estebenet and Martin 2002 2003; Martin and Estebenet 2002 ) Therefore, alternate mea ns of identification are sought such as characteristics of egg clutches. The clutch
15 dimens ions (i.e., total length) number and size of eggs, and color of th e eggs are possible spe cies identifiers (Figure 1 3). Egg clutches of P. maculata tend to have smaller diameter eggs that are more densely packed; however, the overall length of the clutch is generally larger than that of other species and may have up t o 1 200 eggs (ASE 2006). Clutches of P. canaliculata have larger eggs and fewer total eggs (up to 600) but are the same color as P. maculata (ASE 2006). Variation of clutch characteristics within a species caused by food quality, age of the snail and habitat conditions may result in overlapping characteristics, thereby making identification through egg masses difficult. For many Pomacea species, including all of those in Florida, another difficulty in using eggs for species identification is that the eggs tu rn white as they near hatching. Although field identification can be difficult genetic analysis can distinguish among s pecies (Rawlings et al. 2007). Rawlings et al. (2007) confirmed the presence of one population of P. canaliculata in Florida, whereas o ther tested populations assumed to have been P. canaliculata were in fact, P. maculata ( Dr. T. Collins personal communication 2006) Although molecular methods provide a more reliable means of identification they are not applicable for immediate field i dentification, which is required for survey assessments. Additionally, if more than one species is present, and is unidentified or underrepresented, it will probably not be sampled for genetic analysis ( Dr. T. Collins personal communication 2006 ). Therefo re, a means of definitive morphological identification is necessary to distinguish among species. Biology Habitat and Environmental Tolerances Pomacea canaliculata complex snails have not been extensively studied in terms of preferred habitat and environm ental tolerances. The majority of research from their
16 native range has been on shell morphology and variation within P. canaliculata populations (Estebenet and Martin 2002 2003; Martin et al. 2001; Estebenet et al. 2006 ; Martin and Estebenet 2002). At the out set of my study, there was one paper focused on distribution and environmental limits; the snails have a wide range of tolerance to many environmental variables ( Martin et al. 2001 ), but no definitive physiological or ecolo gical limitations are defined Only l imited research has been conducted on temperature tolerance (Alb recht et al. 1999) and other references to environmental factors that influence life histor y were from anecdotal observations. Research on other members of the co mplex (e.g. P. macul ata ) was even less frequent in the literature at the o ut set of this project. Feeding Channeled and Island apple snails are opportunistic feeders that consume a variety of aquatic vegetation, terrestrial fruits and vegetables, algae, remains of other dec aying organisms and conspecific eggs ( personal observation ). The snails maintain an air bubble inside the shell (i.e lung) for buoyancy regulation. Observations in the laboratory and in the field suggest the snails are able to use this mechanism to reach vegetation that would otherwise be inaccessible. Observations on feeding behavior hav e revealed the use of the foot to manipulate foo d into the radula (Figure 1 4). P. canaliculata during which the foot acts a s a funnel to gather food from the surfac e water film with pedal cilia. This action has been observed with smaller snails of other Pomacea species ( personal observation ; Saveanu and Martin 2013) The primary reason P. canaliculata is deemed a pest spe cies in some coun tries is not because of damage to natural ecosystems but because of the damage to
17 aquaculture crops (i.e., rice and taro) (Esteb e net 1995; Yusa et al. 1 999, 2006; Carlsson et al. 2004, 2005; Yu 2005 ; Nakamura 2006 ). I ndiscriminate rapid c onsumption of large amounts of aquatic vegetation however, could alter aquatic ecosystems (Carlsson et al. 2004). Although there have been no reports in Florida of P. canaliculata cause of ecosystem alteration and it has been implicated in the decline of the native apple snail, P. paludosa and the federally endangered Florida Snail Kite ( Rostrhamus sociabilis plumbeus) ( Cattau et al. 2010 ) Research support ed the hypothesis that yo ung birds have difficulty handling larger snails (Cattau et al. 2010) ; unfortunately, this became an accepted n otion for the birds in general. foraging behavior, however, suggests older birds are capable of handling snail s > 60 mm TL (Bernatis 201 3 ). Reproduction Apple snails are dio eci ous and both sexes of P. canaliculata and P. maculata as well as P. paludosa reach sexual ma turity at approximately 25 mm length (about 2 3 months of age) (Lach et al. 2000; Estoy et al. 2002a, 2002b). Females only need to mate once to lay e ggs throughout their lifespan. Albrecht et al. (1999) found that temperature is a limiting factor for Channeled app le snail reproductive efforts. As temperature decreases, the number of clutches deposi ted decreases. In Florida, Channeled and Island apple snails reproduce throughout the year, although effort slows during the winter months (personal observation) Estoy et al. (2002a, 2002b) suggested that food is a limiting factor for reproductive efforts I have observed this i n the laboratory as well, with increases in the number of egg clutches deposited after periods
18 of starvation (> 6 days) followed by f eeding During the starvation period egg deposition ceases, but resumes after the snails begin to feed again. During the summer months when reproductive efforts are highest, females may la y a clutch of eggs once a week. Hatching time varies with environmental conditions (Albrecht et al. 1999; Estoy et al. 20 02a, 2002b). In Florida, I have observed P. canaliculata to have the shortest developmental duration; as few as 10 days in outdoor mesocosms and as short as seven days in controlled laboratory cond itions. Hatch time is fastest in the summer months and is gene rally about 11 14 days with h atching su ccess near 100% Likewise, Island apple snails hatch in outdoor mesoco s ms in 10 days and in labor atory conditions in nine days, but is generally 12 1 6 days Hatch time and number of emerging offspring may be affected by environmental conditions such as in undation and temperature fluctuations (Albrecht et a l. 1999; Estoy et al. 2002a, 2002b). H atched clutches I have examined in the wild however, in dicate a hatch rate of nearly 100% This is based on observations of un hatched eggs in r elation to the enti re clutch. T he ability of a single female to produce several thousand offspring in one season, along with limited numbers of predators in the environments where snails have been introduced (Yusa et al. 2006), are two reasons that nonindigenous snail popula tions are expanding rapidly At the out set of my research, it was unknown if introduced apple snails could hybridize with each other or with the native P. paludosa In a laboratory setting, I observed copulation between field collected P. canaliculata and P. paludosa and P. canaliculata and P. maculata but in each instance the resulting egg clutch did not hatch Previous copulation could have led to egg clutches being laid, but inadequate
19 fertilization may have cause d offspring to be unviable In South A merica, P. canaliculata has been observed mating with P. maculata and M. cornuarietis the latter of which will mate with almost any snail species ( Dr. P. Martin personal communication 2008 ), but again no known hybridization occurred. Furthermore, trials designed to encourage hybridization res ulted in zero egg clutches laid, although copulation was observed (Bernatis 201 3 ). Age and Growth Currently, there is no way to accurately determine lifespan of the Channeled and Isla nd apple snails generally ranges from 2 4 years (Estebenet and Cazzinaga 1992) Arnold (2011) used stable oxygen is otopes in shell carbonate to try and determine the age of Pomacea maculata, but had mixed results. Several studies have demonstrated that tem perature and food resources limit the growth of apple snails (Albrecht et al. 1999; Estoy et al. 2002a, 2002b). Furthermo re, Tanaka et al. (1999) reported that as P. canaliculata density inc reases, growth rate decreases. I have also observed this in holdin g tanks in where tanks with snail > 500 g/64 liters When divided in half ( i.e., 250 g/64 liters) and each new tank population displayed a net gain of 60 70 g within one week. However, when the populations started at 250 g/64 liters and wer e combined into a single larger population ( i.e., 500 g/ 64 liters) mass gain was only 12 1 4 g for the entire population. Growth is rapid and snails can reach 30 mm within 2 3 months (personal observation; Lach et al 2000; Estoy et al 2002a 2002b). In controlled envir onment s with low density (< 3 snails/ 4 liters) and abundant food, growth rates can be up to 1 mm / 24 hours in snails < 40 mm in TL (personal observation; Figure 1 5).
20 Pathways of Introduction Members of the P. canaliculata complex are native throughout the Amazon Basin and the La Plata Basin in southeast Brazil, Argentina, Bolivia, Paraguay, Peru, Uruguay, and other locations throughout South America. Introductions resulting in established populations have occurred in Taiwan, Indonesia, Thailand, Cambodi a, Hong Kong, southern China, The Philippines, Japan, Australia, Hawaii, California Arizona, Florida, and Texas. At the beginning of this study, isolated reports in the U.S had come from Georgia, North and South Carolina, Ohio and Indiana ( Howel ls 2001; G FMFS 2003; IDNR 2005; ASE 2006; CDA 2006) E stablished populations now occur through 10 southern states (Figure 1 1) Channeled apple snails were introduced in Taiwan in the 1980s, to develop an escargot industry. T he market however, was not as successful as anticipated, and through intentional or accidental means, snails escaped into the rice fields and bec a me serious crop pest s (ASE 2006). Distribution has also been linked to hobbyist release from the aquarium industry, as well as deliberate introduction s of snails for nuisance plant control (personal observation) Channeled apple snail s were probably brought to Florida for human consumption and the aquarium trade but escaped or were released, and became established by 1978 (Thompson 1997). I ntroducti on dates, locations and recent taxonomic revisions of the species in Florida (Hayes et al. 2009, 2012) makes the exact place and time of introd u c tion difficult to determine. Originally, the Florida Fish and Wildlife Conservation Commission (FWC) and the F lorida Department of Environmental Protection (FDEP) stated P. canaliculata was introduced sometime in the mid dle or late 1970s in South Florida ( Mr. Gary Warren, FWC, personal communicati on 2006 ; Denson and Eby 2004). Later introductions occurred in the T ampa Bay area in th e 1980s. For the next two
21 decades there was little information a bout the habitat of the snail. When I began my research, there was a substantial increase in reports and confirmations of snail populations (at least 23 counties), ranging f rom Miami to Tallahassee (Mr. Scott Hardin, FWC, personal communication 2006). As of 201 3 Florida had nonindigenous Pomacea populations in more than half its counties (38 of 67) and 29 watersheds. P omacea canaliculata can move many meters a day in the water or on land, and may float downstream thereby facilitating dispersal Additionally, intentional unsanctioned introductions for aquatic plant control in the 1980s ( Mr. G. Warren personal communication 2006 ) increased the ra te of spread in South Flor ida. Pomacea canaliculata sp. was discovered in the St. Johns River watershed in 2004 and subsequently observed in Duval C ounty northernmost Florida, in 2005, (Denson and Eby 2004; Frank 2005). P. maculata has been unintentionally introduced into water bo dies through restoration activities E gg masses were present on aquatic macrophytes planted as part of the restoration process ( Mr. Beachum Furse, FWC personal communication 2008). Egg masses h ave also been observed on boats. If these eggs are not removed prior to hatching spread is possible if the boat is launched into a different water body (Figure 1 6). Effects of Introduction Effects of introduction range from beneficial to innocuous to potentially fatal for humans if snails are consumed raw and har bor specific parasites. The negative impacts, discussed in de tail below, are not only system dependent (e.g. natural vs. urban), but species d ependent as well. Although, there is no scientific documentation in Florida, numerous anecdotal accounts from res idents and state agencies implicate the snail in chang ing lake conditions ( Mr. Dana Denson, Reedy Creek Improvement District,
22 personal communication 2006 ). The large size of P. canaliculata complex snails and their preference for consuming aquatic vegetati on makes them potential threat s to aquatic ecosystems. These systems are home to many endemic invertebrate species that are dependent on stable habitats (e.g., Floridobia mica -Ichetucknee Silt Snail and Procambarus econfinae -Panama City Crayfish ) Any or ganism that depletes the abundance of aquatic plants has the potential to alter ecosystems and threaten other species (Carlsson et al. 2004). Like other gastropods, P. canaliculata complex snails are hosts to numerous parasites that are transfer able to humans through consumption of raw or undercooked snail meat One such parasite, Angiostrongylus cantonensis a nematode known as rat lung worm is capable of infecting humans and c ausing eosinophilic meningitis, a potentially fatal condition. Furthermore, the parasite has been found to reach high concentrations in the edible part of the snail, the head foot portion (Chao et al. 1987). Regardless of the health risks associated with consuming the snails, they are now after failed initial attempts to develop an escargot industry, considered a delicacy in East and Pacific Asian markets (ASE 2006). Research on both P. canaliculata and P. maculata in their native habitat is limited at best, and few predators have been identified in the laboratory (Hendarsih et al. 1994; Yusa, et al. 2006) Predators that consume eggs and snails tend to be unwanted organisms in urban settings e.g. fire ants, rats, and alligators. Predation on P. canaliculata sp. by Florida predators (native or introduced) has been minimally inve stigated and most information is anecdotal; some studies suggest that introduced tilapia, Oreochromis spp., consume juvenile P. canaliculata (Hendarsih et al. 1994). The
23 key predator of the native apple snail, the endangered Florida Snail K ite, feeds spari ngly on the invasive species, possibly because of snail siz e constraints (Cattau et al 2010). Birds have been observed to take as long as eight hours to open one nonindigenous snail, and juvenile birds appear to be less success ful foraging on large nonind igenous snails ( Mr. Jeff Kline, U. S. National Park Service, Dr. T. Collins, and Dr. Phil Darby, University of West Florida, personal communication s 2006 ). Snyder and Kale (1983) however, provide evidence that failure to handle larger snails is nothing mo re than clumsiness on the part of the bird. Observations of kites on Lake Okeechobee, also contradict the conclusions of Cattau (Bernatis 201 3 ). Furthermore, limpkins have been observ ed successfully feeding on P. maculata (J. Kline and P. Darby personal co mmunication 2006 ). In some locations, presence of P. maculata has even been credited with the return and success of limpkins ( Mr. Charles Lee, Florida Audubon Society, personal communication 2011 ). Pomacea paludosa is the only native apple snai l in the United States and is the primary food source of the endangered Florida Snail Kite (Darby 2005; Kushlan 1975). The native snail exhibits similar life history patterns as the nonindigenous (e.g. laying eggs on emergent vegetation, aestivation in drought), bu t differs in critical aspects such as foraging behavior, number of offspring and lifespan. Th e native snail feeds predominant ly on periphyton and particles in the surface water film (McClary 1964) I have observed frequent consumption of the plants Hydr illa verticillata and Vallisneria americana T he invasive apple snails fee d primarily on macrophytes and the refore consumes the attached periphyton. Competition for preferred food likely occurs once macrophytes have been largely consumed and the
24 main food source is the suspended particles in the surface water film. In some systems (e.g. urban ponds) where native and nonindigenous apple snails coexist food may be a limiting factor for native apple snail survival T he number of offspring per clutch diff ers more than 10 fold between P. p al udosa and the invasive snails. Clutches of P. paludosa have an average of 28.3 eggs, with average egg size ranging from 3.7 5.6 mm (Hanning 1979; Figure 1 3). Variation in clutch sizes of these invasive species is grea t, with estimates of 200 12 00 eggs per clutch (ASE 2006). Individual egg sizes of the invasive species are much smaller, generally ranging from 0.9 1. 9 mm (personal observation). Native apple snails are much larger at hatching, have rows of pila, and are pigmented, whereas the nonindigenous are smaller, typically do not posses s pila, and shell pigmentation may or may not be present (Figure 1 7). Decline of the native apple snail and the Florida Snail Kite may simply be a result of prolonged natural and anthropogenic perturbations. Darby (2005) suggests prolonged drying caused by water management decisions has impacted native apple snail populations in some locations. The Florida Snail Kite is also impacted by dry conditions and is known to move in respo nse to fluctuating water levels ( Dr. Jim Rodgers, FWC, personal communication 2011 ). Although the introduction of invasive snails has been suggested as the reason for declining native apple snail populations and Florida Snail Kites there is no quantitativ e evidence to support this hypothesis. Summary and Research Goals Florida has numerous aqua tic ecosystems. These systems regulate water discharge and recharge, support freshwater fisheries and wildlife, and some are dominated by submerged an d emergen t aquatic vegetation. Any organism that forages
25 indiscriminately and continuously on aquatic plants has the potential to alter Florida aquatic ecosystems. For instance, macrophytes dominated communities could be transformed into phytoplankton dominated com mu nities (Carlsson et al. 2004). Given the recent range expansion of P. canaliculata complex snails in Florida, the potential impacts of this organism require investigation. Although there are a many questions to be examined regarding P. canaliculata comp lex snails three general questions have arisen: 1) what are the physiological tolerances that affect the range and habitat use of P. canaliculata spp. in Florida, 2) is there a morphological means to distinguish among the species, and 3) what are the impa cts of P. canaliculata spp. on the flora used in restoration and managed for ? Each of these questions is addres sed in the following chapters. In Chapter 2, variation in shell morphology between hatchlings, juveniles, and adults of P. maculata and P. canali culata is examined to determine if external morphological characters are reliable for field identification. Chapter 3 aims to fill basic research gaps in physiological tolerances to common environmental factors that affect distribution and success of estab lishment. A series of LC 50 trials to determine lethal levels of pH, salinity, and desiccation are explored Chapter 4 reports on a series of feeding rate and feeding preference trials focused on determining impacts on vegetation by P. maculata and P. c analiculata Chapter 5 presents results of a project designed to evaluate hand removal as a means of controlling or eradicating P. canaliculata comp lex snails from urban systems. Answering these questions fulfills two important objectives First, there is a dearth of general ecological, physiological, and morphological data on introduced Pomacea i n Florida, and my research help s fill this void. Second, my research will help
26 management agencies develop guidelines for regulation of these nonindigenous apple s nails, assist in developing lake r estoration plans (i.e. plants to use/avoid), and provide insight s for developing control programs in residential or small urban ponds.
27 Figure 1 1. States with Pomacea in North America Florida (in black ) has four spec ies of nonindigenous Pomacea ( P. maculata P. canaliculata P omacea sp. and P. diffusa ) one native species ( P. paludosa ) All other states (in grey) have at least one nonindigenous Pomacea species.
28 Figure 1 2. Nonindigenous Pomacea populations throu ghout Florida as of 201 3. These are point locations and do not fully represent the spread of the snails in canal and river systems or around the periphery of lakes. The P. maculata location in the panhandle j ust north of the state line, in Georgia, represe nts records from Lake Seminole which extends a cross the Florida Georgia border
29 Figure 1 3. Egg masses of Pomacea spp. in Florida. Top left: P. paludosa (native), bottom left: Pomacea sp., middle: P. maculata top right: P. canaliculata and bottom right: P. diffusa Photos of P. diffusa and Pomacea sp. courtesy of Dr. Ken Hayes. Other photos by Jennifer Bernatis.
30 A B Figure 1 4 Holding tank observation of: ( A ) P. maculata using its foot to manipulate root materi al for consumption; (B ) P. canaliculata using its foot to manipulate lettuce for consumption Photos by Jennifer Bernatis. Figur e 1 5. Measurement of added shell (dark line on right side of shell 3.4 mm) a fter 80 hrs in a low population density, hig h food abundance tank of small P. maculata (< 40 mm TL). Photo by Jennifer Bernatis.
31 Figure 1 6. Pomacea maculata egg masses in various stages of development and hatching. Photo by Jennifer Bernatis. Figure 1 7 Two day old hatchling apple snails, o n the left P. paludosa and on the right P. canaliculata Photo by Jennifer Bernatis.
32 CHAPTER 2 MORPHOLOGY Background Nonindigenous f reshwater gastropods of the genus Pomacea (Ampullariidae) are native to South and Central America and have recently undergo ne several taxonomic revisions ( Cazzaniga 2002; Hayes et al. 2009, 2012). Pomacea have been introduced into many countries and become established in natural and agricultural aquatic ecosystems. Pomacea were introduced for biocontrol, the aquarium trade, an d the escargot industry ( Naylor 1996; Halwart 1998 ). In the last decade, Pomacea have received much attention because foraging by some species has devastated taro and rice crops ( Yusa and Wada 1999; Joshi 2001; Sin 2003 ) and because they are considered th reats to natural we tlands (Carlsson et al. 2004). Two commonly introduced species implicated in crop damage and ecosystem alteration are Pomacea canaliculata Pomacea maculata ( Perry 1810) ( Lowe et al. 2000; Rawlings et al. 2007, Carlsson e t al. 2004; Hayes et al. 2009, 2012 ). Unfortunately, because of multiple introductions, inherent variation within species, shared common names, and limited knowledge of the ecophysiology of the genus, identification can be both complex and problematic. Recent studies have focused on the genetic identities of Pomacea in both native and introduced ranges ( Rawlings et al. 2007; Hayes et al 2009, 2013 ; Matsukura et al. 2013 ). These studies helped c larify species identities, but there was little emphasis on field identification methods that can be used by non malacologists (e. g., natural resource managers). (Estebenet and Cazzaniga 199 8; Estebenet and Martin 2003; Estebenet et al. 2006), as
33 most recent studies have focused on the genetics of the introduced p opulations in Asia and Hawaii. Except for the work by Rawlings et al. (2007) that established the identities of many U.S. populatio ns, research on differentiating species within the continental U.S. is scarce. In the U.S., nonindigenous apple snails are known from 11 states (Figure 1 1 ). Genetic analysis revealed that Florida has more Pomacea species than any other state (Rawlings et al. 2007). These include the native, P. paludosa (Say 1829) and at least four nonindigenous species: P. canaliculata P. maculata P. diffusa (Blume 1957), and P omacea sp (formerly considered P. haustrum ; Hayes et al. 2009) As of 2012 Florida ha d nonin digenous Pomacea populations in at least 29 watersheds covering 38 of 67 counties (Figure 1 2). In the U.S., the pet industry is responsible for the introduction of numerou s nonindigenous and invasive species. Recognizing the damage that can be caused by P. canaliculata in April 2006 the U.S. Department of Agriculture, Animal and Plant Health Inspection Services, restricted importation and interstate transport of all nonindigenous Pomacea spp. except for the species complex Pomacea bridgesii / diffusa Sna ils in t he P. bridgesii/diffusa complex are algavores and likely of little threat to agriculture or natural systems. Both P. canaliculata and P. maculata however, continue to be sold in pet stores and on line, thus violating the interstate transport restr iction (personal observation). The rationale provided by the pet industry is that it is not possible to disting uish one species from another. Whereas it may be true that it is impossible to distinguish between the prohibited species P. maculata and P. cana liculata the argument does not apply to differentiating these species from P. diffusa or P. paludos a
34 Shell morphology can be used to separate P. paludosa from P. diffusa and both from P. maculata and P. canaliculata (Thompson 2000). External morpho logical characteristics can be influenced by environmental factors (Tanaka et al. 1999; Martin and Esteben e t 2002; Estebenet et al. 2006) or selective breeding (Yusa 2004). Internal anatomy is not altered by environmental factors and is useful for identifi cation. This approach, however, requires examination of male specimens, as differences are typically observed in the glands of the male reproductive system (Hayes et al. 2012). Moreover, live specimens may not always be available for identification. Egg m asses can assist with identification because egg mass characters vary among Pomacea spp. in North America. Nonetheless, all Pomacea spp. found in North America lay eggs that turn white/grey before hatching, thus hindering identification among species with egg masses that appear similar ( e.g., P. maculata and P. canaliculata ). General characteristics of Pomacea eggs are provided in Table 2 1. Variation in egg clutches, however, is common and may be influenced by factors such as temperature, habitat quality, population density, and snail age (personal observation; Lacani lao 1990; Tanaka et al. 1999). In addition, egg masses may not always be available to facilitate species identification. In 2006, a population of nonindigenous Pomacea in Duval County was selec ted for research. Personal observation of other Pomacea populations (i.e ., P. maculata ) suggested that shell shape and egg masses at this location were different. Genetic analysis confirmed that this was the first population of P. canaliculata in Florida ( Dr. T. Collins, personal communication 2006 ) and led to the hypothesis that shell variation
35 between P canaliculata and P. maculata could be used to d istinguish between the species. If shell morphology can be used to differentiate among nonindigenous snail species, it should provide a valuable tool to help reduce the spread of nonindigenous Pomacea This study was undertaken to identify shell morphological characteristics, in combination with reliable identification methods (e.g., egg mass), to provide a me ans of rapid field identification of the two most commonly misidentified Pomacea species in the southeast U.S., P. canaliculata and P. maculata Correct identification of the species may help determine potential impacts on a system. For instance, basic bi ological attributes of the species, such as reproduction and feeding differ, thus resulting in different impacts. Second, because of differences in biology, correct identification may help determine the best metho ds for control or eradication. Perhaps most importantly, correct identification may help reduce the spread of these invasive species through misidentification in the pet industry or elsewhere. Methods Collection and Holding Procedures Adult and juvenile Pomacea maculata were collected from mul tiple locations across Florida. The locations chosen had verified species identification as reported in Rawlings et al. (2007) and with Dr. Timothy Collins (personal communication 2006 ). At the beginning of the study, in 2006, the only known P. canaliculat a location was in Duval County, Florida. A number of shells from a California population were, however, supplied by Dr. T. Collins and 156 shells from Argentina were provided by Dr. Pablo Martin ( Universidad Nacional del Sur ). Total numbers of adult snai ls used in the study were 739 P. maculata and 741 P. canaliculata
36 bias (i.e., undue representation from a single clutch) no more than three snails were collected within 25 m of each other during a sampling event. Ninety three juvenile P. maculata and 86 juvenile P. canaliculata were used. Upon return to the laboratory, snails were placed in 120 liter, re circulating aquaria (filled with 100 L water), separated by collection loca tion, and fed hydrilla, pennywort, and lettuce (romaine and iceberg) three times per week Snail density never exceeded 15 snails per tank. A 50/50 mix of well water and pond water was used. Conditions similar to those in Florida lakes were maintained: tem perature was 23 26 o C, dissolved oxygen was 90 100% saturation, salinity was < 0.20 ppt and pH was 8.2 8.4 (YSI 560). Hatchlings were from egg clutches collected in the field and returned to the lab for hatching. Measuring Procedures Using digital caliper s, 10 measurements (e.g. total length [TL], total width [TW], vertical height [VH]) on shells of adult and juvenile snails were made to the nearest 0.01 mm (Fig. 2 1). These 10 measurements, modified from Estebenet and Martin (2003) were used to generate 4 5 ratios and analyzed for differences between: 1) populations of P. maculata adults, 2) populations of P. canaliculata 3) adult P. maculata and P. canaliculata 4) juveniles of both species and 5) hatchlings of both species. Snail shells were sometimes da maged and only shells for which at least seven of the measurements were obtainable were used. Shells with non uniform growth patterns were retained because such patterns represent natural variation. Measurements were taken within two weeks of returning to the laboratory. Measurements on hatchlings were limited to TL and made within three days of hatching using a lens micrometer on a dissecting microscope. The relation between shell TL and mass was also investigated, with shell mass determined to the nearest 0.1 g (Ohaus Navigator Pro).
37 Statistical Analysis A preliminary cluster analysis and non parametric MANOVA was conducted on the P. maculata populations. Results revealed no significant differences among populations and therefore these data were pooled. L ikewise, preliminary analysis indicated that the Argentina and California P. canaliculata were comparable to the Florida population and data from the three localities were pooled. Preliminary cluster analyses suggested snail size can be used to distinguish species. Size groups (TL) 49.99 mm, and a group combining both size classes. The results of all three groupings, all adults, large adults and small adults, are present ed. All analyses were conducted between species of each age/size group and within species and between age/size groups. Each size class of each species was treated and labeled as a distinct group in the analyses: P. maculata large adults = 1, small adults = 2, juveniles = 3; P. canaliculata large adults = 4, small adults = 5 and juveniles = 6. For analyses in which adult size groups were combined, the species designation was P. maculata = 1 and P. canaliculata = 4. The data were analyzed with several non p arametric multivariate techniques. Multiple approaches were used to test the findings and identify the simplest method for species identification. First, a traditional non parametric analysis of variance (Proc GLM with ranked data) was conducted. This met hod follows classical morphometric approaches and yields results that are easily interpreted by non specialists. A second approach compared length/mass ratios. This was conducted with adult shell material, using non parametric analysis of variance (proc NP AR1WAY).
38 Several classification techniques were also used. A hierarchical cluster test for species variation, with classification as a post test, using results provided by th e discriminant procedures described below (proc CLUSTER). The cluster frequency does not require a priori knowledge of species classification. In this study, the number of potential clusters was limited to four, accounting for both species and age groups. The following discriminant methods differed from the cluster analysis, as they require a classification variable (i.e., species). A canonical discriminant analysis was used to find the linear combinations of quantitative variables that best revealed the di fferences among classes (proc CANDISC). Next, forward selection stepwise discrimina n t analysis was conducted to reduce the total number of variables (proc STEPDISC). Only those that maintained that level of significance through subsequent iterations are part of the final model. Finally, discrimina nt function analysis was used to classify the observations into a species (proc DISCRIM). All analyses were conducted on the full 45 ratio model and the reduced model derived from the stepwise procedure to evalu ate the utility of the reduced model. All analyses were performed with SAS 9.3 (SAS Institute Incorporated, Cary, North Carolina). Results Nonparametric MANOVA The mean ratio value between size groups within species, and between species within size group s were analyzed (Table 2 2). The full model analysis of the combined adult class (CAC, all snails > 30.00 mm TL) resulted in a significant difference between species (P < 0.0001). All ratios for the two species, except TLVH, TLPA and VHPA,
39 were significant ly different. In the large adult class (LAC, > 50 mm TL), the analysis resulted in a significant difference between the species (P < 0.0001). In this comparison, all but TLVH and ALBL were significantly different. In the small adult class (SAC, 30.00 49. 99 mm TL), the analysis resulted in a significant difference between species (P < 0.0001). Ratios that were not significantly different included TWAW, TWBL, VHSH, and AWBL. Generally, as size decreased, so too did the difference between species. Significa nt differences were observed w ithin species and between adult size groups (P < 0.0001) for both P. maculata and P. canaliculata Ratios that were not significantly different for P. maculata were TLSH, TWVH, VHBL and AWAP The co mparison between size groups of P. canaliculata had a markedly different outcome. In this analysis only 22 ratios were significantly different Between species analysis of the juvenile group resulted in a significant difference be tween species ( P < 0.0001 ; Table 2 3). Significant di fferences for 28 ratios were observed between juveniles of both species. Juvenile P canaliculata were significantly different for 35 ratios compared with the CAC, 34 ratios with the LAC and SAC (all P values < 0.0001). Juvenile P. maculata were significan tly different from the CAC for 37 of the ratios, 38 ratios in the LAC and 33 for the SAC (all P values < 0.0001). Although the majority of the variables differed for each species between juvenile and adult classes, many of the actual differing variables we re species specific. Only TL was measured for hatchling s snails. A total of 750 three day old hatchlings were measured (N=375 for each species) in May, July and September. Data were used to conduct three analyses: 1) test TL differences between species w ithout regard to time of season, 2) test TL differences between species for each sampling period, and 3) the overall effect of time of year on TL. In the analysis that disregarded date, there was a significant difference in mean TL between species (P < 0.0 001). The
40 overall mean TL for P. canaliculata was 1.8 mm (SE = 0.007) and for P. maculata was 1.5 mm (SE = 0.006) The effect of season on TL was also significant for both species (P < 0.0001). P. canaliculata hatchlings increased from a mean TL of 1.7 mm (SE = 0.006) in May to 1.8 mm (SE = 0.007) in July and to 2.0 mm (SE = 0.007) in September. Pomacea maculata increased throughout the season from 1.4 mm (SE = 0.008) in May, to 1.5 mm (SE = 0.008) in July, to 1.6 (SE = 0.007) in September. Analysis of t he relation between TL (mm) and M (g) was conducted on adult shells. In the CAC analysis, the Kruskall 0.0001). The class mean for P. maculata adults was 5.34 (SE = 0.190) and for P. canaliculata 12.28 (SE = 0. 225). There was overlap in the ranges of the species, P. maculata ranged from 1.44 34.97 and P. canaliculata from 4.38 37.82. The LAC Kruskall P. maculata the class mean and range were 3.87 (SE = 0.079) and 1.44 13.74. For P. canaliculata the class mean and range were 9.14 (SE = 0.167) and 4.38 16.14. In the .4464). The class mean and range for P. maculata were 15.42 (SE = 0.731) and 2.33 34.97; P. canaliculata mean and range were 15.57 (SE = 0.334) and 6.43 37.82. Preliminary Canonical Di scriminan t (CDF) and Discriminant Function Analyses The initial hypothesis was that species and shell morphology ratios are correlated. Analysis of the full model (45 ratios) using both species and large/small adults, juveniles resulted in a significant model 0 .050, F 225, 7403.7 = 27.08, P < 0 .0001) with five significant canonical correlations ( P <0.0001; Table 2 4, Figure 2 2A). The first canonical variable (Can 1) was well correlated with the class means and exp lained 94.97% of the variation, where as Can 2 accounted for much less of the
41 variation explaining only 74.80%. Discriminant function resubstitution classification accurately classified species regardless of size, 97.82% of the time for P. maculata (N= 719 of 735) and 88.77% for P. canalicula ta (N = 712 of 802). In the cross validation a ccurate placement decreased to 89.79% and to 86.65% respectively The full model for the CAC analysis was significant 0 .196 7 F 45, 1294 = 117.49, P < 0 .0001, Figure 2 3A). In this analysis Can 1 (0.89630) is well related ( R 2 = 0.803 ) with the class mean s P. maculata = 2.12 and P. canaliculata = 1.9 2 Discriminant function resubstitution classification accurat ely placed 96.06% of P. maculata and 86.64% of P. canaliculata whereas the cross validation accurately placed 93.55% of P. maculata and 85.94% of P. canaliculata Individuals that were misclassified tended to be outliers in Figure 2 3 A and were generally snails of the smaller size class. Results of the full model SAC were significant 0 .31 6 F 45, 407 = 19.58, P < 0 .0001; Figure 2 4A). Can 1 ( 0.827 1), although significant was weakly related ( R 2 = 0.684 ) with the class means, P. maculata = 2.87 and P. canaliculata = 0.75. In the discriminant function resubstitution, 100.00% o f P. maculata were correctly placed and 99.16% of P. canaliculata were successfully placed. However, in the cross validation, P. maculata dropped to 62.77% successful placement but P. canaliculata remained at 99.16%. Results of the LAC and full model wer e significant 0 .16 5 F 45, 841 = 94.68, P < 0 .0001 ; Figure 2 4 C). Can 1 ( 0.913 9) was strongly related ( R 2 = 0.835 ) with the class mean s, P maculata = 1.79 and P. canaliculata = 2.8 2. Accurate placement of the snails increased when using large snails. Discrimi nant function resubstitution
42 correctly placed P. maculata 98.52% of the time and P. canaliculata had an 88.99% placement. In the cross validation correct placement declined slightly for each species. For P. maculata the placement rate dropped to 97.23% an d P. canaliculata fell to 87.25%. Results of the juvenile full model were significant 0 .32 6 F 45, 151 = 6.94 P <0.0001; Figure 2 4 E). In this analysis Can 1 (0.821) was not strongly related ( R 2 = 0.674 ) with the class means, P. maculata = 1.423 and P. canaliculata = 1.43 8 The results of the classification analysis are simila r to the overall analysis. Discriminant function resubstitution correctly placed P. maculata 98.52% of the time and P. canaliculata 88.99% In the cross validation correct placement declined slightly for each species. For P. maculata the placement rate d ropped to 97.23% and P. canaliculata fell to 87.25%. The final CDF compared wi thin species and between adults (as separate size classes ) and juveniles. The overall model results for differences between age groups of P. maculata were significant 90, 1376 = 23.99, P < 0 .0001) and coupled with the output in Figure 2 5A, which indicates that a shift in growth pattern may occur between juveniles and adults T he Can 1 score (0.8557) was not strongly related ( R 2 = 0.7322 ) with the cla ss mean s, adult large = 0.8 9 adult small = 1.0 7 and juveniles = 3.8 5. R esults of the overall model for differences between age groups of P. canaliculata analysis were also significant 90, 1510 = 12.77, P < 0 .0001). However, Can 1 ( 0. 788 ) was much lower than observed in the P. maculata analysis. Furthermore, the model was only weakly related ( R 2 = 0.621 6) with the class mean s, adult large =
43 0.57, adult sma ll = 0.38, and juvenile = 3.42 As for P maculata the adult size groups are si milar, but the juveniles are a distinct group (Figure 2 5 B). Forward Stepwise Analysis Combin in g all size groups required 32 of the 45 ratios for species separation. Those not included were: TLAL, TLAW, TLSH, TLPA, TWBL, VHPA, ALSW, ALAP, SHSW, SHBL, SW PA, APBL, and PABL. Twenty one of the variables entered and remaining in the model had P values of < 0.0001. N o variables were removed once added to the model. A nalysis of juveniles found that far fewer variables were nec essary to distinguish species. Only five variables met model criteria and therefore stayed: SHS W, ALAP, VHSW, AWSH, and SWAP. N o variables were removed in the process and the first three were significant at P < 0.0001. AWSH and SWAP were included in the model with significance values of P = 0 .0084 and 0 .0014 respectively. In the CAC, 14 variables were necessary for the model. F ive variables initially entered into the model were subsequently removed: SWPA, ALSW, SHSW, AWAP, and AWSH. The final va riables in the model included: TLTW, VHAP, SWBL, TLAP, TWPA, TWSW, TLSW, SWAP, SHBL, ALA P, SHPA, AWPA, VHSW, and TLVH. The majority of the variables were significant at P < 0 .0001, but three were only significant at P < 0.05 ( TLAP, SHBL and VHSW). Small adults required 11 variables for species sep aration. The variables SWAP and VHAW were entered and removed from the model; SWAP was re entered into the model at step 12. The final model variables were: TLTW, TWAL, VHSH, ALAP, SWAP, VHSW, TWAW, T WSW, AWSW, TWVH, and ALBL. The variables were all highly significant most with P < 0 .0001 TWVH w as the least significant at P = 0 .0063
44 Fifteen variables were required to separate large adults. The variables SWPA and SHAP were entered, removed, and re entered (step 17 and step 19, respectively) for the final model. The final variables included in this model were: TLTW, SHSW, ALBL, VHBL, TWAP, TWSW, TLSW, TLAP, SWBL, TLBL, ALPA, PABL, SWAP, TLPA, and SHAP. Follow up Canonical Discriminant and Discriminant Function Analyses Using the results of the forward ste pwise analysis the canonical discriminant and discriminant fu nction analyses were repeated. The reduced model results using each size 1 F 160, 7432.4 = 37.43, P < 0.0001). The model was based on 32 ratios, therefore, the similarity in results to the full model is not surprising (Figure 2 2B). The Can 1 score (0.8899) was f airly strongly related ( R 2 = 0.79 66) with the class means (Table 2 4). The results of the discriminant function cross validation were simil ar to the whole model results. Overall P. maculata were placed within the correct species 91.29% of the time an inc rease of 1.5% in placement success. There was no change in the placement success rate ( 95.20% ) of large adults into species. There was an increase in placement success for both small adults (89.36% vs 80.85%) and juveniles (71.71% vs. 68.68%). The results for P. canaliculata were not as successful, showing a slight decrease in overall species placement (85.28% vs 86.65%). Successful placement of large adults into species decreased from 84.05 % to 83.76%; small adults decreased from 88.30 to 86.07%; and ju veniles decreased from 89.79% to 87.75% Overall, there were an additional 11 P. canaliculata misclassified as P. maculata, but 21 fewer P. maculata were misclassified In the CAC analysis 3 F 14, 125 = 350.00, P < 0.0001). This model used 14 ratios, and although Can 1 ( 0.887 ) was slightly
45 lower than the 45 ratio model, the actual results of the cross validation improved. In this model, Can 1 was less well related ( R 2 = 0.787 ) with the class means than the full model ( P. maculata = 2.02 2 and P. canaliculata = 1.826; Figure 2 3B). C ross validation had a 94.81% accurate placement of P. maculata an increase of 1.26%, and a 92.61% accurate placem ent of P. canaliculata an increase of 6.67%. The results of the SAC analysis were more similar to those of the juveniles than large adults. 1 F 11, 442 = 77.69, P < 0 .0001) (F igure 2 4B). However, Can 1 ( 0.8118 ) was only moderately related ( R 2 = 0.659 ) with the class means, P. maculata = 2.7 2 and P. canaliculata = 0.7 1. In the cross validation placemen t 88.30% of P. maculata were correctly placed and 94.72% of P. canaliculata were successfully classified. Placement of P. maculata increased by 25.53%, but a slight decrease by 4.44% was observed for P. canaliculata The final CDF model for small adults c an be expressed as Can1 = .00292(TLTW) + .00442(TWAL) + .00250(VHSH) + .01217(ALAP) .00659(SWAP) .00838(VHSW) .00865(TWAW) + .01384(TWSW) .01250(AWSW) .00630(TWVH) .00251(ALBL). R esults for the LAC analysis were more definitive for identifying species. The 15, 875 = 274.93, P < 0 .0001) The Can 1 score ( 0.9082 ) was strongly related ( R 2 = 0.82 5) with the class means P. maculata = 1.72 and P. canaliculata = 2.7 3. The cross val idation had the greatest success with this test group, and both species exceede d a 90% correct placement rate. The 97.07% result for P. maculata was a non signficant slight decrease, 0.16%, from the original model However, P. canaliculata increased by 5.2 1% to 92.46% correctly classified (Figure 2 4D) The final large adult CDF model can be expressed as
46 Can1 = .00598(TLTW) .00239(SHSW) + .00078(ALBL) + .00041(VHBL) .01557(TWSW) .00591(TLSW) + .01542(TLAP) + .00862(SWBL) .00143(TLBL) .00106(ALPA) .00235(PABL) .00457(SWAP) .00161(TLPA) + .00334(SHAP). 0 .469, F 5, 191 = 43.21, P < 0.0001). However, Can 1 ( .072 9) was much lower than the full model In this analysis Can 1 was weakly related ( R 2 = 0.53 1) with the class means, P. maculata = 1.05 and P. canaliculata = 1.06. Although the overall strength of the relation was reduced, using this model had a slight increase in overall classification rate (83.24% vs 82.74% ; Figure 2 4F ). T he placement for P. maculata remained unchanged at 81.81% and P. canaliculata increased s lightly to 84.69% from 83.67%. The final model for juveniles can be expressed as Can1 = .03368(SHSW) + .00973(ALAP) .01349(VHSW) + .01062(AWSH) .00166(SWAP). Pre and Post Cluster Frequencies Results of the preliminary cluster analysis for the full model and all six classifications suggested that 2, 3, or 6 clusters could be used; the decision to use 2 clusters was used because two species were present. It was the visual output, however, that provided the support for variation between ages, as three groups emerged, one for each species and a separate group that consisted primarily of juveniles from both species (Figure 2 6). In the full model preliminary cluster an alysis, 91.07% of P. maculata were found in cluster 2 and 84.16% of P. canaliculata were found in cluster 1 In the post cluster analysis, using the 32 variabl e model 89.88% of the P. maculata were in cluster 2, but 85.47% of P. canaliculata were in clust er 1 In the juvenile analysis, the preliminary run placed 76% of P. maculata into cluster 2 and 57% of P. canaliculata were placed into cluster 1 The follow up model
47 using only five variables, had a decrease in clust er placement for both species. The o ccurrence of P maculata was only 67% in cluster 2 and P. canaliculata reduced to 45% occurrence in cluster 1. Removing juveniles from the analysis improved the cluster results. In the preliminary run of the CAC, 95.91% of P. maculata were in cluster 2 a nd 90.62% P. canaliculata were in cluster 1. In the post CAC run, placement of P. maculata remained the same in cluster 2, but only 85.51% of P. canaliculata were in cluster 1. The SAC preliminary run placed 98.93% of P. maculata in cluster 1, but P. canal iculata were nearly evenly divided (cluster 1 = 58.49% and cluster 2 = 41.51%). The reduced model placed 100% of P. maculata into cluster 1, but also placed 88.61% of small adult P. canaliculata into cluster 1. Results from the LAC analysis were the highes t obtained of any test scenario, and occurred with both models. In the preliminary model 97. 60 % of P. maculata were in cluster 2 However only 82.31% of P. canaliculata were in cluster 1. With the use of the reduced model, the frequency of P. maculata in cluster 2 decreased slightly to 96.33% whereas the frequency of P. canaliculata in cluster 1 increased to 95.65%. Discussion The genus Pomacea has undergone extensive taxonomic revision, therefore problems with misidentification are not surprising. Pomac ea maculata has often been confused with P. canaliculat a The implications of misidentification have yet to fully emerge, but it has resulted in conflicting information in the literature. Several studies looked at morphological characteristics of Pomacea but typically within species and between populations (Estebenet and Martin 2003; Estebenet et al. 2006). Hayes et al. (2012) addressed identification of several Pomacea species using genetic analysis,
48 internal morphology and limited shell morphology. They concluded that there are characteristics that can be used to discern among species. That study, however, focused predominantly on populations outside North America. My study is the first to focus on shell morphology for introduced Pomacea in North America that considers multiple species and size classes. I found distinct differences in shell morphology between species in different size classes, as well as differences within species and between size classes. Because all P. maculata were from Florida, as wer e the majority of P. canaliculata care should be taken in applying the results beyond the southeast U.S. Furthermore, in this study, sexual dimorphic growth patterns were not included as a variable, though they have been reported elsewhere (Estebenet and Cazzaniga 1998; Estebenet and Martin 2002; Estoy et al. 2002a). Because shells of individuals are often the only material collected, sex of snails was not considered, making these findings more applicable for people confronted with the organisms (e.g. wild life managers, pet trade professionals). The multiple approaches taken in this study encompass the many techniques used in species identification. Each method found differences between the species, with varying degrees of success depending on snail size. T he following discussion emphasizes those inter species differences that are readily observed within each size class. Prior observations of egg size and newly hatched snail size indicate differences between species ( Barnes et al. 2008 ; Tambur i and Martin 2011) In long term studies of Florida populations, however, this difference was not observed. In Florida, perhaps because of the relatively warm climate and extended breeding season (all year in south Florida), differences in egg size and hatchlings were observed. As the summer
49 progressed, the size of eggs and hatchlings in both species increased. High variability in hatch size, however, showed the species overlap, making species identification by hatchling difficult. If hatchlings are obtained in the fie ld and no other means of identification (i.e., egg masses, larger specimens) are available, identification only to the genus level ( Pomacea sp.) is suggested. Juvenile snails exhibit a narrow range of morphological variation, within and between species, a nd identification is difficult, as indicated by the results of the cluster frequency. MANOVA analysis, however, demonstrated that there are distinguishing characteristics, as more than half the morphometric measures were significantly different between spe cies. The final model generated by the CDF required only five variables. Unfortunately, the statistical analysis indicates that at best, the snails will be properly classified < 85% of the time. Furthermore, lab reared snails were not included in this stud y. Personal observations suggest that, depending on rearing conditions, distinguishing between these two species is extremely difficult for snails <20 mm TL, as growth patterns are nearly identical. To correctly identify individuals in the juvenile age class, additional steps are necessary. First, ensuring that specimens from multiple clutches are utilized will help prevent the possibility of relying on specimens that may be from the same clutch and share inherent traits. Another option is to raise snai ls to a larger size. Because growth can be affected by habitat conditions, it is important to control population densities and provide adequate nutrition (Estebenet and Cazzaniga 1992; Tanaka et al. 1999). Furthermore, the presence of multiple species may impact snail growth. Connor et al. (2008) found that presence of nonindigenous Pomacea mixed with the native P.
50 paludosa reduced the growth of the native apple snail. Finally, although this work was designed to minimize the effort required to identify sp ecies, the results suggest that larger rather than smal ler numbers of measures promote accurate identification. As the snails increase in size, identification based on shell morphology is more reliable. Once the snails are > 30 mm TL, accurate species id entification with shell material is possible, as each species has developed specific characteristics. The size of the snail, however, determines which variables are more important. Models for distinguishing small and large adult taxa share only four variab les (TLTW, SWAP, TWSW, and ALBL). This lack of commonality between the models indicates that necessity for larger snails in identification, with larger individuals yieldin g the best model. Likewise the correct assignment of small snails in the cross validation varied between species and was low, and placement in the frequency clusters was a combined 75.17% success. Although this was low, it was still greater than for juveni les (56%), again pointing to shifts in growth through life stages. The models used for large adults had consistently h igher classification success (> 90.00%). Fourteen variables were used, but one measure was singled out as easy to use and was common in all models for adults (TL:TW). This measure is particularly useful because it is easily obse rved in the larger specimens (> 50.00 mm TL), and the larger the specimen the more obvious. While holding the specimen with the spire up and the aperture facing the observer, P. maculata is generally wider than it is long, whereas the opposite is true for P. canaliculata Specifically, if the ratio is < 0.97 the
51 snail is likely P. maculata but if the ratio is > 1.0 5 it is typically P. canaliculata In smaller adults, with TL 30.00 49.99 mm, this guideline value is 1.02. Other defining aspects of shell shape are found in characteristics associated with spire width (SW). The spire width, in relation to several other measures (i.e. TL, VH), is different between spe cies. The relationship of SW to other variables indicates that the spire width is greater in P. maculata than it is in P. canaliculata Given that P. maculata is generally a wider snail, this is not surprising. Several variables, TLSW, TWSW and AWSW, are s ignificantly different between species, regardless of size class. In large adults these ratios differ between species by at least 0.6 mm, P. maculata being larger. The value is still larger in P. maculata for small adults, but the difference was reduced t o as low as 0.26 (AWSW). Variables related to aperture shape (AL, AW, and AP) should all be used with caution. Previous work indicated that aperture shape is a function of sexual dimorphism in P. canaliculata ( Cazzaniga 1990). Another shell characterist ic that was hypothesized to be useful for identification was the shell length to mass ratio. Live snails are influenced by many variables (e.g., habitat, reproductive status) that can affect weight. For example, the mass of female P. maculata has been obse rved to decrease by 10 20 g after laying eggs (personal observation). To avoid complications related to this factor, yet not restrict specimens to males, shells were used. Large adult P. maculata were significantly heavier, relative to TL, than any other t est group. In general, P. maculata shells tend to weigh more than P. canaliculata shells. Overall, the data suggest that snails > 50 mm TL, with a TLM value of <4.0, are likely to be P. maculata Adults 30.00 49.99 mm TL, with a TLM value < 6.0 are likely to be P. maculata Although these results suggest this is a good means for
52 identification, it should not be used as the sole criterion. As previously stated, growth rates of snails can vary tremendously based on environmental conditions. Although the result s of this work provide a strong framework for differentiating P. maculata from P. canaliculata sole reliance on shell morphology leads to a high error rate in identification. All of the variables used in this study may be impacted by environmental conditi ons. Multiple specimens should be collected from a site, preferably from multiple seasons, for identifications. Combining shell morphology with other species identifiers (i.e., egg masses) increases the likelihood of making a correct identification. Only t hrough correct identification of a species can there be a true understanding of its biology and role in ecology.
53 Table 2 1. General egg characteristics of Pomacea spp. in North America. V alues are ranges commonly reported in the literature and personally ob served. They probably do not include the true range of variation. Species Color Eggs/clutch Egg diameter (mm) Clutch length (mm) Layers P. paludosa pale salmon 20 80 4 6 35 65 1 P. diffusa pale salmon 150 250 2 3 25 50 1 2 P. canaliculata pink to red 300 600 2 4 35 70 2 4 P. maculata pink to red 400 > 1,000 2 3 40 130 3 6 P. sp. green 200 300 3 5, but irregular in shape 40 60 2 3
54 PA Figu re 2 1 Shell measurements for adult and juvenile snails of P. canaliculata and P. maculata : Total length (TL), Total width (TW), Ap er tur e length (AL), Body length (BL), Ap er tur e width (AL), Spire width (SW), Spire height (SH), Ap er tur e projection (AP), Pa ri e tal attachment (PA), Vertical height (VH). Hatchling snails were only measured for total length. Photos by Jennifer Bernatis. VH TW BL SH TL SW AW AL AP
55 Table 2 2. Results of the 45 ratio measurements. Means ( X ) and range s are provided for each ratio for all test groups. P group and betwe en species comparisons. P within species and between larg e and small group comparisons. Only P values > 0 0001 are given ( see text for further explanation ) Adults Overall Adults Large Adults Small P. canaliculata P. maculata P. canaliculata P. maculata P. canaliculata P. maculata X (Range) X (Range) X (Range ) X (Range) X (Range) X (Range) TLTW 1.05 (0.94 1.22) 0.97 ( 0 .89 1.07) 1.05 (0.94 1.22) 0.97 ( 0 .89 1.05) 1.05 (0.96 1.22) 1 0.1806 1.00 (0.89 1.07) TLVH 1.42 (1.21 1.67) a 0.3319 1.41 (1.19 1 .65) 1.41 (1.21 1.58) a 0.1130 1.40 (1.19 1.65) 1.43 (1.22 1.67) 1 0.3070 1.45 (1.28 1.62) TLAL 1.35 (0.99 1.60) 1.30 (1.14 1.49) 1.35 (1.04 1.60) 1.30 (1.14 1.49) 1 0.0002 1.35 (0.99 1.60) 1 0.8480 1.28 (1.16 1.47) TLAW 2.07 (1.57 2 .60) 1.84 (1.43 2.32) 2.07 (1.59 2.60) 1.82 (1.43 2.32) 2.06 (1.57 2.49) 1 0.3376 1.98 (1.72 2.26) TLSH 1.76 (1.09 2.11) 1.82 (1.37 2.59) 1.74 (1.13 2.02) 1.82 (1.37 2.59) 1 0.4824 1.78 (1.09 2.11) 1.81 (1.56 2.14) a 0 .0005 TLSW 3.71 ( 2.24 5.03) 4.33 (3.26 5.61) 3.67 (2.24 4.71) 4.36 (3.45 5.61) 3.74 (2.78 5.03) 1 0.0060 4.12 (3.26 4.73) TLAP 4.01 (2.11 8.08) 2.64 (1.44 4.31) 3.91 (2.16 8.08) 2.60 (1.44 3.82) 4.09 (2.11 7.99) 1 0.0035 2.88 (1.98 4.31) TLPA 3.91 ( 2.06 8.23) a 0.3845 3.92 (2.79 5.84) 3.87 (2.85 5.21) a 0 .0344 3.97 (2.79 5.84) 3.96 (2.06 8.23) 1 0.0498 3.60 (2.91 5.17) TLBL 1.15 (1.00 1.64) 1.08 (0.98 1.32) 1.14 (1.06 1.59) 1.07 (0.98 1.21) 1.16 (1.00 1.64) 1 0.0339 1.10 (1.03 1.32) TWVH 1.35 (1.07 1.58) 1.45 (1.26 1.72) 1.35 (1.14 1.58) 1.45 (1.26 1.72) 1 0.9368 1.35 (1.07 1.57) 1 0.6196 1.44 (1.26 1.59) TWAL 1.28 (0.84 1.51) 1.33 (1.13 1.50) 1.29 (0.87 1.51) 1.34 (1.15 1.50) 1.28 (0.84 1.51) 1 0.0326 1.27 (1.13 1.45) TWAW 1.97 (1.49 2.59) 1.89 (1.54 2.32) 1.98 (1.57 2.59) 1.87 (1.54 2.32) 1.97 (1.49 2.39) 1 0.2470 1.97 (1.65 2.21) a 0 .6160 TWSH 1.68 (1.06 2.06) 1.87 (1.37 2.53) 1.66 (1.10 1.97) 1.88 (1.37 2.53) 1.70 (1.06 2.06) 1.79 (1.56 2.11) TWSW 3.53 (2.03 4.79) 4.45 (3.41 5.88) 3.50 (2.03 4.58) 4.50 (3.45 5.88) 3.56 (2.68 4.79) 1 0.0212 4.09 (3.41 4.66) TWAP 3.81 (2.01 8.05) 2.69 (1.51 4.24) 3.73 (2.16 7.71) 2.67 (1.51 3.75) 1 0.0073 3.89 (2.01 8.05) 1 0.00 36 2.83 (1.90 4.24)
56 Table 2 2 continued Adults Overall Adults Large Adults Small P. canaliculata P. maculata P. canaliculata P. maculata P. canaliculata P. maculata X (Range) X (Range) X (Range) X (Range) X (Range) X (Range) TWPA 3.73 (1.90 7.45) 4.02 (2.89 5.96) 3.69 (2.80 5.06) 4.09 (3.01 5.96) 3.77 (1.90 7.45) 1 0.1691 3.56 (2. 89 5.09) a 0.0002 TWBL 1.09 (0.93 1.44) 1.01 (0.99 1.31) 1.09 (0.96 1.36) 1.10 (0.99 1.26) 1 0.0001 1.09 (0.93 1.44) 1 0.9511 1.09 (1.02 1.31) a 0 .2129 VHAL 0.95 (0.64 1.21) 0.92 (0.72 1.14) 0.95 (0.69 1.21) 0.93 (0.72 1.14) 0.95 (0.64 1.15) 1 0.3461 0.88 (0.80 1.08) VHAW 1.46 (1.08 1.88) 1.31 (0.91 1.69) 1.46 (1.12 1.88) 1.29 (0.91 1.69) 1.45 (1.08 1.79) 1 0.2348 1.36 (1.20 1.62) VHSH 1.24 (0.77 1.50) 1.29 (0.94 1.83) 1.23 (0.77 1.50) 1.29 (0.94 1.83) 1.25 (0.85 1.48) 1.24 (1.08 1.54) a 0 .2120 VHSW 2.61 (1.51 3.51) 3.07 (2.36 3.98) 2.59 (1.51 3.25) 3.10 (2.45 3.98) 2.63 (2.04 3.51) 1 0.0468 2.83 (2.36 3.31) VHAP 2.82 (1.55 5.78) 1.86 (1.02 2.73) 2.76 (1.55 5.78) 1.84 (1.02 2.73) 1 0.0003 2.88 (1.55 5.76) 1 0.0064 1.97 (1.38 2.69) VHPA 2.75 (1.48 5.25) a 0.6510 2.78 (1.92 4.47) 2.73 (1.97 3.79) a 0 .0077 2.82 (2.02 4.47) 2.78 (1.48 5.25) 1 0.1027 2.48 (1.92 3.60) VHBL 0.81 (0.66 1.10) 0.77 (0.64 0.96) 0.81 (0.68 1.10) 0 .76 (0.64 0.96) 1 0.3969 0.81 (0.66 1.09) 1 0.8992 0.76 (0.67 0.91) ALAW 1.54 (1.18 2.13) 1.42 (1.17 1.81) 1.54 (1.20 2.13) 1.39 (1.17 1.71) 1.54 (1.18 2.10) 1 0.2774 1.55 (1.36 1.81) a 0 .0136 ALSH 1.31 (0.83 1.91) 1.40 (1.01 1.80) 1.29 (0.83 1.80) 1.40 (1.01 1.80) 1 0.0409 1.33 (0.83 1.91) 1 0.0007 1.41 (1.15 1.59) ALSW 2.77 (1.49 4.49) 3.33 (2.51 4.58) 2.74 (1.49 4.19) 3.35 (2.51 4.58) 1 0.0005 2.79 (2.06 4.49) 1 0.0861 3.21 (2.62 3.83) ALAP 2.98 (1.54 6.64) 2.04 (1 .11 3.15) 2.90 (1.61 5.60) 1.99 (1.11 2.97) 3.05 (1.54 6.64) 1 0.0108 2.24 (1.49 3.15) ALPA 2.91 (1.45 5.20) 3.00 (2.15 4.37) 2.87 (2.10 4.21) 3.03 (2.19 4.37) 2.94 (1.45 5.20) 1 0.0374 2.80 (2.15 3.73) a 0.0124 ALBL 0.86 (0.71 1.42) a 0.0007 0.83 (0.73 1.00) 0.85 (0.71 1.34) a 0 .1955 0.82 (0.73 0.99) 0.87 (0.72 1.42) 1 0.2015 0.86 (0.79 1.00) AWSH 0.85 (0.56 1.17) 0.99 (0.73 1.45) 0.84 (0.56 1.11) 1.00 (0.73 1.45) 0.87 (0.69 1.17) 1 0.0002 0.91 (0.75 1.06)
57 Table 2 2 continued Adults Overall Adults Large Adults Small P. canaliculata P. maculata P. canaliculata P. maculata P. canaliculata P. maculata X (Range) X (Range) X (Range) X (Range) X (Range) X (Range) AWSW 1.80 (1.07 2.98) 2.38 (1.48 3.31) 1.78 (1.07 2.84) 2.41 (1.48 3.31) 1.82 (1.35 2.98) 1 0.0094 2.08 (1.68 2.55) AWAP 1.93 (0.95 4.27) 1.43 (0.85 2.09) 1.88 (1.10 3.55) 1.42 (0.85 2.09) 1 0.8449 1.99 (0.95 4.27) 1 0.0008 1.43 (0.88 1.95) AWPA 1.90 (0.96 3.73) 2.14 (1.37 3.38) 1.87 (1.35 2.87) 2.20 (1.44 3.37) 1.93 (0.96 3.73) 1 0.0410 1.81 (1.37 2.73) a 0.0001 AWBL 0.55 (0.42 0.78) 0.58 (0.47 0.74) 0.55 (0.42 0.78) 0.59 (0.47 0.74) 0.56 (0.42 0.78) 1 0.0456 0.55 (0.47 0.67) a 0 .7107 SHSW 2.10 (1.37 3.25) 2.37 (1.81 3.53) 2.11 (1.37 3.25) 2.39 (1.81 3.53) 2.10 (1.51 3.19) 1 0.3646 2.27 (1.94 2.61) SHAP 2.28 (1.22 4.87) 1.45 (0.74 2.28) 2.25 (1.22 4. 87) 1.43 (0.74 2.18) 2.30 (1.22 4.43) 1 0.1810 1.59 (1.05 2.28) SHPA 2.22 (1.24 4.41) 2.15 (1.50 3.36) 2.22 (1.64 2.99) a 0.0005 2.17 (1.50 3.36) 2.22 (1.24 4.41) 1 0.4711 1.99 (1.52 2.81) SHBL 0.65 (0.52 0.99) 0.59 (0.43 0.77) 0.66 (0.55 0.99) 0.59 (0.43 0.77) 0.65 (0.52 0.96) 1 0.0018 0.61 (0.54 0.75) SWAP 1.08 (0.54 2.27) 0.61 (0.32 1.04) 1.07 (0.54 2.27) 0.59 (0.32 0.96) 1.09 (0.57 1.96) 1 0.2144 0.70 (0.46 1.04) SWPA 1.06 (0.62 2.36) 0.90 (0.54 1.43) 1.06 (0.72 1.56) 0.91 (0.54 1.43) 1 0.0239 1.06 (0.62 2.36) 1 0.7524 0.88 (0.67 1.32) SWBL 0.31 (0.22 0.52) 0.25 (0.18 0.34) 0.31 (0.23 0.52) 0.24 (0.18 0.33) 0.31 (0.22 0.44) 1 0.1655 0.27 (0.22 0.34) APPA 1.01 (0.40 2.65) 1.53 (0.82 2.8 8) 1.02 (0.47 1.45) 1.58 (0.85 2.88) 1.01 (0.40 2.65) 1 0.1157 1.27 (0.82 1.91) APBL 0.29 (0.13 0.53) 0.42 (0.27 0.75) 0.30 (0.14 0.52) 0.42 (0.27 0.75) 1 0.0002 0.29 (0.13 0.53) 1 0.0134 0.39 (0.27 0.55) PABL 0.29 (0.14 0.56) 0.28 (0. 18 0.38) 0.29 (0.21 0.42) 0.27 (0.18 0.38) 0.29 (0.14 0.56) 1 0.3326 0.30 (0.21 0.38) a 0.0003
58 Table 2 3. Results of the 45 ratio measurements for juveniles. Means ( X ), the min max range, and P values for juvenile between species co mparisons are provided. The second set of columns provides P values for differences within species and between the three adult groups W here P <0.0001 for all three adult group compariso ns, only one P value is presented. Between Species Within Species vs Adult Size Group P. canaliculata P. maculata P. canaliculata P. maculata X (Range) X (Range) P value Overall Small Large Overall Small Large TLTW 1.07 (0.97 1.21) 1.08 (0.93 1.18) 0.0139 0.0010 0.0046 0.0007 < 0.0001 TLVH 1.36 (1.18 1.51) 1.43 (1.19 1.68) < 0.0001 < 0.0001 0.0128 0.1480 0.0006 TLAL 1.36 (1.18 1.52) 1.31 (1.18 1.46) < 0.0001 0.1309 0.0747 0.2569 0.3700 0.0046 0.9876 TLAW 2.09 (1.77 2.55) 2.03 (1.68 2.44) 0.0108 0.3757 0.2929 0.5392 < 0.0001 0.0120 < 0.0001 TLSH 2.12 (1.64 4.21) 1.72 (1.30 1.90) < 0.0001 < 0.0001 0.0 116 < 0.0001 < 0.0001 TLSW 3.67 (2.90 4.66) 4.14 ( 2.93 4.95) < 0.0001 0.3653 0.0841 0.9543 < 0.0001 0.3464 < 0.0001 TLAP 3.28 (1.51 8.77) 2.79 (1.36 5.60) 0.3151 < 0.0001 0.8740 0.0003 0.9389 TLPA 3.48 (2.49 5.16) 3.54 (2.67 4.97) 0.2610 < 0.0001 < 0.0001 0.0573 < 0.0001 TLBL 1.11 (0.97 1.23) 1.12 (1.03 1.19) 0.5118 0.0027 0.0005 0.0285 < 0.0001 0.0044 < 0.0001 TWVH 1.27 (1.05 1.42) 1.33 (1.20 1.58) < 0.0001 < 0.0001 < 0.0001 TWAL 1.27 (1.05 1.42) 1.21 (1.09 1.44) < 0.0001 0 .0004 0.0022 0.0002 < 0.0001 TWAW 1.97 (1.66 2.42) 1.88 (1.67 2.42) 0.0005 0.3030 0.6147 0.1418 0.5983 < 0.0001 0.8476 TWSH 1.99 (1.40 4.03) 1.60 (1.31 1.88) < 0.0001 0.0278 0.2804 0.0014 < 0.0001 TWSW 3.45 (2.77 4.12) 3.77 (2.96 4.72) < 0.0 001 0.0268 0.0032 0.2307 < 0.0001
59 Table 2 3 continued Between Species Within Species vs. Adult Size Group P. canaliculata P. maculata P. canaliculata P. maculata X (Range) X (Range) P value Overall Small Large Overall Sma ll Large TWAP 3.11 (1.37 8.63) 2.59 (1.35 5.41) 0.3616 < 0.0001 < 0.0001 TWPA 3.27 (2.23 4.76) 3.28 (2.41 4.86) 0.8342 < 0.0001 < 0.0001 TWBL 1.05 (0.93 1.15) 1.03 (0.93 1.16) 0.0340 < 0.0001 < 0.0001 VHAL 0.99 (0.86 1.15) 0.91 (0.7 2 1.08) < 0.0001 < 0.0001 < 0.0001 0.0004 0.3883 0.0001 0.0306 VHAW 1.53 (1.26 1.84) 1.42 (1.14 1.65) < 0.0001 < 0.0001 < 0.0001 VHSH 1.56 (1.11 3.14) 1.20 (0.96 1.45) < 0.0001 < 0.0001 < 0.0001 VHSW 2.69 (2.16 3.25) 2.83 (2.28 3.41) < 0.0 001 0.0001 0.0018 < 0.0001 < 0.0001 0.7893 < 0.0001 VHAP 2.41 (1.06 6.60) 1.95 (0.93 4.04) 0.8758 < 0.0001 0.5140 0.0013 0.5442 VHPA 2.56 (1.65 3.70) 2.47 (1.89 3.84) 0.0446 < 0.0001 < 0.0001 0.3044 < 0.0001 VHBL 0.82 (0.72 0.91) 0.78 (0.63 0.90) < 0.0001 0.0007 0.0016 0.0011 0.0007 0.0030 0.0029 ALAW 1.54 (1.26 1.88) 1.55 (1.39 1.83) 0.4324 0.3088 0.1747 0.5389 < 0.0001 0.9731 < 0.0001 ALSH 1.57 (1.15 3.25) 1.32 (1.07 1.52) 0.3128 0.0008 0.0220 < 0.0001 < 0.0001 ALSW 2.72 (2.09 3.51) 3.11 (2.39 3.88) < 0.0001 0.5833 0.2966 0.9229 < 0.0001 0.0363 < 0.0001 ALAP 2.40 (1.10 6.26) 2.13 (1.03 4.30) 0.1516 < 0.0001 0.6691 < 0.0001 0.9110 ALPA 2.56 (1.80 3.94) 2.69 (2.12 3.66) 0.0013 < 0.0001 < 0.0001 0.0030 < 0.0001 ALBL 0. 82 (0.72 0.89) 0.85 (0.77 0.91) < 0.0001 0.4355 0.4777 0.3995 < 0.0001 0.0015 < 0.0001
60 Table 2 3 continued Between Species Within Species vs. Adult Size Group P. canaliculata P. maculata P. canaliculata P. maculata X (Range) X (Range) P value Overall Small Large Overall Small Large AWSH 1.03 (0.68 2.24) 0.85 (0.70 0.92) 0.3779 0.1395 0.6958 0.0130 < 0.0001 AWSW 1.77 (1.37 2.27) 2.01 (1.56 2.62) < 0.0001 0.3004 0.0540 0.9902 < 0.0001 0.0131 < 0.0001 AWAP 1.53 (0.76 4.13) 1.36 (0.69 2.64) 0.1692 < 0.0001 < 0.0001 AWPA 1.67 (1.17 2.71) 1.74 (1.28 2.46) 0.0161 < 0.0001 < 0.0001 0.0073 < 0.0001 AWBL 0.54 (0.44 0.61) 0.55 (0.45 0.62) 0.0005 0.0012 0.0001 0.0223 < 0.0001 0.49 78 < 0.0001 SHSW 1.89 (0.77 2.62) 2.36 (1.83 3.03) < 0.0001 0.0003 0.0012 0.0002 0.2817 0.0008 0.1245 SHAP 1.75 (0.39 4.82) 1.61 (0.77 3.01) 0.1016 < 0.0001 0.0210 0.0876 0.0107 SHPA 1.77 (0.74 2.61) 2.05 (1.59 2.89) < 0.0001 < 0.0001 0.0008 0.1566 < 0.0001 SHBL 0.57 (0.24 0.69) 0.65 (0.57 0.80) < 0.0001 < 0.0001 < 0.0001 SWAP 0.90 (0.39 2.56) 0.69 (0.34 1.40) 0.3705 < 0.0001 0.0439 0.0012 0.0105 SWPA 0.95 (0.61 1.44) 0.87 (0.63 1.45) < 0.0001 < 0.0001 0.0033 0.1732 0.0006 SWB L 0.30 (0.23 0.38) 0.27 (0.22 0.35) < 0.0001 0.1314 0.2939 0.0673 < 0.0001 0.1521 < 0.0001 APPA 1.35 (0.38 2.60) 1.36 (0.60 2.72) 0.9624 < 0.0001 < 0.0001 0.1689 < 0.0001 APBL 0.43 (0.12 0.74) 0.43 (0.20 0.81) 0.2795 < 0.0001 0.1297 < 0.0001 0.1437 PABL 0.33 (0.21 0.45) 0.32 (0.22 0.40) 0.2595 < 0.0001 < 0.0001 0.0169 < 0.0001
61 Table 2 4. Results of the initial 45 ratio model and the reduced follow up model for the canonical discriminant function analysis for both species and all size groups. Data for all canonical variables (CV) is provided as well as the class means for the first three CV. Canonical correlation value (CC); squared canonical correlation (R 2 ); eigenvalue (E); likelihood ratio (LR); F values (F), and degrees freedom (DF) CV 1 2 3 4 5 45 Ratio CC 0.89445 0.71522 0.53786 0.44989 0.29945 R 2 0.80005 0.51154 0.28929 0.20240 0.08967 E 4.0014 1.0473 0.4071 0.2538 0.0985 LR 0.05039 0.25204 0.51601 0.72606 0.91032 F 27.08 13.92 8.55 6.16 3.58 DF Num, Den 225, 7403.7 176, 5937.5 129, 4462.6 84, 2980 41, 1491 Class Means 1 2.5 9 0.14 0.1 7 2 0.6 3 0.3 7 1.07 3 0.7 3 2.20 1.6 2 4 1.6 6 0.6 1 0.13 5 1.8 3 0.6 9 0.1 9 6 1.6 5 2.86 1.50 Reduced CC 0.89256 0.71260 0.52607 0.44163 0.28280 R 2 0.79667 0.50780 0.27675 0.19504 0.07997 E 3.9183 1.0317 0.3827 0.2423 0.0869 LR 0.05360 0.26363 0.53561 0.74058 0.92002 F 37.43 19.18 11.59 8.40 4.67 DF Num, Den 160, 7432.4 124, 5971. 8 90, 4495.8 58, 3006 28, 1504 Class Means 1 2.5 6 0.1 5 0.1 7 2 0.60 0.36 1.05 3 0.66 2.1 6 1.58 4 1.6 5 0. 60 0.12 5 1.81 0.6 8 0.18 6 1.62 2.8 7 1.4 4
62 Figure 2 2. Results of the canonical discriminant function analysis: A) Preliminary 45 ratio mod el B) Follow up reduced model. Species classifications are: 1 = large P. maculata 2 = small P. maculata 3 = juvenile P. maculata 4 = large P. canaliculata 5 = small P. canaliculata and 6 = juv enile P. canaliculata A B
63 Figure 2 3. Re sults of the combined adult size class canonical discriminant function analysis: A) 45 ratio model combined size class B) Reduced model combined size class. 1 = P. maculata and 4 = P. canaliculata B A
64 Figure 2 4 Results of the juvenile, small adult, and large adult canonical discriminant function analyses: A) 45 ratio model small adults B) Reduced model small adults C) 45 ratio model large adults D) Reduced model large adults E) 45 ratio model juveniles F) Reduced model juveniles. 1 3 are P. maculata and 4 6 are P. canaliculata A B C D E F
65 Figure 2 5. Results of the within species canonical discriminant function analyses using the 45 ratio model: A) Pomacea maculata B) Pomacea canaliculata 1 = large P. maculata 2 = small P. maculata 3 = juvenile P. maculata 4 = large P. canaliculata 5 = small P. canaliculata and 6 = juvenile P. canaliculata A B
66 Figure 2 6. Results of the preliminary cluster analysis. Cluster 1 is dominated by adult P. maculata (species 1 and 2); cluster 2 is dominated by adult P. canaliculata (species 4 and 5); and cluster 3 is dominated by juveniles of both species (species 3 and 6). 3 1
67 CHAPTER 3 PHYSIOLOGICAL TOLE R ANCES Background N ew W orld apple snails are native to South and Central America with one species, Pomacea paludosa (Say 1829), native to North America. Pomacea have been introduced into many countries and have subsequently become established in natural and artificial aquatic systems. Introduction s have oc curred for a variety of reasons which include bio control, the aquarium trade, and the escargot industry (Halwart 1994; Naylor 1996). In the last decade Pomacea have received much attention because the foraging capabilities of some species have devastated crops (Yusa and Wada 1999; Joshi 2001; Martin 2004) and because they are considered a threat to the functioning of natural wetlands (Carlsson et al. 2004). Two of the most commonly introduced species implicated in crop damage and ecosystem alteration are Pomacea canaliculata al. 2000), and Pomacea maculata (Perry 1810) ( Carlsson et al. 2004 ; Rawlings et al. 2007 ; Hayes et al. 2009, 2012 ). In the U.S. Pomacea spp. continue to spread with populations in at least 10 states (Figure 1 1). Compared to other regions of the U.S. introductions of Pomacea spp. (typically P. maculata ) has been greatest in the Southeast Florida has four nonindigenous Pomacea species : P. canaliculat a P. maculata P. diffusa (Blume 1957), and P omacea sp. (previously described as P. haustrum Reeve 1856) (Figure 1 2) ( Rawlings et al. 2007 ; Hayes et al. 2009, 2012 ). The most prevalent is P. maculata with populations ranging from Miami to Eglin Air For ce Base in the panhandle. Although P. canaliculata is believed to be limited to northeast Florida, the ir continued spread through this region
68 suggests that populations may continue to establish through the S outheast ( Bernatis 2013 ). There are isolated reco rds in north and south Florida for Pomacea diffusa whereas Pomacea sp. is believed to be limited to the Loxahatchee National Wildlife Refuge. Although impacts to large natural lakes and rivers have not been documented quantitatively, a plethora of observa tions have implicated the snails in extensive urban aquatic system damage. This damage has involved the consumption and depletion of aquatic vegetation that was planted for restoration and enhancement. In Florida, the sales value of aquatic vegetation rout inely exceeds $ 13 million annually ( United S tates Department of Agriculture 2006 ), making the impacts of P. maculata and P. canaliculata an economic concern. Despite their abundance and broad distribution in the Southeast, there is little basic biological research on nonindigenous Pomacea particularly in te rms of physiological tolerances in their introduced ranges (Ramakrishnan 2007) Most research has focused on eradication, feeding, reproduction, and genetics ( Lach et al. 2000; Yusa et al. 2006; Hayes e t al. 2009 2012 ; Burks et al. 2011). Limited research has been conducted on phy siological tolerances in the ir native range, with an emphasis on desiccation and temperature effects ( Albrecht et al. 1999; Martin et al. 2001; Estebenet and Martin 2002; Marti n and Estebenet 2002; Seuffert and Martin 20 09a, 2009b ; Seuffert et al. 2010 2012 ) Th e paucity of research is surprising given that key elements of invasion success are the biological limitations in r elation to a novel environment. B yer s et al. ( 2013) a ttempted to model future distribution of nonindigenous Pomacea in the southeast U.S. with an emphasis on pH impacts Unfortunately
69 reliance on limited, third party database s to pr edict snail distribution limits was a significant shortcoming of their appr oach Many behavioral and biological factors are key in the distribution and establishment success of nonindigenous species. One such factor is the frequency of introduction and number of individuals introduced (i.e., propagule pressure; Marchetti et al. 2004; Lockwood et al. 2007). In Florida, reports of increased nonindigenous apple snail populations occurred after the hurricanes in 2004, suggesting a natural spread. T here are also cases of the snails being transported by humans between locations for var ious reasons (e.g. for bait, to feed birds), and there is circumstantial evidence of aquaria release ( personal observation). Depending on the m otive for transport and release introductions are likely to occur into systems that are ideally suited for the snails. (i.e., jump dispersal) into new locations that are suitable habitat, and the succ ess of establishment, increase threat risk s to some aquatic systems. Several abiotic factors are known to influence mollusc distribution including : oxygen, temperature, pH, salinity, Ca + food availability and desiccation events ( Prosser and Health 1991; Somero 1995; Martin et al. 2001; Estebenet and Martin 2002; Martin and Estebenet 2002; Ansart and Vernon 2003 ). Numerous studies are publ ished on the impacts of calcium carbonate (CaCO 3 ) availability and mollusc success, as defined by shell formation, regeneration, and growth ( Wilbur 1964 ; Vanderbo and Vanpuymb 1966; Young 1975; Nduku and Harrison 1976; Lodge et al. 1987; Brown 2001 ; White 2007 ). Karst aquatic environments provide the CaCO 3 required by snails to produce shells and eggshells. In the U.S. about 20% of the land surface is classified as karst and a
70 substantial portion of waterbodies in the S outheast exceed (Scott et al. 2002; US GS 2013) the limiting threshold of 14.8 mg CaCO 3 /L for Pomacea rep orted by Martin et al. (2001). Therefore, spread and establishment of Pomacea throughout the Southeast U.S. and other karstic regions are likely to be minimally influenced by low availabilit y of CaCO 3 Pomacea are amphibious and possess both a lung and gills T herefore, dissolved oxygen levels are not as critical for survival as are other abiotic factors ( Aldridge 1983; Ito 2002; Seuffert and Martin 200 9b ). Several studies have looked at the influence of oxygen during different life stages of Pomacea spp. in particular during aestivation or aerial exposure. In general Pomacea spp. are capable of obtaining oxygen either aerially or from the water and may rely on aerial respiration to compens ate for less than adequate dissolved oxygen concentrations ( Burky et al. 1972 ; Freiberg and Hazelwood 1977; Santos and Mendes 1981; S euffert and Martin 2009a, 2009b ). Temperature regimes are also important in the distribution, growth and survival of anim als (Prosser and Heath 1991). Numerous studies on Pomacea t hermal tolerances have been conduct e d ( Burky et al. 1972; Albrecht et al. 1999; Martin et al. 2001; Ramakrish nan 2007; Matsukura et al. 2008, 2009; Seuffert et al. 2010; and many others). Although environment at time of exposure, Pomacea spp. are capable of surviving temperatures ranging from 0 40 C for at least short periods (1 3 days) (Burky et al. 1972 ; Matsukura et al. 2009 ). In outdoor mesocosms, I observed snails foraging under a thin layer of ice
71 when diel aerial temperatures ranged from 2 to 11 C for 10 days. Similarly, snails were active in summer diel water temperatures ranging from 27 to 40 C for 14 days The objecti ve of my study was to obtain biological information for P. canaliculata and P. maculata t hrou ghout their life history stages by using laboratory survival trials and examining growth patterns under several environmental variables that may limit invasion suc cess. These included s tarvation, s alinity, pH, and desiccation and were tested at levels represent ative of the ambient conditions in and where these species are established. The results provide data for development of distribut ion models and may be helpful for developing species specific control measures. Finally, this information will provide heretofore missing data on the basic biology of the se snails. Methods Collection and Holding I collected Pomacea maculata by hand from no rth Florida (Leon County) and central FL (Polk County). At the beginning of this project only one population of P. canaliculata was known to exist in FL (Duval County) and that is where I collected individuals for my experiment. Species identifications w ere confirmed by Dr. Timothy Collins (Florida International University) and Rawlings et al. (2007). Experiments were conducted in an indoor facility from July 2006 to June 2007. I used 113 L tanks with a 50:50 mix of pond and well water to house snails p rior to using them in experiments I maintained t ank s under the following conditions: temperature 23 25 C; dissolved oxygen 80 90% ( via air stones ); pH 7.5 8.0; ammonia, nitrate, and nitrite were within normal ranges for aquacultur e (Swann 1993) Water wa s continuously filtered through mechanical and chemical media. Snails were fed ad libitum
72 ( Hydrilla verticillata Hydrocotyle sp., cantaloupe and pears ) every 48 72 hours Snails were kept in holding tanks for no more than two weeks prior to trials. The sm all size of hatchlings and juveniles required for testing precluded field collection, therefore these age groups were reared in the laboratory Egg masses deposited by captive adults during the first week of captivity were hatched and I used these hatchli ngs in my trials. Individual snails were used in only one experimental procedure. General Procedures As described below, I tested snail survivorship using four treatments: starvation, pH, salinity and desiccation tolerance. I used t hree age groups in ea ch treatment : hatchlings ( < 4 days old and < 3.0 mm total length), juveniles (10 25 mm total length), and adults ( 30 mm total length). The total length (TL) and mass (M) of each snail was recorded at the start of the trial. All experiments were conducted to determine the lethal level for 50% of the snails (L C 50) with in a 28 day trial. Snails were randomly selected a nd randomly distributed into treatment tanks. Experimental tanks varied in volume for each snail size class: adult tanks were 79.7 L, juvenile tanks were 39.9 L, and hatchling tanks were 9.5 L. All variables had multiple levels (e.g., salinity had five tre atment levels). Therefore, the e xperimental design was a replic ated, complete randomized block; for example, a replicate adult P. canaliculata salinity block included one tank for each treatment level. A dult treatments had five replicated blocks, whereas j uveniles and hatchlings had either five or six replicated blocks. Treatments followed ASTM Standards E 729 and E 1022 (American Society for Testing and Materials 2002), utilizing a continuous mechanical/chemical filtration system. Mortality was checked dai ly and was determined by examining the operculum for resist ance and visibility of tissue. When tissue was visible and the snail was not moving, I further stimulated the foot for
73 retraction response. If foot retraction did not occur, the animal was classifi ed as deceased and removed from the test tank. There are known metabolic effects of feeding and digestion in animals and as a result animals are often starved throughout trials ( Berezina 2001 ; Wang et al. 2001 ) Therefore, I did not feed the snails in this study. Starvation Starvation trials were conducted on adults and juveniles. Snails were either starved or fed ad lib the same diet as in the holding tanks. The total numbers of snails used varied by age and species, but were distributed evenly throughout the tanks. There were 60 adult P. canaliculata per treatment (N = 12/t ank), with starting mean TL = 42.57 mm (30. 12 59.36 mm ) and starting mean M = 18.41 g ( 6.01 45.77 g ) There were 48 P. maculata per treatment (N = 8/ta nk) with starting means of TL = 70. 27 mm ( 51.24 90.03 mm ) and M = 114.79 g ( 66.9 180.2 g ) There were 40 juvenile P. canaliculata per treatment (N = 8/tank) and 35 P. maculata per treatment (N = 7/tank). R espectively, the initial means were: TL = 22.25 mm ( 15.82 24.91 mm) and 19.11 ( 13.90 2 3.94 mm) and, M = 3.28 g ( 1.63 4.38 g) and 2.82 g (1. 32 4.73 g) Salinity Salinity tolerance trials were conducted over a range of 0 32 ppt at 8 ppt intervals. Artificial seawater (Instant Ocean, Blacksburg, VA) was used and salinities were maintained wit h in 2 ppt. At the outset of the experiment, all tank salinity levels were at 0 ppt and were increased at a rate of 1 ppt / 30 minutes until the desired salinity level was reached, a rate slightly sl ower than high tide intrusion. This was accomplished throu gh equal volume removal and replacement with higher salinity water ; water changes were also made in the 0 ppt condition. Thirty snails were used for all trials (N = 6/ta nk). The initial means for adult P. canaliculata and P. maculata respectively, were:
74 T L = 43.87 mm (30.05 63.75 mm ) and 64.17 mm ( 32.99 0 83.96 mm ) and M 18 .99 g (5.20 52.40 g ) and 94.54 g ( 5.40 170.61 g ) The initial means for juvenile P. canaliculata and P. maculata, respectively, were: TL = 16.17 mm (10.37 23.89 mm ) and 20.78 mm (16.17 2 4.91 mm ) and M = 2.03 g (0.92 3.49 g ) and 2.54 g (1.41 3.57 g ) For hatchl ings the initial TL mean for P. canaliculata was 2.2 mm (2.0 2.5 mm ); and for P. maculata was 2.4 mm (1.6 2.8 mm ). pH Trials investigating pH tolerance were conducted at a range o f 5.5 9. 5 at intervals of 1.0 pH unit. The pH levels were maintained at 0.1 pH unit s Commercially available, mollusc safe, pH adjusters (pH Up and pH Down, Aquarium Pharmaceuticals, Pennsylvania, USA) were used to regulate pH levels. Test conditions w ere established in the tanks prior to the introduction of the snails. Thirty adults of each species were used per treatment level (N = 6/tank) The initial means for adult P. canaliculata and P. maculata, respectively, were: TL = 41.73 mm (31.17 55.80 mm) and 65.80 mm (35.50 88.94 mm) and, M = 15.59 g (4.08 40.81 g) and 103.77 g (9.64 169.10 g) Thirty six j uvenile s and hatchling s of each species were used per treatment level (N = 6/tank) The initial means for juvenile P. canaliculata and P. maculata, resp ectively, were: TL = 17 .41 mm (11.39 24.48 mm) and 17.78 mm (10.35 26.45 mm) and M = 2.26 g (0.95 4.16 g) and 2.79 g (0.93 5.04 g) For hatchli ngs the initial TL mean for P. canaliculata was 2.2 mm (1.9 2.5 mm), and for P. maculata was 2.5 mm (1.8 2.9 mm) Desiccation Desiccation tolerance trials were conducted under two test conditions with an aquatic envir onment serving as the control; snails in the control group were not fed. Test conditions were either desiccation, defined as < 60% relative humidity and dry sand
75 substrate, or semi desiccation defined as > 80% relative humidity and moist sand Allen, LA). Desiccation tanks had a 5 cm layer of dry sand in the bottom and a desiccant pouch was used if humidity levels exceeded 60%. Semi desiccation treatments had a 5 cm layer of sand that was kept wet with up to a 1.6 mm of standing water. The control tanks had the same water quality conditions as the holding tanks. M ortali ty was easier to detect in this trial (i.e., odor ). Snails with closed opercul a were tested by gently prying the operculum. If there was resistance, the snail was considered to be alive; if the operculum was easily movable and soft tissue was readily expel led, the snail was deceased. There were 45 adult P. canaliculata in the wet and dry conditions (N = 9/tank) and 40 snails in the control (n = 8/tank) ; adult P. maculata had 40 snails in each condition (N = 8/tank) The initial means for adult P. canalicula ta and P. maculata respectively, were: TL = 46.86 mm (35.89 64 mm) and 66.64 mm (30.04 87.61 mm) and M = 25.84 g (12.66 48.52 g) and 99.69 g (4.90 179.40 g) Both species of juveniles had 60 snails per treatment The initial means for juvenile P. canalic ulata and P. maculata respectively, were: TL = 17.36 mm (10.03 28.19 mm) and 19.07 mm (10.20 27.18 mm) and M = 3.28 g ( 1.63 4.38 g) and 2. 91 g (0.93 3.95 g) Both species of h atchling s had 40 snails per treatment. The initial TL mean for P. canaliculata was 1.8 mm (1.7 1.9 mm) and for P. maculata was 1.5 mm (1.4 1.6 mm). After 28 days, I continued the study in an effort to ascertain survival time under wet and dry conditions. Minor changes in methodology were implemented beginning with week five, to min imize snail disturbance. Snails continued to be checked daily for mortality (i.e. tissue expulsion). E very two weeks, one of the snails assumed to be
76 alive was placed in a tank for revival. If the snail did not respond, it was classified as dead and th e time of death was recorded as seven days prior (i.e., halfway between the revival episodes). This process continued until a live snail was found, or until no more live snails were present Live snails were not returned to test conditions reducing the to tal number of snails available every two weeks Statistical Analyses Summary statistics were calculated for all data. ANOVAs were used to ensure tank parameters did not differ throughout the course of the study. Mortality was evaluated as time to event da ta using standard survival analysis methods (Daniel and Cross 2013; Allison 2010, Kleinbaum and Klein 2005) Median survival time was estimated directly from the survivorship curve using the Kaplan Meier procedure Survival curves were compared using Ge o n t est; follow up pairwise comparisons controlling for Type I error were done in the event that a significant difference was found among curves. Between species factor comparisons were calculated using the Mantel Cox log rank tes t. In trials where survival exceed 50% for the 28 days, the actual survival rates are provided. All tests were performed in either SPSS 17.0 (SPSS Inc., Chicago, Illinois) or SAS 9.2 (SAS Institute, Cary, North Carolina). Results Starvation Both species and age groups had a median survival time of 28 days for t he starved and fed treatment s Survival of adult P. maculata was not significantly different between treatments (Table 3 1 ), with a 96% survival rate in each. Similarly, the fed adult P. canaliculat a had a 95% survival rate. But for P. canaliculata there was a significant
77 difference from the 65% survival rate of the starved snails ( G W = 16.65, P <0 .001; Table 3 1 ). Between adult species and within treatment there was not a significant difference fo r either condition (Fed: X 2 = 0.33, P = 0.569 and Starved: X 2 = 0.22, P = 0.636). Juveniles of both species had a median survival of 28 days for both conditions Juvenile P. canaliculata had 87% survival when fed and 72% survival when starved. Juvenile P. maculata had 83% survival when fed and 68% survival when starved. Between treatment survival within a species was not si gnificantly different (Table 3 1 ). However, whereas the fed treatment was not different between species, the starved treatment was signi ficantly different ( 2 = 15.11 P < 0 .001). Salinity Adult median survival varied between treatments, with lower salinities having higher survival (Figure 3 1). In the global analyses of treatments within species, there were significant differences ( P. canaliculata : G W* = 13. 03 P <0.001; P. maculata : G W = 123.4, P < 0.001 ; Table 3 1). For both species, t here were no significant differences for 0 ppt vs 8 ppt or 24 ppt vs 32 ppt but all were different from 16 ppt In 0 ppt, 100% of both species survived the entire 28 days. I n 8 ppt, 93.3% of the P. canaliculata and 86.7% of the P. maculata survived for 28 days. At 16 ppt median survival time was 5.7 days for P. canaliculata and 3.6 days for P. maculata All P. canaliculata were deceased at 17 days and all P. maculata were d eceased at nine days. For both species, 100% mortality occurred by day three in the 24 and 32 ppt conditions. Survival at 16 ppt, however, was significantly different from the other four salinities. Between species and within treatment, median survival was not significantly different for 0 ppt, 8 ppt, 24 ppt, or 32 ppt. O nly at 16 ppt was survival significantly different between species 2 = 3.89, P = 0.049 )
78 Juvenile median survival varied between treatments, with lower salinities having higher survival (Figure 3 1). In the global analyses of treatments within species all survival curves differed ( P. canaliculata : G W = 134.8, P <0.001; P. maculata : G W = 129.0 P <0.001; Table 3 1 ). In 0 ppt both species had 100% survivorship for 28 days The median survival at 8 ppt for P. canaliculata (56.7%) was 28 days and for P. maculata only 19.6 days Median survival at 16 ppt for P. canaliculata was four days and for P. maculata, three days Both species had a 100% mortality rate at three days for 24 ppt and two days for 32 ppt. Within treatment there were no significant differences in survival b etween species 2 = 10.78, P <0.001). Hatchling median survival varied between treatments, with lower salinities having higher survival (Figure 3 1). In the global analyses of treatments within species all survival curves differed ( P. canaliculata : G W = 140.2, P <0.001; P. maculata : G W = 140.0, P <0.001 ; Table 3 1 ) There were, however, no significant differences between species at any level. Although m edian survival at 0 ppt for both species was 28 days only 53% of P. maculata and 73% of P. canalic ulata were alive on day 28. At 8 ppt, median survival for both species was 3.4 days The remaining three treatment levels had a median survival of 1.5 days. pH Adult median survival was 28 days for both species and all treatments (Figure 3 2). In the globa l analyses of treatments within species, m edian survival was not significantly different ( P. canaliculata : G W = 7.3, P = 0.120; P. maculata : G W = 8.8, P = 0.067 for P. maculata ; Table 3 1 ). Survival rates of P. canaliculata and P. maculata respectively were: 93.3% and 90.0% at pH = 5.5; 93.3% and 96.6% at pH = 6.5; 100%
79 (both species) at pH = 7.5 and 8.5 ; and 86.7% and 100% at pH = 9.5 Between species, only pH 9.5 was significantly different ( 2 = 4.21, P = 0.040) Juvenile median survival was 28 days for both species and all treatments (Figure 3 2). In the global analyses of treatments w ithin species P. maculata were not significantly d ifferent (G W = 5.9, P = 0.253), h owever, P. canalicul ata were significantly d ifferent (G W = 13.4, P = 0.010; Table 3 1 ). Survival rates of P. canaliculata and P. maculata respectively were: 58.3% and 86.7% at pH = 5.5; 55.5% and 86.7% at pH = 6.5; 88.8 % and 77.7% at pH = 7.5; 88.8% and 75% at pH = 8.5; an d 66.6 % and 86.7 % at pH = 9.5 Between species, survival estimates were significantly different at pH of 5.5 ( 2 =9.00, P = 0.003 ), 6.5 ( 2 = 10.62, P = 0.001 ), and 9.5 ( 2 = 5.66, P = 0.017 ); in each treatment P. maculata had the greater survival. Hatchling median survival varied between treatments and species, with survival generally increasing as pH increase d (Figure 3 2). In the global analyses of treatments within species, survival curves of both species were significantly different ( P. canaliculata G W = 64.3, P < 0 .001 ; P. maculata G W 72.3, P < 0 .001 ; Table 3 1 ) For both species, the lowest median surviv al time, and the only significantly different curve between species, was at pH 5.5 ( P. canaliculata : 3.1 days and P. maculata : 3.0 days; 2 = 5.70, P = 0.017 ). Survival increased as pH increased, reaching 10.3 days for P. canaliculata at pH 8.5 and 10 days for P. maculata at pH 9.5. With the exception of pH 9.5, median survival was greater for P. canaliculata but t he maximum survival for a ll snails regardless of treatment, was 14 days Desiccation Adult median survival was 28 days for both species and all treatments (Figure 3 3). In the global analyses of treatments within species, there w ere no significant
80 difference s in survival curves f or P. maculata ( G W = 5.5, P = 0.062 ; Table 3 1). Although median survival for all treatments was 28 days for P. canaliculata there was a significant difference in that the wet and dry were equal, but both differed from the control ( G W = 10.7 P = 0. 0058 ). Between species and within treatment there were no significantly different median survival times. The 28 day s urvival rates of P. canaliculata and P. maculata respectively, were: control = 82.5% and 80%; wet = 53.3% and 80.0%; and dry = 55.5% and 57.5% At the end of one year snails of both species in the wet conditions were alive whereas in dry conditions all P. maculata were dead at 153 days and all P. canaliculata at 154 days. Juvenile median survival was 28 days for both species and all treatment s (Figure 3 3). In the global analyses of treatments within species, survival curves were significantly different ( P. canaliculata : G W = 15.8, P < 0 .001; P. maculata : G W = 115.9, P < .001 ; Table 3 1 ). In both cases the control and wet treatments were equi valent and differed from the dry treatment, the latter having the shorter survival times. The 28 day s urvival rates of P. canaliculata and P. maculata respectively, were: control = 78.0% and 77.0%; wet = 92.0% and 83.0%; and dry = 60.0%, with all P. macul ata dead by day 14. Between species only the dry treatment was significantly different ( 2 = 76.40, P <0.001). In the extended trial, P canaliculata survived until day 147 and 98 in wet a nd dry conditions, respectively, and P. maculata survived in wet co nditions until day 133. Hatchling median survival varied between treatments and species, with survival greatest in the control and lowest in the dry condition (Figure 3 3). In the global analyses of treatments within species, a ll curves differed for P. can aliculata (G W = 50.02, P <0.001; Table 3 1 ). However, for P. maculata the control and wet treatments
81 were equal and both differed from dry (G W = 57.2, P <0.001). Between species, the significant differences were 2 2 = 26.38, P <0.001) treatments. The 28 day survival rates of P. canaliculata and P. maculata respectively, were: control = 80.0% and 45.0% (although 55% on day 27); wet = 17.0% (47% on day 27) and 27.0% (6 0.0% on day 27); and dry = 0.5%, with all P. maculata dead by day 10 In the extended wet trial, P. canaliculata survived to day 49 and P. maculata until day 42. In the dry treatment P. canaliculata survived to day 42. Discussion The goal of this investiga tion was to test snail survival thresholds under different conditions of starvation, salinity, pH, and desiccation. Such data can be used to develop predictive models for the likeli hood of establishment, provide data for species specific management protoco ls and contribute to the knowledge base on invasive Pomacea Survival under each of these variables was linked to the age of the snail. Older snails had greater survivorship. In each case, mortality was greatest and most rapid for hatchlings. This was ant icipated, as many organisms are most vulnerable at a young age. The fact that hatchling snails survived as long as they did in some conditions indicates how physiologically robust the y are for invasion success In most of the treatments, adults and juvenil es had 28 day survival rates, although the overall proportion of juvenile survival was slightly lower than that for adults. Pomacea maculata and P. canaliculata consume a wide variety of macrophytes and given their voracious appetites removing food shoul d have an impact on survival. Several authors have shown that quality of food is important to growth and reproduction of Pomacea ( Burks et al. 2011; Boland et al. 2008; Carlsson and Br nmark 2006; Carlsson et al. 2004; Tanaka et al. 1999), two important fa ctors for establishment
82 success. R esults of my study however, indicate adults and juveniles can survive without feeding for at least 28 days. In this trial conditions were highly controlled, restricting the ability of snails to feed on anything (e.g., pe riphyton) but this is not a common condition in nature. Pomacea which allows them to feed on minute particles in the surface film of the water by using their foot as a funnel (Saveanu and Martin 2013; McClary 19 64; personal observation). Although there is no evidence to suggest they can obtain all of their energy needs from surface film feeding, it reduces the likelihood of starvation and potentially allow s snails to survive in low food quality habitats. I have o bserved i n locations where all vegetation appears to have been extirpated and snails persist, populations, egg deposition s, and hatch rates are reduced The results of my study demonstrate the snails are capable of surviving prolonged periods without prefe rred food and that starvation is likely not an effective means of management. Restricting or removing vegetation from a system may help to reduce nonindigenous snail populations, but depending on the system and the number of snails present, the increase in decaying biomass may prove detrimental for other organisms. Another factor that influenc es the distribution of aquatic organisms is salinity (Remane and Schlieper, 1971) and it can be of particular importance for freshwater organisms exposed to estuarine conditions (Costil et al. 2001). In coastal environments periodic changes in salinity are normal although frequency and intensity can vary (e.g., tidal cycles or flooding) C hanges in salinity can cause osmotic stress in organisms resulting in a variet y of behavioral and physiological responses (Hart et al. 1991; Piscart et al. 2006). However, t his does not mean freshwater gastropods are
83 incapable of tolerating moderately saline conditions as Jacobsen and Forbes (1997) found that 5 ppt is optimal for th e invasive freshwater gastropod, Potamopygrus antipodarum (New Zealand mudsnail). A limited body of information is available on salinity tolerances of apple snails (Cowie 2002). The only Pomacea species thought to have salinity tolerance i s P. bridgesii ( Jordan and Deaton 1999); however these snails may actually have been P. diffusa (Rawlings et al. 2007). Jordan and Deaton (1999) demonstrated that P. bridgesii had an 80% survival rate for two weeks, when exposed to osmolality levels ranging from 20 200 mOsm ( 1 8 ppt); however, when levels reached 400 mOsm ( 12 ppt) 100% mortality occurred within three days. Similar to the findings of Jordan and Deaton (1999), my results indicate that 8 ppt is a threshold for survival. Snails in my study, however, wer e able to withstand higher salinity (16 ppt) for several days. The results of my study indicate that systems only w eakly influenced by tid es (e.g. < 8 ppt) may be acceptable habitat for P. maculata and P. canaliculata This is confirmed in the lower St. Johns River, Duval County, FL where I observed established P. maculata in a tidally influenced location and routinely experience a salinity of 4 ppt T he ability to withstand higher salinity even for the short term, is an important factor when considering future distributions as snails could potentially move between coastal streams via estuaries. The success of the snails in saline systems would then be dependent on other biotic and abiotic factors (e.g. food, predators, temperature). The s uccess and d iversity of aquatic organisms is dependent, in part, on pH ( Bemvenuti et al. 2003; Berezina 2001; Herrman e t al. 1993; Hall et al. 1980). Molluscs typically have low abundance in acidic habitats with pH < 6.0, because they are unable
84 to secrete and maintai n their shells (Berezina 2001 ; Okland 1992; Mackie, 1987; McK illop a nd Harrison 1972). Unfortunately, there have been few controlled studies on the effects of pH on freshwater gastropods (Glass and Darby 2008; Hunter 1989), but there are many in situ studi es that evaluate pH as a factor for distribution ( Karunaratne et al. 2006; Watson and Ormerod 2004; Prescott and Curteanu 2 004; Harvey and McArdle 1986). Studies of pH influences on nonindigenous Pomacea are scarce and typically focused on distribution (By ers et al. 2013; Martin et al. 2001). In North America, the majority of pH research has been limited to distribution impacts on native P. paludosa (Perera et al. 1989; Hurdle 1974; Gleason et al. 1975) but there have been few laboratory studies ( Glass and Darby 2008 ; Ramakrishanan 2007 ). At the time I began my study, there had been no other studies on pH effects on the survival of P. maculata and P. canaliculata Shortly after I finished my study, Ramakrishanan (2007) published results on P. maculata Rama krishanan (2007) findings were comparable to results for both species in my study. Both studies indicate that common pH levels in surface waters (e.g., 6 9) have no impact on nonindigenous Pomacea survival. Sublethal effects (e.g., suppressed growth, r eproduction, shell strength) of pH have been reported for freshwater gastropods (Hunter 1989) A decrease in shell strength increases susceptibility to predation (Stein et al. 1984) or accidental death through cr ushing (Glass and Darby 2008). Decreases in shell strength under acidic conditions have been reported for P. paludosa (Glass and Darby 2008). They observed erosion and brittleness at pH < 6.5 I observed these effects in adults and juveniles of both species in my study but only at a pH of 5.5. The e rosion I observed typically included the spire wearing or breaking off and discoloration (e.g., color pattern
85 disappearance) and brittleness of the outer whorl. The ability of nonindigenous Pomacea to tolerate and grow in a wide range of pH environments i mplies that there are abundant potential habitat s for colonization. Results of these studies provide a possible explanation for why the snails are invasive in one system, but only present in others, as the effects of reduced growth and weakened shells may prevent the populations from becoming a threat in lower pH environments. In aquatic systems that experience periodic desiccation, animals that are unable to retreat to pools of water have evolved several coping techniques. One of the most common response s to aerial exposure by freshwater gastropods is aestivation (Darby et al. 2008; Brown, 2001; Aldridge 1983; McMahon, 1983; Burky et al. 1972; Little, 1968; Meenakshi 1964). Animals that use aestivation/hibernation as a means of surviving until o nditions return, often display a reduc ed metabolic rate, and freshwater gastropods are no different (Guppy and Withers 1999; Aldridge, 1983; McMahon 1983). In my study, a reduction in heart rate, which serves as a proxy for reduced metabolism, was observed through the transparent shells of juvenile and small P. canaliculata adults. At the start of the desiccation trials, there were 64 68 beats per minute (bpm) but by week four that had declined to 2 6 bpm. Burky et al. (1972) also observed a reduction in m etabolic rate in P. urceus during an emersion trial that concluded at 526 days, with snails still alive. Likewise, Pomacea lineata s urvived 13 months out of water (Little 1968). T he native apple snail, P. paludosa however, is not as tolerant to desiccat io n, as adult mortality reached 73% by 18 weeks (Darby et al. 2008). Ramakrishanan (2007) found that P. maculata was able to tolerate emersion for 70 days at 30C and >95% relative humidity (RH) and >308 days at 20 25C and 75%
86 >95% RH. The results of my st udy were similar for both P. maculata and P. canaliculata As RH increased to >80%, survival rates for adults of both species reached 365 days, at which point the trial was discontinued. In lower R H conditions ( < 60%) survival was 22 weeks for both species. Similar to the results of Ramakrishanan (2007) I found that smaller snail s had shorter survival times, regardless of the test condition. Pomacea spp. can apparently tolerate prolonged emersion and aestivat e for periods of time that protect them t hrough normal seasonal patterns. Furthermore, the ability to withstand drought conditions and water management draw downs gives the snails an advantage in some systems. Management strategies for control of invasive species include chemical, biological and mecha nical approaches. Results of my study suggest that complete control or eradication of P. maculata and P. canaliculata from all systems will not be possible. For example, dry down events are often used in the management of wetlands and other water bodies to contro l unwanted plants and animals. But given the ability of nonindigenous Pomacea to survive such conditions this would not be an effective control strategy and could even be detrimental to the native apple snail (Darby et al. 2008). Similarly, chemic al approaches may have limited success. In response to physical disturbances, P. canaliculata can bury into the sediments in just seconds If the application of a chemical triggers a defense response, the snails may bury or simply withdraw into their she it conditions. Biological controls may help stabilize populations in some larger systems. Yusa et al. (2006) looked at numerous potential predators for P. canaliculata and many of the test predators or congener s, occur in Florida. Other strategies, such as hand
87 removal of eggs and live snails, are successful in some situations (e.g. urban ponds), and may be more effective than drawdowns or chemical applications (Bernatis 2013). Understanding the relationship of an organism to its envir onment throughout its life history is important for many reasons. Invasive Pomacea have the ability to colonize a variety of lakes and rivers and other bodies of water (e.g., canals). Given their physiological tolerances combined w ith their unique biological traits (i.e., water and air breathing, rapid growth rates and generalist feeding behavior ), few obst acles will block colonization. Ultimately, the risk of spread and establishment throughout the U.S. is great. In temperate regio ns, colonization may be limited to systems with thermal refuges (e.g., power plants). I n southern regions however, many freshwater and low salin ity systems may be at risk for future colonization. This research provide s a basis for developing more complex experiments that can be initiated to understand how the snail s respond to characteristics of non native habitats which may in turn be used to predict establishment success and develop successful control mechanisms.
88 Table 3 1 Comparison of survival curve (using the Mantel Cox Log rank Test). Condition Adults Juveniles Hatchlings P. canaliculata P. maculata P. canaliculata P. maculata P. canaliculata P. maculata Starvation G W = 16.65 G W = 0.00 G W = 2.8 G W = 2.5 P < 0.001 P = 0.995 P = 0.093 P = 0.115 Within Species Comparisons Within Species Comparisons Starved Control Not performed owing to non significance of global test Not performed owing to non significance of global test Not performed owing to non significance of global test Between Species Comparisons Between Species Comparisons Co ntrol 2 = 0. 33 P = 0. 569 Control 2 = 0. 05 P = 0. 831 Starved 2 = 0. 22 P = 0. 636 Starved 2 = 15.11 P < 0.0001 Salinity G W = 130.3 G W = 123.4 G W = 134.8 G W = 129.0 G W = 140.2 G W = 140.0 P < 0.001 P < 0.001 P < 0.001 P < 0.001 P < 0.001 P < 0.001 Within Species Comparisons Within Species Comparisons Within Species Comparisons = 32) 32) All curves differed All curves differed All curves differed All curves differed Between Species Comparisons Between Species Comparisons Between Species Comparisons 0 Not calculable 0 Not calculable 0 2 = 1.71, P = 0.191 8 2 = 0.78, P = 0.378 8 2 = 1.95, P = 0.162 8 2 = 0.22, P = 0.643 16 2 = 3.89, P = 0.04 9 16 2 = 10.78, P = 0.001 16 Not calculable 24 2 = 0.16, P = 0.690 24 2 = 1.05, P = 0.305 24 Not calculable 32 2 = 0.11, P = 0 .741 32 Not calculable 32 Not calculable
89 Table 3 1 cont inued : Comparison of Surv ival Curves within and between s pe cies Condition Adults Juveniles Hatchlings P. canaliculata P. maculata P. canaliculata P. maculata P. canaliculata P. maculata pH G W = 7.3 G W = 8.8 G W = 18.0 G W = 5.3 G W = 64.3 G W = 72.3 P = 0.120 P = 0.067 P = 0.010 P = 0.253 P < 0.001 P < 0.001 Within Species Comparisons Within Species Comparisons Within Species Comparisons Not performed owing to non significance of global test Not performed owing to non significance of global test (5.5 = 6.5 = 9.5) Not performed owing to non significance of global test = 8.5 = 9.5) 8.5 = 9.5) Between Species Comparisons Between Species Comparisons Between Species Comparisons 5.5 2 =0.25, P = 0.619 5.5 2 =9.00, P = 0.003 5.5 2 =5.70, P = 0.017 6.5 2 = 0.37, P = 0.544 6.5 2 = 10.62, P = 0.001 6.5 2 = 0.08, P = 0.775 7.5 Not calculable 7.5 2 = 2.17, P = 0.141 7.5 2 = 0.002, P = 0.964 8.5 Not calculable 8. 5 2 = 1.96, P = 0.162 8.5 2 = 0.001, P = 0.974 9.5 2 = 4.21, P = 0.040 9.5 2 = 5.66, P = 0.017 9.5 2 = 0.22, P = 0.643 Desiccation Tolerance G W = 4.5 G W = 5.5 G W = 15.8 G W = 115.9 G W = 50.2 G W = 57.2 P = 0.104 P = 0.062 P < 0.001 P < 0.001 P < 0.001 P <0.001 Within Species Comparisons Within Species Comparisons Within Species Comparisons Not performed owing to non significance of global test Not performed owing to non significance of global test (wet = c dry dry All curves differed dry Between Species Comparisons Between Species Comparisons Between Species Comparisons Control 2 = 25.69, P < 0.001 Control 2 = 0.29, P = 0.590 Control 2 = 9.48, P = 0.002 Wet 2 = 1.68, P = 0.195 Wet 2 = 1.82, P = 0.178 Wet 2 = 1.48, P = 0.224 Dry 2 = 0.21, P = 0.649 Dry 2 = 76.40, P < 0.001 Dry 2 = 26.38, P < 0.001
90 Figure 3 1 A C. Salinity sur vival curves for P. canaliculata and P. maculata adult (A), juvenile (B), and hatchling (C) snails. Overlapping curves occur in adults at 24 and 32 ppt and in hatchlings at 16, 24 and 32 ppt. Survival p r oportion Days Days 0 ppt 8 ppt 16 ppt 24 ppt 32 ppt P. canaliculata P. maculata A B C
91 Figure 3 2 A C. pH survival curves for P. canaliculata and P. maculata adult (A), juvenile (B), and hatchling (C) snails. Note overlapping curves for duration of treatments in adults and scale difference fo r hatchlings. Survival proportion P. canaliculata P. maculata Days Days 5.5 6.5 7.5 8.5 9.5 A B C
92 Figure 3 3 A C. Wet/dry survival curves for P. canaliculata and P. maculata adult (A), juvenile (B), and hatchling (C) snails. P. maculata P. canaliculata Days Survival proportion Dry Wet Control Days A B C
93 CHAPTER 4 FEEDING RATES AND PREFERENCES Background Established populatio ns of nonindigenous apple snails in North America are increasing. Once confined to Hawaii, Texas, and Florida, nonindigenous Pomacea spp. are c urrently present in at least 10 states (Figure 1 1). The most commonly occurring nonindigenous species is Pomacea maculata (Perry 1810; formerly described as P. insularum P. canaliculata (Lamarck 1819), P. haustrum (Reeve 1856), P diffusa (Blume 19 57). Marisa cornuarietis (Linnaeus 1758), the giant rams ho rn snail, belongs to the same family, Ampullariidae, and this native of South America and the southern Caribbean is now found in several areas of North America. Pomacea canaliculata and P. maculata are known to consume a variety of macrophytes, periphyton, and terrestrial v egetation (e.g. grass, fruit). Although, generally considered herbivores, they are opportunistic feeders consuming the remains of other decaying organisms as well as other Pomacea spp eggs (personal observation). The primary reason the s nail s have been deemed a pest species in some countries is because of the damage they cause to rice and taro crops (Estebnet, 1995; Yusa et al. 1999; Carlsson et al. 2004; Carlsson et al. 200 5; Yu 2005; Nakamura 200 6 ; Yusa et al. 2006). E stimated damage to these crops amounts to millions of U.S. dollars per year (Cowie 2002) In the U.S., reports of invasive Pomacea feeding on crops are limited to Hawaii and Texas. In Florida numerous anecdotal reports from citizens and state agencies implicate these sna ils in changing lake conditions through their consumption of aquatic
94 macrophytes, which cause s a shift to phytoplankton dominance. Aquatic ecosystems and associated wetlands provide ecological and economic ben efits (Burlakova et al. 2009). These systems a re richly diverse and provide many services including: water purification, discharge and recharge, retention of pollutants, control of erosion, sources of energy and nutrient s and support of recreational activit i es ( Brenner et al. 1990 ; Brenner et al. 199 0 ; Barbier et al. 1996 ). Economically, wetland ecosystems are estimated to contribute $4.9 trillion annually in ecosystem s ervices (Costanza et al. 1997). In Florida, millions of dollars have been spent by state agencies on aquatic and wetland habitat rest oration ( State of Florida 2013 ). Wetlands are some of the most threatened habitats and much of the threat is attributed to invasive species (Zedler and Kercher 2005). Herbivores in wetland habitats may impact community and ecosystem structure and functio ns ( Lodge 199 1; Lodge et al. 1998 ; Carlsson et al. 2004 ). Given their documented impacts to crops, P. maculata and P. canaliculata are among those herbivores that could potentially affect wet lands ( Cowie 2002; Burl a kova et al. 2008). In Florida, impacts ha ve been observed in urban systems, but no quantitative data exist to suggest wide scale wetland damage. However, because of crop damage elsewhere, and the demonstrated potential for damage, the U.S. prohibits the interstate transport and importation of al l nonindigenous Pomacea except P. bridgesi Texas and Mississippi even prohibit the possession of nonindigenous Ampullariidae. In 2012, the European Commission of the European Union (ECEU) ban ned all Pomacea and included provisions for requiring inspectio n of transported plants and erad ication of the snails (ECEU 2012).
95 Negative impacts such as the consumption of most vegetation, may have been an unintended consequence of using Pomacea spp. as a bio control for nuisance aquatic weeds ( Howells 2001 ; Oku ma et al. 1994; personal observation). As a result, most feeding studies have focused on consumption rates and preferences of nonindigenous or invasive plants ( Cowie 2002 ; Carlsson et al. 2004 ; Boland et al. 2008; Carlsson and Bronmark 2006; Cowie 2002 ). G ettys et al. (2008) examined the effect of temperature on P. maculata feeding using both native and nonindigenous submersed plants, and found that both are readily consumed. Burlakova et al. (2009) looked at feeding rates o n native and nonindigenous plant s that are considered nece ssary for wetland restoration. Their study provided further evidence that P. maculata consume vegetation at a rate that may cause damage to we tland habitat structure and they concluded that the snails should not be use d as a bioco ntrol agent. The objective of my study was to evaluate the foraging activity of P. maculata and P. canaliculata on commonly co occurring native and nonindigenous vegetation. I accomplished this by conducting experiments that 1) compared no choice feed ing rates for adult and juvenile P. maculata and adult P. canaliculata and 2) use d a multi choice feeding experiment with plant species that are commonly managed (e.g. Hydrilla ) or used in aquatic habitat restoration The results can be used in developin g plant restoration activities for systems with nonindigenous Pomacea populations in an effort to minimize the snail impacts and thereby decrease economic loss and increase habitat quality.
96 Methods General Procedures Three series of experiments were us ed to determine feeding preference of snails and rates of consumption of common aquatic plants: 1) single plant species feeding rates, 2) six week preference with two snail age groups, and 3) two week mixed snail age population preference trials. Twelve plant species were identified by the Florida Fish and Wildlife Conservation Commission (FWC) as species of interest for evaluating snail feeding preferences and rates (Table 4 1). All trials used adult (> 30 mm total length) and juvenile (10 25 mm total len gth) snails. Adult P maculata were collected from Tallahassee, Lakeland, and Okeechobee, FL. Preliminary trials found no significant difference in feeding rates between the se populations, therefore snails from the three sites were combined in further tria ls. Adult P. canaliculata were collected from the only known Florida population, a retention pond in Duval County, FL. Adult snails were maintained in tanks for 2 weeks prior to trials and were fed an ad lib diet of lettuce, cantaloupe, tomatoes, H. verti cillata and Hydrocotyle sp. Juvenile snails were lab reared from egg clutches collected in the field and fed a diet of lettuce, cantaloupe, tomatoes, H. verticillata Hydrocotyle sp., and algae. S nails were used in one treatment only Preliminary observat ions demonstrated a difference in feeding rates of snails when offered portions of plant material (e.g. leaves) vs. intact plants. Therefore, in each trial plants were presented as whole plants (roots, stalks, leaves) to simulate natural systems. Variati on in growth patterns ( e g., Hydrilla verticillata vs. Schoenoplectus californicus ) prohibited using equal masses of material across all plant species. Therefore, p lant material mass was only equal within plant species
97 Damp dry weights were used for bo th snail mass and vegetation. Snails and plant material were blotted dry with towels prior to weighing. All weights were taken within 10 minutes of removal from water. Mass (M) was recorded in grams to the nearest 0.1 g using an OHAUS Navigator digital sca le (Ohaus Corporation, New Jersey, USA). Feeding Rates All 1 2 plant species were used for testing feeding rates of adult and juvenile P. maculata and eight 1) were used with adult P. canaliculata Plants were e ither field collected from Lake Okeechobee or purchased from Aquatic Pl ants of Florida (Sarasota, FL). Test chambers were 17 L plastic tanks (Kritter Keepers) with a 3 4 cm substrate of sand and aquatic plant fertilizer filled with 14 L of well water. Sub strate and water were allowed to stabilize for 24 hr. prior to introduction of plants. Treatments were single plant species; plants were damp dried, weighed, and placed into the tanks, ensuring roots were adequately buried. Tanks were allowed to stabilize for two hours prior to introduction of snails. The facility allowed for natural light and temperature conditions but did not allow for influence of other environmental conditions (e.g. rain). Trials were conducted from May to July in 2008 2009, and 2010 To prevent seasonal e ffects within snail species/size and plant species, the trial for any given snail group and plant species was conducted in one block of time (e.g., all adult P. canaliculata and H. verticillata trials were completed in May 2008 ) S nails were randomly selected ( P. maculata N = 30 adults and juveniles; N = 20 adult P. canaliculata ) damp dried and weighed, after which a single snail was placed into each tank. The mean starting mass of adult P. maculata was 69.5 g (SE = 1.93), P.
98 cana liculata was 43.7 g (SE = 0.64), and juvenile P. maculata was 5.0 g (SE = 0.10). Tanks were checked daily I n the event th at vegetation was depleted or near ly deplet ed, more weighed plant material was added. A fter 72 hr snails and plant material were rem oved, damp dried, and the mass es of both were recorded. Between 10 and 15 control tanks (i.e., no snails) were used for each plant species Feeding rates were recorded as raw consumption rates (RCR) and then standardiz ed by snail mass (SCR). Feeding P re ferences Single age g roup F eeding preferences of P. maculata adults and juveniles were conducted with a six week cafeteria style experiment, modified from Gettys et al. (2008) and Brown ( 1982) using the plants indicated in Table 4 1 Snails were maintain ed and reared as described above. Trials were conducted in an outdoor facility to provide natural light and weather conditions Tanks were 91.4 cm in diameter and filled with 310 L of water. S and and aquatic plant fertilizer w ere mixed and spread across th e bottom of the tank to a depth of 8 10 cm. Tanks were allowed to stabilize for 72 h r. prior to introduction of plants and for another six hours before introducing snails. The trial was conducted from 21 June 2010 to 1 August 2010. There were three tre atment levels with 13 tanks per treatment: no snails (control), adult snails (2 snails/tank), and juvenile snails (5 snails/tank) Densities of snails were representative of natural densities I observed in Florida lakes S nails were marked and weighed, and total length (TL) and total width (TW) in mm were recorded. The mean starting mass was 137.6 g (SE = 4.2) for adults and 6.3 g (SE=0.1) for juveniles. A fter six weeks snails and plant material were removed, damp dried, and measurements repeated
99 Mixed age g roup A mixed age group trial was designed to test consumption preferences of mixed age populations. To remove multi season effects, and because of the quick rate at which snails grow, it was possible to conduct the trial over a period of only two wee ks, from 26 August 2010 through 9 September 2010. Protocols for tank set up and plant/snail introduction are as described in the single age group trial. This trial was comprised of four treatments (N = 32 with 8 tanks per treatment): no snails (control), 1 :1 of adults:juveniles (3 of each age), 2:1 of adults:juveniles (4 adults, 2 juveniles), and 1:2 of adults:juve niles (2 adults, 4 juveniles). In this trial only four plant species were used: V. americana H. verticillata N. advena and P. repens These s pecies commonly co occur in Florida systems. Based on the growth rates in the 6 week trial, only mass (M) was recorded for adults, but total length (TL), and M were recorded for juveniles. The mean starting adult M was 99.5 g (SE = 0.49). Juvenile starting mean M and TL were 1.9 g (SE = 0.06) and 16.36 mm (SE = 0.17). At the end of 14 days plants and snails were removed and measurements were repeated Statistical Analysis Feeding rates (both RCR and SCR) were analyzed using ANCOVA. To test for differences between treatments and plant species these variables were included as fixed effects in the model. To account for differences in the plant weights, the starting weight was included in the analysis as a covariate. No choice feeding rates were analyzed as ra w values and also standardized by snail mass, providing additional information on the effect of snail size and consumption. The preference trials were analyzed with a linear mixed model. Treatment and plant species were included as fixed effects to test fo r differences between groups and to account for differences of plant
100 weights, starting plant weights were included as a covariate. Also, to account for the fact that multiple plant species were presented simultaneously in each tank a repeated measures str ucture was modeled. Survival analysis (Proc Lifereg) was used in the 6 week and 2 week trials to compare plant survivorship. 0 .05. All analyses were carried out with SAS 9.2 (SAS Institute, Cary, North Carolina). Results The statistical significance of individual comparisons of all plant species are presented in Appendix A 1 for raw consumpti on rates (R CR ) and Appendix B 1 for standardized consumption rates ( SCR ) The global test for differences between all treatments was highly significant P < 0.0001; specific differences between treatments are discussed below. Feeding Rates Feeding r ates o f a dult P. canaliculata H ydrilla verticillata V. americana and S. californicus (P < 0 .0001), B. caroliniana (P = 0 .0045), and P. repens (P = 0 .0303) mass was significantly less in the snail treatment than the control tanks. The first four of these plants were also th e top four plants consumed (Table 4 2). Results of the RCR between plant species showed consumption of H. verticillata was greatest ( X = 22.3 g, SE 0.93) and was the only plant that differ ed significan tly from all other plants. Con sumption of H. verticillata was more than double that of the next most consumed plant, Bacopa caroliniana ( X = 10.8 g, SE = 1.32). The mean consumption rate for all plants combined over 72 hrs was < 10 g/plant material. In the SCR, H. verticill ata was consumed the most ( X = 0.47 g, SE 0.01) and was the only plant to differ significantly from all other plants. In both the RCR and SCR consumption of Sagittaria lancifolia and N advena ranked lowest. In both analyses, the
101 consumed amoun t of each plant differed significantly from at least one other plant. Significant differences in consumpt ion changed between RCR and SCR for t wo comparisons ; N. advena N. guadalupensis and S. californicus were not significantly different in the RCR, but w ere in the SCR. Four of the top five plants consumed were submerged vegetation. Feeding r ates of a dult P. maculata Hydrilla verticillata S. californicus and B. caroliniana (P < 0 .0001), V. americana (P = 0 .0011), C. esculant a (P = 0 .0046), P. repens (P = 0 .0077), and P. hemitom o n (P = 0 .0089) mass was significantly less in the snail treatment than in the control tanks These same plants were also the top seven plants c onsumed (Table 4 2). The RCR of H. verticillata ( X = 33.6 g/ 72 hrs SE = 3.36) was si gnificantly different from all other plants species. Ranking second was S. califo r nicus ( X = 19.4 g/72 hrs SE 3.29) but it was consumed at a rate of ~ 60% of H. verticillata The least consumed plant was P. illinoensis ( X = 0.6 g SE = 0.15) and consumptio n appeared restricted to the margins of the leaves. The SCR were similar to the RCRs and H. verticillata was again significantly different from all species and ranked highest in consumption ( X = 0.88 g SE = 0.04). Two plant species were equally consumed t he least, P. illinoensis and Z. aquatic, ( X = 0.01 g / g of body weight ; SE 0.01). In both analyses, the snail consumption rate of each plant species differed significantly from at least one other plant. Significant differences in consumpt ion c hanged between RCR and SCR for several comparisons : Z. aquatica vs. V americana N. advena P. hemitom o n P. repens C. esculant a ; P. illinoensis vs. N. guadalupensis V. americana N. advena P. repens S. lancifolia P. hemitom o n C. esculant a ; P. repen s vs. N. guadalupensis V. americana S. lancifolia ; V. americana vs. B. caroliniana P. hemitomen ; C. esc ulanta vs. B. caroliniana S. californi cus ; and S.
102 lancifolia vs. P. hemitom o n consumption was not significantly different in the RCR, but was in the SCR. Comparing consumption rates of submerged versus emergent vegetation showed mean consumption of submerged vegetation was slightly greater than consumption of emergent vegetation, 7.9 g (SE = 2.82) and 7.7 g (SE = 3.15), even after removing H. verticill ata which accounted for 51.06% of the total vegetation consumed. Feeding r ate s of j uvenile P. maculata Hydrilla verticillata V. americana S. californicus and C. esculant a (P < 0 .0001), V. americana (P = 0 .0002), P. repens (P = 0 .0012), N. gu a da l upe nsis (P= 0 .0029), N. advena (P = 0 .0337), and P. hemitom o n (P = 0.0449) mass in the snail treatment was significantly less than in the control tanks. Differing from adults the juvenile RCR showed that H. verticillata was consumed at the second highest rat e ( X = 4.61 SE 0.60) and C. esculant a was consumed at the greatest rate ( X = 5.0 g, SE 0.62; Table 4 2 ). T he least consumed plant was P. illinoensis ( X = 0.7 g SE 0.15). However, i n the SCR analysis H. verticillata was consumed at the greatest rate ( X = 1.03 g SE = 0.17). All other plants were consumed at < 1.0 g/g body weight. The least consumed was P. illinoensis at X = 0.19 g /g body weight (SE 0.04) Significant differences in consumpt ion changed between RCR and SCR for several comparisons : H. verti cillata vs. N. guadalupensis V. americana and Z. aquatica ; P. repens vs. N. guadalupensis V. americana N. advena S. lancifolia L. and P. illinoensis ; P. illinoensis vs. N. guadalupensis V. americana and P. hemitomon ; and P. hemitomo n vs. N. advena and S. lancifolia L. were not significantly different in the RCR, but were in the SCR. The mean consumption of submerged vegetation was less than emergent vegetation, 2.1 g
103 (SE = 0.55) and 3.1 g (SE = 0.48); and H. verticillata was the only submersed speci es to make the top five in consumption rates. Comparison of feeding r ates of adult s nail s Eight plant species were compared for differences in adult feeding rates (Table 4 2). Results of the RCR were significantly different for all model terms ( plant, sn ail species, and plant*snail species ; P <0.0001). Pomacea maculata consumed significantly more than P. canaliculata for three plant species: P. repens (P = 0 .0062), B. caroliniana (P < 0 .0001) and S. californicus (P < 0 .0001). The greatest difference was obs erved with S. californicus for which P. canaliculata consumed a mean of 3.7 g (SE = 1.07) and P. maculata consumed nearly 3x more ( X = 19.4 g, SE = 3.29). Three plants, H. verticillata V. americana and B. caroliniana were consumed by P. canaliculata at approximately two thirds the rate of P. maculata In the SCR, snail species was not a significant model effect (P = 0.0802), but plant (P < 0 .0001) and plant*snail species (P = 0 .0016) were significant terms. The only SCR that differed significantly betwee n P. maculata and P. canaliculata was for S. californicus (SCR = 0.25 and 0.09, respectively; P <0.0001). Comparisons of feeding r ates of P. maculata age g roups The overall RCR model for differences between adult and juvenile snails was significant, as wer e all model terms (P <0.001 for all). Comparison of RCR between P. maculata adults and juveniles indicated five of the plant species were consumed at significantly different rates (Table 4 2). Although adults consumed significantly greater absolute quantit ies, standardizing consumption to account for snail mass showed that juveniles ate more vegetation per gram of body mass, with the except ion of V. americana The overall SCR model for differences between adult and juvenile snails
104 was significant, as were a ll model terms (P < 0.001 for all). The differences in SCR were significant for all plant species except H. verticillata V. americana and B. caroliniana The greatest difference in SCR was observed with P. illinoensis with juveniles consuming 19x more t han adults. The most similar consumption rate was for V. americana with adults consuming only 12% more than juveniles. Adjusted for mass juvenile P. maculata consume d greater amounts of all plant species tested. Feeding Preferences Six week feeding p refe rence s tudy There was an overall loss of plant biomass in both the adult and juvenile treatments and an increase in the control group ( Table 4 3). During the course of the trial there was a total decrease in combined plant species biomass of 5473.9 g ( X = 421.1 g, SE = 36.1) in the adult tanks and 4989.9 g ( X = 383.8 g, SE = 28.3) in the juvenile tanks; this difference was not significant (P = 0.2103). However, the control tank plan t biomass increased by 1741.6 g ( X = 133 .9 g, SE = 28.2), which was significantly different from tanks containing both snail age groups (P <0.0001; Figure 4 1 ). Two plant species had an overall loss of mass in the control treatments. Sagit t aria lancifolia had an overall decrease in biomass ( X = 1.6 g SE = 1 8 4 ) and B. caroliniana ( X = 3.9 g SE = 1.1). Overall, the control group had significantly greater plant mass for all plants than the juvenile treatment (P <0.0001 ) One exception was the mean final mass of S. lancifoli a, which was not signif icantly different (P = 0.1168). Masses of all plants at the end of the study were significantly greater in the control treatments than the adult treatments. The least significant was S. lancifolia (P = 0.0420), whereas the remainder had significance values of P < 0.0087.
105 In the adult treatment there was an overall significant effect of plant species (P <0.0001). For all plant species, consumption was significantly different from at least one other species (Table 4 4). An increase in plant biomass was obse rved in at least one tank for four species: S. lancifolia P. repens and P. hemitomon ; f ive tanks had an increase in N. advena A single tank had an increase in three plant species, Panicum spp. and N. advena but this tank also had one snail die during th e trial. Similar to the adults, the effect of plant was also significant in the juvenile treatment (P <0.0001). In the juvenile treatment all plant species except H. verticillata V. americana and N. guadalupensis had at least one tank with a plant biomas s increase. In the overall model comparing the change in mass within plant species and between age groups, there was not a significant difference (P = 0.0949 ; Table 4 4 ). However, significant differences were observed with B. caroliniana (P = 0.0213) and S californicus (P = 0.0232) The change in plant mass was 55.8% and 62.6% greater, respectively, in the adult treatment. The change in mass for H. verticillata and N. guadalupensis was greater in the juvenile treatment, but neither w as significantly diffe rent from the change in mass from the adult treatment S urvival analysis conducted on the presence/absence of plants (above the sediments) was significantly different between treatments (P <0.0001 ; Table 4 5 ). The post hoc adjusted multiple comparison log rank test found significant differences between all treatments: adult and juvenile (P = 0.0056), adult and control (P <0.0001), and juvenile and control (P = 0.0006). In the c ontrol tanks all plant species except B. caroliniana were present throughout the trial ( X = 42 days, range = 42 days). Bacopa caroliniana was no longer visible in one tank at day 14, and subsequently disappeared
106 from other tanks ( X = 33.7 days; range = 14 42 days). In both age groups, N. guadalupensis was deple ted in at least one tank in nine days. In the adult tanks N. guadalupensis was present throughout the trial ( X = 22.6 days), however in the juvenile tanks it was gone by day 27 ( X = 20.1 days). Two week feeding preference t rials T here was an overall significant difference in consumption between plants and among the treatments (P <0.0001). Treatment 1 (4 adults:2 juveniles) was significantly different from treatment 2 ( 3 adults: 3 juveniles; P = 0.0129) and treatment 3 ( 2 adults: 4 juveniles; P = 0.0102) but treatment 2 was not significantly different from 3 (P = 0.9237). All treatments were significantly different from the control (P <0.0001). For each vegetation type t reatment 1 had the greatest total consumption (Figure 4 2 ). Tr eatment 1 consumption rate for Hydrilla was significantly different from t reatment 2 (P = 0.0099). Treatment 1 consumption of V. americana was significantly greater, nearly twice t reatments 2 or 3 (P = 0.0318, 0.0124, respectively). No other within plant s pecies consumption rates were significantly different between snail tank treatments. T he control tanks however, had greater end amounts of vegetation compared to snail treatments; and the final mass was significantly greater for both Hydrilla and V. ameri cana (P < 0.0001 ) Survival analysis on presence/absence was not conducted as all tanks had some vegetation of all species at the end of the trial. Four snails died during the trial, one adult in treatment 2, two juveniles in treatment 2, and one juvenile in treatment 3. In the overall model for differences in growth, the initial snail mass was the significant effect (P <0.0001). Although the effect of treatment was not significant (P = 0.0826), there was a significant difference between
107 treatment 1 and 2 (P = 0.0268). In all treatments, juveniles had the greater change in mass: treatment 1, 3.54 and 4.81 g (P = 0.0663); treatment 2, 3.31 and 4.35 g (P = 0.4666); and treatment 3 3.21 and 3.78 g (P = 0.0083). Discussion Nonindigenous apple snails consume a variety of emergent and submergent plants These snails have been implicated in the decimation of pond vegetation and are a demonstrated threat to certain crops (Esteb e net, 1995; Yusa et al. 1999; Carlsson et al. 2004; Carlsson et al. 2005; Yu 2005; Nakamu ra 2006 ; Yusa et al. 2006; personal observation) The impacts of these snails are in the millions of U.S. dollars, and potential threats to ecosystem function placed P. canaliculata Invaders list and prompted the banning of the ent ire Pomacea genus in the European Union (Lowe et al. 2000 ; Cowie 2002 ). Implementing proper management strategies for systems where P. maculata and P. canaliculata are present requires a thorough understanding of their ecology. The results of my feeding st udies expand on the knowledge base of the foraging rates and food preferences of both species. My results support previous studies demonstrating that smaller P. maculata and P. canaliculata have appetites equivalent to, or exceeding, larger snails at leas t on a mas s adjusted basis ( Carlsson and Lacoursie re, 2005; Carlsson and Br nmark, 2006; Burlakova et al. 2009; Fang et al 2009; Qiu and Kwong 2009; Burks et al. 2011). My research, however, provides evidence that larger adults ultimately consume more tha n juveniles, making population size class structure a key factor in determining ecosystem impacts. Furthermore, I found that feeding is not restricted to nonindigenous vegetation. Several plant species that are used for aquatic ecosystem restoration are co nsumed by
108 introduced snails. Finally, in addition to snail age/size, plant age is also an important factor in foraging behavior. Although the overall rate of consumption is greater for adult snails, when standardized for snail mass, juveniles consume d as much or more of the plant species tested in my study, except V. americana Previous feeding preference tr ials used small adult snails (< 16 g) and juveniles (Carlsson and Br nmark 2006; Burks et al 2011 ; Morrison and Hay 2011 ). In my study, P. maculata juv eniles were of similar starting size to those in the Burks et al. (2011) study, but the mean adult s tarting mass in my study was much greater, approximately 69 g in feeding rate trials and 120 g in long term trials. My statistical analysis took into consid eration the starting weight of the snail and found that this was a significant variable in feeding rate. One plant species, C. esculant a, which was tested in both my study and Burks et al. (2011), yielded similar results when comparing consumption between snail size classes. Results for the small size class in my study were different from the findings of Burks et al. (2011), perhaps because my small size class encompassed both size classes in their study. But my findings were in agreement with theirs insofa r as they showed that larger snails consume less on a body weight basis. Results from P. hemitomon trials by Morrison and Hay (2011) were very similar to my results. Snails in both trials did not show a predilection for eating Panicum spp. Other studies have also demonstrated that P. canaliculata size is also a factor in plant consumption rates (Carlsson and Br nmark 2006; Morrison and Hay 2011). These studies, however, used smaller snails than I did. Bacopa carolini a na was tested by Morrison and Hay (201 1) and in my study, with similar results despite different snail
109 size, which may indicate that consumption rates of certain plants may not change with snail size. The mean starting mass in Morrison and Hay (2011) was 0.4 1.5 g and in my study was 43.7 g, y et adjusted for snail body weight, plant consumption rates were within 0.02 g of each other. Two other plant genera were common to both studies, Panicum and Sagittar i a Results were similar for the former, but markedly different for the latter. Whether the difference in Sagittar i a spp. consumption was driven by snail size or test methods is unknown. Differences in methodology may account for some of the variation, as preliminary observations demonstrated different feeding behaviors based on the portion of t he plant offered to the snail. Qiu and Kwong ( 2009 ) also noted differences in feeding and growth dependent on quality of the vegetation. I offered snails whole plants fixed in a substrate, to better mimic natural environments. Size may also be important in determining the impacts of nonindigenous appl e snails on aquatic vegetation. Although results from previous studies and my study suggest that small snails have a greater rate of consumption per unit body weight, t his only holds for snails <30 mm TL or < 16 g M. The adult test snails from my study were much larger than those in previous feeding trials, and because of the much larger biomass, they ultimately consumed more vegetation than juveniles. For example, in my study, the mean standardized consumption for H. verticillata for adults was 0.88 g/g snail mass and 1.03 g/g snail mass for juveniles; however, the mean starting mass for each snail was 69.5 g and 5.0 g, respectively. Using the standardized consumption rates and the mean starting mass adults co nsumed 61.16 g/72 hr. vs. 5.15 g/ 72 hr. for juveniles. The observed higher consumption rates of juveniles may be important when
110 extrapolated to population size class structure (e.g., in situations with few adults and numerous young snails). The feeding pr eference trials were designed to look at community feeding rates of P. maculata Preferential consumption of H. verticillata was observed for both adults and juveniles. No ontogenetic shift in preference was observed, except in the case of Panicum spp. Juv eniles consumed greater amounts of P. hemitomon whereas adults consumed more P. repens Otherwise, adults and juveniles consumed all plants in the same order, ranked by consumption rates, supporting previous studies that also found no ontogen e tic shifts (C arlsson and Br nmark 2006; Burks et al. 2011). Overall, adults consumed more of all plant species, but only two plants were significantly different between age groups, B. caroliniana and S. californicus It is important to note that juveniles had reached a dult status by the end of the trial, but were still smaller than the large adults. Given that there were more juveniles in th e tanks (N=5 /tank ), if they had the same appetite as adults, consumption of vegetation should have been much greater than in the ad ult tanks (N=2/tank). Greater impact by adult dominated communities was observed in the 2 week mixed age trial. For all vegetation types, the treatment with four adults and two juveniles out consumed the other age combinations. Snails in the juvenile dom inated treatment ranked second in consumption of H. verticillata and V. a mericana but otherwise consumption by this population structure was ranked third. It was also only in this treatment where a plant, P. repens increased in biomass. My short term mes ocosm study, while contradicting previous work regarding snail size impact, demonstrates that
111 beyond size, snail size class distribution needs to be evaluated to determine potential impacts in nature. The results of my feeding rate and preference stu dies confirm p reliminary observations regarding plant quality as a factor for consumption, and provide additional support of similar observations by Qiu and Kwong ( 2009 ). Neither species of snail restricted their feeding to just stems and leaves. They cons umed roots and root structures (e.g. bulbs, rhizomes) of many of the species tested when this portion of the plant was available. In the case of C. esculant a and to an extent N. advena the root structures are readily consumed by both native and nonindige nous Pomacea even in the presenc e of other preferred foods. Overall, there did not appear to be a preference for emergent or submerged vegetation, but more of a preference based on plant age and size. For example, cut S. californicus stems that were offer ed as food in holding tanks and preliminary feeding rate trials were virtually ignored; however, small whole plants were readily consumed. This was also observed with S. lancifolia N. advena and both Panicum species. When new growth emerged the snails r eadily fed on the tender new growth, but once stems were above the water line or greater than about 0.5 m long, foraging activity was minimal. N ew growth of emergent and floating vegetation tested in the study appears to be highly preferred A s the stems/l eaves age and rise above the water line however, it appears to be less susceptible to Pomacea consumption. One exception to this occurred with V. americana In both preference trials, foraging behavior on V. americana was markedly different than for any other plant species tested. In all snail treatment tanks the snails would eat the stems just above
112 the sediment thereby cutting off the blades. Over time cut blades were gradually consumed either by snails that migrated to the water surface to consu me floating blades, or by snails on the bottom, once the leaves sank to the sediment. In all cases the only portion of the plant remaining above the sediment was the node, yet the plants recover ed New growth never exceeded 0.25 m and retained red pigment ation. The snails were seldom observed feeding on this new growth, even when no other food was available. After the trials, the remaining plant material was placed into snail holding tanks where it has retained the red pigmentation and short stature two ye ars later. It is not uncommon for plants to alter their composition to become unpalatable in the presence of herbivores ( Sunell and Healy 1985 ; Qiu and Kwong 2009 ) This may be a response mechanism of V. americana and should be further investigated as it suggests the plant can withstand nonindigenous apple snails. B ecause the long blade s are replaced by short blades, however, the habitat is altered for other organisms. And, if the red pigmentation is a defense mechanism, this could have adverse affects on other aquatic fauna as well. There is little doubt that in controlled settings, a consumption preference occurs between juvenile and adult P. canaliculata and P. maculata (Lach et al 2000; Burks et al 2011). R esults of my feeding preference studies, in con junction with prior studies provide ample evidence that in certain systems (e.g. retention ponds) with preferred plants, nonindigenous Pomacea may decimate plant communities Unfortunately, as research continues more plants are identified as being highl y palatable to the snails and as a result determining which plants are suitable for restoration efforts becomes more difficult, particularly if eradication of snails prior to planting is not feasible.
113 T he results of my feeding preference stud ies provid e information that may be helpful for reduc ing the impacts of the nonindigenous snails. S nails appear to prefer new stem and shoot growth of emergent vegetation. For instance, planting vegetation that is at least 0.5 m tall or above the water level at the planting site may reduce the impact of snail feeding long enough for some plants to become established. In locations where snails are abundant ensuring plants are above the water line is recommended. Second, choosing plants with known herbivore defense mechanisms may also restrict snail foraging activity. This may not completely eliminate consumption, as was the case for V. americana but could reduce consumption Third, because the snails do not completely deplete any one plant type before moving on, pl anting a variety of species may reduce complete consumption of any one plant species. Finally, planting in multiple locations, with an emphasis in areas with few or no snail s may enhance the possibility of plant establishment. Aquatic r estoration activit ies in Florida cost millions of dollars annually ( State of Florida 2013) In some cases, these expenditures have occurred in locations with known P. maculata populations, ultimately resulting in an economic loss as a consequence of snail foraging. Recogniz ing that the snails have a preference for particular plant speci es and exhibit certain feeding behaviors (i.e., feeding on new growth) will be useful in selecting restoration options that are less impacted by snails. Ideally, snails should be eradicat ed or reduced in numbers prior to restoration, but this is not always feasible. Techniques to reduce the fiscal and environmental impacts of snails should be implemented in restoration efforts where nonindigenous Pomacea occur
114 Table 4 1 Plant species used in the three feeding trials. Common names (Wunderlin and Hansen 2003) are provided in parentheses. Those marked with an were used in the adult P. canaliculata 72 hr feeding rate trial. Plant (Common name) Adult Juvenile 72 hr 6 week 2 week 72 hr 6 week 2 week Hydrilla verticillata 1 (Waterthyme) Vallisneria americana 1 (Tapegrass) Nuphar advena 1 (Spatterdock) Panicum hemitomon (Maidencane) Panicum repens (Torpedograss) Bacopa ca roliniana 1 (Lemon Bacopa) Schoenoplectus californicus (Giant Bulrush) Potamogeton illinoensis 1 (Illinois Pondweed) Najas guadalupensis* 1 (Southern Waternymph) Sagittaria lancifolia L. (Bulltongue Arrowhead ) Colacasia esculanta (Wild Taro) Zizania aquatica (Annual Wild Rice) 1 Majority of the plant structure is submerged
115 Table 4 2 C onsumption ( X C) and body weight adjusted ( standardized ) X consumption ( X SC), in grams, for the 72 hour feeding trials. Within plant species P values are provided. P values in the P. canaliculata column are comparisons with adult (A) P. maculata P values in the P. maculata juvenile (J) column are comparisons with adult P. maculata Treatment P. canaliculata (A) X C X SC SE SE P value P value P. maculata (A) X C X SC SE SE P. maculata (J) X C X SC SE SE P value P value Plant H. verticillata 22.3 0.93 0.2093 0.47 0.01 0.1316 33.6 3.36 0.88 0.05 4.61 0.60 < 0.0001 1.03 0.17 0.1525 V. americana 9.3 0.59 0.1521 0.22 0.01 0.1698 12.1 1.83 0.51 0.10 1.8 0.16 0.0002 0.43 0.06 0.3175 N. advena 1.1 0.25 0.983 6 0.03 0.01 0.1128 1.8 0.35 0.07 0.02 1.3 0.60 0.6800 0.26 0.10 0.0329 P. hemitomon 4.3 1.09 0.07 0.02 1.9 0.30 0.0011 0.64 0.13 < 0.0001 P. repens 1.9 0.49 0.0062 0.04 0.01 0.1418 4.7 0.47 0.07 0.01 3.2 0.36 0.0857 0.98 0.14 < 0.0001 B. carolinia na 10.8 1.32 < 0.0001 0.30 0.04 0.8599 14.9 1.44 0.38 0.06 2.4 0.31 < 0.0001 0.64 0.09 0.1026 S. californicus 3.7 1.07 < 0.0001 0.09 0.02 < 0.0001 19.4 3.29 0.25 0.03 3.4 0.26 < 0.0001 0.73 0.09 < 0.0001 P. illinoensis 0.6 0.15 0.01 0.00 0.7 0.15 0 .5426 0.19 0.04 < 0.0001 N. guadalupensis 2.3 0.31 0.1599 0.05 0.01 0.1952 2.8 0.40 0.03 0.00 1.7 0.22 0.1204 0.45 0.07 < 0.0001 S. lancifolia L. 1.6 0.34 0.2553 0.03 0.01 0.4659 1.8 0.32 0.02 0.00 1.8 0.27 0.3394 0.34 0.05 < 0.0001 C. esculant a 1 5.3 2.48 0.38 0.08 5.0 0.62 0.2839 0.88 0.13 0.0039 Z. aquatica 0.9 0.22 0.01 0.00 3.0 0.36 0.2008 0.53 0.07 < 0.0001
116 Table 4 3 Results of the 6 week feeding preference for adults and juveniles. Total consumption (Total C), mean end mass ( X EM), mean consumption ( X C) and standard error (SE) of plant material consumed /tank in grams, are provided The third column is the mean end mass and SE in the control group. Adult Juvenile Control Total C X EM (SE) X C (SE) Total C X EM (SE) X C (SE) X EM (SE) H. verticillata 1835.3 10.4 (7.0) 141.2 (7.0) 1916.3 4.6 (0.9) 147.4 (29.8) 191.4 (6.2) V. americana 1060.2 18.5 (9.2) 81.5 (9.2) 936.1 29.2 (6.2) 72.0 (14.8) 101.5 (7.21) N. advena 126.5 148 .4 (15.7) 9.7 (17.6) 81.2 146.2 (16.2) 6.2 (18.7) 224.1 (7.0) P. hemitomon 142.9 40.8 (4.2) 10.99 (4.6) 222.5 34.2 (4.5) 17.12 (4.9) 54.0 (3.2) P. repens 181.1 38.5 (3.4) 181.1 (3.0) 116.7 43.2 (4.6) 8.9 (5.1) 60.1 (2.0) B. caroliniana 137.0 1.7 (0.5) 1 0.5 (0.8) 76.5 5.3 (1.5) 5.9 (1.7) 8.2 (1.2) S. californicus 176.6 4.7 (2.6) 13.6 (2.7) 110.6 9.6 (3.3) 8.5 (4.8) 32.6 (2.2) P. illinoensis 294.4 6.7 (1.7) 22.6 (2.1) 227.6 12.0 (3.4) 17.5 (5.3) 39.0 (3.8) N. guadalupensis 575.6 5.4 (3.2) 44.3 (2.9) 619 .9 1.9 (0.2) 47.7 (9.2) 62.6 (2.8) S. lancifolia L. 944.3 67.3 (18.8) 72.7 (21.1) 682.5 83.3 (17.6) 52.5 (19.4) 142.6 (16.6)
117 Table 4 4. P value comparisons of foraging rates in the 6 week preference trial. Top row is within adult and between pla nt species. Bottom row is within juvenile and between plant species. Schoenoplectus californicus was significantly different from all at P < 0.0001 and is not shown. The third line is the within plant species and between age group P value. Plant 1 2 3 4 5 6 7 8 9 Hydrilla verticillata (1) 1.0000 1.0000 0.8155 0.7120 0.3352 0.1126 0.5615 0.0002 < .0001 0.1524 0.0001 < .0001 < .0001 <. 0001 < .0001 0.0130 0.4974 0.0365 0.0032 Najas guadalupensis (2) 1.0000 1.0000 0.2360 0.180 3 0.6077 0.0006 < .0001 0.1298 0.0213 < .0001 < .0001 < .0001 0.0003 0.0011 0.0198 0.2051 0.0004 Vallisneria americana (3) 1.0000 1.0000 0.0943 < .0001 < .0001 0.0123 0.0087 < .0001 < .0001 < .0001 < .0001 0.8053 0.0866 0.0347 < 0001 Nuphar a dvena (4) 1.0000 1.0000 0.7647 0.0064 0.0082 0.7877 0.1782 0.9288 0.0262 < .0001 < .0001 < .0001 < .0001 Bacopa caroliniana (5) 1.0000 1.0000 0.0213 0.0009 0.0527 0.0010 0.4158 < .0001 0.0006 0.0044 < .0 001 Panicum r epens (6) 1.0000 1.0000 0.5421 0.6030 0.0185 < .0001 < .0001 < .0001 < .0001 Sagittaria lancifolia L. (7) 1.0000 1.0000 0.2510 < .0001 < .0001 < .0001 < .0001 Panicum hemitomon (8) 1.0000 1.0000 0.1919 0.0879 0.1710 Potamogeoton illinoensis (9) 1.0000 1.0000 0.1986
118 Table 4 5 Results of the presence/absence of plant species for the 6 week fe eding preference trial. The mean ( X P) days present and minimum/maximum range of plant material above the sediments are the last two columns for each age group. The X days present and range for the control group for all plants was 42. P values are provided for differences within plant species for control vs. adult (A) and juvenile (J) and between age groups. Adult Juvenile P values X P Range X P Range Control vs. A, J Adult/Juvenile H. verticillata 28.1 17 42 30.1 20 42 0.0006, 0.0077 0.9469 N. guadalupensis 22.6 9 42 20.1 9 27 < 0.0001, < 0.0001 0.9329 V. americana 34.5 2 8 42 38.2 25 42 0.0004, 0.4975 0.0332 N. advena 42 42 42 42 N/C* B. caroliniana 15.5 10 42 23.5 9 42 0.0035, 0.1060 0.5187 P. repens 42 42 41.6 37 42 1.0000, 0.5267 0.5267 S. lancifolia 39.1 28 42 39.9 30 42 0.1198, 0.3719 0.9250 P. hemitomon 42 42 4 1.0 29 42 1.0000, 0.5267 0.5267 P. illinoensis 36.8 19 42 39.0 31 42 0.0091, 0.1347 0.7220 S. californicus 36 31 42 36.4 26 42 < 0.0001, 0.1217 0.0344 N/C = Not calculated because all tanks ha d equal survival.
119 A B C Figure 4 1 Re presentative tanks at the end of the six week trial (A = Control; B = Juvenile; C = Adult). Photos by Jennifer Bernatis.
120 Figure 4 2 Total consumption during the 2 week trial. Values by bars are the tank mean and standard error (in parentheses), for the three treatment groups. Treatment 1 = 4 adults:2 juveniles; Treatment 2 = 3 adults: 3 juveniles; and Treatment 3 = 4 juveniles: 2 adults 164.4 (6.4) 132.2 (20.2) 25.5 (12.4) 10.9 (11.5) 3.7 (7.1) 148.6 (16.9) 74.4 (7.5) 40.2 (8.9) 40.5 (6.8) 0.9 (1.7) 0.3 (1.7) 1.5 (2.3) 3
121 CHAPTER 5 EFFECTIVENESS OF MANUAL REMOVAL FOR CONTROL OF NONINDIGENOUS APPLE SNAILS Background In the last decade apple snails (Ampullariidae: Pomacea ) have received much attention because of their ability to damage crops and vegetated habitats in freshwater systems ( Yusa and Wada 1999; Joshi 2001; Carlsson 2004; Martin 2004 ) Pomacea are native to South and Central America and one species, P omacea paludosa (Say 1829), is native to North America. Nonindigenous Pomacea species have been introduced into North America and Asia (Rawlings et al. 2007; Hayes et al. 2009) In the U.S., nonindigenous apple snail s are establi shed in at least 10 states (Figure 1 1 ; Rawlings et al. 2007; Byers et al. 2013 ). Florida has more Pomacea species than any other state. These include native, P. paludosa, and at least four nonindigenous species: P. canaliculata (Lamarck 1819), P. maculata (Perry 1810), P. diffusa (Blume 1957), and an unknown Pomacea species previously described as P. haustrum (Reeve 1856) (Rawlings et al. 2007; Hayes et al. 2009, 2012). As of 2013, Florida has nonindigenous Pomacea populations in at least 29 watersheds and in 38 of th (Figure 1 2; Bernatis unpublished data) The majority of the locations are inhabited by P. maculata but two locations are inhabited by P. canaliculata which was listed as one of the Apple snails belonging to the P. canaliculata group (e.g., P. maculata ) are a documented crop threat in we tland agriculture, especially to rice paddies in Southeast Asia and Japan, and to a lesser degree to taro productio n in Hawaii (Yusa and Wada 1999 ; Joshi 2001; Martin 2004). The potential for agricultural impacts in Florida is low and has yet to be documented. The overall potential for adverse impacts in natural
122 aquatic systems however, may be high. In large systems ( e.g., Lake Okeechobee) the impacts of the nonindigenous snails are unknown, as no evidence conclusively demonstrates a cause and effect relationship between the presence of nonindigenous apple snails and the loss of vegetation. T he impacts on aquatic veget ation in urban ponds have been described qualitatively by community residents and government agency employees. Follow up on such reports provides evidence that nonindigenous apple snails are capable of complete removal of aquatic plant communities un der ce rtain conditions ( personal observation ). Urban ponds are often community focal points and hundreds of thousands of dollars have been spent on their restoration. Nonindigenous Pomacea may negate restoration efforts through their rapid consumption of purpose ly planted aquatic vegetation, resulting in economic loss and potential ecosystem alteration (Carlsson et al. 2004). Many me thods (e.g., chemical, predator, mechanical and integrated) have been used in attempts to control the spread and impacts of nonindi genous apple snails, each with varying degrees of success and most with a focus on crop protection (Estebenet and Cazzaniga 1990; Litsinger and Estano 1993; Calumpang et al 1995; Halwart et al. 1998 ; Wada 2004 ; Yusa et al. 2006 ). In Florida where crop da mage has not occurred, eradication has focused upon limiting damage to habitat in natural and urban systems. C hemical eradication with copper sulfate (CuSO 4 ) was attempted on nonindigenous apple snails in a 14.57 ha area of Newnans Lake (2,700 ha), Alachua County north central Florida Following the treatment s, live snails and egg masses remained and the snail population subsequently expand ed throughout the lake ( personal observation). Several chemical treatments have been tested and are moderately effecti ve, but
123 impacts upon non target organisms, lingering residual toxins and economic considerations h ave led most authors to conclude that other methods such as hand removal water level manipulation, or natural predators are preferable to chemical control (Litsinger and Estano 1993; de la Cruz et al. 2000; Sin 2003 ). Various aquatic insect larvae (i.e. Sphaerodema molestum -water beetle) can consume small snails < 40 days old at the rate of 30 per day (Chanyapate 1997), and larvae of the firefly Lucio loa lateralis may consume snails up to three week s old (Kondo and Tanaka 1989). Numerous fish ( e.g., carp, tilapia, bass) are predators of snails, and appear to be limited only by their gape (Caguan and Joshi 2002; Yusa et al. 200 6 ). Bluegill s ( Lepomis mac rochirus ) have been observed attacking and feeding on injured P. canaliculata (personal observation). ecosystems known to prey on Pomacea spp include, the S nail K ite ( Rostrhamus s o c iabilis plumbeus ) limpkins ( A ramus guarauna ) American alligat or s ( Alligator mississippiensis ), largemouth bass ( Micropterus salmoides ), red eared turtle s ( Trachemys scripta elegans ), raccoon s ( Procyon lotor ) and rats ( Rattus norvegicus ) ; while ants ( Solenopsis invicta ) forage on eggs (Yusa et al. 2006; personal observation). Aquatic plant control programs may have direct and indirect impacts on apple snails. Undesirable emergent aquatic vegetation (e.g. Typha sp.) is commonly sprayed with herbicide/surfactant combinations. Typical ly, within 2 3 weeks post treatment plants are decaying, resulting in submersion of attached Pomacea egg masses. Submersion of eggs in water has been shown to increase mortality of developing snails, which could impact recruitment (Pizani et al 2005; Horn et al. 2008;). Nonindigenous apple snails have short incubation periods (7 14 days) and may not be as affected by
124 submersion. Native apple snails require longer to hatch, 21 days or more, potentially impacting recruitment when exposure to submersion incre ases. Direct application of herbicides or surfactants on apple snail eggs has not been extensively studied; however, there is evidence that chemicals (e.g., morpholine) reduce hatch rates (Wu et al. 2005 ). Florida has numerous aquatic ecosystems. These ec osystems regulate surfac e water discharge and aquifer recharge, and support endemic and economically important freshwater fishe ries and wildlife populations. Many of the aquatic ecosystems throughout Florida are dominated by submerged an d emergent aquatic vegetation. Any organism that forages indiscriminately and continuously on aquatic plants has the potential to significantly alter Florida ecosystems; for instance, by turning macrophyte dominated communities into phytoplankton dominated commu nities (Carls son et al., 2004). Developing an adequate nonindigenous apple snail control program is difficult as each aquatic ecosystem is unique. For instance, in urban settings, such as residential and municipal retention ponds, use of chemical treatments may not be an option. Therefore, the objective of my study was to investigate the effectiveness of a three year manual removal program in the management of P. canaliculata snails and egg masses in an urban retention pond. Maximal effectiveness would be achieved thro ugh complete eradication. We consider the population controlled when snail numbers are substantially reduced to minimize impacts on the system (Netherland and Schardt 2009). There is uncertai but without knowing the actual s tarting population number, there is no way to determine a meaningful value for final snail numbers. If found to be effective, a properly designed manual removal program could reduce the
125 ecological damage and economic costs associated with other types of r emoval programs. Methods Removal Protocols A residential neighborhood retention pond in south Jacksonville, FL was chosen as a site to evaluate the use of manual removal for control of nonindigenous apple snails (Figure 5 1). The pond has an area of 1.62 ha, a perimeter of 850 m and a maximum depth of 8.2 m. The perimeter of the pond has a submerged shelf extending approximately 3.5 m from the shore. Maximum depth of the shelf is 1.7 m. Beyond the shelf the pond reaches a maximum depth of 8.2 m. The area to the north and west of the pond is ephemeral sw amp and pine forest that draines into Julington Creek (approximately 2.5 km from the pond) and ultimately to the St. Johns River (approximately 12.5 km from the pond). The area to the east and south of the pond is primarily residential and commercial development. Five surveys of this Pomacea canaliculata population were conducted during 2006 2008. From the shoreline only snails on the peripheral shelf were visible; all snails were counted around the pond During this period, densities on the shelf ranged from 1 3 snails/m 2 V egetation in the pond included Hydrilla verticillata Hydrocotyle sp., Ceratophyllum demersum and Sagittaria sp. No other apple snail species (i.e., P. paludosa ) were present, an d the only other gastropods were Planorbella duryi Planorbella scalaris and Physella cubensis In May 2008, a three year study was implemented with the purpose of evaluating the efficacy of controlling the apple snail population by nonchemical means. On e day per week a Florida Fish and Wildlife Conservation Commission (FWC) field technician
126 walked the perimeter of the pond along the shelf and hand removed all observed P. canaliculata (e.g., hatchling through adult) and egg masses within reach Based on the each week. Because of a reduction in snails and eggs during winter and an overall decrease in snail and egg mass numbers as the project progressed, collection frequ ency was reduced Weekly removal continued until fall of 2009, when the frequency of collection was decreased to twice a month. During the third study year the time between collections was extended to three weeks. Collected snails were killed by freezing a t 12 o C for at least 2 weeks and were then buried; egg masses were scraped from the substrate and crushed at the pond. Cost of the project was monitored, including supplies (i.e., bags for snails and eggs), salary, and gas (calculated as mileage rate). During the first year, a second neighborhood in Jacksonville was identified as having P. canaliculata Because a control site with the same species was not available at the onset of the trial, and introducing snails elsewhere was not an option, this served as a reference site for the duration of the trial, i.e., to track the fate of snail populations where no eradication methods were used. The ponds in this neighborhood were smaller than 1 ha on average; however, four of the five ponds were found to have sn ails. No snail removal by the FWC was undertaken at this location. The ponds in this neighborhood were checked quarterly for the presence and absence of the snails. In July 2009, P. canaliculata were found in an ephemeral drainage ditch located adjacent to the primary study pond ( Figure 5 3 ). The ditch was connected to the pond by culverts and was considered a source for pond re colonization by apple snails.
127 Subsequently, snails and egg masses were removed from the ditch on the same schedule as the main pond After the discovery of the ditch population, all ponds (N= 17) within a 3 km radius of the primary study pond were surveyed quarterly for snails and egg masses. Because of this discovery and the difficulty in ensuring complete removal of snails from this location, a consequence of dry/wet cycles and self burying behavior of the snails, four post study surveys were conducted from June 2011 June 2012 to determine whether snails were repopulating the pond. The same study protocols (e.g., walking the shelf) w ere used. Dive Survey During study year 2, on April 16, 2010, two scuba divers (Karst Environmental Services Inc., High Springs, FL) conducted a benthic transect survey of the retention pond with the purpose of determining the presence and depth distributi on of apple snails in the pond. Eight transects crossing the pond perpendicular to the long side and evenly dividing the pond, were surveyed (Figure 5 1 ). Visibility was sufficient for the divers to survey 3 3.5 m on both sides of the transect line, for a 6 7 m wide sweep of each line per diver. The average depth of transects 1 through 6 was 6.4 m, and the average depths of transects 7 and 8 were 4.6 m and 3.0 m, respectively. Overall, the divers covered approximately one third of the total bottom surface area during the dive. Results Primary Location During the three year study a total of 107 collections were conducted. Total time spent collectin g snails and egg masses was 245 person hrs The total number of snails
128 and eggs removed from the pond were, 21 ,343 and 20,244, respectively. The overall cost of the project was $10,475, and covered 29.5 ha. The manual removal control program was initiated on May 29, 2008. On the first collection day, four FWC staff collected 2,948 snails and 1,737 egg masses. Thr oughout the first year (May 29, 2008 May 26, 2009) 49 collections were made, and 20,961 snails (Figure 5 2A ) and 18,934 egg masses (Figure 5 3A ) were removed. Three collections were missed because of in clement weather (e.g., storms). The average time spe nt per collection in year one was 3.28 person h rs. In year two (June 3, 2009 May 11, 2010) the number s of snails and eggs collected were markedly decreased from year one. The number of snails removed, N = 327, decreased by 98.5% from year one (Figure 5 2B ). Likewise, the number of egg masses removed, N = 1,249, was a 93.5% decrease from year one (Figure 5 3B ). On the first collection day the number of snails removed, N = 30, was 99% less than year one; and 97.3% fewer egg masses were removed, N = 48 D uring this collection year, all 38 scheduled collections were made, with an average collection time of 1.68 person hrs. During year three (May 27, 2010 June 1, 2011) a total of 55 snails (Figure 5 2C ) and 61 egg masses (Figure 5 3C ) were removed ; both wer e less than 1% of the first year totals. On the day of the final collection, no snails were observed, and two egg masses were removed. In the final year, all 20 scheduled collections were made and the average time per collection was 1.00 person hr. Result s of the dive survey conducted in April of year two indicated that snails were either absent from the pond or present in such low numbers that they were not
129 detectable. During the dive survey, apple snail shell material was retrieved at all depths on all t ransects, but no live snails were retrieved. The divers described the vegetation as patchy and sporadic. The dominant vegetation was an unidentified Sagittaria sp. that had been observed in the pond since the May 2006 evaluation. Three follow up visits co nfirmed the success (e.g., no live snails) of the manual removal method. On the first of these visits, in early September 2011, only one old and damaged egg mass was observed on a drain culvert and no snails were found. Two months later, on the second visi t, no egg masses or snails were observed. On the third visit, in February 2012, again no snails or egg masses were seen. Howeve r, in May 2012, after a storm, two sna ils and six small egg masses (< 20 eggs) were observed. As of June 2012, no other ponds or c reeks within a 3 km radius of the main pond and drainage ditch contained Pomacea canaliculata Reference Ponds Snails remained in the reference ponds, with no apparent declines, throughout the 3 year removal program at the main study site. As of March, 20 11 these snails had spread to another three ponds within the neighborhood. Neighborhoods around the main removal site have also been checked periodically for the presence of snails, but as of March 2011, no other ponds or creeks in the nearby (within 3 km) neighborhoods have nonindigenous Pomacea Discussion Nonindigenous Pomacea are abundant in many and the results of this study provide a framework for other infested locations. Because the impacts of nonindigenous Pomacea a ppear to be system depend ent, control programs must be tailored to the specific system. The goal of this study was to evaluate
130 the effectiveness of hand removal in the management of Pomacea canaliculata in a small urban pond. The results of this study foun d that regular removal of snails and eggs eventually controlled (e.g., negligible P. canaliculata presence) and possibly eradicated, this population. The greatest impact was made in year one, but the continued removal, helped ensure missed snails were even tually removed from the system. The follow up year surveys indicate control was achieved. The reappearance of a few snails in the primary study pond one year after the conclusion of the study is noteworthy and several explanations are plausible First, t he drainage ditch behind the pond continued to harbor snails for the duration of the study, and the rains of May 2012, which exceeded 40 cm in this area, reconnected the culvert between the two systems, allowing snail movement between the locations. Second it is known that the snails were originally introduced (by homeowners) into the pond to control vegetation; a second introduction may have occurred, for similar reasons or other reasons (e.g., emptying of an aquarium). As the snails did not reappear unti l after the May 2012 rains, the first explanation is more plausible and highlights the need to ensure removal in adjacent locations and to adapt the removal schedule to take into account the prese nce of surrounding populations. Finally, several sewer drain s empty into the pond. Therefore, there is a possibility that snails were located in these structures and avoided all collection attempts. Determining the most appropriate method of control is system dependent, and in some situations formal control may n ot be necessary. For example, in large natural systems, observations suggest that the impacts of noni n digenous Pomacea on aquatic vegetation are minimal (e.g., no observable change in vegetation) though they have not
131 been quantified (Bernatis personal obse rvation). Furthermore there have been no harmful effects reported in that have nonindigenous Pomacea populations. A likely explanation for differences in snail population dynamics between small urban systems and large natural lake a nd marsh systems is that in the latter snails are preyed upon by alligators, birds, fish, and raccoons and egg masses are preyed upon by animals such as fire ants (Halwart et al. 19 98; Yusa 2001; Yusa et al. 2006) This is further supported by the results of a recent study that demonstrates a relationship between predators and reduced P. canaliculata density in an urban river in Japan (Yamanishi et al. 2012). But in urban pond systems, specifically those that are community focal points, reliance on predato rs as a control method is not realistic. Ecologically, the impacts of hand removal compared with chemical methods were negligible. Some studies using chemical control have succeeded, but success has come with unintended ecological costs, such as high mort ality to non target organisms ( Maini and Rejesus 1993; Calumpang et al. 1995; de la Cruz et al. 2000). Chemical treatment of apple snails has proven most effective when the chemical directly contacts snails (de la Cruz and Joshi 2001). In many lake and riv er systems water clarity and flow prevent direct contact application. Also, Pomacea spp. have strong chemosensory abilities that coupled with survival mechanisms may allow them to survive well past the treatment viability period (Aizaki and Yusa 2009 ; per sonal observation ). Molluscicides generally have a half life of several days, but P. canaliculata and P. maculata depending on age, can survive more than a year without feeding or emerging from their shell (see Chapter 3 ). Although these approaches can be successful, particularly when combined with other methods (e.g., manual removal ), the use of manual removal alone
132 avoids the use of chemicals. This reduces impacts on non target organisms such as insects, fish and plants, prevents general contamination of the water, and guarantees removal of the snails from the system. Cost is often a consideration in management strategies. In this study, exclusive of the dive survey, the cost of manual removal was substantially less than the use of chemical treatments. Over the course of the project 107 collections were made, for a total cost of $10,475 or just under $100 per visit. Comparing this to the Newnans Lake treatment, the cost per ha was at least 35% lower at ($354.44/ha vs. $540.34/ha). The cost of the three copper sulfate applications on Newnans Lake included the chemical only which was provided at a reduced rate; salaries, gasoline, and other support equipment/staff expenses were not available. This study was designed to evaluate a three year manual remo val program for an urban pond. A substantial impact was made by the end of the first year, and control was successful by the completion of the study. Furthermore, success of this program was achieved in part, because of access to the entirety of the shore line. Systems in which access to emergent vegetation (e.g., egg laying substrate) or a large portion of the shoreline is limited, may not have the same level of success. The actual size of the system may not be a factor assuming adequate access, but resu lts may not be as quick and the scheduling protocols may need to be implemented. Such modifications could include twice per week collections during summer centered on peak reproductive periods. The primary benefit of multiple collections per week is the f aster removal of snails and egg masses from the system; this protocol may work for organizations with limited funding (e.g., one year grants). Although the cost of this program was relatively
133 low, a ny cost may be too great for some municipalit ies or homeow but manual removal has the b enefit of being implemented free through organized volunteer events. The manual removal method allows for constant monitoring, is adaptable based on the number of egg masses and snails present, and the r esults are quantifiable. Additional benefits can be gained from a weekly removal program, such as observations on the general health of the system (e.g., algal blooms, changes in vegetation). An effective manual removal and monitoring program can be establ ished at minimal cost. Such a program may ultimately lead to better management practices and overall improved system health. Regardless, management of invasive apple snails must be tailored to the system to avoid fiscal misuse and unwarranted, unnecessary ecological impacts.
134 Figure 5 1 Illustration of primary s tudy site in Duval County, FL. The pond is 850 m perimeter and is approx imately 1.62 ha in total area. The maxi mum depth is 8.2 m. The drainage creek study area is indicated by the shaded area. The dashed line between the two areas is a drain culvert connecting the areas. The numbered lines are the transects of the dive survey.
135 Figure 5 2 The number of snail s collected in each year A = year one (49 collections) ; B = yea r two (38 collect ions); and C = year three (20 collections). The in year one indicates the collection was not made (see text) ; the absence of a bar indicates zero snails were collected. For years one and two, dates of the first and last collection, as well as the dates of the first collection of each month are provided. Year three are all of the collection dates. A B C
136 Figure 5 3 The number of egg masses collected on each collection. A = year one (49 collections) ; B = yea r two (38 collections); C = year three (20 collec tions). The in year one indicates the collection was not made (see text) ; the absence of a bar indicates zero egg masses were collected. For years one and two, dates of the first and last collection, as well as the dates of the first collection of each m onth are provided. Year three are all of the collection dates. A B C
137 CHAPTER 6 CONCLUSIONS Pomacea spp. were intro duced in to North America and Asia for a variety of reasons that include d bio control, the aquarium trade, and the escargot industr y (Halwart 1994; Naylor 1996). In the last decade Pomacea canaliculata and Pomacea maculata have received much attention because of their potential to damage wetland crops through foraging and P. canaliculata 1 00 Invaders (Lowe et al. 2000). Nonindigenous apple snails in the U.S. occur in at least 10 states and Florida has more introduced apple snai l species than any other state. Although there are many questions to be addressed regarding invasive Pomacea the goal of my work wa s to provide information on the morphology, physiology, feeding behavior, and potential eradication techniques for P. canaliculata and P. maculata in Florida. This information can be used to identify species, develop removal protocols and minimize impacts. Nonindigenous apple snail species although similar in appearance, have growth patterns that allow the m to be distinguished from one an other. Relationships among 45 ratios w ere evaluated using parametric and non parametric analyses, and provide a range o f options for snail identification. E ase of identification i s depe ndent on the size of the individual. Similarities between species when snail total length is 30 mm limits the reliability of shell morphology for taxonomic identification Reliability of shell morphology for identification improves with larger snails, ranging from 3 0 50 mm total length, but is still limited. For example, in the cluster analysis accurate placement of P. maculata was 98.93%, but P. canaliculata was assigned nearly evenly between clusters, illustrating shell morphology was not adequate for identifying individuals in this size class.
138 Snails 50 mm in total length are the preferred size for using shell morphology to identify the two snail species. A 14 variable mo de l [ Y = .00598(TLTW) .00239(SHSW) + .00078(ALBL) + .00041(VHBL) .01557(TWSW) .00591(TLSW) + .01542(TLAP) + .00862(SWBL) .00143(TLBL) .00106(ALPA) .00235(PABL) .00457(SWAP) .00161(TLPA) + .00334(SHAP) ] enabled correct species identification 96.33% of the time for P. maculata and 95.65% of the time for P. canaliculata I f Y is negative the snail is likely P. maculata ( X = 2.56) whereas if it is p ositive then it is probably P. canaliculata ( X = 1.65). The relationship between total length and mass of shell material was also hypothesized as a means to distinguish between the two species Overall, shells of adult P. maculata were heav ier than those of adult P. canaliculata M uch as in the other analyses, however, the smaller the snail the less reliable this approach was as there was no significant difference in the ratio between small adults of the two species As size increased P. maculata generally weighed more than P. canaliculata for animals of a given length and thus the ratios were si gnificantly different. Given the overlapping values of the total length:mass ratios for the two species, however, this should not be the sole meth od used for species identification It is possible to distinguish the two snail s pecies from one an other, but the degree of success and reliability is dependent on several factors, most notably the availability of multiple identifiers and specimens. Beca use of the influence of environment on snail growth patterns, reliance on any single method or use of only a few specimens, particularly if small, will decrease the rel iability of the identification.
139 R esults of this work yielded basic of information that s hould provide stimulus for future studies. My work focused primarily on populations of nonindigenous snails in Florida Future research should focus on e xpanding the range of study locations e.g., all southeastern states. This will provide a larger base f or refining the model and perhaps identifying characters particular to specific to population s. Furthermore, o utliers were observed in the models T hese may have been a consequence of water or food quality issues a s both influence growth rates of Pomacea spp. Studies covering multiple environmental factors will provide additional information that may be used to generate more reliable identification methods. Nonindigenous apple snails are capable of tolerating broad ranges of environmental conditions. Alt hough many freshwater organisms have narrow tolerance ranges, these snails survive across a wide range of pH and moderate levels of salinity. They tolerate extended periods of desiccation and go without consuming food for extended perio ds of time. This rob ustness enhances their ability to invade and establish populations in numerous types of aquatic eco systems. S nail age is a determinant of survival. Hatchlings of both species had the lowest survival rate under all test co nditions, followed by juveniles. A d ults displayed the longest survival rates under all test conditions. With few exceptions survival was identical between species. Both P. maculata and P. canaliculata adults are able to tolerate salinity ranges from 0 8 ppt and can survive for at least 28 days at 8 ppt Juveniles of both species c an also tolerate 8 ppt, but median survival for the two species differed ( P. canaliculata = 28 days; P. maculata = 19.6 days). Hatchlings survived for 28 days at 0 ppt, but
140 survival at 8 ppt did not exceed 7 day s for either species. The pH of the water also appears to have little effect on the survival of the snails over the range found in many Florida aquatic systems. A 28 day survival was observed for adults and juveniles of both species for pH 5.5 9.5. Hatchli ngs however, displayed different median survival and all snails were deceased by day 14. Corrosion of the shell was observed for juveniles and adults of both species at pH 5.5. Adults of both species can survive for a year in high humidity, semi wet cond itions D ry conditions reduce survivorship to 154 days. Juveniles of both species of juveniles tolerate semi wet conditi ons for up to 147 days, but dry condition s affect species survival differently ( P. canaliculata survived 98 days, but P. maculata surviv ed only 14 days). Hatchling P. canaliculata survived to day 49 in the semi wet condition and day 42 in the dry condition whereas P. maculata survived 42 and 28 days respectively. The starvation survival trial demonstrated that nonindigenous Pomacea do no t need to consume macrophytes on a regular basis, and are able to withstand poor habitat conditions. Juveniles and adults of both species had median survival times of 28 days without feeding on macrophytes. High t olerances of P. canaliculata and P. macul ata to a variety of physiological stressors make them efficient invaders. Nevertheless many factors and their interactions can affect the survival of organisms, and the factors addressed in this study are only a few of those that probably affect snails an d each was tested individually. The influence that each factor has on long term establishment of snail populations requires more extensive laboratory and field research. Future physiological research should focus on c ontrolled laboratory studies evaluatin g the interactions of specific variables (e.g ., salinity and pH) would provide
141 additional information to help predict the future spread of these species. Furthermore, f ield evaluations of Pomacea population dynamics along with the physi c o chemical proper ties of the environment would provide in situ data to use with laboratory studies to predict future snail distribution s and develop means for control. Nonindigenous apple snails are opportunistic feeders and primarily consume macrophytes and algae. The ir high consumption rates and diverse palate s make P. maculata and P. canaliculata threat s to wetland s. They also have the capacity to damage the normal function and structure of aquatic eco systems. Previous studies compiled a long list of plants consumed an d results of my study expand ed that list My study also highlighted the importance of offering whole plants as they occur in situ and illustrated that different outcomes can occur if snails are offered whole, versus parts of plants. My feeding rate s tud ies indicate that the snails readily consume both submerged and emergent vegetation. In the single plant consumption rate trials, H. verticillata was overall, the most preferred plant followed by C. esculant a Panicum spp. are not readily consume d and co ntrary to previous work neither was P. illinoensis In my study entire plants were offered as opposed to selected plant parts. Snails consumed stems, leaves, roots and bulbs. Offering entire plants is more representative of natural conditions and provide s additional information on potential snail impacts on vegetation When consumption was normalized for snail size (gram consumed per gram snail mass) juveniles ate greater amounts of vegetation than the adults. Large adults however, consumed the greatest mass of vegetation simply because they are larger than juveniles.
142 I evaluated feeding preferences in two experiments, a six week single age group and a two week mixed age g roup. Overall, adults and adult dominated test groups consumed the most of al l plant species. Only one plant species, N. gu a dalupensis was consumed to eradication, and this occurred with juveniles in the single age trial. These experiments showed that snails readily consume all parts of the plants with a preference for new growth Aquatic ecosystem restoration, often involving plant re establishment costs millions of U.S. dollars annually. In some cases, planting occurs in locations with known P. maculata populations, ultimately resulting in economic loss es because of snail for aging. Recognizing that snails prefer particular macrophytes and like to consume new growth should make it easier to develop restoration plans that will minimize the impact of snails. For example, using mature plants or plants with new growth that is abov e the water line may minimize consumption by the snails My research expanded the list of plants that are consumed by nonindigenous snails. A dditional feeding rate and preference studies on plants used for restoration are necessary to develop restoration protocols for ecosystems in which P. maculata and P. canaliculata are present. Long term in situ studies are needed to further evaluate the impact of nonindigenous Pomacea on aquatic ecosystems. In situ studies will allow for the natural senescence and em ergence of the plants enabling evaluation of true snail impact. Finally, b oth laboratory and in situ studies which incorporate nonindigenous apple snail predators will provide additional information of snail behavior and facilitate development of impact prediction models.
143 Numerous methods have been evaluated for the control and eradication of nonindigenous Pomacea Most previous studies focused on agricultural areas and employed chemical control agents. In Florida, P maculata and P canaliculata of ten occur in urban aquatic ecosystems. My study therefore, evaluated the use of manual r emoval of snails and egg masses over the course of three years, to control P. canaliculata in a suburban pond. Hand removal of eggs and snails was conducted on a regu lar basis, beginning with weekly removal, followed by removal every two weeks, and finally every three weeks. The reduction in removal frequency was made possible by the rapid decline in numbers of snails and egg masses in the system at the end of the fir st year. Although complete eradication could not be confirmed, a dramatic decline in the snail population was documented A second site was monitored for presence/absence of P. canaliculata with no attempt at removal. During the three year study, snails e xpanded from this reference site into several surrounding ponds. A total of 21,343 snails and 20,244 egg masses w ere removed from the study pond, with 98.2% of the snails and 93.5% of the egg masses removed in year one This method had no apparent ecolog ical costs ( e.g. non target organism mortality) and did not interfere with human use of the system. The economic cost was minimal compared with other treatments and manual removal by hand is a method that can be carried out by volunteers. Such a n approac h may constitute a better management practice and may lead to improved eco system health. The management of invasive apple snails should be tailored to the specific system to avoid unnecessary costs and ecological impacts.
144 My research suggests that con trol of nonindigenous apple snails can be accomplished without the harmful impacts of chemical treatments. The extent to which this method may be used in larger systems needs to be investigated. Certain criteria such as wetland size and access to littoral habitats should still be considered when undertaking such a study, as these factors will determine the likelihood of success. Collection frequency should be determined through modeling techniques. Using these factors will help determine the cost of such a control program. In conclusion, invasive species cost taxpayers millions of dollars annually. Increasing our knowledge of an invasive snail species will ultimately enable us to eradicate control or at least mitigate its impacts Gaining such knowledge sh ould be a priority for aquatic ecologis ts My research fill ed some fundamental gaps in our knowledge of nonindigenous Pomacea Numerous recommendations for future research resulted from my work and will help advance our understanding of Pomacea in North Am erica. I hope that my findings and future studies will help minimize economic and ecologic al costs of invasive apple snails
145 APPENDIX A P VALUE COMPARISONS OF RAW CONSUMPTION FEEDING RATES Appendix A 1 P value comparisons of RCR feeding rates. Top p val ues are adult P. canaliculata feedin g rates between plant species. Middle p values are adult P. maculata feeding rates. Bottom p values are juvenile P. maculata feeding rates between plant species. The number in parenthesis under the plant name represents the same plant species in the top row. Plant 1 2 3 4 5 6 7 8 9 10 11 12 Hydrilla verticillata (1) 1.0000 1.0000 1.0000 < .0001 < .0001 0.1046 < .0001 < .0001 0.0831 < .0001 < .0001 0.0112 < .0001 0.0003 0.4789 < .0001 < .0001 0.1111 <. 0 001 < .0001 0.0195 < .0001 0.0142 < .0001 0.0011 < .0001 0.0002 0.2637 < .0001 0.0103 < .0001 0.9034 Najas guadalupensis (2) 1.0000 1.0000 1.0000 < .0001 0.0005 0.8743 0.9752 0.8922 0.1591 < .0001 < .0001 0.0203 0.4339 0.0825 0 .5272 0.9655 0.7762 0.1446 0.0385 0.0755 0.0847 0.0678 0.1743 < .0001 0.0017 0.0001 < .0001 0.0135 0.0450 Vallisneria americana (3) 1.0000 1.0000 1.0000 < .0001 0.0004 0.2054 0.9329 0.0003 0.0152 0.0003 0.0675 0.6063 < .0001 0. 0010 0.1800 0.1357 0.0930 0.2051 0.0958 0 .0015 0.0006 0.0010 0.6258 < .0001 0.7630 0.0301 Nuphar A dvena (4) 1.0000 1.0000 1.0000 < .0001 < .0001 0.0013 0.4110 0.0841 0.5530 0.9898 0.8606 0.8356 0.0385 0.4834 0.0636 0.7233 0. 1632 < .0001 < .0001 < .0001 < .0001 0.0065 0.0006 Bacopa caroliniana (5) 1.0000 1.0000 1.0000 0.0002 < .0001 0.0352 < .0001 < .0001 0.0034 < .0001 0.0027 < .0001 < .0001 0.0011 0.7662 0.6067 0.0014 0.0425 0.0001 0.6279 Panicum R epens (6) 1.0000 1.0000 1.0000 0.3988 0.1244 0.3893 0.7331 0.1489 0.7233 0.4227 0.0060 < .0001 0.0021 .0205 < .0001 0.1928 0.0189 Sagittaria lancifolia L. (7) 1.0000 1.0000 1.0000 0.0626 0.5790 0. 0647 0.8947 0.1612 < .0001 < .0001 0.0001 < .0001 0.0049 0.0008
146 App endix A 1 C ontinued Plant 1 2 3 4 5 6 7 8 9 10 11 12 Panicum hemitomon (8) 1.0000 1.0000 0.9636 0.7495 < .0001 < .0001 0.0479 < .0001 0.3225 0.0004 Potamogeoton illinoensis (9) 1.0000 1.0000 < .0001 < .0001 0.0827 < .0001 0.2767 0.0002 Schoenoplectus californicus (10) 1.0000 1.0000 1.0000 0.0033 0.0927 0.0011 0.2626 Colacasia esculanta (11) 1.0000 1.0000 0.4610 0.0054 Zizania Aquatica (12) 1.0000 1.0000
147 APPENDIX B P VALUE COMPARISONS OF STANDARDIZED CONSUMPTION FEEDING RATES Appendix B 1 P value comparisons of standardized (g consumed/ gram o f snail weight) feeding rates. Top p values are adult P. canali culata feedin g rates between plant species. Middle p values are adult P. maculata feeding rates. Bottom p values are juvenile P. maculata feeding rates between plant species. The number in parenthesis under the plant name represents the same plant species in the top row. Plant 1 2 3 4 5 6 7 8 9 10 11 12 Hydrilla verticillata (1) 1.0000 1.0000 1.0000 < .0001 < .0001 0.0015 < .0001 < .0001 0.0015 < .0001 < .0001 < .0001 0.0049 0.0010 0.0907 < .0001 < .0001 0.8832 <. 0001 < .0001 < .0001 < .0001 0.0418 < .0001 < .0001 < .0001 < .0001 0.4245 < .0001 0.6392 < .0001 0.0288 Najas guadalupensis (2) 1.0000 1.0000 1.0000 < .0001 < .0001 0.9914 0.0380 0.4489 0.1242 < .0001 < .0001 0.1312 0.4410 0.0218 0.0024 0.1750 0.5085 0. 4068 0.0115 0.2443 0.0046 0.0137 0.5328 < .0001 0.0165 < .0001 0.0065 0.0069 0.3133 Vallisneria americana (3) 1.0000 1.0000 1.0000 < .0001 < .0001 0.1216 0.3009 0.0592 0.1340 < .0001 0.0036 0.0025 < .0001 < .0001 0.4007 0.0074 0.2486 < .0001 0.0133 < .0001 0.9581 0.0170 0.7422 0.0067 < .0001 0.3164 Nuphar Advena (4) 1.0000 1.0000 1.0000 < .0001 < .0001 0.0024 0.1885 0.1230 < .0001 0.4657 0.1565 0.4778 0.0755 0.0071 0.0004 0.3493 0.0073 < .0001 < .0001 < .0001 < .0001 0.0006 0.0110 Bacopa caroliniana (5) 1.0000 1.0000 1.0000 < .0001 < .0001 0.1222 < .0001 < .0001 0.0197 < .0001 0.7291 < .0001 < .0001 < .0001 0.0666 0.3705 0.1188 0.2206 < .0001 0.6182 Panicu m Repens (6) 1.0000 1.0000 1.0000 0.5562 0.0032 0.0001 0.8133 0.0588 < .0001 < .0001 0.1642 0.0031 0.5144 0.0012 0.7474 < .0001 0.0413 Sagittaria lancifolia L. (7) 1.0000 1.0000 1.0000 0.0015 0.0466 0.0293 0.1003 0.0 487 < .0001 0.0013 < .0001 0.0004 0.0406 0.0659
148 Appendix B 1 Continu ed Plant 1 2 3 4 5 6 7 8 9 10 11 12 Panicum hemitomon (8) 1.0000 1.0000 1.0000 0.4725 0.3015 0.1384 0.1037 0.2803 <.0001 0.2422 <.0001 Potamogeoton ill inoensis (9) 1.0000 1.0000 < .0001 < .0001 < .0001 < .0001 0.8942 0.0005 Schoenoplectus californicus (10) 1.0000 1.0000 1.0000 0.7822 0.7413 < .0001 0.1637 Colacasia esculanta (11) 1.0000 1.0000 < .0001 0.0852 Zizania Aquatica (12) 1.0000 1.0000
149 LIST OF REFERENCES Aizaki, K., and Y. Yusa. 2009. Field observations of the alarm response to crushed conspecifics in the freshwater snail Pomacea canaliculata : effects of habitat, vegetation, and body si ze. Journal of Ethology 27(1):175 180. Albrecht, E.A., N.B. Carreno, and C.Vazquez. 1999. A quantitative study of environmental factors influencing the seasonal onset of reproductive behaviour in the south American apple snail Pomacea canaliculata (Gast ropoda:Ampullariidae). J ournal of Mollu scan Studies 65:241 250. Alderson EG 1925. Studies in Ampullaria. W. Heffer and Sons, Cambridge, England. Aldridge, D.W. 1983. Physiological ecology of freshwater prosobranchs. I n, W.D. Russell Hunter (ed.), The M ollusca, Vol. 6, Ecology Academic Press, New York. Allison, P. 2010. Survival analysis using SAS: A practical guide, 2nd ed. Cary, NC: SAS Institute. American Society for Testing and Materials (ASTM). 2002. E729 96 Standard guide for conducting acute to xicity test materials with fishes, macroinvertebrates, and amphibians. ASTM International, Pennsylvania, USA. American Society for Testing and Materials (ASTM). 2002. E1022 94 Standard guide for conducting bioconcetration tests with fishes and saltwater b ivalve mollsks ASTM International, Pennsylvania, USA. Ansart, A., and P. Vernon. 2003. Cold hardiness in mollusks. Acta Oecologica Internat ional Journal of Ecology 24(2): 95 102. Apple Snail Website (ASE) Re trieved September 2006 http://www.applesnail.net/content/species/pomacea_canaliculata.htm Arnold, T.E. 2011. Using oxygen isotopes to determine growth patterns in the invasive gastropod Pomacea insularum M.S. Thesis U niversity of Florida. Barnes M A R.K. Fordham, R.L Burks, and J.J.Hand. 2008. Fecundity of the exotic applesnail, Pomacea insularum Journal of the North American Benthological Society 27: 738 745 Barbier, E. B., M. C. Acreman, and D. Knowler. 1996. E conomic valuation of wetlands: a guide for policy makers and planners. Ramsar Convention Bureau, Gland, S witzerland. Bemvenuti, C.E., Rosa Filho, J.S., Elliott, M. 2003. Changes in soft bottom macrobenthic assemblages after a sulfuric acid spill in the Ri o Grande harbor (RS, Brazil). Brazilian Journal of Biology 63: 183 194.
150 Berezina, N.A. 2001. Influence of ambient pH on freshwater invertebrates under experimental conditions. Russian Journal of Ecology 32: 343 351. Bernatis, J.L. 2013. Ecology and Impacts of Nonindigenous Apple Snails (Ampullariidae: Pomacea ) in Florida Florida Fish and Wildlife Conservation Commission Final Report. Boland, B.B., M Meerhoff, C. Fosalba, N. Maz zeo, M.A. Barnes, and R.L. Burks. 2008. Juvenile snails, adult appetites: Con trasting resource consumption between two species of applesnails ( Pomacea ) J ournal of Molluscan Studies 74: 47 54. Brenner, M., M.W. Binford, and E.S. Deevey (1990). In Ecosystems of Florida Chapter Central Florida Press, Orlando, FL. Brown, K.M. 2001. Mollusca. Gastropoda. I n, J. H. Thorp, A. P. Covich (eds.), Ecology and Classification of North American Freshwater Invertebrates Academic Press, New York. Brown, K.M. 1982. Resource overlap and co mpetition in snail ponds: An experimental analysis. Ecology 63(2) 412 422. Burks, R.L., S.A. Hensley, and C.H. Kyle. 2011. Quite the appetite: Juvenile island apple snails ( Pomacea insularum ) survive consuming only exotic invasive plants. Journal of Moll uscan Studies 77:423 428. Burky, A.J., J. Pacheco, and E. Pereyra 1972 Temper a ture, water, and respiratory regimes of an amphibious snail, Pomacea urceus (Mller), from the Venezuelan savannah. Biological Bulletin 143 :304 316. Burlakova, L.E., A.Y. Ka ratayev, D.E. Padilla, L.D. Cartwright, and D.N. Hollas. 2009. Wetland restoration and invasive species: Apple snail ( Pomacea insularum ) feeding on native and invasive aquatic plan ts. Restoration Ecology, 17(3): 433 440. Byers J E W.G. McDowell, S.R. Dod d, R.S. Haynie, L.M. Pintor and S.B. Wilde. 2013. Climate and pH p redict the potential r ange of the i nvasive a pple s nail ( Pomacea insularum ) in the s outheastern United States. PLoS ONE 8(2): e56812. doi:10.1371/journal.pone.0056812 California Department of Agriculture (CDA ). Retrieved September 2006. http://www.acgov.org/cda/awm/resources/channeledapplesnail.pdf#search='chan neled%20apple%20sn ail
151 Calumpang, S.M.F., M.J.B. Medina, A.W. Tejada, and J.R. Medina. 1995. Environmental impact of two molluscicides: Niclosamide and metaldehyde in a rice paddy ecosystem. Bulletin of E nvironmental Contamination and Toxicology 55: 494 501. C arlsson N.L and C. Brnmark 2006. Size dependent effects of an invasive h erbivorous snail ( Pomacea canaliculata ) on macrophytes and periphyton in Asian wetlands. Freshwater Biology 51:695 704. Carlsson, N.O.L., C. Brnmark, and H. Lars Anders. 2004. Invading her bivory: The golden apple snail alters ecosystem functioning in Asian wetlands. Ecology 85(6): 1575 1580. the introduced golden apple snail ( Pomacea canaliculata ) in Lao P. D.R. Biological Invasions 7: 233 241. Cattau, C.E., J. Martin, and W.M. Kitchens. 2010. Effects of an exotic prey species on a native specialist: Example of the snail kite. Biological Conservation 143:513 520. Cagauan, A.G., and R.C. Joshi. 2002. Freshwat er fishes as biological control for golden apple snail Pomacea canaliculata Lam. in rice fields. Terminal Report Philippine Rice Research Institute and Freshwater Aquaculture Center and College of Fisheries Central Luzon State University Sci ence City of Mu oz, Nueva Ecija. Cazzaniga, N J 1990 Sexual dimorphism in Pomacea canaliculata (Gastropoda: Am pullariidae). Veliger, 33 (4): 390 394. Cazzaniga N J 2002. Old species and new concepts in the taxonomy of Pomacea (Gastropoda: Ampullariidae) Workshop: Ampullari BIOCELL 26(1): 71 81 Chanyapate, C. 1997. The golden apple snail, Pomacea canaliculata research and management in Thailand. Ecology and Management of the Golden Apple Snail in Rice Prod Pitsanulok. Thailand. Chao, D., C.C. Lin, and Y.A. Chen. 1987. Studies on growth and distribution of Angiostrongylus cantonensis Larvae in Amplullarium canaliculatus Southeast Asian Journal of Tropical Medici ne and Publi c Health 18(2). Connor, S.L., C.M. Pomory, and P.C. Darby. 2008. Density effects of native and exotic snails on growth in juvenile apple snails Pomacea paludosa (Gastropoda: Ampullarii dae): a laboratory experiment. J ournal of Molluscan Studies 74: 355 362.
152 d natural capital. Nature, 387:253 260. Costil, K., Dussart, G. B. J., Daguzan, J. 2001. Biodiversity of aquatic gastropods in the Mont St Michel basin (France) in relation to salinity and drying of habitats. Biodiversity and Conservation 10: 1 18. Cowie, R.H. 2002. Apple snails (Ampullariidae) as agricultural pests: their biology, impacts and management. Pages 145 192 in, G. M. Barker (Ed.), Molluscs as Crop Pests CABI International, Wa llingford and New York. Daniel, W.W. and Cross, C. L. 2013. Biostatistics: A foundation for analysis in the health sciences. New York, NY: Wiley. Darby, P. 2005. Apple snail abundance in Snail Kite foraging sites on Lake Okeechobee in 2005. Annual Report for Florida Fish and Wildlife Conservation Commission. Darby, P.C., R.E. Bennetts, and H.F. Percival. 20 08. Dry down impacts on apple snail ( Pomacea paludosa ) demography: implications for wetl and management. Wetlands 28(1): 204 214. De nson, D. and G. Eby 2004. Invasive exotic snails in Lake Brantley, Seminole County. Fla. Dept. Environ. Protection, Central District, Orlando, FL. http://www.dep.state.fl.us/central/ [accessed 09 13 2005]. de la Cruz, M.S., and R.C. Joshi. 2001. Efficacy of commercial molluscicide formulations against the golden apple snail Pomacea canaliculata (Lamarck). Philippin e Agricu ltural Scientist 84(1): 51 55. de la Cruz, M.S., R.C. Joshi, and E.C. Martin. 2000. Potential effects of commercial molluscicides used in controlling golden apple snails on the native snail Vivipara costata (Quoy & Gaimard). The Philippine Entomologist 14( 2): 149 157. Estebenet, A.L. 1995. Food and feeding in Pomacea canaliculata (Gastropo da: Ampullariidae). Veliger 38: 277 283. Estebenet, A.L., and N.J. Cazzaniga. 1990. Effect of short term exposure to copper on survival of apple snails in an integrated co ntrol program. Journal of Aquatic Plant Management 28: 103 105. Estebenet, A.L., and N.J. Cazzaniga. 1992. Growth and demosgraphy of Pomacea canaliculata (Gastropoda: Ampullariidae) under laboratory condit ions. Malacological Review 25: 1 12.
153 Estebenet, A. L., and N.J. Cazzaniga 1998. Sex related differential growth in Pomacea canaliculata (Gastropoda: Ampullariidae). J ournal of Molluscan Studies 64: 119 123. Estebenet AL, and P.R. Martin. 2002. Pomacea canaliculata (Gastropoda: A mpullariidae): Life histo ry traits and the ir plasticity. Biocell 26 (1): 83 89 Estebenet A.L., and P.R.Martin. 2003.Shell interpopulation variation and its origin in Pomacea canaliculata (Gastropoda : Ampullariidae) from Southern Pampas, Argentina. J ournal of Molluscan Studies 6 9: 301 310 Estebenet, A.L., Martin, P.R., and Burela, S. 2006. Conchological variation in Pomacea canaliculata and other South American Ampullariidae (Caenogastropoda, Architaenioglossa) BIOCELL 30 (2): 329 335. Estoy, G. F., Jr., Y. Yusa, T. Wada, H. Sakurai and K. Tsuchida 2002a Size and age at first copulation and spawning of the apple snail, Pomacea canaliculata (Gastropoda: Ampullariidae). Appl ied Entomol ogical Zool ogy 37:199 205. Estoy, G. F., Jr., Y. Yusa, T. Wada, H. Sakurai and K. Tsuchida 2002b Effects of food availability and age on the reproductive effort of the apple snail, Pomacea canaliculata (Lamarck) (Gastropoda: Ampullariidae) Appl ied Entomol ogical Zool ogy 37(4):543 550 European Commission of the European Union (ECEU) 2012. C omm ission Implememnting Decision of 8 November 2012 as regards to measures to prevent the introduction into and the spread within the Union of the genus Pomacea (Perry). Official Journal of the European Union notified under document C (2012) 7803. Fang L., P.K. W ong, P.K., Lin L., C. L an and J.W. Q iu 2009. Impact of invasive apple snails in Hong Kong on wetland macrophytes, nutrients, phytoplankton and filamentous algae. Freshwater Biology 55:1191 1204. Frank, W. Retrieved 2006. Pomacea canaliculat a (Lamarck, 1819) channeled applesnail in Duval County, Florida Jacksonville Shell Club, Jacksonville, FL. http://www.jaxshells.org Freiberg, M.W. and Hazelwood, D.H. 1977. Oxygen consumption of two amphibious sna ils: Pomacea paludosa and Marisa cornuarietis (Prosobranchia: Ampullariidae). Malacologia 16: 541 548. Gettys, L.A., W.T. Haller, C.R. Mudge, and T.J. Koschnick. 2008. Effect of temperature and feeding preference on submerged plants by the island apple sn ail, Pomacea maculata (d'Orbigny, 1839) (Ampu llariidae). Veliger 50(3): 248 254.
154 Glass, N.H. and P.C. Darby. 2008. The effect of calcium and pH on Florida apple snail, Pomacea paludosa (Gastropoda: Ampullariidae), shell growth and crush weight. Aquatic Ec ology doi 10.1007/s10452 008 9226 3. Gleason, P.J., P. Stone, P. Rhoads, S.M. Davis, M. Zaffke, and L. Harris. 1975. The impact of agricultural runoff on the Everglades marsh located in the conservation areas of the central and southern Florida flood con trol district. Resource Planning Department of the South Florida Water Management District, West Palm Beach. Guppy, M. and P. Withers 1999. Metabolic depression in animals: physiological perspectives and biochemical generalizations. Biological Reviews 74 : 1 40. Hall, R.J., Likens, G.E., Fiance, S.B., Hendrey, G.R. 1980. Experimental acidification of a stream in the Hubbard Brook experimental forest, New Hampshire. Ecology 61:976 989. Halwart, M. 1994. The golden apple snail Pomacea canaliculata in Asi an rice farming systems: Present impact and future threat International Journal of Pest Management, 40: 199 206. Halwart, M., M.C. Virayi, and G. Kaule. 1998. Cyprinus carpio and Oreochromis niloticus as biological control agents of the golden apple snail Pomacea canaliculata : effects of predator size, prey size and prey density. Asian Fisheries Science 11 : 31 42. Hanning, G. 1979. Aspects of reproduction in Pomacea paludosa (Mesogastropoda: Pilidae). Master of Science Thesis, Florida State University. Hart, B.T.m P. Bailey, R. Edwards, K. Hortel, K.James, A. McMahon, C. Meredith, and K. Swaddling. 1991. A review of the salt sensitivity of the Australian freshwater biota. Hydrobiologia 210:105 144. Harvey, H. H., McArdle, J.M. 1986. Composition of the benthos in relation to pH in the LaCloche lakes. Water, Air & Soil Pollution 30: 529 536. Hayes, K.A., R.H. Cowie, and S.C. Thiengo. 2009. A global phylogeny of apple snails: Gondwanan origin, generic replationships, and the influence outgroup choice (Ca enogastropoda: Ampullariidae). Biological Jour nal of the Linnean Society, 98: 61 76. Hayes K A R.H. Cowie, S.C. Thiengo, and E.E. Strong 2012 Comparing apples with apples: C larifying the identities of two highly invasive neotropical Ampullariidae ( Ca enogastropoda). Zoological Journal of t he Linnean Society 166 : 723 753
155 Hendarsih, H., S. Suriapermana, A. Fagi, and I Manwan. 1994. Potential of fish in rice fish culture as a biological control agent of rice pests. Role of fish in enhancing ricefield eco logy and in integrated pest management. ICLARM Conference Proceedings 43: 32 33. Herrman, J., E. Degerman, A. Gerhardt, C. Johansson, P. Lindgell, and I. Muniz. 1993. Acid Stress Effects on Stream Biology. Ambio 22(5): 298 307. Horn K C S.D. Johnson, K.M. Boles A. Moore, E. Siemann, and C.A. Gabler 2008 Factors affecting hatchling success of golden apple snail eggs: effects of water immersion and cannibalism. Wetl ands 28(2): 544 549 Howells, R.G. 2001. History and status of applesnail ( Pomacea spp. ) introductions in Texas. Management Data Series No. 183. Plant Pest Diagnostics: Entomology, California Department of Agriculture, Sacramento, CA. Hunter RD. 1989 Effects of low pH and low calcium concentration on the pulmonate snail Planorbella trivolv is : a laboratory study. Canadian Journal of Zoology 68:1578 1583. Hurdle, M.T. 1974. Life history studies and habitat requirements of the apple snail at Lake Woodruff National Wildlife Refuge. Proceedings of the Southeast Association of Game and Fish Commissions 27:215 224. Indiana Department of Natural Resources (IDNR) Retrieved September 2006. http ://www.in.gov/dnr/invasivespecies/CHANNELED%20APPLE%20SNAIL.pdf#s earch='channeled%20apple%20snail' Ito, K. Environmental factors influencing overwintering success of the golden apple snail, Pomacea canaliculata (Gastropoda: Ampullariidae), in the norther nmost population of Japan. Applied Entomological Zoology 37(4):655 661. Jacobsen, R., Forbes, V.E. 1997. Clonal variation in life history traits and feeding rates in the gastropod, Potamopyrgus antipodarum : performance across a salinity gradient. Functi onal Ecology 11: 260 267. Jordan, P.J. and L.E. Deaton. 1999. Osmotic regulation and salinity tolerance in the freshwater snail Pomacea bridgesi and the freshwater clam Lampsilis teres Comparative Biochemistry and Physiology Part A 122 :199 205. Joshi R C de la Cruz, M.S., E.C. Martin EC, J.C. Cabbigat, R.F. Bahatan, A.D. Bahatan, J. Choy Awon, N.P. Chilagan, and A.B. Cayong 2001 Current status of the golden apple snail in the Ifugao Rice Terraces, Philippines. J ournal of Sustain able Agric ulture 18 (2/3): 71 90
156 Karunaratne, L.B., P.C. Darby, and R.E. Bennetts. 2006. The effects of wetland habitat structure on Florida apple snail density. Wetlands 26(4): 1143 1150. Kleinbaum, D. G., Klein, M. 2005. Survival analysis: A self learning text. New York, N Y: Springer. Kondo, A., and Tanaka, F. (1989). An Experimental Study of Predation by the Larvae of the Firefly, Luciola lateralis, Motschulsky, (Coleoptera: Lampyridae) on the Apple Snail, Pomacea canaliculata, Lamarck, (Mesogastropoda: Pilidae). A pplie d Ent omology and Zool ogy 33 :211 216. Kushlan, J. 1975. Population changes of the apple snail, Pomacea paludosa in the southern Evergaldes. The Nautilus 89(1) : 21 23. Lacanilao, F. 1990. Reproduction of the golden apple snail (Ampullaridae): Egg masses, hatching, and incubation. Philippine Journal of Science 119(2). Lach, L., D.K. Britton, R.J. Rundell, and R.H. Cowie. 2000. Food preference and reproductive plasticity in an invasive freshwater snail. Biol ogical Invasions 2: 279 288. Litsinger J A D.B. Estano 1993 Management of the golden apple snail Pomacea canalicu l a ta (Lamarck) in rice. Crop Prot ection 12(5): 363 370 Little, C. (1968) Aestivation and ionic regulation in two species of Pomacea (Gastropoda : Prosobranchia). Journal of Experimental Bi ology 48 :569 585 Lockwoo d, J.L., M.F. Hoopes, and M.P. Marchetti. 2007. Invasion Ecology. Blackwe l l Publishing, Malden, MA. Lodge, D.M., K.M Brown, S.P. Klosiewski, R.A. Stein, A.P. Covich, B.K. Leather, and C. Bronmark. 1987. Distribution of Freshwater Snails: Spatial scale and the relative importance of physiochemical and biotic factors. American Malacological Bulletin 5( 1): 73 84. Lodge, D. M. 1991. Herbivory on freshwater macrophytes. Aquatic Botany 41:195 224. Lodge, D. M., G. Cronin, E. Van Donk, and A. J. Froelich. 1998. Impact of herbivory on plant standing crop: comparisons among biomes, between vascular and nonvascular plants, and among freshwater herbivore taxa. Ecological Studies: Analysis and Synthesis 131: 149 174. Lowe, S.J., M. Browne, an d S. Boudjelas. 2000 100 of the World's Worst Invasive Alien Species. IUCN/SSC Invasive Species Specialist Group (ISSG), Auckland, NZ.
157 Mackie, G.L. 1987. Effects of acidifying environments on freshwater mollusks in southern Ontario, Canada. American Ma lacological Bulletin 5: 31 39. Maini P N and B. Morallo Rejesus 1993 Antifeedant activity of the crude and formulated products from Azadirachta Indica to golden snail (Pomacea spp.). Philipp ine J ournal of Sci ence 122(2):177 128 Marchetti, M.P., P. B, Moyle, and R. Levine. 2004. Alien fishes in California watersheds: characteristics of non native fishes across invasion stages in Californ ia. Ecological Applications 14: 587 596. Martin P.R., and A.L. Estebenet. 2002. Interpopulation variation in life history traits of Pomacea canaliculata (Gastropoda: Ampullariidae) in Southwestern Buenos Aires province Argentina. Malacologia, 44(1): 153 163. Martin, P.R., A.L. Estebenet, and N.J. Cazzaniga. 2001. Factors affecting the distribution of Pomacea canalic ulata (Gastropoda: Ampullariidae) along its southernmost natur al limit. Malacologia 43(1 2): 13 23. Matsukura, K., M. Okuda, N.J. Cassaniga, and T.Wada. 2013. Genetic exchange between two freshwater apple snails, Pomacea canaliculata and Pomacea maculata invading East and Southeast Asia. Biological Invasions, doi 10.1007/s10530 013 0431 1. Matsukura, K.H., Tsumuki, Y. Izumi, and T. Wada. 2009. Physiological response to low temperature in the freshwater apple snail, Pomacea canaliculata (Gastropoda: Ampul lariidae). Jouranl of Experimental Biology 212: 2558 2563. Matsukura, K. H. Tsumuki, Y. Izumi, and T. Wada. 2008. Changes in chemical components in the freshwater apple snail Pomacea canaliculata (Gastropoda: Ampullariidae), in relation to the developmen t of its cold hardiness. Cryobiology 56: 131 137. McClary, A. 1964. Surface inspiration and ciliary feeding in Pomacea paludosa (Prosobranchia: Mesogastropoda: Ampullariidae). Malacologia 2(1):87 104. McKillop, W.B., Harrison, A.D. 1972. Distribution of aquatic gastropods across an interface between the Canadian Shield and limestone formations. Canadian Journal of Zoology 50: 1433 1445. McMahon, R.F. 1983. Physiological ecology of freshwater pulmonates. In, W.D. Russell Hunter (Ed.), The Mollusca, Volume 6, Ecology Academic Press, London: 359 430.
158 Meenakshi, V.R. 1964. Aestivation in the Indian apple snail Pila I, adaptation in natural and experimental conditions. Comparative Biochemistry and Physiology 11: 379 386. Mitsch, W. J. and J. G. Gosselink. 2007. Wetlands, 4th edition. W iley and Sons, New Jersey, USA. Morrison, W.E., M.E. Hay. 2011. Feeding and growth of native, invasive and non invasive alien apple snails (Ampullariidae) in the United States: invasives eat more and grow more. Biological Inv asions 13: 945 955. Nakamura, R. 200 6. Taro production hits record low Hawaii Taro Annual Report. U.S. Department of Agriculture, National Agricultural Statistics Service Honolulu, Hawaii Naylor, R. 1996. Invasion in agriculture: assessing the cost of the golden ap ple snail in Asia. Ambio 25(7): 443 448. control. In L.A. Gettys, Haller, W.T., and Bellaud, M. (Eds.), Biology and Control of Aquatic Plants A Best Managem ent Practices Handbook. Nduku, W.K. and A.D. Harrison. 1976. Calcium as a limiting factor in the biology of Biomphalaria pfeifferi (Krauss) (Gastropoda: Planorbidae). Hydrobiologia 49:143 170. O kland, J. 1992. Effects of acidic water on freshwater snail s: results from a study of 1000 lakes throughout Norway. Environmental Pollution 78:127 130. Okuma, M., K. Tanaka, and S. Sudo. 1994. Weed control method using apple snail ( Pomacea canaliculata ) in paddy f ields. Weed Research, Japan 39: 114 119. Perera G Yong Cong M, Ferrer JR (1989) Influencia de la vegetation acuatica en la distribucion de los moluscos fluviales de la Isla de la Juventud. Revista Cub ana de Medicina Tropical 42:182 191. Piscart, C., P. Usseglio Polatera, J.C. Moreteau, and J.N. Beise l. 2006. The role of salinity in the selection of biological traits of freshwater invertebrates. Archiv fur Hydrobiologie 166( 2): 185 198. Pizani N V A.L. Estebenet, P.R. Martin 2005 Effects of submersion and aerial exposure on clutches and hatchling s of Pomacea canaliculata (Gastropoda: Ampullariidae). Am erican Malacol ogical Bull etin 20(1 2):55 63 Prescott, D.R.C., Curteanu, M.M. 2004. Survey of aquatic gastropods in the central Parkland subregion of Alberta. Alberta Species at Risk Report 92:1 50.
159 Prosser, C.L. and Heath J. E. 19 91. Temperature. In, C.L. Prosser (ed.), Environmental and Metabolic Animal Physiology John Wiley and Sons Inc, New York. Q iu J. W. and K.L. K wong 2009. Effects of macrophytes on feeding and life history traits of the invasive apple snail ( Pomacea canaliculata ). Freshwater Biology 54:1720 1730. Ramakrishnan V. 2007. Salinity, pH, temperature, desiccation and hypoxia tolerance in the invasive freshwater snail Pomacea insularum Ph.D. Dissertation. Texas, US: University of Texas at Arlington. Rawlings T A K.A. Hayes, R .H. Cowie, and T.M. Collins 2007 The identity, distribution, and impacts of non native apple snails in the continental United States. BMC Evol utionary Biology 7: 97 Remane, A., and C. Schlieper. 1971. Biology of Brackish Water John Wiley and Sons, New York. Santos, C.A.Z. and Mendes, E.G. 1981. Oxygen consumption of the amphibious snail Pomacea lineata : influence of weight, sex and environments. Comparative Biochemistry and Physiology A 69:595 598. Saveanu, L. P.R. and P.R. Martin. 2013. Pedal surface collecting as an alternative feeding mechanism of the invasive apple snail Pomacea canaliculata (Caenogastropoda: Ampullariidae). Journal of Molluscan Studies 79:11 18. Scott T.M., G.H. Means, R.C. Me ans, and R.P. Meegan. 2002. First magnitude springs of Florida. Open file report no. 85, Florida Department of Environmental Protection, Florida Geological Survey, Tallahassee Seuffert, M.E.,Marti n, P.R.,2009a.Influence of temperature, size and sex on aer ial respiration on Pomacea canaliculata (Gastropoda: Ampullariidae)from southern Pampas, Argentina Malacologia 51:191 200. Seuffert, M.E.,Marti n, P.R.,2009b.Dependenceonaerialrespirationandits influence on micro distribution in the invasive freshwater snail Pomacea canaliculata (Caenogastropoda, Ampullariidae). Biological Invasions doi:10.1007/ s10530 009 9582 5. Seuffert, M. E ., S. Burela, and P.R. Martin. 2010. Influence of water temperature on the activity of the freshwater snail Pomacea canaliculat a ( Caenogastropoda: Ampullariidae) at its southern most limit (Southern Pampas, Argentina) Journal of Thermal Biology 35: 77 84
160 Seuffert, M.E., L. Saveanu, and P.R. Martin. 2012. Threshold temperatures and degree day estimates for embryonic development of the invasive apple snail Pomacea canaliculata (Caenogastropoda: Amp ullariidae). Malacologia 55(2): 209 217. S in T 2003 Damage potential of the golden apple snail Pomacea canaliculata (Lamarck) in irrigated rice and its control by cultural approaches. Int ernational J ournal of Pest Manag ement 49(1):49 55 Snyder, N.F.R., and H.W. Kale, 1983. Mollusk predation by snail kite s in Colombia. The Auk, 100(1): 93 97. Somero, G.N. 1995. Proteins and temperature, Annual Review of Physiology 57: 43 68. State of 2000 2013/14. http://floridafiscalportal.state.fl.us/ Stein RA, Goodman CG, Marschall EA (1984) Using time and energetic measur es of cost in estimating prey value for fish predators. Ecology 65:702 715. Sunnell, L.A., and P.L. Healey 1985. Distribution of Calcium Oxalate Crystal Idioblasts in Lea ves of Taro (Colocasia esculent a ) American Journal of Botany 72 (12): 1854 1860 Swan n, L. 1993. Water quality water sources used in aquaculture. Indiana Sea Grant #IL IN SG Fs 93 1. Aquaculture Extension Fact Sheet AS 486. Tamburi NE, Martn PR. 2011. Effects of food availability on reproductive output, offspring quality and reproductive efficiency in the apple snail Pomacea canaliculata Biological Invasions 13: 2351 2360. Tanaka, K., T. Watanabe, H. Higuchi, K. Miyamoto, Y. Yusa, T. Kiyonaga, H. Kiyota, Y. Suzuki, and T. Wada. 1999. Density dependent growth and reproduction of the apple snail, Pomacea canaliculata : a density manipulation experiment in a paddy field. Res earch in Popul ation Ecol ogy 41:253 262. Thompson, F.G. 1997. Pomacea canaliculata (Lamaarck 1822) (Gastropoda, Prosobranchia, Pilidae): A freshwater snail introduced in to Florida, U.S.A. Malacological Review 30: 91. Thompson F.G. 2000. An Identification Manual for the Freshwater Snails of F lorida. Walkerana 10(23). United States Department of Agriculture. 2006. Florida aquaculture sales total $75 million in 2005. Na tional Agricultural Statistics Service and Florida Department of Agricult ure and Consumer Services.
161 United States Geological Survey (USGS ) Retrieved 2013. Concentration of hardness as calcium carbonate in milligrams per liter. http://ga.water.usgs.gov/edu/characteristics.html Vanderbo, O. and Vanpuymb, S. 1966. Calcium metabolism in freshwater mollusc Quantitative importance of water and food supple for calcium during growth. Nature 210 do i: 10.1038/210791a0 Wada T 2004 Strategies for controlling the apple snail Pomacea canaliculata (Lamarck) (Gastropoda: Ampullariidae) in Japanese direct sown paddy fields. J apan Agric ultural Res earch Q uarterly 38(2):75 80 Wang. T. 2001. Physiologi cal consequences of feeding in animals. Comparative Biochemistry and Physiology Part A 128 : 395 396. Watson A.M., and S.J. Ormerod 2004 The distribution of three uncommon freshwater gastropods in drainage ditches of British grazing marshes Biological Conservation 118:455 466 White, M., M. Chejlava, B. Fried, and J. Sherma. 2007. The concentration of calcium carbonate i n shells of freshwater snails. American Malacological Bulletin 22(1 2): 139 142 doi: 10.4003/0740 2783 22.1.139. Wilbur, K.M. 1964. Shell formation and regeneration. Pages 243 282 in K.M. Wilbur and C.M. Yonge (eds). Physiology of Mollusca: Volume I Aca demic Press New York and London. Wu D. C J.C. Yua, B.H. Chena C.Y. Lina, and W.H. Ko 2005. Inhibition of egg hatching with apple waxsolvent as a novel method for controlling golden apple snail ( Pomacea canaliculata ). Crop Prot ection 24:483 486 Wunde rlin, R.P. and Hansen, B.F. 2003. Guide to the Vascular Plants of Florida. University of Florida Pres s, Gainesville, FL, USA. Yamanishi Y K. Yoshida, N. Fujimori and Y. Yusa 2012 Predator driven biotic resistance and propagule pressure regulate the invasive apple snail Pomacea canaliculata in Japan. Biol ogical Invasions 14(7):1343 1352 Young, J.O. 1975. Preliminary field and laboratory studies on survival and spawning of several species of gastropoda in calcium poor and calcium rich waters. Proceed ings of the Malacological Society of London 41:429 437. Yu, A.M. 2005. Golden kuhol eats weeds in transplanted rice Ecoport ID 123. Ecoport Foundation, Inc. Institute on Governance, Ottawa, Ontario, Canada.
162 Yusa Y 2001 Predation on eggs of the appl e snail Pomacea canliculata (Gastropoda: Ampullariidae) by the fire ant Solenopsis geminate J ournal of Moll uscan Stud ies 67: 275 279 Yusa, Y. 2004. Inheritance of colour polymorphism and the pattern of sperm competition in the apple snail Pomacea canaliculata (Gastropoda: Ampullariidae) J ournal of Molluscan Studies 70: 43 48. Yusa Y N. Sugiura and T. Wada 2006 Predat ory potential of freshwater animals on an invasive agricultural pest, the apple snail Pomacea canaliculata (Gastropoda: Ampullariidae), in southern Japan. Biol ogical Invasions 8: 137 147 Yusa Y and T. Wada 1999 Impact of the introduction of apple sn ails and their control in Japan. Naga 22(3): 117 128 Zedler, J. B., and S. Kercher. 2005. Wetland resources: status, trends, ecosystem services, and restorability. Annual Review of Environment and Resources 30:39 74.
163 BIOGRAPHICAL SKETCH Jennifer grew up i n Jacksonville, Florida and graduated from The Bolles School. Her background is diverse with a Bachelor of Science in Exercise Science ( 1994) and Master of Science in e xercise p h ysiology with a concentration in n utrition (1996), both from Fort Hays State U niversity in Hays, KS. After graduation she owned a restaurant, worked with developmentally disabled adults at a not for profit company, was a corporate trainer and taught high school P.E. A second Master of Science was earn ed in o rganismal Biology (2005 ) from the University of Nevada Las Vegas. In 2005 Jennifer and Aquatic Sciences for her doctoral research. In 2008, she changed departments to the School of Interdisciplinary Ecology in an e ffort to broaden academic and professional exp erience She was employed full time with the Florida Fish and Wildlife Conservation Commission, while pursuing her degree. Jennifer earned her Doctor of Philosophy in 201 4 and continues to work for Florida Fish and Wildlife Conservation Commission.