Ecology of Freshwater Fish 2001: 10: 105–110 Copyright C Munksgaard 2001 Printed in Denmark Á All rights reserved ISSN 0906-6691 Parasitism, water temperature and life history characteristics of the freshwater fish Gobiomorphus breviceps Stokell (Eleotridae) Hamilton WJ, Poulin R. Parasitism, water temperature and life history W. J. Hamilton, R. Poulin characteristics of the freshwater fish Gobiomorphus breviceps Stokell Department of Zoology, University of Otago, (Eleotridae). Dunedin, New Zealand Ecology of Freshwater Fish 2001: 10: 105–110. C Munksgaard, 2001 Abstract – Life history traits can vary among populations of the same fish species in response to selective pressures exerted by the local environ- ment. Here we used inter-population comparisons to investigate the ef- fects of parasitism and variability in water temperature on life history traits in a New Zealand freshwater fish, the upland bully. Although age and size at maturity varied among populations, they were not significantly re- lated to either temperature regime or average parasite load. Trade-offs between egg size and clutch size were only apparent in one of the seven populations when populations were analyzed separately but became clear when all data were pooled. However, neither average population clutch Key words: age at maturity; clutch size; egg size; size or egg size was related to the population’s mean parasite load or the inter-population comparison; trade-offs; local temperature regime. Although the latter two parameters may in- trematodes fluence other life history variables, such as the number of clutches pro- Robert Poulin, Department of Zoology, University duced per season and the partitioning of eggs among clutches, there was of Otago, P.O. Box 56, Dunedin, New Zealand; no evidence that they influenced life history strategies of the upland bully e-mail: robert.poulin/stonebow.otago.ac.nz populations investigated here. Accepted for publication November 11, 2000 Un resumen en espan˜ol se incluye detra´s del texto principal de este artı´culo. Variation in life history characteristics among In fish, parasites are known to lower physical con- populations of the same species has been attri- dition or to reduce growth and reproductive out- buted to factors ranging from latitudinal clines put (e.g., Pennycuick 1971; McPhail & Peacock (Fleming & Gross 1990; Vøllestad & Abe´e-Lund 1983; Lemly & Esch 1984; Poulin 1993). These 1990) to differences in predation regime (Crowl & proximate effects can influence fish fitness, and if Covich 1990; Reznick et al. 1990; Fraser & Gilliam levels of parasitism are relatively constant over 1992; Belk 1998). Effects of parasitism on life his- time in given populations, parasites can play selec- tory traits have mostly been ignored. Theoretical tive roles in fish evolution. Therefore, we might ex- models predict that a host that matures earlier or pect fish in heavily parasitized populations to at a smaller size in response to parasitism could evolve different life history strategies, to compen- reduce the impact of parasites on its reproductive sate for the effects of parasitism, than fish in other success (Hochberg et al. 1992). Such a change in populations. life history strategy could give the host a selective The optimal age at maturity is a trade-off be- advantage over conspecifics that mature later or at tween present reproduction and the probability of a larger size. For instance, in a comparison among future reproduction (Stearns 1989, 1992). Because populations, Lafferty (1993) found that snails ma- fecundity usually increases with fish size (Wootton tured at a smaller size in populations where the 1991), individuals allocating energy to growth can prevalence of larval trematode parasites was high. expect an increased future reproductive success. In 105 Hamilton & Poulin populations where parasites that lower host sur- variation in life history traits among bully popula- vival are abundant, however, the probability of fu- tions. ture reproduction is reduced. We might thus expect individuals from heavily parasitized populations, Methods where future reproduction is compromised, to ma- ture earlier and/or at a smaller size than individ- Upland bullies were collected from 7 distinct uals from populations experiencing weaker selec- stream populations in the Otago province of the tive pressures from parasitism. Phenotypic plas- South Island, New Zealand, during three breeding ticity could, to some extent, obscure evolutionary seasons (austral spring and summer, 1994–1995 to divergence between populations, but it is unlikely 1996–1997). All streams were similar with respect to mask it completely. Also, when fish populations to width (2–4 m) and depth (Ͻ50 cm), and all had are compared to evaluate the effect of a variable the same predator species (brown trout Salmo trut- such as parasitism, it is difficult to exclude the po- ta and eels, Anguilla spp.) and similar faunas of tential influence of other factors that also vary invertebrate prey. Water temperature ranges were among populations. At the very least, one should collected monthly in all streams, using maximum- attempt to control for variables for which data is minimum thermometer. Severe flooding events pre- readily obtained. Among abiotic factors, for in- vented complete annual thermal records. Fish were stance, water temperature is easily quantified and segregated by population and sex and kept in large it is known to have important effects on fish tanks for 1–2 days before being killed by an over- growth rates, age and size at maturity, and other dose of anesthetic (2-phenoxy ethanol). Fish were life history traits (Wootton 1991). In particular, wet weighed and measured to the nearest mm be- fish egg sizes vary across populations with respect fore dissection. The number of cysts of T. opis- to water temperature, either on a latitudinal scale thorchis found in each fish was recorded, as was (Fleming & Gross 1990) or on a seasonal basis the number of eggs in each female. Using only fe- (Mills & Eloranta 1985), and temperature can males from which eggs could be expelled by a therefore influence the trade-off between egg size gentle pressure on the body sides prior to dissec- and egg numbers. It is thus important to assess the tion, we also took measurements of egg sizes. A role of this variable along that of parasitism. random sample of 15 well-developed eggs from Here, we used inter-population comparisons to each of these clutches was measured using a bin- investigate how age and size at maturity, clutch size ocular microscope to obtain an average egg size and egg size in upland bullies (Gobiomorphus brev- per clutch. Since bully eggs are slightly ovoid in iceps Stokell) vary in response to variation in para- shape, egg size was taken as the maximum egg sitism and water temperature. Upland bullies are length. one of the most common fish species in streams of Scales and otoliths were removed from each fish New Zealand’s South Island (McDowall 1990). All for age analyses. Age was determined using annu- the bully populations we studied are parasitised by lus formation in both scales and otoliths. Upland larvae of the digenean trematode parasite Telogas- bullies display scale annuli formed when growth ter opisthorchis (MacFarlane 1945). The larvae en- slows down during winter (Staples 1975). When cyst in the visceral organs, muscles and body cav- circuli followed an annulus, fish were classified into ity of bullies, reducing the fish’s physical condition a‘‘¹’’ age class. Analysis of bully otoliths was used and affecting its responses to predators and its par- as a confirmation of ages determined using scales. ental care behavior (Poulin 1993; Stott & Poulin Otoliths were ground by sanding until the annuli 1996). The abundance of this parasite varies could be discerned under a light microscope. When greatly among fish populations (Hamilton et al. possible (some otoliths disintegrated during grind- 1997). Furthermore, its abundance has remained ing), both otoliths were used for age determi- stable in the only one of our study streams for nation. There was no significant difference, in any which we have long-term data: in the Linnburn of the populations, between ages calculated with stream, prevalence of T. opisthorchis has remained either left or right otoliths (paired t-tests, all 100% and the mean number of cysts per fish has PϾ0.50). There were also no significant differences not varied significantly between 1992 and 1997 between ages obtained from scales or otoliths from (Poulin 1993; Hamilton et al. 1997). Variation the same fish, in any of the populations (paired t- among fish populations and stability over time tests, all PϾ0.20). Still, the average of scale and make this parasite a potential agent of selection. otolith age was used to categorize fish into various For instance, it appears to have caused differences age classes (0¹,1,1¹, etc.). Age and size at ma- in the evolution of secondary sexual traits among turity, for each population, were taken as the age bully populations (Hamilton & Poulin 1999). It is and length of the youngest/smallest fish in which therefore also likely to be a key determinant of mature gonads were present. 106 Life history traits of upland bully We used parametric statistics after log10 trans- formation of the data that needed normalization (parasite load, total fish length, fish weight, clutch size). Using total length (gravid female fish only) and egg size as independent variables, we performed multiple regressions with clutch size as the depend- ent variable. This was done separately for each population as well as on pooled data. This pro- cedure allowed us to test for a trade-off between egg size and egg numbers while controlling for female body size, both within particular streams and among all fish independently of stream of origin. Then, pooling data from all populations, we re- gressed clutch size on total length and used the re- siduals as measures of clutch size standardized for body size.