Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Sexual size dimorphism in two populations of threespine stickleback (Gasterosteus aculeatus) : female… Hooker, Laura Jayne 1988

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1988_A6_7 H66.pdf [ 4.82MB ]
Metadata
JSON: 831-1.0097673.json
JSON-LD: 831-1.0097673-ld.json
RDF/XML (Pretty): 831-1.0097673-rdf.xml
RDF/JSON: 831-1.0097673-rdf.json
Turtle: 831-1.0097673-turtle.txt
N-Triples: 831-1.0097673-rdf-ntriples.txt
Original Record: 831-1.0097673-source.json
Full Text
831-1.0097673-fulltext.txt
Citation
831-1.0097673.ris

Full Text

SEXUAL SIZE DIMORPHISM IN TWO POPULATIONS OF THREESPINE STICKLEBACK (Gasterosteus aculeatus): FEMALE BODY SIZE AND SEASONAL FECUNDITY IN A MULTIPLE SPAWNING SPECIES by LAURA JAYNE HOOKER B.Sc, The University of British Columbia, 1983 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTERS OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Zoology) We accept this thesis as conforming to the required standard UNIVERSITY OF BRITISH COLUMBIA September 1988 © Laura Jayne Hooker In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of ZoOLO^ M, The University of British Columbia Vancouver, Canada Date >SfpV. a«|\<Kg  DE-6 (2/88) ii ABSTRACT To date, models of sexual size dimorphism do not explain selection for small females, and they are also limited in their ability to explain intraspecific variation in sexual size dimorphism. I propose that small females, in species which produce multiple clutches in a breeding season, could have a selective advantage if the interval between clutches is shorter for small clutches of eggs. When the breeding season is long, small females may produce more eggs in total than large females by producing more clutches, and thus small size could be selected for. Two populations of threespine stickleback (Gasterosteus aculeatus) showing divergence in the sexual bias of size dimorphism were used to determine if large or small females had a seasonal fecundity advantage in these multiple spawning fish, and whether the two populations had diverged in life-history characteristics (age at first reproduction, number of clutches, length of breeding season). In addition, the mechanisms by which the differences in size were achieved was investigated. Size-frequency diagrams obtained from field samples indicated that the Lewis Slough population was an annual one, while fish at the Angus Campbell site apparently survived for more than one breeding season. The larger size of females at the Angus Campbell site resulted primarily from continued growth with age, while males stopped growing in about one years time. In an environment chamber female fish from Lewis Slough grew more slowly, as they approached maturity, than males and were therefore were smaller than males. Data from field collections, fry raised to maturity in an environment chamber, and females individually monitored in captivity over the course of a breeding season indicated that the populations have diverged in life-history characters. Females from the Angus Campbell ditch site produced fewer clutches and eggs over the breeding season (a measure of reproductive effort), delayed maturity and matured at a greater size, and had a longer iii life-span than Lewis Slough females. These observations are more in accordance with the predictions from bet-hedging theory than r & K selection theory. Data from individually monitored females held in a common environment indicated that clutch size and interclutch interval increased with increasing body size but small females still did not attain the seasonal fecundity advantage predicted by the model. However, these results suggest that small females are capable of achieving a greater seasonal fecundity relative to large females than would be predicted by the difference in average clutch size alone. Actual counts of the total numbers of eggs produced by individuals in a breeding season showed seasonal fecundity to be independent of body size. Female body size and fecundity are more weakly linked than previously realized and this confers an increased flexibilty for responding to diverse selective pressures. iv TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES AKNOWLEDGEMENTS INTRODUCTION 1 SITE DESCRIPTION 9 Physical 9 Chemical 15 Biotic 16 MATERIALS AND METHODS 23 Reproductive Characteristics 23 Growth Under Laboratory Conditions 26 Growth and Age Structure in the Field 30 RESULTS 32 Reproductive Characteristics 32 Growth Under Laboratory Conditions 56 Growth Age Structure in the Field 69 DISCUSSION 77 Age Structure and Growth Patterns 77 Population Specific Reproductive Characteristics and Life-History Evolution 80 Model Testing 85 Correlates of Female Reproductive Success 89 Concluding Remarks 91 LITERATURE CITED 93 V LIST OF TABLES Table Title Page 1 Oxygen concentration in percent saturation for the two study sites during 1985. 15 2 Estimated percent of stomach fullness in sticklebacks from the Angus Campbell and Lewis Slough sites on two dates per site. 16 3 List of fish species trapped at the two study sites over a three year period. 17 4 Mean size (mm), standard deviation, and sample size of fish obtained in survey collections from the vicinity of the Angus Campbell and Lewis Slough study sites. 22 5 Mean egg diameter in millimeter (X) and standard deviation (S.D.) in clutches produced by a single female, between females, and between populations. 25 6 Sample size, mean, standard deviation and median for a suite of reproductive characters including body size at the beginning of the season and summer growth rate. 33 7 Regression of reproductive characters on body size for each population. 34 8 Probability values for a significant difference in reproductive characters between females from the two sites. 46 9 Matrix of correlation coefficients for reproductive characters in each population. 50 10 Summary of Analysis of Covariance on linearized growth curves of laboratory bred and raised fish. 56 11 Slopes and standard errors for linearized growth curves of replicates within a population. 58 12 Mean size (mm), standard error, and sample size for each sex within a population in laboratory bred and raised fish at the end of the experiment. 58 13 Summary of 2-way ANOVA for the effects of sex and population on body size in laboratory bred and raised fish. 59 14 Summary of 2-way ANOVA for the effects of sex and population on body size in wild-caught and laboratory raised fry. 62 15 Mean body size (mm), standard error, and sample size of each sex within a population with probability levels for a Least Means a posteriori test on the cells. 62 vi Total number of wild-caught and laboratory raised females from each site catagorized into maturity status and two size classes. Results of marking experiment. vii LIST OF FIGURES Fig. Title Page 1 Sexual size dimorphism in breeding Lewis Slough and Angus Campbell fish over a three year period. 8 2 Lower and central Fraser Valley of British Columbia showing location of two study sites. 11 3 Fluctuation in water depth at the two study sites over a three year period. 13 4 Seasonal water temperature profiles at the two study sites over a three year period. 14 5 Seasonal variation in the proportion of breeding females in catch at the two study sites (a & b). and proportion of anadromous marine fish in catch at the Lewis Slough site (c). 18 6 Lower and central Fraser Valley of British Columbia showing location of survey sites. 21 7 Artificial seasonal profile of temperature and photoperiod used to induce maturity in wild-caught fry. 29 8 Mean cluch size of an individual plotted against body size. 35 9 Length of an individual's reproductive period (days between first and last clutch) plotted against body size. 36 10 Date of first clutch (as numbered from the first of January) plotted against body size. 37 11 Date of last clutch (as numbered from the first of January) plotted against body size. 38 12 Total number of clutches produced in the reproductive season plotted against body size. 39 13 Total number of eggs produced in a reproductive season plotted against body size. 40 14 Mean interclutch interval of an individual plotted against body size. 41 15 Log of an individual's mean interclutch interval plotted against body size. 42 16 Square root of summer growth rate plotted against body size. 44 17 Length of an individual's reproductive period (time between first and last clutch) plotted against total number of clutches. 52 viii 18 Mean interclutch interval of an individual plotted against total number of clutches. 53 19 Diagrammatic summary of significant correlations between reproductive characters in females from the Angus Campbell site. 54 20 Diagrammatic summary of significant correlation between reproductive characters in females from the Lewis Slough site. 55 21 Growth in laboratory bred and raised fish. 57 22 Growth in wild-caught fry raised to maturity in an environment chamber under simulated conditions. 61 23 Distribution of timing of first ovulation in Angus Campbell females, wild-caught as fry and raised in the laboratory under simulated photoperiod and temperature conditions. 65 24 Distribution of timing of first ovulation in Lewis Slough females, wild-caught as fry and raised in the laboratory under simulated photoperiod and temperature conditions. 67 25 Size-frequency histograms for Angus Campbell males, females, and unknowns as proportion of catch obtained at sampling dates spanning from May 1986 - May 1987. 71 26 Size-frequency histograms for Lewis Slough males, females, and unknowns as proportion of catch obtained at sampling dates spanning from July 1986 - June 1987. 75 ix AKNOWLEDGEMENTS I would like to thank my supervisor, Dr. J. D. McPhail, for his support and continued tolerance (in the face of vehicular Acts of God). Drs. J. Myers and W. E. Neill provided invaluable guidance throughout all stages of this thesis, and extremely thorough editing jobs. Field assistants, A. Simons and T. Suzuki, were highly appreciated, as was everyone who "rode up the valley" with me. Some of whom were slightly disappointed (Vivian) that biological research was not necessarily a trip on the Calypso. Much thanks to my PC pals on the third floor, particularily M. L. Burleson, for allowing me countless hours of PC mooching. Finally, I would like to thank my Mom and Dad for their love, endless emotional support, and seemingly endless financial support. 1 INTRODUCTION Darwin (1874, pg.332-375) proposed that the female bias in body size observed in many animals evolved f rom the selective advantage associated with the increased fecundity of large females. Large females tend to have larger clutches or broods than small females, presumably because there is more space inside their body (Shine 1988). Darwin ' s "fecundity advantage" model has been widely invoked and is especially attractive when applied to many ectotherms where clutch size can increase dramatically with increasing body mass (e.g. Shine 1979; Berry & Shine 1980; Veul le 1980; Semlitsch & Gibbons 1982; Woolbright 1983). This model, however, makes no provision for the evolution of small female body size, and those species and populations where males are larger than females requires an explanation. In birds and mammals, and to a lesser extent reptiles, variation in the bias of sexual size dimorphism has been attributed to the type of mating system and selection that result f rom intrasexual competition for mates (e.g. Trivers 1972; Singer 1982; Clutton-Brock 1983; Woolbright 1983; Ris ing 1987). In these models the presence of small females is predicted under conditions which select for large males. None of the models actually propose a mechanism that selects for small female body size. The application of these sexual selection models to ectotherms is l imited because little provision for ^determinant growth and the associated changes in number of eggs per clutch with size and age is made (number of eggs per clutch is referred to as "instantaneous fecundity" by Shine (1988), this term w i l l be used throughout this thesis to provide continuity with existing literature). In addition, these models have been developed from comparisons across species. Therefore, unless the populations exhibit radical differences in mating systems and behaviour, their usefulness to explain population divergence in the bias of sexual size dimorphism is l imited (but see Weatherhead 1980; Price 1984; and Ris ing 1987). Studies on populations are 2 important because they are less confounded by phylogenetic constraints, and are thus helpful in elucidating the initial action of natural selection on sexual traits and the evolutionary pathways of their development. Differential niche utilization has also been suggested to account for variation in sexual size dimorphism (Selander 1966). It is hypothesized that dimorphism allows a breeding pair to more efficiently use the resources in a territory. However investigators have found little to support this idea (Hays 1972; Rothstein 1973; Howe 1982; and Jehl & Murray 1984). Therefore evolution of small female body size in ectotherms must be explained in the face of the seeming fecundity advantage of large body size. A model is also required that is suitable for the analysis of interpopulation divergence in sexual size dimorphism. Neither the fecundity advantage model nor the sexual selection models address the capacity of females to produce multiple broods, or clutches, in a breeding season (within season iteroparity). Yet it is recognized in many species of small teleost fish that repeat spawning within a season can dramatically affect the potential seasonal fecundity of a female (Mann et aL 1984; Ware 1984; Wootton 1984; Hubbs 1985; and Burt et al 1988). This observation illustrates a fundamental flaw underlying Darwin's fecundity advantage model. Natural selection should work to increase life-time reproductive success, whereas in Darwin's model, body size increases because selection acts to increase instantaneous reproductive success (see discussion by Shine 1988). Instantaneous fecundity may increase as a function of body size in species where females produce multiple clutches, but seasonal or life-time fecundity may be independent of (Gale 1983, Burt et aL 1988), or inversely correlated with, body size. The process of maximizing life-time reproductive success as a function of either increasing the number of eggs per clutch (clutch size) or number of clutches provides a mechanism for the selection of small female body size and the evolution of intraspecific variation in the direction of sexual size dimorphism. If small females are better able to 3 spawn repeatedly within a reproductive season then given sufficient time the clutch size advantage of large body size may be offset. In this way small females may actually enjoy a greater seasonal fecundity than larger females, and thus be selected for. If, however, conditions curtail the length of the breeding season so that one or very few clutches are produced, then seasonal reproductive success approaches instantaneous fecundity, which is a function of body size. Directional selection should then act to increase body size. Number of clutches must decrease with increasing body size in order for the above scenario to work, and quantitatively the ratio of interclutch interval to clutch size must increase with larger body size so that given sufficient time small females can exceed the egg production of larger females. Alternatively, providing the reproductive season is long enough, small females must be able to continue producing clutches for a longer period of time than large females. Burt et aL (1988) predict that the number of clutches per unit time decreases with increasing body size because of the allometry of reproduction in aquatic ectotherms. Clutch mass scales as adult body mass to the power of 0.92 (Blueweiss et aL 1978) while absolute metabolic rates, including growth rates, generally scale to the 0.75 power (Peters 1983). As a result energy is more efficiently converted to the production of gametes than somatic tissue (Townsend & Calow 1981, pg. 251). Alternatively, interclutch interval may vary as the number of eggs in a maturing clutch varies, and thus be correlated with body size. A likely component of oogenesis that could affect variation in interclutch interval is vitellogenesis (deposition of true yolk). Vitellogenin (the protein precursor of yolk) is synthesized in the liver and secreted into the bloodstream. This material is then sequestered from the bloodstream by the oocyte and deposited in the cortex of the cell (Wallace 1978). Given that each egg receives the same amount of yolk - i.e. that egg size does not vary with clutch size - it may simply require more time to metabolize and deposit materials in a larger number of eggs. 4 Experimental evidence for the existence of a relationship between body size and number of clutches produced in a breeding season is difficult to locate. The information that is available is usually collected incidentally to the main thrust of an investigation, and consequently is often incomplete, mixed or confounded with food ration, population density and population source effects. This is particularly true for field studies. For example, Dingle et al (1982) found size variation among females from two populations of milkweed bugs (Oncopeltus fasciatus). The larger females produced more clutches and had smaller interclutch intervals, but they also produced smaller clutches. In contrast, Hegmann & Dingle (1982) looking at correlations between life history characters in a population of the same species of milkweed bugs, found that body size, clutch size, and interclutch interval were positively correlated. Gibbons et al. (1982) in a field study of two species of turtles (Kinosternon subrubrum and Pseudemys scripta) found no apparent relationship between individual body size and clutch frequency. In an interpopulation study Reznick & Endler (1982) found that larger female guppies (Poecilia reticulata') from one population had larger offspring and more developing embryos produced at longer intervals than the smaller females from a second population. Mann et al (1984) observed that larger sculpins (Coitus  gobio) produced fewer clutches than their smaller counterparts; however their data are confounded with density effects. There is, therefore, theoretical and experimental evidence suggesting that the necessary qualitative relationships do exist, while the quantitative aspects of these relationships have yet to be fully explored. Repeat spawning and seasonal production of eggs are components of reproductive effort. Reproductive effort is in turn a component of life history theory (Stearns 1980; Bell 1980), therefore it is predicted that populations which have diverged in reproductive parameters, and sexual size dimorphism, also have different life histories. The conditions for testing Darwin's fecundity advantage model on multiply spawning fish, and my model of sexual size dimorphism appear to be well met in two 5 populations of the freshwater form of the threespine stickleback Gasterosteus aculeatus) found in the central Fraser Valley, British Columbia. These two populations are found in the same geographic area and exhibit different biases in sexual size dimorphism (see Figure 1). Initial observations on captive females fed ad libitum diets over the reproductive period (late March to early September) suggested that smaller females produce more clutches more frequently than large females, both within and between populations. Historically, these two populations were probably part of a single population since the two sites are located in an area formerly occupied by a large shallow lake. This lake (Sumas Lake) was drained circa 1912 to form farmland, consequently the separation between the two populations is recent (<80 years) and divergence may have been quite rapid. Divergence in this instance describes the relative differences between phenotypes because there is no actual knowledge of the founder population. The close geographic proximity of the sites exposes the two populations to similar climatic conditions (e.g. air temperature, photoperiod, and precipitation) so that causative factors for the divergence are likely to be found within the immediate habitats. These circumstances make the two populations amenable to the study of processes driving divergence in sexual size dimorphism, reproductive parameters, and general life histories. Gasterosteus aculeatus on the west coast of North America is well suited for studies on patterns of variation in behaviour, morphology, life histories etc. because of the high degree of phenotypic diversity found both within and between freshwater population (reviewed by Wootton 1976, 1984). Much of this diversity is directly attributable to the local selective regime (McPhail 1969; Hagen & Gilbertson 1972, 1973; Bentzen & McPhail 1984; Bell 1984; Lavin & McPhail 1985). The purpose of this thesis is to investigate the mechanisms by which the size differences are achieved (i.e. inherent or environmentally induced differences in growth rate, population age structure, etc.) and to determine if 1) the populations have diverged in life history characters, and if so, to interpret the pattern with respect to current life history 6 theory; 2) variation exists in reproductive parameters as a function of female body size; 3) a fecundity advantage to large females exists in these multiply spawning fish and 4) the quantitative variation in reproductive characters is such to support my model for divergence in sexual size dimorphism. 7 Figure 1. Sexual size dimorphism in breeding Lewis Slough (LS) and Angus Campbell (AC) fish over a three year period. Boxplots show size distribution of fish. The length of the box represents the interquartile range of the values, the middle line indicates the median value, the area above the line is 75% of the interquartile range, and below the line is 25% of the interquartile range. The whiskers are drawn to the limit of the standard range (1.5 X interquartile range), and stars represent points beyond the standard range (Chambers et al 1983). f - female, m - male. CO o body size (mm) body size (mm) o CO 00 0> o - r — cn o OS o —r— 1-CO o body size (mm) > o > o co oo cn to CO 3 9 SITE DESCRIPTION PHYSICAL Figure 2 shows the location of the two study sites in the lower Fraser Valley of southwestern British Columbia. The geographic proximity of the two sites subjects them to the same general climatic conditions, but even though precipitation is similar, fluctuations in water level differ greatly between the two sites (Fig.3). Angus Campbell is a roadside ditch interrupted by pools on the downstream side of culverts. It is most likely fed by rain, irrigation runoff, and small springs common to this area. During periods of heavy rainfall the pools can be up to 2m and wide and 2m deep. The water flow can be turbulent and laden with suspended material. During late summer, however, when rainfall is greatly reduced the pools all but dry up and leave as little as 10cm of water. The intervening sections of the ditch/stream between the pools can dry completely. Angus Campbell ditch flows into Marshall Creek and the section of the ditch near the mouth often dries up and completely bars access to the creek from the ditch. In contrast, Lewis Slough is part of a large interconnected system of drainage and irrigation canals. These range from about 1 - 5m in width and have maximum depths ranging from about 1 - 3m. At the Lewis Slough collection site the maximum measured fluctuation is considerably less than that observed at the Angus Campbell collection site (Fig. 3) and at no time was the water depth shallow enough to expose the minnow traps used for collection. In contrast during the summer at the Angus Campbell site it was frequently difficult to find water deep enough to cover the traps. The Lewis Slough site also experienced less variation in water flow, and was seldom as turbulent as the high flows observed at the Angus Campbell site. Although water depth and flow conditions are different for the two sites the seasonal profiles of water temperature (Fig. 4) are similar. 10 Figure 2. Lower and central Fraser Valley of British Columbia showing location of two study sites. AC - Angus Campbell, LS - Lewis Slough. 11 12 Figure 3. Fluctuation in water depth at the two study sites over a three year period. The data for the Angus Campbell site represents maximum depth. The data for the Lewis Slough site indicates fluctuation in water level at the location of the measuring stick, the maximum depth is one meter deeper than this. Dotted lines represent a period of greater than two months over which no measurements were taken. 200 1 1 1 1 1 1 1 1 1 1 t ' • < ' i i i i i 5/85 6 7 8 9 4/86 5 6 7 8 9 10 11 12 1/87 2 3 *6 7 8 9 10 11 D A T E (month/year) 14 Figure 4. Seasonal water temperature profiles at the two study sites over a three period. * - Angus Campbell, • - Lewis Slough. 15 The Angus Campbell site is also subject to human disturbances. Once every three or four years Angus Campbell is ditched by Abbottsford municipal workers. In March 1987 a backhoe was used to deepen the water course and remove emergent vegetation, as well as the nests of any males breeding at the time. The banks of the ditch also are regularly mowed during the summer months. This throws grass into the pool areas, and at least once, in June of 1987, the mower knocked the cap off a water pipe causing a large flow of very silty water into, the ditch, again, probably destroying the nests of any breeding males. C H E M I C A L Oxygen (Table 1) was the only chemical parameter measured. It was often difficult to obtain repeatable readings because of particulate material fouling the meter's membrane. It was therefore abandoned during sampling in 1986. In general, oxygen readings are at or just below saturation level except in Angus Campbell during the summer when both water levels and oxygen levels were low. Table 1. Oxygen concentration in percent saturation for the two study sites during 1985. DATE 5/22 6/2 6/14 7/4 7/24 8/13 9/7 80.4 50.5 60.1 50.3 30.9 5.1 6.4 50.4 50.9 61.3 62.3 63.0 60.5 63.5 Because both sites are in areas of intense agricultural activity, and they are subject to applications of fertilizer and irrigation, I suspect that levels of phosphorous and nitrogen, particularly in the summer, probably are high and it is unlikely that primary production at Angus Campbell Lewis Slough 16 either site is nutrient limited. The suggestion of high nutrient loading is supported by the turbid appearance of the water and organic sudsing in Lewis Slough, and by the smell of cow manure in the water of Angus Campbell. BIOTIC From early spring to mid-autumn both sites support luxurious stands of emergent and floating vegetation. At the Angus Campbell site during mid-summer a mat of macrophytes approximately 20cm thick covers the water surface of the pools. During the rest of the year plant cover was reduced but neither site was ever completely devoid of plants and there were always small stands of cover present. At both sites large numbers of invertebrates were associated with the vegetation. Sweeps with a dipnet through, and next to, the plants yielded an abundance of amphipods, water beetles, leeches, dragon and damselfly nymphs, and snails. The occurrence of these decreased when the vegetation died out during the winter. Adult stickleback stomachs were filled with amphipods, chironomid and ceratopogonid larvae, ostracods, and some plant material. In the stomachs examined there was no obvious difference in the estimated percent of fullness (Table 2) between the two sites. Table 2. Estimated percent of stomach fullness in sticklebacks from the Angus Campbell and Lewis Slough sites at two dates per site. Values for each date are a proportion of the total sample examined. Samples were obtained by trapping. Traps were left in for four hours and fish were immediately killed upon collection of traps. Angus Campbell Lewis Slough  % of fullness april 25 may 22 may 2 July 4 0 - 25 .45 .56 .48 .65 25 - 50 .27 .21 .26 .14 51 - 75 .22 .21 .21 .18 76 - 100 .06 .02 .05 .03 N 49 43 42 34 17 Table 3. L ist of f ish species trapped at the two study sites over a three year period. X indicates presence, - indicates absence. A C denotes the Angus Campbel l site, and L S the Lewis Slough site. A C L S Threespine stickleback (freshwater form) (Gasterosteus aculeatus) Threespine stickleback (anadromous form) (Gasterosteus aculeatus) Peamouth chub (Mylocheilus caurinus) Redside shiner (Richardsonius balteatus) Cutthroat trout (Salmo clarkii) Squawfish (Ptychocheilus oregonensis) Brassy minnow (Hybognathus hankinsoni) Carp (Cyprinus carpio) X X X X X X X X X X X X Table 3 lists the fish species trapped at the two sites over the three year study period. Excluding the resident freshwater form of Gasterosteus aculeatus. Angus Campbel l supports three other fish species, however, these species only appeared during early spring or late autumn and never occurred in large numbers. During the summer sticklebacks appear to be the only resident fish. In contrast Lewis Slough supports eight species of fish, including both the resident freshwater and the anadromous marine forms of Gasterosteus aculeatus. L i ke Angus Campbel l resident freshwater sticklebacks were the predominant form in Lewis Slough, both in numbers and persistence in the area. The only fish that exceeded freshwater sticklebacks in numbers was the anadromous marine form in Lewis Slough. This form comes into the area to breed during A p r i l and early May. Figure 5(c) shows seasonal variation in the proportion of the anadromous form in the catch at the Lewis Slough site over the three year study period. There is some indication that their presence is cycl ic, in 1985 numbers were low, in 1986 they were high, 18 .8 , 2 / 1 9 4 / 1 0 5 / 3 0 7 / 2 0 9 / 7 DATE Figure 5. Seasonal variation in the proportion of breeding females in catch at the two study sites (a & b), and proportion of anadromous marine fish in catch at the Lewis Slough site (c). Arrows in 5(a) indicate date of drying up or extreme low water at the Angus Campbell site, o - 1985, • - 1986, * - 1987. 19 and 1987 they were again low. In addition, in Lewis Slough the peak in abundance of the anadromous form appears to coincide with the peak reproductive period of the sticklebacks in Angus Campbell (Fig. 5(a)), whereas the primary reproductive period of the freshwater sticklebacks in Lewis Slough appears to occur later (Fig. 5(b)). Figure 5 also indicates that the stickleback (hereafter meaning the resident freshwater form unless otherwise specified) in Angus Campbell begin breeding earlier and stop breeding sooner than Lewis Slough. The Angus Campbell breeding season is almost two months shorter than that in Lewis Slough because the ditch dries up in most years. In total, the breeding season in Lewis Slough is longer than that in Angus Campbell. The sexual size dimorphism observed in the Angus Campbell ditch (Fig.l) was not found in samples from the drainage system surrounding this site (Fig.6), rather, these populations tended to have females and males of approximately the same size range (Table 4). These collection sites were physically larger than the Angus Campbell site and tended to resemble either sloughs or creeks. In addition, they did not show the same kind of temporal variation as Angus Campbell. Nothing analogous to the Angus Campbell site was found in the immediate vicinity. Sticklebacks trapped in the sloughs surrounding the Lewis Slough site (Fig.6) yielded roughly the same bias in sexual size dimorphism (Table 4) as observed in Lewis Slough (Fig.l), although some larger females were caught. One exception was a small spring fed stream (Site 17). A large range in females size was observed at this site and very large females were common. In addition, the marine anadromous form was more frequently trapped in larger numbers at this location than at the Lewis Slough site. 20 Figure 6. Lower and central Fraser Valley of British Columbia showing location of collection sites. Numbers correspond with those in Table 3. 22 Table 4. Mean size (mm), standard deviation, and sample size of f ish obtained in survey collections from the vicinity of the Angus Campbel l and Lewis Slough study sites. Site numbers correspond with those in Fig.6. Values for a site may represent data collected over more than one sampling period. female male Site n mean S.D. n mean S,D, 1 35 45.7 7.5 22 43.1 3.2 2 21 42.1 5.2 15 44.9 4.6 3 15 44.2 7.4 10 45.7 4.7 4 23 39.5 6.4 17 42.2 5.4 5 32 41.1 7.3 21 43.3 3.1 6 28 43.2 6.8 17 42.8 5.4 7 15 34.3 5.2 10 35.0 3.4 8 23 38.7 5.9 12 42.4 5.1 9 33 47.3 4.9 25 44.2 4.3 10 24 43.3 6.5 18 41.3 3.7 11 23 44.7 6.9 20 43.6 3.1 12 41 35.5 5.5 22 38.7 3.7 13 32 36.1 5.9 17 42.1 3.8 14 10 40.2 3.1 8 41.2 3.8 15 52 36.8 7.5 "33 43.7 4.4 16 43 39.4 7.5 27 44.2 4.9 17 30 41.2 8.1 21 40.3 3.8 18 25 37.2 5.8 16 45.6 3.4 19 21 35.9 5.5 12 44.3 4.1 23 MATERIALS AND METHODS  REPRODUCTIVE CHARACTERISTICS My goal was to evaluate the reproductive performance of females as a function of both body size and population. Since environment can influence growth rate and ultimately both body size and fecundity, it was necessary to rear both populations in a common environment. A procedure was designed where individual wild-caught females could be monitored over the course of a spring and summer. Comparative data from the field would have been desirable. Unfortunately, however, I could not track wild individuals. Females were trapped from each site just prior to that population's breeding season (early March for Angus Campbell, and early April for Lewis Slough) and transported back to the laboratory. At this time their abdomens were beginning to swell and they were clearly distinguishable from males, but since the breeding season had not yet started it was unlikely that any clutches had already been produced. Because the widest possible range of body sizes for breeding females in each population was desired, and Lewis Slough females are small, a sample of Lewis Slough females was collected in the summer of 1986 and held in captivity for use in 1987. This allowed the fish to grow to sizes larger than those that normally occur in Lewis Slough and into the size range typical for Angus Campbell fish. Thirty females from each population (for Lewis Slough the 30 fish were 15 fish from 1986 and 15 fish from 1987) were measured and individuals placed separately into 16 X 12 X 11.5cm guppy breeding nets. These nets were suspended in a 295 X 80 X 30cm fiberglass tray located in a courtyard exterior to the Biosciences building at U.B.C. Thus, the experimental fish experienced natural Vancouver temperature and photoperiod conditions. Dechlorinated water was run through the tank at the rate of about 12 ml/s (although this varied somewhat depending on other users of the water system). This flow replaced the tank's volume about once every sixteen hours. To monitor temperature a maximum-minimum thermometer was suspended in the tank. 24 The fish were fed ad. libitum once or twice a day on Tubifex worms, adult Artemia. and Chaoborus larvae, all obtained from a local pet food supplier. With this feeding regime, food was constantly available in each net and typically some food was present in the net at the time of the next feeding. In addition, aquatic vegetation taken from the study sites was placed in each net to provide cover as well as some natural food (e.g. dipteran larvae, snails and leeches). The females were monitored daily to determine if they had matured a clutch of eggs. Ovulation was judged to have occurred when eggs could be readily stripped by hand from the female's abdomen. When a fish ovulated, it was measured, the date recorded, and the number of eggs was counted. At first egg diameter was also measured but it was found that means varied little between individuals (Table 5). Student's t-test of the data showed no significant difference in mean egg diameter between the two populations (p = 0.213). Consequently, egg diameter was not measured in later experiments. All the experimental fish were killed and measured on September 4, 1987. At this time no female had produced a clutch for about three weeks. The data from this experiment allowed each individual's growth rate to be determined over the summer. In addition, the total number of clutches produced, the mean number of eggs in a clutch (mean clutch size), the length of the reproductive period (date of first clutch to date of last clutch), the total number of eggs produced over the summer, and the mean interclutch interval (the average time between clutches) for an individual could be determined. The measurements from each female were treated as statistically independent. Even though the fish were held in a common tank of water it was assumed that the flow rate was sufficient to remove any diffusible hormones that may affect ovulation frequency. Neighbors did not noticeably ovulate at the same time. Within each population all of the above reproductive characters were regressed against initial body size to determine if they were related (Model I regression). Where necessary square root or log transformations were applied to normalize the data. 25 Table 5. Mean egg diameter in millimeters (X) and standard deviation (S.D.)in clutches produced by a single female, between females, and between populations. Each mean was calculated from a sample of 10 eggs chosen without obvious bias, and measured with a micrometer under a dissecting microscope at medium power. AC - Angus Campbell, LS -Lewis Slough. indr# bodv sizefmm-) pop. clutch# X S.D. 1 35.3 LS 1 1.5 0.07 2 1.5 0.06 3 1.5 0.11 4 1.5 0.12 5 1.5 0.05 2 35.8 LS 1 1.5 0.08 3 36.8 LS 1 1.5 0.09 4 37.4 LS 1 1.5 0.07 2 1.4 0.08 3 1.4 0.08 5 40.4 LS 1 1.5 0.08 2 1.5 0.01 3 1.4 0.09 4 1.5 0.03 "'5 1.5 0.13 6 40.5 LS 1 1.5 0.02 7 42.0 AC 1 1.6 0.04 8 43.0 AC 1 1.5 0.07 9 45.9 AC 1 1.5 0.05 10 47.0 AC 1 1.4 0.18 11 47.4 LS ' 1 1.9 0.10 12 48.5 AC 1 1.7 0.12 13 52.0 AC 1 1.5 0.05 14 53.9 AC 1 1.5 0.05 15 56.8 AC 1 1.6 0.05 To investigate differences in these reproductive characters between the two populations the appropriate regression lines were compared. Because only two lines are involved this constitutes a simplified case of ANCOVA and the two coefficients can be tested with a Student's t-test (Sokal & Rohlf 1981 pg.506). Intercepts were not tested because a body size of zero has no biological significance. Instead two Y values at X = 40mm and X = 50mm were compared (Zar 1984 pg.299) with a t-test. These body lengths were chosen because they roughly represent the average size of breeding females in the two populations. 26 If the characters showed no significant regression on body size, measurements of fish from the two populations were also compared with Student's t-tests (two-tailed). The possibility of associations among the reproductive characters within each population was explored with correlation analysis. The significance of correlation coefficients were determined by the technique given in Sokal and Rohlf (1981 pg.585). GROWTH UNDER LABORATORY CONDITIONS The observation of a striking difference in the size of breeding females between the two sites suggested the possibility of growth rate differences between Angus Campbell and Lewis Slough fish. To investigate this possibility, an experiment was conducted using laboratory bred and reared fish. Since these animals experienced the same feeding and environmental regime, any growth rate differences detected between the two populations are probably inherited. In addition, size, age, and length of photoperiod at maturity for each population was determined by rearing wild-caught fry under identical conditions in an environment chamber. In the first experiment it was important to reduce family effects in order to assess a generalized growth curve for each population. Therefore, 12 separate crosses were made in each population and individuals from eaph family within a population were mixed and raised together. Adult males and gravid females from each site were trapped and transported alive back to the laboratory. Eggs were stripped from ovulated females and placed in plastic petri dishes with a drop of water. Each clutch of eggs was fertilized by a single male and each male was used only once. The testes were removed from pithed males and macerated with fine-tipped forceps. The resultant slurry was placed on top of the eggs for about twenty minutes. The remains of the testes were then removed and the eggs placed in a 500ml jar with dechlorinated water conditioned with a hay infusion. The jars were placed in an environment chamber under controlled light (12h light/12h dark) and temperature : 27 conditions (20 °C) with air bubbled vigorously through the water. Unfertilized and fungus-ridden eggs were removed daily. After hatching the air flow was reduced and upon complete absorption of the yolk sac a subsample from each cross was measured from the tip of the nose to the tail with a micrometer under a dissecting microscope (at the highest magnification that included the whole fish in the field of view). Subsequent subsamples were measured in this fashion until the fish were large enough to have their standard length determined with calipers. Fifteen days after hatching 150 individuals from each population were chosen without conscious bias. These were then divided among six 13.21 aquaria (six tanks per population, 25 fish per tank). The aquaria were filled with dechlorinated water and supplied with a filter, straw, and aquatic plants native to the collection sites. After five days the fish density in each aquarium was reduced'to 20, and at the age of 100 days the density was reduced to 10. Prior to this age a subsample from each tank had been taken for measurement, but at this time all fish in each tank were measured. Half were killed and sexed. The experiment was terminated on day 140. By this date it appeared that the growth rates were declining. The remaining fish were then killed, measured, and sexed. Initially, newly hatched fry were fed with a hay infusion which supplied small protozoans, subsequently they were fed Artemia nauplii, and finally weaned to live Tubifex worms. Fish were fed at least once a day and to excess (again there was almost always living food available in the tanks when the next day's feeding occurred). This was done to reduce competition for food and the effects that this might have on growth. The resultant growth curves (size as a function of age) were transformed by applying the linear form of the exponential sigmoid equation Y = ln(X/(A-X)), where A is an estimate of the asymptote (Spain 1982)! Slope differences between tanks and population were then tested with an ANCOVA (GENLIN on the MTS system) using age as the covariate. Bonferroni and Mirumal a posteriori tests were applied to detect non-homogeneous subsets at the 5% probability level. Sizes at the termination of the experiment were analyzed as a function of sex and population, and the sex by population 28 interaction, using a 2 X 2 ANOVA design available on the Statistical Analysis Systems program. A Least Squares Means a. posteriori test was then applied to identify any cells which might deviate from the rest. In the second experiment wild-caught fry were obtained by dipnet at the study sites during early August and then transported alive back to the laboratory. Here standard lengths were measured to the nearest millimeter with calipers and the fish were divided into 1mm size catagories (17-18mm, 18-19mm etc.). Keeping the populations separate, I divided the size catagories into groups of 5 fish and placed each group into a 13.2 1 aquarium The tanks were supplied with dechlorinated water, air filters, some gravel and aquatic vegetation native to the study sites. The tanks were situated in an environment chamber and temperature and photoperiod were manipulated to roughly simulate, yet accelerate, a seasonal cycle (Fig. 7). The temperature and light regimes were fabricated to induce maturity in sticklebacks. This species responds to iBumination beyond a photoperiod threshold level that shifts with past photoperiod conditions (Baggerman 1985). The fish were fed ad libitum on a diet of Anemia nauplii, Chaoborus larvae, and Tubifex worms. Periodically, standard length was measured with calipers; initially only one or two individuals from each tank were sampled and measured because when they were small it was difficult to locate all individuals. Later all individuals within a tank were measured. During the course of the experiment females were judged to be mature when they ovulated their first set of eggs. Ovulation was determined by the criteria previously described. The date and size of the female was recorded. Each fish was then removed to another aquarium where she could be monitored for individual growth and the production of subsequent clutches. Using a Mann-Whitney U test with tied ranks (Sokal & Rohlf 1981, pg. 433) the two populations were analyzed for differences in time of reproduction and size at first maturity. 29 40 80 120 160 200 240 280 320 TIME (days) Figure 7. Artificial seasonal profile of temperature and photoperiod used to induce maturity in wild-caught fry. The solid line represents photoperiod (hours of illumination out of 24), and the dotted line indicates temperature. The dashed line indicates a period when the environment chamber broke down and photoperiod went to zero and temperature fluctuated. Time on the X-axis corresponds to the estimated age of the fry (see text for details). 30 The experiment was ended and all fish killed after temperature and photoperiod in the environment chamber had been held at a maximum of 20°C and 16 h light for about a month. Final standard length was measured and sex was determined by dissection. Final size as a function of population, sex, and population by sex interaction, was analyzed as in the previous experiment. GROWTH AND AGE STRUCTURE IN THE FIELD Initially I attempted to age the sticklebacks by removing otoliths from wild-caught fish. This technique failed, however because although there were clear bands on the otoliths their pattern was too inconsistent for a reliable estimate of age. These populations may deposit hyaline and non-hyaline zones on a sub-seasonal basis (see Borland 1986 for a discussion of the difficulties of aging 'sticklebacks using otoliths). This may be a population specific problem as some investigators claim to have no difficulty with this technique (Giles 1987). Clues to the age structure of my populations came only from changes in the size frequency distribution over time. The same changes also indicated growth patterns in the field. Samples of fish were trapped at the study sites throughout the course of each year. I attempted pole seining in Angus Campbell during a period of very low water but the extremely dense vegetation made this difficult. Samples were either preserved in 5% formalin and examined in the laboratory, or the fish were sexed by external features when possible, measured and returned to the water. In the latter circumstance some females could be determined because of their distended abdomens, whereas males often sported nuptial colouration and were much leaner than the females. Many individuals, however, could not be positively sexed in the field and these were placed in a third category designated unknown. Histograms of size frequency for each sex, and the unknowns, in each population were produced throughout the course of a year. During the early spring of 1987 a marking experiment was performed at the Angus 31 Campbell site to determine sex specific growth in two size classes. The population sample was drawn in March and fish measuring < 40mm had their second dorsal spine clipped. Those measuring >40mm had their first dorsal spine clipped. After clipping the fish were then returned to the water. Forty rmilimeters was chosen as the demarcation point because the available information suggested this was approximately the cut-off size between adults of the previous year and the young of the present year. The area was subsequently re-trapped twice and the samples assayed for sex, size, and marks. The sex of individuals that grew from the smaller size class into the larger was noted. A similar experiment was not attempted at the Lewis Slough site because at this site fish were unavailable in March. 32 RESULTS REPRODUCTIVE CHARACTERS Table 6 presents the mean, standard deviation, and median for body size, a suite of reproductive characters, and summer growth rate for the 29 individually monitored females (one female from each site died before the end of the experiment) from each site. In Angus Campbell, sample size for date of first and last clutch differs from 29 because one female did not produce a clutch. In both populations, sample size for mean interclutch interval differs from 29 because some individuals ovulated only one clutch and therefore an interclutch interval could not be calculated. Regressions of reproductive characters and summer growth rate on female size for each population are summarized in Table 7. Figures 8-16 present scatter diagrams of these reproductive characters plotted against body size. A regression line is presented only when its' slope is significantly different (at p<.05) from zero. In the analysis for length of reproductive period (Fig.9) individuals with a calculated value of zero (i.e. first and last clutch were the same) were not included. Individuals (3 from Angus Campbell, 1 from Lewis Slough) with a mean interclutch interval greater than 60 days were regarded as outlyers and were not included in this analysis. The only character shared by both populations that shows a slope significantly different from zero (rx.OOl) is clutch size (Fig.8). Analysis of the slopes and estimates of Y at two X values (Table 8) shows that the relationship is not different in the two populations (p>.5). The regression lines indicate that 40mm females produce about 75 eggs per clutch, while 50mm females produce nearly twice as many. The quantitative aspects of this relationship is in close agreement with the observations of Wootton (1973), Hagen 33 Table 6. Sample size, mean, standard deviation and median for a suite of reproductive characters including body size at the beginning of the season and summer growth rate. The experimental females from each site were monitored in a common environment. Reproductive period indicates number of days between the first and last clutch. Dates represent the number of days counted from January 1. Angus Campbell CHARACTERISTIC n mean S.D. median body size (mm) 29 49.2 7.5 47.7 mean clutch size 29 146.28*' 64.81 143.00 reproductive period (d) 29 46.21 43.95 45.00 date of first clutch 28 107.71 24.31 105.00 date of last clutch 28 152.89 41.24 154.00 total number of clutches 29 3.24 2.29 3.00 total number of eggs 29 443.14 285.04 420.00 mean interclutch interval (d) 21 23.81 17.24 18.00 summer growth rate (mm/day) 29 .09 .02 .03 Lewis Slough CHARACTERISTIC n mean S.D. median body size (mm) 29 46.0 5.0 46.8 mean clutch size 29 127.52 41.88 127.00 reproductive period (d) 29 67.31 26.38 72.00 date of first clutch 29 136.41 19.23 135.00 date of last clutch 29 202.72 16.16 203.00 total number of clutches 29 5.17 2.20 5.00 total number of eggs 29 651.90 306.72 648.00 mean interclutch interval (d) 28 • 18.93 11.45 16.00 summer growth rate (mm/day) 29 .03 .02 .03 34 Table 7. Regression of reproductive characters on body size for each population. The p represents the probability that the slope of line is not different from zero (a = intercept, b = slope). Reproductive period indicates number of days between the first and last clutch. Dates represent the number of days counted from January 1. Angus Campbell characteristic df a b r2 p mean clutch size 27 -171.34 6.45 .56 .001 reproductive period*,b date of first clutch' 19 5.22 0.05 .02 .5 26 11.85 -0.31 .04 .5 date of last clutch 26 198.78 -0.85 .04 .5 total number of clutches 27 6.98 -0.08 .06 .2 total number of eggs 27 -138.45 11.81 .10 .1 mean interclutch interval(d)c,d 16 1.81** 0.02 .24 .05 summer growth rate(rnm/day)' 27 0.16 -0.02 .50 .001 Lewis Slough characteristic df a b r2 p mean clutch size 27 -177.95 6.64 .63 .001 reproductive period*b date of first clutch* 27 13.10 -0.11 .07 .2 26 10.50 0.02 .03 .5 date of last clutch 27 238.98 -0.79 .06 .2 total number of clutches 27 10.41 -0.11 .07 .2 total number of eggs 27 -171.05 17.88 .08 .2 mean interclutch interval(d)c,d 25 1.89 0.02 .09 .2 summer growth rate(mm/day) •27 0.08 -0.01 .02 .5 a-square root transformation c-outlyers removed b-zero values removed d-log transformation 35 O to CO Figure 8. Mean clutch size of an individual plotted against body size. The regression lines for both populations is significant at p = .001 (solid - Angus Campbell, dashed -Lewis Slough). Slopes are indicated on the lines. The r 2 value for Angus Campbell is .56, and for Lewis Slough it is .63. Dots - Angus Campbell, triangles - Lewis Slough. 36 40 50 60 body size (mm) 70 F i i g . U u N ?• t 6 1" 8^ °la? ^dividuals reproductive period ( days between first and last clutch) plotted against body size. Dots - Angus Campbell, triangles -Lewis Slough. c 37 o LO CvJ O o CNJ sz o o o CO TJ to h o o o LO 30 40 50 60 body size (mm) 70 Figure 10. Date of first clutch (as numbered from the first of January) plotted body size. Dots - Angus Campbell, triangles - Lewis Slough. 38 O i n CM o o CM o o GO O CO LO o CD •4—* "D O O O LO 30 40 50 body size (mm) 60 70 Figure 11. Date of last clutch (as numbered from the first of January) plotted against body size. Dots - Angus Campbell, triangles - Lewis Slough. 39 O • A 00 - A • A A itches A • • clu CO — A A A A • o CD A A AAVv. A mb c • A ^ • A A • • • •4—» o • • • A A A A i • • C\J A • A» • • • • A • • • o 1 ' 1 1 1 30 40 50 60 70 body size (mm) Figure 12. Total number of clutches produced in the reproductive season plotted against body size. Dots-Angus Campbell, triangles - Lewis Slough. 40 CD J Q E r3 rz 75 o o o o o cvj o o o o o CO o o CD o o o o — • A A A • A A A A * .A • A A A A • LA A A A • ft A *A A • A • • • A A • • . A • A . A • - A • A • • A-T 1 • 1 1 l 30 40 50 body size (mm) 60 70 Figure 13. size. Dots Total number of eggs produced in a reproductive season plotted against body - Angus Campbell, triangles - Lewis Slough. 41 Figure 14. Mean interclutch interval of an individual plotted against body size Outlyers have been removed. Line represents a locally weighted regression line on combined data from both populations. Dots - Angus Campbell, triangles - Lewis Slough. 42 O a _ 1 0 rt co JC o *-» 3 o o i— • Q ) C O c c rt E co o LO c\i o C M A • A A • • 0 . 0 2 ^ t ^ ' A A A — A • • «A • • • ^ r * ^ A AA • A A A • A >*^*^A A A • A • A ' A • r i 1 1 1 30 40 50 60 body size (mm) 70 Figure 15. Log of an individuals mean interclutch interval plotted against body size. Outlyers removed. Regression for Angus Campbell is significant at p = .05 (solid line, r2 = .24). Regression for combined data from both populations is significant at p = .01 (dotted line, r2 = .16). Slope is indicated on the line. Dots - Angus Campbell, triangles - Lewis Slough. 43 Figure 16. Square root of summer growth rate plotted against body size. Fig. 16(a) -Angus Campbell; regression is significant at p = .001, r 2 = .50. Slope is indicated on line. Fig. 16(b) - Lewis Slough (slope is not different from zero). 40 50 body s i z e (mm) b) L E W I S S L O U G H A A A A A A A A A 30 40 50 body s i z e (mm) i 60 70 45 (1967), and Kynard (1972) who report the expected instantaneous fecundity of a 50mm female to be 133 for the low plated morph, 154 for the partially plated morph, and 163 for the fully plated morph. The greatest clutch size for females in this size range is reported by Crivelli & Britton (1987) who estimated an average 259 eggs for an average female of 52.3mm. This instantaneous fecundity relationship with body size is not translated into a significant seasonal fecundity advantage for large females (Fig.13). Small females achieve a greater seasonal fecundity than would be expected from the slope in Figure 8, and large females achieve a lesser fecundity. The maximum number of eggs appears to be produced by mid-sized females. Figure 14 represents the mean interclutch interval plotted against body size. A robust locally weighted regression line calculated on the data (Cleveland (1979)) indicates a curvilinear relationship roughly approximating an exponential curve, therefore the values were log transformed (Fig. 15). A significant regression is indicated for Angus Campbell fish (p = .05), but not Lewis Slough fish (p = .2). An ANCOVA indicates that the regression slopes, intercepts, and estimates of Y at two body sizes in the two populations are not significantly different (p>.5)(Table 8). Therefore, data from both populations were pooled and a significant regression at a probability of .01 was obtained (df = 43, intercept = 1.82, slope = .02, r2 = .16). Interclutch interval increases with increasing body size: a 40mm female has an average interval of about 13 days whereas a 50mm female averages about 18 days. This relationship between interclutch interval and body size translates into a tendancy for large females to produce few clutches; however, Figure 12 makes it clear that small fish can have either a few or many clutches. The minimum interclutch interval observed in this study is greater than the 3-4 days recorded by Wootton & Evans (1976), and by McPhail (personal communication) for Paxton Lake females. The direction of this relationship is opposite to the findings of Wootton (1977). He determined that larger females produced more clutches more frequently than smaller females. However, his differences in female size were obtained by 46 Table 8. Probability values for a significant difference in reproductive characters between females from the two sites. Included are tests for the difference in slopes obtained from regressing characters against body size (covariate), estimates of Y at 40 mm and 50 mm, and where a non-significant regression slope was found, a comparison of means. All comparisons are tested with a Student's t-test. body size(covariate) clutch size mean interclutch interval(d)c,d total number of clutches total number of eggs length of reproductive period(d)*,b date of first clutch* date of last clutch summer growth rate(mm/d)' Slope >.5 >.5 .005 <001 <.001 <.001 >.5 <.001 estimates of means Y at X = 40 X = 50 .06 >.5 >.5 ->.5 >.5 -.05 .05 .002 .1 .005 .01 .2 >.5 >.5 .05 <.001 <001 <.001 <001 <.001 <.001 <.001 -a-square root transformation b-zero values removed c-outlyers rembVed d-log transformation rearing fish on different food regimes. Thus, his small females were the product of a history of low ration, and were food stressed during the reproductive period. Consequently, his results are confounded by food effects and are not completely comparable to my study. It is unlikely that ration had an affect in my experiment because all fish were fed to excess, and all but one fish grew in length during the reproductive season. Figure 16 presents the square root of summer growth rate as a function of body size. Angus Campbell females (Fig. 16(a)) show a significant negative relationship with body size (p<.001). Because of length/volume considerations this is the expected observation in fish with indeterminate growth (Ursin 1979). In Lewis Slough females, however, growth rate appears to be independent of body size (Fig. 16(b)). This absence of a relationship may be partially due to a lack of large females in Lewis Slough, since it is the large females that depress the line in the Angus Campbell plot. . 47 The data were tested to determine differences in the suite of reproductive characters resulting from population source. Table 8 gives the results of the between population comparison of slopes obtained from the regression analysis, along with a comparison of two estimated Y values for each character, and two-tailed t-tests on means for those characters not significantly linear with body size. A two-tailed Student's t-test was also performed to compare the mean body sizes of the females used in the experiment. This indicates that the mean body size of the experimental fish did not differ at a significance level of .06. This probability is very close to the standard rejection level of .05. Therefore the possibility that the body sizes in the two experimental populations are different should not be totally dismissed, particularly since the standard rejection level is arbritrary. As noted earlier the relationships between body size and clutch size, and body size and interclutch interval do not differ between the two populations. There is, however, a difference between the populations in average female body size. Thus, it is likely that any observed difference between the two populations in clutch size and interclutch interval results not from divergence between the populations but from this difference in mean female body size. In contrast, the total number of clutches produced per population does differ (slopes at p=.005, Y estimates at p=.05, and means at .002). Lewis Slough females produce more clutches over the summer than Angus Campbell females (Fig.12). Lewis Slough females appear to also continue producing clutches for a longer period of time than Angus Campbell fish (Fig.l 1). When reproductive period is regressed on body size the slopes differ between the populations (p<.001); however, neither the Y estimates (p=.2, and p=.5) or the means (p>.5) are different. If zero values (fish that had only one clutch and therefore no reproductive period could be calculated) are included and a one-tailed normal approximation to the Mann-Whitney U test performed (Sokal & Rohlf 1981 pg.436), then the populations do differ (p=.005). The regressions of total number of eggs produced on body size (Fig. 13) also give slopes that differed between populations (at p<.001), but estimates of Y at the two different sizes yield different results. At a body size of 40mm the estimate of total egg number is 48 the same (p=.l), yet at 50mm it is different (p=.05). A t-test on means suggests that overall there is a significant difference in total egg production between the two populations (p=.01). I conclude that Lewis Slough females are capable of producing more eggs over the reproductive season than Angus Campbell females. Dates of first (Fig.9) and last clutch (Fig. 10) also differ between the populations (p<.05) for both the estimates of Y and the means. The regression slopes, however, differ only for the date of first reproduction (p<.001) and not for the date of last reproduction (p>.5). This analysis indicates that Angus Campbell fish tend to start breeding earlier and stop breeding sooner than Lewis Slough fish. As suggested from the regression analysis the slopes for growth rate over the summer (Fig. 16) differ significantly (p<.001) between the two populations, as do the two Y estimates at 40 and 50 mm body sizes (p<.001). This indicates that Lewis Slough females in this size range are growing at a faster rate than their Angus Campbell counterparts. Correlation analysis was used to investigate how this suite of reproductive characters covary and to determine their interaction with total egg production (a measure of reproductive success). The results are given in Table 9. The only character significantly correlated with body size in both populations is clutch size (p<001). In the Angus Campbell population growth rate over the summer is negatively correlated with body size (p<.001), as is the log of the mean interclutch interval (p=.05). Combined data from both populations shows a correlation (p=.01) between log of the mean interclutch interval and body size. The total number of eggs an individual produces over the reproductive period is positively correlated with the number of clutches the fish produces, and to a lesser extent (Angus Campbell at p=.05, and Lewis Slough at p=.01) with the average clutch size. Thus an increase in the total number of eggs produced is most effectively achieved by increasing the number of clutches than by increasing clutch size. The number of clutches is correlated with the length of an individual's reproductive period (p<.001) (Fig. 17). Thus, the time that elapses between an individual's first and its' last clutch is determined by the number of 49 Table 9. Matrix of correlation coefficients for reproductive characters measured in each population. Significance levels indicated by asterisks, * - .05; ** - .01; *** -.001. 1 2 3 4 5 6 7 8 9 A 1) S i z e (mm) - -.203 -.206 .116 - .248 .490* .310 *** .745 *** -.641 N 2) Date o f F i r s t C l u t c h 3 - -.131 *** -.643 - .538 -.156 *** -.609 - .090 .215 G U 3) Date o f L a s t C l u t c h *** .748 *** .675 .335 ** .556 - .207 -.187 S 4) L e n g t h o f Re p r o d u c t i v e P e r i o d ( d ) a ' ^ - *** .701 .376 *** .792 .080 -.287 c A M 5) 6) T o t a l Number o f C l u t c h e s , , , c , d Mean I n t e r c l u t c h I n t e r v a l IQJ -.325 *** .711 --.055 .184 .115 -.028 -.253 P B 7) T o t a l Number o f Eggs - .385 -.249 L 8) C l u t c h S i z e - -.281 L 9) Summer Growth Rate (mm/d)a -1 2 3 4 5 6 7 8 9 1) S i z e (mm) - .107 -.244 -.338 -.258 .300 .292 *** .793 -.184 L E 2) Date Qf F i r s t C l u t c h * - .246 *• *** -.823 *** -.685 .208 ** -.544 - .020 .081 W 3) Date o f L a s t C l u t c h !*** .686 .435 .281 .418 - .034 -.200 I S 4) Leng t h o f Rep r o d u c t i v e P e r i o d ( d )a ' ^ _ *** .714 .047 ** .560 - .072 -.090 s 5) T o t a l Number o f C l u t c h e s ** -.528 *** .763 - .088 -.257 L f~\ 6) c d Mean I n t e r c l u t c h I n t e r v a l (d) ' - .424 .244 .153 u U 7) T o t a l Number o f Eggs -** .540 -.301 G H 8) C l u t c h S i z e - .056 9) Summer Growth Rate (mm/d)a — a -square r o o t t r a n s f o r m a t i o n c - o u t l y e r s removed b - z e r o v a l u e s removed d - l o g t r a n s f o r m a t i o n © 51 clutches produced, and not because the interclutch intervals are long. A robust, locally weighted regression line calculated on the data indicates that this relationship is curvilinear. Initially reproductive period increases as clutch number increases, but the relationship reaches a plateau at about 75 to 100 days. This suggests that a further increase in clutch number must be accompanied by a decrease in interclutch interval. This relationship is observed in Figure 18 where the robust locally weighted regression lines show an initial plateau and then a decline with increasing clutch number. These two characters are not correlated in Angus Campbell females but are correlated in Lewis Slough fish (p=.01). The length of the reproductive period is also significantly correlated with the date of first and last clutch (p=.001) in both of the populations. Thus, if an individual starts breeding early it tends to continue breeding for a longer period of time. A summary of the significant correlations between reproductive characters in each population is illustrated diagrammatically in Figures 19 and 20. 5 2 Figure 17. Length of an individual's reproductive period (time between first and last clutch) plotted against total number of clutches. Curves indicate robust locally weighted regressions (solid line - Angus Campbell, dashed line - Lewis Slough). Dots - Angus Campbell, triangles - Lewis Slough. 53 Figure 18. Mean interclutch interval of an individual plotted against total number of clutches. Curves indicate robust locally weighted regressions (solid line - Angus Campbell, dashed line - Lewis Slough). Dots - Angus Campbell, triangles - Lewis Slough 54 ANGUS CAMPBELL length of reproductive period I total number of clutches total number of eggs clutch size I body size I sqrt summer growth rate log mean Interclutch Interval ZT Figure 19. Diagrammatic summary of significant correlations between reproductive characters in females from the Angus Campbell site. Fat black lines indicate a significance at rx.OOl, medium striped lines significant at p=.01, and thin white lines significant at p=.05. No line represents p>.05. (-) indicates a negative correlation. 55 LEWIS SLOUGH body size sqrt summer growth rate Figure 20. Diagrammatic summary of significant correlations between reproductive characters in females from the Lewis Slough site. Fat black lines indicate a significance at rK.OOl, medium striped lines significant at p=.01, thin white lines significant at p=.05. No line represents p>.05. All correlations are positive. 56 GROWTH UNDER LABORATORY CONDITIONS Table 10. Summary of Analysis of Covariance on linearized growth curves of laboratory bred and raised fish. Source df SS F-ratio p-level population 1 0.197 1.92 .167 tank(population) age * replicate 9 3.199 3.45 .0004 10 2.906 2.82 .002 residual 737 75.906 Figure 21 illustrates change in length with age of fish raised from eggs in an environment chamber. The results of the ANCOVA on the linearized growth curves (Table 10) indicates that size, when adjusted for age, is not different (p=.167) between the two populations. Both the Bonferroni and Multiple Range a posteriori tests support this conclusion (p>.05). There are, however, significant tank effects associated with age corrected size (p=.0004). Both the Bonferroni and Multiple Range tests show two homogeneous subsets shared between the two populations (at p=.05), with replicates 2, 3, and 5 (all Angus Campbell tanks) breaking the sets apart. This tank effect is interpreted as primarily due to the result of slightly different intercept values. The analysis also shows a difference in the age by replicates (slope) at a probability of .002 : however the a posteriori tests indicate that the slopes constitute one homogeneous subset. This apparent discrepancy occurs because the a posteriori test takes into account the standard error around the lines and this tends to eliminate the differences found by the ANCOVA. The slopes and standard errors of the linearized growth curves for each tank in a population is presented in Table 11. 57 30 20 10 u 1 1 1 1 1 i i I 0 20 4 0 60 80 100 120 140 160 AGE (days) Figure 21. Growth in laboratory bred and raised fish. Each line represents a tank (replicate). Each point prior to 100 days represents the mean size and standard error of a tank sub-sample, while points at 100 and 140 days are the means and standard errors of all fish within a tank. 58 Table 11. Slopes and standard errors for linearized growth curves of each replicate within a population. The calculated common slope for all tanks is .017 with a standard error of .240e'\ Angus Campbell Lewis Slough  1 2 3 4 5 6 1 2 3 4 5 slope .017 .015 .019 .017 .018 .018 .016 .017 .016 .017 .015 S.E. .761 .758 .910 .868 .845 .813 .733 .800 .868 .769 .746 (no-3) n 79 83 56 59 58 63 83 67 66 71 74 I conclude that when raised under identical laboratory conditions fry from the two populations do not exhibit an overall difference in growth rate during the early period of their life. The mean size and standard error of each sex within a population at the termination of the experiment is shown in Table 12. By the age of 140 days these fish have attained a size just short of the natural (field) breeding size of Lewis Slough females. Table 12. Mean size (mm.), standard error, and sample size for each sex within a population in laboratory bred and raised fish at the end of the experiment. mean S.E. n Angus Campbell female 35.4 .51 24 Lewis Slough female 34.7 .64 21 Angus Campbell male 36.0 .37 31 Lewis Slough male 35.1 .57 26 A two-way ANOVA performed on the data in Table 12 examined the sex by population interaction. The results are presented in Table 13. They indicate no significant effect of sex (p=.332), population (p=. 129), or sex by population interaction (p=.838). Therefore any inherent sexual dimorphism in size in these two populations is not evident under laboratory conditions by this age. The growth of wild-caught fry raised in the laboratory to maturity under simulated conditions is shown in Figure 22. The fry used in this experiment were between 16 and 59 Table 13. Summary of 2-way ANOVA for the effects of sex and population on body size in laboratory bred and raised fish. source df SS F-ratio p-level model sex population sex * population error 3 1 1 1 98 0.230 0.060 0.156 0.003 6.525 1.15 .332 0.90 .345 2.34 .129 0.04 .838 20mm in length when they were caught. They were one to two months old when captured if they were growing at a rate similar to that estimated for laboratory bred fish. This size range, however, can be observed in laboratory bred fish at a single age, therefore for the purpose of graphing the data a conservative estimate of age (45 days) was made. One or two individuals from each tank were measured from the ages of 80 to 159 days. These data are plotted as individual points on the graph (Fig. 22). Overlaid on these data is the generalized growth curve from Figure 21 with the mean size and standard error of each sex within each population at the termination of the experiment (data from Table 12). The curve from laboratory bred fish and the points from wild-caught fry appear to closely coincide. At the ages of 220, 259, and 323 days all individuals in each tank were measured. The mean and standard errors for each population at these dates are also plotted on Figure 22. The mean body size of the two populations appears to start diverging after the age of 200 days. At the termination of the experiment (day 323) the fish were killed, measured and sexed. The mean body size and standard error for each sex within a population is plotted and it appears that the body size of Lewis Slough females is much smaller than either Lewis Slough males or Angus Campbell fish (all of which are similar). These small females are responsible for reducing the mean size of Lewis Slough fish relative to Angus Campbell fish. 60 Figure 22. Growth of wild-caught fry raised to rfmturity in an environment chamber under simulated seasonal conditions. The, mark at 45 days indicates size range of fry used in the experiment. Three subsequent dates show sizes of individuals sampled from each tank (x - Angus Campbell, o - Lewis Slough). Overlaid is a dashed line representing a generalized growth curve from Figure 21, with the mean size and standard error of fish from each sex within a population at the termination of that experiment (from Table 10). The two subsequent dates show the mean size and standard error of all individuals in a tank within a population. The final date indicates the mean size and standard error of each sex within a population at the end of the experiment. AC - Angus Campbell, LS - Lewis Slough. See text for further details. BODY SIZE (mm) 62 Table 14. Summary of 2-way ANOVA for the effects of sex and population on body size in wild-caught and laboratory raised fry. source df SS F-ratio p-level model 3 3.482 11.88 .0001 sex 1 0.710 7.26 .0088 population sex * population 1 1.174 12.01 .0009 1 2.058 21.01 .0001 error 69 6.742 The summary of a two-way ANOVA performed for sex and population effects on these data is presented in Table 14. There appears to be a significant effect in all categories. The results of a Least Means a posteriori test between all cells is shown in Table 15, along with the mean, standard error and sample size of each cell (data used in Fig.22). The Least Means test shows that Angus Campbell males, females, and Lewis Slough males are not different from each other (with a minimum probability level of .15 between Angus Campbell males and females), but Lewis Slough females are significantly different from these three (with a maximum probability of .0002). Table 15. Mean body size (mm), standard error, and sample size of each sex within a population with probability levels for a Least Means a posteriori test on the cells. AC - Angus Campbell, LS - Lewis Slough. p - level  _n mean S.E. ACfemale LSfemale ACmale LSmale AC female 29 43.0 .94 - .0001 .1543 .5273 LS female 12 36.8 .99 - - .0002 .0001 AC male 14 41.5 .99 - - - .4379 LS male 18 42.4 .52 - -63 Thus, given identical laboratory conditions, both sexes of Angus Campbell fish and Lewis Slough males tend to grow along'a similar trajectory. In contrast, Lewis Slough females follow the same trajectory early in life but then begin to grow more slowly so that after about one year they are significantly smaller than the other fish. Note, however, that if fish are raised individually they achieve a much greater body size, and this includes Lewis Slough females. In addition, the growth pattern of none of the fish appears to approach a classic asymptote in size over the time observed. Wild-caught fry raised in the laboratory were brought to maturity under simulated temperature and photoperiod conditions. This provides information on inherent differences in time of maturity and body size at maturity. Both are important components of life-history theory. Figures 23 and 24 show the distribution of fiming of first reproduction for the two populations. The X axis represents the experimental day at first ovulation. Below this the experimental photoperiod length at that date is recorded, and this is translated to the corresponding natural date of that photoperiod at the latitude of the sites (information from Environment Canada 1987 sunrise sunset Tables). The distributions appear to be roughly similar but there is some indication that Lewis Slough females mature earlier (or at a shorter photoperiod). The non-parametric one-tailed Mann-Whitney U test was used to test the null hypothesis of equality between the times of maturation because the distributions were not normal, and transforming the data did not appear to achieve normality. The result of the test indicates a significant difference (p<.05) between the two populations, with Lewis Slough females maturing earlier than Angus Campbell females. This result is contrary to the prior observations that females at the Angus Campbell site tend to start their breeding earlier in the season than females at the Lewis Slough site. There is no linear relationship between size and timing of reproduction either in Angus Campbell or Lewis Slough females (p >.05, r .^007). Both graphs show a second mode, and perhaps some of these second mode fish represent smaller individuals that invested in growth (or had to grow) before becoming reproductivly mature. 64 Figure 23. Distribution in timing of first ovulation in Angus Campbell females, wild-caught as fry and raised in the laboratory under simulated photoperiod and temperature conditions. The X-axis represents the experimental day of first ovulation translated to the photoperiod regime at that day, further translated to the natural date corresponding to that photoperiod at the latitude at which the fish were sampled. 255 265 275 285 295 305 315 TIME (days) 12:15 13:00 14:00 14:45 15:30 1 6 : 0 0 » - > - PHOTOPERIOD (hours of illumlnatlon/24) 3/22 4/3 4/20 5/4 5/20 8/4*%-*- CORRESPONDENT DATE OF PHOTOPERIOD 66 Figure 24. Distribution in timing of first ovulation'in Lewis Slough females, wild-caught as fry and raised in the laboratory under simulated photoperiod and temperature conditions. The X-axis represents the experimental day of first ovulation translated to the photoperiod regime at that day, further translated to the natural date corresponding to that photoperiod at the latitude at which the fish were sampled. T 255 265 275 285 295 305 315 12:15 13:00 14:00 14:45 15:30 16:00~»-*» 3/22 4/3 4/20 5/4 5/20 6 / 4 » - * * TIME (days) PHOTOPERIOD (hours of illumination/24) CORRESPONDENT DATE OF PHOTOPERIOD 68 Table 16. Total number of wild-caught and laboratory raised females from each site catagorized into maturity status and two size classes. Maturity status determined by whether or not a female had ovulated a clutch of eggs. Size classes chosen because 37 mm is the size below which no Angus Campbell female ovulated eggs. Data are shown as the number of individuals out of the total for that site, as well as the percentage of the total. Angus Campbell Lewis Slough # % # % mature >37 mm 42 50.0 21 53.8 <37 mm 0 0 9 23.1 immature >37 mm 33 39.3 6 15.4 <37 mm 9 10.7 3 7.7 TOTAL 84 100 39 100 Fifty-one percent of the total number of Angus Campbell females raised (all densities; ranging from 1 fish per tank - 25 fish per tank) ovulated, while 78% (same range of densities) of Lewis Slough females ovulated. There was no difference in the size of Lewis Slough females that ovulated and those that did not (Student's t-test, df=37, p=.225). There was also no difference in Angus Campbell females at p=.08 (df=82). This probability level is close to the conventionally prescribed level of .05, and therefore there may be biological, if not statistical, significance. Again the suggestion is that Angus Campbell females may invest more in growth before becoming reproductivly mature than Lewis Slough females. This proposition is supported by data from Table 16 which implies that there is a higher size threshold for sexual maturity in Angus Campbell females than Lewis Slough. Of the total number of Lewis Slough females raised, 31% were less than 37mm at the end of the experiment; 75% of these produced a clutch of eggs. In contrast, 10.7% (9 fish) of the total number of Angus Campbell females raised were less than 37mm at the end of the experiment and none of these ovulated. Since 51% of Angus Campbell fish ovulated in the laboratory, by proportions four or five of these nine fish should have matured. 69 GROWTH AND AGE STRUCTURE IN THE FIELD Figure 25 shows the size frequency histograms for Angus Campbell females, males, and unknowns for spring 1986 and spring 1987. During May 1986 in the midst of the breeding season, the majority of individuals were between 40 and 50mm except for a few females that skew the graph towards the 50 to 60mm region. Because of low water level trapping became difficult in subsequent months and the next sample of substantial size was not obtained until November. These were brought back to the laboratory and sexed by dissection. At this time a large number of fish were in the 30 to 40mm range. Because of their absence in the earlier sample, I interpret these as fish bom during the summer of that year. Fish in the 40 to 50mm range are still present; however, the females that were in the 50 to 60mm range in the last sample are not evident. I assume they died or left the area. The histogram of sizes of fish caught in February (also sexed in the laboratory) shows roughly the same distribution but more fish are in the 40 to 50mm region, and some individuals are now 50 to 60mm long. This distribution indicates the growth of 30 to 40mm fish (young from the previous summer) into the 40 to 50mm range, but also includes some adults from the last summer. Adults from the previous summer also experience growth and this fills in the 50mm length region. By April, when breeding began, those females in reproductive condition were in the upper 40mm range. Males were somewhat smaller than this, while the fish that could not be sexed were distributed in the 30 to 40mm range. The small fish were probably young from the last summer, and the breeding fish are primarily composed of individuals that reached that size region in February. Thus, adults from the previous summer were entering a second year of reproductive activity. A spurt of growth appears to have occurred between April and May in the females and fish of unknown sex, but not the males. In May both the males and the unknowns 70 Figure 25. Size-frequency histograms for Angus Campbel l males, females, and unknowns as proportion of catch obtained at sampling dates spanning f rom May 1986 to May 1987. If fish were sexed in the field, and an individual 's sex could not be confidently determined, then that individual was categorized as an unknown. MAY 1966 71 NOV. 1986 FEB. 1987 J C o CO o c o o a o .25 .20 .15 .10 .05 0 10 20 30 40 50 60 70 1 0 20 30 40 50 60 70 10 20 30 40 50 60 70 body size (mm) b o d y 8 l z e ( m m ) n=36 J C o as o c o o a o .25 .20 .15 .10 .05 0 .50 .40 .30 .20 .10 0 10 20 30 40 50 60 70 body size (mm) APRIL 1987 n»36 MAY 1987 n»75 10 20 30 40 50 60 70 10 20 30 40 50 60 70 na12 rl«29 .25 .20 .15 .10 .05 0 10 20 30 40 50 60 70 10 20 30 40 50 60 70 .30i .25 n-109 na 70 10 20 30 40 50 60 70 10 20 30 40 50 60 70 body size (mm) body 8 l z e ( m m ) 72 were distributed in the 40 to 50mm range, and there were no smaller fish. In females there is a definite second mode around 60mm of size. This distribution could result from growth of females that were in the unknown sex category in April into the 40 to 50mm size range by May. At this time they begin their first season of breeding, while females of 40 to 50mm that were breeding in April have now grown into the 60 mm region. It is possible, however, that these larger females could have just returned to the site. During the reproductive season at the Angus Campbell site apparently there are at least two age classes of breeding females and this is reflected in their size distribution. In contrast, males appear to be growing determinantly (or at a much lower asymptote than females) so that size structure does not reflect age structure. Small males in the 30 to 40mm region in November can be observed growing into the 40 to 50mm region in February and April. There is, however, no observed growth of males beyond the 40 to 50mm range, and there is no reduction of individuals in this size region during the autumn as would be expected if this were an annual population of males (as will be seen in Lewis Slough fish, (Fig.26)). Further evidence that females, and not males, experience a growth spurt in early spring is provided by the marking experiment (Table 17). Of the 295 fish marked on March 16 1987, 52 were recaptured on April 8 (total catch corresponds to the April histogram in Figure 23), and 11 of these grew into the next size category. All of these (except for one fish which could not be sexed) were female. However, a total of only 7 males were caught, so these results on their own are not conclusive. They are, however, supported by similar results obtained from the catch on May 25. In addition, Crivelli and Briton (1987) also observe a growth spurt in the spring in their females. An alternative explanation for the distribution of females in Angus Campbell is that there is only one age class, but two growth forms: one fast growing and one slow growing. This could give rise to the two size modes; however under laboratory conditions no evidence of different growth forms was found. A contrasting pattern of size distributions over time is observed in Lewis Slough 73 Table 17. Results of marking experiment. 295 fish were caught on March 16 1987 and marked by clipping their dorsal spines. Those fish greater than 40mm. had their first dorsal spine clipped, while those less than 40mm. had their second dorsal spine clipped. April 8 and May 25 represent total number of fish caught and of those fish the ones supporting marks. Each mark category is sub-divided into size classes and sex. Those fish in the third hierarchical category represent individuals who have grown into the next size class. number of fish with clipped spines . first spine second spine  >40 mm <40 mm >40 mm date total female male ? female male ? female male ? 4/8 232 20 5 1 4 2 9 10 - 1 5/25 166 6 2 2 - 2 1 7 - 4 fish (Fig. 26). During July, in the midst of breeding, most males were between 40 and 50mm in length, while the females and unknowns were somewhat smaller. In October no fish could be accurately sexed in the field, but the size distribution of the total catch shows a distinct bi-modality. The smaller mode is interpreted as young produced in the summer and the second mode as adults from the summer. By November individuals (sexed in the laboratory) in the 40 to 50mm region have almost entirely disappeared and the size of fish present corresponds roughly to the smaller mode seen in October. In December, and through to early spring, no fish were caught in the traps. Apparently, they either left the area or were not active enough to be trapped. The next catch of significant numbers occurred in April. At this date individuals were distributed between 30 and 50 mm, indicating growth from November. The observed size range in May is similar to that in April although some individuals grew into the 40mm length category. This contrasts with the large jump in size observed with Angus Campbell females and unknowns during this time period. There was little change in size distribution between May and June. 74 Figure 26. Size-frequency histograms for Lewis STough males, females, and unknowns as proportion of catch obtained at sampling dates spanning from July 1986 to June 1987. If fish were sexed in the field, and an individual's sex could not be confidently determined, then that individual was categorized as unknown. 75 sz o CO o c o o a o i_ a .30 .25 .20 .15 .10 .05 0 .30 .25 .20 .15 .10 .05 0 .30 .25 .20 .15-1 .10 .05^  0 JULY 1986 OCT. 1986 .35 .30 .25 .20 .15 .10 .05 10 20 30 40 50 60 70 NOV. 1986 ns51 10 20 30 40 50 60 70 .25 .20 .15 .10 .os-i o 20 30 40 5b 60 7b .20 n»68 n=48 .15 .10-1 .05 n=82 10 20 30 40 50 60 70 b o d y size (mm) 10 20 30 40 50 60 70 10 20 30 40 50 60 70 b o d y size (mm) b o d y size (mm) SZ o CO o c o o a o .30 .25 .20 .15 .10 .05 • 0 APRIL 1987 MAY 1987 n-41 JUNE 1987 n = 36 10 20 30 40 50 60 70 I c T l ^ C - l o ^ ^ ^ i ^ V l r f ^ ^ h a T T b .40, .30 .20 .10 0 n = 24 .20 .10 .05 na15 .40 .30 .20 .10 0 n=36 10 20 30 40^ 0~io~7o 10 20 3'0 4 T ^ c r 7 o b o d y size (mm) 10 20 30 40 50 60 70 b o d y size (mm) .50 .40 .30 .20 .10 0', n=*64 10 20 30 40 50 60 70 body size (mm) 76 To summarize, few if any adults from Lewis Slough seem to live from one summer to the next. The size class representing adults in October completely disappears in November. The subsequent shifts in size distribution with time can be interpreted as growth of young over the winter and spring and into the breeding season. By this time they resemble last year's adults in body size. These interpretations lead to the conclusion that the Lewis Slough population differs from Angus Campbell in that Lewis Slough adult females are primarily composed of one age class, unless large individuals leave the system before November and exhibit virtually no growth until they return around April. This seems unlikely in light of the growth observed in Lewis Slough females held in captivity over the winter period. These data only span a one year period; however, samples from other dates in other years fit the same pattern (although the means can*shift somewhat). These data have not been presented because they are incomplete. 77 DISCUSSION AGE-STRUCTURE AND GROWTH PATTERNS Understanding divergent patterns of sexual size dimorphism within a species requires a knowledge of the processes that give rise to the differences. Dimorphism in size can be produced through a variety of mechanisms: sex specific growth rates, differential mortality rates, and different growth forms (e.g. determinant versus d^eterminant growth, or growth asymptotes). All of these could conceivably be subject to selection. Interpreting the significance of size dimorphism depends upon knowing if selection has occurred, and if so, on what traits and in which sex. The contrasting direction of sexual size dimorphism in the two study populations appears to arise from two separate mechanisms. The large size of Angus Campbell females is a function of a sex specific growth pattern and the age structure of the population. Many fish in this population apparently survive for more than one breeding season and females appear to continue growing throughout their life. In contrast, males either exhibit deterministic growth or grow toward a much lower asymptote than females. My experiments did not continue long enough to distinguish between these two alternatives. Thus, further research is required to clarify this point. Determinate growth in males, however, is indicated by the observations of Crivelli & Britton (1987). They determined that males in a population of Mediterranean stickleback stopped growing after reaching sexual maturity. The females in this population, however, continued to grow after maturity, and also had a higher growth rate prior to maturity than the males. Thus a female bias in sexual size dimorphism 78 became evident at an early age. Laboratory experiments on Angus Campbell fish indicate that up to the age of first maturity there is no inherent difference between the sexes in growth rates. In the field, however, sexual size dimorphism becomes evident earlier because of a growth spurt experienced by females during the early spring just prior to the breeding season. Lack of such a growth spurt in laboratory raised females indicates that this spurt is environmentally induced. Males do not experience this period of rapid growth, perhaps because they have already reached the end point of their growth. Alternately, there may be large energy demands on mature males at this time of year because of territory acquisition, defense, and nest building, all of which divert energy from growth, or inhibit feeding. This suggestion is weakened, however, by the observation that males raised in the laboratory where territorial and nest building behaviour did not occur, and where feeding continued, still corresponded closely in body size to the size of males observed in the field. If determinant growth in stickleback males is a general phenomenon, then a female bias in size dimorphism would be predicted in age structured populations. Lewis Slough is an annual population. Here monomorphism in size is predicted if sex specific growth patterns were the only mechanisms operating. My growth studies indicate, however, that there is an inherent decrease in growth rate in Lewis Slough females as they approach maturity. Thus, dimorphism in size only emerges as the fish age and is not evident during the early stages of their life. Mature Lewis Slough females held in captivity until a second breeding season continued to grow, but it is not known if growth in Lewis Slough males mirrors that of Angus Campbell males. No males from this population were kept for more than one season and it is unlikely that two year old males occur in nature. Certainly, growth rate and pattern of growth are traits in these 79 populations which selection could act on to produce sexual dimorphism in size. Lewis Slough females are smaller than Lewis Slough males, and in this case small size appears to have been selected for in females because their growth rate as they approach maturity deviates from the trajectory typical of Lewis Slough males and Angus Campbell fish of both sexes. In Angus Campbell females are larger than males and it is not clear if it is female size that has shifted up or male size that has shifted down. The ramifications of male body size and reproductive success are not clear in the threespine stickleback. Studies have shown that there is no correlation between male size and territory quality, reproductive success and fry care (van den Assam 1967; Kynard 1972; Presley 1976; Sargent 1982) and this suggests that male size is not important to reproductive success. However the variance in body size in these experiments is small, as it tends to be relative to female variance (Table 1, J.D. McPhail pers.comm), which may imply that stabilizing selection has confined body size to a very narrow range. Rowland (1983) investigated interspecific aggression and dominance between Gasterosteus  aculeatus and Gasterosteus wheatlandii and found that the larger CL aculeatus (mean standard length 4 9 - 5 1 mm) males displaced the smaller CL wheatlandii (mean standard length 32 - 33) males from their nests and established their own territories. These results indicate the large penalty a small CL aculeatus male would experience in trying to establish and defend a territory against larger conspecifics (without which he could not successfully court and mate). Selection against large male size could result from these individuals being energetically less efficient at competing for territories and performing epigamic displays (as Searcy 1979 observed in red-winged blackbirds). Sex specific patterns of growth in CL aculeatus may not be the result of current selective pressure. Different growth forms could simply reflect phylogenetic history and past selective events. Data on sex specific growth 80 patterns in the threespine stickleback are lacking, yet they are essential to predictions about the direction of sexual size dimorphism. Data from other species in the family Gasterosteidae are required to determine if phylogenetic patterns exist (i.e. is determinant or d^eterminant growth the "derived character") before the role of selection in determining the extent of sexual dimorphism in size can be assessed. P O P U L A T I O N S P E C I F I C R E P R O D U C T I V E C H A R A C T E R I S T I C S A N D L I F E - H I S T O R Y E V O L U T I O N The two study populations differ in the direction of sexual size dimorphism (which sex is largest) and in age structure. In addition, my results suggest that females in the two populations have diverged with respect to reproductive and life-history parameters. I will attempt to discuss how these populations differ with respect to their reproductive traits, and whether this divergence is as predicted by life-history theory. Given a common environment and ad libitum diets, Lewis Slough females ovulated more clutches and produced more eggs over the course of the reproductive season than did Angus Campbell females. The total number of clutches is greater in Lewis Slough fish because a larger proportion of Angus Campbell females produced no, or only one clutch (0.27 versus 0.03). The relationship between body size and clutch size, and body size and interclutch interval does not differ between the two populations. When exposed to the same environment Angus Campbell females started producing clutches earlier in the season and ended breeding sooner than their Lewis Slough counterparts. This mirrors the field observation that Angus Campbell fish begin breeding earlier than Lewis Slough fish. Thus, these Angus Campbell fish cease producing 81 clutches earlier than Lewis Slough fish even when environmental conditions do not force them to stop. As a result, individual Lewis Slough females continue to ovulate clutches for a longer period of time than Angus Campbell females. If reproductive effort can be measured by the total number of eggs produced (Williams 1966; Vitt & Congdon 1978; and Cabana et aL 1982), then Lewis Slough females are investing more in seasonal reproduction than Angus Campbell females. These conclusions are derived from data collected on fish held under artificial conditions. Field values may differ as a result of such environmental variables as food availability and temperature; however, the results indicate the relative capability of each population given a common environment. There is as yet no satisfactory method for marking large numbers of stickleback in the field and monitoring the progress of individuals. Therefore, determining the number of clutches a female produces in situ is, at this time, impossible. Crivelli & Britton (1987) attempted to determine the number of clutches a female was capable of producing by counting pre-ovulatory egg stages. Under most circumstances this method probably underestimates the number of clutches because stickleback ovaries contain a large proportion of undifferentiated and immature eggs. Only a portion of these are recruited to pass through the maturation process as a series of clutches (Wallace & Selman 1979). My laboratory raised fish showed a significantly earlier date of first ovulation in Lewis Slough females. This difference is statistically significant; however, since it is only about five days it is not likely to be biologically significant in terms of early versus delayed maturity. More important is the observation that 51% Angus Campbell females reared in the laboratory did not reach maturity after a seasonal cycle; whereas under the same conditions 78% of the Lewis Slough females did mature. Upon dissection at the 82 end of the experiment it was found that many of the females that had not ovulated eggs (and thus were not classified as reaching maturity) did possess ovaries containing immature eggs. These results may be a laboratory artifact or they may indicate that a high proportion (about 50%) of the Angus Campbell females delay maturation for a year. An age-at-maturity of more than a year has been recorded in other populations of CL aculeatus (freshwater and marine: Greenbank & Nelson 1959; Munzing 1959; Aneer 1973). If female maturity in the Angus Campbell population is delayed, the delay is not because the ovaries remain in an immature state. In this population the ovaries develop and eggs begin to mature but they are not ovulated and may be re-absorbed throughout the course of, or at the end of, the reproductive season. Ovulation is basically a two-step process: the pituitary gland secretes gonadotropin which in turn causes the secretion of steroids that cause the eggs to ovulate (Wasserman & Smith 1978). Perhaps the delay in maturation is because the hormonal system is not fully developed. In both populations the females that did not ovulate were no smaller than ovulated females (although Angus Campbell females at a p=.08 are close to being significant at the standard rejection level of p=.05). There is evidence, however, that the minimum size for maturity in Angus Campbell females is larger than that for Lewis Slough, and size at maturity is known to be a heritable trait in the threespine stickleback (McPhail 1977). Life-history theory predicts suites of survivorship and reproductive traits that increase the probability of maximizing lifetime reproductive success in particular environments. In animals 'bet-hedging' and 'r & K selection' are the most popular theories, and at this time bet-hedging is particularly favored. Density dependant versus independent processes are the primary variables driving the evolution of life-histories under r & K selection. "K" selection is thought to occur under conditions of high population density 83 where resources are limited. The '"K" refers to the carrying capacity of a population. "K" selection is envisaged as favoring individuals with lower total resource or energy demands and therefore small body size, long life span, delayed reproduction, and low reproductive effort. Resources are thought not to be limiting when population densities are kept low as the result of random, catastrophic events in the environment. As a result increased resource acquisition can be used to enhance fecundity and survivorship via large body size and/or increased reproductive output. Under these conditions it is predicted that selection will favor large body size, early age at first reproduction, semeloparity and a high reproductive effort. Population size is determined primarily by "r", the intrinsic rate of increase. (MacArthur 1972; Steams 1975; Calow 1978, 1982; Green 1980; reviewed by Boyce 1984). Bet-hedging theory proposes age-specific mortality rates as the selective agent in the divergence of life-histories. Differential reproductive effort with age, balanced by the cost of reproduction, increases the probability of maximizing the numbers of surviving young. The cost of reproduction is defined as the general deleterious effect of present reproduction on future survival, or fecundity, or both (Pianka & Parker 1975; Bell 1980). High, variable, or unpredictable adult mortality rates select for increased reproductive effort early in life, therefore for early maturity and a short life span. Small adult body size is predicted as a consequence of an early age of maturity. Delayed maturity, low reproductive effort, and a longer lifespan are selected for by the same mortality conditions when they occur in juveniles (Steams 1975, 1976, 1980, 1983; Reznick & Endler 1982). The contrasting environmental conditions necessary for r and K selection appear to be present in the two study sites. The Angus Campbell site is subject to large fluctuations in water level and major perturbations from human 84 activity; whereas the Lewis Slough site is much more stable. As a result Angus Campbell fish should be more subject to "r" selection than "K", and Lewis Slough fish more subject to "K" selection than "r". Except for body size, however, the predicted direction of life history characters is not in agreement with theory. The suite of observed characters are, in fact, the reverse of predictions. For example, it would be predicted that Angus Campbell fish should have a higher reproductive output, earlier maturation, and a shorter life-span than Lewis Slough fish. My data, however, suggests that Angus Campbell females have a lesser reproductive effort, may delay maturity and mature at a greater size, and have a longer life-span than Lewis Slough fish. The pattern of age structure in the two populations resembles one which could result from the age related mortality "schedules described in bet-hedging theory. Mortality rates were not measured in the field but it appears that adult fish in Lewis Slough do not survive to a second year of breeding, whereas many Angus Campbell fish appear to. The source of mortality in Lewis Slough is not known, but it is not inherent (as in Pacific salmon). Lewis Slough females held in captivity continued to grow and survive to breed in a second season. At the Angus Campbell site in the late summer the density of fry in the few remaining pools of water is extremely high. Under these circumstances competition probably is intense for resources such as food, space, and oxygen. At this time the mortality rates in fry could be high. Small fish also are susceptible to predation by invertebrates, such as dragonfly nymphs, and these are abundant at this site. Bet-hedging theory predicts that Angus Campbell fish should have a smaller reproductive effort, should delay maturity (and thus exhibit larger body sizes), and have a longer life-span than Lewis Slough fish. The predictions are supported by the data. Generally it appears that the life history traits of the two populations 85 have diverged more in accordance with bet-hedging theory than r & K selection theory. However, neither theory is mutually exclusive. Mortality schedules and demographic events could both influence life history responses. MODEL TESTING No specific mechanism is elucidated in life history theory to account for sexual size dimorphism. Life history theory and my model, however, are related through the concept of reproductive effort and maximizing the number of young produced in a reproductive season and an individual's lifetime. This is in accordance with Dobzhansky's (1956) suggestion that natural selection does not work on life-history traits in isolation, but rather on the combined traits of the whole individual. My model proposes that a positive relationship exists between clutch size, interclutch interval and body size. The quantitative aspects of these relationships are such that given an extended reproductive season the "fecundity advantage" of large females is an illusion. The greatest number of eggs (reproductive effort) in a season is achieved through the production of many small clutches rather than a few large clutches. Hence, small body size in females, and the resultant increase in clutch number give small females a "fecundity advantage". If, on the other hand, conditions limit the possible number of clutches in a season to one or two, then seasonal production of eggs will approximate instantaneous fecundity, and thus large females will be at an advantage. The Lewis Slough population is annual and has a relatively long reproductive season. Since most females in this population only breed for one season the individuals that produce the most young in that season (i.e. those 86 with an increased reproductive effort) should predominate. The mechanism that increases the seasonal production of young is an increase in the number of clutches. This in turn is achieved through small female body size. In contrast, large female body size probably is advantageous at the Angus Campbell site, since the reproductive season is short and many females survive to breed in a second season. To test this model it is necessary to establish that the appropriate relationships exist between body size and clutch number and clutch size, and that the quantitative variation in these traits is sufficient to overcome the fecundity advantage of large size. A 40 mm female ovulates about 75 eggs per clutch and produces about half as many eggs as a 50 mm fish, and only 38% as"*many eggs as a 60 mm female. To achieve equality over a season a 40 mm females must produce twice as many clutches as a 50 mm female and almost three times as many clutches as a 60 mm fish. In my data the total number of clutches does not decrease significantly with body size. The regression is significant, however, if data from both populations are combined, but this result is dubious because analysis of covariance indicates that the relationship between clutch number and body size is significantly different between the populations. The graph does indicate that large females tend to have only a few clutches, while smaller females can have few to many clutches. Mean interclutch interval appears to vary in a curvilinear fashion with body size and the logarithm of mean interclutch interval varies significantly with body size when data from both populations are combined (covariance analysis indicates that the slopes and intercepts do not vary between the two populations). An interclutch interval of approximately 13 days is indicated for a 40 mm female, 18 days for a 50 mm fish and 24 days for a 60 mm female. The mean 87 interclutch interval of the small fish is greater than that required for them to overcome the clutch size advantage of the larger females. Thus, my model is insufficient to provide a mechanism to explain the evolution of small female size in Lewis Slough females. However, given a breeding season of about 150 days (such as observed in Lewis Slough) a 40 mm female could produce up to 865 eggs. This is about 70% of the total number of eggs a 50 mm or 60 mm female could produce in the same time period. Contrasted with the proportion achieved at a single spawning, the fecundity disadvantage of small size is greatly reduced. If a straight line is drawn through the bulk of the points in Figure 14 a mean interclutch interval of about 10 days is indicated for a 40 mm fish. Under these circumstances such a female could achieve 90% of the eggs produced by a larger female. The result of this effect is observed in Figure 13 where total number of eggs produced over the breeding season is seen to be independent of body size. This suggests that there is no seasonal fecundity advantage associated with large body size in either of these populations of stickleback. The compensatory effect of increased clutch number is apparent at any time period long enough to permit multiple spawning in small females. Therefore, large body size is only advantageous where the probability of only a chance at single spawning is high, or, where fry survival is strongly dependant upon being hatched within a narrow time frame. Although the breeding season in Angus Campbell ditch is approximately 100 days long, several observations suggest that individual females may be limited to one or at the most two clutches. Trap samples from the field indicate a sex ratio biased towards females in the order of almost 2:1 (although laboratory bred fish show a 1:1 sex ratio). Thus, it is possible in Angus Campbell that males may be limiting resource. This effect is exacerbated by the long male reproductive cycle, particularly during the earlier part of the season when water temperatures are colder and egg development slower. Angus Campbell 88 fish begin breeding at water temperatures of about 10 degrees Celsius. This is lower than in Lewis Slough, and near the minimum temperature at which breeding occurs (Baggerman 1980). If there is severe competition for available resources in Angus Campbell during the late summer, and the acquisition of these resources is a positive function of body size, then older fry may have an advantage because of their greater size. At this site, early breeding is constrained by poor and unpredictable weather conditions, and if late breeding results in fry that are too small to successfully compete when conditions become severe in late summer, then females have a narrow window in which to optimize the probability of their offspring surviving. Given such a brief chance of successfully breeding, large size (and thus high instantaneous fecundity) would maximize the number of young produced during this short interval. Similar mechanisms may operate in a population of Mediterranean threespine stickleback studied by Crivelli and Britton (1987). This annual population also exhibits a pronounced female bias in size, and male growth is determinant. The environment is similar to Angus Campbell in that the ditches where breeding occurs dry up early in the summer and thus curtails the potential reproductive season. These authors estimate that females produce only one or two clutches. They interpreted this as an adaptation to the Mediterranean climate, but I propose that this pattern is a more general phenomenon associated with seasonally ephemeral habitats that place a premium on large clutch sizes. The evolution of small body size and intraspecific variation in the direction of sexual size dimorphism still remains a problem. My data suggest that there is no quantitative advantage to either large or small body size, except under the special conditions previously described. Therefore, neither my model nor the fecundity model is sufficient to explain variation in sexual size dimorphism in these multiple spawning fish. A general model is still required 89 and it may be fruitful to look at some of the sexual selection models that have been developed in the ornithology literature (for a review see Jehl and Murray 1986). Some changes are necessary to adapt these models to fish (and ectotherms in general). Specifically, the model must account for indeterminant growth, the increase in instantaneous fecundity with increasing body size, and multiple spawning within a season. The model proposed by Murray (1984) and Jehl and Murray (1986) seems promising. The mating system and the degree of dimorphism is predicted from the ratios of breeding males to non-breeding males, and the ratio of breeding males to breeding females. The actual direction of dimorphism is dependant upon the form of territorial display (i.e. areal versus ground). This model may be useful if the type of display and its consequences can be interpreted for the behaviour of fish. It is particularly applicable to intraspecific patterns because such factors as sex ratios and ratios of breeding adults to non-breeding adults can be easily influenced by environmental conditions. CORRELATES OF FEMALE REPRODUCTIVE SUCCESS Given that seasonal production of eggs is independent of body size, what are the correlates of this measure of reproductive success? Total seasonal production of eggs is correlated with clutch size and total number of clutches, but the correlation is strongest with total number of clutches. The number of clutches and length of a female's reproductive period is positively correlated in each population. The relationship is curvilinear and the curves resemble those resulting from hyperbolic saturation equations, with a maximum of 100 days at a total clutch number of 6. Any further increase in the number of clutches must correspond with a decrease in interclutch interval. Individuals that 90 produce larger numbers of clutches are able to do so because they can produce clutches at smaller intervals, and this is related to the size of the individual. Mean interclutch interval is positively correlated with clutch size; however the relationship is not significant. This suggests that the positive relationship between interclutch interval and body size arises from body size (or body mass) itself, and not as a function of increasing clutch size. In the absence of food effects, variation amongst individuals of approximately the same size may result from variation in metabolic activity and efficiency in oogenesis or hormone production. Total egg production co-varies with both number of clutches and clutch size, hence those individuals that produced the greatest number of eggs over the season tend to be mid-size females that were capable of producing a large number of clutches (quickly and for a long period of time) and had intermediate clutch sizes. This suggests that there is an optimum size for females in terms of potential egg production, and interestingly this size closely corresponds to normal size range of males. Why then have females diverged from what appears to be a body size that optimizes the probability of them producing large numbers of eggs? By negating the fecundity advantage model, my data challenges the traditional assumption that selection favors large females and provides an alternate way of thinking about the consequences of female body size. If fecundity considerations do not restrict body size, then body size is "freer" to change in response to other selective pressures. For example, Borland (1986) proposed that the males in a population of resident freshwater stickleback were constrained from a normal choice of larger, more fecund females by the presence of breeding marine anadromous females. Marine anadromous fish are larger than the freshwater morph, and if males exhibited normal behaviour some "mistakes" would be made and hybridization would occur. But males in this population 91 display a preference for females of small body size (or something correlated with small body size). This provides males with a mate recognition symbol and a cue to the females genotype. Under these circumstances, sexual selection could act to favor the expression of small female body size in this population. Alternately, in some populations the target of selection may not be seasonal fecundity or body size, but rather the number of clutches produced. The probability of an individual male successfully raising a clutch of eggs must vary depending on such unpredictable events as nest predation or random environmental fluctuations. If the probability of such events is high, a female would increase the likelihood of leaving surviving offspring by distributing her seasonal allotment of eggs among the nests of many males. Thus as the probability of egg and fry survival 'varies, s*6 should the degree of iteroparity within a season and hence body size. It might also be expected that those taxa (such as some cyprinids) that partially spawn a single clutch of eggs with many males may exhibit similar correlations. CONCLUDING REMARKS Miller (1979) suggested that the type of iteroparity exhibited by the sticklebacks probably developed under tropical to semi-tropical conditions where persistent but relatively low levels of production provided resources for both larval survival and repeated gonadal maturation. Once this property was developed in a phylogenetic line it might enable derivatives to colonize higher latitudes where environmental changes require the maximization of reproductive effort on a seasonal basis. The benefits of multiple spawning to a small teleost in terms of increased reproductive output are obvious. Less obvious is the flexibility this 92 reproductive pattern gives an organism to adjust to a variety of environments that may impose different selective pressures. The observations made in this study, therefore, contribute to the understanding of the diversity and success of the threespine stickleback throughout the North Temperate region. Stickleback are not unique among temperate freshwater fish in producing multiple clutches within a breeding season. Some cyprinids such as shiners and minnows, as well as some darters, show this pattern of reproduction (Gale 1978, 1983, 1985) and some suggestion of a similar relationship between body size, total egg production, and clutch frequency (Gale 1983). In these groups, however, these relationships have yet to be explicitly explored. Most of these fish were derived from warm-water ancestors (Miller 1979) but now exist over a wide geographic range and in a variety of environmental conditions. Cyprinids, and other species of stickleback, would be particularly useful in deterrnining the generality, or pervasiveness, of the patterns of covariation between reproductive parameters and body size that were found in this study. Inter- and intra-specific studies with these taxa, as well as further studies on other threespine stickleback populations, would also be useful in determining the conditions under which sexual dimorphism in size varies. Such data would aid in the development of a general model to describe the evolution of this phenomenon in fish. 93 L I T E R A T U R E C I T E D Aneer G. 1973. Biometrical characteristics of the three-spined stickleback (Gasterosteus  aculeatus L.) from the northern Baltic proper. Zool. Scr. 2:157-167. Baggerman B. 1980. Photoperiodic and endogenous control of the annual reproductive cycle in teleost fishes, in: Environmental Physiology of Fishes Ali M. A. (ed.) Plenum Publishing Corp., New York. Baggerman B. 1985. The role of biological rhythms in the photoperiodic regulation of seasonal breeding in the stickleback Gasterosteus aculeatus L. Neth. J. Zool. 35:1431. Berry J. F. and R. Shine 1980. Sexual size dimorphism and sexual selection in turtles (order Chelonia). Oecologia 44:185-191. Bell G. 1980. The costs of reproduction and their consequences. Amer. Nat. 116:45-76. Bell M. A. 1984. Evolutionary phenetics'and genetics, the three-spined stickleback Gasterosteus aculeatus amd related species, in: Evolutionary Genetics of Fishes. Turner B. J. (ed) Plenum Press, New York."* Bentzen P. and J. D. McPhail 1984. Ecology and evolution of sympatric stickleback (Gasterosteus'): specialization for alternate trophic niches in the Enos Lake species pair. Can. J. Zool. 62:2280-2286. Blueweiss L., Fox H., Kudzma V., Nakashima D., Peters R., and S. Sams 1978. Relationships between body size and some life-history parameters. Oecologia 37:257-272. Borland M. 1986. Size-assortative mating in three-spine sticklebacks from two sites on the Salmon River, British Columbia. MSc. Thesis University of British Columbia. Boyce M. S. 1984. Restitution of r- and K- selection as a model of density-dependant natural selection. Ann. Rev. Ecol. Syst. 15:427-447. Burt A., Kramer D. L., Nakatsura K., and C. Spry 1988. The tempo of reproduction in Hyphessobrycon pulchripinnis (Characidae), with a discussion on the biology of "multiple spawning" in fishes. Env. Biol. Fish. 22:15-27. Cabana G., Frewin A., Peters R. H. and L. Randall 1982. The effect of sexual size dimorphism on variations in reproductive effort of birds and mammals. Amer. Nat. 120:17-25. Calow P. 1978. Life Cycles Wiley, New York. Calow P. 1982. Homeostasis and fitness. Amer. Nat. 120:416-419. Chambers J. M., Cleveland W. S., Kleiner B. and P. A. Tukey 1983. Graphical Methods  for Data Analysis Wadsworth. Cleveland W. S. 1979. Robust locally weighted regression and smoothing scatter plots. JASA 74:829-836. 94 Clutton-Brock T. H. 1983. Selection in relation to sex. in: Evolution from Molecules to  Men Bendall D. S. (ed) Cambridge University Press,Cambridge. Crivelli A. J. and R. H. Britton 1987. Life history adaptations of Gasterosteus aculeatus in a Meditteranean wetland. Env.Biol.Fishes 18:109-125. Darwin C. R. 1874. The Descent of Man and Selection in Relation to Sex Second edition. Appleton, New York. Dingle H., Blau W. S., Braun C. K. and J. P. Hegmann 1982. Population crosses and the genetic structure of milkweed bug life-histories, in: The Evolution and Genetics of  Life Histories Dingle H. and J. P. Hegmann (eds), Springer-Verlag, New York. Dobzhansky T. 1956. What is an adaptive trait? Amer. Nat. 90:337-347. Gale W. F. 1983. Fecundity and spawning frequency of caged bluntnosed minnows -fractional spawners. Trans. Amer. Fish. Soc. 112:398-402. Gale W. F. and G. L. Buynak 1982. Fecundity and spawning frequency of the fathead minnow - a fractional spawner. Trans. Amer. Fish. Soc. 111:35-40. Gale W. F. and W. G. Deutsh 1985. Fecundity and spawning frequency of captive tesselated darters -fractional spawners. Trans. Amer. Fish. Soc. 114:220-229. Gale W. F. and C. A. Gale 1977. Spawning habits of spotfin shiner (Notropis spilopterus) - a fractional crevice spawner. Trans. Amer. Fish. Soc. 106:170-177. Giles N. 1987. Population biology of three-spined sticklebacks, Gasterosteus aculeatus. in Scotland. J. Zool. Lond. 212:255-265. Gibbons J. W., Greene J. L. and K. K. Patterson 1982. Variation in reproductive characteristics of aquatic turtles. Copeia 4:776-784. Green R. F. 1980. A note on K-selection. Amer. Nat. 116:291-296. Greenbank J. and P. Nelson 1959. Life history of the three-spined stickleback Gasterosteus aculeatus Linnaeus in Karluk Lake and Bare Lake, Kodiak Island, Alaska. U.S. Fish. Wildlife Serv. Bull. 153:537-559. Hagen D. W. 1967. Isolating mechanisms in three-spine sticklebacks (Gasterosteus aculeatus). J. Fish. Res. Bd. Can. 24:1637-1692. Hagen D. W. and L. G. Gilbertson 1972. Geographic variation and environmental selection in Gasterosteus aculeatus L. in the Pacific Northwest America. Evol. 26:32-51. Hagen D. W. and L. G. Gilbertson 1973. Selective predation and the intensity of selection acting upon the lateral plates of three-spined sticklebacks. Heredity 30:273-287. Hays H. 1972. Polyandry in the spotted sandpiper. Living Bird 11:43-57. Hegmann J. P. and H. Dingle 1982. Phenotypic and genetic covariance structure in Milkweed Bug life history traits, in: The Evolution and Genetics of Life Histories Dingle H. & J. P. Hegmann (eds), Springer-Verlag, New York. 95 Howe M. A. 1982. Social organization in a nestling population of Eastern Willets (Cataptrophorus semipalmatus). Auk 99:88-102. Hubbs C. 1985. Darter reproductive seasons. Copeia 1985:56-68. Jehl J. R. Jr. and B. J. Murray Jr. 1986. The evolution of normal and reverse sexual size dimorphism in shorebirds and other birds, in: Current Ornithology Vol. 3 Johnston R. F. (ed) Plenum Press, New York. Kynard B. E. 1972. Male breeding behaviour and lateral plate phenotypes in the threespine stickleback (Gasterosteus aculeatus L.) Ph.D. Thesis University of Washington. Lavin P. A. and J. D. McPhail 1985. The evolution of freshwater diversity in the three -spine stickleback (Gasterosteus aculeatus): site-specific differentiation of trophic morphology. Can. J. Zool. 63:2632-2638. MacArthur R. H. 1972. Some general theorems of natural selection. Proc. Natl. Acad. Sci.  U. S. A. 48:1893-1897. McPhail J. D. 1969. Predation and evolution in a stickleback (Gasterosteus aculeatus). L Fish. Res. Bd. Can. 26:3183-3208. McPhail J. D. 1977. Inherited interpopulation differences in size at first reproduction in the three-spine stickleback Gasterosteus aculeatus. Heredity 38:53-60. Mann R. H., Mills C. A. and D. T. Crisp 1984. Geographic variation in the life-history tactics of some species of freshwater fish, in: Fish Reproduction Potts G. W. and R. J. Wootton (eds) Academic Press Inc., London. Miller P. J. 1979. Adaptiveness and implications of small size in teleosts. Symp. zool. Soc.  London 44:263-306. Munzing J. 1959. Polymorphe populationen von Gasterosteus aculeatus L. (Pisces, Gasterosteidae) in sekundaren intergradioszonen der Deutschen buch und benachberter gebiete. Faun. Okol. Milt. 4:69-84. Murray B. G. Jr. 1984. A demographic theory on the evolution of mating systems as exemplified by birds, in: Evolutionary Biology Vol. 18 Hecht B., Wallace B. and G. Prance (eds) Plenum Press, New York. Peters R. H. 1983. The Ecological Implications of Body Size Cambridge University Press, Cambridge. Pianka E. R. and W. S. Parker 1975. Age specific reproductive tactics. Amer. Nat. 109:453-464. Presley P. H. 1976. Parental investment in threespine stickleback Gasterosteus aculeatus. MSc. Thesis University of British Columbia. Price T. D. 1984. The evolution of sexual size dimorphism in Darwin's Finches. Amer. Nat. 123:500-518. Reznick D. and J. A. Endler 1982. The impact of predation on life history evolution in Trinidadian guppies (Poecilia reticulata). Evol. 36:160-177. 96 Rising J. D. 1987. Geographic variation of sexual dimorphism in size of savannah sparrows (Passerculus sandwichensis): a test of hypothesis. Evol. 41:514-524. Rothstein S. I. 1973. The niche-variation model - is it valid? Amer. Nat. 107:598-620. Rowland W. J. 1983. Interspecific aggression and dominance in Gasterosteus. Env. Biol.  Fish. 8:269-277. Sargent R. C. 1982. Territory quality, male quality, courtship intrusions, and female nest choice in the threespine stickleback, Gasterosteus aculeatus. Anim. Behav. 30:364-374. Searcy W. A. 1979. Sexual selection and body size in male red- winged blackbirds. Evol 24:311-319. Selander R. K. 1966. Sexual dimorphism and differential niche utilization in birds. Condor 68:113-151. Semlitsch R. D. and J. W. Gibbons 1982. Body size dimorphism and sexual selection in two species of water snakes. Copeia 1982:974-976. Shine R. 1979. Sexual selection and sexual dimorphism in the Amphibia. Copeia 1979:296-306. Shine R. 1988. The evolution of large body size in females: a critique of Darwin's "fecundity advantage" model. Amer. Nat 131:124-131. Singer M. C. 1982. Selection for small size in male butterflies. Amer. Nat. 119:440-443. Sokal R. R. and F. J. Rohlf 1981. Biometry W. H. Freeman & Co., San Francisco. Spain J. D. 1982 BASIC Microcomputer Models in Biology Addison-Wesley Publishing Company, London. Steams S. C. 1975. A comparison of the evolution and expression of life-history traits in stable and fluctuating environments of Gambusia affinis in Hawaii. Ph.D. Thesis. University of British Columbia. Steams S. C. 1976. Life-history tactics: a review of the ideas. Quart. Rev. Biol. 51:3-47. Steams S. C. 1980. A new view of life-history evolution. Oikos 35:266-281. Steams S. C. 1983. A natural experiment in life-history evolution: field data on the introduction of mosquitofish (Gambusia affinis) to Hawaii. Evol. 37:601-617. Townsend C. R. and P. Calow 1981. Physiological Ecology: an Evolutionary Approach to  Resource Use Blackwell Scientific Publications. Trivers R. L. 1972. Parental investment and sexual selection, in: Sexual Selection and the  Descent on Man. 1871 - 1971 Campbell B. (ed). Aldine, Chicago. Ursin E. 1979. Principles of growth in fishes. Symp. zool. London. 44:63-87. van den Assem J. 1967. Territory in the three-spined stickleback Gasterosteus aculeatus L.: an experimental study in intra-specific competition. Behaviour suppl. 16. 97 Veulle M. 1980. Sexual behaviour and evolution of sexual dimorphism in body size in Jaero (Isopoda: Asellota). Biol. J. Linn. Soc. 13:89-100. Vitt L. J. and J. D. Congdon 1978. Body shape, reproductive effort and relative clutch mass in lizards: resolution of a paradox. Amer. Nat. 112:595-608. Wallace R. A. 1978. Oocyte growth: nonmammalian vertebrates, in: The Vertebrate Ovary: Comparative Biology and Evolution. Jones R. E. (ed) Plenum Publishing Corp., New York. Wallace R. A. and K. Selman 1979. Physiological aspects of oogenesis in two species of sticklebacks, Gasterosteus aculeatus (L.) and Apeltes quadricus (Mitchell). J. Fish  Biol. 14:551-54. Ware D. M. 1984. Fitness of different reproductive strategies in teleost fishes, in: Fish  Reproduction Potts G. W. and R. J. Wootton (eds) Academic Press Inc., London. Wasserman W. J. and L. D. Smith 1978. Oocyte maturation: nonmammalian vertebrates, in: The Vertebrate Ovary: Comparative Biology and Evolution Jones R. E. (ed.) Plenum Publishing Corp., New York. Weatherhead P. J. 1980 Sexual dimorphism in two savannah sparrow populations. Can. J.  Zool. 58:412-415. Williams G. C. 1966. Adaptation and Natural Selection Princeton University Press, Princeton. Woolbright L. L. 1983. Sexual selection and size dimorphism in anuran amphibia. Amer.  Nat. 121:110-119. Wootton R. J. 1973. Fecundity of the three-spine stickleback Gasterosteus aculeatus L. J.Fish Biol. 5:683-688. Wootton R. J. 1976. The Biology of Sticklebacks Academic Press, London. Wootton R. J. 1977. Effect of food limitation during the breeding season on the size, body components and egg production of female sticklebacks (Gasterosteus aculeatus). L Anim. Ecol. 46:823-834. Wootton R. J. 1984. The Functional Biology of Sticklebacks Croom Helm, London. Wootton R. J. and G. W. Evans 1976. Cost of egg production in the three-spined stickleback (Gasterosteus aculeatus). J. Fish Biol. 8:385-395. Zar J. H. 1984. Biostatistical Analysis second edition. Prentice-Hall Inc., New Jersey. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0097673/manifest

Comment

Related Items