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  degree freely  this  at the  thesis  in  partial  fulfilment  of  University of  British  Columbia,  I agree  available for reference  copying  of  department publication  this or of  and study.  thesis for scholarly by  this  his  or  her  representatives.  ZoOLO^M,  The University of British Columbia Vancouver, Canada  DE-6 (2/88)  >SfpV. a«|\<Kg  that  may be It  thesis for financial gain shall not  Department of  requirements  I further agree  purposes  permission.  Date  the  that  the  advanced  Library shall make it  by the  understood be  an  permission for extensive  granted  is  for  allowed  that without  head  of  my  copying  or  my written  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 atfirstreproduction, 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 predictionsfrombet-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  T A B L E OF CONTENTS LIST OF TABLES LIST OF FIGURES AKNOWLEDGEMENTS INTRODUCTION  1  SITE DESCRIPTION Physical Chemical Biotic  9 9 15 16  MATERIALS AND METHODS Reproductive Characteristics Growth Under Laboratory Conditions Growth and Age Structure in the Field  23 23 26 30  RESULTS Reproductive Characteristics Growth Under Laboratory Conditions Growth Age Structure in the Field  32 32 56 69  DISCUSSION Age Structure and Growth Patterns Population Specific Reproductive Characteristics and LifeHistory Evolution Model Testing Correlates of Female Reproductive Success Concluding Remarks  77 77  LITERATURE CITED  93  80 85 89 91  V  LIST OF TABLES  Table 1 2 3 4  5  6  7 8 9 10 11 12  13 14 15  Title  Page  Oxygen concentration in percent saturation for the two study sites during 1985.  15  Estimated percent of stomach fullness in sticklebacksfromthe Angus Campbell and Lewis Slough sites on two dates per site.  16  List of fish species trapped at the two study sites over a three year period.  17  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  Mean egg diameter in millimeter (X) and standard deviation (S.D.) in clutches produced by a single female, between females, and between populations.  25  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  Regression of reproductive characters on body size for each population.  34  Probability values for a significant difference in reproductive characters between females from the two sites.  46  Matrix of correlation coefficients for reproductive characters in each population.  50  Summary of Analysis of Covariance on linearized growth curves of laboratory bred and raised fish.  56  Slopes and standard errors for linearized growth curves of replicates within a population.  58  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  Summary of 2-way ANOVA for the effects of sex and population on body size in laboratory bred and raised fish.  59  Summary of 2-way ANOVA for the effects of sex and population on body size in wild-caught and laboratory raised fry.  62  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. 1 2 3 4 5  6 7  Title  Page  Sexual size dimorphism in breeding Lewis Slough and Angus Campbell fish over a three year period.  8  Lower and central Fraser Valley of British Columbia showing location of two study sites.  11  Fluctuation in water depth at the two study sites over a three year period.  13  Seasonal water temperature profiles at the two study sites over a three year period.  14  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  Lower and central Fraser Valley of British Columbia showing location of survey sites.  21  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. Date of first clutch (as numbered from the first of January) plotted against body size.  36 37  Date of last clutch (as numbered from the first of January) plotted against body size.  38  Total number of clutches produced in the reproductive season plotted against body size.  39  Total number of eggs produced in a reproductive season plotted against body size.  40  Mean interclutch interval of an individual plotted against body size.  41  10 11 12 13 14 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 19 20  Mean interclutch interval of an individual plotted against total number of clutches.  53  Diagrammatic summary of significant correlations between reproductive characters in females from the Angus Campbell site.  54  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. 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  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  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  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  23  24  25  26  61  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  D a r w i n (1874, pg.332-375) proposed that the female bias i n body size observed i n many animals evolved f r o m the selective advantage associated w i t h the increased fecundity of large females. presumably  Large females tend to have larger clutches or broods than s m a l l females,  because  there is more  space inside their body  (Shine  1988).  Darwin's  "fecundity advantage" m o d e l 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; V e u l l e 1980; Semlitsch & 1982;  Woolbright 1983).  Gibbons  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 i n the bias of sexual size dimorphism has been attributed to the type of mating system and selection that result f r o m intrasexual competition for mates (e.g. Trivers 1972; Singer 1983; Woolbright 1983; R i s i n g 1987).  1982;  Clutton-Brock  In these models the presence of s m a l l females is  predicted under conditions w h i c h select for large males.  N o n e of the models actually  propose a mechanism that selects for small female body size.  The application of these  sexual selection models to ectotherms is limited because little provision for ^determinant growth and the associated changes i n number of eggs per clutch w i t h 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). addition, these models have been developed f r o m comparisons across species.  In  Therefore,  unless the populations exhibit radical differences i n mating systems and behaviour, their usefulness to explain population divergence in the bias of sexual size dimorphism is limited (but see Weatherhead 1980; Price 1984; and R i s i n g  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.  body size (mm) CO o  o  body size (mm) cn OS o o o —r— -r—  body size (mm) CO o  > o  1-  > o  cooo cn  CO  00 0>  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 5/85  1  1  1  6  7  8  1  9 4/86 5  1  6  1  7  1  1  8  1  9  t  10 11  '  •  <  12 1/87 2  D A T E (month/year)  '  i  3 *6  i  7  i  8  9  i  i  10  11  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.  CHEMICAL  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 Angus Campbell Lewis Slough  5/22 80.4 50.4  6/2 50.5 50.9  6/14 60.1 61.3  7/4 50.3 62.3  7/24 30.9 63.0  8/13 5.1 60.5  9/7 6.4 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  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. % of fullness 0 - 25 25 - 50 51 - 75 76 - 100 N  Angus Campbell april 25 may 22 .45 .56 .27 .21 .22 .21 .06 .02 49  43  Lewis may 2 .48 .26 .21 .05 42  Slough July 4 .65 .14 .18 .03 34  17  Table 3.  L i s t of fish species trapped at the two study sites over a three year period. X indicates presence, - indicates absence. A C denotes the Angus C a m p b e l l site, and L S the L e w i s Slough site.  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)  AC  LS  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.  E x c l u d i n g the resident freshwater f o r m of Gasterosteus aculeatus. A n g u s C a m p b e l l  supports three other fish species, however, these species only appeared during early spring or late autumn and never occurred i n large numbers. D u r i n g the summer sticklebacks appear to be the only resident fish. In contrast L e w i s Slough supports eight species of fish, including both the resident freshwater and the anadromous  marine forms o f Gasterosteus  aculeatus.  Like  C a m p b e l l resident freshwater sticklebacks were the predominant f o r m i n L e w i s  Angus Slough,  both i n numbers and persistence i n the area. The only fish that exceeded freshwater sticklebacks in numbers was the anadromous marine f o r m i n L e w i s Slough. early M a y .  This form comes into the area to breed during A p r i l and  Figure 5(c) shows seasonal variation in the proportion of the anadromous f o r m  i n the catch at the L e w i s Slough site over the three year study period.  There is some  indication that their presence is cyclic, i n 1985 numbers were l o w , i n 1986 they were high,  18  .8 ,  2/19  4/10  5/30  7/20  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. M e a n size (mm), standard deviation, and sample size of fish obtained i n survey collections f r o m the vicinity of the Angus C a m p b e l l and L e w i s Slough study sites. Site numbers correspond with those i n Fig.6. Values for a site may represent data collected over more than one sampling period. Site 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19  n 35 21 15 23 32 28 15 23 33 24 23 41 32 10 52 43 30 25 21  female mean S.D. 45.7 7.5 42.1 5.2 44.2 7.4 39.5 6.4 41.1 7.3 43.2 6.8 34.3 5.2 38.7 5.9 47.3 4.9 43.3 6.5 44.7 6.9 35.5 5.5 36.1 5.9 40.2 3.1 36.8 7.5 39.4 7.5 41.2 8.1 37.2 5.8 35.9 5.5  n 22 15 10 17 21 17 10 12 25 18 20 22 17 8 "33 27 21 16 12  male mean S,D, 3.2 43.1 44.9 4.6 45.7 4.7 42.2 5.4 43.3 3.1 42.8 5.4 3.4 35.0 42.4 5.1 44.2 4.3 41.3 3.7 43.6 3.1 38.7 3.7 42.1 3.8 41.2 3.8 4.4 43.7 44.2 4.9 40.3 3.8 3.4 45.6 4.1 44.3  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. maximum-minimum thermometer was suspended in the tank.  To monitor temperature a  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 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.  Table 5.  ind# bodv sizefmm) pop. 1 35.3 LS -  r  2 3 4  35.8 36.8 37.4  LS LS LS  5  40.4  LS  6 7 8 9 10 11 12 13 14 15  40.5 42.0 43.0 45.9 47.0 47.4 48.5 52.0 53.9 56.8  LS AC AC AC AC LS ' AC AC AC AC  clutch# 1 2 3 4 5 1 1 1 2 3 1 2 3 4 "'5 1 1 1 1 1 1 1 1 1 1  X 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.4 1.4 1.5 1.5 1.4 1.5 1.5 1.5 1.6 1.5 1.5 1.4 1.9 1.7 1.5 1.5 1.6  S.D. 0.07 0.06 0.11 0.12 0.05 0.08 0.09 0.07 0.08 0.08 0.08 0.01 0.09 0.03 0.13 0.02 0.04 0.07 0.05 0.18 0.10 0.12 0.05 0.05 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 withfine-tippedforceps.  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 fungusridden 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 retrapped 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  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 femalesfromeach 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 countedfromJanuary 1. Table  CHARACTERISTIC  n  body size (mm) 29 mean clutch size 29 reproductive period (d) 29 date of first clutch 28 date of last clutch 28 total number of clutches 29 total number of eggs 29 mean interclutch interval (d) 21 summer growth rate (mm/day) 29 CHARACTERISTIC  n  body size (mm) 29 mean clutch size 29 reproductive period (d) 29 date of first clutch 29 date of last clutch 29 total number of clutches 29 total number of eggs 29 mean interclutch interval (d) 28 • summer growth rate (mm/day) 29  Angus Campbell mean S.D.  median  49.2 146.28*' 46.21 107.71 152.89 3.24 443.14 23.81 .09  7.5 64.81 43.95 24.31 41.24 2.29 285.04 17.24 .02  47.7 143.00 45.00 105.00 154.00 3.00 420.00 18.00 .03  mean  Lewis Slough S.D.  median  46.0 127.52 67.31 136.41 202.72 5.17 651.90 18.93 .03  5.0 41.88 26.38 19.23 16.16 2.20 306.72 11.45 .02  46.8 127.00 72.00 135.00 203.00 5.00 648.00 16.00 .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.  df  characteristic  mean clutch size 27 reproductive period* 19 date of first clutch' 26 date of last clutch 26 total number of clutches 27 total number of eggs 27 mean interclutch interval(d) 16 summer growth rate(rnm/day)' 27 ,b  c,d  Angus Campbell a b  r  2  p  -171.34 5.22 11.85 198.78 6.98 -138.45 1.81** 0.16  6.45 0.05 -0.31 -0.85 -0.08 11.81 0.02 -0.02  .56 .02 .04 .04 .06 .10 .24 .50  .001 .5 .5 .5 .2 .1 .05 .001  a  b  r  p  -177.95 13.10 10.50 238.98 10.41 -171.05 1.89 0.08  6.64 -0.11 0.02 -0.79 -0.11 17.88 0.02 -0.01  .63 .07 .03 .06 .07 .08 .09 .02  Lewis Slough characteristic  df  mean clutch size 27 reproductive period* 27 date of first clutch* 26 date of last clutch 27 total number of clutches 27 total number of eggs 27 mean interclutch interval(d) 25 summer growth rate(mm/day) •27 b  c,d  a-square root transformation b-zero values removed  c-outlyers removed d-log transformation  2  .001 .2 .5 .2 .2 .2 .2 .5  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 value for Angus Campbell is .56, and for Lewis Slough it is .63. Dots - Angus Campbell, triangles - Lewis Slough. 2  36  40  50  60  70  body size (mm)  ?• t " ^ °l ? 61  8  a  ^dividuals reproductive period ( days between first and last clutch) plotted against body size. Dots - Angus Campbell, triangles -Lewis Slough. F  i  i g  . uN U  c  37  o LO CvJ  O  o  CNJ  sz  o  o to  h  o CO TJ  o o  o LO  30  40  50  60  body size (mm)  Figure 10. Date of first clutch (as numbered from the first of January) plotted body size. Dots - Angus Campbell, triangles - Lewis Slough.  70  38  O  in  CM  o o  CM  o o  O CO LO GO  o  CD •4—*  "D  O O  O LO  30  40  50  60  70  body size (mm)  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 •  cluitches  00  -  A  A •  A  A •  A CO  —  o  A  A  A  •  A  A AAVv. A  mb  CD  A  •  •A ^ • A A  c  • •  •  •4—»  •  o C\J  • •A A A  A  •  •  o 30  A» •  Ai • •  • •  •  A  •  •  1  ' 1  1  1  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  o o  —•  o o  A  cvj  A A  o o o  A  A A  A  o o  CO  E r3 rz  75  o  A  A  CD JQ  •  o o  o o  A  A  • ft  A  • •A  A •  • -  A  A  •  o o  •  A  *A  CD  *  .A  A  A LA  A  •  A  •  . A .  •  •  A •  A  •  •  AT  1  30  40  •  1  1  l  50  60  70  body size  Figure 13. size.  (mm)  Total number of eggs produced in a reproductive Dots - Angus Campbell, triangles - Lewis Slough.  season plotted against body  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  10  A  rt co  •  •  AA  • JC o *-» 3  o  i—  Q)  o  •  A A A  —  CO  •  c c rt  • • LO  c\i  >*^*^A  A  A A  o  A A • A  •  AA A  •  • A  CM  • •  ^ r * ^ A  E co o  «A  0.02^t^'  A  '  A •  r  i  1  1  1  30  40  50  60  70  body size (mm)  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, r = .24). Regression for combined datafromboth populations is significant at p = .01 (dotted line, r = .16). Slope is indicated on the line. Dots - Angus Campbell, triangles - Lewis Slough. 2  2  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 = .50. Slope is indicated on line. Fig. 16(b) - Lewis Slough (slope is not different from zero). 2  b) L E W I S  A A  A  A A  SLOUGH  A  A A A  40  50 b o d y s i z e (mm)  i  30  40  50 b o d y s i z e (mm)  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, r = .16). Interclutch interval increases with increasing body size: a 2  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. Wootton (1977).  The direction of this relationship is opposite to the findings of  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 nonsignificant regression slope was found, a comparison of means. All comparisons are tested with a Student's t-test.  Slope  estimates of Y at X = 40 X = 50  body size(covariate) clutch size mean interclutch interval(d) , total number of clutches total number of eggs length of reproductive period(d)* date of first clutch* date of last clutch summer growth rate(mm/d)'  >.5 >.5 .005 <001 <.001 <.001 >.5 <.001  >.5 >.5 .05 .1 .2 .05 <.001 <.001  a-square root transformation b-zero values removed  c-outlyers rembVed d-log transformation  c d  ,b  rearing fish on different food regimes.  >.5 >.5 .05 .005 >.5 <.001 <001 <.001  means .06 -  -  .002 .01 >.5 <001 <.001 -  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 A  1)  Size  N  2)  Date o f F i r s t  3)  Date o f L a s t  G  (mm)  Clutch  3  4) L e n g t h  c  5)  Total  6)  ,,,c,d Mean I n t e r c l u t c h I n t e r v a l IQJ  o f Reproductive Number o f  7) T o t a l 8)  Clutch  L  9)  Summer G r o w t h  I S  Date o f L a s t  4)  Length  6)  f~\  u U G H  7)  7  R a t e (mm/d)  2  3 .107  Clutch*  -  -.244 .246  Clutch  o f Reproductive Number o f  Period  Total  (d) '^ a  Clutches  Size  a -square r o o t b -zero values  a  transformation removed  ***  9 *** -.641  -.187  -.055  .115  -.253  -  .385  -.249  4  5  -.338 -.258 *** *• * * * -.685 -.823 !*** .435 .686 *** _ .714  6  7  .292 ** .208 -.544  .281 .047 ** -.528  -  .215  -.287 -.028  -.281  8  9 *** -.184 .793  .300  - .020  .081  .418 - .034 ** .560 - .072 *** .763 - .088  -.200  .244  .153  .424  -  Number o f E g g s  Summer G r o w t h R a t e (mm/d)  8  .116 - .248 .310 .490* .745 *** *** -.643 - .538 -.156 -.609 - .090 *** *** ** .748 .675 .335 .556 - .207 *** *** .701 .376 .792 .080 *** -.325 .711 - .184  a  c d Mean I n t e r c l u t c h I n t e r v a l (d) '  8) C l u t c h 9)  6  -  Clutches  -  3)  L  -.131  Period ( d ) ' ^  (mm)  Date Qf F i r s t  5) T o t a l  -  5  Size  2)  s  -.206  a  1 1) S i z e  -.203  4  Number o f E g g s  L  L E W  3  Clutch  U S  A M P B  2  ** .540  -.090 -.257  -.301  -  .056 —  c - o u t l y e r s removed d-log transformation  ©  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. Initiallyreproductiveperiod 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.  52  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  log mean Interclutch Interval total number of eggs  ZT  clutch size  I I  body size  sqrt  summer growth rate  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  population tank(population) age * replicate residual  1 9 10 737  0.197 3.199 2.906 75.906  1.92 3.45 2.82  p-level .167 .0004 .002  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  0  1  20  1  40  1  1  1  i  i  60  80  100  120  140  AGE  I  160  (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  slope S.E.  (no-)  1  2  3  4  5  .017 .761  .015 .758  .019 .910  .017 .868  .018 .845  79  83  56  59  58  Lewis Slough 6  1  2  3  4  5  .018 .813  .016 .733  .017 .800  .016 .868  .017 .769  .015 .746  63  83  67  66  71  74  3  n  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.  Angus Campbell Lewis Slough Angus Campbell Lewis Slough  female female male male  mean  S.E.  n  35.4 34.7 36.0 35.1  .51 .64 .37 .57  24 21 31 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  model sex population sex * population error  3 1 1 1 98  0.230 0.060 0.156 0.003 6.525  F-ratio p-level 1.15 0.90 2.34 0.04  .332 .345 .129 .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. A C - 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  model sex population sex * population error  3 1 1 1 69  3.482 0.710 1.174 2.058 6.742  F-ratio p-level 11.88 7.26 12.01 21.01  .0001 .0088 .0009 .0001  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  AC LS AC LS  female female male male  _n  mean S.E. ACfemale LSfemale  ACmale  LSmale  29 12 14 18  43.0 36.8 41.5 42.4  .1543 .0002 -  .5273 .0001 .4379  .94 .99 .99 .52  -  .0001 -  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 lifehistory 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 12:15 3/22  265  275  13:00 14:00 4/3  4/20  285  295  14:45  15:30  5/4  5/20  305  3 1 5 TIME (days)  16:00»->8/4*%-*-  P H O T O P E R I O D (hours of illumlnatlon/24) C O R R E S P O N D E N T 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 12:15 3/22  265  275  13:00 14:00 4/3  4/20  285  295  305  315  14:45  15:30  16:00~»-*»  5/4  5/20  6/4»-**  TIME ( d a y s ) P H O T O P E R I O D (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.  mature  >37 mm <37 mm  immature  >37 mm <37 mm  TOTAL  Angus Campbell # % 42 50.0 0 0  Lewis Slough # % 21 53.8 9 23.1  33 9  39.3 10.7  6 3  15.4 7.7  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 C a m p b e l l males, females, and unknowns as proportion of catch obtained at sampling dates spanning f r o m M a y 1986 to M a y 1987. If fish were sexed i n the field, and an individual's sex could not be confidently determined, then that individual was categorized as an unknown.  71 MAY 1966  NOV. 1986  FEB. 1987  J C  o  CO  o c o o a o  10  20  30 40  50  60 70  1  0  20  .25  30 40  50  60 70 10 20  body size (mm)  b  .20 .15  n=36  .10 .05 0  10 20  30 40 50  60  70  body size (mm) APRIL 1987  MAY 1987  .25 .20 n»36  .15  n»75  .10  JC  o  .05 0 10  20  30 40  50  60 70 10 20  30 40  50  30 40  50  60  70  as o .50 .40  c o o a o  .30  na12  rl«29  .20 .10 0  10  20  30 40  50  .25  60 70 10 20 .30i  70  60  70  .25  .20 .15  60  n-109  na 70  .10 .05 0 10  20  30 40  50  60 70 10  body size (mm)  20  30 40 50  body  8 l z e  (  m m  )  30 40 o  d  y  8  l  z  e  50 (  m  60 m  )  70  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  date  total  4/8 5/25  232 166  fish (Fig. 26).  first spine >40 mm female male 20 6  5 2  .  second spine ? 1 2  <40 mm female male 4 -  2 2  ? 9 1  >40 mm female male 10 7  -  ? 1 4  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. occurred in April.  The next catch of significant numbers  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 .30 .25 .20 .15 .10 .05 0  JULY 1986  OCT. 1986  10 20 30 40 50 60 70  10 20 30 40 50 60 70  sz  .30 o .25 CO .20 o .15 .10 c .05 o 0  o a o i_ a  NOV. 1986  .35 .30 .25 .20 ns51 .15 .10 .05 .25 .20 .15 .10  n»68  .osio  20 30 40  5b 60 7b  10 20 30 40 50 60 70  .20  .30 .25 .20  body  size (mm)  .15  .15-1 n=48 .10-1 n=82 .10 .05 .05^ 0 10 20 30 40 50 60 70 10 20 30 40 50 60 70 body  SZ  .30 .25 .20 .15 n-41 .10 .05 • 0  size (mm)  APRIL 1987  size (mm)  MAY 1987  JUNE 1987  n = 36  10 20 30 40 50 60 70 I c T l ^ C - l o ^ ^ ^ i ^ .40, .20 .30  o  CO  o  .20 n = 24  c o o a o  body  .10  .10  Vlrf^^haTTb  na15  .05  0 10 20 30 40^ ~io~7o 10 20 ' b o d y size (mm) .50 0  .40  3  .30  .40  .20 n=36  .30 .20  .10 0  0  4T^cr7o  n=*64  .10 10 20 30 40 50 60 70 body  size (mm)  0', 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. November.  The size class representing adults in October completely disappears in 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). been presented because they are incomplete.  These data have not  77  DISCUSSION  A G E - S T R U C T U R E AND G R O W T H 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  determinant versus ^determinant growth, or growth asymptotes). could  conceivably  be  subject  to  selection.  Interpreting  the  forms  (e.g.  All of these 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. for  Many fish in this population apparently survive  more than one breeding season and females appear to continue growing  throughout their  life.  In contrast,  males  either  exhibit  or grow toward a much lower asymptote than females. continue  long  enough  to  distinguish  deterministic growth  My experiments did not  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). determined growing  that  after  males reaching  in  a  population  sexual  maturity.  of  Mediterranean stickleback The  females  in  this  They stopped  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. females indicates experience  this  Lack of such a growth spurt in laboratory raised  that this period  of  spurt is rapid  environmentally  growth,  reached the end point of their growth.  perhaps  induced. because  Males do not  they  have already  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. predicted if sex My  growth  specific  Here monomorphism in size is  growth patterns were the only mechanisms operating.  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  males and Angus Campbell fish of both sexes.  Lewis Slough  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 clear in the correlation  of male body size and reproductive success are not  threespine  stickleback.  between male  size  Studies  have  shown  and territory quality,  that there  reproductive  is  no  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. (1983) investigated aculeatus  interspecific  and Gasterosteus  aggression  Rowland  and dominance between 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 conspecifics Selection  in trying to establish  (without against  energetically  less  which  large  male  efficient  he size at  could could  and defend not  successfully  result  competing  a territory against larger  for  from  court  these  and  mate).  individuals being  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  phylogenetic  pressure.  Different  growth  history and past selective events.  forms  could  Data on sex  simply  reflect  specific growth  80 patterns  in the  threespine  stickleback  are lacking, yet  predictions about the direction of sexual species in the family Gasterosteidae patterns  exist  (i.e.  character") before  is  the  of  are essential  size dimorphism.  to  Data from other  are required to determine if phylogenetic  determinant  role  they  or  selection  ^determinant  growth  in determining the  the  "derived  extent of sexual  dimorphism in size can be assessed.  POPULATION HISTORY  SPECIFIC  REPRODUCTIVE  CHARACTERISTICS  A N D LIFE-  EVOLUTION  The  two  study  populations  differ  in  the  direction  dimorphism (which sex is largest) and in age structure. suggest that  females in the  reproductive and life-history populations  differ  two  populations  have  to  sexual  size  In addition, my results  diverged with respect  parameters. I will attempt  with respect  of  to  discuss  their reproductive traits,  to  how these  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  reproductive season than did Angus Campbell females.  the  course  of  the  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. reproductive  effort  can be measured  by the  If  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 artificial  conclusions  conditions.  environmental results  variables  indicate  the  are derived from data collected  on fish held under  Field  a  as  values  food  relative  may  availability  capability  of  differ and each  as  result  temperature; population  of  such  however,  given  the  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 clutches  because  undifferentiated  this method probably underestimates  stickleback  ovaries  contain  a  large  the number of proportion  of  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 ovulation  laboratory raised fish showed in  Lewis  significant; however, biologically  Slough  females.  a significantly This  earlier date of first  difference  is  statistically  since it is only about five days it is not likely to be  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 containing immature eggs.  as reaching maturity) did possess ovaries  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 is not because the ovaries remain in an immature state.  delayed, the delay  In this population the  ovaries develop and eggs begin to mature but they are not ovulated and may be reabsorbed 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 traits  that  success selection'  in  theory  increase  the  predicts  probability  particular environments. are  the  particularly favored.  most  suites  of  of In  dependant  maximizing  animals  popular theories,  Density  survivorship  and  versus  reproductive  lifetime  reproductive  'bet-hedging' at  this  time  independent  primary variables driving the evolution of life-histories "K" selection  and  and  'r & K  bet-hedging  is  processes are the  under r & K selection.  is thought to occur under conditions of high population density  83 where resources population. resource  are limited.  The '"K" refers to the carrying capacity of a  "K" selection is envisaged as favoring individuals with lower total  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. result  increased  resource  survivorship via these  conditions  early  age  at  acquisition  can be  used  to  enhance  As a  fecundity and  large body size and/or increased reproductive output. Under it  is  first  predicted  that  selection  reproduction, semeloparity  will  favor  large  body  size,  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 agent  in  with age,  the  theory  divergence  proposes of  age-specific  life-histories.  Differential  balanced by the cost of reproduction, increases  maximizing the numbers of surviving young. as the general deleterious fecundity,  mortality rates as the selective  adult  the  effort  probability of  The cost of reproduction is defined  effect of present reproduction on future survival, or  or both (Pianka & Parker 1975;  unpredictable  reproductive  mortality  rates  select  Bell for  1980). increased  High, variable, or reproductive  early in life, therefore for early maturity and a short life span.  effort  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. to  large  fluctuations  in  water  level  and  The Angus Campbell site is subject 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,  Slough fish.  earlier  maturation,  and  a  shorter  life-span  than  Lewis  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. susceptible  Small fish also are  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  accordance work  a  reproductive  season  and  an  individual's  lifetime.  This  is  in  with Dobzhansky's (1956) suggestion that natural selection does not  on  life-history traits in isolation, but rather on the combined  traits of  the whole individual. My  model  interclutch  proposes  that  interval  and  a  positive  body  relationship  size.  The  exists  between  quantitative  clutch  size,  of  these  aspects  relationships are such that given an extended reproductive season the "fecundity advantage"  of  large  females  is  an  illusion.  (reproductive effort) in a season is small clutches females,  and  "fecundity  rather than the  a few  resultant  advantage".  If,  The  greatest  achieved through large  of eggs  the production of  clutches. Hence, small body  increase  in  on  other  the  number  clutch number  give  hand, conditions  small limit  many  size  in  females  a  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  reproductive season.  Slough  population  is  annual  and  has  a  relatively  long  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. increases  the seasonal  clutches.  The mechanism that  production of young is an increase in the number of  This in turn is achieved through small female body size.  large female body size probably is advantageous  In contrast,  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. populations  are  The regression combined,  but  is  this  significant,  however,  result  dubious  is  if data from both 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  populations).  the  slopes  and  intercepts  do  not  vary  between  the  two  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  limited to one or at the most two clutches. a  sex  ratio biased  towards  females  laboratory bred fish show a 1:1  individual females  may be  Trap samples from the field indicate  in the  sex ratio).  Campbell that males may be limiting resource.  that  order of  almost  2:1  Thus, it is possible  (although in Angus  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  unpredictable weather conditions,  breeding  is  constrained  and if late breeding results  by  poor and  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 threespine  mechanisms  stickleback  may  studied  by  operate  in  a  population  Crivelli  and Britton  of Mediterranean  (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. two  clutches.  They  These authors estimate that females produce only one or 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  direction of sexual size dimorphism still remains a problem. that there is  no quantitative  advantage  to  either large  except under the special conditions previously described.  variation in the My data suggest  or small body size, 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). increase  Specifically, the model must account for indeterminant growth, the  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. form of  The actual direction of dimorphism is dependant upon the  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. clutches and length  of a female's reproductive  in each population.  The relationship is  period is  The number of positively  curvilinear and the  correlated  curves resemble  those resulting from hyperbolic saturation equations, with a maximum of 100 days at a total clutch number of 6. must  correspond  with  a  Any further increase in the number of clutches  decrease  in  interclutch  interval.  Individuals  that  90 produce larger numbers of clutches are able to do so because they can produce clutches  at  individual. size;  smaller Mean  however  positive  intervals,  and  interclutch the  relationship  this  interval  relationship is between  is  is  not  interclutch  related  positively significant.  interval  and  to  the  size  correlated  of  with  the  clutch  This  suggests that the  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  normal size range of males.  interestingly  this  size  closely  corresponds  to  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 traditional alternate  negating assumption  the  fecundity  that  advantage  selection  favors  model, large  my  data  females  way of thinking about the consequences of female  challenges  and  provides  body size.  the an 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, seasonal  in  fecundity  some  populations  the  target  of  selection  may not be  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  environmental fluctuations.  unpredictable  events  as  nest  predation  or random  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 sticklebacks  (1979) suggested that probably  where persistent  developed  the  under  type  of  tropical  iteroparity exhibited to  semi-tropical  by the conditions  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. study, therefore,  contribute to  the  The observations made in this  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, total  1985) egg  and some suggestion of a similar relationship between body size, 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. and other species of the  generality,  or  stickleback, pervasiveness,  would be particularly useful of  the  patterns  of  Cyprinids,  in deterrnining  covariation  between  reproductive parameters and body size that were found in this study.  Inter- and  intra-specific  on other  threespine  studies  stickleback  with  these  populations,  taxa,  as  would  well also  as be  further useful  conditions under which sexual dimorphism in size varies.  in  studies  determining  the  Such data would aid in  the development of a general model to describe the evolution of this phenomenon in fish.  93 LITERATURE CITED  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 for Data Analysis Wadsworth.  1983. Graphical Methods  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. 30:364374. 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 infishes.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