c-l THE EFFECTS OF INTRASPECIFIC COMPETITION FOR FOOD ON REPRODUCTION IN THE GUPPY (POECILIA RETICULATA) AND THE IMPLICATIONS FOR THE STOCK AND RECRUITMENT PROBLEM by MICHAEL ROWLAND MORRELL B. A., Stanford U n i v e r s i t y , 1 9 6 8 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS' FOR THE DEGREE OF MASTER OF•SCIENCE i n the Department of Zoology We accept t h i s t h e s i s as conforming t o the required standard. THE UNIVERSITY OF BRITISH COLUMBIA September, 1973 In presenting t h i s thesis i n p a r t i a l f ulfilment of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t freely available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. I t i s understood that copying or publication of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of ~~&>Q I y The University of B r i t i s h Columbia Vancouver 8, Canada i i ABSTRACT Laboratory experiments were conducted with female guppies to determine the effects of intraspecific competition for food on fecundity, fry size, and gestation period. Two groups of females which had been raised at different levels of intensity of competition were -compared. Gestation period was significantly longer i n the high competition group. The effects on fecundity and fry size were dependent on the size of the adult female. The largest females i n the high competition group did not differ with respect to these parameters from individuals of the same size i n the low competition treatment; the smaller females produced fewer but larger fry i n each brood under high competition. The t o t a l weight of fry per brood produced by females of a given size was not different between treatments. These results are discussed from the standpoint that the reproductive characteristics of the individuals of a population can be viewed as t a c t i c a l components of a strategy whose objective i s to maximize the present value of the number of surviving progeny produced by each individual i n the course of i t s lifetime. The optimal distribution over time of energy devoted to reproductive ends and the optimal distribution at one point i n time of a given amount of energy among many or few offspring are expected to vary among populations according to the particular age- and size-specific mortality rates faced by each population. Both magnitude and v a r i a b i l i t y of mortality are important. It i s suggested that by measuring the appropriate reproductive parameters i i i of the adult members of a population, i t should be poss i b l e to make pr e d i c t i o n s about the shape and expected v a r i a b i l i t y of the s t o c k / r e c r u i t r e l a t i o n s h i p . i v TABLE OF CONTENTS Page TITLE PAGE . . i ABSTRACT. i i TABIiE OF C ONTENTS o e e e s o * « e v e « o e o e * o » » * « * e * < » » * O 0 e o o 0 o » » o * o o * * « e « L I S T O F TABXJES© • « e a o e « o * * o * > a o e a o 0 * « * « o * « « o e o « o « « * * o o o o o a e o o * o « LIST OF FIGURES. v i i ACKNOWLEDGEMENTS. .... i x INTRODUCTION. '. 1 MATERIALS AND METHODS ' 9 Experiment 1...... • 10 Experiment 2...... 12 Experiment 3«»« » 15 RESULTS o o o o o o . « . . o o . o 0 « . . . . . . o . . o e « . o . . . . o . . . . . o . . . . . . . . . . . . . . . 17 Experiment 1 17 . Experiment 2............ • 21 Experiment 3............*.....0.**..*.000*0*0000**0.****** 3^ Comparison of Experiments 1 and 3« 31 DISCUSSION. . o . . o . . o © . . . . . . o . . « . « . o . . . . © « . o o . e o e . . . . . © o . o f t . . . . . . 45 Reproductive strategy.... 46 Field studies on poe c i l i i d reproduction................... 5^ Other laboratory studies of fecundity and nutrition....... 65 Studies on exploited marine populations................... 67 Reproductive strategy theory and stock/recruitment........ 82 CONCLUSIONS. SUMMARY BIBLIOGRAPHY APPENDIX v i LIST OF TABLES Table Page I Analysis of variance of standard length of fry. Experiment 3 — high food vs. low food ..... 30 I I " The calculated fecundity of female plaice 37 cm. i n length and estimates of reproductive l i f e span at different l o c a l i t i e s . « . . o o o e o o e e o . . . . o e o . o « o . . . . . . e o o . o o o 71 I I I Reproductive parameters of Atlantic herring tribes. o 76 IV Environment encountered by larval Atlantic herring 81 v i i LIST OF FIGURES Figure Page 1 A family of Ricker s t o c k / r e c r u i t curves 3 2 A family of Beverton and Holt s t o c k / r e c r u i t c u r v e s 0 . 3 3 Schematic drawing of the apparatus used i n Experiments 2 and 3 t o i s o l a t e experimental females, 13 4 " The regression of the logarithm of numbers of f r y per brood on maternal l e n g t h . Experiment 1 . . . • 19 r 5 The regression of the logarithm of numbers of embryos obtained by d i s s e c t i o n on maternal l e n g t h . Experiment 1 . 20 6 The regression of f r y length on maternal l e n g t h . Experiment 1 . . 22 7 The regression of d a i l y consumption of dry food on wet weight of female. Experiment 2 24 8a The regression of numbers of f r y per brood on maternal l e n g t h . Experiment 3 — high food treatment 26 8b The regression of numbers of f r y per brood on maternal l e n g t h . Experiment 3 — low food treatment 27 9a The regression of f r y length on maternal l e n g t h . Experiment 3 — high food treatment. 28 9b The regression of f r y length on maternal l e n g t h . Experiment 3 — low food treatment 29 10 Comparison of the regression l i n e s of the logarithm of numbers of f r y per brood on maternal l e n g t h . Experiment 1 v s . Experiment 3». • 33 11 Comparison of the regression l i n e s of f r y length on maternal l e n g t h . Experiment 1 v s . Experiment 3«« • • • 35 12 The regression of the dry weights of broods of Experiment 3 on ^ ( S L ^ * 8 ) . * . 37 13a The regression of estimated brood dry weight on maternal l e n g t h . Experiment 1. 39 v i i i Figure Page 13b The regression of a c t u a l brood dry weight on maternal l e n g t h . Experiment 3•••• • • 40 14a The regression of index of reproductive e f f o r t on maternal l e n g t h . Experiment 1..... 42 14b The regression of index of reproductive e f f o r t on maternal l e n g t h . Experiment 3**««« • • 43 15 The r e l a t i o n s h i p between the instantaneous rate of population growth, m, and population density f o r r -s t r a t e g i s t s and K- s t r a t e g i s t s . . . . 50 16 Average numbers of embryos c a r r i e d by female Gambusia manni c o l l e c t e d at two l o c a t i o n s i n the Bahamas at d i f f e r e n t seasons 64 17 The regression of fecundity of p l a i c e from d i f f e r e n t areas on reproductive l i f e span... 73 18 Average dry weight of r i p e eggs of various stocks of Clupea harengus of the northeast A t l a n t i c p l o t t e d against month of spavining 77 19 Fecundity-egg s i z e diagram f o r some winter-spring and summer-autumn spawning stocks of Clupea harengus. 78 i x ACKNOWLEDGEMENTS My supervisor, Dr. Norman J . Wilimovsky, gave generously of h i s time and t a l e n t throughout the study. F i n a n c i a l support came from National Research Coun c i l of Canada grants t o Dr. Wilimovsky. Dr. Donald McPhail f i r s t suggested that I consider my problem from the point of view of r - and K- s e l e c t i o n . My understanding of the theory of reproductive strategy increased considerably as a r e s u l t of conversations with Dr. E r i c Charnov and Mr. Stephen Stearns. In the f i n a l phases of the study, I received more than a l i t t l e help from my f r i e n d s . Ms. V a l e r i e Best and Ms. Sharon Heizer typed the manuscript, and Ms. Stephanie Judy prepared several f i g u r e s . Dr. Robin Harger gave encouragement and sound, advice at a c r i t i c a l juncture; without h i s help the work would c e r t a i n l y not have been completed. To Robin, I dedicate t h i s t h e s i s . 1 INTRODUCTION One of the c e n t r a l problems of f i s h e r y biology i s the nature of the r e l a t i o n s h i p between the stock of spawning adults of a population and the strength of the r e c r u i t y e a r - c l a s s which they produce. The problem can be resolved i n t o two components, both of i n t e r e s t to the f i s h e r y manager: 1) What i s the average number of r e c r u i t s produced by a spawning stock of a given size? This r e l a t i o n s h i p over a range of stock s i z e s defines the shape of the f a m i l i a r s t o c k - r e c r u i t curve and allows the manager to determine the optimal spawning stock s i z e . 2) How much v a r i a t i o n i s t o be expected about the average curve, and what are the s p e c i f i c causes of departures from the average? V a r i a t i o n i n strength of year-classes produced by s i m i l a r - s i z e d spawning stocks i s often very large i n n a t u r a l populations. The causes of the v a r i a t i o n are u s u a l l y very d i f f i c u l t to i d e n t i f y and quantify f o r reasons reviewed by Gulland (1965). I f t h i s question could be answered, i t would be p o s s i b l e t o p r e d i c t the strength of r e c r u i t i n g year-classes some time before the a c t u a l time of recruitment, and therefore t o t a l stock s i z e s could be f o r e c a s t . The t y p i c a l approach to estimation of the s t o c k - r e c r u i t r e l a t i o n -ship has been to c o l l e c t estimates of spawning stock si z e and r e s u l t a n t 2 recruitment for a series of years and to f i t a curve to the data. Larkin (1973) has discussed the different types of curves which may be used. Curves described by Ricker (1954, 1958) and Beverton and Holt (1957) are the two types most often used i n management; examples of these curves are shown i n Figures 1 and 2 . In both families of curves, recruitment increases with stock size at low population levels; the major difference between them i s that at high stock levels the Ricker curve indicates an absolute drop i n recruitment, whereas the Beverton and Holt curve approaches an asymptote. In both families of curves recruitment per unit stock reaches a maximum at intermediate stock sizes and decreases with further increases i n spawning stock; thus both curves imply the existence of density-dependent processes regulating population size. The viewpoint adopted i n this study i s that the stock/recruit curve for a population represents the result of a l l the interactions of the reproductive characteristics of the population with particular aspects of the environment. Reproductive characteristics of individual members of the population are assumed to have evolved by natural selection; the combination of reproductive t r a i t s shown by a given individual - age at f i r s t , maturity, fecundity, reproductive lifespan, etc. - can be considered as tactics i n a reproductive strategy whose objective i s to maximize individual fitness. The hope i s that from an understanding of these reproductive tactics i n an evolutionary context w i l l come an understanding of their population consequences - that i s , the stock/recruitment relationship of the population. The observed year-to-year variation i n recruitment indepent of stock density i s assumed to 2.0 SPAWN E R S - W Figure 1. A family of Ricker stock-recruit curves. The general equation i s Z = W e a ( 1 - W ) . In curve a, a = 0.667; i n b, a = 1.0; i n c, a =""2.678. A f t e r Ricker (195S).. a J 1 i : 1 1 2 3 4 Number of eggs x 10-'°(E) Figure 2. A family of Beverton and Holt stock/ r e c r u i t curves. The general equation 1 i s R = cc+fi/E. A f t e r Beverton and Holt (1957). 4 r e s u l t from density-independent f l u c t u a t i o n s i n the relevant environmental v a r i a b l e s ; i f t h i s v a r i a b i l i t y has been present i n much the same form over evolutionary time, the reproductive t a c t i c s of the population should r e f l e c t some accommodation to i t . The experimental part of the present study deals with absolute fecundity and c l o s e l y r e l a t e d v a r i a b l e s . Bagenal(l973) has reviewed the p o s s i b l e relevance of fecundity t o the stock and recruitment problem. He c i t e s several f i e l d studies showing decreases i n average absolute fecundity with i n c r e a s i n g population density, namely Bagenal (1963b, I 9 6 5 ) , Hodder (1963), K i p l i n g and Frost (1969), R a i t t (1968).- Laboratory studies by Bagenal (1969a) , Hester ( 1964) , and Scott (1962) provide evidence that fecundity of i n d i v i d u a l females i s p o s i t i v e l y r e l a t e d to n u t r i t i o n a l l e v e l . The hypothesis which r e s u l t s from t h i s i s that observed reductions i n fecundity at high population density r e s u l t from reduced i n d i v i d u a l feeding l e v e l mediated by i n t r a s p e c i f i c competition; t h i s hypothesis has been suggested several times i n connection with the f i e l d studies c i t e d above. Bagenal (1973) p o i n t s out that the observed r e l a t i o n s h i p between fecundity and population density i s of the form necessary to produce the stock-recruitment curves of Beverton and Holt and of Ricker, although i t i s evident that other f a c t o r s , such as compensatory m o r t a l i t y of eggs and larvae,can produce the same e f f e c t . 5 The objectives of this study, then, are the following: 1) to describe quantitatively the effects of intraspecific competition for food on fecundity and related reproductive parameters, and 2) to relate these findings to the stock and recruitment problem. Herein competition i s taken to mean the demand by two or more organisms for the same resource i n excess of immediate supply; this definition i s essentially that of Milne ( 1961) . It i s of interest to distinguish two different courses that competition may follow: "contest" and "scramble", i n the terminology of Nicholson (1954)• In contest competition, some of the competitors are able to satisfy completely their requirements for the scarce resources, while others secure l i t t l e or none. In scramble competition, on the other hand, the available resource i s divided more or less randomly among the competitors, and a l l receive less than their f u l l requirements. The effect of competition, regardless of the course i t takes, i s to reduce consumption of the scarce resource (relative to f u l l requirements) for some or a l l of the competitors. When scarce food i s the object of competition, at least some of the competitors v a i l be forced to exist on reduced rations. This reduced energy input to an individual may be allocated among the conflicting internal demands of maintenance, growth, and reproduction i n different ways. Energy expenditure i n a l l categories may be reduced, or expenditures 6 i n one category may be maintained at a high level at the expense of the others. In this study attention i s focussed on reproduction, but i t i s important to remember that reproduction i s only one of several processes which may be affected by food shortage. It would be extremely d i f f i c u l t to pursue the objectives set out-above i n a natural f i s h population. The intensity of competition must depend on population density and the available supply of scarce resources; the identification of c r i t i c a l resources and measurement of available supply would be very involved. Furthermore, i t i s probable that f i s h fecundity at a given time i s significantly influenced by events occurring from one to several reproductive seasons previously (Bowers and Holliday, '1961; Hester, 1 9 6 4 ; Hodder, 1 9 6 3 ; Kipling and Frost, 1 9 6 9 ) ; therefore, i t i s desirable to know the history of a given individual over several reproductive seasons. Both of these problems are far more easily handled i n the laboratory than i n the f i e l d . The guppy Foecilia reticulata (Peters) was chosen as the experimental animal for several reasons. Guppies are relatively easy to maintain i n the laboratory and reproduce readily throughout the year. In addition, their short gestation period (about one month) makes i t possible to collect a considerable amount of data on reproduction i n a reasonable time. 7 In Poecilia f e r t i l i z a t i o n i s internal, and the embryos develop inside the ovary. The fry are released by the female at an advanced state of development; they can swim actively at birth and normally begin to feed soon after release. The schedule of stages of oogenesis and embryonic development have been worked out for Poecilia reticulata by Stolk (1951)• The ovary of a mature female normally contains oocytes and embryos of three discrete stages: (l) many small undifferentiated oocytes of less than 0.1 mm. diameter, (2) a smaller number of medium-sized oocytes up to 0.3 mm, diameter containing some yolk, (3) either a set of large yolky oocytes (or ova) or a similar number of developing embryos. I f we take the day of parturition of one brood as day 0 and assume a constant brood period of 2.8 days, we may follow the development of the oocytes that w i l l eventually constitute one brood. On day 0 some of the many small oocytes begin' to differentiate and accumulate yolk. By day 28 these oocytes have become medium-sized oocytes as defined above; they join a pre-existing group of medium-sized oocytes, according to Stolk. Following the next parturition (day 56 in this model), some of the medium-sized oocytes lay down the bulk of their yolk supply very rapidly, maturing to large ova, and are f e r t i l i z e d . The resultant fry are born on day 84. On the basis of Stolk's work, i t seems possible that.events, occurring up to three brood periods before parturition may affect the numbers and sizes of fry produced. In an experimental study of the 8 effects of feeding on guppy reproduction, Hester (1964) found that brood size was significantly affected by feeding regime of the female during the two brood periods immediately preceding parturition. Hester vras unable to rule out the possibility of three period effects. I f an individual guppy reduces the amount of energy dedicated to reproduction, one would expect some or a l l of the following effects: 1) a reduction i n number of fry per brood, 2) a reduction i n size of individual fry, 3) an increase i n time between broods. A l l of these potential effects are examined i n this study. . 9 MATERIALS AND METHODS The guppies used i n this study were descendants of f i s h collected i n August, 1967j i n Trinidad, West Indies. The i n i t i a l collection consisted of something over 100 adults and immatures taken by B. Seghers from the Tacarigua River about four kilometers northeast of i t s confluence with the Caroni.River. The fish were shipped by a i r to Vancouver, where the stock was maintained i n aquaria i n the laboratory of Dr. N.R. Liley at the University of Bri t i s h Columbia. At the time of the i n i t i a l collection the Tacarigua at the collecting site was three meters broad with a rather uniform depth of about ,15m. Current was measured, at about .4m/sec, and the water was clear; however, after a rainstorm, the stream becomes turbid, and the flow may increase by an order of magnitude. Aside from an algal film on the stream bottom, there were no aquatic plants at the collecting s i t e . Water temperature was not recorded i n 19°7» but at a similar site on the nearby Aripo River, temperature varied from 25.5 to 26.5°C i n the summer of I969. The. c i c h l i d Crenicichla alta and the characid Hoplias malabaricus, both of which prey on adult guppies, were present at the Tacarigua site. Other fishes present included the c i c h l i d Aequidens pulcher, the characids Astyanax bimaculatus and Corynopoma r i s e i , the l o r i c a r i i d Hypostomus ro b i n i i , and a pimelodid of the genus Rhamdia (Seghers, 1973» and pers. comm.). 10 From September, 19°9i several hundred of the descendants of the original Tacarigua stock were maintained i n the laboratory as a source of experimental animals. The f i s h were kept i n 64 l i t e r capacity glass aquari with floating plants, Ceratophyllum sp., for cover. Water temperatures were maintained at 26 + 1° C by the use of thermostatically controlled aquarium heaters. Aquaria were not exposed to natural l i g h t ; the^main illumination for the tanks was provided by 20 watt fluorescent lights about 35 cm. above the surface of the water. Lights were automatically controlled to maintain a constant day length of 14 hours throughout the study. A l l aquaria were f i l l e d with tap water to which was added one gram of rock salt per l i t e r 1 ; water lost by evaporation was replaced at frequent intervals with tap water. On a weekly basis 28 1. of water were removed from each tank and replaced with fresh tap water plus salt. Each tank.was equipped with a glass wool and charcoal f i l t e r to circulate, aerate and f i l t e r the water. - The main food during most of the study was l i v e Tubifex sp. (Oligochaeta); f i s h i n stock tanks were fed once daily i n amounts sufficient to maintain an almost continuous supply of l i v e Tubifex on the tank bottom. This basic diet was supplemented with small quantities of Tetramin, a commercial dry f i s h food. Departures from this basic diet w i l l be discussed below. Experiment 1: Measurement of Reproductive Parameters of Isolated i Females Fed on Excess Tubifex _______________ In August, 197O5 24 healthy mature females were selected from stock tanks as experimental animals. Fish were selected so as to 1 1 represent, equally a l l lengths of mature females present i n the stock " tanks (14-45 mm. standard length). Individual females were isolated i n glass jars of 1.95 !• capacity containing 1.3 1. of water with floating Geratophyllum as cover for fry. Jars were kept i n 26° C water baths. The fi s h were fed once daily, six days a week with excess Tubifex. Each day before feeding, feces and food remaining from the previous day were siphoned from the jars, and the water was passed through a charcoal and glass wool f i l t e r and returned to the jar. At weekly intervals half the water i n each jar was removed and replaced with tap water plus salt.. , At lease once every 24 hours, jars were carefully examined for the possible presence of fry. Upon discovery, fry were removed from the jar, anaesthetized with MS-222, and measured, to the nearest 0.1 mm. standard length under a compound microscope, using a technique adapted from Counts ( l 9 o l ) . After a l l fry had been released, the female was anaesthetized with MS-222 and measured to the nearest 0.1 mm. standard length with d i a l calipers. The female was then quickly blotted on a paper towel and placed i n a pre-weighed beaker of water; the change i n weight of the beaker plus f i s h was recorded to the nearest .01 g. The female was then returned to her original jar, to which two males had been added. The males were removed after five days, during which time they were presumed to have inseminated the female (see Rosenthal, 1952). 12 Experiment 2 ; Measurement of Ration of Dry Food Experiment 1 provided descriptive information on reproduction of f i s h fed i n excess; the next step was to measure the effects of reduced rations. In order to set up a treatment of reduced rations for fis h of different sizes, i t was necessary to establish a curve of f u l l rations as a function of body size. Two separate attempts were made to measure guppies* actual consumption of Tubifex fed i n excess. Neither .experiment produced usable results; the main problems were the dif f i c u l t y , of accurately weighing the l i v e Tubifex before feeding and of recovering a l l of the uneaten . portion after feeding. To circumvent the d i f f i c u l t i e s of weighing l i v e Tubifex, dry food (Tetramin) was used.exclusively i n a l l experiments and i n the stock tanks beginning i n June, 1971• In this and subsequent experiments'a different system was used to isolate individual females i n such a way as to.minimize predation on fry and to provide improved circulation and f i l t r a t i o n of water. Nylon mesh was attached to a plexiglass frame, using Pliobond glue or silicone aquarium sealant, to form 10 separate compartments, each measuring 10.2 x 12.5.x 2 7 . 8 cm. Each framework was placed i n a 64 1..capacity aquarium with a f i l t e r system as described above. The water level was maintained so that each individual compartment enclosed about 3»5 1« of water. Figure 3 i s a schematic drawing of the apparatus. Each compartment which contained an adult female (A.C,D,G,I,J) had fine mesh on three sides; even newborn guppies were too large to penetrate 13 A j B y C D ^/ ////// F G 1 J Figure 3« Schematic drawing of the apparatus used i n Experiments 2 and 3 to isolate experimental females. The outside dimensions are 20.4 x 62.5 x 27.8 cm. a) Lateral view b) Top view. Solid lines indicate small mesh; broken lines indicate large mesh. 14 this mesh. The fourth side consisted of larger mesh through which fry-but not adults could pass. Thus fry being pursued by their mothers were able to escape through the large mesh into the adjacent refuge compartment (B,E,H~). Each refuge compartment was divided diagonally by a fine mesh barrier i n order to separate the fry of different females. Fry were observed to use the refuges i n the intended manner. Twenty experimental females were selected from stock tanks so as to give equal representation to a l l size classes present i n the stock tanks (14-36 mm.).. The experimental animals were placed i n the individual mesh compartments of the isolation apparatus. After a four day acclimation period, during which f i s h were fed and excess food was removed once daily, the experimental procedure was begun. For four consecutive days a weighed quantity, of dry food was placed in'each compartment, and the f i s h were allowed to feed undisturbed for one hour. (A p i l o t study showed that v i r t u a l l y a l l feeding occurs i n the f i r s t 15 minutes after food i s presented, and that after an hour none of the f i s h studied showed any further interest i n the remaining food.) At the end of the hour the amount of food remaining i n each enclosure was recorded qualitatively (0, trace, or +), and any excess was removed. On the following day the amount of food offered was increased for f i s h which had consumed a l l or nearly a l l of the previous day's ration, and food was reduced for the others. In this manner, i t was possible to bracket roughly the actual amount of food consumed i n one feeding by an individual f i s h . 15 Experiment 3 * Measurement of -Reproductive Parameters of Isolated Females at F u l l and Half Rations of Dry Food . In August, 1971> two groups of 10 healthy females were selected from the stock tanks for use i n this experiment. Before the f i s h were actually selected, 10 values of standard length were picked which covered the range of sizes available i n the tanks (14-28 mm.). Fish from the stock tanks were then measured, and the f i r s t ones f a l l i n g within 0.5 mm.- of the predetermined lengths were assigned to f u l l or half rations by the toss of a coin. Each selected f i s h was placed i n the isolation compartment which had been previously designated at random for that specific length and ration. Thus f i s h wefe randomly assigned to food treatment and tank position, with the restriction that each treatment group contained the same range and distribution of sizes. The isolated guppies were daily fed weighed quantities of Tetramin representing f u l l or half rations for their particular weight. Feces and food from the previous day were removed prior to each feeding. Following each parturition, rations were adjusted to the appropriate percentage of post parturn weight of the female ( 3 . 0 % for high, 1,5%> for low food treatment) i n order to allow for weight changes due to growth during the experiment. Tanks were checked at least once daily for newborn fry, and when births occurred the same procedures were followed as described for Experiment 1. In addition, i n this experiment after the standard 16 length of each fry had been measured, the entire brood was placed i n a drying oven at 60° C, and the dry weight of the brood as a whole was measured to the nearest 0.1 mg. after 48 hours, by which time the weights were constant. 17 RESULTS In Experiments 1 and 3 mortality of experimental f i s h was high. The causes of the mortality are not known; possibly the restriction of normal activity imposed by the isolation chambers was somehow involved. Typically f i s h stopped feeding and remained inactive for a few days or a week before dying. Dead fi s h showed no external growths, lesions, or parasites, and post-mortem dissections revealed no internal abnormalities. Because of the losses of experimental animals, both experiments were terminated earlier than anticipated. In Experiment 1, only three fi s h survived to day 50; of these, one died on day 60, and the other two were alive on day 70, when the experiment was stopped. Experiment 3 was terminated on day 51'with four survivors. As discussed i n the Introduction, i t was anticipated that reproductive effects of nutrition would be apparent only after a time lag of up to three months. Thus, most, i f not a l l , of the data on reproduction collected i n the two experiments reflect feeding conditions i n the stock tanks prior to the experiments themselves. This idea i s supported by the results of Experiment 3t i n which there are no significant differences between the high and low food treatments with respect to either numbers or sizes of fry produced. Experiment 1 I t was anticipated that maternal size would have significant effects on both numbers and sizes of fry produced. Figure 4 shows 18 the relationship between the natural logarithm of number of fry i n a brood and the standard length of the mother, the points a l l l i e near a straight line with positive slope significantly different from zero (p< .001). Eighteen of the experimental females which died were dissected, and their ovaries were examined. Fifteen of these contained developing embryos, althouth i n one case decomposition had progressed so far that i t was impossible to count the embryos. Of the three females without embryos, one had produced a brood five days before dying, and the next batch of large oocytes had not yet been f e r t i l i z e d ; i n the other two cases there was no development of oocytes beyond the smallest, least differentiated stage. Figure 5 shows the regression of the natural logarithm of number of embryos on maternal length. The positive slope i s significantly different from zero (p< .001). A comparison of the regressions of Figure 4 and Figure 5 may provide a measure of the extent of maternal cannibalism, since the embryos counted by dissection had not, of course, been exposed to predation. The comparison was carried out by means of analysis of covariance. The hypothesis that the two regressions have a common slope was accepted at a low level of probability (p=.08). There was a nonsignificant tendency for the excess of embryos over fry actually born and recovered to be greater for smaller females than for larger ones; the largest females had about the same numbers of progeny i n both cases (the regression lines intersect at a standard length of 3 9 « 7 mm.). The hypothesis that the elevation of the two regression lines i s the same was rejected (p=.01); 19 EOCK„ ±00 °± U . L L a m ID STANDARD LENGTH OF FEMALE a ) Figure 4» The regression of the logarithm of numbers of f r y per brood on maternal length. Experiment 1. ±00°.. STANDARD LENGTH OF FEMALE ( Figure 5. The regression of the logarithm of numbers of embryos obtained by dissection on maternal length. Experiment 1. 21 i f the regressions are given the same slope, then for a l l lengths of females there were significantly more embryos than fry. This result suggests that the numbers of a l l broods actually born may have been reduced by cannibalism before the fry were removed and counted. However, the differential effect of maternal size i s the opposite of what would be expected i f predation were the cause of the difference. Significant pre-natal mortality could also account for this effect. The data collected i n Experiment 3 shed more light on this problem. The relationship of length of fry to length of the mother i s positive (Figure 6), but there i s considerable variation of length of fry within broods. A straight line f i t to the points has a positive slope significantly different from zero (p< .001), but accounts for only about one-fourth of the t o t a l sum of squares. Addition of higher order terms (up to quartic) to the regression equation does not significantly reduce the residual. Two females produced two broods each, and two others produced three and four broods. Thus four individuals gave seven values of time between broods; the mean interval was 24.8 days (standard deviation:2.9; range:22-28 days) e Experiment 2 After four days of experimental feeding, five f i s h had not eaten a significant amount on any day; these were a l l very inactive and were excluded from the analysis. Of the remaining 15, three produced 22 7»0. . s 10» , 15* SO" S5» 30° 35= 40" 45« 50* STANDARD LENGTH DF FEMALE ( M M O Figure 6. The regression of f r y length on maternal length. Experiment 1. The regression l i n e was c a l c u l a t e d on the b a s i s of the i n d i v i d u a l lengths of 343 f r y . For c l a r i t y , only the mean length of each brood + two standard e r r o r s has been p l o t t e d . 23 broods during the i n i t i a l acclimation period; one of these died the day after giving birth. One male was added to each of the other two compartments i n order to maintain a supply of fresh sperm for the reproductive study which was to follow; the data on feeding for these compartments were also excluded from the analysis. The remaining twelve individuals a l l consumed a l l or nearly a l l of the food offered on at least one day. Where different amounts were eaten on different days, the higher value was used i n the analysis. The solid line of Figure 7 shows the results of the regression analysis of weight of dry food eaten on li v e weight of female; the linear regression accounts for almost 9&f0 of the t o t a l sum of squares. The Y-intercept of the line i s at 0.62 mg. A regression line passing through the origin (broken line i n Figure. 7) f i t s the data almost as well; for convenience this second line was chosen as the basis for calculations of rations for Experiment 3« Experiment 3 This experiment was designed to measure the effects of reduced . rations on reproduction. I t was intended that the experiment should last for three brood periods i n order to allow for the pos s i b i l i t y that brood size i s determined early i n the process of oogenesis. However, mortality of experimental f i s h was very high (half the fish died i n the f i r s t 28 days), and the experiment was terminated on day 51 with four survivors. 24 0« 200* 400» 600° 600° 1000" WEIGHT OF FEMALE CMGO Figure 7 . Hie regression of d a i l y consumption of dry food on wet weight of female. Experiment 2, Solid l i n e : best lin e a r f i t (Y = .6236 + .0287 X). Broken l i n e : best lin e a r f i t passing through o r i g i n (Y = .02984 X). 25 In the course of the experiment, 11 different females produced 16 broods--eight broods i n each treatment group. Within each treatment group there i s a significant positive regression of the natural logarithm of number of fry on maternal length; the hypothesis of zero slope was rejected i n each treatment group (high food:p< . 0 1 ; low food .05>p>.02). The two regressions are shown in Figure 8 . The residual error about the regression line i s significantly greater for low food than for high; this violates the assumptions of analysis of covariance, so s t a t i s t i c a l comparison of the two treatments was carried out as recommended by Snedecor (1956:97) by means of a t-test using 6 rather than 12 degrees of freedom. Hypotheses of equal slopes and equal intercepts were both accepted (p>.5)« In Figure 9 fry size i s plotted as a function of maternal size for each treatment group. Two broods from the high food group were born prematurely; fry from these broods had large yolk sacs which imposed an appreciable curve on the body, preventing accurate measurement. These two broods were excluded from this part of the analysis. For neither treatment i s the slope of the regression line significantly different from zero. Possible differences i n average size of fry between treatment were tested by means of a two-level nested analysis of variance (Table I ) . There were highly significant differences i n average standard length of fry among broods within treatments (p<.005), but the difference i n average fry length between treatments was not significant i n comparison (.25>p>.10). 26 a o o * T STANDARD LENGTH DF FEMALE (MM-) Figure 8a. The regression of the logarithm of numbers of fry per brood on maternal length. Experiment 3 — high food treatment. Figure 8b. The regression of the logarithm of numbers of fry per brood on maternal length. Experiment'3 — low food treatment. 28 — N n 10« 15° SO- E5» 30« 35= 40» 45» 50« STANDARD LENGTH DF FEMALE (MM*) Figure 9a. The regression of fry length on maternal length. Experiment 3 — high food treatment. The regression line was calculated on the basis of the individual lengths of 47 fry. The mean length of each brood + two standard errors has been plotted. 29 7°0. w 6»a. >-L L LL a z a < D < CJJ B°SJ-5*4.. B«0. 5»a. 5oa(_ f~ 10« ±5« so< E5« 30 < 40« 45« 50 < STANDARD LENGTH DF FEMALE ( Figure 9b. The regression of fry length on maternal length. Experiment 3 — low food treatment. The regression line was calculated on the basis of the individual lengths of 5° fry. The mean length of each brood + two standard errors have been plotted. \ 30 TABLE I. Analysis of variance of standard.length of f r y — high food vs. low food Source of Variation d. f. SS MS F Between treatments 1 .879 -879 2.149(1,12) NS Among broods within treatments 12 4.912 .409 12.58(12,89) ** Within broods; error 89 2.890 .0325 Total 102 8.681 31 Comparison of Experiments 1 and 3 The failure of the fish of Experiment 3 to demonstrate significant differences i n reproduction under different feeding regimes i s not surprising in view of the response time-lags predictable on the basis of Turner's (1937) and Stolk's (1951) work on oogenesis and Hester's (1964) experiments on reproduction and nutrition. Most of the broods of Experiment 3 were born before the experimental treatment was expected to have taken effect. The data from this experiment probably result from conditions of l i f e i n the stock tanks prior to the experiment; the same i s probably true for the fi s h of Experiment 1, although i n this case the experimental feeding regime was th*e same as that of the stock tanks. The fi s h of Experiment 1 were fed l i v e Tubifex i n excess, both i n the stock tanks and i n the experiment. For two months prior to the beginning of Experiment 3» fi s h i n the stock tanks were fed exclusively on Tetramin given i n a single daily feeding. I have no information on the relative nutritional value of Tubifex compared to Tetramin, but experienced aquarists are generally of the opinion that l i v e food i s better for fish than dry preparations. After introduction to the aquaria, uneaten Tubifex survived indefinitely on the bottom, and fis h fed on them throughout the day. The dry food, on the other hand, was taken primarily from the surface or as i t sank through the water column; uneaten flakes lying on the bottom were usually ignored. Thus the Tubifex-fed fish had the opportunity to feed at w i l l throughout the day, whereas the dry food group effectively was restricted to less than one hour of feeding each day. 32 Maximum sizes attained by f i s h fed on Tubifex were considerably greater than those of f i s h raised on dry food. In selecting the experimental fish for Experiment 1, the largest female measured had a standard length of 45«2 mm.; females larger than 30 mm. were common i n a l l stock tanks. In setting up Experiment 3» on the other hand, the largest female encountered measured only 27.8 mm. This i s an indication that growth and/or survival rates were lower when only dry food was provided. Densities of f i s h i n the stock tanks were not measured, but they were similar under both feeding regimes. Each 64 1. capacity stock tank contained approximately 100 to 150 adult guppies plus a similar but more variable number of fry and immatures. It appears that the stock tank histories of the f i s h used i n Experiments 1 and 3 are different enough that comparison of the results of the two experiments provides valid information on the central problems of this study. Unfortunately, densities and feeding regimes i n stock tanks were not carefully controlled; however, the qualitative history of the f i s h of Experiment 1 was of good nutrition and low intensity of competition, whereas the f i s h of Experiment 3 experienced poorer nutrition and consequently more intense competition i n the months preceding the experiment. On this basis the results of the two treatment groups of Experiment 3 have been pooled for comparison with Experiment 1. In Figure 10 are reproduced the lines of best f i t of the regressions of the natural logarithm of numbers of fry on maternal 33 > LL LL a ra z 10- IS" so 45" 50' STANDARD LENGTH OF FEMALE ( M M O Figure 10. Comparison of the regression lines of the logarithm of numbers of fry per.brood on maternal length. Broken l i n e : Experiment 1 (points are shown i n Figure 4) • Solid l i n e : pooled data from Experiment 3 .(poi^s are shown i n Figures 8a and 8b). 34 length for the two experimental groups; the points on which the lines are based are shown i n Figures 4 and 8. Of course, only the section of the lines corresponding to females under 28 mm. i s valid for comparative purposes, since no females above that size produced broods i n Experiment 3« The data were compared by analysis of covariance, and the hypothesis that the two regressions were samples from populations having the same slope was rejected (p=.05). The smaller females i n Experiment 3 produced significantly fewer fry per brood than did females of the same size i n Experiment 1. The lines intersect where maternal length i s 26„9 mm.; i n other words, the largest females of the dry food group produced about the same numbers of fry per brood as females of corresponding size i n the Tubifex group. This i s the same sort of effect that appeared as a non-significant trend i n Experiment 1, when numbers of embryos carried by fish of a given size were compared with numbers of fry actually born and recovered for counting. Figure 11 shows the regression lines of fry lengths on maternal lengths for Experiment 1 and the pooled data for Experiment 3 ; the points are shown i n Figures 6 and 9« The two regressions were compared by analysis of covariance, and the difference i n the slopes was found to be highly significant (p=.004). Again, i n the size range of interest (maternal length 28 mm.), the smaller females i n the two groups show a substantial difference, while the larger ones perform about the same. The lines intersect at a maternal length of 2 6 . 6 mm. (almost identical to the point of intersection of the lines representing numbers per brood). 35 7«0U-S e»a. >-(Z LL, LL. a I ID —™ro /Cm LJ a rz < n < I— UJ 6°a "I————+• 4™ 10" 15« SO" 30» 33° 40" 4tj< STANDARD LENGTH DF FEMALE (MMO Figure 11. Comparison of the regression lines of fry length on maternal length. Broken l i n e : Experiment 1 (points are shown i n Figure 6). Solid l i n e : pooled data from Experiment 3 (points are shown i n Figures 9a and 9b). 36 and length of f r y ) are not s u f f i c i e n t to describe unambiguously the diff e r e n c e s i n reproductive response of the two experimental groups. I t seems obvious that the smaller females responded d i f f e r e n t l y under the d i f f e r e n t n u t r i t i o n a l regimes, but i t i s not c l e a r whether the t o t a l commitment of energy to reproduction by the small females i n Experiment 3 a c t u a l l y decreased or was simply d i s t r i b u t e d among fewer and l a r g e r f r y per brood. By the beginning of Experiment 3» i t had become evident that the t o t a l weight of. each brood should be measured i n add i t i o n to the lengths of i n d i v i d u a l f r y ; accordingly, i n Experiment 3 the dry weight of each brood as a whole was measured, as described above under Mat e r i a l s and Methods. In order to obtain an estimate of the dry weights of broods produced i n Experiment 1, the following technique was used. Assuming that dry weight of an i n d i v i d u a l f r y i s p r o p o r t i o n a l to i t s length r a i s e d to some exponent, i t follows that D V o o d = a + ^ S L X ) ( 1 ) where — b r o o d 1 3 t h e w e i & h t of an e n t i r e brood, (ISL ) i s the sum of the standard lengths of the fry-making up the brood, i n d i v i d u a l l y raised"to an exponent x, and a and b are the constants of a l i n e a r regression equation. Equation ( l ) was f i t to the data of Experiment 3 using t r i a l values of x ranging from 0 . 1 to 1 0 . 0 . The r e s i d u a l sum of squares was minimal when x was 4 . 8 ; f o r t h i s value of the exponent, the equation accounted f o r almost 97$ of the observed v a r i a t i o n i n dry weight of broods (the f i t to the points i s shown i n Figure 1 2 ) . This equation Figure 12. The regression of the dry weights of broods of Experiment 3 on 2_(SL +* )• See text f o r explanation. 3 8 was then used to calculate estimate of the dry weight of broods produced i n Experiment 1 from the data on standard length of fry i n those broods. In Figure 13 are plotted the natural logarithms of brood dry weights as a function of maternal length for the two experimental groups: the positive slopes are obviously significant ( p < . 0 0 1 i n both cases). When the two regressions are compared by analysis of covariance, the tendency for the smaller females of Experiment 3 to produce.lighter broods than those of Experiment 1 i s found not s t a t i s t i c a l l y significant; the hypothesis that the regressions have a common slope i s accepted (p=12). Additional variance may have been introduced into the data of Experiment 1 by the use of estimates rather than direct measurements of dry weight of broods; i t i s possible that this added error plus the small sample sizes have resulted i n the erroneous acceptance of a false n u l l hypothesis that the slopes are equal. In the same analysis the hypothesis that, given the regressions have a common slope, the elevation of the lines i s not different, i s also accepted (p= . 4 2 ) . Gadgil and Bossert ( 1 9 7 0 ) suggest that reproductive effort, defined as the fraction of an organism's energy resources devoted to reproduction at a given time, be measured as the ratio of weight of sex products to to t a l body weight. I have calculated the ratio of dry weight of each brood to the wet weight of the mother just after parturition; this index of reproductive effort should be nearly proportional to Gadgil and Bossert's measure. 39 Figure 13a. The regression of estimated brood dry weight on maternal length. Experiment 1. i 40 EOCK 100* —N B STANDARD LENGTH DF FEMALE CMMO Figure 13b. The regression of a c t u a l brood dry weight on • maternal length. Experiment 3.— pooled data. 41 In both Experiment 1 and Experiment 3 the index of reproductive effort tended to increase as maternal length increased (Figure 14) . Although there i s a good deal of scatter about the regression lines shown i n Figure 14, the positive slope of the line i s significantly different from zero for Experiment 1 (p=.03) . In Experiment 3 the increase of reproductive effort with size i s not significant (p=. l8)„ Comparison of the two sets of data by analysis of covariance showed that the two regressions were indistinguishable with respect to either slope (p=.99) or elevation (p=.31) of the lines of best f i t . In Experiment 3 five different females produced two broods each. The mean time between broods was 2 9 . 8 days (standard deviation: 2 . 9 5 ; range: 25-32 days). Although sample sizes are small, a comparison of these results with those of Experiment 1 (mean gestation period= 2 4 . 8 days) leads to the conclusion that the mean time between broods was significantly smaller i n Experiment 1 than i n Experiment 3 (t= 2 . 8 6 , 10 d.f., p<.02). Following i s a brief summary of the conclusions from the comparison of the results of Experiment 1 with those of Experiment 3« Under conditions of relatively poor nutrition and, presumably, more intense competition, time between broods increases for a l l females of the sizes studied. The smaller females i n an aquarium respond to deteriorating nutritional conditions by producing fewer but larger fry i n each brood, while larger females are essentially unaffected. As 42 STANDARD LENGTH DF FEMALE ( M M - ) Figure 14a. The regression of index of reproductive effort on maternal length. Experiment 1, 0»OHJL Figure 14b. The regression of index of reproductive effort on maternal length. Experiment 3 — pooled data. accurately as can be determined from the present study, females of the same size produced about the same t o t a l weight of fry per brood under both nutritional regimes, but there i s a suggestion that smaller females produced lighter broods under the poorer feeding regime. 45 DISCUSSION Svardson (1949) noted that, a l l other things being equal, natural selection should favor genotypes producing higher numbers of eggs, and so average fecundity within a population should increase with every generation. Since natural populations do not show this steady increase i n fecundity over.time, he reasoned that there must exist some selective pressure towards decreasing egg number and that observed fecundities must represent a compromise between opposing forces. Svardson suggested four selective forces which might operate to limit fecundity i n f i s h ; his discussion i s summarized i n the following paragraphs. Fi r s t , a f i s h must dedicate some energy to growth and maintenance i n addition to reproduction. Increased energy devoted to growth and maintenance processes should increase the probability that an individual w i l l survive u n t i l the next reproductive season. Also, since f i s h fecundity often increases with body size, energy put into growth during one time period may result i n increased fecundity i n subsequent spawnings. Thus i n an evolutionary sense the advantages of increased fecundity at a given time may be counterbalanced by the future reproductive advantages l i k e l y to result from alternative uses.of available energy. Second, there may be a physiological strain involved i n producing increasing numbers of eggs. This strain may directly reduce an individual's chance of surviving u n t i l the next reproductive season. 4 6 Third, f o r species which provide some form of parental care f o r t h e i r eggs and young, the l i m i t s on fecundity may be set not by the number of eggs which a female can produce but by the number of eggs or young which can be successfully cared f o r . The argument here follows that of Lack (1948) f o r b i r d s . F i n a l l y , i f there i s a premium on large size of i n d i v i d u a l eggs, some compromise must be made between egg size and egg number, since, for a given input of energy to reproduction, any increase i n energy per egg must reduce the t o t a l number of eggs produced. Svardson c i t e d evidence from several sources to show that large eggs produce large larvae, which have an advantage over smaller i n d i v i d u a l s i n i n t r a s p e c i f i c competition. Svardson*s paper generates the hypothesis that f i s h populations whose larvae face intense i n t r a s p e c i f i c competition w i l l produce fewer and larger eggs than populations of the same species f o r which i n t r a s p e c i f i c competition i s a less important source of l a r v a l mortality; i n these l a t t e r populations predation and unfavorable physical factors, which are assumed to act more or l e s s independently of l a r v a l size, would be the important causes of l a r v a l deaths. Reproductive Strategy Ideas sim i l a r to Svardson's have been developed i n more general and detailed form i n the recent l i t e r a t u r e on r - and K- selection. MacArthur and Wilson (1967: 145-180) coined the terms i n discussing optimal reproductive strategies i n uncrowded and crowded environments. They argued that i n an uncrowded environment with abundant resources, those genotypes w i l l be favored which can u t i l i z e the most resources 47 and produce the largest numbers of progeny per unit time,, maximizing r, 1 / the intr i n s i c rate of population increase of the logistic equation. —> Such individuals w i l l be expected to predominate, for example, i n a situation where climate i s seasonally severe and a few survivors of the unfavorable season each year recolonize under conditions of abundant supply of resources. On the other hand i n crowded environments with a relatively uniform and favorable climate, genotypes which can replace themselves at the lowest density of available resources (maximizing K, carrying capacity i n logistic terminology) w i l l be favored; these genotypes are termed K-strategists. Pianka (1972) has summarized the development of the concepts of r- and K~ selection. He states that r- strategies w i l l evolve i n populations which are usually maintained at low density relative to carrying capacity by a high rate of density-independent mortality,, In such a situation resources w i l l be abundant relative to the population's demand, and competition w i l l be.light, so each individual w i l l have a high t o t a l income of energy. A high proportion of this energy income w i l l be channeled into reproduction since alternative demands w i l l be low; there i s l i t t l e benefit to be derived from energy devoted to gaining competitive advantage over conspecifics, and the high, often catastrophic, mortality makes i t a better bet for an individual to l/ The logistic equation defines a sigmoid curve of population growth over time. Its differential form i s dN „ / K - N s where N = dt ~ r K K ; population size i n numbers, r = int r i n s i c rate of population increase, and K = maximum attainable value of N , or carrying capacity. The equation was f i r s t derived by Verhulst (1838). 48 maximize present reproduction than to invest energy i n growth, for example, i n order to gamble on reproductive benefits i n an uncertain future. Since the progeny w i l l also face light competition but potentially high density-independent mortality, the optimal strategy for the parent i s to minimize the amount of energy input per offspring so as to maximize the number of individuals produced. The net result i s that r- selected populations w i l l show rapid development, early maturity, high fecundity, and a tendency towards semelparity (a single reproductive period followed by death as i n Onehorhynchus); individual offspring w i l l be of "low quality" - small, without well-developed competitive or anit-predator mechanisms. Pianka theorizes that high population density, with consequent increased intensity of competition and reduced resource a v a i l a b i l i t y , produces K- selection. Individual energy income i s low, and a smaller proportion (relative to the r- selected case) i s available for repro-duction since more energy must be diverted to competitive ends. I f the offspring w i l l also encounter intense competition, more energy input per offspring i s favored, and numbers per parent must be reduced. The reproductive consequences of K- selection should include slower develop-ment, greater competitive a b i l i t y , lower resource requirements, delayed maturity, low fecundity, iteroparity (more than one reproductive attempt i n a lifetime), and high quality offspring. It i s important to emphasize that natural selection always favors those genotypes which maximize m, the birth rate minus the death 49 rate. The difference between r- strategists and K- strategists l i e s i n the response of this parameter to changes i n population density, or more precisely, a v a i l a b i l i t y of resources. Figure 15, modified slightly from Gadgil and Bossert (1970: Fig. l ) , i l l u s t r a t e s the point. When population density i s low and resources are abundant, the r- strategist w i l l have a higher m as a result of i t s higher birth rate, since mortality under these conditions i s presumably primarily density independent and affects r- strategists and K- strategists approximately equally. However, the high birth rate of _r- strategists i s achieved at the cost of "quality" of individuals; therefore, as population density rises and competition becomes more intense, the death rate of r- strategists rises rapidly, and m i s reduced accordingly. K- strategists have low birth rates as a consequence of dedicating a significant proportion of their energy budget to non-reproductive a c t i v i t i e s , which act i n one way or another to keep death rates relatively low even under conditions of high population density and low resource a v a i l a b i l i t y . Thus m for a K- strategist f a l l s relatively slowly with increasing population density. Above some density (marked A i n Figure 15) the output of surviving offspring i s higher for the K- strategist than for the r- strategist. If population density (or resource availability) i s consistently on the l e f t side of A, then r- strategists w i l l predominate i n the population; K- strategists w i l l be favored to the right of A. I f conditions fluctuate to either side of A, one might predict that the population would be polymorphic for reproductive strategy or that some intermediate strategy would evolve. A third alternative would be the evolution of the a b i l i t y to switch strategies according to conditions; this would seem particularly l i k e l y 50 Population Density : — > <= Resource Availability Figure 1 5 . The relationship between the instantaneous rate of population growth, m, and population density for r-strategists and K- strategists. See text for explanation. From Gadgil and Bossert ( 1 9 7 0 ) . 51 i f the fluctuation showed a regular periodicity. Pianka (1970) and Gadgil and Solbrig (1972) stress that ideal r-and K- strategies represent the poles of a continuum. The latter workers state that the position of a population on this continuum relative to other closely related populations can be quantified according to the proportion of to t a l energy income dedicated directly to reproduction. They suggest that one measure of this proportion i s the t o t a l weight of offspring produced relative to weight of the parent. Gadgil and Solbrig emphasize that increased reproductive output does not by i t s e l f constitute evidence for r- strategy since increased a v a i l a b i l i t y of resources to a population, occasioned perhaps by a sudden catastrophic mortality, results i n increased energy income for the survivors, which i n turn can cause increased reproductive output without any change i n the proportion of t o t a l energy used for reproduction; they term this the population dynamic effect. The literature on "r- and K- selection so far discussed represents one approach to the problem of the evolution of l i f e history characteristics. I find i t an int u i t i v e l y appealing introduction, but as so far stated, i t needs some qualification. It i s oversimple to think that a l l reproductive strategies can be placed somewhere along a one-dimensional continuum. This concept implies that l i f e history characteristics such as age at f i r s t maturity, growth rate, and fecundity are so highly inter-dependent that they w i l l always vary together i n a predictable way i n response to population density, resource a v a i l a b i l i t y , and mortality schedules. 52 In f a c t , the r e l a t i o n s h i p s are so complex that we should not be surprised to discover many populations d i s p l a y i n g mixtures of " r - s e l e c t e d " and "K-selected" t r a i t s . Tinkle and Wilbur (1973) review several instances of occurrence of such "mixed s t r a t e g i e s " - e.g. extremely high reproductive e f f o r t expended by populations whose adult density i s quite stable and presumably near car r y i n g capacity. Such departures from what would be expected on the ba s i s of r -and K- s e l e c t i o n arguments can best be understood i f we consider the basic objective of reproductive strategy. By d e f i n i t i o n n a t u r a l s e l e c t i o n favors those genotypes which make the greatest c o n t r i b u t i o n to the gene pool of succeeding generations; i t follows that the objec t i v e of reproductive strategy, as of any other evolving system, must be to maximize t h i s c o n t r i b u t i o n , or f i t n e s s . 3. + 1 Schaffer (1972) has shown that the quantity to be maximized i s : b. + p.• where b^ i s the number of progeny produced by a female at age i which survive to reproduce, and p_. i s the p r o b a b i l i t y that the female w i l l v i + 1 survive from age i to age i + 1, and — — — — i s F i s h e r ' s (1930) ^0 reproductive value f o r a female age i + 1. This l a s t quantity i s a device f o r expressing a l l expected future b i r t h s to an i n d i v i d u a l i n terms of. t h e i r equivalent value i n present b i r t h s ; the concept i s analogous to that of present discounted value of future income i n economics. Schaffer's formulation i s deceptively simple; each term can be subdivided i n t o several components, many of which are i n t e r r e l a t e d . 53 The term b^ i s the absolute fecundity of an adult female at age i m u l t i p l i e d by the proportion of those eggs which survive t o become reproducing adults. Fecundity i s an increasing f u n c t i o n of reproductive e f f o r t , the proportion of the organism's t o t a l energy income devoted to reproduction. Fecundity f o r a given reproductive e f f o r t i s also l i k e l y to be an i n c r e a s i n g f u n c t i o n of parental si z e i n organisms l i k e fish", which continue to grow a f t e r reaching sexual maturity. Thus, fecundity at age i can be a decreasing f u n c t i o n of reproductive e f f o r t i n previous seasons to the extent that growth has been reduced. Pre-reproductive s u r v i v a l of o f f s p r i n g i s probably often a f u n c t i o n of f a c t o r s beyond the c o n t r o l of the adult (e.g. population density, resource a v a i l a b i l i t y , p h y s i c a l environmental f a c t o r s ) ; however, s u r v i v a l of young stages may i n some cases be a function of egg or l a r v a l s i z e , which i n turn i s negatively r e l a t e d to fecundity at any given l e v e l of reproductive e f f o r t . The term i s the p r o b a b i l i t y of adult s u r v i v a l from the present reproductive season to the next. I t i s l i k e l y to be a decreasing f u n c t i o n of present reproductive e f f o r t . To the extent that adult s u r v i v a l i s r e l a t e d to s i z e , i t may also be a function of reproductive e f f o r t i n previous seasons. A d d i t i o n a l l y , i t i s l i k e l y t o be a g e - s p e c i f i c . The f i n a l term i s Fisher's reproductive value. For an i n d i v i d u a l adult t h i s i s the expected production of surviving o f f s p r i n g i n a l l future reproductive seasons, reduced by the p r o b a b i l i t y of the adult's surviving to each season. A l l of t h i s i s f u r t h e r discounted according to the 54 population growth r a t e . I f the population i s i n c r e a s i n g , then one surviving o f f s p r i n g i n the present represents a higher proportion of the population than one produced i n the future and i s e v o l u t i o n a r i l y worth more; t h i s i s s i m i l a r to the standard economic model wherein money which can begin earning i n t e r e s t now i s worth more than, the same amount to be received i n the fu t u r e . However, i f the population i s d e c l i n i n g , i t i s as i f there were a negative i n t e r e s t rate, and future reproduction i s worth more than present. In summary, a reproductive strategy i s a combination of l i f e h i s t o r y c h a r a c t e r i s t i c s manifested by an organism which may be thought of as representing a set of t a c t i c a l choices by the organism with respect to several i n t e r r e l a t e d v a r i a b l e s . One group of these choices has to do with the magnitude and d i s t r i b u t i o n of reproductive e f f o r t of the i n d i v i d u a l over i t s l i f e t i m e . V a r i a b l e s included i n t h i s group are ages of f i r s t maturity and onset of senescence, number of reproductive attempts i n the l i f e t i m e (the choice of semelparity or i t e r o p a r i t y i s included here), and energy t r a d e o f f s among reproduction and other energy-consuming functions. The magnitude of the energy input to each i n d i v i d u a l egg or l a r v a i n a p a r t i c u l a r reproductive attempt represents a separate category of choice; f o r a given t o t a l energy input the organism must choose one of the i n f i n i t e number of combinations of absolute fecundity and energy content of i n d i v i d u a l eggs or larvae; i . e . the same t o t a l energy can be d i s t r i b u t e d among many small young or a few large ones. D i f f e r e n t combinations of choices (that i s , d i f f e r e n t strategies) w i l l r e s u l t i n d i f f e r e n t numbers of young which themselves survive to breed. Strategies which optimize the sum of present output plus the present value of future output w i l l be favored by n a t u r a l s e l e c t i o n i f they are h e r i t a b l e . 55 Species of the Family Poeciliidae have several characteristic l i f e history features which must be considered i n discussing their reproductive strategies.. The shallow water areas they inhabit are often quite unstable over time, and population levels may fluctuate considerably. Generation time i s typically short, varying from several weeks to a few months. Each adult female may produce several broods over a relatively long reproductive season. These features combine to make i t very l i k e l y • that successive generations, or even the same individual at different times, w i l l face quite different sources and levels of mortality. Under such conditions the a b i l i t y to make rapid adjustments of reproductive strategy to immediate and anticipated conditions should be of great selective value. Aside from the purely descriptive information on guppy reproduction, the most interesting fact which emerges from my data i s that under conditions of poorer nutrition and presumably more intense competition, the smaller females redistributed the same reproductive effort among fewer, larger fry (the same size as fry of larger females), while the larger females showed no change i n strategy. This result prompts several questions for which I have hypothetical answers, most of which are testable. What i s the advantage to the smaller females of increased size of fry? When competition among fry i s intense, larger fry may have a large enough survival advantage that, i n spite of the reduction i n 56 absolute fecundity, the fitness of the adult i s increased by putting more energy into each individual offspring 0 This i s -essentially the same as Svardson's (1949) fourth hypothesis (see page 2 3 ) . Magnuson (1962) found i n a laboratory study that when competition for food was intense, larger juvenile medaka (Oryzias latipes) successfully dominated smaller ones and showed significantly, faster growth. Bagenal (1969b) i n a f i e l d experiment showed that brown trout (Salmo trutta) fry hatching from larger eggs had a higher survival rate than those from smaller eggs li v i n g i n the same enclosures; he suggested that the mechanism producing this effect might have been that the larger fry were more effective at maintaining t e r r i t o r i e s . I f the explanation of the f i r s t question i s correct, then why did only the smaller females change strategy? One hypothesis i s that reduction i n rations acted as the signal for the change i n strategy and that, as a result of contest-type competition, only the smaller females actually suffered reduced rations i n the experiment. The d i f f i c u l t y with this hypothesis i s that i f the change of strategy i s adaptive, and i t s effects are related to differential survival of fry of different sizes, then i t i s only the smaller females which are receiving the messag about what conditions w i l l be l i k e for their fry. Since fry of both larg and small females presumably face a similar competitive situation, the fitness of the larger females should be considerably reduced by their i n a b i l i t y to read the environment and respond with a corresponding increase i n size of fry. It i s possible that the difference i n response of large as opposed to small females i s an artifact of the experimental 57 situation, i n which food was localized i n time and space, and where the population was spatially restricted i n an environment offering l i t t l e visual isolation among competitors: Magnuson (1962) found that both of these conditions increased the intensity of contest competition. Possibly under natural conditions competition i s closer to the scramble type, and a l l sizes of adults are sufficiently affected by deteriorating conditions to receive the message and make the appropriate switch of strategy. Why did reproductive effort not decline when the nutritional regime worsened? A simple hypothesis i s that the level of reproductive effort i s genetically fixed and cannot vary much except on an evolutionary time scale. Alternatively, i t seems l i k e l y that when adult mortality i s high and/or growth i s slow, as i n Experiment 3» there i s l i t t l e benefit to be gained by postponing reproductive effort u n t i l such time as . probability of survival of fry might increase. One of the major deficiencies of the present study i s that I have only scanty information on the ecology of my experimental animals under natural conditions, where the characteristics under investigation evolved and presumably have adaptive value. I do have two pieces of information from Ben Seghers (personal communication) which seem relevant. F i r s t , although the Tacarigua River i s relatively stable i n comparison with other rivers of the area, i t does undergo regular fluctuations i n flow and related variables as a result of marked seasonal variations i n r a i n f a l l . Second, predation on adult guppies by cichlids i s probably relatively 1 58 intense. In an attempt to make up f o r the lack of f i e l d data on my own organisms, I w i l l review i n some d e t a i l f i e l d i n v e s t i g a t i o n s on s i m i l a r problems i n r e l a t e d species. F i e l d Studies on P e o c i l i i d Reproduction J.D. McPhail (unpublished data) has made preliminary f i e l d and laboratory studies of reproduction of another small P e o c i l i i d , Neoheterandria t r i d e n t i g e r , i n Panama; h i s r e s u l t s show some s t r i k i n g p a r a l l e l s with mine. He worked p r i m a r i l y with populations i n h a b i t i n g small f o r e s t streams which d r i e d up almost completely during the dry . season. High wet season populations are reduced dramatically as the streams dry up; a very small proportion of these populations survives the dry period i n i s o l a t e d pools i n deeper parts of the stream bed. When the ra i n s come again, f l a s h f l o o d i n g causes a sudden expansion of a v a i l a b l e habitat, which the survivors r e c o l o n i z e . I t seems probable that the onset of the wet season also brings about an increase i n av a i l a b l e food, which i n these streams i s mostly of t e r r e s t r i a l o r i g i n . The changes i n reproductive pattern through the year suggest that Neoheterandria populations switch strategy according to the season, although McPhail's data do not show d i r e c t l y that si n g l e i n d i v i d u a l s use d i f f e r e n t s t r a t e g i e s at d i f f e r e n t times. Late i n the dry season females i n the i s o l a t e d streambed pools suspend reproduction e n t i r e l y . E a r l y i n the. wet season females from the expanding populations produce r e l a t i v e l y large numbers of small f r y at approximately 10 day i n t e r v a l s ( t h i s species shows superfoetation, so the gestation period f o r an 59 i n d i v i d u a l brood i s about 20 days during t h i s season). Females taken from the dense populations of the l a t e wet season and ea r l y dry season produce much smaller broods of large young at 15 day i n t e r v a l s . I t looks as though these populations are using r - strategy when population density i s low and K- strategy when the a v a i l a b l e habitat i s r e l a t i v e l y f u l l . McPhail also examined females from l a r g e r permanent streams i n the same area. In most c o l l e c t i o n s females tended towards the low fecundity/large f r y end of the spectrum, as one would expect i n a stable, benign environment. However, i n some l o c a l i t i e s the females showed the r - selected pattern (high fecundity and small f r y ) ; these populations a l l l i v e d i n l o c a l i t i e s where f i s h predators were abundant and adult m o r t a l i t y was probably high. Krumholz (1948) studied the fecundity of populations of the P o e c i l i i d Gambusia a f f i n i s i n northern I l l i n o i s and southern Michigan, where they had been introduced f o r mosquito c o n t r o l . These populations, at the northern l i m i t of the species' range, suffered high winter m o r t a l i t y followed by rapid increase to high d e n s i t i e s during the spring and summer months. During one reproductive season Krumholz made monthly c o l l e c t i o n s of Gambusia from established populations i n two ponds, one of which appeared to have low and the other moderate b i o l o g i c a l p r o d u c t i v i t y . E a r l y i n the following spring he introduced mature v i r g i n females and males to a very productive stock-watering pond; he made c o l l e c t i o n s from t h i s population throughout the summer. From the preserved c o l l e c t i o n s Krumholz measured length frequencies, reproductive state of both sexes, 6 0 and numbers of embryos c a r r i e d by gravid females. Unfortunately he made no measurements of size of young. Females of two d i f f e r e n t l i f e h i s t o r y types occurred i n a l l of Krumholz' populations. Females born i n mid to l a t e summer grew t o moderate size and overwintered as immatures; these females began to reproduce i n ea r l y spring and produced three to f i v e broods before dying, sometimes a f t e r a period of s t e r i l e s e n i l i t y i n t h e i r second summer or autumn. The spring progeny of the over-wintered females showed ra p i d growth and matured at a much smaller siz e than d i d t h e i r mothers; they produced from one t o three broods during l a t e spring and summer and then died. Thus within the same population, females born i n spring d i s t r i b u t e d t h e i r reproductive e f f o r t according to r - strategy, and females born i n summer behaved l i k e K- s t r a t e g i s t s ; choice of strategy depended not on parentage but on season of b i r t h . In both l i f e h i s t o r y types i n a l l ponds fecundity at a given length decreased with each successive brood from a maximum at the f i r s t or second brood; another way of st a t i n g the same th i n g i s that average fecundity f o r length showed a more or l e s s steady decrease wi t h i n a given pond as the season progressed. Comparisons among the three ponds showed that f i s h i n the newly stocked, productive pond matured at a smaller size and produced more broods i n the course of the season r e l a t i v e to the other ponds. For comparable l i f e h i s t o r y types and times of year, average fecundity at a given length increased with ••increasing pond p r o d u c t i v i t y . 61 Krumholz attributed the decrease i n fecundity over time to physiological changes associated with senescence i n the females. However, his data do not rule out the possibility that decreasing fecundity (which may have been accompanied by an increase i n size of individual fry) represented a switch i n reproductive strategy i n response to increasing population density and competition as the season advanced.. It i s suggestive that the drop i n fecundity with time was most pronounced i n the least f e r t i l e pond and that fecundity i n the newly stocked, productive pond continued at a high level throughout the summer. Krumholz (1963) reported the results of a study on Gambusia manni in two ponds i n the Bahamas. The techniques were similar to those described for the previous study. In Daniels' Pond there were no aquatic predators; the pond at Calloo Mangrove was connected to the sea at high tide and contained several predatory marine species, including juvenile barracuda. Although he made no direct measurements of population densities, Krumholz f e l t that density was much higher i n Daniel's Pond. The Daniels' Pond population contained a high proportion of large senile individuals, particularly i n late summer and autumn at the end of the reproductive season; the maximum size observed at Calloo Mangrove was much smaller, and there was no evidence of sen i l i t y . Krumholz reasoned that predators eliminated senescent Gambusia at Calloo Mangrove. Krumholz concluded that growth was faster i n the Daniel's Pond population due to presumed higher biological productivity of the pond. As indirect evidence for this hypothesis, he cited his observations that 62 both maximum size and size at f i r s t maturity were higher i n Daniels' Pond. The plots of length frequencies for successive collections at Daniels' Pond show a clear advance of modal length frequencies from month to month, from which growth of individual cohorts can be estimated. The modal length frequencies for the Galloo Mangrove population are much less clear within any collection, and they vary erratically with time; possibly this i s due to rapid changes i n the age and size composition of the population due to high predation, but very rapid growth could have the same effect. Since the difference i n predation mortality probably also.accounts for the observed differences i n maximum length between ponds, Krumholz' contention about differences in.growth rates rests on the.fact that the Calloo Mangrove f i s h mature at a smaller size. It has been shown for many species of fi s h that early maturity i s often associated with fast growth; Svardson (1943) showed that fast-growing male guppies matured at an earlier age and smaller size than slower growers. In sum, on the basis of the limited evidence, i t seems more l i k e l y that growth i s more rapid at Calloo Mangrove. The two populations showed marked differences i n reproductive strategy. Calloo Mangrove, females had higher average fecundity throughout the year i n spite of being considerably smaller on the average than Daniels* Pond f i s h . In addition, peak reproduction, as measured by average fecundity and proportion of females which were gravid, started earlier i n the year and continued longer at Calloo Mangrove. 63 In Figure 16 I have plotted fecundity against size of female for the two populations at three different times, which are representative of collections made during the early, middle, and late parts of the reproductive season. In a l l Calloo Mangrove collections fecundity increased exponentially with maternal length; the regression of fecundity on length for Daniels' Pond has zero slope early and late i n the season, but-at the peak of the reproductive period the slope and height of the regression were quite similar to Calloo Mangrove. My explanation of the data i s that the Calloo Mangrove population follows r- strategy at a l l seasons due to high predation pressure, which keeps population density low, resource availability high, and competition low. The explanation of the Daniels' Pond data i s more complicated and involves some speculation. I suggest that i n January and July the population was resource-limited and was following K- strategy; the January population density was probably lower but so was biological productivity i n the moderately seasonal Bahamas. In late January and early February a period of extremely low temperature caused heavy mortality i n Daniels* Pond. The reduced population density, which coincided with seasonally increasing productivity of the pond, must have reduced the intensity of competition, and the population appears to have switched towards r_-strategy for several months (February through May). In the absence of data one can only speculate, that the increased fecundity i n A p r i l was accompanied by a reduction i n average size of fry and that i n the early and late collections average size of fry was greater. 64 Daniels' Pond 10 19 Jan. o o »o °9 °eo J I I J I Gal loo Mangrove 23 Jan. 1 ' I 20 25 May « 15 > E a) 0) XI E I 5 J ( L 24 May o • o 00 J I I I I 15 10 0. 19 July ° 0 •°° 0 ° « J L 18 Ju ly o 0 ° , J L 20 25 30 35 4 0 45 20 25 30 35 4 0 Total length of female (mm,) 45 Figure 16. Average number of embryos carried by female Gambusia manni collected at two locations i n the Bahamas at different seasons. Each point i s an average of several individuals. Data from Krumholz ( 1963: Tables 15 and 1 7 ) . 65 Other Laboratory Studies of Fecundity and N u t r i t i o n Hester (1964) i n v e s t i g a t e d the e f f e c t s of reduced r a t i o n s on fecundity of a domestic, s t r a i n of P o e c i l i a reticulata,. His r e s u l t s i n d i c a t e d that low r a t i o n s during one brood period s i g n i f i c a n t l y reduced the numbers of f r y i n that brood and the one f o l l o w i n g . He observed no trend i n si z e of f r y with feeding l e v e l or maternal s i z e , and there were no changes i n length of the gestation period, The stock of guppies used by Hester was somewhat l e s s fecund f o r length than my stock, and the f r y he measured averaged much smaller' than those i n my study. There are several discrepancies between Hester's r e s u l t s and those of the present study. Hester found that reduced r a t i o n s a f f e c t e d large and small females s i m i l a r l y ; t h i s i s probably due to the f a c t that h i s experimental f i s h were i s o l a t e d i n i n d i v i d u a l aquaria f o r the e n t i r e experiment. Thus there was no competition f o r food i n Hester's experiments, whereas i n my stock tanks l a r g e r i n d i v i d u a l s may have been able t o garner a disproportionate share of food at the expense of the smaller f i s h . In the l i g h t of my r e s u l t s i t i s s u r p r i s i n g that Hester reported no r e l a t i o n between size of f r y and e i t h e r r a t i o n or maternal s i z e ; however, i t appears that he measured only 79 f r y and that they were much more v a r i a b l e i n length than those of my study. I t i s p o s s i b l e that with a l a r g e r sample he might have found a r e l a t i o n s h i p between f r y si z e and e i t h e r r a t i o n or maternal s i z e . With regard to gestation period I have no s p e c i f i c explanation of the constancy reported by Hester as opposed to the lengthening of the period under poor conditions which I observed. In general i t should be emphasized that the two s t r a i n s of P e o c i l i a compared here have probably been under quite d i f f e r e n t s e l e c t i v e regimes 66 for some time, and i t would be unrealistic to expect them to respond i n the same way even to identical situations. Bagenal (1969a) reported on laboratory experiments on the relationship between feeding level and fecundity i n Salmo trutta. In the f i r s t experiment f i s h were kept i n two large tanks and fed on commercial dry-food - one tank at f u l l rations and the other at half that level. At the end of eight months fecundities of the two groups of ripening f i s h . were compared. Fish from the high food treatment had grown faster during the experiment, but when fecundities were adjusted for length there was no difference between the treatments. The slope of the regression of fecundity on length was significantly steeper i n the low food group; the larger low food individuals were about as fecund as individuals of the same size i n the high food treatment, but the smaller low food f i s h had very few eggs for their size. This i s essentially the same situation that I observed i n comparing my Experiments 1 and 3« Bagenal hypothesized that i n the low food treatment the large individuals were securing a disproportionate share of the food; he set up a second set of experiments to explore this idea. In his second experiments Bagenal set up four separate treatment groups of f i s h ; two of these were fed at f u l l and the other two at one-third rations. After five months of this regime growth depensation was evident i n a l l tanks; to reduce the size range i n each tank, the f i s h i n the two tanks within each food level were sorted by size and reassigned to tanks so that each food level after the sorting consisted of one tank of larger f i s h and one of smaller ones. After four more months 67 the f i s h from the high food tanks were more fecund for their length than the low food f i s h , but this time there was no difference between treatments i n the slope of fecundity on length. In both sets of experiments Bagenal found that well-fed f i s h produced smaller eggs by dry weight than the food-deprived ones. Scott (1962) i n a similar study on Salmo gairdneri found no difference i n egg size by wet weight between food levels. Bagenal (1969a) pointed out that the water content of eggs increases as they mature, and that there-fore the appropriate measure of egg size for comparative purposes i s dry weight. Scott's well-fed f i s h were i n a more advanced state of maturity than the starved ones; consequently, the similarity i n wet weight between treatments suggests that the low-food f i s h had larger eggs i n terms of dry weight. Studies on Exploited Marine Populations Several f i e l d studies have documented year-to—year changes i n fecundity within populations: Bagenal (1963b, 1965) for three species of f l a t f i s h i n the Clyde Sea, Hodder (1963) for haddock (Melanogrammus aeglefinus) of the Grand Banks, and Raitt (1968) for Norway pout (Trisopterus esmarkii) i n the North Sea. In a l l the above cases fecundity increases were apparently associated with either lowered population density or increases i n available food per individual. In the absence of data on egg size or some other measure of quality i t seems most probable that these cases represent Gadgil and Solbrig's (1972) population dynamic effect rather than any change i n reproductive 68 strategy ( i . e . change i n magnitude or distribution of reproductive effort). Bagenal (1966) gives a comprehensive review of his own and others' f i e l d studies on the fecundity of plaice (Pleuronectes platessa) throughout i t s range i n the eastern North Atlantic and adjacent seas. Populations i n different l o c a l i t i e s vary considerably i n age composition and growth rate; since fecundity of plaice increases exponentially with length, different populations are compared i n terms of the expected fecundity of a 37 cm. f i s h (F^y) t as calculated from the equation for the regression of fecundity on length for*a given area. F^y i s lowest (84,000 eggs) i n the Southern Bight of the North Sea; fecundity increases f a i r l y evenly with distance i n a l l directions from thi s area. Thus F^ i s high i n the Bay of Biscay ( 1 8 4 , 0 0 0 ) , the west coast of Ireland ( 1 5 3 , 0 0 0 ) , and northern Norway ( 1 5 6 , 0 0 0 ) . In two l o c a l i t i e s s t i l l farther away, Iceland and the Barents' Sea, F^ drops again (106,000 and 107,000 respectively). Fish from inside Trondheim Fjord i n Norway and from the Baltic are much more fecund than those from other l o c a l i t i e s and do not f i t the general pattern described above; i n both cases, however, there i s evidence from marking experiments and meristic studies to suggest that the populations involved are r a c i a l l y distinct from the other populations considered. Bagenal (1966) postulates that over most of the area of plaice distribution, with the exception of Trondheim Fjord and the Baltic, there exists a single basic fecundity type and that differences i n fecundity 69 are caused by local differences i n food supply. In support of this hypothesis he notes that populations with high fecundity also tend to have the following characteristics: 1) low population density, 2) fast growth, and 3) high gutted weight at a given lengths It i s consistent with the above analysis to say that most populations of plaice follow the same reproductive strategy and that variations i n fecundity are due to the population dynamic effect. In addition to his basic point about food supply, Bagenal (1966) notes that fecundity i s higher i n areas where the prevailing currents are l i k e l y to carry the planktonic plaice eggs and larvae far from suitable nursery grounds, and conversely, that low fecundity i s associated with favorable currents. Bagenal points out that the survival value of high fecundity i s obvious where mortality of young stages i s l i k e l y to be high; however, the selective mechanism favoring lower fecundity under apparently better conditions i s not immediately clear, since even under these circumstances more fecund individuals should on average make a larger contribution to the next generation. This i s the same problem which Svardson confronted i n 1949» Bagenal considers Svardson's (1949) hypothesis that reduced fecundity can result from selection for larger eggs, which produce larvae, of greater competitive quality. Unfortunately, Bagenal's plaice samples were taken at different times of year i n different l o c a l i t i e s , and there-fore f i s h from different areas were at different stages of maturity. 7 0 Because of t h i s , d i r e c t comparison of egg s i z e s among populations was hot p o s s i b l e . He approached the question i n d i r e c t l y by comparing two populations on the b a s i s of density of oven-dried eggs, which does not vary much within a population over a considerable range of developmental stages. He found that eggs from the more fecund population were also more dense; since a l l p l a i c e eggs are p o s i t i v e l y buoyant, he reasoned that the denser eggs contained a higher proportion of f a t . Bagenal considered that t o be evidence against Svardson's hypothesis; however, i t seems t o me that Bagenal's point i s i r r e l e v a n t to the question of absolute si z e or energy content of an i n d i v i d u a l egg and that Svardson's hypothesis i s s t i l l v i a b l e i n t h i s case. There i s another p o s s i b l e explanation of the v a r i a t i o n i n p l a i c e fecundity reported by Bagenal. As Murphy (1968) and Schaffer (1972) have noted, i f l a r v a l m o r t a l i t y v a r i e s appreciably from year to year, the f i t n e s s of an adult i s increased i f reproductive e f f o r t can be spread over more years, even at the cost of reduced fecundity i n a given year. Data on annual v a r i a t i o n i n l a r v a l s u r v i v a l are not a v a i l a b l e f o r most of the p l a i c e populations studied; however, i t i s p o s s i b l e t o make a crude estimate of reproductive span i n d i f f e r e n t populations from the age composition of d i f f e r e n t stocks. In Table I I are assembled estimates of reproductive l i f e s p a n f o r most of the l o c a l i t i e s studied by Bagenal (1966), together with t h e i r respective F^ from h i s Table 2. There i s considerable uncertainty about the values f o r reproductive span, due t o small sample s i z e s i n many cases as w e l l as to the f a c t that d i f f e r e n t populations are subject to d i f f e r e n t l e v e l s of f i s h i n g m o r t a l i t y , which Table I I . The calculated fecundity of female p l a i c e 37 cm. i n length (F^y) and estimates of reproductive l i f e span at d i f f e r e n t l o c a l i t i e s . Data on F^r, from Bagenal (1966). Other data from sources i n d i c a t e d . F L o c a l i t y 37 Age Reproductive Sample Source (thousands) range span (years) siz e of data Barents Sea 107 V-XXII 18 7 Wimpenny (1953) S. Bight, N. Sea 84 I-XVII .17 223 Simpson (1959) Flamborough Gd., N. Sea 96 II-XV 14 33 Simpson (1959) Clyde Sea 137 III-X 8 55 Bagenal (1958) Clyde Sea 159 II-XI 10 156 Bagenal (1963c) Rye Bay 127 n - x r a 16 84 Bagenal (1960a) Plymouth 137 III-X 8 60 Bagenal (1960a) Wexford, S. .Ireland 139 II-IV • 3 30 Bagenal (1960b) S c h u l l , S. Ireland 150 II-IV 3 9 Bagenal (1960b) Dingle, W. Ireland 153 II-VTII 7 47 Bagenal (1960b) Galway, W. Ireland 146 II-V 4 20 Bagenal (1960b) Kellybegs, W. Ireland 132 II-VII 6 31 Bagenal (1960b) Table I I . The calculated fecundity of female p l a i c e 37 cm. i n length (^7) estimates of reproductive l i f e span at d i f f e r e n t l o c a l i t i e s . Data on F^r, from Bagenal (1966). Other data from sources ind i c a t e d , (continued) L o c a l i t y F37 Age Reproductive Sample Source (thousands) range span (years) size of data Arendal, S. Norway H I III-XI 9 12 Bagenal (1962) Bergen, W. Norway 134 IV-XV 12 24 Bagenal (1962) Trondheim Coast 150 IV-IX 6 11 Bagenal (1962) Troms