UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Analysis of mortality in a population of tadpoles of the red-legged frog (Rana aurora) Calef, George Waller 1971

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


831-UBC_1971_A1 C34.pdf [ 4.56MB ]
JSON: 831-1.0101377.json
JSON-LD: 831-1.0101377-ld.json
RDF/XML (Pretty): 831-1.0101377-rdf.xml
RDF/JSON: 831-1.0101377-rdf.json
Turtle: 831-1.0101377-turtle.txt
N-Triples: 831-1.0101377-rdf-ntriples.txt
Original Record: 831-1.0101377-source.json
Full Text

Full Text

AN ANALYSIS OF M O R T A L I T Y IN A POPULATION OF T A D P O L E S OF T H E RED - L E G G E D F R O G (Rana aurora) by G E O R G E W A L L E R C A L E F B . S c . University of Chicago (1964) A THESIS SUBMITTED IN P A R T I A L F U L F I L M E N T OF T H E REQUIREMENTS FOR T H E D E G R E E OF DOCTOR OF PHILOSOPHY in the Department of Zoology We accept this thesis as conforming to the required standard. T H E UNIVERSITY OF BRITISH COLUMBIA March 1971 In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r an a d v a n c e d d e g r e e a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e a n d s t u d y . I f u r t h e r a g r e e t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by t h e H e a d o f my D e p a r t m e n t o r by h i s r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l n o t be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . D e p a r t m e n t o f The U n i v e r s i t y o f B r i t i s h C o l u m b i a V a n c o u v e r 8, C a n a d a D a t e McuscM- SO (9?j T A B S T R A C T The survivorship, distribution, growth rates, and natural predation rates on the population of Rana aurora tadpoles in Marion Lake, British Columbia, were studied during the summers of 1969 and 1970. Natural survivorship appeared to have two phases: a rapid decline in numbers during the first four weeks after hatching was followed by a less rapid decline, until approximately 5% of the population remained at metamorphosis, after 11-14 weeks. The size attained by tadpoles living at any depth in the lake was strongly correlated with the accumulated degree-days since hatching (r = .93). The regression equation (degree-days vs. size) obtained from data on animals in the lake could accurately predict the growth rate of tadpoles kept in the laboratory at maximum food supply. It was concluded that animals in the lake grew at the maximum rate permitted by the temperature. Alterations of the density of tadpoles and predatory salamanders were performed in enclosures in the lake. The percentage of tadpoles surviving in the enclosures containing no salamanders was twice as high as it "was in the natural population. The percentage of tadpoles surviving in the experimental enclosures was independent of density, at least up to 75 tadpoles/m . Except at the lowest density (5/m^), no significant effects were noted at any density, on the rate of growth or time to metamorphosis. These experimental observations support the notion that food is not limiting to the tadpoles at their normal densities. Most of the mortality of tadpoles observed in the natural population was attributed to predation because: 1. Tadpoles kept in the laboratory on starvation diets survived for many weeks, even though they did not grow. 2. Tadpoles living at densities up to 100 times normal could survive and grow to metamorphosis in the absence of predators. 3. Predation rates observed in the field, and simulated on the basis of laboratory experiments, were sufficient to account for much of the mortality observed in the natural population. Laboratory studies of salamanders preying on tadpoles in multi-prey systems showed that the number of tadpoles eaten was proportional to the density of tadpoles present except at very low density, where fewer tadpoles than expected were eaten. The tadpoles also had a refuge from some predators after they had grown to a certain size. Salamanders were observed to move into areas of high tadpole density, thereby increasing locally the intensity of predation. It was concluded from the laboratory and field observations that mortality of tadpoles is density-dependent. However, the limiting factors for the Rana aurora population are unknown, since the frogs were not studied after they left the lake. The numbers of eggs laid in the two years of the study were very similar, but no social mechanisms for controlling the breeding population were observed. T A B L E OF CONTENTS A B S T R A C T o LIST OF T A B L E S LIST OF FIGURES A C K N O W L E D G E M E N T S INTRODUCTION MATERIALS AND METHODS The study area Estimation of numbers of eggs Estimation of numbers of tadpoles Enclosure experiments Measurement of growth rates Laboratory experiments with predators RESULTS Life history of the natural population Survivorship in the natural population Survivorship in the enclosure experiment Survivorship of starved tadpoles Growth rate in the natural population Growth rate in the experimental enclosures Tadpole predators Estimation of predation on tadpoles in the lake Laboratory experiments with predation Page DISCUSSION 38 Comparison with results from other studies 38 Functional response of predators 43 Numerical response of predators 45 Interactions of growth rate and predation 46 Population regulation in Rana aurora 46 Effect of tadpoles as benthic grazers 48 S U M M A R Y 51 L I T E R A T U R E CITED 53 APPENDIX 1 56 APPENDIX 2 67 LIST OF T A B L E S Table 1. Table 2a. Table 2b. Table 3. Table 4. List of treatments in the enclosure experiment. Dispersal of tadpoles in 1969. Dispersal of tadpoles in 1970. Observed vs. predicted growth rates in the laboratory. Results of eight-day predation experiment. Page 8 16 16 27 35 i i i LIST OF FIGURES Page Frontispiece. Adult female of Rana aurora. Figure 1. Maps of Marion Lake. 4 Figure 2. Survivorship in the natural population. 18 Figure 3. Survivorship in the experimental 19 enclosures. Figure 4. A regression of Tog weight against log 22 length. Figure 5. Growth rate of tadpoles in Marion Lake 23 during three summers. Figure 6. Growth rate of tadpoles from different 24 depths and different areas of the lake in 1970. Figure 7. Size vs. accumulated degree-days. 25 Figure 8. The effect of population density on 28 growth of tadpoles in the experimental enclosures. Figure 9. The relationship between density of tadpoles 33 and the number eaten by Taricha, in a multi-prey system. Figure 10. Grazing rate of the tadpole population 49 under three mortality schedules. iv A C K N O W L E D G E M E N T S A n ecologist who believes that his science is actually " t h e study of the inter-relationships among living things" must welcome t h e opportunity of participating in an integrated study of a biological community. I thank Ian Efford, my supervisor, for providing me with t h e opportunity to do my research as part of the Marion Lake Project of the Canadian International Biological Program. Throughout my study D r . Efford has provided facilities, advice, and encouragement in both a professional and personal capacity, for which I am grateful. Many people, both students and faculty, have provided valuable ideas and dialogue during my studies at the University of British Columbia. I remember them fondly. I would especially like to mention Dennis Chitty, Barry Hargrave, and Iain Neish. D . H . Chitty, C . V . Finnegan, C S . Holling, and D . J . Randall have read the manuscript and made valuable suggestions concerning the writing and the presentation of data. I thank Brian Henderson for assistance with the field work in 1969, and Conrad Wehrhahn and Neil Gilbert for advice on statistical techniques. Victor McCauley kindly provided temperature data for the area of the lake in which he was working. Finally, I thank Kim Hyatt, Annice Thompson, and Susan Robinson for making the autumn of 1970, which could have been one of the worst times of my life, one of the best. v 1 INTRODUCTION Although the tadpoles of frogs and toads are among the most common inhabitants of lakes, ponds, and streams, little is known about their population dynamics or their ecological role in these communities. Turner (1962) reviewed the literature and found that "the survival rates of larval anurans, whether development is direct or indirect, are virtually unknown." Or again, "the estimation of survival of larval anurans remains one of the most difficult problems in anuran demography, and the discovery of a reliable technique would be an important contribution. " Herreid and Kinney (1966) studied the survivorship of Rana sylvatica larvae after they discovered a reliable method of marking the tadpoles. They observed a mean mortality rate of 96% in four populations. Brockleman (1968), who studied the growth and survivorship of a population of Bufo americanus tadpoles, found a mortality rate similar to that found by Herreid and Kinney. By a series of experiments in screen enclosures, with altered densities of tadpoles, predators, and food supply, Brockleman concluded that mortality of Bufo tadpoles was density-dependent, and mediated by the supply of food. Other studies have dealt with growth rates, food habits, tadpole predators, and other aspects of the natural history of various anuran species, but no attempt has been made to integrate these approaches for a single species. (For example, see Gosner and Black, 1955--growth rates; Jenssen, 1968--food habits; Anderson, 1968--tadpole predators; and Savage, 1952--general natural history.) My study (as part of the Canadian International Biological Program study of Marion Lake) was intended to: 2 1. Estimate the size and distribution of the population of Rana aurora tadpoles, and to measure the growth rate of tadpoles in different areas of the lake; 2. Identify and quantify the sources of mortality; 3. Explore the controls which might operate in main-taining the population size. Since Brockleman's study stressed the importance of food supply and its density-dependent effect on growth rates, while others have emphasized the potential effect of predation (e.g. Anderson, 1968), these were the two ecological factors on which I concentrated. 3 MATERIALS AND METHODS The study area: Marion Lake is a small (10 hectares), oligotrophic lake situated at an elevation of about 300m on the edge of the Coastal Mountains in southern B r i t i s h Columbia, (see Efford, 1967, for a detailed description). Although the lake fluctuates considerably during the year, owing to seasonally heavy rains, it is relatively stable in summer, and has approximately the depth profile shown in F i g . IA. The lake bottom is a soft, flocculent mud, with patches of macrophytes growing in certain areas. Seven sub-habitats were defined within the lake, in terms of type of vegetation and depth of water (Fig. IB). The areas termed "submerged vegetation" supported some of the following species: Chara  globular is (Thuill), Isoetes occidentalis (Henders.), and Potamogeton  epihydris (Raf.). Since Potamogeton natans (L.) and Equis etum sp. sometimes occur together in the lake, areas containing these two species were considered as one sub-habitat. The l i l y pad, Nuphar polysepala (Englem.), was common along the shore in clumps on the open mud, and often grew inassociation with P. natans. Estimation of numbers of eggs: I determined the numbers and distribution of egg masses by systematically searching the entire 2 lake. To aid in this search, I divided the lake bottom into 1000m quadrats by placing an anchored buoy at the corners of each quadrat (Fig. IA). The shallow areas (less than l m in depth) were searched from a boat and each egg mass was marked with a stake. The deeper areas were searched by diving, and a systematic pattern was assured by following a boat which was rowed back and forth through the quadrat. The clarity of the water and the similarity of repeated counts in selected areas suggest that few clutches were missed. F i g . 1 Maps of Marion Lake (after Neish, 1970). A . The grid system shown was marked in the lake by means of floats at the intersection of the transects. B. A map of the sub-habitats, which were defined in terms of vegetation and depth of water. Open mud; O-lm. 1 f;1-2m. | I l l l ; 2 - 3 m . Nuphar polype pal a Submerged vegetation [ ; Potamogeton n a tans & £q i ; Dry areas 1151111 . 3+m. turn sp.E^3 5 Estimation of numbers of tadpoles: Since the size of the lake and the difficulties of obtaining large numbers of tadpoles made a mark-and-recapture estimate unsatisfactory, I estimated the numbers of tadpoles in the lake by using a grid of funnel traps. A funnel trap, constructed of a 1-gallon plastic jug and an insect screen funnel, was attached to each buoy in the grid system. A long cord allowed the trap to be placed anywhere in the quadrat. Additional traps were placed at intervals along the shore in the shallow water. One hundred and thirty-two traps were in operation. The traps were hauled once a day at approximately the same time in mid-morning. The numbers of tadpoles and potential predators --for example, salamanders or diving bugs--were recorded, the animals were released, and the trap was replaced. This trapping procedure yields only relative numbers of animals at each trapping site. To arrive at an estimate of the absolute number present, one needs to find the ratio between the number of animals known to be present within a given area and the number caught per day by a trap in that area. Then one can reverse the procedure to interpret the trapping results. Each trap was assumed to represent a 2 sample of the population density in one entire quadrat (1000m ). • To find this ratio, I placed known numbers of tadpoles in enclosures within the lake and trapped inside. I used a wide range of tadpole densities in the enclosures to be sure of encompassing the range of densities normally found in the lake. The procedure was repeated at intervals throughout the summer to allow for changes in the behavior of tadpoles as they grew and as conditions in the lake changed. In 1969, I used plywood enclosures of 10m^, and in 1970, I used plastic cages of 2 4m (see page 7), since the plywood enclosures had been damaged during the winter. Both sets of enclosures were in approximately l m of water. This calibration method has two faults. F i r s t , the tadpoles may tend to aggregate at the sides of the enclosure, thus reducing the numbers of animals caught by a trap in the middle. Second, only one habitat type and depth is calibrated. A s a check on this calibration method, I used two other methods. 1. Direct counts: I swam random transects of known length and width through the major habitats, recording the numbers of tadpoles observed. Since tadpoles are cryptically colored and often bury themselves in the mud, this method can be used to give only a minimum estimate of numbers present. In no case was the number of tadpoles estimated by the calibration method less than the number estimated by direct counts. 2. Quadrat samples: A heavy metal frame with a screen top and sliding bottom was thrown randomly from a boat. The frame sank rapidly to the bottom, capturing any tadpole within a one-fourth square meter area. The mud was then shaken out through the screen top, leaving the tadpoles in the sampler. The numbers estimated by this method were comparable to those estimated by the trap-calibration method. Two implicit checks on the calibration method exist. F i r s t , the calibration method never resulted in impossible estimates, e.g. more tadpoles than eggs. Second, although the calibration factors varied from year to year, the corrected survivorship curves are very similar. Thus the method used to calibrate the trapping results appears to be valid. Some idea of the variance associated with trapping can be obtained from a 64-trap grid which I placed in quadrat 91 on June 11, 1969. The mean number caught per trap was 2.95, with a variance of 0.5 and a standard error of 0.28. The day-to-day trapping results also indicated consistency in the trapping success in any particular area. If a trap caught few tadpoles on one day from a quadrat, it usually caught few tadpoles every day from that same quadrat, and conversely for quadrats which yielded several tadpoles per trap. The variance associated with my estimates of the total population, however, is compounded of the 7 variances of 132 individual traps as well as the variance associated with the calibration method. Therefore, it is not possible to put confidence intervals on the population estimates. Enclosure experiment: alterations of tadpole and predator  density: This experiment was designed to answer the following questions: 1. What is the effect of density on the growth and survivorship of tadpoles? 2. Is there a threshold density beyond which growth is affected, and how does this threshold compare with the actual density of the p o p u l a t i o n in the lake? 3. Can predators (salamanders, in this experiment) affect the density of tadpoles? 4. Is there a density of tadpoles below which they have a refuge from predation? The experiment utilized sixteen enclosures. There were four densities of tadpoles with no salamanders present, and four densities of predators with a constant initial density of tadpoles. Each treatment was replicated (Table 1). The enclosures were constructed of a framework of iron pipes to which walls of 15 m i l polyethylene sheeting were attached. The enclosures were 2 meters on a side and 1.9 meters high. The joints were sealed with polyethylene tape, an unfortunate choice since the tape often came unstuck when exposed to alternate wetting and drying, as in rain storms, although it held if it was constantly submerged. Several replicates were lost in this way. Polyethylene sheeting was chosen in preference to screening because it accumulates less periphyton. The area chosen for the experiment was the open mud bottom between l/2m and l m in depth in quadrats 86 and 92. The open mud habitat was chosen for a variety of reasons. Approximately 47% of the total lake bottom is open mud between 0 and 2m in depth and thus represents the largest single habitat in the lake. The uniformity of the Experiment I - Variable tadpole density - No vertebrate predators. Tadpoles added/Pen Salamanders /Pen Pen # 20 None 12 + 20 " 16 40 " 5 40 " 14 + 100 " 10 100 - " 13 + 300 " 3 300 " 7 Experiment II - Variable predator density - Constant initial tadpole density. Salamanders added/Pen Tadpoles /Pen Pen # 4A 100 2" 4A " 6 2A;2T " * l " 2A;2T " 4 4A;4T " 9 4A;4T " 15 10T " 8 10T " l l " T A B L E 1. L i s t of treatments used in the enclosure experiments. The * signifies that an enclosure developed a leak prior to the June 21 census; the + signifies that the enclosure dev-eloped a leak after June 21. "A" refers to an Ambystoma; " T " refers to a Taricha. The area of bottom in these en-closures was 4 m^. 9 habitat ensured that conditions within the pens would be as similar as possible. The enclosures sank almost l/2m into the soft sediment, thus ensuring that the tadpoles and salamanders did not escape. Finally, trapping results in 1969 indicated that very few tadpoles inhabited the open mud areas and yet sediment from these areas could be used as food for tadpoles in the laboratory. Thus I felt that if it could be shown that these areas could support even moderate numbers of tadpoles, this would be a strong argument for a limiting factor other than food. The cages were floated into position on a boat and dropped into the sediment. This resulted in some disturbance of the mud. However, the enclosures were placed one month prior to the beginning of the experiment, so conditions had time to stabilize. A l l enclosures were trapped during this time to remove salamanders. The enclosures were a l l placed before the tadpoles hatched. Tadpoles rather than eggs were used to stock the pens. Since the population of tadpoles, in the area where the experiments were conducted, normally consists of individuals which have hatched at different times and consequently are of different sizes, tadpoles of different sizes were used to stock the experimental pens. They were taken from the deep, water in quadrats 6 4 and 6 9 , and from the shallower water in quadrat 85. The animals in deeper water were just hatching; the others were approximately two weeks old. A random mixture of the two sizes was used, which unfortunately resulted in a greater size variance in some pens than in others. The tadpoles were introduced into the pens on May 1 5 , 1 9 7 0 , and the salamanders were added two days later. The Taricha  granulosa (Skilton) used were adult males greater than 6 cm in body length, and the Ambystoma gracile (Baird) were large adults greater than 6 cm. The salamanders of both species used in the experiment were collected in M i r r o r Lake, a nearby lake in the University of B r i t i s h Columbia research forest. 10 Growth rates: I measured the body lengths of tadpoles captured from the natural population, of tadpoles kept under constant temperatures in the laboratory, and of the tadpoles kept at the different densities in the enclosures in the lake. A 35mm slide was taken of tadpoles in a white enameled pan which also contained a millimeter ruler and the date and source of the sample. Measurements of body length, to the nearest 0.5mm, could then be taken with dividers from the projected slide. The tadpoles captured by the trapping grid were assumed to be a random sample of the population and were photographed at intervals throughout the summer. Animals from three distinct areas of the lake were suspected of having different growth rates, owing to temperature differences and different hatching times. Tadpoles from each of these three areas were treated as separate samples. The results from the three samples were then combined to obtain a mean length for the entire population in the lake. This procedure results in a stratification of the sample, since the areas in which most tadpoles are caught contributed most to the total mean length (Fig. 6). To examine the effect of temperature on growth in the lake, I estimated the degree-days experienced by different segments of the total tadpole population. The temperature of the mud-water interface was taken three days per week at l/2m intervals from l/2m to 4 l/2m. The temperatures on other days were then estimated by interpolating from these data. The daily temperatures at each depth were then summed for the summer and plotted against the size attained by animals captured in o traps at that depth. A correction factor was applied by subtracting 4 C from each daily temperature because tadpoles do not grow at temperatures below 4°C (Calef, unpublished). In the laboratory, tadpoles captured from the lake were placed for three weeks in dish pans containing mud and abundant filamentous 11 algae and maintained at constant temperatures of 5 , 10 , 15 , and 20 C. The mud and algae were replaced at weekly intervals, which insured a super-abundance of food of the type the tadpoles have available in the lake. Tadpoles of different sizes and from different areas of the lake were included in each experiment. The tadpoles from the various enclosures in the field experiment were measured only twice, in order not to disturb the animals and sediment too much. Laboratory experiments with predators: The laboratory experiments with the predatory salamander Taricha granulosa were designed to answer the question: what is the relationship between the number of tadpoles eaten by the predator and the density of tadpoles present, in a situation where alternate prey are available? Taricha was chosen as the experimental animal because it was known to be an important tadpole predator in Marion Lake (Neish, 1970), was available a l l summer, and could be stomach-pumped to determine food intake. Two types of experiments 2 were performed, in wading pools 2m in area. In the f i r s t type of experiment, a layer of sediment from the lake, containing the normal complement of benthic invertebrates, was placed in the pens. The density and composition of organisms in the sediment was not measured and presumably varied from area to area, and from time to time during the year (Mathias, 1966). After two days to allow the sediment 2 to settle, tadpoles were added to obtain densities ranging from 2.5/m to 150/m^. This range includes most of the densities which occur naturally in the lake (see page 30). Then four adult male Taricha were introduced. The salamanders had been starved for six days or longer, and thus had achieved maximum hunger level (Neish, 1970). The pools were covered with a sheet of plywood to simulate the nocturnal conditions under which salamanders normally hunt. The salamanders were allowed to hunt for three hours and then removed and stomach-pumped. 1 2 The numbers of tadpoles and other food items were recorded. The other type of prey-choice experiment was a longer-term experiment in which Taricha remained in the pools up to eight days. At dusk each day, just prior to the time that salamanders normally begin hunting, various densities of Hyla regilla tadpoles (mean volume of 0.058 c.e.) and standard-sized chunks of horse heart (mean volume of 0.051 ± 0.003 c.e.) were added as prey. In the morning after the salamanders had stopped hunting, and returned to the nest boxes in the corners of the pools, the remaining pieces of meat and tadpoles were removed and counted. Hyla regilla tadpoles were used because, by the time these experiments were conducted, the R. aurora tadpoles in Marion Lake were too large for the salamanders to eat. The Hyla tadpoles used in this experi-ment were collected in the mountains of Manning Park, B r i t i s h Columbia. They were approximately the size of R. aurora tadpoles during the f i r s t two weeks after hatching, when predation by Taricha on Rana tadpoles is most evident in the lake. 1 3 RESULTS Life history of the natural population: The adult frogs appear in Marion Lake within days after the ice leaves the lake. Males begin calling in the submerged weed beds, and mating and egg laying follow immediately. I observed the f i r s t egg masses on March 9, A p r i l 4, and February 21, in the years 1968, 1969, and 1970 respectively. The breeding aggregations last approximately two weeks, although the season appeared to be prolonged in 1970, the year in which frogs entered the lake earliest. Males are often seen calling .underneath a skim of ice, and the eggs are in water of 4°- 5°C for many days. (See Appendix 1 and Licht, 1969a and 1969b, for other aspects of the breeding aggregations.) The figure in Appendix 1 shows the distribution of egg masses in the lake. A comparison of the vegetation map with the egg distribution map reveals that most of the eggs are laid in vegetation, especially in submerged beds of Potamogeton and Isoetes. In 1969, however, when the ice prevented the frogs from entering the lake until A p r i l , a substantial number of egg masses was laid in the very shallow water along the, east side of the lake. . Most of the eggs hatch during the f i r s t two weeks of May. Some of the masses in very shallow water (less than l/3m) or very deep water (greater than 3m) hatch earlier or later. In many areas, for example the Potamogeton bed in quadrat 85, there is initially a high density of 2 tadpoles: up to 500/m . The tadpoles lie on the mud or cling to the vegetation, moving very little for about a week. Then they disperse. I followed their dispersal from two areas by placing two perpendicular lines of funnel traps which intersected in the center of the weedbeds. Each a r m of the resulting cross had 12-15 traps set at 1.5m intervals. At first tadpoles were captured only in the traps closest to the weedbeds, then they appeared in traps progressively further from the 1 4 weedbeds. Two to three weeks after hatching they inhabited an area of several thousand square meters (Table 2). In 1969, tadpoles appeared in the trapping grid in quadrat 85 which were larger than the ones which had hatched there. These must have come from eggs in shallower water, which had hatched earlier. Thus over large areas of the lake the tadpoles must mix freely during the initial dispersal. Once the initial dispersal is over, there is little change in the distribution of tadpoles throughout the summer (see Appendix 2). The animals begin to metamorphose (appearance of fore limbs) in late July or early August. The metamorphosing animals move fir s t to the lakeshore and then into the woods. Tadpoles continue to metamorphose and emerge from the lake until early October. It is possible that in some years a very small number of tadpoles fails to metamorphose before the lake freezes. Tadpoles of this species do not overwinter in Marion Lake. Rana aurora frogs do not return to breed until at least their third year of life (Licht, pers. comm.). Survivorship in the natural population: The initial input of eggs in the population was determined by multiplying the mean number of eggs per mass by the total number of masses. The number of eggs per mass was determined by counting the eggs in 35 randomly selected masses in 1969. The mean number per mass was 531 + 19 (S.E.). The estimated number of eggs per mass was assumed to be the same in 1970. The estimated mortality prior to hatching is a compound of two factors; fungal infection, and death due to desiccation or freezing. The eggs which had fungal infection were not developing. I do not know whether the fungus killed the eggs or infected only defective ones. The mortality attributable to fungal infection, which is easily recognized as a white fuzz on the surface of the egg, was estimated during the 1969 egg count, taken just prior to hatching. A mean of 2.4% + 1.0 3 (S. E.) of the eggs in the 35 masses sampled were infected. Again the percentage mortality due to fungus was assumed to be the same in 1970 as in 1969. 1 5 T A B L E 2a. Dispersal of tadpoles from Potamogeton beds in quadrats 1 85 and 67 in 1969. Traps 1-5 are the five traps on each a r m of the trapping cross which are nearest to the center of the weedbeds. Other traps are progressively further away. The numbers in each column represent the total tadpoles captured on each date in each group of traps. The totals for the two weedbeds are combined. Note the tendency for the percentage of tadpoles in the outer traps to increase with time. T A B L E 2b. Dispersal of tadpoles from Potamogeton bed in quadrat 85 in 1970. Traps 1-4 are the four traps on each arm of the trapping cross which are nearest to the center of the weedbed. The other traps are progressively further away. The numbers in each column represent the total tadpoles captured on each date in each group of traps. Note the tendency for the percentage of tadpoles in the outer traps to increase with timeT 16 T A B L E 2a. Dispersal of Tadpoles (1969). Trap numbers Date in May 1-5 6-10 11-15 15 3 1 -16 - - 1 17 - -18 6 -19 72 12 -20 87 6 13 21 22 23 228 86 64 24 373 323 95 25 420 483 202 26 27 161 130 137 28 41 60 17 29 • - 43 -T A B L E 2b. Dispersal of Tadpoles (1970). Trap numbers Date in May 1-4 5-8 9-12 4 - - -5 8 2 -6 17 1 -7 63 21 3 8 9 594 87 15 10 510 204 81 11 94 26 11 12 111 39 14 13 204 75 31 14 69 22 7 15 95 39 8 16 108 113 81 17 176 120 90 18 107 150 165 19 130 123 199 20 109 110 96 17 In 1969. large numbers of egg masses (approx. 180 ) were laid in the shallows (less than l/4m in depth) on the east shore of the lake. Seven of these were destroyed by drying out when the water level f e l l , and the majority hatched just prior to drying out. In 1970, about twenty egg masses laid in the shallows on the f i r s t two days of the breeding season froze during a cold snap early in the breeding season. Thereafter, no more eggs were laid in the shallows. No predation on egg masses was observed, although Taricha has been observed eating its own eggs and those of Ambystoma gracile in Marion Lake. In most years, few Taricha are present in the lake for more than a few days prior to the hatching of the tadpoles. The survivorship of the larvae appears to have two phases. A rapid decline in numbers, from approximately 300,000 to approximately 75,000, takes place during the fi r s t three to four weeks after hatching. Thereafter, the decline is less rapid until at metamorphosis approximately 15,000, or about 5% of the initial population, remain (Fig. 2). It is possible that the decline in numbers is even more rapid than is indicated, but no population estimate is available for the fi r s t month, since the distribution of tadpoles at that time is so patchy while the tadpoles are dispersing. The point where the decline in numbers becomes less rapid coincides with the period at which the tadpoles become too large for the salamanders to eat. Survivorship in the enclosure experiment: In the enclosures containing no salamanders, a constant percentage (approximately 40%) of the initial number of tadpoles present survived over the fi r s t five weeks, except at the lowest density (5/m^), at which none died (Fig. 3). No tadpoles were surviving in the enclosures containing salamanders at the June 23 census. During this period, the survivorship in the natural population was less than 20%. Unfortunately, the data on survivorship after June 23 are sketchy. Some of the c a g e s d e v e l o p e d leaks, and hence could not be FIGURE 2. Survivorship of Rana aurora tadpoles during the summers of 1969 and 1970. Each point represents the calibrated mean estimate from ten trap-days for 132 traps. Apri l May June -July Aug. Sept. 19 FIGURE 3. Survivorship of tadpoles in the experimental enclosures. The number of tadpoles per trap in each enclosure is plotted against the initial numbers introduced. The dashed line represents the number of tadpoles normally caught in this habitat, at this time of year. The circles represent the values for individual enclosures, and the crosses indicate the mean at each density. 50 100 150 200 250 300 Initial No. of Tadpoles per cage 20 included. Just before the final census a severe rain flooded many of the cages, possibly permitting tadpoles to escape over the top of the cage. After the flood, I collected a l l tadpoles remaining in the intact cages. In the low density cage a l l 20 tadpoles survived and meta-morphosed. Even at the highest density (cage #1), a minimum of 42% 2 of the tadpoles survived, representing a final density of over 30 tadpoles/m . If this density is considered to represent the carrying-capacity of this type of habitat, then the open mud habitat between 0 and 2m in depth alone could support 1,457,000 tadpoles to metamorphosis, or five times the total number of eggs laid by the R. aurora population, and almost 100 times the total number surviving in the natural population. If fact, the 2 data on growth suggest that even 31 tadpoles/m is not the highest density which that habitat can support, since the experimental animals grew as fast and metamorphosed at the same time as the animals in the natural population. Thus we can conclude that Marion Lake is capable of supporting, to metamorphosis, a population of tadpoles at least 100 times the numbers actually surviving, and probably more, since other areas of the lake could certainly support some tadpoles. Survivorship of starved tadpoles: In the spring of 1970, fifty tadpoles were~kept in an aquarium of lake water with no added food supply. The water was changed weekly. Although there may have been slight plant growth on the walls of the aquarium, these tadpoles had much less food than any tadpoles in the lake, and fecal production was negligible. Forty-seven of the 50 tadpoles were st i l l alive at the end of the month. The tadpoles had not grown, and were obviously not healthy, but this experiment demonstrates that starvation alone cannot be the direct cause of the mortality observed in the lake. Growth rate in the natural population: The data on growth rates are presented as increases in body length, rather than as increases in weight, which are more difficult to obtain accurately for large numbers 21 of animals in the field. Since almost 50% of the dry weight of a tadpole is gut contents, great care must be taken in using weight as a measure of growth rate in tadpoles. Variation in weight of gut contents can occur, depending on where the animal has been feeding, and to what extent they evacuate their guts before being killed and weighed. A regression of length on weight is presented (Fig. 4), which shows that length allows a good prediction of weight. It is likely that much of the variance about the regression is due to differences in the weight of gut contents. It can be seen that the growth rate varies from year to year (Fig. 5) and fr o m area to area in the same year (Fig. 6). However, the size attained by a sub-population of tadpoles within the lake is strongly correlated (r = .93) with the degree-days which it has experienced since hatching (Fig. 7). This correlation holds between years and between different depths in the lake, where productivity and standing crop of benthic algae are known to vary by as much as two-fold and five-fold respectively (Hargrave, 1969). The correlation is remarkable when one considers that the size estimates used in the correlation are for a group of animals which are assumed to be living at a mean depth. That is, the animals designated as living at a depth of l m actually are caught in traps from less than 1/"4m to almost 2m in depth. Consider, for example, the size value which contributes most to the variance of the regression is the value for June 15, 1969 (* in F i g . 7). Since many of the egg masses were laid in the shallows that year, it is possible that many tadpoles lived for some time at depths less than lm, which was the depth used for the degree-days calculation for this sub-population. Consequently, they would actually have accumulated more degree-days than those shown. It thus appears that at a l l times and at a l l depths, tadpoles have sufficient food to grow at the maximum rate permitted by the temperature. It should be noted here that this method assumes that the animals captured in a given location have lived there for some time. The high correlation suggests this is true. The data on dispersion (Appendix 2) also support this assumption. FIGURE 4. A regression of log weight against log length. The e e tadpoles were captured from the lake and dried for 48 hours, at 50°C. There was no weight loss after this period. The correlation coefficient is .98. 23 FIGURE 5. Growth rate of tadpoles in Marion Lake during three summers. Each point represents the mean length of the total sample captured in the traps on that date. The standard error of each value was less than 10% of the mean. Body Length in mm. Ol o Ol 03 o 22. CD > C CO O • t > CO CO CD N j O) O) o co oo T J 24 FIGURE 6. Growth rate of tadpoles from different depths and different areas of the lake in 1970. Note that animals from deep water have little effect on the combined mean for the lake, since they are rare. 2 0 , £ E •15. O) c 10 + • A O North End South End Deep Water(>2m) Combined Sample May June July Date Aug. 25 FIGURE 7. The size attained by tadpoles from various sub-populations is plotted against an estimate of accumulated degree-days since hatching. The standard error on a l l size estimates is less than 10% of the mean. F o r the meaning of the asterisk, seepage 21. 200 . 40a 600 800 Degree-days since Hatching 26 In order to test the idea that tadpoles in the lake have max-imum food available to them, I tried to predict the growth rate of tadpoles maintained in the laboratory on maximum rations, by using the regression obtained from the field data. The amount of growth observed at maximal food under constant temperatures was predicted quite well by the equation obtained on lake animals except at the lowest temperature (10°C), which is lower than is ever encountered in the lake after May (Table 3). The error in measurement (0.5mm) also becomes significant at this lower temperature, at which the predicted growth is only 1.7rnm. Growth in the experimental enclosures: The idea that food is not limiting is supported by the results of the enclosure experiment. 2 2 Here density ranged from values of 5/m to 75/m , a density which is over twenty times that normally found in the open mud habitat in the vicinity of the experimental enclosures. Except for the low density enclosure, there was no significant difference (p = .05) in the mean size obtained by individuals at these different densities, nor is the size attained different from the mean size of the tadpoles in the natural population at this depth (Fig. 8). I cannot explain why the tadpoles in the low density enclosures grew faster than the others. Since there were no temperature differences between the inside and outside of the enclosures, temperature cannot . account for the difference. It is possible that slightly larger tadpoles were introduced into the low density enclosures, since so few were used in the low density experiment and because, as previously stated, tadpoles of unequal sizes were pooled and then used to stock the enclosures. A slight difference in the mean size of animals in one pen, at the beginning of the experiment, would be magnified by the time they had grown for five weeks. The animals in the lowest density cages also had a higher survival rate than the tadpoles in the other enclosures, which would be expected if they were larger and hence less delicate and less likely to be eaten by Odonates. Mean Mean Mean Expected Accumulated 0 Initial F i n a l Increase Increase Percent E r r o r Days Temp. Size Size (mm) , (mm) in Prediction 126 10°C 15.25 16.6 1.35 1.70 21% 231 15°C 16.0 19.2 3.2 3. 12 3% 336 20 °C 14.25 18.95 4.70 4.54 4% T A B L E 3. The difference between the actual growth of tadpoles in the laboratory, at constant temperature and maximum food supply, and the growth predicted using the growth vs. accumulated temperature regression for field animals. Ten tadpoles were used at each experimental temperature. 28 FIGURE 8. The effect of density on growth of tadpoles in the experimental enclosures. Body length on June 21, 1970, is plotted against the initial density introduced into the enclosures. The vertical bars indicate 95% confidence intervals. The mean size of animals from the natural population is shown by the solid horizontal line, and the 95% confidence intervals by the horizontal dashed lines. CM 1 8 J 1 7 J cb £ E 1 6 J B 1 5 CD ^ 1 4 o CD 2 0 4 0 6 0 Initial Density of Tadpoles/m' 8 0 29 The f i r s t metamorphosed tadpoles appeared in a l l cages at approximately the same date (July 18-20), which was the same time that animals were beginning to metamorphose in the lake. The animals in the highest density enclosures were slightly smaller, just prior to meta-morphosis, than the tadpoles from the lake (20.4mm vs. 21. 9mm; p = .05). Tadpole predators: The following species eat tadpoles in Marion Lake: 1. Rainbow trout (Salmo gairdneri) (Richardson); estimated population 4, 000 (Sandercock, 1969). Tadpoles were found in the stomachs of two specimens (out of twenty collected) in quadrat 85 in May 1970. Larger individuals of this species, confined in pens in the lake, will take tadpoles added to the pens. 2. Pacific coast newt (Taricha granulosa); estimated population 2,500 (Neish, 1970). There is extensive evidence of predation on tadpoles in the field by this species (see pages 31 -32). 3. Northwestern salamander (Ambystoma gracile); estimated population 14, 500 adults, and 45,000 young of the second year (Neish, 1970). There is evidence of predation in the field (see pages 31 -32). 4. Giant diving bug (Lethocerus sp.); no population estimate is available. However, the fact that approximately fifty individuals of this species were captured in the funnel traps each year suggests a sub-stantial population. Lethocerus has been observed to take tadpoles in the field, and can capture tadpoles of a l l sizes. 5. Odonata spp. ; the larger species of both damselflies and dragonflies have been observed to capture tadpoles in the field. In the laboratory six large Eshna sp. consumed on average three tadpoles per day. Tadpoles soon outgrow the size of vulnerability to Odonates. 6. Garter snake (Thamnophis sertalis) (Cope); no population estimate is available. However, several dozen individuals have been cap-tured in funnel traps in shallow water. The snakes in several of the traps had eaten tadpoles. 30 Estimation of predation on tadpoles in the lake: A n estimation of the effect of salamanders as predators on tadpole population can be made. Neish (1970) has developed a simulation model for predation by Taricha in Marion Lake which gives an excellent prediction of the proportion of various prey types taken by Taricha both in the lake and under experimental conditions in the laboratory. Using this model he estimates that approximately 200,000 tadpoles are eaten by Taricha during May and the f i r s t two weeks in June. His estimate is based on the assumption that the entire population of Taricha is feeding in weedbeds where tadpoles are concentrated in high density, and that the , 2 . 2 . • 2 density of tadpoles is 100/m , 50/m , and 4/m in three successive two-week periods (page 66). Neish's estimate of predation by Taricha is too high for several reasons: 1. There are not 200,000 tadpoles available at these high densities. On the basis of the known numbers of eggs and dispersal rates of the tadpoles, I estimate the following distribution of the tadpole population during May: 63,000 tadpoles at densities between 100 and 2 2 500/m ; 84,000 tadpoles at densities between 25 and 100/m ; and the 2 remaining 150,000 at densities between 5 and 10/m . 2. -The entire Taricha population is not found in the "high-density" weedbeds as Neish assumed. In 1969, I estimated, by a mark-and-recapture method, a population of approximately 500 Taricha in the Potamogeton beds in quadrats 85 and 62. Taricha were also abundant in quadrats 112, 113, 125, 71, and 54 (the areas of intermediate tadpole density). Thus a reasonable breakdown of the Taricha distribution is as follows: 500 Taricha in areas of high tadpole density; 500 Taricha in areas of moderate tadpole density, and the remaining 1, 500 in low density areas. 3. Neish made no correction to his tadpole densities in each 2-week interval, as tadpoles were removed from the population by predation. 3 1 Neish 1 s model re-run with the above corrections results in an estimate of 120,000 to 140, 000 tadpoles eaten by Taricha in May: 50, 000 eaten from areas of high tadpole density, 40, 000 eaten at the intermediate densities, and 30, 000-50, 000 at low densities. Another method which could be used to estimate the mortality due to salamander predation is to capture Taricha and Ambystoma in the field, sample their stomach contents, and project the observed predation rate to the entire population of salamanders. This method has the difficulty that, for two reasons, it tends to under-estimate the number of tadpoles eaten. F i r s t , tadpoles are digested so rapidly that they are indistinguishable after a few hours. Second, salamanders might be captured before they had finished hunting. However, a correction factor from laboratory experiments can be applied to the digestion rate. In two laboratory experiments (see pages 34-35), groups of Taricha consumed 150 and 184 tadpoles on the last day of the experiment. Sixty-three and 92 tadpoles, respectively, were recovered when these Taricha were stomach-pumped the next morning. This indicates that less than 50% of the tadpoles were distinguishable within a few hours after consumption. Thus the field samples may be corrected by multiplying by 2. , . In the.two years of my study, stomach contents were sampled from 38 Taricha and 47 Ambystoma, captured in areas of moderate tadpole density. Thirty percent of the Taricha and 20% of the Ambystoma contained tadpoles. The corrected mean number of tadpoles eaten by the Taricha which contained tadpoles was 3.6, and the corrected mean for Ambystoma was 2. Twenty Taricha were taken from areas of high tadpole density. Virtually a l l of these salamanders contained tadpoles, with a corrected mean of 12.5 tadpoles per salamander. These figures agree well with the prediction from Neish's corrected model. He predicts between 13.2 and 14. 2 tadpoles per Taricha at high tadpole density, as compared with the 12.5 observed, and 2.6 per salamander at moderate densities, as compared with the observed 1.2 (that is, 3.6 x 30% = 1.2). The discrepancy in the 32 latter figures may be due to capturing Taricha before they had finished hunting. Thus the field estimations give a prediction of total Taricha predation slightly over 100,000 tadpoles. The predicted predation by Ambystoma is 48, 000 (6, 000 Ambystoma x 0 .4 tadpoles /day x 20 days). The Ambystoma are assumed to feed for only 20 days because after this time the tadpoles are too large for capture by most of the salamanders. Thus during the 4 weeks following the tadpole hatch, salamanders appear capable of eating at least 150,000 tadpoles, which is approximately 75% of the observed mortality during this period. Predation by fish and insects likely makes up the rest of the mortality. It would be difficult to determine the numbers of tadpoles eaten by these other predators in Marion Lake in the course of a summer. In order to do so, one would need to know the numbers and distribution of the predators and their alternate prey, as well as the relative vulnerabilities and availabilities of each prey species to each predator. Such data are not available for most of the predators, and many of the prey. Laboratory experiments with predation: The results of the experiments in which Taricha fed with several prey-types available, show that there is a linear, or very nearly linear, relationship between the density of tadpoles, and the number eaten per salamander (Fig. 9). Above a certain density, there is no further increase in the tadpoles consumed with increasing density. This asymptote results because the salamanders are able to feed to satiation in three hours at and above this density. The satiation ration for Taricha is known (Neish, 1970), and the Taricha feeding at high tadpole densities in these experiments achieved this satiation ration of 1.6-1.7 cc. Many of the Taricha were so full that they regurgitated tadpoles while being anaesthetized for stomach-pumping. The differences in slope and asymptote of the curves for various experiments presumably represent differences in availability of alternate prey, and in the escape 33 FIGURE 9< The relationship between density of tadpoles and the number eaten by Taricha, in a multi-prey system. Each point represents the mean number of tadpoles eaten by 8-24 salamanders. The standard errors for experiments at the intermediate and high tadpole densities are about 30% of the mean, and are higher for low densities, since many salamanders eat no tadpoles at these low densities. The squares and circles indicate values obtained in experiments with newly-hatched tadpoles as prey, on May 23-25, 1970, and June 1-6, 1969, respectively. The values represented by triangles are for an experiment with 2-week old tadpoles (May 24-25, 1970). The two upper curves have asymptotes because in these two experiments the salamanders could feed to satiation in 3 hours (page 32). — i i T r ~ 5 0 1 0 0 2 0 0 3 0 0 Total No. of Tadpoles per Pool 34 behavior of tadpoles of different ages. The important feature is the linear rise to the "satiation" plateau, which implies that, over most densities, tadpoles will be taken in proportion to their abundance in a multi-prey system. I have compared the stomach contents of Taricha feeding at low tadpole density, in my experiments, with stomach contents of animals caught from the lake. At low tadpole densities, mostly inverte-brates are eaten by Taricha. There is no significant difference either in volume of prey eaten or in prey-types caught between experimental Taricha and animals from the lake. Thus this experimental design seems to mimic the "normal" conditions under which Taricha hunt, which is reflected by their hunting success. The curves (Fig. 9) seem to go to zero not at the origin, but at 2 density of approximately 2.5 tadpoles/m , suggesting that there is a density below which a rare prey is rarely eaten in a multi-prey system. Although this conclusion is not statistically justifiable with this experiment, it is confirmed by data from other experiments (see below). The results of the eight-day experiments confirm two important points (Table 4): . 1. Reduced vulnerability of rare prey: In three experiments (pens 1, 2, and 5), one prey-type (either tadpoles or pieces of heart) was present at low density, and in low proportion. A chi-square analysis was done on each experiment to determine the similarity of the proportion of the two prey-types eaten, as in the following example (pen #2): # eaten # not eaten Pieces of Heart 324 876 Tadpoles 6 74 Heart Hyla Pool # Density Density Heart Pieces Hyla Heart Hyla (#/m2) (#/m2) Available Available Eaten Eaten 1 75 1 1200 16 314 1 2 75 5 1200 80 324 6 3 75 25 1200 400 280 91 4 5 75 80 1200 79 1170 5 15 300 240 4800 37 1477 T A B L E 4. Results of the eight-day predation experiment. The numbers in the last four columns represent totals for the experiment. The prey eaten are totals for the eight Taricha feeding in each pool, over the eight-day period. 36 In a l l three cases the number of rare-type individuals eaten is less than one would expect by random encounter (p = .001). This result is compatible with the results obtained in the short-term experiments, which suggested that at low population density very few tadpoles were taken. The fact that in one case horse heart was the rare prey, and in the other two cases tadpoles were rare, eliminates the possibility that some characteristic of tadpoles as prey accounted for their low vulnerability at low density. It appears that any prey whose density and proportion of the total prey is low, is less likely to be eaten than one would expect, and at some densities this reduced vulnerability amounts almost to a refuge from predation by Taricha. 2. Differential digestibility of prey: An estimate of the rate at which Taricha digest different prey can be obtained from these experiments as follows: At high density of prey, salamanders eat a constant quantity, which is equal to the maximum amount which they can digest each 2 day (Neish, 1970). F o r example, at 300 tadpoles/m (pen #5) the eight salamanders consumed 1477/8 x 8 = 23.07 tadpoles per day per salamander 2 over the eight-day experimental period. At 75 heart-pieces/m (pens #1 and 2) the Taricha ate 314=324/8 x 8 x 2 = 5.0 pieces of heart per day. The amount of alternate prey consumed in each case is negligible (see Table 4), but the actual amount of tadpoles and heart-pieces eaten would have been very slightly higher if the tiny quantity of alternate prey had not been eaten. Tadpoles thus seem to be passed through the gut at a rate of about 4.7 times as fast as heart-pieces. These differences in digestive rate can be seen if Taricha are stomach-pumped shortly after feeding. The pieces of heart are almost unaffected by the digestive process, whereas the tadpoles are in an advanced state of digestion within 2-3 hours after being eaten. Most of the other prey in the lake seem to be digested at a rate comparable to that of horse heart. F o r example, on several occasions I have found live amphipods in the stomachs of Taricha containing highly-digested tadpoles. 3 7 The "digestibility" of a given prey type will influence the extent to which it can protect another prey population from predation, by acting as an alternate prey. This can be seen in the results of the reciprocal experiments in pens 2 and 4 (that is, experiments in which the proportions of the two prey types were reversed) (Table 4). In the case where tadpoles were the abundant prey (pen #4), 79 out of 80 pieces of horse heart were eaten; whereas, when horse heart was the abundant prey, only 6 out of 80 possible tadpoles were eaten. Thus in a situation where the proportions of two alternate prey were identical, the degree of protection afforded by the abundant prey depended on its relative digest-ibility. F i v e times as many tadpoles as horse heart pieces (or amphipods presumably) would have to be consumed by the Taricha to achieve the same degree of satiation. This can have important consequences in a situation where total prey is limited. 38 DISCUSSION Comparisons with results from other studies: Although quantitative studies of survivorship in anuran larvae are sti l l rare in the literature, the results obtained in my study are comparable to those obtained in studies of some other species populations. Herreid and Kinney (1966) found a mean survival to metamorphosis of 4% in four populations of Rana sylvatica tadpoles. Moreover, their survivorship curves demon-strate the same shape that I found, with 60-80% of the animals dying in the f i r s t three weeks, and a lower death rate thereafter. Other published figures for survival of tadpoles include 4% for Bufo americanus (Brockle-man, 1968), and less than 8.5% for Rana pretiosa (Turner, I960). Thus, it appears that a type III, or concave, survivorship curve (Deevey, 1947) characterizes the few tadpole populations which have been studied, and that a survival rate of less than 10% to metamorphosis may be typical. However, with the exception of Brockleman, none of the authors attempted to explain the observed mortality. Turner (1962) stressed the variability of survivorship in tadpoles, althoughJie maintains that too much attention is paid to the "wholesale disasters" which sometimes befall l a r v a l populations in the temperate regions owing to meteorological catastrophes of various sorts. During the two years of my study, both the shape of the survivorship curve and the final percentage metamorphosing were similar. This lack of variation is probably due to the relative stability of a lake such as Marion Lake as compared with some of the more fluctuating and ephemeral ponds and puddles where frogs and toads often lay their eggs. Most of the egg masses in Marion Lake were deposited in deeper water and thus were not exposed to the possibility of either desiccation or freezing. In fact, there appeared to be a strong selection pressure against egg laying in 3 9 shallow water (see page 17). Several lines of evidence support the conclusion that most of the tadpole mortality observed in this study is attributable to predation. F i r s t , the predation rates by salamanders observed in the field and estimated from Neish's (1970) predation model account for at least 75% of the observed mortality during the firs t weeks after hatching. Moreover, salamander predation ceases at the time that the mortality curve for tadpoles becomes less steep. Tadpoles kept in the laboratory survive for weeks, even on starvation diets and at high densities. In the enclosures without salamanders, survivorship was higher than in the natural population, while in the enclosures with salamanders, the tadpoles a l l died. The mortality observed in the later part of the summer, after the tadpoles had become too large for salamanders to capture, is probably due to predation by Lethocerus and Thamnophis. Both species inhabit the shallow weedy edges of the lake, towards which the tadpoles migrate as they near metamorphosis, and they are capable of attacking even large tadpoles. It would be very difficult to estimate the Lethocerus population, since they cannot be seen or trapped alive for a mark-recapture estimate. Moreover, analysis of stomach contents could not be used if one wanted to estimate actual predation rates in the field, since Lethocerus does not ingest its prey whole. However,. just because we can show that greater numbers of tadpoles survive in the absence of predators during their aquatic stage--as appears to be the case in my enclosure experiments--we cannot infer that greater numbers of frogs would return to breed three years thereafter if predators were removed from the lake. We must always keep in mind the concept of inter compensatory mortality, which in its simplest form states that surplus animals which don't succumb to one source of mortality will succumb to another (Errington, 1946). 40 My interpretation of the cause of tadpole mortality disagrees with Brockleman's (1968). He concluded from his studies on Bufo americanus tadpoles that both growth and survivorship were inversely dependent on density of tadpoles, and that if predation is important, it is important only as a factor interacting with variations in growth rate of tadpoles, which were density-dependent. However, the validity of his conclusions can be challenged on several points. Perhaps the most important weakness, in light of my work and that of Anderson (1968), is Brockleman's omission of salamanders from his experimental design. He writes (page 17), "Ambystoma  tigrinum larvae were found to be voracious tadpole predators during a pilot study in 1965, when a few individuals in an open-bottomed pen consumed every tadpole introduced (nearly 100 per Ambystoma larva in two weeks). Thereafter, an attempt was made to exclude a l l such larvae from the pens, as they would produce an uncontrollable error in the experimental design." He made no effort to estimate the number of salamanders present in his ponds or to estimate predation rates occurring naturally. Brockleman also dismisses Lethocerus on the basis of seeing very few. Very few were seen at Marion Lake, but trapping showed them to be quite abundant. Brockleman admits that the mechanism of density-dependent mortality is not revealed by his study, and notes (page 50), "It is doubtful that starvation by itself could be a sufficient source, since tadpoles in the laboratory, smaller than those in the pens, live for some time. If this is true, then food limitation, interacting with some other source of mortality, is a cause of density-dependent mortality." Since Brockleman has no continuous record of the numbers of tadpoles surviving in the natural population, he cannot say whether the bulk of natural mortality occurs before or after sufficient time has passed for differences in tadpole growth rate to become important. In my study, as in Herreid and Kinney's, 41 and in Macan's (I960), most of the mortality occurred very early in l a r v a l life, before significant differences in growth rate due to density would manifest themselves. Temperature differences did exist between the experimental pens in Brockleman's study, and these could be sufficient to account for the small differences in growth rates observed. Brockleman does not know the effect of temperature on the growth of Bufo tadpoles. Even though Brockleman claims that food was limiting the growth of tadpoles at higher densities in his study, he noted that tadpoles rarely ventured more than a few feet from shore, and thus were utilizing only a small proportion of the total habitat available to them. However, it should be noted that the natural densities which 2 Brockleman encountered in his study ponds were approximately 1000/m , many times higher than the highest natural densities (after dispersal) in my study. This figure indicates the potential impact of tadpoles as grazers in certain habitats (see page 48). Macan ( I960) , who studied the effects of introducing Brown Trout (Salmo trutta) into a tarn, reported that tadpoles of Rana and Bufo were among the species "greatly reduced" in numbers following the introduction of the fish. He states (page 437) , "When there were no fish, tadpoles could be seen swimming all over the tarn until the end of June, when tiny adults were common round the edge. After the fish had been introduced, tadpoles could be found at a few places at the extreme edge, where the vegetation prevented to approach of fish. No tadpole was recorded inside a fish. Most disappeared within a few days of starting to swim, and dates on which fish were netted could easily have fallen on either side of this short period. Moreover, a tadpole probably does not remain recog-nizable long in the digestive juices of a trout." This again illustrates the difficulties in estimating predation rates by sampling stomach contents in the field, especially with a creature as digestible as a tadpole. 42 In considering the relationship between the tadpoles and their predators in Marion Lake, one sees that the potential impact of predation on the tadpole population is very great. The fish and salamanders alone have a population of over 50,000; whereas, the tadpoles number approximately 300,000. Any of these predators can eat more than 6 tadpoles in a single day. In order to see just how rare the tadpoles are in comparison with other prey, let us look at the density of amphipods Hyallela azteca and Crangonyx richmondensis. Mathias (1966) gives the , 2 . 2 mean density of these species as approximately 1, 500 /m and 300/m , respectively, at l m depth in July. Thus in one quadrat of the lake 2 (1, 000m ), there are many more amphipods than the entire population of tadpoles put into the lake each year. In fact, excluding chironomids, the number of amphipods in the benthic community exceeds the numbers of a l l other macroscopic prey species combined, and they account for well over 50% of the total prey volume consumed by salamanders (Neish, 1970). This pattern of abundance, with one or two species accounting for as many as 50% of the individuals present, seems characteristic of many commun-ities (e.g., Calef and Grice, 1967; Fager, 1968), and has important consequences for predator-prey interactions. Thus, we are faced, not so much with the problem of whether the predators can have a significant effect on the tadpole population (as we might be in considering birds feeding on forest insects, or predators feeding on small mammals (Elton, 1942) ), as with the problem of how any tadpoles survive at a l l . Much attention has been given in the ecological literature to this question of how predator-prey systems are stabilized and extinctions prevented. Instabilities are seen most clearly in laboratory predator-prey systems having the following characteristics (e.g. Gause, 1934; Huffaker, 1958; Huffaker et al, 1963): 1. A single species of predator depending for its food on a single species of prey. 43 2. Predator and prey reproducing continually. 3. Prey being available at a l l times. 4. P r e y being relatively evenly distributed in a uniform habitat. Since many of these conditions do not obtain in natural communities, and since many predator-prey systems seem to be relatively stable in nature (at least extinctions are rarely seen), much study has been devoted to discovering the feedback mechanisms which might confer stability on natural predator-prey systems. I would like to examine some of these hypothesized mechanisms to see which seem to apply to the tadpoles and their predators in Marion Lake. Functional response of predators to prey density: Holling (1959) has emphasized that the relationship between the density of a prey species and the number eaten by a predator is not always a simple linear function. He distinguishes two possible alternative types of response by a predator to increasing prey density: the "invertebrate" type, which is a negatively accelerated rise to the "satiation plateau", and the "vertebrate" type response, which is a sigmoid relationship in which fewer prey are taken than would be expected at low density, and greater numbers are taken at high density. A predator with a "vertebrate" type response has a poten-tial stabilizing effect in a predator-prey system, in that the predation is operating as a negative feedback, reducing the mortality of the prey population when prey density is low and increasing mortality when the density is high. The best field data demonstrating a "vertebrate" type response of predators to prey density is Tinbergen's ( I960) work on birds eating forest insects. Tinbergen hypothesized that the birds developed a "specific searching image" for a particular prey after a certain number or frequency of contacts with it. When a particular prey was rare, the birds seldom captured it, since they were not looking for it. As it became more abundant, the searching image developed, and the prey type was then taken in even greater proportion than its abundance would suggest, since the birds were 44 then searching for it specifically. Royama (1969) has challenged Tinbergen's concept of "searching image" as a necessary explanation of the data observed. Royama claims that the same disproportionate representation of a given prey-type in the birds' diet would be predicted if prey-types were localized in different "niches" (he means habitats) and the birds searched predominantly in areas where prey was in highest abundance or "profitability". I will discuss this idea under the heading "numerical response of predators to prey density. " Nevertheless, laboratory studies have shown that experience can change a predator's response to a given prey-type, usually by altering the "reactive" or "recognition" distance (Ware, in prep. ; Beukema, 1968). This change in response with experience is a l l that is required, in theory, to change the proportion of a prey species taken from that expected to be taken, on the assumption of random encounters by the predator. My experiments with Taricha show that a prey, present in both low numbers and low proportion, is under-represented in the salamanders' diet. Because of this response, tadpoles have a refuge at low density from salamander predation. However, my work did not reveal the mechanism behind this reponse. Even if salamanders continued to take tadpoles in the expected proportions at low density, the tadpole population would sti l l have what amounts to a refuge. This is because at low density so few tadpoles would be eaten that the total number eaten would be small and then the tadpoles would grow to a size too large to be eaten by the salamanders. Thus the absolute number of tadpoles eaten would be less at low density than at high density. I am troubled, however, by the results of the enclosure experiments. If a refuge exists at a low density of tadpoles as claimed above, then I should have expected some of the tadpoles to have survived in the enclosures containing salamanders. In fact, none survived. Three possible explanations occur to me. F i r s t , since the initial number of tadpoles 45 was high, perhaps some sort of learning took place while the tadpoles were present at high density, and was retained into the period when tadpoles had become reduced in numbers. Second, the results of the entire experiment in the enclosures may have been an artifact if both salamanders and tadpoles congregated at the sides of the enclosure. This would have the effect of raising the density of tadpoles locally. Such a situation is possible since the tadpoles used in this experiment in the lake were at the stage when dispersal normally occurs, while the Hyla tadpoles used in the laboratory experiments showed no obvious tendency to move towards the sides of the pools. Third, the salamanders may have decimated the alternate prey in the enclosures to such an extent that the tadpoles became proportion-ately more abundant. This seems unlikely since in several of the enclosures the salamanders' density was only slightly above normal. Numerical response of predators to prey density: Many of the instabilities in predator-prey systems result from time-lags in the numerical response of predators to prey density. If the survivorship or reproduction of a predator population is increased by increasing prey density, then the predators will be able to exert increased predation pressure on the prey. However, unless the life span and reproductive rate are similar for prey and predator, an increased burden of predation may fall on an already depleted population of prey. It seems unlikely that tadpoles would produce a "numerical response" in salamanders in Marion Lake, because they grow too large to be captured by the salamanders, and consequently they are not available for much of the summer. Any increase in the number of predators caused by an increase in the numbers of tadpoles would have to be supported by the other prey populations in the lake throughout the remainder of the summer. The total number of other prey, compared with the numbers of tadpoles, is such that tadpoles form a very small proportion of the total prey of salamanders when considered over an entire summer. It seems likely that the numbers of predators, in a community containing many prey species, would be adjusted relative to the most abundant species of prey (amphipods in the case of Marion 46 Lake) especially if the most abundant species of prey were present at a l l times, while the r a r e r ones occurred only periodically, as do tadpoles and certain species of caddis. Migration of predators into an area of high prey density is a mechanism by which predators can effectively increase their numbers to exploit an increased prey abundance without producing population oscillations induced by lags associated with differential reproductive rates. Taricha, especially, is capable of detecting local prey concentrations, since it can follow olfactory gradients to locate a source of prey (Neish, 1970). Taricha were captured in large numbers in the weedbeds in quadrats 85 and 67 during May 1969, when Rana tadpoles were present in high density. As the tadpoles dispersed, the Taricha became much less numerous in these areas. This influx was not as marked in 1970, suggesting that the phenomenon may not occur in a l l years. However, in both years, 2 Taricha were present in other areas of moderate tadpole abundance (50/m 2 to 100/m ). One wonders why the frogs clump their eggs. If the egg masses were dispersed more evenly throughout the lake, mortality should be lower, because there would be no concentrations of tadpoles for Taricha to orient towards, and at low density the total amount of predation would be much less before the tadpoles grew to sufficient size to have a refuge. Interactions of growth rate and predation: If the growth rate of the tadpoles were lowered, the tadpoles would remain vulnerable to predation for a longer time. We would certainly expect that at some density the growth rate of tadpoles would decrease. However, my experiments in the lake show that the density at which growth would be affected is many times higher than the normal densities observed in Marion Lake. In other habitats, such as the one studied by Brockleman (1968), variations in growth rates at different densities may affect the total vulnerability of the tadpoles to predation. Population regulation in Rana aurora: The inevitable question 4 7 which arises in a population study is, what regulates the population? In the case of R. aurora in Marion Lake, I must answer "I don't know, " because I was unable to study the t e r r e s t r i a l stages of the life cycle. Regulation has been defined operationally by ecologists in a variety of ways. One method of operationally recognizing regulation is to increase a supposedly limiting resource, e.g. food, or conversely to reduce a potentially limiting source of mortality, e.g. predation, and see if the population increases. I have inferred, from my experimental observations, that if predation by salamanders were excluded, the tadpoles would have an increased survival rate during their aquatic stage. This does not necessarily mean that more frogs would come back to breed in three years if salamanders were removed from Marion Lake. In fact, since 15,000-30,000 juvenile frogs leave the lake each year, but about 1, 000 new adults may be necessary to maintain the present population level*, a percent mortality as high as that observed in the lake (i.e. 90-95%) must occur between metamorphosis and adulthood. Some ecologists maintain that if mortality is density-dependent, then regulation has been demonstrated (Eisenberg, 1967). I have shown that some aspects of salamander predation depend on the tadpole density; so in this respect,- regulation is occurring. The numbers of frogs leaving the lake were similar during the two years of my study. This in itself is suggestive of regulation, since the concept of regulation implies the mainte-nance of the population about some mean level. However, the initial input This crude estimate of adult survival is obtained from the data on the breeding chorus (Appendix 1). In 1969, a total of 344 adult males was marked, or 20% of the estimated population of 1770 . In 1970, 405 males were cap-tured. If all the frogs marked in 1969 had survived, we would expect .20 x 405 = 81 of them. In fact, 66 were captured, indicating a survival rate of 66/81 = 80%. If 20% of the males in fact die each year, then .20 x 2700 (the mean of the population estimates for 1969 and 1970) = 540 new males would be required to maintain the population. Assuming equal sex ratio in the population, and equal survival of males and females, then approximately 1,000 new frogs would be required each year. Clearly, there are many assumptions involved, but 1,000 frogs is the best estimate possible. 48 of eggs was almost identical in the two years, so we have no information on the tendency of the population to return to a given level after it has been displaced. Perhaps we should view the life cycle of a population as a series of phases (e.g. egg, tadpole, metamorphosis, dispersal after metamorphosis, sub-adult, and adult), each with its own set of density-dependent mechanisms. Thus the amount of regulation that a population would experience in any one phase would depend on what it had experienced in the previous phases. The input into the tadpole population in the two years of my study did not vary by more than 10% . A relatively constant population such as this is characteristic of many species which have behavioral mechanisms of population regulation. Unfortunately, no long-term studies of anuran populations are available as they are for birds (Lack, 1966) and mammals (e.g. Errington, 1963). Among temperate zone frogs, only the Bullfrog Rana catesbeiana is known to exhibit territoriality during the breeding season (Emlen, 1968), although some other species have home ranges (Dole, 1968). Thus it is not possible to exclude some social regulating mechanism of population size in R. aurora, although I found no evidence for such a mechanism during my studies of the breeding chorus in Marion Lake (Appendix 1). Effect of tadpoles as benthic grazers: Tadpoles could have a potentially large effect on the benthic community by utilizing benthic productivity and mechanically distributing the sediment. This is because the rate at which they pass food through the gut is so high (on the order of 2-6 hours, Savage, 1952; Calef, unpublished). However, since a large proportion of the tadpole population dies soon after hatching, when the animal's gut capacity is s t i l l small, this potential is not realized. If most of the tadpoles grew almost to metamorphosis, their contribution to benthic grazing would be increased tremendously (Fig. 10). Thus the mortality rate, as well as the final percentage surviving, influences the ecological role of an animal population in its community. 49 FIGURE 10. Grazing rate of the tadpole population under three mortality-schedules.* Curve #1 - grazing rate of the natural population. Curve #2 - grazing rate assuming a linear mortality from hatching to August 1. Curve #3 - grazing rate assuming no mortality. Curve #4 - grazing rate of amphipod population (from Hargrave, in press).^ * The grazing rate of tadpoles is calculated as follows: F r o m the mor-tality schedule the number of animals in the population at 3-day intervals is computed. The mean size of the tadpoles on each date is calculated from an average growth curve for the three years of my study. This allows a calculation of the weight of gut contents. The animals are assumed to f i l l their guts 7 times/day (as they do at 18 C, Calef, unpublished results). Although there would presumably be some variations in the weight of food eaten at different temperatures, the temperature variation is not great dur-ing most of the time that tadpoles are in the lake, and the animals used in the calculations for average weight are living at different depths and conse-quently at different temperatures. Thus the error in not correcting for temperature is probably not significant. Hargrave (1971) gives a table for the amount of sediment egested per day per meter squared by the amphipods (Hyallela) living at l m depth. F r o m these figures, the amount of sediment ingested can be calculated by correcting for the 15% assimilation of sediment by the amphipods (Har-grave, 1969). To compute the grazing rate of the entire population, the figures for l m 2 are multiplied by 63, 000, the area of the lake between 0m and 2m in depth (Neish, 1970). The population of amphipods below the 2m depth in Marion Lake is insignificant (Mathias, 1966). Comparisons of the population estimates and growth data presented by Hargrave (1969) and Mathias (1966) indicate that the total grazing rate of the amphipod population may vary by as much as 100% from year to year, and the estimate presented here appears to be near the maximum. Dry wt. of Sediment ingested in Kg./day 50 Hargrave (in press) has done studies which allow a calculation of the grazing rate of amphipods in Marion Lake (Fig. 10). Although the total grazing rate by amphipods varies from year to year, owing to popu-lation fluctuations and differences in-temperature, it is clear that the amphipods 1 grazing rate is approximately three times that of the tadpole population, during the three months in which tadpoles are in the lake. Thus, grazing by tadpoles is of only moderate importance when compared with that of the amphipods and other of the invertebrate grazers (for example, ehironomids, by biomass estimates, are suspected to be even more important than amphipods; V. McCauley, personal communication). However, since my study has suggested that considerably more tadpoles than normally occur could be supported in Marion Lake, and because other studies (e. g. Brockle-man, 1968) have shown that tadpoles are often present in high numbers, tadpoles must not be dismissed a p r i o r i as unimportant in the "ecological equation" of fresh water habitats. 51 SUMMARY 1. The number of egg masses laid by the Rana aurora population in Marion Lake was almost identical in the two years of this study (618 vs. 620). It was concluded, from studies on the breeding chorus, that not a l l of the adult frogs breed each year. Egg mortality, prior to hatching, was less than 3% in a random sample of egg masses, although in each year a small number of masses in shallow water were lost by freezing or drying out. Approximately 300,000 tadpoles hatched each year. 2. During the fi r s t 4 weeks after hatching, approximately 200,000 (66%) of the tadpoles died. In the next 7-10 weeks at least 70,000 more died, so that 5-10% of the initial population remained at metamor-phosis. 3. Total predation on tadpoles by the salamander populations was estimated by two methods. The two methods gave similar results and indicated that salamander predation accounted for at least 150, 000 tadpoles, 75% of the mortality, in the fi r s t 4 weeks after hatching. There-after, tadpoles were too large for salamanders to capture. It was "argued that other predators account for significant mortality, after the tadpoles become too large for salamanders, but direct evidence is lacking. 4. Tadpoles grew at a rate dependent on temperature, regard-less of their location in the lake. Their growth rates in the lake could be described by the equation: Length (mm) = 7.25 + .0135 x accumulated degree-days, and this equation was used to predict growth rates in the laboratory of tadpoles provided with maximum food supply. 5. Tadpoles were kept in enclosures in the lake, from which . 2 . 2 salamanders were excluded, at 4 different densities (5/m , 10/m , 2 2 25/m , and 75/m ). The percent surviving, the growth rate, and the time to metamorphosis were unaffected by density, with one exception, the cages with 5 tadpoles/m 2. It was concluded from these experiments, and from 52 item 4 above, that in the absence of predators 100 times the normal density of tadpoles could survive to metamorphosis in the lake. 6. Predation studies in the laboratory with the salamander Taricha granulosa showed that there was a linear relationship between the density of tadpoles and the numbers eaten by salamanders, in a situation where "natural" alternate prey were available. "When tadpoles were present 2 at very low density (5/m ) and constituted a low proportion of the total prey, less were eaten than would be expected. Thus, predation by salamanders has a density-dependent effect on the tadpole population. 7. B e c a u s e such a s u b s t a n t i a l proportion of the tadpole population dies soon after hatching, most of their potential impact as benthic grazers is not realized. 53 L I T E R A T U R E CITED Anderson, J.D. 1968. A comparison of the food habits of Ambystoma macrodactylum sigillatum, Ambystoma macrodactylum croceum, and Ambystoma tigrinum calif orniense. Herpe-tologica 24: 273-284. Beukema, J.J. 1968. Predation by the three-spined stickleback (Gasterosteus aculeatus L.): the influence of hunger and experience. Behaviour 31: 1-136. Brockleman, W.Y. 1968. Natural regulation of density in tadpoles of Bufo americanus. Ph.D. thesis, Univ. of Michigan. 78pp. Calef, G.W., and G.D. Grice. 1967. Influence of the Amazon River outflow on the ecology of the western tropical Atlantic II. Zooplankton abundance, copepod distribution, with remarks on the fauna of low salinity areas. J. Mar. Res. 25: 84-94. Deevey, E.S., J r . 1947. L i f e tables for natural populations. Quart. Rev. B i o l . 22: 283-314. Dole, J.W. 1968. Homing in Leopard frogs, Rana pipiens. Ecology 49: 386-399. Efford, I . E . 1967. Temporal and spatial differences in phytoplankton productivity in Marion Lake, B r i t i s h Columbia. J. F i s h . Res. Bd. Can. 24: 2283-2307. Eisenberg, R.M. 1966. The regulation of density in a natural population of the pond snail Lymnaea elodes. Ecology 47: 889-906. Elton, C. 1942. Voles, mice and lemmings. Clarendon Press, Oxford. Emlen, S.T. 1968. T e r r i t o r i a l i t y in the Bullfrog, Rana catesbeiana. Copeia 2: 240-243. Errington, P.L. 1946. Predation and vertebrate populations. Quart. Rev. Biol. 21: 144-177, 221-245. 54 Errington, P.L. 1963. Muskrat populations. Iowa State University-Press, Ames. 665 pp. Fager, E.W. .1968. The community of invertebrates in decaying oak wood. J. Anim. E c o l . 37: 121-142. Gause, G.F. 1934. The struggle for existence. Williams and Wilkins, Baltimore. 163 pp. Gosner, K.L., andl.H. Black. 1955. The effects of temperature and moisture on the reproductive cycle of Scaphiopus h. holbrooki. Amer. Midi. Nat. 54: 192-203. Hargrave, B.T. 1969. Inter-relationships between a deposit-feeding amphipod and metabolism of sediment microflora. Ph.D. thesis, Univ. of B r i t i s h Columbia. 195 pp. Hargrave, B.T. In press. Prediction of egestion by a deposit-feeding amphipod. Oikos. Herreid, C.F., andS. Kinney. 1966. Survival of the Alaskan woodfrog (Rana sylvatica) larvae. Ecology 47: 1039-1041. Holling, C S . 1959. The components of predation as revealed by a study of small-mammal predation of the European pine sawfly. Can. Entom. 91: 293-320. Huffaker, C.B. 1958. Experimental studies on predation: dispersion factors and predator-prey oscillations. Hilgardia 27: 343-383. Huffaker, C.B., K.P. Shea, andS.G. Herman. 1963. Experimental studies on predation: complex dispersion and levels of food in an acarine predator-prey interaction. Hilgardia 34: 304-330. Jenssen, T.A. 1967. Food habits of the green frog Rana clamitans before and during metamorphosis. Copeia 67: 214-219. Lack, D. 1966. Population studies of birds. Clarendon Press, Oxford. 341 pp. 55 Licht, L . E . 1969a. Comparative breeding behavior of the red-legged frog (Rana aurora aurora) and the western spotted frog (Rana pretiosa pretiosa) in southwestern B r i t i s h Columbia. Canadian Jour, of Zool. 47: 1287-1299. Licht, L . E . 1969b. Unusual aspect of anuran sexual behavior as seen in the red-legged frog, Rana aurora aurora. Can. J. Zool. 47:505-509. Macan, T.T. 1966. The influence of predation on the bottom fauna of a moorland fish pond. Arch. Hydrobiol. 61: 432-452. Mathias, J.A. 1966. Population energetics of two amphipod species in Marion Lake. M.Sc. thesis, Univ. of Br i t i s h Columbia. 74 pp. Neish, I.C. 1970. A comparative analysis of the feeding behavior of two salamander populations in Marion Lake, B.C. Ph.D. thesis, Univ. of B r i t i s h Columbia. 108 pp. Royama, T. 1970. Factors governing the hunting behavior and selection of food by the great tit (Parus major L.). J. Anim. E c o l . 39: 619-668. Sandercock, K. 1969. Bioenergetics of the Rainbow trout (Salmo gairdneri) and the Kokanee (Oncorhynchus nerka) population of Marion Lake, B r i t i s h Columbia. Ph.D. thesis, Univ. of Bri t i s h Columbia. 165 pp. Savage, R.M. 1962. Ecological, physiological, and anatomical obser-vations on some species of anuran tadpoles. Proc. Zool. Soc. Lond. 112: 467-514. Tinbergen, L. I960. The natural control of insects in pine-woods. I. Factors influencing the intensity of predation by songbirds. Archiv. Neerlan. Zool. 13: 266-343. Turner, F.B. 1962. The demography of frogs and toads. Quart. Rev. Bi o l . 37: 303-314. 56 APPENDIX 1 This appendix is the text of a short note submitted to the Canadian Journal of Zoology describing my observations of the breeding chorus of R. aurora in Marion Lake. I include it with the thesis because the observations bear on the distribution of eggs, the numbers of frogs required to maintain the population (see page 4 ), and the possibility of social mechanisms as a regulatory factor for the population. 57 INTRODUCTION Recent studies on the breeding aggregations of certain temperate zone frogs have revealed the existence of t e r r i t o r i a l and agonistic behav-ior (Emlen, 1968) and an "effective" breeding population which is sub-stantially less than the observed total population (Merrell, 1968). The importance of these observations for theories of population regulation and evolution suggests that such studies should be extended to other populations of anurans. This paper reports observations made on the breeding aggrega-tions of the red-legged frog Rana aurora during the springs of 1969-1970 in Marion Lake, B r i t i s h Columbia (see Efford, 1967, for a more detailed description of the lake). Special attention was paid to the movements of male frogs during the breeding season, and to the relationship between the numbers of potentially breeding adults and the actual numbers of egg masses produced by the population. METHODS Licht (1969) has described some aspects of the breeding behavior of R_. aurora at several locations in southwestern B r i t i s h Columbia. The most striking feature of the breeding chorus is that the males call under water, often at depths of several meters. Therefore, I made my observa-tions from a rowboat equipped with tractor headlights, which shone vertically down into the water. On several nights I visited the submerged weedbeds (see F i g . lb) where most of the calling and egg-laying took place. The calling males were captured with dip nets, individually toe-clipped and released immediately. 58 FIGURE l a . Distribution of egg masses in Marion Lake in 1969 and 1970. The number in the upper left-hand corner of each quadrat indicates the number of egg masses found in that quadrat in 1969; the lower right-hand number indicates masses found in 1970. FIGURE l b . Map of habitat types found in Marion Lake. F o r key, see page 3 in the main body of the thesis. Open mud; O- lm. 1 l ;1-2m. l l l l l ; 2 - 3 m 3+m. E§±±±1 Nuphar po l ysepa la | L Submerged vege ta t ion ; Potamogeton natans & Equisetum spMMM ; Dry areas 59 P r i o r to capture, the spatial relationship of males was observed. Toe-marking allowed me to estimate the population of males by means of a mark-recapture procedure, to follow their movements from weedbed to weedbed, and to record the reproductive activities of males later captured in amplexus. After the frogs left the lake, I determined the numbers and distribution of egg masses by systematically searching the entire lake. 2 To aid in this search, I divided the lake up into 1000m quadrats by placing an anchored buoy at the corners of each quadrat. The shallow areas (less than l m in depth) were searched from a boat and each egg mass was marked with a stake. I searched the deeper areas by diving and a systematic pattern was assured by following a boat which was rowed back and forth through the quadrat. The clarity of the water and repeatability of counts in selected areas suggest that few clutches were missed. I may have altered the behavior of the frog population, but I doubt it because: (1) only 10-15% of the estimated population was captured and marked during the study each year, (2) males which were captured and marked later were found in amplexus, (3) the number of egg masses in 1969 and 1970 did not differ significantly from the number found in 1968, when the breeding aggregations were not disturbed, (4) only a short period was spent each evening in any given area of the lake. Hydrophone observa-tions made by Licht (personal communication) indicated that at least some of the frogs within a weed bed continued to cal l even when the boat was pr e s ent. R ESULTS AND DISCUSSION Licht's (1969 ) descriptions of R. aurora breeding aggregations are applicable, with two exceptions, to the Marion Lake population. Licht 60 states "that water temperature of 7.0 C is sufficient for spawning and that once this minimum is reached, most eggs are deposited." At Marion Lake frogs call and breed at water temperatures between 4° and 5°C, and many times males were observed calling under the ice. The number of males calling appears to be slightly reduced on these cold nights, but o egg-laying s t i l l occurs. The eggs remain in water of less than 5 C for many days during the spring. Egg-laying in Marion Lake occurs at a l l depths from a few inches to 5m, the maximum depth found in the lake, and not just "up to 5 feet", as stated by Licht. The population estimates for calling males are shown in Tables I and II. Females are captured much less frequently than males (less than 10% of frogs captured were females) and a l l females captured were in amplexus. Thus, females must mate almost immediately upon entering an area where males are calling, and unlike the males, they do not remain in the lake after copulation. Consequently, no population estimate was attempted for the females. The substantial difference between the numbers of egg masses laid and the numbers of males estimated (Tables I and II) could be due either to an unequal sex ratio (with more males than females) or to some mechanism by which a large fraction of the adult population of both sexes is excluded from breeding. In either case, the "effective" breeding population, that is, the number of adults actually contributing to the next generation, is less than would be supposed from an estimate of the adult male population. Many males clearly fai l to breed. M e r r e l l (1968) gives a good discussion of the importance in evolutionary theory of distinguising between the number of potentially breeding adults in a population and the actual number contributing to the next generation. The small sample (7 females, 5 males) of frogs caught in the vicinity of Marion Lake during the summer of 1969 is not adequate to establish a sex ratio for the population. M e r r e l l in his study found equal Capture Date 1 2 3 4 5 6 7 8 9 1 27 ' - 4 - 2 - 1 1 >ture 2 42 5 2 1 2 - 1 ->ture 3 78 1 2 4 - 1 1 Maximum Likelihoo d U 4 j 61 1 2 - - - estimate of breedin CO o 5 34 - - - - males = 1770 + 280 > 6 37 . - 1 CO u PH 7 14 - - Total egg clutches o CO 8 • 36 - = 618 Dat 9 15 # Marked 0 9 3 6 8 1 4 2 # Unmarked 27 42 78 61 34 37 14 36 15 # Released 27 42 8 7 64 40 . 45 15 40 17 Total Marked in Population 27 69 147 208 242 279 293 329 344 Pop. estimate - - 667 3136 1386 1361 4 1 8 5 2980 2790 T A B L E I. Population estimate of breeding males for 1969. The successive population estimates are shown in the bottom row. An overall estimate by the maximum likelihood method (assuming no mortality or emigration of males during the breeding season) is shown in the right column. Capture Date 1 2 3 4 5 6 7 8 9 1 54 3 2 1 0 0 2 0 1 4) 2 77 1 0 0 2 0 1 0 +e 3 a O 4 55 i 1 0 0 0 2 0 Maximum Likelihood 1 59 1 0 1 0 2 Estimate of breeding 19 0 0 0 0 males = 3600 + 775 S £ 6 t . 11 0 0 0 H PH 7 m 48 0 0 Total egg masses = 6 ; 8 24 0 & 9 39 Total Marked 0 3 3 2 1 2 3 3 3 Total Unmarked 54 77 55 59 19 11 48 24 39 Total Released 54 81 58 61 20 13 51 27 41 Total Marked in Population 54 135 193 254 274 287 338 365 406 Pop. estimate - 1485 2610 5886 5080 1781 4879 3042 4988 T A B L E II. Population estimate of breeding males for 1970. The successive population estimates are shown in the bottom row. An overall estimate by the maximum likelihood method (assuming no mortality or emigration of males during the breeding season) is shown in the right column. 63 sex ratios aways from the breeding ponds, but many fewer egg masses than males in the breeding populations. He did not mention the possibility that females do not breed every year. If females required more than one year to produce mature eggs, then there would be more males than females in the breeding pond, but not at other times of the year. There is no evidence on this point from my study. However, Licht (pers. comm.) has found that in another population of R. aurora near Vancouver, at least some of the females develop a new clutch of eggs each year. The frogs recaptured in my study were often found in locations distant from their place of original capture. In fact, these movements often resulted in an animal moving from one type of weedbed to another, e.g. from Potamogeton to Isoetes, and from one depth to another. Of the 34 recaptures in 1969, 23 were found in different weedbeds from the one .in which they were fi r s t captured. Three of these frogs moved less than 100m, fourteen moved 100 to 300m, and six moved more than 300m. In 1970, of the twenty-four recaptured, thirteen moved less than 100m, two moved 100-300m, and three others moved more than 300m. Twenty-five males were captured in amplexus in 1969. Three of these were in amplexus after previous capture in another weedbed, four frogs were recaptured up to 3 days after having been marked while in amplexus, and one frog was captured in amplexus twice with a different female each time. Thus, frogs which move away from a given area can st i l l be successful in breeding. Frogs marked during the 1969 season were recaptured in 1970. Of 66 such recaptures, 38 had returned to the weedbed where they had been captured a year earlier. Nine others were less than 100m away from the area of previous capture. Thus, males seem to return to a particular area of the lake. No displays of aggressive behavior between males were noted. The spatial distribution between frogs was not the relatively even spacing which one associates with t e r r i t o r i a l behavior in other vertebrates. Often 64 several males were clustered within inches of each other, while others were several feet apart. On one occasion five males were captured in amplexus with one female. Such interference with the mating of an individual would not be expected in a t e r r i t o r i a l species. Indeed, protec-tion from interference with mating is one of the functions commonly-attributed to territoriality (Hinde, 1956). Although males mostly aggregated in the major weedbeds, isolated males were often observed near logs, sticks, or l i l y pads in the open mud areas. The fact that egg masses were often found in these l o c a l -ities indicates that these isolated males must often breed successfully. The above observations suggest the following conclusions: (1) males of Rana aurora do not defend or remain in a specific territory throughout the breeding season. Males which move from one area to another can sti l l be successful in mating. However, there is a tendency for males to return to a given area of the lake each year, (2) male frogs do not leave the lake immediately after mating and may mate more than once, (3) only a fraction of the potential male breeding population appears to reproduce. Thus, there may exist in some R. aurora populations a social or physiological mechanism which regulates reproductive behavior, but this mechanism is not the classic type of an aggressively defended territory. 65 A C K N O W L E D G E M E N T S I thank Iain Neish and i l l the o t h e r g r a d u a t e students who spent chilly nights on Marion Lake h e l p i n g m e catch f r o g s . Neil Gilbert offered advice on statistical techniques. This work was s u p p o r t e d by C a n a d i a n International Biological P rogram Grant #656218 to Dr. I.E. E f f o r d , 66 R E F E R E N C E S Efford, Ian E. 1967. Temporal and spatial differences in phytoplankton productivity in Marion Lake, B r i t i s h Columbia. J.Fish.Res. Bd.,Can. 24(11): 2283-2307. Emlen, S.T. 1968. T e r r i t o r i a l i t y in the Bullfrog, Rana catesbeiana. Copeia, 2: 240-243. Hinde, R.A. 1956. The biological significance of the territories of birds. Ibis, 98: 340-369. Licht, L . E . 1969. Comparative breeding behaviour of the red-legged frog (Rana aurora aurora) and the western spotted frog (Rana pretiosa pretiosa) in south-western B r i t i s h Columbia. Can. Jour. Zool. 47: 1287-1299. Me r r e l l , D.S. 1968. A comparison of the estimated size and the "effective size" of breeding populations of the Leopard frog Rana pipiens. Evolution, 22: 274-283. 67 APPENDIX 2 These maps indicate the mean number of tadpoles captured per trap in each quadrat during 1969 . The figures presented are aver-ages for ten trap-days. May 21-June 5, 1969 T o t a l Tadpoles/day = 84.5 June 6-June 16 T o t a l Tadpoles/day = 91.1 June 17-June 29 T o t a l Tadpoles/day = 50.8 June 30-July 11 T o t a l Tadpoles/day = 28.6 J u l y 14-July 25 T o t a l Tadpoles/day = 36.0 J u l y 27- August 7 T o t a l Tadpoles/day = August 8-August 21 T o t a l Tadpoles/day = 6.6 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items