umvflh.‘ . . .. . . . . . . .. u: ..uvvm. , 4..rn.......4\»1 .. . ., . . . 3.-.. . , a) «4.414.... 4&9... . . " ‘1, ll‘.‘ . \ ‘I'14‘.",';'..:; ‘ I51" 2-: . -2 . F» v 4 34.4.... . . 4.4 4m? ,4... ..I4.. 33:" ...g...sl -. .1I . o>t ._ .7, 4.4.4.43. . . . ....x....w..4.,.4.. 4%,”: . f T AT .leD. . .mnr ~14» NF ".3“ . w: . . . . . . . "MLWVMWVIUIPLHL1 A”? . . . . _ J . - . 4-. , rwkmkfifiumfii . . a-.. L ?I«»I.‘t\...u.«r. . y. Nonmflunuhunlt Jl!1\ PW“: ' *a .- . "P .EJ . u . .4 3:2 . fifiéfi. ..u = . .541 ..v vfiC‘ r. . . .5444. Jet, .. .-. . . awn}. :Ls ,\ Iww . . , ..D. . . .7. A .l .-.. 3 . ( .m. . uuyVfiLfifi.3fif . .. , x . ,1u1muM...w.udnl..r.wflh.ul¥v. ~ ..C? . . . . . . . ., . . Innfl‘ 3%... x. . . . .F snfkyr . 4 . . . .. .u‘ s , .. ...n|u..mwml.wu . 4. . . . . H. ,. V .. .I.._. L... 43.4?“ . . _ 2.41;} I .Mmmmwmmm gamma?” ....&.¢.§m , WU“ \H||\\\|\\\J|U\\|\|\illMW“WJIM 31293 104483 1 I ,r ‘J 4.44 ' \AA. J’J a J}. _ . _ ‘3 Vm‘ m isn'- in i? ii} ~-;. 4.1 mg. UETW erg-i. L —.—— This is to certify that the thesis entitled Foraging Efficiency and Size-Class Competition in the Bluegill Sunfish (Lepomis macrochirus) presented by Gary George Mittelbach has been accepted towards fulfillment of the requirements for Ph .D . . Zoology degree in 2...g§g.).....4 Major professor Date 425’ 5 big} 315 0-7 639 “5' .445 ..,; l ({"".“\\\ .. 5.. if if?“ r 6-. .9... .19).. :4} 5995‘. $5 .113.“ ' OVERDUE FINES: 25¢ per day per item RETURNING LIBRARY MATERIALS: Place in book return to remove charge from circulation records FORAGING EFFICIENCY AND SIZE-CLASS COMPETITION IN THE BLUEGILL SUNFISH (Lepomis macrochirus) By Gary George Mittelbach A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY W. K. Kellogg Biological Station and Department of Zoology 1980 fi—vvq_‘h.—VWA‘__ _ .. . ._ __ __ V. _ . ABSTRACT FORAGING EFFICIENCY AND SIZE-CLASS COMPETITION IN THE BLUEGILL SUNFISH (Lepomis macrochirus) By Gary George Mittelbach A foraging model was developed to predict the optimal diet breadth and maximum energetic intake of a given-sized fish foraging in each of three aquatic habitats; the open water, vegetation and bare sediments. Mbdel parameters of prey encounter rates and prey handling times were quantified as functions of fish size, prey density, and prey size through a series of laboratory feeding experiments using the bluegill sunfish (Lepomis macrochirus). Results of these experiments show both searching ability and prey handling efficiency to increase with in- creasing fish size. Predictions of prey size selection and optimal habitat use based upon maximizing energetic gain were then examined in a small, Michigan lake for three size classes of bluegills. Prey abundances were measured in the open water, vegetation and bare sediment habitats during the months of May through August 1979 and size-frequency distributions of prey available in each habitat determined. Bluegills 2>100 mm SL were highly size selective in their feeding and their diets closely matched predictions of an optimal diet model. From two estimates of relative prey visabilities I show that these fish selected larger prey items ‘than would be predicted if prey were consumed "as encountered". Habitat use of large bluegills was also shown to maximize foraging return as Gary George Mittelbach fish switched from utilizing vegetation prey to utilizing open water zooplankton as relative foraging profitabilities in the two habitats changed across the summer. Bluegills <<100 mm SL were restricted in their habitat use, remaining in or near the vegetation despite demon- strated increases in foraging return available in the open water habitat. Size-related predation risk apparently accounts for differences in habitat use between bluegill size classes. The consequences of size- specific foraging abilities and predation risks to the outcome of competition in size-structured populations are discussed, with parti- cular attention paid to the problem of "stunting" in fish. To George, Dorothy, Kay, and Lydia ii ACKNOWLEDGMENTS First and foremost, I thank my major advisor, Earl Werner, for his guidance, support, and unselfish sharing of insights and information. His conduct of science was an inspiration and his friendship a pleasure. This study profited greatly from comments by Donald Hall, Donald Beaver, Richard Marritt, David Hart and Katherine Gross on earlier drafts of the thesis and from discussions with James Gilliam, Cader Olive, Leni Wilsmann and the other members of the Ecology group at Michigan State. The rigors of field work were often made enjoyable by the assis- tance of James Gilliam, Katherine Gross, Leni Wilsmann and George Mittelbach. I thank John Gorentz for expert computer programming and Robert Wetzel and William Crumpton for kindly providing unpublished data on Lawrence Lake. Charlotte Seeley and Art Weist helped in many ways, for which I'm grateful. Finally, I thank William Platt and Robert Bovbjerg for the beginning and my wife, Kay, for it all. This work was supported by NSF grants DEB 7620106 and DEB 7824271 to Drs. Werner and Hall. iii TABLE OF CONTENTS LIST OF TABLES 00...... OOOOOOOOOOOOOOOOOO O. ........... LIST OF FIGURES ........ I. O. O O 0.. O. ..... C 00000 0 DEFINITION OF SYMBOLS USED IN THE TEXT . .............. CHAPTER 1: FORAGING EFFICIENCY AND BODY SIZE IN BLUEGILLS ....... INTRODUCTION OOOOOOOOOOOOOO ........ O ........... 0...... AFORAGING MODEL 0......0.0.0.0...OOOOOOOOIOOOOOOOOOOO IABORATORY EDERIMENTS O... ..... OOOOOOOOOIOOOOOOOOOOOO MethOdS .0...0.0.0.000...O...OOOOOOOOOOOOOOOOOOOOOO. Results .......... .............................. .... AFIELD STUDY 0...... ....... O. ...... 0.000.... ...... 0.. Site ..OOOOOCOIOOOOOOIOOOOOOOOOOOOOOO ..... 0.00.00... Bluegill Diet and Habitat Use ......... ..... ........ Habitat Profitabilities and Use .. .................. DISCUSSION ........................................... CHAPTER 2: SIZE-CLASS COMPETITION ............................... INTRODUCTION 0... OOOOOOOOOOOOOOOOO OOOOOOOOOOOOOOOOOOOO INTRASPECIFIC COMPETITION ..... .............. ......... APPENDIX A: DISTRIBUTION OF BODY SIZE IN THE PREY OF FISHES ...... APPENDIX B: CALCULATION OF PREY ENERGETIC CONTENTS .... ........... LITERATURE CITED ................................................. iv 0000 24 24 29 40 49 58 58 58 69 97 98 Table A1 A2 A3 A4 A6 A7 A8 A9 LIST OF TABLES Bluegill Swimming Speeds ............................. Handling Time Relations for Bluegills Feeding on Each of Three Prey Types .......................... Prey Encounter Rate Equations ....... ....... .. ........ Prey Size-selection by Large Bluegills ...... . ..... ... Distribtuion of Daphnia in Lawrence Lake ............. Size Distribution of Invertebrate Prey in the Vegetation of Lawrence Lake .......................... Size Distribution of Invertebrate Prey in the Bare Sediments of Lawrence Lake ............... ....... Average Size Distribution of Invertebrate Prey in the Open water of Lawrence Lake .......... ....... .. Parameters of the Normal Distribution Fit to the Logarithm to Base e of Prey Length in Three Habitats ......OOOOOOOOOO00.0.0.0... ......... .0. Seasonal Pattern in the Density of Prey Organisms =’3.0 mm Body Length in the Vegeta- tion and Sediment Habitats O O O O O O O O O O O O O O O O O ...... O O O 0 Mean Number and Dry Weight (per m3) of Major Prey Groups Found in the Vegetation .................. Total Dry Weight (mg/m3) and Mean Size (mg dry wt.) of Vegetation Prey (exclusive of gastro- pods) ................. . .............................. Total Dry Weight (mg/m3) and Mean Size (mg dry wt.) of All Vegetation Prey ................. ......... Mean Number and Dry Weight (per m2) of Major Prey Groups Found in the Sediments .. ................. Page 16 20 36 48 74 75 76 77 85 87 88 89 90 Table A10 All Page Total Dry Weight (mg/m2) and Mean Size (mg dry wt.) of All Sediment Prey ............................ 91 Total Biomass (mg dry wt./m2) of Invertebrate Prey Found in the Littoral Habitats of Small Lakes by Various Researchers ......................... 94 vi Figure 6 a,b,c 10 LIST OF FIGURES Increase in Capture Rate with Experience for Bluegill Sunfish Feeding on Chironomus and Baal-11118000000000.0000... ........ O ..... 000...... ..... Prey Handling Time as a Function of Relative Prey Size in Each of Three Habitats ..... . ........... Laboratory Prey Encounter Rate (E t 1 SE) in Relation to Prey Density, Prey Length and Fish Length 0000000000 0...... ..... 00...... ........... Rate of Prey Encounter as a Function of Prey Density and Prey Length for Three Sizes of Bluegills (20, 60, 11011“) O O O O O O O I O ..... O C O 0000000000 Bathymetric Map of Lawrence Lake, Michigan ...... Size-frequency Distribution of Prey Available in a Habitat, Predicted Optimal Diets, and Actual Diets of Three Size Classes of Bluegills ... .............. . ......................... Average Foraging Return (En/T) of Three Sizes of Bluegills (125, 75, 30mm) as a Function of Prey Sizes in the Diet ........................... Linear Relationships Between Predicted Foraging Return from a Habitat and Stomach Contents (mg dry wt.) of Bluegills Foraging >'80‘2. of Their Diet from that Habitat for the Open Water (x) and Vegetation (o) Habitats ...... Seasonal Pattern in Predicted Habitat Profit- ability (left) and Actual Habitat Use (right) for Three Size Classes of Bluegills .......... ....... Hypothetical Distribution of Net Caloric Gain from the Diets of Two Bluegill Size Classes Utilizing the Same Prey Resource .................... vii Page 13 15 19 23 27 32 39 43 46 60 Figure Page 11 Hypothetical Distribution of Net Caloric Gain from the Diets of Two Bluegill Size Classes Feeding at High Resource Lavels (upper panel) and Low Resource Levels (lower panel) ..... ...... ... 63 12 Calculated Net Caloric Intake (En/T) for Bluegills Feeding on Three Different Distri- butions of Vegetation Prey ......................... 66 A1 Distribution of Prey Sizes (mm body length) in the Vegetation Habitat Averaged Over the Season O00.0.0000.........OOIOOOOOOOOO0.00.... ...... 80 A2 Distribution of Prey Sizes (mm body length) in the Bare Sediment Habitat Averaged Over the Season . ............ .. ...... ........... ......... 32 A3 Distribution of Prey Sizes (mm body length) in the Open Water Habitat Averaged Over the season 0.000............OOOOOOOOOOOOOOO.... ...... 0.0 84 viii e(i) Hm B(i) DEFINITION OF SYMBOLS USED IN THE TEXT Energetic content of prey size i (cal) Handling time for prey size i (sec) Encounter rate for prey size i (number/sec) Prey length (mm) Assimilatible fraction of energetic content Energetic cost of searching (cal/sec) Energetic cost of handling (cal/sec) Fish standard length (mm) Swimming speed (m/min) Water temperature (°C) Fish live weight (g) Prey density (units appropriate to each habitat) Prey dry weight (mg) Net energy gained while foraging (cal) Time spent foraging (sec) Search time (sec) ix CHAPTER 1 FORAGING EFFICIENCY AND BODY SIZE IN BLUEGILLS INTRODUCTION At the heart of the rapidly expanding theory of optimal foraging is the goal of predicting the diet and habitat use of organisms as a function of resource availability and utility to the consumer. To date, this theory has largely been concerned with the processes of individual consumer choice, however, its potential applicability to the study of community level questions is large (e.g. Werner 1977). Clearly, the ability to predict the diet and habitat use of organisms in nature represents a potential foundation upon which more mechanistic theories of competition and species packing may be built. The eventual usefulness of optimal foraging theory in unraveling the nature of community structure will depend in large part on the degree to which the theory can quantitatively predict forager behavior in the field (Pyke et al. 1977). There have been relatively few field studies of optimal foraging (Goss-Custard 1977, Davies 1977) and only the work of Belovsky (1978) on moose has actually attempted.to predict an animal's diet in nature through optimal foraging criteria. This slow progression from theory to empirical test is understandable as major difficulties stand in the way of extending optimal foraging models to the field. Not the least of these difficulties is the need to measure the availability of various prey types to the consumer and the quantification of actual resource use by the foraging animal. Crucial in the testing of any foraging model is a means by which prey abundances, measurable in the field, can be translated into actual rates of prey encounters by a forager. This chapter develOps an optimal foraging model for the bluegill sunfish (Lepomis macrochirus) and examines its predictions in a natural environment. Since bluegills capture prey individually and swallow them intact, the size and type of prey consumed by a foraging fish are easily determined from gut contents. Further, the close association between prey type and habitat which exists for many groups of aquatic animals (e.g. open water zOOplankton, vegetation dwelling insects, etc.) permits the determination of a fishes' habitat use through dietary analysis. In these respects, the bluegill represents an ideal organism for the study of diet selection and habitat use in nature. The size-distributed nature of fish populations also allows an examination of the functional relationships between body size and foraging efficiency so that relative differences in resource utilization can be related to potential competitive interactions between size classes. It has become increasingly clear that the development of a predictive theory of community structure in fish and other size- distributed species will depend on a thorough understanding of size- specific interactions (Neill 1975, Werner 1977), and Chapter 2 represents an initial attempt at utilizing foraging theory to explore questions relating to size-class competition in fish. In this study I utilize a series of laboratory foraging experiments to quantify foraging parameters of prey encounter rates and handling times as functions of fish size, prey size and prey density. These experiments were designed to simulate the physical structure and prey types found in each of three aquatic habitats (the open water, bare sediments and vegetation) which previous studies have demonstrated to be key divisions of the bluegill's natural environment (Werner and Hall 1976, 1979, Werner et a1. 1980). Seasonal prey abundances were then measured in these three habitats in a small Michigan lake and a discussion of the distribution and dynamics of this prey resource is presented in Appendix A. Prey abundances were translated into prey encounter rates and profitabilities using results of the laboratory experiments. A foraging model was then used to make predictions of optimal diet breadth and habitat use for a given-sized fish and these predictions were compared to the actual diets and habitat use of three size classes of bluegills occurring in the lake. A FORAGING MODEL To predict the range of prey sizes eaten and maximum net energetic intake for a given sized bluegill foraging in habitat type j an optimal foraging model similar to those developed by Charnov (1976), Werner and Hall (1974), Pearson (1976) and others was formulated. Briefly, let E(ij) = A e(ij) - Ch H(ii) where A = assimilable fraction of energetic content, 3(ij) = energetic content of prey size i in habitat type j (sec), and Ch = energetic cost of handling prey (cal/sec). Thus E(ij) represents the net energetic gain from eating prey size i. If prey are consumed over time T (sec) which is divided into time spent searching for prey (P) and time spent n handling all prey 2 i=1 prey size i encountered per second of search, then the net rate of B(ij)H(ij), where 3(11) equals the number of energy intake (En/T) from habitat j is n P 2 (3(ij) E(ij)) - c8 p En/T = i=1 or n P + P 2 B(ij) H(ij) i=1 n ,2 (”(15) E(ij) ) - Cs En/T = 1:1 (1) n 1 + Z 3(15) "(15) i=1 where C8 = energetic cost of searching (cal). The optimal diet for a predator (i.e. that subset of available prey sizes which maximizes its net rate of energetic intake (En/T)) can now be determined by ranking prey sizes from highest to lowest E(ij)/H(ij) and then adding prey sizes to the diet until the ratio En/T is maximized. A proof of this theorem can be found in Charnov (1976). The inclusion of the energetic costs of searching and handling in equation 1 is of course necessary when comparing the energetic intakes of foragers of differing body size. In fish and many other organisms, searching ability and efficiency at handling prey will also be functions of body size. Therefore, in order to predict the diet and net energetic intake of different size classes of the bluegill it is necessary to determine the functional relationships between predator size and the parameters in expression (1). For fish, the assimilatible fraction of energy ingested (A) appears to be independent of body size (Elliot 1976) and a value of 0.7 is appropriate for most prey (Ware 1975, Elliot 1976). The energetic costs of searching and handling (Cs and Ch) can be estimated for bluegills using the data of Wdhlschlag and Juliano (1959). These investigators measured the oxygen consumption of bluegills as a function of body weight (W), swimming speed (8) and water temperature (t) using a rotating respirometer suspended in a lake. Oxygen consumption was converted to calories expended by assuming 1 mg oxygen consumed equals 3.25 calories (Elliot and Davison 1975). Since other parameters of the foraging model were determined as functions of fish standard length, the measure of fish size contained in Wdhlschlag and Juliano's energetic equation was converted from live weight in g (W) to standard length in mm (L) using a length-weight relationship for bluegills of W = .000026 L3-043 (unpublished data). The resultant equation: Log 0 = -7.8512 + 2.5847 Log L + 0.0142 s + 0.0198 c (2) was used to calculate Cs and Ch- CS was calculated using the swimming speed exhibited by a fish while searching. Ch should include the costs of masticating prey and swallowing as well as any movement which occurs. At present, no measures of the costs of processing prey exist for bluegills. Since most of the bluegill's prey are small, soft bodied invertebrates which the fish swallows whole, it is probably reasonable to assume its energetic expense of chewing and swallowing is small relative to that of swimming. Therefore, Ch was approximated using equation 2 and the swimming speeds exhibited by fish while handling prey. Swimming speeds of bluegills engaged in searching and handling were 6 determined from a series of laboratory foraging experiments designed also to measure prey encounter rates and handling times (a complete description of the laboratory system can be found in the following section). Average swimming speeds were calculated for each experiment by dividing the total distance traveled while searching or handling by the total time spent in the activity. Table 1 lists the swimming speeds (R':_l SE) of fish while searching or handling prey in laboratory representations of vegetation and sediment habitats. As would be expected, fish swam slower while handling prey than while actively searching. Also, fish size and habitat type had a significant effect on swimming speeds (two-way ANOVA (searching); fish size F = 48.3, p < .001; habitat type F = 10.5, p < .002; two-way ANOVA (handling); fish size F = 15.9, P < .001; habitat type F = 3.2, P > .05; n = 96). However, the total range of swimming speeds observed across all fish sizes or habitats was small and over this range in swimming speeds only minor changes in energetic costs occur. For example, a 50 mm bluegill at 18°C expends .00081 cal/sec swimming at 1.0 m/min and .00087 cal/sec swimming at 3.0 m/min (by eq. 2). Therefore, to simplify calculation of Optimal diets, average swimming speeds taken across all bluegill sizes and both habitats were used to calculate the energetic costs of searching and handling. Bluegill swimming speeds while searching and handling were 1.7 :_.08 and .08 :_.05 m/min (x :_1 SE, n = 96) respectively; values which fall within the range of swimming speeds observed for feeding fish in nature (Ware 1975). The energetic content of prey size i (9(i)) was determined by (:onverting prey lengths to dry weights and then multiplying by the appropriate caloric equivalent (Cummins and Wuycheck 1971). Specific on. + m.o co. + o.~ 50..“ s.o sc..H o.a wcm~vcwn wcmsuuwwm Am Ea mm 03..“ m.o ea. H.z.a mamdvcm: mamsuuwmm Am 55 mm Na..u w.o as. + m.~ no..u s.c as. + m.~ acmdmcwc magnuumow am as ms «a..H s.~ ca. u.~.~ we. “.m.~ oz..u s.~ acadvcm: wcmnoummm Am SE m¢ me. + 5.0 dfi. + ¢.d so..u c.o mo. H.m.~ Mawdvcm: wcmnuumwm Am 55 cm “women: acmemvmm mo..H ~.o oc. H.c.~ as. “.m.~ as. H.~.N McMchms wcmsuumom Am as no uwumnmz cemumumaw> ucoemwmm paw cemumuomo> osu mo mCOMumucmmmuamu mucumuoamfi am zone use mcmsoumom maawmwsan you Amm a “w amsxe cm mummam wCMEEwsm coo: so. + m.o Na. + «.3 m.0a no..u s.o mo. “.m.a z.zs waddvcm: wcwsoummm AEEV :uw:w~ zoum em as cam cameos swam am..u w.o s~..H m.~ z.ma mu..n m.o am..n o.a m.m w:Mdvcm£ mamcuummm AEEV Luwcmfi xwum mm as odd saucofi same .mumumam: mo wmumm osu womavcmz cam .mvowam acmEEm3w Hfimmmaam 4 3an 8 length-weight regressions and caloric equivalents used are detailed in Appendix B. To quantify prey encounter rates (B) and handling times (H) a large number of laboratory experiments were performed using various combinations of prey size, fish size and prey density. These experiments were designed to simulate the structure and prey type found in each of three distinct aquatic habitats; the open water, bare sediments and vegetation. Experimental field studies (Werner and Hall 1976, 1979, Werner et a1. 1980) have shown that the bluegill treats these habitats as distinct divisions of the environment; each habitat differs markedly in physical structure and associated prey and therefore requires unique modes of foraging by the fish. Within a habitat type, prey were characterized by body size and density since the responsiveness of fish to both parameters is well recognized (Ivlev 1961, Werner and Hall 1974, Eggers 1977). LABORATORY EXPERIMENTS Methods In the laboratory, realistic approximations of each of the three habitat types were constructed as follows: Habitat Laboratory Open water structure = open aquaria prey = zo0p1ankton (Daphnia pulex) Bare sediments structure = aquaria containing 4-5 cm layer of marl sediments prey = midge larvae (predominantly Chironomus plumosus) Vegetation structure = anchored, live Elodea (100 plants/m2) prey = damselfly naiads (Coenagrionidae) The particular prey types chosen commonly occur in each habitat in nature and in general represent the degree of crypsis and mobility characteristic of prey in the habitat (i.e. tube-dwelling midges in the sediments, free-swimming zOOpIankton in the open water, clinging damselfly naiads in the vegetation). The physical structure of each habitat was represented by natural substrates where apprOpriate. The general format for each experiment involved dividing a 214 liter aquarium into two unequal sized sections, 26 liters and 188 liters, by a removable glass partition. The larger volume contained a prey distribution of known composition (prey type, size and density). Daphnia were size-sorted by washing them through a series of U.S. standard sieves and a known number of a specific size class were then introduced into the larger division of the aquaria immediately prior to the initiation of a feeding experiment. Two sizes of Daphnia (2.20 :_ .04 and 1.14 :;.02 mm body length) and five prey densities (0.1, 0.5, 2.5, 5, 15 individuals/l) were used. For the sediment habitat Chironomus larvae were sized, counted, and then distributed over the surface of the larger section of the aquaria. Midges were introduced during the late afternoon of the day prior to an experiment as this permitted the larvae to burrow into the sediments. Two midge sizes (19.53 1 .46 mm and 11.07 :_.34 mm body length) were used at densities of densities of 50, 300, and 1000 individuals/m2. Experiments simulating the vegetation habitat used Coenagrionidae naiads (9.51 1 .27, 13.15 I .33, and 22.30 :_.28 mm body length) at densities of 38 and 192 individuals/m3 for large damselflies, 192 individuals/m3 for medium 10 damselflies and 192, 385 and 1,538 individuals/m3 for small damselflies. Damselfly naiads were introduced into an aquarium the day prior to an experiment. All prey were either collected from local ponds or cultured in the laboratory. After the introduction of prey a fish which had been starved for 24 h was placed into the smaller section of the aquarium and allowed to acclimate for 30-60 minutes. An experiment was then initiated by removing the glass paritition and allowing the fish to feed. Data recorded were handling time for each prey captured, time between strikes, success or failure of a strike, and the distance traveled by a fish while handling prey and while searching (excepting the plankton experiments where no distance measurements were taken). Handling times for Chironomus and Coenagrionidae were measured as the time from prey capture until the reinitiation of search. Handling times for Daphnia were too short to be measured accurately in this manner. For these prey, handling times were determined by examining the time per prey item captured as a function of prey density. The asymptotic value of time per prey item as density increased was used as a measure of baseline or minimum handling time (Ware 1972, Werner 1977). Experiments were performed over short time periods (30 sec. for Daphnia, 3 min for Coenagrionidae, and 10 min. for Chironomus) to minimize the effect of prey depletion. Maximum prey depletion was < 25% of total excepting the lowest Daphnia and Coenagrionidae densities where prey depletion was about 45%. Over 500 feeding experiments were conducted. Ten bluegills (33-109 mm SL) were used in the open water experiments, nine (21-115 mm SL) in the vegetation experiments and six (21-109 mm SL) in the sediment 11 experiments. Each combination of prey size, fish size and prey density was replicated from 3 to 6 times. Water temperatures ranged from 17-22°C. Results Naive bluegills exposed to these laboratory habitats show a marked increase in prey capture rate with successive feeding experiments (Figure 1). Since I wished to determine the maximum energetic return available to a fish in a given habitat, measurements of prey encounter rates and handling times were taken after fish were experienced in a habitat and prey capture rate had levelled off (generally after 6-8 trials). Figure 2 shows handling time per prey item (H) plotted as a function of relative prey size (prey length/fish length) for prey in each of the three habitats. For each prey type handling time increases exponentially above a critical ratio of prey length/fish length (i/Lcrit)- Below this ratio handling time remains approximately constant. These results conform to the general relationship postulated by Schoener (1969) and are in close agreement with the data obtained by Werner (1974) for bluegills and green sunfish (Lepomis cyanellus). For each prey type an exponential equation was fit by least-squares regression to all values greater than the minimum handling time per individual prey. The fitted equations, minimum handling times, and critical ratios of prey length/fish length are listed in Table 2. It is interesting to note the high minimum handling time for fish feeding on Chironomus (9.6 sec compared to 1.0 sec for Daphnia). When a fish captures a midge larvae it engulfs a mouthful of sediment along with the prey. The cost of separating the prey from these sediments is 12 Figure 1. Increase in Capture Rate with Experience for Bluegill Sunfish Feeding on Chironomus and Daphnia. One exper- iment was performed per day. Excepting one point, points represent the mean i 1 SE for 2-6 fish/experiment. 13 H muswwm .OZ ._.Zm<<_~_mmxm_ m m o m380cotgu ON om cow BLVH 'XVW lNBDHHd 14 Figure 2. Prey Handling Time as a Function of Relative Prey Size in Each of Three Habitats. Prey used were Chironomus larvae for the sediments, Coenagrionidae naiads and Daphnia for the vegetation, and Daphnia for the open water. Daphnia were included with Coenagrionidae to represent the size and shape of cladoceran prey common to the vegetation. Fish sizes ranged from 22-110 mm. Curves were fit by the regression equations in Table 3. HANDLING 15 SEDIMENTS VEGETATION OPEN WATER l 1 I l r l 8 4 8 4 .4 .8 PREY LENGTH/ FISH LENGTH X10 Figure 2 l6 mqo.o NN.¢ qmc.o .umuoq\a mmEmu acm~vcm£ EsEmCMZ mammEmu memo waddvcm: sows? 3o~mn :uucmH :mmw\£uucmfi mono mo memumu ucmmouamu .umuoq\a NC. + Nc.~ u 1 CN 0 : HN. + no.0 u m ..NNHC NC. + Nc.~ u I .aatug\m v a\m Axe memo «unseen: mad 0 : em. 0 u amusemc cam A\M oco.om mmc.c n m mmpw:0muumcooo no u a ca. 0 u q\m ~o¢.om dom.~ n z masocoufino as u a No. u t a\a mms.xlm smm.o u z «eczema .omuua\m M.g\m was“ sate Ame mama seasons: .5 a Hm. 8:. .zadmmucocoaxm mommmuocM memo mam~pcmz cowgz o>onm mam ucmumaoo mmsfim> use .mma>9 zoom mouse mo Lomm co mammmmm mfiamamsdm mow mnemumfimm mEmH mam~vcmz .N mfinmy 17 significant and functions to increase the handling time for sediment prey. The rate at which a fish encountered prey (B) was a function of prey density (D), length (i), and fish size (L) in each habitat (Figure 3). Prey encounter rate was defined operationally as the mean number of prey captured per unit search time; this rate was calculated for each experiment by dividing the number of prey captured by the total experimental time minus total handling time. The relationships shown in Figure 3 can be qualitatively interpreted from known information on the visual abilities of fish. Numerous studies (Ware 1971, Werner and Hall 1974, Vinyard and O'Brien 1976, and Confer et al. 1978) have demonstrated a positive relationship between reactive distance, the distance at which a fish can detect a prey, and prey body length. Schmidt and O'Brien (referenced in O'Brien 1979) have also shown that reactive distance increases with fish body length. However, while reactive distances may be measured as functions of prey length and fish length in simple environments, the translation of these reactive measures into actual rates of prey encounter in habitats containing physical structure and cryptic prey is a complex if not hopeless task. A regression approach was chosen instead as a means of predicting prey encounter rates in each of the habitats studied. The variables B, D, i, and L, as well as their logarithmic transforms, were examined using stepwise multiple regression to determine which combination of variables provided the best predictor of prey encounter rate (B) in a habitat. A regression model of the form: Log B = a + b1 Log D + b2 Log i + b3 Log L 18 Figure 3. Laboratory Prey Encounter Rate (E f 1 SE) in Relation to Prey Density, Prey Length and Fish Length. Results shown in the first panel are from two 65 mm bluegills feeding on Daphnia (1.14 mm body length). The second panel shows results from two 11 mm bluegills feeding on Coenagrionidae (192 naiads/m ) and the third panel shows data from 8 bluegills eeding on 11.1 mm Chiron- omus larvae at 1000 larvae/m . Curves were fit by the regression equations in Table 3. m ouawwm l9 2:5 2E5 Isozma Ihwzwq Iw_m mo commmmuumu some mo Show och cemmwouumu ofiamuaas osu mo ml. + soc.o oa. n.~om.o am..u ooo.~ no mm. Cu. mm. mqo.~ fino.o ~No.o Np ac..u ¢a~.o me..u aas.o oc..H aso.o an .A uoq mp + H mo; Na + a mo; fin + m .umumnm; 50mm cw moumu umucsoocm amua acmsfiuummv mnemumsvo Amm a “J muwumEmuma pmuumm .mCOmumsvm mumm umucsoocm xoum eq..“ moo.m- cm. n.saa.m- om. n.sam.m- Q m uoq ma cemumaco comumuowm> mucmsmvom umuma code swamps; .m mange 21 yielded the best overall fit to the data and the fitted constants for each habitat are listed in Table 3. The complete regression for each habitat is highly (p < .001) as are each of the variables within the regressions (p < .05). Inspection of residuals revealed no strong bias in the equations. A hierarchial F test was used to test for significant interactions between the variables. No 3-way interactions were found to be significant (p > .05). Of the nine possible 2-way interactions only one, D x L for the vegetation habitat, was significant (p < .05). Therefore, a simple model excluding interaction terms was used. To illustrate the relative effects of prey density, prey length, and fish length on the rate of prey encounter by bluegills, response surfaces constructed using the regression equation for the vegetation habitat are shown in Figure 4. As can be seen, fish size has a dramatic effect on prey encounter rate, larger fish encountering more prey per unit time than smaller fish. This result is consistant with the visual physiology of fish which predicts an "increased acuity or sensitivity or both in larger fish of any species" (Hester 1968). Qualitatively, the effects of prey size, fish size, and prey density were similar in all three habitats. Important quantitative differences, however, exist between habitats (Table 3). Notably, encounter rates for the same size and density of prey are much greater in the open water than in the vegetation or sediments. The open water provides no structural refuges for prey and consequently zooplankton occurring in the lighted epilimmion are considerably more available to fish than littoral prey. The effect of fish size on prey encounter rate is also greatest in the Open water habitat where no environmental structure limits visibility. The basic tradeoffs involved in determining habitat utility to the Figure 4. 22 Rate of Prey Encounter as a Function of Prey Density and Prey Length for Three Sizes of Bluegills (20, 60, 110 mm) Foraging in the Vegetation. Response surfaces were constructed from the equation Log B = 5.905 + 0.779 Log 1) + 1.045 Log 1 + 0.694 Log L {Table 3). Prey densities of 10-45 prey/m3 and prey lengths of 5-12 mm were used for each fish size. 23 ES 00— q ouswwm SE 00 >h.w2wo Ihmvzmq >wma >mma BLVU HaanOONB EE ON 24 bluegill are thus apparent from the relationships of prey handling times and encounter rates characteristic of each habitat. Prey found in the open water habitat are highly visible and are encountered at high rates by a foraging fish. However, since these prey are small and must be handled individually, the profitability (E/H) of any single zooplanktor is low. Vegetation dwelling prey are generally larger than zooplankton and will have higher individual profitabilities (E/H); however the structure of the vegetation reduces the rate at which bluegills can find these prey. Encounter rates with sediment dwelling prey are similar to those found in the vegetation, but the time required by fish to extract these prey results in higher handling times and lowered prey profitability. The foraging model (eq. 1) can now be used to weigh the relative magnitudes of these effects and determine the net energy available from a habitat. Because parameters of the model are calculable functions of prey size, fish size and prey density, the net energetic return and Optimal diet of a given sized bluegill can be predicted from a knowledge of available prey sizes and abundance alone. Measures of the prey sizes and abundances naturally available in each habitat were determined in a small Michigan lake. These data were used to generate predictions of Optimal diets and habitat use for bluegills in this lake, which were then compared to actual diets and habitat use of the fish. A FIELD STUDY Site Lawrence Lake, a mesotrOphic, marl lake, 4.9 ha in surface area, located 2.1 km east of Hickory Corners, Barry County, Michigan was <fl1osen as the study site because it contains a large and apparently 25 stable population of bluegills (Hall and Werner 1977), a simple vegetational community (79% Scirpus subterminalis by weight (Rich et al. 1971)) and distinct habitat types similar to those used in the laboratory experiments. Prey resources and fish were sampled from each of three habitats (barren marl bench, vegetated bench and slope, and open water) on the east shore of Lawrence Lake (Figure 5) during the months of May-August, 1979. Prey living in the bare, marl sediments were sampled with a corer (6.5 cm dia.; 8 cm deep) operated by a diver wearing mask and snorkel or SCUBA. Five random samples were taken on each date. Samples were gently washed through a series of 3 sieves (2.83, 0.710, 0.180 mm), from which prey were collected and preserved in 10% formalin. Contents of the smallest mesh sieve were mixed with a 20% solution of M2304, allowing the separation of live prey from inorganic sediments and debris. The vegetation habitat was sampled by a diver using a modified version of Gerking's sampler (Gerking 1957), described fully in Appendix A. This sampler has proven to be effective in capturing prey from the size of small cladocerans to large, motile prey such as damselflies, mayflies and occasionally small fish. Sampling was stratified with five samples taken randomly along the bench (.5-1.5 m depth) and five samples taken randomly along the slope (2-4 m depth) on each date. Vegetation samples were carefully washed into a series of 3 sieves (2.83, 0.710, 0.180 mm). Contents of the 2 largest sieves were sorted by hand and the live prey removed and preserved in 10% formalin while contents of the smallest sieve were separated using 20% Mgsoa solution. Open water prey were sampled at two locations in the lake using a 26 Figure 5. Bathymetric Map of Lawrence Lake, Michigan. Sampling location is indicated. 27 E... mmuhui 2. Mudguhg 50:85 puma OOn 00m. 09 o on i. mmwhg 24030.2 .> .5500 >¢m 4 years respectively (Werner and Hall unpublished data). Stomach contents of fish in each size class were counted, measured and identified to the lowest taxonomic level possible (42 different prey groups, about 50% to genus or species). Length-weight regressions were Obtained for each prey taxon permitting the conversion of counts to unaight. The prey were then categorized according to habitat types, i.e. prwey found in (1) the Open water (plankton), (2) sediments, (3) 30 vegetation, or (4) other, which included prey not specific to any of the first three habitats. The resource samples were used as guides in assigning prey to specific habitat types. The optimal diet and average net energetic return (En/T) available in each habitat was calculated for each bluegill size class using equation 1 and the size-frequency distribution of prey available in the field. In predicting optimal diets from equation 1 and the laboratory derived parameters, encounter rates for different sized prey were assumed to act independently; i.e. the rate at which a fish encounters a given sized prey is unaffected by the size and/or density of other prey in the environment. This assumption is necessary in extending prey encounter rates experimentally determined with single prey size classes to the field situation where multiple prey sizes occur concurrently. Figure 6 compares available prey distributions, predicted optimal diets, and the actual diets exhibited by the three size classes of bluegills in the field. Optimal diets were computed for fish sizes corresponding to the midpoint of each size class. The largest bluegills were highly size-selective in their feeding and the distribution of prey sizes eaten corresponded quite closely to that predicted by the optimal diet model (Figure 6). The correspondence with the predicted optimal diet is especially impressive in light of the extreme size selection predicted in two different habitats and across three sampling dates. Bluegills (> 100 mm) utilizing the plankton fed almost exclusively on large Daphnia, ignoring capepods, small cladocera (e.g. Ceriodaphnia and Bosmina) and the majority of smaller Daphnia present (which ranged (flown to 0.65 mm body length). Since the distance at which a fish can detxect a zooplanktor increases with prey size (see Confer et al. 1978 Figure 6 a,b,c 31 Size-frequency Distribution of Prey Available in a Habitat, Predicted Optimal Diets, and Actual Diets of Three Size Classes of Bluegills. Data presented represent those dates where prey availability in a habitat could be matched with the diets of 4 or more fish which had obtained :>90% of their diet by weight from that habitat. Sample sizes ranged from 4-9 fish/date. 32 o ouswwm .U AEEV IkOZM... >mmn_ VN A: w. VN o; m. 0. g w v JEH ION. ION. m:_mw3.m ION. H 6.: of- .o. I. . 3 new row. ..ov. N b. F 'P b P I: = o H .0 row new you I. D H 5836 W n 5:. mm. V ...—..— TOV. 10V. 10V. 1 N n P h b O lflfr—JJ A V .2. f .2. .3. W 8 no .a=< o. 33. «N >32 m... co.xcm.u vow. c0323; ION. c2339.; . Vow. N . a: .. 33 o; m. .ucoo o muswwm n~.u=< c0252.. 1...: AEEV Ihwzmu >mmn. O. a. a v o. a. a v p p p h h b 5 J Flirllu HJIH you. now. 2:935 1cm. :2: oo....m 50‘. fl \WOV. '0‘ .R oo. p r wow. wow. :685 vow. 5:. I HOV. 10V. '0‘ P11 IL P j you. 2 :3. you. 2 >22 now. co..o.omo> c3333; ..ov. row. We... “has. NBLVS '.IV|N|.LdO AONBOOBHd lNSIQWV 34 .ucoo c muswwm AEEV IPOZm... >mmn. 0.. m. 0— N— m V 0— NP 0 V p P p p b [Eln— - - now. ...om. 2.68:6 you. E:.onuo_ IOV fiJ.WOV IOY Pfion. - p p s - W1 L, b b ION. ION. .zmmam ION :2: On IOV. fiOV. IOV. - p - 1%.H_ r. rlrg mm da< ION. o. .23. ION. mm >32 ION. co.xcm.d co_.m.ooo> .::.m.oam> IOV. IOV. IOV. of. ‘IVWIldO AONBOOB 8:1 N31V3 1N3|8WV 35 for a review), such observed size selection may simply reflect an increased encounter rate with larger, more visible prey. For the open water habitat, where visibility is unhindered by environmental structure, two methods are available to compute the average prey size ingested by a bluegill consuming prey "as encountered". The first uses a formula derived by Eggers (1977) who, assuming a cylinderical visual field swept out by a swimming fish, calculated the expected prOportion of prey size i ingested as where Pi = proportion of ingested prey of size-class i, Di = ambient density of size-class i, Li = length of prey in size-class i, and n = number of classes. A second estimate of the average prey size encountered can be obtained using the regression equation for the open water habitat presented in Table 3. From this equation the number of prey size-class i encountered/sec foraging can be calculated knowing the body length and density of prey size-class i and the size of the fish. Table 4 compares the mean length of Daphnia in the environment with predicted mean length of Daphnia if consumed "as encountered", the mean length of Daphnia in the optimal diet, and the mean length of Daphnia consumed on two sampling dates when at least 4 large bluegills collected had been eating only Daphnia. On both dates the average size of prey consumed was significantly greater (p < .01) than predicted if prey were consumed "as encountered". Thus, the prey size-selection shown by these fish represents actual choice and not simply an increased encounter with large Daphnia. Secondly, the optimal foraging model predicted quite accurately the mean prey size eaten; the predicted means falling within 36 Table 4. Prey Size-selection by Large Bluegills. Comparison of mean body length of Daphnia in the environment; eaten "as encountered"; in the optimal diet (125 mm SL. bluegill); and actually eaten by bluegills (101-150 mm SL.) (4 fish each date). All means 1.1 SE except where indicated. Ambient and "encountered" means based on n = 2 prey samples from each date. 2 Prey Length (mm) July 19 August 3 Ambient 0.88 :_.06 1.05 :_.01 "Encountered" 1.03 + .17 1.21 + .01 (Egger's eq.) — —T — — "Encountered" 1.04 i .15 1.18 :_.04 (regression) * * Optimal Diet 1.75 * l 1.78 * l Eaten 1.74 _+_ .04 _[ 1.86 i .04 __l_ 1 .13 957. 0.1. 1 .13 7. 0.1 * p < .01 37 the 95% confidence intervals for the average prey size eaten on both dates (Table 4). The two smaller size classes of bluegills were also size-selective in their feeding, although to a much lesser extent than bluegills > 100 mm (Figure 6b,c). These smaller fish also regularly included in their diet some prey sizes below those predicted by the Optimal foraging model. The inclusion of these sub-Optimal prey sizes may well represent the fact that the net energy intake (En/T) of these fish shows little change over a broad range of prey sizes. In Figure 7, En/T is plotted as the diets of the three fish sizes expand from including the nost profitable to least profitable prey; the optimal diet being defined as including all prey sizes 2 the prey size at which En/T is maximized. The En/T curve has a distinct peak for the largest size class of bluegills (125 mm) and the optimal diet is clearly defined. For 75 mm bluegills the peak in the En/T curve has become a plateau, and for 30 mm bluegills the En/T curve has flattened to a gentle rise and fall. Thus, while a unique lower bound on the diet which maximizes En/T exists, there is obviously a range of prey sizes of lower profitability, the inclusion of which has only a very slight effect on a small fishes' average rate of energetic intake. Therefore, if all fish sizes have equal abilities to estimate average energetic intake, one would expect the same error in estimation of En/T to cause an increasing number of sub-Optimal prey sizes to be included in the diet as fish size decreases. For the data shown in Figure 6, the percent error in estimating En/T which would account for the consumption Of sub-optimal prey sizes by the Lawrence Lake fish is on average 11.: 3.5% for large bluegills, 21.: 13% for medium bluegills and 6 I 3.5% for small Figure 7. 38 Average Foraging Return (E /T) of Three Sizes of Blue- gills (125, 75, 30 mm) as 3 Function of Prey Sizes in the Diet. Curves were constructed from equation 1 and the size-frequency distribution of zooplankton available in Lawrence Lake on July 19. Over this range of prey sizes, prey profitability (E /H ) is directly related to prey length for all three sizes of fish. The optimal diet of each fish size is defined as including all prey sizes _>_- that prey size which maximizes En/T (indicated by an arrow). 39 n muswwm >>On_ >._._.._.m<._..n.OIn. 1.0—I .EEV xeozmu >mma V. O. Né 04 ON - L . p - o\o\o \ . o\o 01/ \ O/o/o O\O \ villwnuumflnuvllr e Q\\\O|IIO\\1 x /-/ /O|O/O\ \ 1V0. I O I. n j/ /. \ / . / 2 .\ .... /./.I.\.\ . E... on all /./ \ EE mu 1 o 0 25 mm. [II /./+\ IN... (oas/Ieo) .L/uI-l 40 bluegills. Thus, these limited data suggests that all three bluegill size class did in fact optimize their diet to approximately the same degree. Clearly, the Optimal foraging model developed in this study, which is based upon simple body size relations determined in the laboratory, quite accurately predicted the diet breadth and prey selection of the largest size class of bluegills in the field. Moreover, the actual and predicted size selection of these fish was greater than that resulting from a simple visibility bias alone. Thus, one can confidently use the model to examine the degree to which habitat use by these fish is based upon maximizing foraging intake. The diets of the two smaller bluegill size classes often included some prey sizes smaller than those predicted. However, it was shown that the inclusion of these prey sizes generally has only a small effect on the fishes' average rate of energetic intake. Consequently, the model's predictions of net energetic return should be good estimates of habitat profitability for these fish sizes also. In the next section the seasonal pattern of habitat value is determined for Lawrence Lake and compared to the actual habitat use of the three bluegill size classes. Habitat Profitability and Use If bluegills are feeding so as to maximize their energetic return one would expect not only to find close correspondence of optimal and actual diet breadths but also each size class foraging in the habitat yielding the highest net energetic gain. Further, as prey abundances change across the summer one would expect to see fish shift habitats in response to changes in relative habitat profitabilities. Over the summer bluegills fed extensively from the vegetation and 41 open water habitats. Therefore, we can plot the predicted energetic return from these habitats against actual foraging intake of fish feeding in the habitats to determine 1) whether predicted foraging returns are related to actual amounts eaten and 2) whether any initial bias exists in the estimation of energetic return between habitats. Plotted in Figure 8 are the model's predicted energetic returns from the vegetation and plankton versus the actual prey biomass ingested by bluegills feeding in these habitats. For both the vegetation and plankton there is a significant positive relationship (p < .05) between predicted and realized intakes, demonstrating that the nodel does provide a good index of habitat value to the fish. Note, however, that points for the vegetation fall below those for the Open water plankton, indicating that energetic intakes for the vegetation habitat have been overestimated relative to those for the plankton (since the times available for foraging each habitat were approximately equal on each sampling date). The basis for this discrepency may lie in the fact that the vegetation densities used in the laboratory experiments were less than those actually occurring in the field (100 stems/m2 vs % 900 stems/m2), resulting in estimates of prey encounter rates higher than those actually obtainable by the fish in the field. Glass (1971) has shown that increasing environmental structure decreases the rate at which largemouth bass can capture prey in the laboratory. When prey encounter rates in the vegetation are reduced 50% the regression lines sl1own in Figure 8 coincide removing the bias in estimated return rates between the vegetation and open water habitats. This reduction in encuaunter rate was applied to the optimal foraging model to generate unbwiased predictions for the profitability of the vegetation habitat Figure 8. 42 Linear Relationships Between Predicted Foraging Return from a Habitat and Stomach Contents (mg dry wt.) of Bluegills Foraging >»80% of Their Diet from that Habitat for the Open Water (x) and Vegetation (o) Habitats. Fit- ted regression equations are Y = -2.35 + 252.1 X (r - .47, p <:.05, n = 24) for the open water and Y = -4.02 + 117.14 X (r - .84, p <:.01, n - 59) for the vegetation. IN STOMACH (mg) AMOUNT 60- 4o- 43 PREDICTED RETURN (cal/sec) Figure 8 44 across the season. The predicted pattern of habitat use is similar, however, using either corrected or uncorrected prey encounter rates in the vegetation. The first column in Figure 9 shows the seasonal pattern in predicted net energetic return (cal/sec) for a given sized bluegill foraging in each of the three habitats. These predictions were generated using the Optimal foraging model and the habitat specific prey abundances measured in the field. Clearly, there were marked differences in habitat profitability. The vegetation habitat was by far the most profitable habitat initially, declining steadily in value across the summer. This decline was predominantly caused by a reduction in the abundance of large insect prey (Appendix A) and had the greatest proportional effect on the largest size class of bluegills. Plankton profitabilities were intially low, rose in late-June, and remained above those of the vegetation habitat throughout July and August. The rise in plankton profitability resulted from an increase in the abundance of large Daphnia (Daphnia #/l > 1.5 mm body length: 0 on June 14; .04 i .01 on June 25; .18 :_.14 on July 19; .32 i .05 on July 29 (§::_l SE, n = 2)). Throughout the summer the sediment habitat remained very low in profitability, reflecting low prey abundances and a high minimum handling time for bluegills feeding on these prey (recall Figure 2). Therefore, the predicted seasonal pattern of habitat use resulting in maximum energy intake for all three size classes is as follows; exclusive use of the vegetation habitat in May, a shift to utilizing the plankton in late June, and a continued exclusive use of plankton throughout July and August. Note that the magnitude of difference in Figure 9. 45 Seasonal Pattern in Predicted Habitat Profitability (left) and Actual Habitat Use (right) for Three Size Classes of Bluegills. Actual habitat use determined from the x amount of prey (mg dry wt.) foraged from each habitat. Dashed lines and closed circles represent the open water, solid lines and closed circles the vegetation, and dotted lines and open circles the sediments. Sample sizes were 3-6 fish/date for large bluegills (101-150 mm), 5-9 fish/date for medium blue- gills (51-100 mm) and 4-9 fish/date for small blue- gills (10-50mm). (cal/sec) RETURN FORAGING 46 PREDICTED ACTUAL Figure 9 '15) Large 5 Large {3% .3 P 0.... . .... L .5 ....o 0.... .. o .0 0.... 3 0 .1 b... 8.; . .0 5 ‘ ... 0.. O. ...9. a... 10 d ' .o c» .......... E e --------------------- : -' iI-I v Medium 2:) 1 Medium 1 :0 < ‘5 ' ' .0. o. 2 .1 ... .... o O '1 L .... o o... O W :- ...... F O: .0 ..... U) .: .05 tin. .‘i ‘-_. Z 5 II o .5: \ - / \,\:: ...".. ........... '— .o 9'...“ \0 ’. 0 .................... C ' . # 2 WM . . .h D P" C) 12 SmaH £3 < SmaH .0 4 '1 ...” . ...... . 2 " \ ......... , .0 2 ‘ D“... .... "-\ 1 - / \ ...... 0%,”... 0.: .o-0 .:\o .'-:i ........ .... ° ' 0 .Q ..................... <3 9 ' ° ' #:h ” H 14 19 23 JUNE JULY AUG 47 predicted profitabilities between the vegetation and plankton habitats early and late in the summer is considerably greater than the percent error in estimated foraging return (En/T) which would result from the inclusion of suboptimal prey sizes in the diets of the Lawrence Lake fish. Therefore, any differences between predicted and actual diet breadths exhibited by the fish are not so large as to affect the prediction of a habitat switch from the vegetation to the plankton across the summer. The actual seasonal diets of the bluegill size classes are shown in column 2 of Figure 9. The habitat use of the largest bluegills was nearly in complete accordance with predictions for maximizing net energetic return, with these fish feeding initially in the vegetation , and then shifting to the open water plankton in July. The shift to the plankton occurring some 2 weeks later than predicted. The two smaller size classes of bluegills also foraged initially from the vegetation as predicted. However, these fish continued to use the vegetation habitat for nearly the entire summer, despite calculations that they would have increased their foraging return by utilizing the plankton after the middle of June. On the single date when these smaller fish feed extensively on plankton, August 23, foraging intakes increased dramatically over previous dates indicating that these fish did profit in utilizing the plankton. August 23 represents a change in the distribution of the plankton resource from previous dates. Table 5 shows that August 23 was the only date in the summer when Daphnia were abundant near the vegetated SIOpe. Consequently, on this date small fish may have been able to feed on Daphnia while remaining in close proximity to the structural protection 48 Table 5. Distribution of Daphnia in Lawrence Lake. Mean density (#/1) :_1 SE of Daphnia sampled at the midlake and nearshore stations for those dates when bluegills were feeding on Daphnia; n = 2 for each case. June 25 July 19 July 29 August 3 August 23 Midlake 0.32 + .08 2.55 + .55 2.25 + .25 2.30 :_.10 3.25 + .25 Nearshore 0.03 + .01 0.18 + .12 0.33 + .01 0.07 :_.01 5.45 + .55 49 of the vegetation. A comparison of the species of Daphnia eaten lends support to this line of reasoning. Three species of Daphnia co-occurred in the plankton on August 23 and they differed markedly in their horizontal distribution. Vertical tows taken from the limnetic station contained on average 64% 2.311431, 30% l_)_. retrocurva, and 6% E. galeata while vertical tows taken at the base of the slope (within .5-1.5 m of the vegetation) contained 6% 2.111335, 51% B. retrocurva and 43% 2: galeata. The two smallest size classes of bluegills had only 2: galeata and 23 retrocurva in their stomachs despite 2, pulex being the most profitable prey due to its large size. Bluegills > 100 mm did feed heavily on D. pulex. These data suggests that large bluegills foraged on Daphnia in the limnetic zone whereas smaller fish would feed on these prey only when they occurred near the vegetation. DISCUSSION Prey size selection by the largest bluegills in the field closely matched predictions of the optimal foraging model, providing strong evidence that diet selection in these organisms is based upon maximizing net energetic intake. This corroborates earlier laboratory work by Werner and Hall (1974) who demonstrated that bluegills select prey so as to maximize biomass (energy) ingested. The diets of smaller bluegills- generally included more suboptimal prey sizes. Other studies of optimal foraging similarly report a fraction of suboptimal prey eaten (Werner and Hall 1974, Krebs et al. 1977) and the inclusion of these prey has been postulated to represent effects of recognition time, "sampling", and/or detection errors (Krebs et al. 1977, Krebs 1978). It is unlikely tfliat large fish are better at recognizing prey than are smaller fish, as thEE same visual cues are available to all fish. However, there may well 50 be size-related differences in the abilities of fish to sample their environment and estimate energetic returns. Small fish encounter and capture fewer prey per unit foraging time than large fish due to an increase in searching ability and handling efficiency with fish size (recall Figures 2 and 4). These differences in capture rate are demonstrated by the diets of bluegills collected on May 22 when all fish sizes had been feeding in the vegetation. On this date the average number of prey items in the stomach was 202 i 20.3 for large bluegills, 32.8 :_ll.5 for medium bluegills and 12.6 :_2.6 for small bluegills (§.: 1 SE). Therefore, if fish must sample their environment and estimate prey availabilities by encountering and/or capturing prey one would expect the sampling efficiency of a fish to increase with body size. Moreover, small fish will be more likely to have their diets affected by small scale patchiness in prey distributions since they encounter fewer total prey per foraging bout. These two factors may in part account for the higher percentage of suboptimal prey sizes included in the diets of small bluegills. A third factor which may contribute to the reduced correspondence between predicted and actual size selection in these fish is that the average foraging return for smaller fish changes little over a broad range of prey sizes. A comparison of the change in energetic return (En/T) as prey sizes are added to the diet (Figure 7) illustrates that the inclusion of suboptimal prey sizes in the diet has a smaller total effect on energetic intake as fish size decreases. Therefore, if all fish are able to estimate foraging return to the same degree (i.e. :_X number of cal/unit time) one would expect any errors in the estimation (If En/T to result in the inclusion of more suboptimal prey in the diets 51 of small fish. If, however, the estimation of foraging return is based on relative values (i.e. :_X% of maximum En/T) decreasing fish size may or may not result in the inclusion of more suboptimal prey in the diet depending on the exact shapes of the En/T curves. The abilities of bluegills or any other organisms to estimate their average energy intake is unknown and the results of this study clearly demonstrate the importance of understanding the means by which organisms sample their environment and estimate parameters in the continued refinement of optimal foraging theory. Bluegills of the largest size class (101-150 mm) were found to forage so as to maximize their energetic gain, switching habitats as relative profitabilities changed across the summer. The importance of this habitat shift to the bluegills, in terms of energetic return and potential growth, is evident from examining Figure 9. In June, prior to feeding on zo0plankton, large bluegills averaged < 7 mg prey dry weight in their stomachs. Upon switching to zooplankton in July gut contents increased to an average of 20-40 mg dry weight. Thus, by utilizing the open water habitat these larger bluegills were able to maintain a high rate of energy intake after the vegetation habitat had dropped in value. / J / I I / The timing of the shift to the Open water habitat was delayed /’ approximately 2 weeks from that predicted by the optimal foraging model. Werner et al. (1980) postulate that such a delay in habitat switching is to be expected when organisms must estimate foraging returns in complex habitats where the effects of learning or experience are important. These workers have shown that bluegills require 6-8 daily exposures to laboratory representations of the Open water habitat before reaching maximum foraging efficiency. Thus, if bluegills sample the open water 52 or other new habitats in an inexperienced state, their estimate of energetic return from the habitat will be considerably below that obtainable once they are experienced. This can result in a delay in the timing of habitat switching relative to predictions based upon experienced foragers (the case in Figure 9). Further, once a habitat switch is made one would expect to find a rapid increase in foraging intake due to an improvement in efficiency in the habitat. The delay in the switch from the vegetation to the open water habitat shown by the large bluegills in this study, plus the marked increase in the amount eaten once the switch occurred, matches these predictions. However, the limited number of samples taken during the period in which the bluegills wererswitching from the vegetation to the plankton precludes a thorodgh evaluation of Werner et al.'s hypothesis. Also, some Of the increase in stomach contents after the switch must be attributed to the increasing profitability of the plankton over this time period as well as to any improvement in foraging efficiency due to experience. Bluegills smaller than 100 mm did not extensively utilize the open“ water habitat except when large Daphnia were abundant near the vegetation (e.g. the August 23 sampling date). The jump in biomass of prey ingested on this date relative to those previous (see column 2, Figure 9) demonstrates that these fish did in fact profit energetically from foraging on Daphnia and would have throughout July and August had these prey always occurred near the vegetation. The habitat use of these smaller bluegills may be compromised by an increased predation risk in more open environments as the degree of “Mm—m» ,- .1... , , , predation risk to the bluegill is strongly size related. Bluegills > 1CK)tmn are beyond the handling capabilities of all but the very largest . ..-. n... 53 individuals of its dominant predator in this system, the largemouth bass (Micropterus salmoides). With decreasing size bluegills < 100 mm become vulnerable to an ever increasing fraction of the bass population with 80% of the bass (excluding young of the year) capable of handling bluegills up to 30-46 mm (Hall and Werner 1977). Therefore, one would expect that bluegills > 100 mm would be totally unconstrained by predation in their habitat selection while bluegills of smaller sizes would be forced to balance their potential energetic gain from a habitat against their risk of being eaten. Glass (1971) has demonstrated that the foraging success of the largemouth bass (Micropterus salmoides) declines dramatically with an increase in environmental structure (e.g. the vegetation habitat versus the Open water). Further, largemouth bass in Lawrence Lake are distributed with the majority of the population occurring in the water column off the slope (Werner et al. 1977). These two factors together suggest that the open water habitat is a relatively risky environment for small bluegills. That bluegills < 100 mm feed on Daphnia occurring within a few meters of the vegetation while apparently ignoring Opportunities to feed on Daphnia located farther offshore suggests that these fish may balance foraging profitability and predation risk when selecting habitats, and that this balance is a dynamic one depending on prey availabilities and relative risks. A size-specific tradeoff between foraging profitability and predation risk is likely to be a common phenomenon among many groups of c>rganisms whose pOpulations are structured by body size. In simple one liabitat systems these tradeoffs can result in a single, optimal body size for organisms occupying that habitat (Lynch 1977). In systems cxnntaining a number of habitats, each with different size-related 54 foraging efficiencies and predation risks, optimal habitat Choice by organisms will involve balancing energetic gain from a habitat against the chance of being eaten in that habitat. In such cases it will be important to consider the time scale over which organisms make decisions as Optimal habitat choice on a day to day basis may not be consistant with optimal habitat use over the entire lifetime of the organism. Also, changes in prey availability and predation risk on a diel basis can dramatically effect the relative values of habitats over time scales of a few minutes or hours (see Hall et al. 1979, Bohl 1980 for examples with planktivorous fish). The existence of such restricted feeding periods in which prey may be maximally available and/or predation pressure reduced would be expected to provide strong selective pressure for the evolution of traits maximizing energy intake during the time available for feeding. Recent laboratory work with notonectids has demonstrated that back- swimmers are apparently able to balance conflicting demands between” foraging profitability and predator avoidance in such a way as to maximize their feeding rate (a measure of relative fitness) (Sih, in prep.). My results similarly suggest the importance of these two factors in determining the habitat use and resultant size-class segregation of bluegills in natural environments. However, this area of ecology represents the somewhat curious case where the development of relevant theory lags behind empirical demonstration. To date, models of Optimal habitat use incorporating both energetic gain and predation risk have been limited to a few, special cases (Rosenzweig 1974, Covich 1976, Pearson 1976). A major difficulty in the development of such theory lies in the fact that foraging profitability and predation risk are 55 basically measured in different units (e.g. energy gained/unit time versus probability of death/unit time, respectively). Only by translating these measures into common units commensurate with fitness will a more general theory of habitat use emerge. The pattern of size-class segregation demonstrated by bluegills in this and other studies (Werner et al. 1977, Hall and Werner 1977, Keast 1977) represents a case where predationflmay function to reduce,“ intraspecific competition by channelling size classes into separate Whabitats. Bluegill size classes show little tendency tO partition the food size dimension of the niche (Hall et al. 1970, Keast 1977) and Werner (1974) has shown from handling time relations alone that bluegills from 50-250 mm in length would be expected to overlap completely in the range of prey sizes eaten. As a result, diet overlaps between bluegill size classes are often large and the potential for intraspecific competition strong. By restricting small fish to habitats of low risk (i.e. the vegetation), predation pressure can result in _de_ fagtg habitat segregation and a reduction in competition between size classes. Unstructured habitats become essentially exclusive resources for individuals large enough to be outside the danger of predation. These exclusive resources can result in large disparities in growth rates between size classes as evidenced by the foraging return of large and small bluegills in Lawrence Lake late in the summer. This phenomenon of predator-generated habitat segregation is likely to be a general one among species where predation risk is related to body size and its potential importance to the outcome of intraspecific interactions remains to be studied. The fact that bluegills and many other species stocked in the 56 absence of predators invariably develop "stunted" populations (Swingle and Smith 1940, Wenger 1972) has long demonstrated the importance of predation in mediating intraspecific competition in fish. Historically, the action of predation in these systems has been viewed as reducing population numbers to the point where large individuals can maintain positive growth rates. It is suggested here that this explanation is probably too simple, that in fact the action of predators in generating habitats of exclusive use to the larger size classes (e.g. the Open water, bare sediments) may actually be the most important factor in stabilizing intraspecific competition between size classes. While predation may function to reduce intraspecific competition in 5‘“ -._.,.._ _‘_ some cases, it is also likely to concentrate interspecific interactions at—specific points in a species life history. In small lakes such as Lawrence, young-of-the-year of all species are found in the littoral vegetation (Werner et al. 1977, personal observation), sometimes after spending a brief period in the limnetic zone as fry (Werner 1966). Adults of these species may eventually come to utilize different resources (e.g. fish, zOOplankton, mollusks), which often occur in separate habitats. Such habitat segregation is precluded among young fish to the extent that predation risk restricts small individuals of these Species to a single habitat, the vegetation (as is apparently the case for bluegills). Moreover, young of different species may exhibit large differences in their abilities to capture vegetation dwelling prey as their morphologies and foraging behaviors are fashioned to efficiently forage other habitats and prey types as adults (Werner 1977). Thus, asymmetries in competitive effects between species occupying the vegetation as juveniles are to be expected. Whether 57 strong interspecific competition actually occurs at this stage in a species life history will depend of course on the degree to which vegetation resources are limiting for small fish. This study has demonstrated that small bluegills could often increase their energy intake by feeding outside the vegetation. Also, juveniles of several fish species have been observed to move from the vegetation into more Open areas at dawn and dusk (Stuntz 1975, Wilsmann 1979, personal Observation) when lowered light levels presumably reduce predation risk. These observations suggest that resources in the vegetation are in fact reduced relative to other habitats and that species confined by predation to the vegetation are in competition. Clearly, the ability of ecologists to evaluate the relative impact of predation risk and foraging return in determining an animal's diet and habitat use will be an important step towards the development of a predictive theory of community structure. The success of simple Optimal foraging models in predicting the diets and habitat use of organisms in the field (Werner 1977, Belovsky 1979, this study) indicates that an Optimal foraging approach will prove extremely useful in examining questions related to resource partitioning, niche overlaps and the extent of competitive interactions both within and between species. CHAPTER 2 SIZE-CLASS COMPETITION INTRODUCTION A major goal in the development of an optimal foraging model which incorporates the effects of predator body size was to utilize foraging theory to provide a mechanistic understanding of the action of competition in size-structured populations. This goal has proven to be both distant and elusive. However, I would like to devote a few pages to some thoughts on the subject of size-class competition and hope that an interested reader might profit from my mental wanderings. Attention is restricted primarily to the Centrarchid fish communities of small lakes, but many of the ideas are generalizable to other systems of size-structured species.. INTRASPECIFIC COMPETITION For simplicities sake, let us first consider a homogeneous en- vironment inwhich all prey types are substitutible resources (sensu Tilman 1980) which can be arrayed by body size. For any given pattern of resource availability one can calculate the optimal resource util- ization for a given-sized fish from the foraging model develOped in Chapter 1. Utilization curves for two sizes of bluegill sunfish feed- ing on a lognormally distributed resource will in general resemble those shown in Figure 10. Having generated utilization curves for different 58 59 Figure 10. Hypothetical Distribution of Net Caloric Gain from the Diets of Two Bluegill Size Classes Utilizing the Same Prey Resource. NET CALORIC GAIN 60 ---- large fish small fish PREY SIZE Figure 10 61 size classes one can then measure overlaps in resource use between size classes. These overlaps will in general be asymmetrical (due to size-specific differences in foraging ability) and the direction and magnitude of these asymmetries estimate the strength of competitive interactions ( msu Cm mmum mumunmuum>cH we :Omusnmuumma mumm .H< oflnmfi 75 +| E. mm... 2. + .2. ac. .. ...... S. + ...... Em . ... £36. ...... 8 + 8. o o S + 8 c... - C... o c c o o... - ca b b o b. ca - c... 8 .. 8 8 H o... b. 8 H 8 o... . cg c 8 H o... S H 8 8 H 8 cg . c... .... an. H .2 i H 8. S. H as c... . c5. 02 H 2m ... H H... mm. H .... ca. H .N. o.m . o... o... H :8 .2 H .8 i H 8. ...N H .8 o... u an .... H can 3. H .8 2m H 8.. Sn H 3.... an 1 ca 2. H was .3. H «2.. 8% H 9...... 3.. H 8.: ca . ....N 2m H gm... 2... H .2} is H 2...... a... H .3... O. . .2 m; H Sag mam H ..mm... .2 H 8a omm H so... c.~ . m... 23.. H 28.8 8m H Sn... mam H 8...... mm. H .8 m. u o. mm H mama a; H 3...... man H 3.... as H 3.... o.. .. m... 2. + 3.. ...m + 8.. i + .... mm. + .8 a... - o mu umsws< a. hash . a. Ocam ~N he: AEEV sumsmfi moan ANE\%V auwmcmv mwum .wumc LOO mOHOEmm m n : “Ammfi HO mmmummcmv cmOE mum mama .Oxmg mucmusmq mo mucoEmpow oumm osu am Noam OumunOuum>cH mo :Omusnmuummn mnmm .N< Oficmh 76 Average Size Distribution of Invertebrate Prey in the Open Table A3. Data are mean densities (+ 18E); Water of Lawrence Lake. calculated from 8 samples taken between May and August, 1979. #/liter Prey length (mm) 861888359962341211 048168832311000000 0LLL00000000000000 n.+.+_+.+.+_+.+_+.+_+_+.+.+_+.+.+_+.+. 8623463193316934111 0295107463420000000 0372232101000000000 /\ I.9.1.A.§,6.7.Ruo,nuI.9.1.A.§.6.7.R.O,o. nonunununvnvnvnunv1.1.1ulululul.1.1.1.1. . . . . . _ . . _ . _ . . . _ . . _ ..> n.1.9.1.4.SJAV7.RVO,n.I.9.2.A.s,£v7.R. 000000000111111111 77 Table A4. Parameters of the Normal Distribution Fit to the Logarithm to Base e of Prey Length in Three Habitats. VEGETATION Modal prey Density at the size (mm) Variance mode (#lm3) ‘RE_ May 21 0.21 1.243 34,658 .90 June 14 0.28 0.668 63,059 .99 July 19 0.59 0.460 32,357 .99 August 23 0.60 0.344 19,808 .99 Seasonal 0.18 1.562 32,794 .91 average SEDIMENTS Modal prey Density at the size (mm) Variance mode (#/m2) 33_ May 21 2.41 0.104 2,611 .68 June 14 1.20 0.474 1,693 .82 July 19 1.16 0.268 6.560 .97 August 23 1.14 0.177 6.921 .98 Seasonal 1.13 0.387 3,672 .98 average PLANKTON Modal prey Density at the size (mm) Variance mode (#/L) 3%_ Seasonal 0.36 0.237 5.24 .76 average 78 parameters of the normal distributions are listed in Table A4. Since the method of Gauch and Chase (1974) is a regression approach one can examine the percent of the total variance explained by fitting the lognormal distribution, R2, to determine the goodness of fit. Examination of the R2's in Table A4 indicates that the distribution of prey sizes in each habitat was quite adequately described by the lognormal curve as generally greater than 90% of the variance was accounted for. Since prey weight (W) can be related to length (L) by a power function such as L = ch; this implies that prey weights are also distributed lognormally in each habitat (Aitchison and Brown 1957, Schoener and Janzen 1968). The range of prey sizes available in the three habitats is reflected in the variances of the fitted distributions, the variance being greatest in the vegetation and least in the plankton (Table A4). These differences in variances between habitats result in large differences in the relative biomass of small and large prey and can have marked effects on the feeding preferences of fish (see Chapter 1). Figures Al-3 illustrate the fitted lognormal curves and actual prey size distributions found in each habitat averaged across the season. Seasonal Patterns There was a strong decline in the abundance of large prey in the vegetation habitat across the summer (Tables A1,5). This change in the vegetation prey size distribution is evident in the steady decrease in the variance of the fitted lognormal distribution from May to August (Table A4). However, the seasonal loss of large prey in this habitat is not reflected in either a change in the mean or modal prey size indicating these measures alone are inadequate for determining the 79 Figure A1. Distribution of Prey Sizes (mm body length) in the Vegetation Habitat Averaged Over the Season. Density is in #/m3. Fitted curve is the normal distribution of 1n prey length with mode = 0.18 and variance = 1.56. DENSITY 80 VEGETQTION PREY o A O O V- (U 0 O O mn—i H O 0 O- m A A A11 A Am A o I — —I— I “r 0.0 4.0 8.0 12 0 PREY SIZE Figure Al D 81 Figure A2. Distribution of Prey Sizes (mm body length) in the Bare Sediment Habitat Averaged Over the Season. Density is in #lmz. Fitted curve is for the normal distribution of 1n prey length with mode = 1.13 and variance = 0.39. 2400 DENSITY 1mm 82 SEDIMENT PREY I I 8b 110 ISO PREY SIZE Figure A2 20. 0 83 Figure A3. Distribution of Prey Sizes (mm body length) in the Open Water Habitat Averaged Over the Season. Density is in #lliter. Fitted curve is for the normal distri- bution of ln prey length with mode 8 0.36 and variance 8 0.24. DENSITY 4 84 PLQNKTON PREY I I I .0 8p 1&0 PREY SIZE Figure A3 I 16 .0 20 .0 Table A5. Vegetation Sediments 85 Seasonal Pattern in the Density of Prey Organisms 2 3.0 mm Body Length in the Vegetation and Sediment Habitats. Densities are #/m3 for the vegetation and #lm2 for the sediments. Density of prey 2 3.0 mm in length May June July August 6,555 3,062 4,020 599 2,593 1,323 1,488 1,123 ‘3" w—r' 86 availability of prey sizes to fish. Table A6 lists the mean number and dry weight of the major prey groups found in the vegetation across the summer. The total biomass of insect prey declined from May to August, as did the biomass of amphipods and ostracods, with the sharpest reduction occurring between May and June. The decline of these prey groups was offset by the increasing biomass of gastrOpods, so that the total biomass of all prey in the vegetation remained fairly constant for nmch of the summer. Since most fish species do not possess the specialized adaptations required to feed effectively on gastrOpods (e.g. a crushing pharyngeal "mill" with large molariform teeth (Keast 1978, Laughlin and Werner 1980)), it is informative to examine the patterns in total prey biomass exclusive of gastrOpods. Table A7 compares the biomass of non-gastropod prey in the vegetation, separated into the bench and slope regions of the habitat. The total biomass of these prey declined steadily and by more than 2.5 fold from May to August. A reduction in the total volume of prey in the vegetation over this time period has also been reported by Ball and Hayne (1952) and Anderson and Hooper (1956). Note that prey biomass and prey size were generally larger in the bench region (.5-1.5 m deep) than on the slope (2-4 m deep). These differences between regions of the vegetation also exist if gastrOpods are included in the prey distribution (Table A8). Prey numbers and biomass in the bare sediments (Table A9) were considerably less than those available in the vegetation (Table A6) (prey densities in #/m3 for the vegetation can be converted to #/m2 of lake bottom by multiplying by 0.5 as vegetation height was approximately 0.5 m). Differences in prey abundance between vegetated and barren 87 Table A6. Mean Nunber arrl Dry Weight (per m3) of Major Pey Groups Found in the Vegetation. S inooephalus S ida Bosmina Cydoridae Total Cladocera Covepoda Ostracoda Hyalel 1a Hydr acarina Chironanidae Other Diptera Tricoptera Ephemeroptera Odonata Total Insecta N= 10 sanples per date. Vegetation Prey #/m3 (mg dry wt.) Ey 21 4,361 (88) 236 (10) 4,224 (6) 22,817 (66) 31,638 (170) 21,340 (81) 11,637 (257) 2,728 (235) 1,133 (86) 6,750 (372) 4,579 (108) 473 (293) 1,254 (595) 192 (228) 13,293 (1,534) Junel4 4,950 (211) 357 (12) 33,814 (37) 31,930 (411) 70,993 (411) 23,067 (126) 6,050 (105) 3,399 (132) 1,743 (138) 5,345 (140) 1,532 (27) 154 (42) 1,270 (285) 121 (248) 8,448 (747) July 19 2,271 (38) 10,109 (383) 951 (1) 6,281 (53) 19,601 (482) 22,302 (92) 4,634 (82) 3,465 (166) 1,633 (135) 21,180 (363) 1,598 (83) 451 (37) 2,100 (123) 220 (45) 22,907 (653) _AEEEEEJ51 2,068 (33) 3,366 (71) *9: 4,276 (33) 9,702 (142) 11,599 (32) 1,603 (60) 1,958 (49) 1,243 (87) 5,422 (140) 898 (4) 390 (25) 1,704 (82) 231 (276) 8,811 (535) Gastropoda 562 (265) 804 (1,279) 1,174 (1,230) 1,003 (694) Total Prey 82,687 114,669 82,577 38,593 (2,639) (2,946) (2,856) (1,603) “absert 88 Auzumma mot aEV wumm zmua.m me \ AuEv u£awm3 zuv zmum .moo~w 3c...H 8c...” Son ZN H 6.4 .+. m: H «Nc.c mfio.o mmc.c ado 8mm moq.~ mm nasas< moon... 8:.” 8c.” :2 H 62 .+. 83 .+. o~o.o mac.c -c.o 6N6.~ oom.4 ~86.~ 64 ease 88.“ 8:0. «och SN H SN H :3 .... mac.o mfic.o 44o.o 488.3 Nm8.fi 4co.~ «a «cam 25...” 8c...” 2c.“ as H a: H 23 .+. oNc.o hao.o hqo.c 34m.N amm.~ mam.m «N 962 vmcmnfiou onon nocmm vmcmnsoo maofim cocoa new cocoa way soon new mumv\mucwsmu:mmws see A co comma mm ~ H.x mum mmumm zmun .maodm cam Locos may nuon now wumv\mmaasmm m :0 vomws Mm a H.M.mum munumo3 kuOH .Amvooouumma mo m>mm=~oxmv zoom cofiumumum> mo A.u3 xuv uEv mumm :mwz vcm AmE\uEv ucwmmz ago fimuoe .~< mdnmfi 89 mm d H.x mum mwnmm >mum musuwos AnuOH acct” ~¢o.o coon” «mo.c macaw o~o.o ocean mmo.o 66:48560 Noon“ mqo.o ocean ~mo.o Noon“ o~c.o macaw wdo.c vacam macaw wmc.o moan“ o¢o.o coon“ “no.0 mfic4H «no.0 cocoa Aunwwwa hut wEv mama mmuo x .mnon cam Locos wnu :uon new mumc\moHaEmm n so woman mm fl H.x mum m8fi.u moo.4 new.“ 6mm.~ N84.“ 886.“ :8 H one.“ cmcmneoo maa.n omm.a won.“ mwh.4 mom.” cow.a cod.“ wqm.a moodm mm~.H 64m.~ oma.n mua.n owe.“ mmo.¢ 3mm.“ a-.m nocmm mE \ Auev usumm3 zuv amum .waodw cam Loco; wzu Luca wow muwv\mucmeu:mmoE com A co woman .mmum acmumumum> -< «o A.u3 xun wEv owwm com: com AmE\mEv uswmoz mun Hmuoe MN um=u=< an 9426 Qd MESH. NN >6: .w< mfinmh 90 Table A9. Mean Nunber arfl Dry Weight (per mz) of Maior Prey Groups Found in the Sedinents. N = 5 sanples per date. Sedimert Prey #I‘m2 (mg dry wt.) Ex 21 June 14 July 19 August 23 CIBCbCEra 181 (1) 120 (1) 843 (9) 310 (9) Ostracoda 783 (22) 60 (2) 1,385 (33) 181 (5) Hyalella 120 (18) ** 181 (2) ** Hydracarina 241 (18) 241 (18) 241 (16) 602 (34) Chironanidae 2,769 (92) 5,538 (106) 11,378 (190) 12,522 (99) Other Diptera 4,515 (25) 2,167 (17) 6,742 (18) 2,829 (183) Tricoptera ** ** ** 60 (l) Emeneroptera 421 (296) 60 (63) ** 783 (130). Odonata 120 (542) ** *1: *2? Total Insecta 8,067 (542) 7,766 (129) 18,602 (208) 16,194 (417) Gastropoda 361 (356) 181 (147) Mr 60 (12) Total Prey 10,595 8,669 22,876 18,361 (962) (299) (278) (486) “absert 91 Aucamm3 >uw «EV wNmm xmuo x + 43c.c hfim + omo.c em3 :55 H mm H Nfic.c mum 44c.c + c4~.H qmc.c cow cmc.o + 63m.“ fimc.c moo NE \ AuEv usumw3 xuv zmum .mumv\mucmsmu:mmwe cqfiA co comma mm ~.H x who momma moum .mumv\mmHaEmm m co comma mw fl H.x mum manuams dquH .xmum ucmfimvmm -< mo A.u3 mac mEv mumm cum: vcm ANE\aEv uSMMmB zun HmuOH mm umsao< on 9456 84 66:4 Nu hm: .c~< ofinmfi 92 habitats have also been recorded in other studies (Wohlschlag 1950, Gerking 1962) and presumably these differences reflect both a lack of micro-habitat diversity and an increased predator vulnerability for invertebrates in open areas. Prey types notably reduced or absent in the bare sediments relative to the vegetation were cladocera, copepoda, Hyalella and many insect groups. Total biomass in the sediments was at a peak in May, declined by approximately 3 fold in June and July and then increased in August. This pattern in total biomass was paralleled by changes in the mean dry weight of prey over this same time period (Table A10). . DISCUSSION This study represents the first attempt to quantify the total size distribution of invertebrate prey found in a natural lake. Results demonstrate that within each major habitat type the distribution of prey sizes closely matches the lognormal curve. The ability to characterize prey availabilities by a simple, two parameter distribution should prove extremely useful for general examinations of prey size selection and potential overlaps in resource utilization by different fish species or size classes (see Werner 1977 for a good example). The work of Werner (1977) represents the only other study which has attempted to determine the size distribution of prey in a freshwater system (a series of small experimental ponds). Werner also found the distribution of prey sizes (weights) to be lognormal and Schoener and Janzen (1968) and Hemmingen (1934) both report body length in terrestrial insect communities to be lognormally distributed. Thus, there is growing evidence that the body size distributions of insects and other invertebrates in both aquatic and terrestrial environments corresponds to the lognormal curve. 93 Whether this correspondence to the lognormal represents a common set of biological interactions or simply reflects statistical laws of large numbers, as May (1975) has shown to be the apparent case for many species-abundance relationships, remains to be determined. The seasonal patterns in prey size and abundance found in this study can be compared to the results of other workers who have measured prey availabilities in small temperate lakes. Table All summarizes the results of those studies which have estimated the total biomass of invertebrate prey found in littoral habitats. Results most directly comparable to this study are those of Wohlschlag (1950) who examined prey abundances in a small marl lake in Indiana. His estimates of total prey biomass are considerably less than those reported here for both the vegetation and bare marl habitats. WOhlschlag was unable to measure the abundance of gastrOpods and cladocera using his sampling technique, but it is doubtful that the inclusion of these prey would be sufficient to make up the difference in biomass between his results and those of this study. It is probable that his method of collection, which used a standard Eckman dredge, was much less efficient at sampling vegetation prey than was the device developed and used here. There is good agreement, however, between Wohlschlag's data and mine concerning the paucity of prey biomass in the bare marl as compared to vegetated habitats. Estimates of total prey biomass determined by Phoenix (1976) are also far below those measured in Lawrence Lake. Phoenix used a very small plexiglass cylinder (3.8 cm I.D.) to sample individual Eleeocharis stems and it is quite likely that large motile prey (e.g. Coenagrionidae, Baetis) were able to escape as this cylinder was lowered 94 .Esumuum suamc mic Eouw vmfioEwm Acmofi .mmofiv mummx osu wow mmmEOMn mumuo>< N .mmHaEmm cm cmv=~ucm uoc mcoaouummo .c~.o x; mamzfiamufisE >£ usumwB zap Cu vmuum>c00 muxumo3 uwB mm pmcmEuwuwv wmmEOMm H ow¢ flow saw and Nmm we umaus< mam quz owmz zed “mm mm saga oom mnqfi cm amm fixm mash New o~m~ am 0mm 4%«.2 Aqu mumm Amsaumowv Awuwxov Hume mum: Amaaumumv Ammumzoomamv cemuMuwum> acmumumum> camuMumum= cemumumwm> 96:86 muse Nfimoazv acaxumc zxamozv «635684563 Aokazc xfiamonm ANE\.u3 xuc uEv mmmEOMm xwum HmuoH .muwsoumwmwm mnemum> >n mwxmd HHmEm mo mumumnm: fimuOuumA mcu :« canon zmpm mumunouum>cH uo ANE\.u3 zap aEv mmmEOMm Hmuoe .-< macaw 95 over a plant. Phoenix also calculated prey biomass using "average weights for representative sizes of organisms" and suggested for this reason that his estimates of biomass are "useful only for showing general trends". The work of Gerking (1962) represents probably the best attempt at determining the biomass of available prey in terms of employing an efficient sampling device and a large number of samples. Clearly from Table All, there is good agreement between Gerking's estimates of prey biomass (measured in a northern Indiana lake) and those determined in this study. This agreement may be superficial, however, as Gerking studied a shallow eutrOphic lake whereas Lawrence Lake is mesotrophic and has a very low primary productivity (Wetzel 1975). Differences in productivity alone would suggest that prey biomass should have been much lower in Lawrence Lake as compared to Gerking's work. Again, it would appear that the sampling program employed in this present study was simply more efficient in collecting the total distribution of available prey species. Therefore, comparisons with other studies of aquatic invertebrate communities demonstrate the guccess of the present study in sampling the entire prey fauna and the sound basis upon which the general patterns of habitat specific prey abundances and size distributions are constructed. The high degree of predictability in the seasonal pattern of abundance of littoral invertebrates both within and between lakes (Ball and Hayne 1952, Gerking 1962, Keast 1977) indicates that the prey data obtained for the vegetation and sediment habitats should be generalizable to other aquatic systems. Plankton abundances and size distirbutions, however, are extremely variable yearly within a lake and 96 between lakes. Therefore, comparisons of the size and abundance of plankton in Lawrence Lake with those of other lakes would not be particularly informative. The average size distribution of the plankton has been included in this study primarily to demonstrate its truncated nature with respect to large prey sizes and to provide a complete picture of prey availabilities in all the major habitats within a single lake. The seasonal reduction in the abundance of large prey in both the vegetation and sediment habitats and the seasonal decline in the total biomass of vegetation prey (exclusive of gastrOpods) are striking features of the Lawrence Lake data. Other workers have reported a reduction in the total volume or biomass of littoral prey from May to August (Ball and Hayne 1952, Anderson and Hooper 1956, Phoenix 1976); but few studies have demonstrated how these changes can shift the size distributions of available prey. In Chapter 1 I have shown that these shifts in prey size distributions can dramatically effect the energetic intake and habitat use of a generalized forager, the bluegill sunfish (Lepomis macrochirus). Given the importance of prey size to the feeding efficiency and prey selection of many groups of predators (Schoener 1969, Wilson 1975, Werner 1977) it is clear that a thorough understanding of the patterns of prey sizes in aquatic communities will be crucial in interpreting resource utilization and potential competitive interactions between co-occurring fish species. This study provides initial data on these patterns of prey availability in a natural system. APPENDIX B CALCULATION OF PREY ENERGETIC CONTENTS Relationships between prey length (i) and energy content (e(i)) were determined for prey characteristic of each of the three habitats studied. Lengths were converted to weights by the following regressions: open water habitat, w = 0.012 i2-63; bare sediment habitat, w - 0.003 i2-35 (r 8 .98, n = 64); vegetation habitat, w = 0.005 i2°l6 (r - .93, n = 179). The open water regression is for 4 species of Daphnia (Burns 1969), the bare sediment regression is for Chironomus larvae (Mittelbach unpublished data), and the vegetation' regression contains a number of prey groups including: amphipods, caddisfly larvae, damselfly naiads, midge larvae, Sida, Simocephalus, copepods, Bosmina, and mayfly nymphs (Dumont et al. 1975, Werner and Hall unpublished data, Mittelbach unpublished data). Prey dry weights were then converted to calories using the following caloric equivalents: 5.0 cal/mg dry wt. for Daphnia (Richman 1958, Cummins and Wuycheck 1971), 5.5 cal/mg dry wt. for Chironomidae (Elliot 1976, Cummins and Wuycheck 1971), and 5.0 cal/mg dry wt. for vegetation prey (Cummins and Wuycheck 1971). 97 LITERATURE CITED LITERATURE CITED Aitchinson, J. and J. A. C. Brown. 1957. The lognormal distribution. Cambridge University Press, New York. Anderson, R. 0. and F. F. Hooper. 1956. Seasonal abundance and produc- tion of littoral bottom fauna in a southern Michigan lake. Transactions, American Microscopical Society 75:259-270. Ball, R. C. and D. W. Hayne. 1952. Effects of the removal of the fish population on the fish-food organisms of a lake. Ecology 33: 41-48. Belovsky, G. E. 1978. Diet optimization in a generalist herbivore: The moose. Theoretical Population Biology 14:105-134. Bohl, E. 1980. Diel pattern of pelagic distribution and feeding in planktivorous fish. Oecologia 44:368-375. Burns, C. W. 1969. Relation between filtering rate, temperature, and body size in four species of Daphnia. Limnology and Oceanography 14:693-700. Charnov, E. L. 1976. Optimal foraging; attack strategy of a mantid. American Naturalist 110:141-151. Confer, J. L., G. L. Howick, M. H. Corzette, S. L. Kramer, S. Fitz- gibbon, and R. Landesberg. 1978. Visual predation by planktivores. Oikos 31:27-37. Covich, A. P. 1976. Analyzing shapes of foraging areas: some ecological and economic theories. Annual Review of Ecology and Systematics 7:235—257. Cummins, K. W. and J. C. WUycheck. 1971. Caloric equivalents for invest- igations in ecological energetics. Mitteilungen, Internationale Vereinigung fUr Theoretische und Angewandte Linmologie. No. 18. Davies, N. B. 1977. Prey selection and the search strategy of the spot- ted flycatcher (Muscicapa striata): A field study on optimal foraging. Animal Behavior 25:1016-1033. Dumont, H. J., I. Van de Velde, and S. Dumont. 1975. The dry weight estimate of biomass in a selection of cladocera, copepoda and roti— fera from the plankton, periphyton and benthos of continental waters. Oecologia 19:75-97. 98 “i E— 99 Eggers, D. M. 1977. The nature of food selection by planktivorous fish. Ecology 58:46-59. Elliot, J. M. 1976. The energetics of feeding, metabolism and growth of brown trout (Salmo trutta L.) in relation to body weight, water temperature and ration size. Journal of Animal Ecology 45:923-948. Elliot, J. M. and W. Davison. 1975. Energy equivalents of oxygen consumption in animal energetics. Oecologia 19:195-201. Flannagan, J. F. 1970. Efficiencies of various grabs and corers in sampling freshwater benthos. Journal, Fisheries Research Board of Canada 27:1691-1700. Galbraith, M. G. 1967. Size-selective predation on Daphnia by rainbow trout and yellow perch. Transactions, American Fisheries Society 96:1-10. Gauch, H. G. and G. B. Chase. 1975. Fitting the gaussian curve in ecological applications. Ecology 55:1377-1381. Gerking, S. D. 1957. A method of sampling the littoral macrofauna and its application. Ecology 38:219-225. Gerking, S. D. 1962. Production and food utilization in a population of bluegill sunfish. Ecological Mbnographs 32:31-78. Glass, N. R. 1971. Computer analysis of predation energetics in the largemouth bass. In: B. C. Patten (ed.) Systems Analysis and Simulation in Ecology. Academic Press, New York. Goes-Custard, J. D. 1977. Optimal foraging and the size selection of worms by redshank (Tringa totanus). Animal Behavior 25:10-29. Hall, D. J., W. E. Cooper, and E. E. Werner. 1970. An experimental approach to the production dynamics and structure of freshwater animal communities. Limnology and Oceanography 15:829-928. Hall, D. J. and E. E. Werner. 1977. Seasonal distribution and abun- dance of fishes in the littoral zone of a Michigan lake. Trans- actions, American Fisheries Society 106:545-555. Hall, D. J., E. E. Werner, J. F. Gilliam, G. G. Mittelbach, D. Howard, C. G. Doner, J. A. Dickerman, and A. J. Stewart. 1979. Diel foraging behavior and prey selection in the golden shiner (Notemigonus crysoleucas). Journal, Fisheries Research Board of Canada 36:1029-1039. Hemmingsen, A. M. 1934. A statistical analysis of the differences in body size of related species. Videnskabelige Meddelelser Dansk naturhistorisk Forening Kobenhavn. 98:125-160. f‘_i_ 100 Hester, F. J. 1968. Visual contrast thresholds of the goldfish (Carassius auratus). Vision Research 8:1315-1336. Ivlev, V. W. 1961. Experimental ecology of the feeding of fishes. Yale University Press, New Haven, Connecticut. Keast, A. 1965. Resource subdivision amongst cohabiting fish species in a bay, Lake Opinicon, Ontario. Great Lakes Research Division, University of Michigan. Pub. No. 13. Keast, A. 1977. Mechanisms expanding niche width and minimizing intra- specific competition in two centrarchid fishes. In: M. K. Hecht, W. C. Steere and B. Wallace (eds.) Evolutionary Biology. Vol. 10. Plenum Press, New York and London. Keast, A. 1978. Feeding interrelations between age groups of pumpkin- seed sunfish (Lepomis gibbosus) and comparisons with the bluegill sunfish (L. macrochirus). Journal, Fisheries Research Board of Canada 35:12-27. Keast, A. and L. Welsh. 1968. Daily feeding periodicities, food uptake rates, and dietary changes with hour of day in some lake fishes. Journal, Fisheries Research Board of Canada 25:1133—1144. Kerr, S. R. 1971. Prediction of fish growth efficiency in nature. Journal, Fisheries Research Board of Canada 28:809-814. Krebs, J. R. 1978. Optimal foraging: Decision rules for predators. In: J. R. Krebs and N. B. Davies (eds.) Behavioural Ecology. Sinauer Associates, Inc., Sunderland. Krebs, J. R., J. T. Ericksen, M. I. Weber, and E. L. Charnov. 1977. Optimal prey selection in the Great Tit (Parus major). Animal Behavior 25:30-38. Laughlin, D. R. and E. E. Werner. 1980. Resource partitioning in two coexisting sunfish (Lepomis gibbosus and Lepomis megalotis pelt- astes). Journal, Fisheries Research Board of Canada (in press). Lynch, M. 1977. Fitness and optimal body size in zooplankton popula- tions. Ecology 58:763-774. Martin, N. Y. 1970. Long-term effects of diet on the biology of the lake trout and the fishery in Lake Openingo, Ontario. Journal, Fisheries Research Board of Canada 27:125-146. May, R. M. 1975. Patterns of species abundance and diversity. In: M. L. Cody and J. M. Diamond (eds.) Ecology and Evolution of Communities. Harvard University Press, Cambridge, Massachusetts. Neill, W. E. 1975. Experimental studies of microcrustacean competition, community composition and efficiency of resource utilization. Ecology 56:809-826. 101 O'Brien, W. J. 1979. The predator-prey interaction of planktivorous fish and zooplankton. American Scientist 67:572-581. Paloheimo, J. E. and L. M. Dickie. 1966. Food and growth of fishes. III. Relations among food, body size, and growth efficiency. Journal, Fisheries Research Board of Canada 23:1209-1248. Pearson, N. E. 1976. Optimal foraging: some theoretical consequences of different feeding strategies. Ph.D. Thesis, University of washington. Phoenix, D. R. 1976. Temporal dynamics of a natural multipredator- multiprey system. Ph.D. Thesis. University of Pennsylvania. Pyke, G. H., H. R. Pulliam, and E. L. Charnov. 1977. Optimal foraging: A selective review of theory and tests. Quarterly Review of Biology 52:137-154. Rich, P. H., R. G. Wetzel, and N. V. Thuy. 1971. Distribution, production and role of aquatic macrophytes in a southern Michigan marl lake. Freshwater Biology 1:3-21. Richman, S. 1958. The transformation of energy by Daphnia pulex. Ecological Monographs 28:273-291. Rosenzweig, M. L. 1974. On the optimal above ground activity of banner- tail kangaroo rats. Journal of Mammology 55:193-199. Sarker, A. L. 1977. Feeding ecology of the bluegill, Lepomis macrochirus, in two heated reservoirs of Texas. III. Time of day and patterns of feeding. Transactions, American Fisheries Society 106:596-601. Schoener, T. W. 1969. Mbdels of optimal size for solitary predators. American Naturalist 103:277—313. Schoener, T. W. 1974. Competition and the form of habitat shift. Theo- retical Population Biology 6:265-307. Schoener, T. W. 1976. Alternatives to Lotka-Volterra competition: models of intermediate complexity. Theoretical Population Biology 10: 309-333. Schoener, T. W. 1978. Effects of density-restricted food encounter on some single-level competition models. Theoretical Population Biology 13:365-381. Schoener, T. W. and D. H. Janzen. 1968. Notes on environmental determ- inants of tropical versus temperate insect size patterns. American Naturalist 102:207-224. Sih, A. (in prep.). Optimal behavior: Can foragers balance two conflicting demands? m 102 Stuntz, W. E. 1975. Habitat selection and growth of bluegills. Ph.D. Thesis. University of Wisconsin. Swingle, H. S. and E. V. Smith. 1940. Experiments on the stocking of fish ponds. Transactions, American Wildlife Conference 5:267-276. Tilman, D. 1980. Resources: A graphical-mechanistic approach to compet- ition and predation. American Naturalist (in press). Vinyard, G. L. and W. J. O'Brien. 1976. Effects of light and turbidity on the reactive distance of bluegill (Lgpomis macrochirus). Journal, Fisheries Research Board of Canada 33:2845-2849. Wankowski, J. W. J. and J. E. Thorpe. 1979. The role of food particle size in the growth of juvenile Atlantic salmon (Salmo salar L.). Journal of Fisheries Biology 14:351-370. Ware, D. M. 1971. Predation by rainbow-trout (Salmo gairdneri): the effect of experience. Journal, Fisheries Research Board of Canada 28:1847-1852. Ware, D. M. 1972. Predation by rainbow trout (Salmo gairdneri): the influence of hunger, prey densities and prey size. Journal, Fish- eries Research Board of Canada 29:1193—1201. Ware, D. M. 1975. Growth, metabolism, and Optimal swimming speed of a pelagic fish. Journal, Fisheries Research Board of Canada 32: Wenger, A. 1972. A review of the literature concerning largemouth bass stocking techniques. Texas Parks and Wildlife Department: Technical Series No. 13, Sheldon, Texas. Werner, E. E. 1974. The fish size, prey size, handling time relation in several sunfishes and some implications. Journal, Fisheries Re- search Board of Canada 31:1531—1536. Werner, E. E. 1977. Species packing and niche complementarity in three sunfishes. American Naturalist 111:553-578. Werner, E. E. and D. J. Hall. 1974. Optimal foraging and the size selection of prey by the bluegill sunfish (Lepomis macrochirus). Ecology 55:1042-1052. Werner, E. E. and D. J. Hall. 1976. Niche shifts in sunfishes: experimental evidence and significance. Science 191:404-406. Werner, E. E. and D. J. Hall. 1977. Competition and habitat shift in two sunfishes (Centrarchidae). Ecology 58:869-876. Werner, E. E. and D. J. Hall. 1979. Foraging efficiency and habitat switching in competing sunfishes. Ecology 60:869-876. 103 Werner, E. E., G. G. Mittelbach, and D. J. Hall. 1980. Foraging profitability and the role of experience in habitat use by the bluegill sunfish. Ecology (in press). Werner, E. E., D. J. Hall, D. R. Laughlin, D. J. wagner, L. A. Wilsmann, and F. C. Funk. 1977. Habitat partitioning in a freshwater fish community. Journal, Fisheries Research Board of Canada 34:360-370. Werner, P. A. and H. Caswell. 1977. Population growth rates and age versus stage-distribution models for teasel (Dipsacus sylvestris Huds.) Ecology 58:1103—1111. Werner, R. G. 1966. Ecology and movements of bluegill sunfish fry in a small northern Indiana lake. Ph.D. Thesis. Indiana University. Wetzel. R. G. Limnology. W. B. Saunders Company, Philadelphia. Wilsmann, L. A. 1979. Resource partitioning and mechanisms of co- existence of blackchin and blacknose shiners (Notropis:Cyprinidae). Ph.D. Thesis. Michigan State University. Wilson, D. S. 1975. The adequacy of body size as a niche difference. American Naturalist 109:769-784. Wohlschlag, D. E. 1950. Vegetation and invertebrate life in a marl lake. Investigations of Indiana Lakes and Streams 3:321—372. Wohlschlag, D. E. and R. 0. Juliano. 1959. Seasonal changes in bluegill metabolism. Limnology and Oceanography 4:195-209. "111111111111111