THE PREDATORY BEHAVIOUR OF RAINBOW TROUT (SALMO G A I R D N E R I ) by DANIEL M. WARE B . S c , U n i v e r s i t y o f B r i t i s h C o l u m b i a , 1 9 6 6 A T H E S I S SUBMITTED IN P A R T I A L FU L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY i n t h e D e p a r t m e n t o f ZOOLOGY Ule a c c e p t t h i s t h e s i s a s c o n f o r m i n g t o t h e r e q u i r e d s t a n d a r d THE U N I V E R S I T Y OF B R I T I S H COLUMBIA J A N U A R Y , 1 9 7 1 In p r e s e n t i n g t h i s t h e s i s in p a r t i a l f u l f i l m e n t o f the requirements f o r an advanced degree at the U n i v e r s i t y of B r i t i s h Columbia, I agree that the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r re ference and study . I f u r t h e r agree t h a t permiss ion fo r e x t e n s i v e copy ing o f t h i s t h e s i s f o r s c h o l a r l y purposes may be granted by the Head of my Department or by h i s r e p r e s e n t a t i v e s . It i s understood that copying o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l ga in s h a l l not be a l lowed wi thout my w r i t t e n p e r m i s s i o n . Department of Z 6 D L O 6 V The U n i v e r s i t y o f B r i t i s h Columbia Vancouver 8. Canada Date ABSTRACT The predatory behaviour of rainbow trout was studied to ident i fy some of the major factors that inf luence the i r response to prey. Two b e n t h i c - l i v i n g amphipods Cranqonyx sp. and Hyale l la sp. were selected as representative prey. In some experiments, a r t i f i c i a l food was u t i l i z e d . Rainbow trout adopt a searching pos i t ion some 10 to 15 cm from a substrate and locate food v i s u a l l y . As a re su l t , they can detect only organisms that are exposed. In the presence of a complex substrate, trout were able to recognize moving prey with greater success (74%) than stat ionary targets (39$) with the same v i sua l c h a r a c t e r i s t i c s . The distance from which trout w i l l react to food was shown to be dependent upon the s i ze , inherent contrast and a c t i v i t y of the object as well as the ambient i l l u m i n a t i o n , t u r b i d i t y of the water and complexity of the substrate. After 6 to 7 days of experience with a new but palatable food, trout can increase t h e i r react ive distance through l ea rn ing . A general system of equations was developed to describe the effect of each of these parameters on react ive d i s tance . On the average, trout successful ly capture 82$ of the prey they attack. In the laboratory, the rate of capture reached a maximum leve l when the density of prey was increased to 240 animals per sq. m. Irrespect ive of the abundance of food, however, decreasing hunger motivation was found to depress the predator ' s rate of capture as was the presence of a substrate in which the-prey could conceal themselves. The effect of water temperature on the v e r t i c a l and hor izonta l movements of Cranqonyx and Hya le l l a was also examined. The v e r t i c a l a c t i v i t y of both prey increased exponential ly with a r i s e in temperature. In contrast , 1 0 ° C. was suggested to be the optimum temperature for the movement of exposed animals. A general s imulation model was developed to test the hypothesis that the s e l ec t ive exp lo i ta t ion of 4 major invertebrate groups in Marion Lake, by t rout , occurs at the perceptual l e v e l . The model considered the predatory behaviour of the f i sh as well as the density and phys ica l c h a r a c t e r i s t i c s of the i r prey, and was able to predict with some accuracy the occurrence of d i f ferent foods in trout stomachs. The model was also able to account for the s ize -se l ec t ive exp lo i ta t ion of Cranqonyx and H y a l e l l a , the seasonal changes in the v u l n e r a b i l i t y of these species, and the fact that the less numerous Cranqonyx was captured just as frequently as H y a l e l l a . Trout require a threshold rate of capture (about 2 captures / min.) to maintain a s p e c i f i c pattern of search. If they do not a t ta in th i s threshold they w i l l switch the i r at tent ion to other hunting patterns . As a re su l t , the population should converge, temporari ly, into areas in which food i s r e l a t i v e l y more abundant. Since trout can also learn to increase the i r responsiveness to prey, both of these c h a r a c t e r i s t i c s would improve the i r hunting e f f i c i e n c y . The resul t s of th i s study indicate that v i sua l predators w i l l discover and,subsequently may explo i t ,1arge prey that tend to be exposed and ac t ive , with greater success than smaller, less act ive or less conspicuous species . Moreover, i f a v i sua l predator maintains a searching pos i t i on , i t may not detect benthic-l i v i n g food organisms less than a c r i t i c a l s i z e . The s ign i f i cance of these conclusions i s discussed. ABSTRACT i LIST OF TABLES v i i LIST OF FIGURES ix ACKNOWLEDGEMENTS x i i GENERAL INTRODUCTION 1 SECTION I THE EFFECT OF PREY DENSITY, PREY SIZE AND THE PRESENCE OF A SUBSTRATE ON THE FEEDING BEHAVIOUR OF TROUT INTRODUCTION 3 METHODS' AND MATERIALS 5 RESULTS 1 1 GENERAL FEEDING BEHAVIOUR 11 The Effect of Hunger and Prey Density on the Rate of Attack 12 The Relationship Between the Rate of Attack and Prey Size 18 THE EFFECT OF A SUBSTRATE ON THE ATTACK RATE 22 PREY CAPTURE SUCCESS 30 THE THRESHOLD RATE OF PREY CAPTURE AND THE SEARCHING PATTERN 3 4 DISCUSSION 39 SUMMARY 4 4 THE E F F E C T OF E X P E R I E N C E ON THE RESPONSE OF TROUT TO UNFAMIL IAR PREY INTRODUCTION 47 METHODS AND M A T E R I A L S 48 RESULTS 50 THE C H A R A C T E R I S T I C S OF THE I N I T I A L RESPONSE OF TROUT TO UNFAMIL IAR PREY 50 THE E F F E C T OF E X P E R I E N C E ON REACT IVE D ISTANCE , 53 THE S P E C I F I C I T Y OF THE RESPONSE OF CONDITIONED TROUT. 57 THE E X T I N C T I O N AND RE -DEVELOPMENT OF R E A C T I V E DISTANCE 63 ATTENTION COMPETIT ION 63 D ISCUSSION 66 SECTION I I I PREY A C T I V I T Y AND V U L N E R A B I L I T Y INTRODUCTION 73 METHODS AND MATERIALS 7 3 F I E L D STUDIES 73 LABORATORY STUDIES 7 5 RESULTS 76 THE E F F E C T OF WATER TEMPERATURE ON THE V E R T I C A L A C T I V I T Y OF CRANGQNYX AND H Y A L E L L A 7 6 THE E F F E C T OF TEMPERATURE ON THE A C T I V I T Y OF EXPOSED AMPHIPODS 83 A SIMULATION MODEL OF THE PREDATORY BEHAVIOUR OF TROUT INTRODUCTION 87 METHODS AND MATERIALS ....89 RESULTS 92 THE CHARACTERISTICS OF THE VISUAL RESPONSE OF TROUT TO PREY ,92 The Relation Between Prey Size and Contrast Threshold 92 The Relation Between the Ambient Illumination, Visual Angle and Contrast Threshold 99 The Effect of Prey Movement on Reactive Distance 102 The Relation Between the Background, Reactive Distance and Target Recognition Success ...105 The Effect of Prey Activity and the Searching Position on the Width of the Path of Search 1° 9 THE ATTACK MODEL H5 APPLICATIONS OF THE MODEL 1 1 8 The selective Exploitation of Amphipods, Odonates, Planorbids and Caddis 118 The Size Selective Exploitation of Amphipods 124 Seasonal Changes in the Exploitation of Amphipodsl29 DISCUSSION 131 SUMMARY i 4 1 BIBLIOGRAPHY 144 APPENDIX 152 I. LIST OF SYMBOLS FOR THE ATTACK MODEL 152 II. PHYSICAL CHARACTERISTICS OF MARION LAKE ..154 III. POPULATION CHARACTERISTICS OF CRANGONYX AND HYALELLA 156 LIST OF TABLES SECTION I Table 1. Charac te r i s t i c s of the experimental substrates 8 Table 2. A representation of the experiments conducted 9 Table 3. Average lengths and dry weights of prey used in the predation experiments 10 Table 4. Regression constants for the rate of attack at time (t) with respect to the hunger l eve l of the trout and the density of prey 30 seconds e a r l i e r 17 Table 5. The re la t ionsh ip between react ive distance and prey s ize 21 Table 6. A comparison of the effect of the d i f ferent substrate treatments on the proportion of Crangonyx and Hyale l l a that were exposed and captured during an experiment 28 Table 7. The proportion of the to ta l number of s t r ikes which success ful ly terminated with the capture of prey 32 SECTION II Table 1. The re l a t ionsh ip between the i n i t i a l and conditioned react ive distance 56 Table 2. The effect of hunger on the react ive distance of 3 trout conditioned to white prey 71 SECTION III. Table 1. The v e r t i c a l d i spersa l a c t i v i t y of Crangonyx and H y a l e l l a , in Marion Lake, with respect to several environmental condit ions 78 Table 2. A. The re la t ionsh ip between the ambient water temperature and the instantaneous proportion of Cranqonyx and Hya le l l a exposed at or above the mud-water interface 86 B. The re la t ionsh ip between the ambient water temperature and the proportion of time exposed amphipods spend moving 86 LIST OF TABLES SECTION IV Page Table 1. The re la t ionsh ip between prey s ize and the average react ive distance 95 Table 2. A comparison of several documented values of the minimum detectable contrast and the minimum vi sua l angle of) d i f ferent animals . . . A u u Table 3. The effect of background d i v e r s i t y on prey recognit ion success J-07 Table 4. The effect of background d i v e r s i t y on react ive distance J-08 Table 5. Values for the parameters of the attack modelJ.14 Table 6. The population c h a r a c t e r i s t i c s of the odonates, caddis , and planorbids for several selected months during an 'average' year 120 Table 7. A. A comparison between the expected and observed percentage occurence of d i f ferent prey groups in trout stomachs 122 B. A comparison between the observed and predicted percentage occurence of d i f ferent prey groups in trout stomachs 123 Table 8. A comparison of the f i t between the expected, predicted and actual s ize composition of Cranqonyx found in trout stomachs in the month of November 128 Table 9. S e n s i t i v i t y of the attack model to selected parameters -133 APPENDIX Table 1. Temporal changes in the size structure of amphipods in Marion Lake .158 L I S T OF F IGURES 1 X SECTION I P a g e F i g u r e 1. The p r o g r e s s i v e d e c l i n e i n t h e r a t e o f a t t a c k 1 6 F i g u r e 2 . The s e a r c h i n g v e l o c i t y o f Rainbouu t r o u t i n s e v e r a l c o n t r o l e x p e r i m e n t s 20 F i g u r e 3 . The a v e r a g e t i m e r e g u i r e d f o r 4 e x p e r i m e n t a l p o p u l a t i o n s o f C r a n g o n y x t o r e a c h an e q u i l i b r i u m l e v e l o f e x p o s u r e 23 F i g u r e 4 . The r e l a t i o n s h i p b e t w e e n t h e n u m b e r o f C r a n g o n y x c a p t u r e d i n d i f f e r e n t h a b i t a t s a f t e r 50 m i n u t e s e x p o s u r e t o t r o u t p r e d a t i o n and t h e i r i n i t i a l d e n s i t y 26 F i g u r e 51 The r e l a t i o n s h i p b e t w e e n t h e n u m b e r o f H y a l e l l a c a p t u r e d i n d i f f e r e n t h a b i t a t s a f t e r 50 m i n u t e s e x p o s u r e t o t r o u t and t h e i r i n i t i a l d e n s i t y . . . 27 F i g u r e 6 . The r e l a t i o n s h i p b e t w e e n t h e d e n s i t y o f e x p o s e d p r e y a n d . t h e a t t a c k r a t e 33 F i g u r e 7 . A s c h e m a t i c r e p r e s e n t a t i o n o f t h e p a t t e r n o f b e n t h i c s e a r c h i n g b e h a v i o u r by t r o u t d u r i n g an e x p e r i m e n t 36 F i g u r e 8 . The r e l a t i o n s h i p b e t w e e n t h e t h r e s h o l d r a t e o f p r e y c a p t u r e a n d t h e p r e d a t o r ' s s t a t e o f h u n g e r 37 F i g u r e 9 . The r e l a t i o n s h i p b e t w e e n t h e e x p e r i m e n t a l s u b s t r a t e a n d t h e r a t e o f e x t i n c t i o n o f t h e b e n t h i c s e a r c h i n g p a t t e r n a f t e r t h e i n i t i a l p h a s e o f c o m p l e t e a t t e n t i o n 38 SECTION I I F i g u r e 1. The e f f e c t o f e x p e r i e n c e on t h e f e e d i n g t i m e and r e a c t i v e d i s t a n c e o f 2 g r o u p s o f t r o u t . . . . 52 F i g u r e 2 . The e f f e c t o f e x p e r i e n c e on t h e r e a c t i v e d i s t a n c e o f 6 t r o u t 55 F i g u r e 3 . The e f f e c t o f s w i t c h i n g t r o u t , c o n d i t i o n e d t o w h i t e p r e y , t o p r e y w i t h d i f f e r e n t l e v e l s o f c o n t r a s t 62 F i g u r e 4 . The r e - d e v e l o p m e n t o f t h e r e a c t i v e d i s t a n c e o f 4 t r o u t 6 4 SECTION III Page Figure 1. The effect of water temperature on the proportion of Cranggnyx that are exposed at or above the mud-water interface 79 Figure 2. The effect of water temperature on the proportion of Hyalella that are exposed at or above the mud-water interface 80 Figure 3. The relationship between the actual density of amphipods and their vulnerable density 82 Figure 4. The relationship between the water temperature and the average amount of time exposed Cranqonyx and Hyalella spend moving 85 SECTION IV Figure 1. A diagramatic representation of the relationship between the contrast threshold and the visual angle 97 Figure 2. A comparison between the observed reactive distance of 4 trout, exposed to different sizes of prey, and the calculated reactive distance .. 103 Figure 3. The effect of target movement on reactive distance 1 0 4 Figure 4. The geometric relationship between the radius of the reactive f i e l d , the trout's searching position and the width of their searching path along the sediment HI Figure 5. A schematic flow diagram of the parameters and computational steps in the attack model H3 Figure 6. A comparison of the observed, expected and the predicted distribution of different size classes of Hyalella found in trout stomachs at two different sampling periods 127 Figure 7. A comparison of the simulated and observed trend in the exploitation of Cranqonyx and Hyalella, by trout in Marion Lake .130 Figure 1. The average seasonal trend in water temperature in Marion,, Lake 155 Figure 2. The r e l a t i v e density of Hya le l l a and Crangonyx in Marion ' Lake 157 ACKNOWLEDGEMENTS I am espec ia l ly grateful to Dr. Ian E . E f ford , my supervisor , for his interes t and encouragement during the course of th i s study. I mould also l i k e to thank Drs . P. Lark in , C. H o l l i n g , R. L i l e y , T. Northcote and Mr. N. G i lber t for o f fer ing valuable c r i t i c i s m of the manuscript. Many people have aided me throughout th i s study, I would p a r t i c u l a r l y l i k e to acknowledge the help of Mr. P. Pearlstone and Mr. B . Delury. F inanc ia l support came from the Canadian International B i o l o g i c a l Program, the National Research Council of Canada and the Univers i ty of B r i t i s h Columbia, Department of Zoology. GENERAL INTRODUCTION Many laboratory studies have been conducted to describe the feeding behaviour of animals (De Ruiter, 1966); very few of these,however, have attempted to predict how natural populations of predators will exploit different species of prey. Until very recently, the method of how to systematically relate laboratory studies to the field was rather elusive. In 1966 however, Holling described what he called the "experimental components analysis". This approach is based upon the premise that a biological process, such as predation, can be broken down into fragments. These components can then be studied under controlled conditions to elucidate their importance and relationship with other parts of the process. The structure of the resulting system is not assumed to be a complete explanation or description but rather is designed to be continually modified in the face of new observations and experimental results. In essence then, the system becomes a working hypothesis that can be tested either in the laboratory or on natural populations. In this study I have attempted, through laboratory experiments, to analyse the feeding behaviour of rainbow trout. The f i r s t section of the study is devoted to a general description of their behaviour and some of the major components which affect i t . This analysis is continued in the second section and takes the form of a single question: can trout learn to alter their response to prey? In the third section, I will describe the effect of water temperature on the activity patterns of two species of amphipods (Cranqonyx richmondensis and Hyalella azteca) that are natural prey of rainbow trout , and develop a v u l n e r a b i l i t y submodel. The f i n a l section w i l l examine the re la t ionsh ip between the v i sua l c h a r a c t e r i s t i c s of prey in general and the process of prey de tec t ion . Various aspects from the manuscript w i l l then be coupled into a s imulation model to attempt to account for ' the s e l ec t ive exp lo i ta t ion of several major invertebrates , but e spec ia l ly the amphipods, by the trout population in Marion Lake. THE EFFECT OF PREY DENSITY, PREY SIZE AND THE PRESENCE OF A SUBSTRATE ON THE FEEDING BEHAVIOUR OF TROUT INTRODUCTION In most aquatic systems i t appears as a genera l iza t ion that f i s h tend to exploit exposed prey to a greater extent than less conspicuous species (Grimas, 1963; A l l e n , 1941). There are other factors , however, besides the degree of exposure which w i l l determine the rate predators exploi t prey. These factors can be divided into 3 s p e c i f i c categor ies : 1) the c h a r a c t e r i s t i c s of the prey ( i e . s i ze , density and behaviour), 2) the feeding behaviour of the predator (i.e. searching behaviour, and the mechanisms i t u t i l i z e s to locate and capture food), 3) the c h a r a c t e r i s t i c s of the environment (i.e. the ambient i l l u m i n a t i o n , temperature and physical s t ruc ture ) . Al len (1941) stressed the importance of the c h a r a c t e r i s t i c s of prey. He suggested that since d i f ferent species have d i f ferent d i spersa l and behavioural patterns, are d i f ferent s izes and exist at d i f ferent dens i t i e s , they are not l i k e l y to have the same p r o b a b i l i t y of being detected and captured by predators . The t a c t i c s that predators employ to locate and handle food also contribute to determining predation rates (Ivlev, 1961; H o l l i n g , 1966). These t a c t i c s , however, may be modified through hunger motivation and in some cases through learning (Ho l l ing , 1965; H o l l i n g , 1966; Beukema, 1968). There i s also considerable evidence to indicate that predation is affected by environmental factors such as the ambient i l lumina t ion ( A l i , 1959; Hunter, 1968; Blaxter, 1968a) and the physical complexity of the environment ( i v l e v , 1961; Macan, 1966; Johannes and Lark in , 1961). The purpose of th i s study was to examine the effect of 4 f ac tor s : l ) prey densi ty , 2) prey s ize , 3) hunger motivation and 4) the presence of a substrate on the feeding behaviour of rainbow trout (Salmo q a i r d n e r i ) . Feeding experiments were rep l i ca ted with 2 d i f fe rent species of amphipods as prey (Cranqonyx r.ichmondensis and Hyale l la azteca) . The study was divided into 4 parts to systematical ly construct a descr ip t ive equation of the effect of each of these components on the behaviour of t rou t . The f i r s t 3 sect ions w i l l consider the influence of the aspects mentioned above on the rate of prey capture, while the f i n a l section w i l l examine the re l a t ionsh ip between the t r o u t ' s rate of capture and searching behaviour. F o u r r a i n b o w t r o u t , o b t a i n e d f r o m M a r i o n L a k e , B r i t i s h C o l u m b i a , w e r e u s e d i n t h i s s t u d y . They r a n g e d i n l e n g t h f r o m 1 3 . 4 t o 1 7 . 0 c m . E a c h f i s h was h e l d i n i s o l a t i o n b e t w e e n e x p e r i m e n t s i n a 227 l i t r e (50 g a l l o n ) g l a s s a q u a r i u m . Two o t h e r i d e n t i c a l t a n k s w e r e u t i l i z e d f o r p r e d a t i o n e x p e r i m e n t s . In b o t h t h e h o l d i n g a n d e x p e r i m e n t a l t a n k s , t h e w a t e r t e m p e r a t u r e was m a i n t a i n e d a t 10 C ( - 2 C ) . A l t h o u g h t h e b a c k g r o u n d i l l u m i n a t i o n was n a t u r a l , i t was ' c o n t r o l l e d ' by a v o i d i n g d i r e c t l i g h t a n d c o n d u c t i n g e x p e r i m e n t s a t t h e same t i m e o f d a y ( 1 2 0 0 t o 1 4 0 0 h r s P . S . T . ) . F o u r d i f f e r e n t s u b s t r a t e s w e r e e m p l o y e d t o t e s t t h e e f f e c t o f p h y s i c a l c o m p l e x i t y on p r e d a t i o n . The c h a r a c t e r i s t i c s o f e a c h o f t h e s e t r e a t m e n t s a r e d e s c r i b e d i n T a b l e 1 . The c o n t r o l s u b s t r a t e was s i m p l y a b a r e , g r e y c o l o r e d s u r f a c e . The l i t t e r I a n d I I t r e a t m e n t s , o n t h e o t h e r h a n d , w e r e c o m p o s e d o f l a r g e p i e c e s o f s t i c k l i t t e r a n d c o v e r e d , r e s p e c t i v e l y , 6% a n d 15% o f t h e t a n k b o t t o m . The f i n e l i t t e r t r e a t m e n t was t h e most c o m p l e x u t i l i z e d a s i t c o v e r e d t h e e n t i r e f l o o r o f t h e t a n k . T h i s m a t e r i a l was o b t a i n e d by s c r e e n i n g s e d i m e n t f r o m M a r i o n L a k e t o r e m o v e b o t h t h e v e r y f i n e a n d t h e v e r y c o a r s e p a r t i c u l a t e m a t t e r . T h i s was n e c e s s a r y t o f a c i l i t a t e t h e r e c o v e r y o f p r e y a n d t o i n s u r e t h a t t h e s u b s t r a t e w o u l d s e t t l e r a p i d l y i f i t was d i s t u r b e d by t h e f e e d i n g a c t i v i t y o f t h e p r e d a t o r . The b e h a v i o u r o f t r o u t was e x a m i n e d a t a m i n i m u m o f 3 di f ferent dens i t ies of prey in each of the 4 substrate treatments. Predation experiments were rep l ica ted independently for both species of amphipods (Table 2) . The response of trout to juveni le Cranqonyx was also invest igated but only in the l i t t e r II treatment. As a re su l t , the effect of prey s ize on predation could be assessed by comparing the resul t s of these experiments with those obtained when adult Cranqonyx were in the same s i t u a t i o n . To observe the feeding behaviour of t rout , in d i f ferent states of hunger motivation, and yet insure that they would not capture enough prey to reach s a t i a t i o n , d i f ferent dens i t ies of Cranqonyx and Hyale l la were u t i l i z e d . Prel iminary resul t s indicated that the trout would become sat iated after consuming about 90 standard sized adult Cranqonyx (Table 3) . To reach the same state, they would have to capture over 200 smaller, H y a l e l l a . Therefore, the density of each prey was regulated so the number captured during an experiment did not exceed these respective l i m i t s . The range in density of juveni le Cranqonyx was i d e n t i c a l to that choosen for Hya le l l a , as they are both of s imi l a r s ize and to some extent weight (Table 3) . The hunger leve l of the trout was standardized by depriving them of food for 48 hours before an experiment (in a few instances t h i s period lasted as long as 96 hour's). Some prel iminary resul t s indicated that the f i sh required 50 to 60 hours at 10 C. to completely digest a s a t i a t ion r a t i o n . Since the amount of food they were able to ingest in v i r t u a l l y every experiment was well below the i r s a t i a t ion l e v e l , 48 hours of depr ivat ion should have been adequate to c lear the digestive- t ract of a l l food material between succesive feedings. The experimental procedure consisted of gathering the reguired number of prey from Marion Lake and holding them, without food, in p l a s t i c containers for up to 24 hours. The feeding tank was prepared by adding a standard sample of substrate material and spreading i t uniformly over the bottom. The prey were then introduced and allowed 60 minutes to disperse before a predator was released. Spec i f ic aspects of the troutJs feeding behaviour were recorded chronolog ica l ly on a Rustrack, 4 channel recorder (Model 921). The experiments were terminated after 50 minutes at which time any prey remaining in the tank were recovered and counted. This res idual density was subtracted from the i n i t i a l density to determine the number of amphipods captured. The recovery technique was tested and was found to be 97% to 100^ e f f i c i e n t in recovering prey, therefore, no correct ion was made for any loss of animals during th i s operat ion. Treatment No. obj e c t s Average Surface2 Area of tank or depth * object area (cm ) bottom covered s i z e (cm 2) Con t r o l L i t t e r I L i t t e r II Fine L i t t e r 0 34 47 4mm, (6x1x1) cm. (9x2x1) cm. (0.4-0.7) mm. 4180 4700 5487 4180 + 0 253 640 4180 *see t e x t CD TABLE 2 . A r e p r e s e n t a t i o n o f t h e e x p e r i m e n t s c o n d u c t e d t o d e t e r m i n e t h e e f f e c t o f a s u b s t r a t e and p r e y d e n s i t y on t h e f e e d i n g b e h a v i o u r o f t r o u t . The c o l u m n f i g u r e s i n d i c a t e t h e number o f r e p l i c a t e e x p e r i m e n t s c o n d u c t e d w i t h d i f f e r e n t f i s h . P r e y D e n s i t y A d u l t C r a n q o n y x A d u l t H y a l e l l a J u v e n i l e C r a n q o n y x S u b s t r a t e 20 40 70 100 2 0 0 20 40 70 100 200 40 100 200 C o n t r o l - 2 2 - - 2 2 1 2 2 - - -L i t t e r I 1 2 2 1 - - 2 - 2 2 - - -L i t t e r I I - 2 2 2 - - 2 - 2 2 2 2 2 F i n e L i t t e r - 2 - 2 2 - 2 - 2 2 - - -N o . e x p . 22 27 6 TABLE 3. Average lengths and dry weights of i n d i v i d u a l prey used in the predation experiments. Prey *X Length (mm) * X Dry weight (mg) Adult Crangonyx 8.1 +_ 1 2.6 Juvenile Crangonyx 4.6 _+ 1 0.5 Adult Hya le l l a 5.7 + 1 1.0 * the range in length i s indicated RESULTS GENERAL FEEDING BEHAVIOUR Rainbow trout appear to locate food v i s u a l l y . This was v e r i f i e d by observing the behaviour of both wild and experimental f i s h . In e i ther case, t h e i r attack response comprised three d i s t i n c t steps; o r i e n t a t i o n , v i sua l f i x a t i o n , and a rapid , d i rec t attack. This i s a common pattern and has been observed for other v i sual predators (Messenger, 1968; 'Hoi1ing, 1966). When trout hunt for benthic prey, e i ther in the laboratory or in the f i e l d , they search from a pos i t ion some 10 to 15 cm above the substrate but or ient downward to face i t . This c h a r a c t e r i s t i c i n c l i n e (about 10 to 20 degrees) might be to d kirect t h e i r v i sua l axis onto the sediment, since i t i s s l i g h t l y obligue to the long i tud ina l body axis (Polyak, 1957). A f i sh w i l l search from thi s c h a r a c t e r i s t i c pos i t ion with monocular v i s i o n . - As a r e su l t , i t w i l l detect most targetsat an angle ( l a t e r a l ) to i t s path of search. Once a target has been detected the trout w i l l or ient to face the object and then pause momentarily to f ixate i t with binocular v i s ion before a t t ack ing . When the attack occurs, i t i s rapid and i s followed immediately by a s t r i k e and engulfment of the prey. The predator w i l l then return to i t s searching pos i t ion before i t resumes hunting. A complete attack sequence may require only 2 seconds. Rainbow trout are not always successful in d i scr iminat ing between prey and ' s i m i l a r ' objects , they w i l l s t r ike at pieces of s t i ck l i t t e r . When they react to an inanimate object they w i l l s t r ike at and re ject i t several times before the i r attack response i s terminated. This indicates that trout probably re ly upon ei ther t a c t i l e or chemical d i sc r iminat ion to d i s t i n g u i s h between food and ined ib le ob jects . Even th i s very cursory descr ip t ion of t h e i r feeding behaviour implies that as long as trout hunt from a pos i t ion some distance from the sediment and re ly upon v i s ion to locate food they w i l l detect only exposed animals. This supposit ion w i l l be examined in more d e t a i l below. The Effect of Hunger and Prey Density on the Rate of Attack .(Control Experiments) It i s well documented that the feeding motivation of many animals i s affected by food depr iva t ion . Ishiwata (1968b) found that the amount of food rainbow trout would consume, feeding acj l ib i tum after various periods of food depr ivat ion, was inverse ly related to the amount which remained in the stomach from the previous meal. If hunger i s defined as an animal 's motivation to feed, i t may be expressed as: (1) H t = 1 - ( Vt / V m a x ) where (H ) i s thelevel of hunger at time ( t ) , (V ) i s the t max stomach capacity and (Vt) i s the amount of food in the stomach at time ( t ) . Hunger w i l l , therefore , be greatest (numerical ly, l ) when the stomach i s empty, and minimal (numerical ly, 0) when i t i s f u l l . This expression of hunger i s convenient in that i t allows one to quantify an animals feeding motivation at any time i f the amount of food in i t s gut i s known. It i s a l so , however, extremely s i m p l i s t i c with respect to the complex mechanisms which are knowito affect feeding motivation (Ruch and Patten, 1965). For example, i t does not account for time lags in phys io log ica l feedbacks such as blood sugar l e v e l s but i m p l i c i t l y assumes that the state of hunger w i l l change immediately af ter any food is ingested or a l t e rna te ly , cleared from the stomach. With these r e s t r i c t i o n s in mind, eguation ( l ) can be used to investigate the re la t ionsh ip between hunger and the feeding behaviour of rainbow t rout . The s a t i a t ion r a t i o n , or the maximum stomach capacity of the f i sh was determined by holding them in i s o l a t i o n u n t i l a l l the food consumed during t h e i r previous meal had been passed from the digest ive t r a c t . They were then allowed to feed ad 1ibitum u n t i l they reached s a t i a t ion (stopped attacking prey) . Since the number of amphipods captured during th i s period and the average weight of a s i n g l e animal were known, the amount of food each f i s h consumed could be est i m a t e d . The r e s u l t s of these experiments i n d i c a t e d that the t r o u t became s a t i a t e d a f t e r consuming 230 mg. dry weight, ( 1 S.E. _+ 12 mg. ), r e g a r d l e s s of whether they were f e d a d u l t Cranqonyx or H y a l e l l a . In the p r e d a t i o n experiments, captured prey were not r e p l a c e d . Therefore, both the d e n s i t y of prey as well as the predator's l e v e l of hunger d e c l i n e d as the experiment progressed. E i t h e r of these aspects could a f f e c t the rate at which t r o u t a t t a c k , t h e r e f o r e , the c o n t r o l experiments were pooled according to the type of prey and were analysed by m u l t i p l e l i n e a r r e g r e s s i o n . For t h i s a n a l y s i s , an experiment was d i v i d e d i n t o 30 second i n t e r v a l s so the d e n s i t y of prey and s t a t e of hunger of the t r o u t at one i n t e r v a l ( t ) could be c o r r e l a t e d with the r a t e of attack i n the succeeding i n t e r v a l ( t + l ) . The l e v e l of hunger at the beginning of each time p e r i o d could be estimated (equation l ) as both the number of prey that had been captured up to that p o i n t and the average weight of each of the t e s t prey were known (Table 3 ) . A regression analys i s indicated that prey density and hunger were p o s i t i v e l y corre lated with the attack rate (Table 4). In other words, as the density of prey decreased the rate of attack diminished but at a rate that was dependent upon the t r o u t ' s l e v e l of hunger. An example of the progressive decl ine in the attack rate from one ser ies of experiments i s i l l u s t r a t e d in Figure 1. Observation suggested that the negative feedback between hunger and the rate of attack might have been due to an increase in the amount of time trout use to handle food as the i r hunger diminishes . Hunger has a s imi l a r effect on handling time in mantids (Hol l ing , 1966). On the basis of the control experiments, the re la t ionship between prey density (PD), hunger ( ) and the rate of attack (RA) was found to be adequately described by the regression eguation: (2) RA = b 1 (PD) + b 2 (H t ) + K The effects of prey density and hunger are indica ted , re spect ive ly , by the constants (b^) and (b^) > (K) i s the Y-intercept of the regress ion. In the control experiments the value of (b^) was s i g n i f i c a n t l y larger when Crangonyx were prey (Table 4). This means that when both species were at comparable dens i t i e s , trout attacked Crangonyx more rapidly than H y a l e l l a . One poss ible explanation for th i s result i s explored below. ure 1. The p r o g r e s s i v e d e c l i n e i n the rate of attack (30 second i n t e r v a l s ) . The r e s u l t s presented are from 2 c o n t r o l experiments i n which the i n i t i a l d e n s i t y of prey was 200 H y a l e l l a . The c l o s e d c i r c l e s i n d i c a t e the r a t e of a t t a c k of one f i s h ; the open c i r c l e s , the a t t a c k r a t e of another f i s h . 9. .CO O -f—• c o o CD > X o CO o ro o o © o o •• oi o .CM .O o o o o o o o •CO •CD i — T - m m — r r CD CM • « • o o o • C M TABLE 4. Regression constants for the rate of attack at time (t) with respect to the hunger l e v e l of the trout and the density of prey 3D seconds e a r l i e r . The constant (b^) indicates the effect of prey density on the rate of attack; (b^) the effect of hunger and (K) the y - in te rcep t . The standard deviat ion of each of the constants and the p a r t i a l 2 c o r r e l a t i o n coe f f i c i en t s (r) are ind ica ted . R i s the to ta l amount of v a r i a b i l i t y accounted for by each regress ion. (n) indicates the number of time in terva l s obtained by pooling a l l the experiments (Exp.) from each treatment. I HULL M Prey Exp n Density (b^) Hunger (b2) K R 2 Control H y a l e l l a 9 93 0.0024 + 0.0002 0.474 + 0.064 -0.207 0.66 r 0.71 0.62 Cranqonyx 4 50 0.0045 + 0.0005 0.360 + 0.044 -0.13 0.83 r 0.78 L i t t e r 0.77 I H y a l e l l a 6 78 0.0023 + 0.0003 0.345 + 0.072 -0.178 0.57 r 0 .62 0.49 Cranqonyx 6 52 0.0075 + 0.0005 0.145 _+ 0.035 -0.05 0.88 r 0.89 L i t t e r 0.51 II H y a l e l l a 6 65 0.0028 + 0.0002 0.241 + 0.067 -0.19 0.62 r 0.81 0.42 Juv. Cranqonyx 6 66 0.0026 + 0.0003 0.326 + 0.162 -0.13 0.50 r 0.68 0.65 Cranqonyx 6 51 0.0053 + 0.0009 0.223 + 0.061 -0.13 0.62 r 0.65 0.46 The Relationship Between the Rate of Attack and Prey Size • The rate a predator w i l l encounter food i s determined large ly by: l ) the distance from which i t w i l l react, 2) the density of prey, and 3) the r e l a t i v e v e l o c i t y between the predator and prey (Hol l ing , 1966). The fact that Cranqonyx was attacked at a faster rate than H y a l e l l a , could be accounted for i f e i ther the searching ve loc i ty of the f i s h , or the distance from which they attacked was dependent upon the type of prey to which they were exposed. Both species of amphipods move r e l a t i v e l y slowly with respect to t rout ; therefore, in th i s case, the ve loc i ty of the predator w i l l ; .contribute most to determining the rate of encounter. The range in ve loc i ty at which trout searched for Cranqonyx, was determined in each of the control experiments ( f i g . 2) . This component was expressed simply as the amount of time the predator took to cover a known distance when i t was in i t s c h a r a c t e r i s t i c searching p o s i t i o n . Although the average ve loc i ty at which the f i s h searched for Hya le l l a was not measured, there was no apparent ind ica t ion that i t changed. The distance from which a predator w i l l react to prey w i l l also determine the rate of attack. Brawn (1969) found that cod can detect large prey from a considerably greater distance than smaller prey. An independent set of experiments was conducted to invest igate the p o s s i b i l i t y that the react ive distance of trout was related to prey s i z e . In order to measure th i s distance, I introduced a s ingle prey of known s ize into one of the experimental tanks before releas ing a f i s h . The react ive distance was defined as the distance between the predator and prey when the trout i n i t i a t e d an at tack. The resul t s of these experiments (Table 5 ) c l e a r l y indica te that the distance of react ion i s dependent upon prey s i z e . This re la t ionsh ip might be s u f f i c i e n t to explain why adult Crangonyx, the larger of the two species , was attacked at almost twice the rate as H y a l e l l a . In a l l other respects, except the i r s ize and a c t i v i t y , these animals are very s i m i l a r . In the control experiments every amphipod was exposed and captured before an experiment was terminated. Under these condi t ions , the density of prey, the i r s i ze , as well as the state of hunger of the trout affected the rate of attack. In the presence of a l i t t e r substrate, however, the feeding behaviour of the predator, the dispersa l behaviour of the' prey, or both might be somewhat a l t e r e d . These p o s s i b i l i t i e s are considered in the next s ec t ion . ure 2. The s e a r c h i n g v e l o c i t y of rainbouu t r o u t i n s e v e r a l c o n t r o l experiments. See tex t f o r f u r t h e r explanat i o n . SEARCHING VELOCITY (cm/sec) TABLE 5. The r e l a t i o n s h i p between r e a c t i v e d i s t a n c e and prey s i z e . The prey were l i v e Cranqonyx. (n) i n d i c a t e s the number of o b s e r v a t i o n s obtained from 2 f i s h . The average r e a c t i v e d i s t a n c e i s expressed to the nearest s i g n i f i c a n t f i g u r e . The 95$ confidence i n t e r v a l s are p resented. Prey Length n Mean r e a c t i v e (mm) d i s t a n c e (cm) 4 9 1 8 + 3 5 17 2 2 + 2 6 9 28 + 5 7 21 2 8 + 3 9 28 3 5 + 3 THE EFFECT OF A SUBSTRATE ON THE ATTACK RATE (LITTER EXPERIMENTS) Before a f i s h was.released the prey were introduced into the experimental tank and were allowed one hour in which to d i sperse . When l i t t e r was present some amphipods would conceal themselves immediately by moving under any object they encountered. Other animals would move about for some time before they took cover or else f a i l e d to move under cover at a l l . The result of t h i s behaviour is that at any instant in time the number of prey that were exposed was determined by the rate at which animals were both leaving and entering concealment. An experiment was conducted with adult Cranqonyx to determine i f they would e s tab l i sh an equi l ibr ium l e v e l of exposure, and i f so, how long i t would take. Four populations of amphipods were introduced into separate containers of water (25 cm in diameter). Each vessel had 18% of the bottom area covered with s t ick l i t t e r i d e n t i c a l to that composing the l i t t e r I and II substrates . The populations were then observed for over an hour. Without exception, the proportion of animals that were exposed decl ined rapidly in the f i r s t 45 minutes after they were introduced u n t i l an apparent equi l ibr ium was reached ( f i g . 3) . Although the length of time concealed animals remained under cover proved to be r e l a t i v e l y long, some did re-expose themselves. The same animals were not cont inua l ly exposed or concealed. On the basis of these observations, i t was concluded that the hour in which the prey were allowed to disperse before an experiment was l i k e l y ' s u f f i c i e n t for the test population to reach an equi l ibr ium leve l of exposure before the ure 3. The average time required for four experimental populations of Cranqonyx to reach an • e q u i l i b r i u m ' l e v e l of exposure. The data points are means of 4 r e p l i c a t e s . The curve was f i t t e d by inspec t ion . TIME (MIN) AFTER INTRODUCTION predator was introduced. The feeding behaviour of trout was e s s e n t i a l l y the same when a substrate was present as i t was in the control s i t u a t i o n . In both cases, the f i s h maintained a searching pos i t ion and responded v i s u a l l y to prey. During the 50 minute duration of the l i t t e r experiments the trout did not disturb the sediment or move pieces of l i t t e r to f ind food, they only captured animals that were exposed. Figures 4 and 5 i l l u s t r a t e the re l a t ionsh ip between the number of prey that were captured and the type of substrate . It i s evident that at each of the experimental dens i t ies a greater number of Cranqonyx and Hyalel1 a were concealed and , therefore.were invulnerable to predation when more cover area was a v a i l a b l e . These resul t s ( f i g s . 4 and 5) were f i t t e d by l i n e a r regression and without exception, were described adequately by a s tra ight l i n e which passed through the o r i g i n (Table 6A). This indicates that the proportion of animals that were exposed in each treatment (PE), which i s given by the slope of the l i n e , was constant over the range of dens i t ies used. In the l i t t e r II ser ies the number of adult and juveni le Cranqonyx that were captured when both populations were of comparable s ize was not d i f ferent so these two sets of data were pooled. The length of time an amphipod requires to locate cover should be d i r e c t l y proport ional to the amount of s t i ck l i t t e r . If the amount of cover area i s increased but the average length of time an animal remains concealed or exposed does not change, then the proportion of prey in the population that are exposed should d i m i n i s h a c c o r d i n g l y . T h i s might e x p l a i n the i n v e r s e r e l a t i o n s h i p between the amount of cover area and the p r o p o r t i o n of prey that were captured when a s u b s t r a t e was present. T h i s may not be a complete explanation,however, as there was a s u b s t r a t e - s p e c i e s i n t e r a c t i o n (Table 6 B). Fewer Cranqonyx were captured i n the l i t t e r II treatment, while H y a l e l l a was l e s s v u l n e r a b l e i n the f i n e l i t t e r experiments. Ne v e r t h e l e s s , the number of prey that were v u l n e r a b l e to a t t a c k and t h e r e f o r e the a t t a c k r a t e were r e l a t e d to the type of s u b s t r a t e . If the r a t e of a t t a c k during the l i t t e r experiments i s analysed i n the same f a s h i o n as the c o n t r o l r e s u l t s , only i n t h i s . : case, the p o p u l a t i o n d e n s i t y i s expressed i n terms of the d e n s i t y of v u l n e r a b l e prey at time ( t ) , i t i s p o s s i b l e to determine i f the presence of a s u b s t r a t e d i r e c t l y a f f e c t e d the f e e d i n g behaviour of the f i s h other than i n d i r e c t l y through prey d e n s i t y . If i t d i d not, then the values of b^, b^, and K that are obtained by a m u l t i p l e r e g r e s s i o n a n a l y s i s of the l i t t e r experiments should be s i m i l a r to the values obtained i n the c o n t r o l s i t u a t i o n . The f i n e l i t t e r treatment was not analysed because i t was d i f f i c u l t to a c c u r a t e l y d i s c r i m i n a t e between a t t a c k s t r o u t d i r e c t e d at prey and other ' s i m i l a r ' o b j e c t s . The r e s u l t s of the l i t t e r experiments are summarized i n Table 4. In both cases, the r e g r e s s i o n constants are not s i g n i f i c a n t l y d i f f e r e n t to those obtained i n the c o n t r o l , with the exception of the l i t t e r I treatment with a d u l t Cranqonyx. Other than t h i s , there i s no i n d i c a t i o n that a l i t t e r s u b s t r a t e ure 4. The r e l a t i o n s h i p between the number of Cranqonyx captured i n d i f f e r e n t h a b i t a t s , a f t e r 50 minutes exposure 2 to t r o u t p r e d a t i o n , and t h e i r i n i t i a l d e n s i t y per 0.42 m . (A) c l o s e d c i r c l e s represent the c o n t r o l s i t u a t i o n , and the open c i r c l e s l i t t e r I. (B) the l i t t e r II treatment, the c l o s e d c i r c l e s i n d i c a t e experiments conducted with a d u l t Cranqonyx ( B . l mm i n length) and the open c i r c l e s , j u v e n i l e Cranqonyx (4.6 mm). (C) the f i n e l i t t e r s e r i e s . INITAL DENSITY Figure 5 . The re l a t ionsh ip between the number of Hya le l l a captured in d i f ferent habitats after 50 minutes exposure to t rout , and t h e i r i n i t i a l density per 0.42 M . (A) the control (B) the l i t t e r I treatment (C) the l i t t e r II treatment and (D) the f ine l i t t e r treatment. 50 100 150 200 INITIAL DENSITY TABLE 6. A comparison of the effect of the d i f ferent substrate treatments on the proportion of Cranqonyx and Hya le l l a that were exposed and subsequently captured during an experiment. (n) indicates the number of experiments; (r) the cor re l a t i on c o e f f i c i e n t ; (PE) i s the slope, or proportion of prey that were captured. In part A, treatment slopes that are not bracketed by the same v e r t i c a l l i n e are s i g n i f i c a n t l y d i f ferent at the 0.05 l e v e l or l e s s . (V ^ ) i s the variance of the s lope. In part B, (t) i s s tudent 's t . A. V u l n e r a b i l i t y within species Hya le l l a Substrate Treatment Vb PE Control L i t t e r I L i t t e r II Fine l i t t e r 9 6 6 6 0.99 0.99 0.98 0.90 .0010 .0003 .0050 .0014 0.98 0 .791 0.661 0.16 Control L i t t e r I L i t t e r II Fine L i t t e r 4 6 12 6 Cranqonyx 1.00 0.97 0.91 0.97 .0000 .0133 .0040 .0016 1.001 0 .90J 0 .431 0.31J B. V u l n e r a b i l i t y between species Control 13 a L i t t e r I 12 1.069a L i t t e r II 18 24.350b Fine l i t t e r 12 2.710b a Not s i g n i f i c a n t b S ign i f i cant at or less than 0.05 l eve l a l tered the behaviour of the t rout . The concealment behaviour of the prey was the primary factor that determined the number of animals that were captured. Therefore, the re la t ionship between the substrate, the vulnerable density of amphipods and the rate of attack can be described by modifying equation ( 2 ) as fo l lows : (3) RA = b 1 (PD) (PE) + b 2 (H ) + K where (PE) i s the proportion of the population exposed (Table 6 A ) . The effect of prey s ize on the attack rate was also evident in the l i t t e r experiments. Regardless of the type of substrate, adult Cranqonyx were attacked cons iderab ly , fa s ter than e i ther juveni le Cranqonyx or H y a l e l l a . I mentioned e a r l i e r that the explanation for th i s observation could be due to the re la t ionship between prey s ize and react ive dis tance . The addi t iona l piece of information that i s consistent with th i s supposit ion i s that Hyafrella and juveni le Cranqonyx are both about the same s ize and were attacked at i d e n t i c a l rates (Table 4 ) . PREY CAPTURE SUCCESS The success predators have in capturing food depends upon three basic components, namely, the i r a b i l i t y to recognize, approach and s t r ike at prey (Ho l l ing , 1966). Trout are not completely successful in d i scr iminat ing between prey and other ' s i m i l a r 1 , targets ; they w i l l attack inanimate objects . This not only indicates that v i sua l cues other than movement w i l l induce trout to attack but also suggests that they could 'waste' a substant ia l amount of searching time i f they attacked many inanimate objects . In the l i t t e r I and II experiments, the f i s h attacked r e l a t i v e l y few pieces of l i t t e r . This was not the case, however, in the f ine l i t t e r treatment as only about 71% of t h e i r attacks were directed toward prey; the rest were directed at pieces of l i t t e r about the same s ize and color as amphipods. Although th i s demonstrates that the d i v e r s i t y of a substrate can inf luence the capture success of t rout , by imparing the i r a b i l i t y to success ful ly d i scr iminate food, th i s aspect i s far beyond the scope of th i s paper and w i l l be treated in more d e t a i l in another sect ion (IV) . Once trout recognize a prey (orient toward i t ) they are always completely successful in approaching to within s t r i k i n g d i s tance . Every s t r i k e they attempt, however, i s not per fect ly executed as some f a i l to capture prey. Table 7 indicates the s t r i k e e f f i c iency of trout in the experiments in which th i s could be determined. The resul t s of an addi t iona l set of control experiments, in which the prey were 11 mm adult Cranqonyx, are also presented. These data (Table 7) show that regardless of the type of substrate trout were extremely successful in capturing amphipods. Since s t r ike success (CS) was the major factor that determined i f a prey would be captured, the rate of capture (RC) in the d i f fe rent treatments can be described simply by treat ing th i s component as a constant and adding i t to equation (3) as fo l lows: (4) RC = [b 2 (PD) (PE) + b 2 (H t ) + K ]CS At th i s point , the effects of prey s i ze , prey density, substrate complexity and capture success have bean incorporated into a s ingle regression eguation which describes the rate at which trout , in d i f ferent states of hunger, can capture food. This descr ipt ion however, i s r e s t r i c t e d by the assumption of l i n e a r i t y , which implies that there i s no l i m i t to the rate of capture. This assumption, of course, i s not true and i s refuted by the data presented in Figure 6. In th i s example, the rate of capture in the f i r s t 30 seconds of the c o n t r o l , l i t t e r I and l i t t e r II experiments i s plotted as a function of the density of exposed prey. There i s l i t t l e question that the capture rate approached a maximum value (average = 0.67) as the density of prey approached about 240/ sq. m. (100 amphipods in the tank) . Therefore, the i m p l i c i t r e s t r i c t i o n of equation ( 4 ) i s that the rate of capture cannot surpass th i s l i m i t regardless of the density of prey, t h e i r s i ze , or the predator ' s state of hunger. TABLE 7. The proportion of the to ta l number of s t r ikes which success ful ly terminated with the capture of prey. indicates the number of experiments. Species Average Length N Mean s t r i k e 95% confidence (mm) success in te rva l of mean Hya le l l a Crangonyx Cranganyx 5.7 8.1 10.8 Control 7 5 9 0.907 0.909 0.902 0.812 0.778 0 .837 1.000 1 .000 0 .967 Hya le l l a Crangonyx 5.7 8.1 L i t t e r I 6 6 0.882 0 .886 0.753 - 1.000 0.716 - 1.000 Hya le l l a Crangonyx Crangonyx 5.7 4.6 8.1 L i t t e r II 6 6 6 0 .820 0.833 0.864 0 .656 0.669 0.716 0 .984 0.997 1.000 Figure 6. The re l a t ionsh ip between the density of exposed prey and the attack ra te . The rate of attack was determined over a 30 second time per iod . In each case, the predator ' s l e v e l of hunger was maximal with respect to equation ( l ) . The data points were obtained from the c o n t r o l , l i t t e r I and l i t t e r 2 experiments with both Cranqonyx and H y a l e l l a . THE THRESHOLD RATE OF PREY CAPTURE AND THE SEARCHING PATTERN Searching behaviour will decrease rapidly once an animal's hunger motivation is satiated (Beukema, 1968; Holling, 1966). Other factors however, might alter both the duration and nature of searching before this occurs. When a trout was released at the start of a feeding experiment i t would immediately move to the bottom, adopt it s characteristic searching position, and begin hunting for amphipods. Figure 7 diagramatically illustrates the pattern of benthic searching behaviour during a typical experiment. As indicated, irrespective of the density of exposed prey the predator would devote a l l of its attention for some time to hunting for amphipods. Eventually however, this attention was disrupted and began to wane. During this phase of the experiment the fish would shift i t s searching position and move higher up into the water column to hunt for food, or else hold a stationary position for a few minutes. In either case i t devoted less time to hunting the substrate for prey. Although in every situation the i n i t i a l searching pattern was disrupted before the experiment was terminated, the duration of this phase was related to the number of vulnerable prey. The trout shifted their attention sooner when the density of prey was low. This poses the question then as to the mechanism which might be responsible for causing a predator to disrupt one searching pattern and switch its attention to another pattern (i.e. hunting for prey in the water column) or behaviour. It seems tenable that an animal would search as long as t h i s behaviour was reinforced but that i t would sh i f t i t s a t tent ion to other forms of hunting or behaviours i f the rate of food intake f e l l toward zero. Perhaps there i s a threshold rate of capture that trout must exceed i f they are to continue hunting for benthic prey. If they cannot a t ta in th i s threshold then they w i l l a l t e r t h e i r behaviour. This hypothesis can be tested by determining i f the rate of capture, when the trout f i r s t disrupted the i r search for amphipods, was r e l a t i v e l y constant (threshold) i r re spec t ive of the predator ' s state of hunger or the type of prey i t was feeding on. These data from the experiments with adult Crangonyx and H y a l e l l a , are presented in Figure 8. When Hyale l l a were prey, the capture threshold was reasonably constant (mean = 0.058 _+ 1 5 .E . .008) i r r e spec t ive of the t r o u t ' s degree of hunger. This was not true however, when the f i sh were exposed to adult Crangonyx; in th i s case, there was a s i g n i f i c a n t pos i t ive re la t ionship ( P = 0.04 ) between the threshold and hunger. Nevertheless, in e i ther instance once the rate of food intake f e l l below an average of 0.051 captures per second the trout moved away from the bottom of the tank. This indicates that there i s a c r i t i c a l rate of capture although the threshold may be modified by hunger. Once the i n i t i a l searching pattern was disrupted, the f i s h cont inua l ly changed t h e i r v e r t i c a l p o s i t i o n . They would move up into the water column for some time and then revert back to ure 7. A schematic representation of the pattern of benthic searching behaviour by trout during an experiment. (CSP) indicates the phase in which the f i s h were completely a t tent ive to hunting for amphipods, (ISP) represents the phase in which the benthic searching pattern waned, (PD) i l l u s t r a t e s the progressive decl ine in the density of exposed prey, and (TRC) i s the density of prey which produces the threshold rate of capture. 0 available time spent searching for amphipods Figure 8. The re la t ionsh ip between the threshold rate of prey capture and the predator ' s state of hunger. The regression l i n e indicated in the experiments with adult Cranqonyx is s i g n i f i c a n t at the 0.04 l e v e l . See text for further explanation. Threshold Rate of Attack • • . • O O - L fNj CD O 4 ^ I I I I i i i Figure 9 . The re la t ionsh ip between the experimental substrate and the rate of ext inct ion of the benthic searching pattern after the i n i t i a l phase of complete a t t en t ion . Only the resul t s obtained with Hyale l l a are presented. The v e r t i c a l bars ind ica te the 95% confidence in te rva l s of each of the means. (A) control experiments, (B) L i t t e r I experiments, (C) L i t t e r II experiments. (A) | 1 © 1 ^ 1 A -k © © © © i i r ~ i — i — r ~ i — r n — j 10 20 30 4 0 50 10 2 0 30 4 0 5 0 T— i — i — i — i — n — I i I 10 2 0 3 0 4 0 5 0 TIME INTERVAL (MIN) AFTER COMPLETE A T T E N T I O N . hunting over the substrate . Since they were no longer being reinforced for searching for amphipods, the amount of time they devoted to th i s behaviour waned throughout the remainder of the experiment ( f i g . 7 ) . The presence of a l i t t e r substrate did influence the feeding behaviour of the trout in one sense, however, because the rate of ext inct ion of the benthic searching pattern was related to the complexity of the substrate ( f i g . 9). The trout would return to the bottom to search for food more frequently when the substrate was d i v e r s i f i e d . DISCUSSION Two fundamental processes, prey detect ion and capture w i l l determine the food organisms which comprise an animals' food supply. The t a c t i c s a predator w i l l u t i l i z e to locate prey w i l l r e s t r i c t not only the types of animals i t can attack but also w i l l determine where and when i t can search e f f e c t i v e l y . If i t can use more than one type of sensory receptor to locate food i t may be somewhat less r e s t r i c t ed in i t s feeding a c t i v i t y than predators which re ly predominantly upon v i s i o n . A l i (1959) demonstrated that salmonids, which feed v i s u a l l y , could not capture prey u n t i l the ambient i l l u m i n a t i o n exceeded the rod threshold . Therefore, th i s c h a r a c t e r i s t i c of the environment w i l l impose r e s t r i c t i o n s both in space and time on the feeding a c t i v i t y of v i sua l predators . Despite th i s apparent drawback, there are d e f i n i t e advantages to using v i s ion to locate food. For example, i t i s a long range and precise mechanism that can accurately f ix the pos i t ion of a target . This i s true even for f a i r l y unsophisticated v i sua l systems such as those possessed by mantids and dragonfly l a rvae . Although these animals can e f f e c t i v e l y detect only moving targets (Pr i tchard , 1965; R i l l i n g et a_L, 1959), th i s i s not true for predators with more developed v i sua l receptors . Most vertebrates can di scr iminate 4 v i sua l propert ies of an object ; s i ze , form, contrast , and motion (Horridge, 1968; Prazordkova, 1969). This implies that they should be able to detect a broader spectrum of food organisms simply because they could effect a decis ion to attack on the basis of some other qua l i ty than prey movement. Many theor i s t s have assumed that predators w i l l encounter food in d i rec t proportion to the abundance of each prey organism. If th i s i s true, then the predator must react from a fixed di s tance . One of the basic c h a r a c t e r i s t i c s of any v i sual system i s that the distance required to d i scr iminate an object i s almost proport ional to the s ize of the target . Therefore, v i sua l predators such as rainbow trout have the opportunity to react to large prey from a greater distance than smaller prey. It was demonstrated that the distance of react ion of trout i s dependent upon prey s ize and that th i s could explain why they attacked adult Cranqonyx faster than e i ther Hyale l la or juveni le Cranqonyx. The searching pos i t ion that rainbow trout adopt when they hunt for benthic prey was only considered s u p e r f i c i a l l y in th i s study. This behaviour,however, has an in tere s t ing i m p l i c a t i o n , A predator can only detect a prey i f the height of i t s searching pos i t ion i s less than the distance i t requires to release an at tack. If th i s condit ion i s not met . then some small s ize classes of prey may be invulnerable to detect ion and subsequently, capture. There i s some evidence that th i s inference may apply to rainbow trout (section I V ) . Both of these c h a r a c t e r i s t i c s , a react ive distance that i s dependent upon prey s ize and a spec i f i c searching pos i t i on , are basic components of the feeding behaviour of t rou t . These same c h a r a c t e r i s t i c s are su f f i c i en t to explain the well documented f i e l d o b s e r v a t i o n that many s p e c i e s of f i s h d i s p r o p o r t i o n a t e l y e x p l o i t l a r g e prey and f r e q u e n t l y f a i l to consume others below a t h r e s h o l d s i z e ( i v l e v , 1961; Brooks, 1968). Ivlev (1961) a l s o pointed out that very e f f e c t i v e bottom feeding f i s h do not r e l y p r i m a r i l y upon v i s i o n to detec t food and do not appear to adopt pronounced s e a r c h i n g p o s i t i o n s . Since other s p e c i e s of salmonids appear to behave s i m i l a r l y to rainbow t r o u t (Schutz, 1969; Sheperd, 1970) t h i s may e x p l a i n uuhy they tend to be r e l a t i v e l y i n e f f e c t i v e p r e d a t o r s of many inconspicuous b e n t h i c -l i v i n g organisms (Smithi, 1961). Capture success a l s o a f f e c t s the types of prey a predator can e x p l o i t . In the present study, there was no i n d i c a t i o n t h a t the s i z e of the prey considered s i g n i f i c a n t l y impared the capture success of t r o u t . In theory, however, there must e x i s t both an upper and lower l i m i t to the s i z e of organism that can be s u c c e s s f u l l y manipulated by predators which swallow t h e i r food. Within t h i s range, there i s l i k e l y to be an optimum s i z e d prey that can be captured most s u c c e s s f u l l y . Both H o l l i n g (1964) and Dixon (1959) have demonstrated that capture success tends to d i m i n i s h i f the prey i s e i t h e r l a r g e r or small than the optimum s i z e . The process of p r e d a t i o n i s not only dependent upon the components of prey d e t e c t i o n and capture success, but also the d e n s i t y of prey, the pre d a t o r ' s hunger m o t i v a t i o n and the degree to which l e a r n i n g can a l t e r the behaviour of e i t h e r the predator or prey. H o l l i n g (1966) has d i s c u s s e d the i n f l u e n c e of prey d e n s i t y and hunger m o t i v a t i o n on the f u n c t i o n a l response of p r e d a t o r s . The a c t i o n of these components proved to be i d e n t i c a l f o r t r o u t , t h e r e f o r e , t h e i r s i g n i f i c a n c e w i l l not be r e i t e r a t e d here beyond s t r e s s i n g that i n c r e a s i n g prey d e n s i t y s t i m u l a t e s the attack rate while d i m i n i s h i n g hunger m o t i v a t i o n J antagonizes t h i s e f f e c t . The p o s s i b i l i t y that the behaviour of t r o u t could be a l t e r e d by l e a r n i n g w i l l be t r e a t e d l a t e r ( S e c t i o n I I ) . In any case, a l l of these c h a r a c t e r i s t i c s w i l l operate to determine the r a t e of capture, which i n turn could i n f l u e n c e the searc h i n g p a t t e r n a predator adopts. Although the c e s s a t i o n of hunger w i l l n a t u r a l l y terminate s e a r c h i n g , t h i s behaviour i s undoubtly s e n s i t i v e to other s i g n a l s as w e l l . Most responses w i l l wane i f they are not r e i n f o r c e d i n t e r m i t t e n t l y ; t h i s i s evident, f o r example, i n the attack response of mantids. If they are not rewarded f o r s t r i k i n g at a 'dummy' t a r g e t they w i l l simply stop responding " ( R i l l i n g et a l , 1959; Holling, 1966). The searc h i n g behaviour of t r o u t i s completely analagous because i f they are not s u f f i c i e n t l y r e i n f o r c e d ( t h r e s h o l d r a t e of capture) they w i l l s h i f t t h e i r p a t t e r n of search. T h i s feedback has been suggested to be r e s p o n s i b l e f o r the changes i n the feed i n g p o s i t i o n of t r o u t i n mountain streams ( J e n k i n s , 1969). This same feedback a l s o suggests that t r o u t could t e m p o r a r i l y converge i n t o areas i n which prey are r e l a t i v e l y more v u l n e r a b l e to a t t a c k . By simply randomly s h i f t i n g t h e i r p o s i t i o n , some i n d i v i d u a l s , w i l l l o c a t e areas of prey abundance. If the d e n s i t y of prey i s high enough to surpass the t h r e s h o l d r a t e of capture, then the predator may remain feeding i n the a r e a . Given enough time, most of the p o p u l a t i o n could converge i n t o a s p e c i f i c r e g ion or at l e a s t adopt the same r e l a t i v e s e a r c h i n g p a t t e r n , s u c h as feeding i n the uuater column or over a s u b s t r a t e . The phenomenon of convergence has been reported f o r other animals (Neish, 1970; Tinbergen, I960) as w e l l as f i s h ( A l l e n , 1941). Holling'(l959a) pointed out that predators that can invoke an immediate numerical response, such as converging, w i l l f u n c t i o n as a s t a b i l i z i n g component of the community because they w i l l tend to counteract any s e r i o u s imbalance i n prey abundance. In c o n c l u s i o n , the o b s e r v a t i o n that t r o u t w i l l s h i f t t h e i r p a t t e r n of search i f they are not being s u f f i c i e n t l y r e i n f o r c e d i m p l i e s that the p o p u l a t i o n w i l l d i s p e r s e through the water column and w i l l tend to converge t e m p o r a r i l y i n t o areas i n which prey are r e l a t i v e l y more v u l n e r a b l e to a t t a c k . In a d d i t i o n , s i n c e t r o u t were found to react to only exposed prey and d i s p l a y e d an a t t a c k response that was dependent upon prey s i z e , t h i s . suggests that they should be e f f e c t i v e p r e d a t o r s of l a r g e , exposed animals but would be r e l a t i v e l y i n e f f e c t i v e i n c a p t u r i n g s m a l l e r or more c r y p t i c s p e c i e s . SUMMARY l ) In experiments without a l i t t e r s u b s t r a t e , i t was shown that both prey d e n s i t y and the hunger m o t i v a t i o n of t r o u t a f f e c t t h e i r r a t e of a t t a c k . These two components are a n t a g o n i s t i c since the former increases the attack rate while the l a t t e r depresses i t . 2) The react ive distance of rainbow trout i s dependent upon prey s i z e . This could explain why they attacked adult Cranqonyx, the largest prey, faster than e i ther Hya le l l a or juveni le Cranqonyx. 3) When a substrate was present, both species of prey concealed themselves. The proportion that were exposed was inverse ly related to the amount of cover area. Since trout w i l l detect only exposed prey, the i r attack, rate was inversely re lated to the d i v e r s i t y of the substrate . 4) There was no consistent ind ica t ion that the presence of a substrate d i r e c t l y a l tered the feeding behaviour of t rout . The concealment behaviour of the prey was the primary factor that determined the outcome of the l i t t e r experiments. 5) The a b i l i t y of rainbow trout to capture prey was shown to be independent of both the s ize of the test prey and the d i v e r s i t y of the substrate . The l a t t e r however, did impare t h e i r success in d i scr iminat ing amphipods. 6) In the laboratory, trout must be re inforced at a rate that exceeds 0.051 captures per second i f they are to maintain a s p e c i f i c searching pat tern . If they do not a t ta in th i s threshold they w i l l switch the i r a t tent ion to other hunting patterns or behaviours. Once th i s occurs, the o r i g i n a l pattern w i l l wane at a rate that i s inversely dependent upon the d i v e r s i t y of the substrate . 7) Due to 4 major c h a r a c t e r i s t i c s of t h e i r feeding behaviour: i ) the dependence of the react ive distance on prey s i ze , i i ) the searching pos i t ion , i i i ) the fact that they w i l l attack only exposed prey, and iv) the threshold rate of capture rainbow trout are l i k e l y to converge into areas in which prey are r e l a t i v e l y abundant,should be e f fect ive predators of large , exposed prey; butshould be considerably less e f fect ive in exp lo i t ing smaller or less conspicuous species . SECTION I I THE E F F E C T OF E X P E R I E N C E ON THE RESPONSE OF TROUT TO UNF AMI L I A R PREY INTRODUCTION The c o n c e p t o f t h e " s e a r c h i n g i m a g e " h a s a t t r a c t e d c o n s i d e r a b l e a t t e n t i o n i n t h e f i e l d s o f a n i m a l b e h a v i o u r and e c o l o g y s i n c e i t was f i r s t p r o p o s e d by T i n b e r g e n ( i 9 6 0 ) . He a n d o t h e r s s i n c e t h e n , h a v e s u g g e s t e d t h a t many a n i m a l s c a n l e a r n t o i n c r e a s e t h e i r r e s p o n s i v e n e s s t o p r e y , b u t t h a t , i f t h e y a r e n o t c o n t i n u a l l y r e i n f o r c e d t h e y w i l l s h i f t t h e i r a t t e n t i o n t o o t h e r o b j e c t s . B o t h o f t h e s e f e a t u r e s w o u l d be a d a p t i v e s i n c e they would e n a b l e p r e d a t o r s t o h u n t w i t h "maximum e f f i c i e n c y " ( C r o z e , 1 9 7 0 ) . The e x i s t e n c e o f t h i s b e h a v i o u r h a s b e e n d e m o n s t r a t e d by . e x p e r i m e n t a l s t u d i e s on v a r i o u s v e r t e b r a t e s ( B e u k e m a , 1 9 6 8 ; De R u i t e r , 1 9 5 2 ; H o l l i n g , 1 9 5 9 a ; C r o z e , 1 9 7 0 ) . To d a t e , t h e a i m o f mos t o f t h i s w o r k h a s b e e n t o r e v e a l t h e c o m p o n e n t s o f f e e d i n g b e h a v i o u r t h a t a p p e a r t o be a f f e c t e d by l e a r n i n g a n d t o e x a m i n e t h e r e s u l t i n g e c o l o g i c a l i m p l i c a t i o n s . W i t h t h e s e c o n s i d e r a t i o n s i n m i n d , H o l l i n g ( 1 9 6 5 ) d e v e l o p e d a g e n e r a l m o d e l t o a c c o u n t f o r t h e l e a r n i n g p r o c e s s . He s u g g e s t e d t h a t e x p e r i e n c e o p e r a t e d t h r o u g h a s y s t e m o f f e e d b a c k s b e t w e e n t h e p a l a t a b i l i t y o f p r e y a n d t h e p r e d a t o r ' s s t a t e o f h u n g e r t o a f f e c t t h e d i s t a n c e o f i t s r e a c t i o n t o p r e y . S i m u l a t i o n s t u d i e s d e m o n s t r a t e d t h a t t h e m o d e l was s u f f i c i e n t t o a c c o u n t f o r t h e p h e n o m e n o n o f t h e s e a r c h i n g i m a g e and had i m p o r t a n t i m p l i c a t i o n s with respect to the se lec t ive advantages of mimicy between palatable and less palatable prey. The experiments described in th i s section were designed to examine some of the factors which might inf luence the response of rainbow trout to a r t i f i c i a l , but palatable prey, and to determine i f a s soc ia t ive learning could be an important component of t h e i r behaviour. The resu l t s were then interpreted in terms of H o l l i n g ' s model to test i f i t was s u f f i c i e n t l y general to account for the effect of learning on the feeding behaviour of t rout . METHODS AND MATERIALS The rainbow trout selected for th i s study ranged in length from 11 to 14 cm ( l to 2 years old) and were obtained from Marion Lake, B r i t i s h Columbia. To insure that the f i sh were completely naive, the experimental prey were formed from commercial chicken l i v e r . The shape and s ize of the prey were standardized by cutt ing c y l i n d r i c a l pieces of l i v e r 3 mm by 5 mm in length . Conditioning experiments were conducted in two e n t i r e l y d i f ferent s i t u a t i o n s . In the f i r s t set of experiments (A), i n d i v i d u a l naive f i s h were placed into a small holding chamber (30 x 12 x 20 cm) that was suspended in a 50 ga l lon (227 l i t e r ) glass aquarium. Six standard prey were then scattered at random through the tank. After the food had been introduced the trout was released from the holding area. An experiment lasted for 20 minutes and was considered to represent one 'day* of experience regardless of whether the animal: fed . or not. Experiments were conducted every 48 hours until the amount of time the fish required to locate and capture a l l 6 prey stabilized. Between successive experiments the predators were held in isolation and without food. All other experiments (B) were conducted in a large rectangul tank (180 x 16 x 30 cm) which had a small holding area at one end. This chamber was separated from the remainder of the tank by an opaque, sliding partition. One side of the tank was marked off in 1 cm intervals so that the distance from which trout would react to prey could be estimated. For each feeding experience a fish was transferred to the holding chamber in the experimental tank. Before the predator was released a single prey was placed near the opposite end of the tank. The reactive distance was defined as the distance between the predator and prey when the fish attacked. After a prey had been captured, the trout was returned to the holding chamber while another piece of food was introduced. A single day of experience consisted of six successive captures. Once these were complete, the predator was returned to its holding tank and deprived of food until the next test period 48 hours later. In some (B) experiments the trout were exposed to prey that contrasted differently with the background. The level of t a r g e t c o n t r a s t was c h a n g e d by s t a i n i n g t h e s t a n d a r d w h i t e p r e y i n a s a t u r a t e d s o l u t i o n o f S u d a n B l a c k B . S e v e r a l d i f f e r e n t d e g r e e s o f c o n t r a s t ( l i g h t g r e y t o b l a c k ) c o u l d be r e p r o d u c e d by v a r y i n g t h e l e n g t h o f t h e s t a i n i n g t i m e . A l t h o u g h t h e t e r m , c o n t r a s t , i s c o m m o n l y d e f i n e d a s t h e d i f f e r e n c e i n t h e a m o u n t o f l i g h t r e f l e c t e d by a t a r g e t w i t h r e s p e c t t o t h e b a c k r o u n d ( l e G r a n d , 1 9 6 7 ) , i t i s u s e d i n a r e l a t i v e s e n s e t h r o u g h o u t t h i s p a p e r . In o t h e r w o r d s , i n b o t h s e t s o f e x p e r i m e n t s , t h e s t a n d a r d w h i t e p r e y h a d a h i g h c o n t r a s t r e l a t i v e t o e i t h e r t h e d a r k g r e y o r b l a c k p r e y b e c a u s e t h e b a c k g r o u n d ( t a n k b o t t o m ) was b l a c k . In b o t h s e t s o f e x p e r i m e n t s , t h e w a t e r t e m p e r a t u r e (10 _+ 2 C . ) t h e b a c k g r o u n d ( t a n k b o t t o m ) i l l u m i n a t i o n ( 0 . 3 , f t - c a n d l e s ) a n d t h e t u r b i d i t y o f t h e w a t e r ( a t t e n u a t i o n c o e f f i c i e n t , 0 . 5 0 ) w e r e c a r e f u l l y c o n t r o l l e d t o i n s u r e t h a t t h e v i s u a l a c u i t y o f t h e f i s h was n o t a f f e c t e d by c h a n g e s i n a n y o f t h e s e c o n d i t i o n s . RESULTS EXPERIMENT A THE C H A R A C T E R I S T I C S OF THE I N I T I A L RESPONSE OF TROUT TO UNFAMIL IAR PRE Y When t r o u t a r e e x p o s e d t o u n f a m i l i a r p r e y s e v e r a l a s p e c t s o f t h e i r f e e d i n g b e h a v i o u r c h a n g e w i t h e x p e r i e n c e . A l t h o u g h some f i s h w i l l i n v e s t i g a t e an u n f a m i l i a r s t i m u l u s t h e f i r s t t i m e t h e y e x p e r i e n c e i t , o t h e r s r e q u i r e r e p e a t e d e x p o s u r e b e f o r e t h e y w i l l r e a c t . The n u m b e r o f s u c c e s s i v e e x p o s u r e s a n i n d i v i d u a l r e q u i r e d b e f o r e i t w o u l d a t t a c k was d e f i n e d as t h e l a t e n t p h a s e . The a v e r a g e d u r a t i o n o f t h i s p h a s e f o r a g r o u p o f 9 t e s t f i s h was 4 days and ranged from 1 to as high as 11. Once the l a t e n t p e r i o d uuas terminated, the behaviour of the f i s h continued to change as they acquired f u r t h e r e xperience. Four separate steps preceed t h e i r capture of prey: l ) o r i e n t a t i o n , 2) f i x a t i o n , 3) a t t a c k , and 4) s t r i k e . Once the f i s h began to react to the t e s t prey many i n d i v i d u a l s would f a i l to complete an a t t a c k sequence. Some animals would v i s u a l l y f i x a t e prey and then f a i l to f o l l o w through with an a t t a c k , or at t a c k , but f a i l to capture prey. The d u r a t i o n of t h i s phase was found to average two f u r t h e r days of experience a f t e r the t e r m i n a t i o n of the l a t e n t p e r i o d . Although t r o u t w i l l a p p a r e n t l y develop a complete a t t a c k sequence r a p i d l y i f a prey i s p a l a t a b l e t h i s may not be true i f i t i s r e l a t i v e l y u n p a l a t a b l e (Sheperd, 1970). A f t e r the t r o u t e s t a b l i s h e d a complete a t t a c k p a t t e r n , the amount of time they took to capture a l l 6 prey diminished as they became more f a m i l i a r with them ( f i g . l ) . E v i d e n t l y , some other component of t h e i r behaviour was s t i l l changing a f t e r 6 days of exposure. H o l l i n g (1966) demonstrated that the attack rate i s determined p r i m a r i l y by three f a c t o r s : l ) the d e n s i t y of prey, 2) the r e l a t i v e v e l o c i t y between the predator and prey, and 3) the predator's d i s t a n c e of r e a c t i o n . In the present experiments experience could have a f f e c t e d e i t h e r the v e l o c i t y of the f i s h or t h e i r r e a c t i v e d i s t a n c e . On the b a s i s of o b s e r v a t i o n , i t d i d not appear as i f t h e i r v e l o c i t y changed c o n s i d e r a b l y from one experiment to the next. Therefore, Figure. 1. The effect of experience on the feeding time and react ive distance of 2 groups of t rou t . The open c i r c l e s indicate the average amount of time i t took 9 f i s h to capture 6 standard (white) test prey; the range i s presented. The closed c i r c l e s show the average change in the react ive distance of 6 d i f ferent t r o u t . Further explanation i s given in the text . _ 7 -c 6 -E — 5 -CD E 4 -•- 2 - 1 "D 0 CD L L 1 -6 I i o o 1 "i—r 1 2 T •3 0 I 1 2 1 T I o T T o T T T o o r r - 7 0 J — 6 0 CD o - 5 0 £ - 4 0 1 - 3 0 a, — 2 0 ^ - 1 0 g cc 4 5 6 7 8 9 Days of exper ience the e f f e c t of experience on t h e i r r e a c t i v e d i s t a n c e was examined. EXPERIMENT B THE EFFECT OF EXPERIENCE ON REACTIVE DISTANCE In these experiments 6 naive t r o u t were exposed to standard prey i n the r e c t a n g u l a r tank. A f t e r t h i s group passed through the l a t e n t phase, t h e i r r e a c t i v e d i s t a n c e was recorded f o r up to 16 c o n s e c u t i v e days of experience. The r e s u l t s i n d i c a t e d t h a t the i n t i t i a l r e a c t i v e d i s t a n c e of every i n d i v i d u a l i n c r e a s e d with f u r t h e r experience before i t f i n a l l y s t a b i l i z e d at a c o n s i d e r a b l y higher l e v e l ( f i g . 2 ) . In most cases the t r o u t r e q u i r e d 6 to 7 days of experience (about 40 exposures to prey) to develop a maximum r e a c t i v e d i s t a n c e . One animal (4) however, r e q u i r e d somewhat more time. Both of these aspects, the i n i t i a l a t t a c k d i s t a n c e and the time r e q u i r e d to develop a response, i n d i c a t e that the process of l e a r n i n g can vary c o n s i d e r a b l y between i n d i v i d u a l s . N e v e rtheless, i f these data are pooled and averaged, i t i s apparent that the change i n r e a c t i v e d i s t a n c e of the second group of f i s h i s i n v e r s e l y c o r r e l a t e d with the d u r a t i o n of the f e e d i n g p e r i o d that was r e q u i r e d by the f i r s t group ( f i g . l ) . T h i s suggests that as the f i s h acguired experience, an i n c r e a s e i n t h e i r d i s t a n c e of r e a c t i o n could have been the causal f a c t o r behind the i n c r e a s e i n t h e i r r a t e of a t t a c k (decrease i n feeding t i m e ) . In the process of learning the f i r s t few experiences trout have with new prey are l i k e l y to have the greatest effect on t h e i r response. This hypothesis can be expressed as: d RD = a (RD - RD) d ~ T M A X where (RD) i s the distance of react ion for a given leve l of experience (E), (a) i s a rate constant and (RD = ) i s the max maximum distance from which a conditioned animal w i l l attack, This expression integrates to, - a (E) M R D = RDmax ( 1 " e } The average value of ( R ^ m a x ) was ca lculated by pooling a l l the data in Figure 2 for the l a s t day of experience (Table l ) . Once th i s parameter i s known, the value of (a) can be estimated by standard regression techniques i f ( l ) i s f i r s t l i n e a r i z e d by a logar i thmic transformation. A regression analys i s was conducted by grouping a l l the data in Figure 2. The resul t s showed that i f these data were transformed they could be described adequately by a s t ra ight l i n e ( r = 0.90 ) but that the l i n e did not pass through the o r i g i n . Therefore equation ( l ) was modified to include an intercept (b) . The value of (a) and (b) are presented in Table 1. In most cases, the trout were able to double the i r i n i t i a l Figure 2-. The effect of experience on the react ive distance of 6 t rou t . The test prey were 'white ' (5 mm in length) . Each data point represents a mean of 6 r ep l i ca te observations, the range i s indicated for several days of experience. 100-8 0 -6 0 -40-20-8 e T • 1 . 1 t E o 100 -1 CD g 8 0 -05 5 60-" O o 40-> o 2 0 -03 CD 0C I I I I I I ! I 1 I I I I I I 3 .1 I 1 I I I I I I I I 1 II ! I I I 10O-I 8 0 -60-4 0 H 20-• # T • • 1 • • T 1 T I I I I I I II M I I ! II 2 4 6 8 10 12 14 9 T '.1 I I I I I I I I I I I I I I I I T . • * * -1 T I • *** • .1 i i i ! I I I I I I I I I I I I 7 JT . T - .I T • • • • 2 • • o 1 1 I I I ! I I 2 4 6 I I II ! I I I I 8 10 1 2 14 16 Days of experience TABLE 1. The r e l a t i o n s h i p between the i n i t i a l (RDT) and c o n d i t i o n e d r e a c t i v e d i s t a n c e (RD ). The simulated max prey were 'white' (high c o n t r a s t ) and 5 mm i n l e n g t h . The average values of (K), (a) and (b) were obtained by p o o l i n g a l l the data. Further e x p l a n a t i o n i s given i n the t e x t . F i s h RDT RD K max (cm) (cm) 1 20 56 2.8 3 26 64 2.5 4 24 56 2.3 7 36 58 1.6 8 32 56 1.8 9 40 74 1.8 Average Parameter Values RDT = 29 RD ? = 61 max K = 2 . 0 a = 0.466 b = -0.038 naive react ive distance after 6 days of experience. If the maximum distance from which a conditioned animal ; w i l l react i s assumed to be a constant function (K) of i t s i n i t i a l response (RDT) (Table l ) then the re la t ionship between these two parameters i s s imply: (2) R D m a x - K (RDT) By subs t i tut ing eguation (2) into ( l ) , the effect of experience can be expressed in terms of an animal 's naive response rather than i t s maximum react ive d i s tance . The resul t i s , (3) RD = K (RDT) ( 1 - e " a ( E ) + b ) THE SPECIFICITY OF THE ATTACK RESPONSE OF CONDITIONED TROUT Since trout can increase the i r react ive distance through learn ing , the question which ar i ses i s just how spec i f i c i s t h e i r response to a prey? Most vertebrate and some invertebrate v i sua l systems receive at least 4 d i s t i n c t pieces of information about any ob ject : l ) s i ze , 2) form, 3) contrast and 4) v e l o c i t y . Therefore a target i s not just one stimulus but rather i s a composite set of at least these 4 v i sua l proper t ie s . Undoubtly, an animal could use several i f not a l l of these cues to form a learned a s soc ia t ion . Some cues however, might be more important than others . To answer th i s question, I decided to condit ion trout to standard white prey and then switch them to another object with i d e n t i c a l physical propert ies except for contras t . IF the response of a conditioned animal i s not s p e c i f i c then i t s react ion to a 'new' object should not change with experience. However, since I choose to a l t e r the l e v e l of prey contrast to test for the s p e c i f i c i t y of the attack response, another var iab le must be taken into cons idera t ion . Hester (1968) as well as others have documented that a v i sua l animal must detect a threshold l e v e l of contrast before i t can discr iminate an object from the background. Underwater, the contrast of a target w i l l appear to attenuate as one moves further away from i t . As a r e su l t , i f an object has a high l e v e l of contrast i t can be detected from a greater distance than one with less contras t . Consequently, i f trout are conditioned to a white target and then switched to one with lower contrast ( i . e . black prey) t h e i r maximum distance of react ion should be d i f f e r e n t . The white prey should be attacked from a greater d i s tance . This does not indica te response s p e c i f i c i t y but i s predic table on the basis of v i sual mechanics. The question i s , however, w i l l the react ion of t rout , conditioned to one target ( i . e . white) , change as they acquire experience with a 'new' prey. If the i r response i s not spec i f i c then they should react to the 'new' object from a "maximum distance" on the f i r s t day; i f t h e i r response i s s p e c i f i c , then t h e i r distance of reaction should improve with experience. In e i ther case, i f the contrast 0 f the 'neui' prey i s lower than the o r i g i n a l then the maximum react ive distance should be l e s s . Nine trout were conditioned to standard white prey u n t i l t h e i r react ive distance s t a b i l i z e d for 4 successive days of experience. They were then assigned, at random, to be switched to e i ther a l i g h t grey, dark grey, or black prey. Two f i s h were assigned to each type of prey, with the exception that 4 f i s h were transferred to black prey. The remaining animal served as a control (white) . On the f i r s t day of exposure the f i s h transferred to the l i g h t and dark grey prey reacted immediately. Their response did not change as they acguired addi t iona l experience, they attacked without hes i ta t ion and from a maximum distance ( f i g s . 3B, 3C). For the trout exposed to dark grey prey, however, the effect of target contrast was apparent because the i r average distance of react ion was considerably less than the contro l ( f i g . 3A). There was a noticeable change in the behaviour of the trout exposed to black prey. On the f i r s t day every ind iv idua l had to be released from the holding area an average of 20 times before i t would react . This lag in te rva l corresponds to the o r i g i n a l la tent per iod, although in th i s case, i t was not as pronounced. Part of the explanation may be due to the fact that the experimental environment was extremely simple and that the 'new' prey retained many of the phys ica l c h a r a c t e r i s t i c s of the o r i g i n a l ob ject . The second and by far the most pronounced change in behaviour was in the react ive distance page 60 omitted in page numbering component. These data are shown in Figure 3 D and c l e a r l y indicate that the attack distance increased s i g n i f i c a n t l y as the f i sh acquired more experience. The behaviour of the f i s h transferred to black prey can be summarized by saying that they i n i t i a l l y appear to be ( searching ' for something e l s e 1 . Once they began to respond to the new prey, however, they were able to increase the i r distance of react ion which demonstrates that they were not t reat ing every object in the tank i d e n t i c a l l y . Their conditioned response, therefore , was somewhat s p e c i f i c . The hypothesis that both the rate of learning (a) and the l e v e l of the conditioned response (K) are independent of contrast can be tested with the data presented in Figure 3 D . Using the estimates of (K), (a) and (b) from previous experiments (Table l ) , and the i n i t i a l distance from which trout attacked black prey (RDT) as a s t a r t ing point , the change in t h e i r response was predicted from equation ( 3 ) . Since there i s a reasonably close f i t between the observed and predicted trend (curve, f i g . 3 D ) , t h i s demonstrates that the rate of learning was not. affected by prey contras t . To summarize, these experiments demonstrate three po int s : l ) the response of conditioned trout i s somewhat s p e c i f i c , 2 ) the maximum distance trout w i l l react to prey i s dependent upon target contrast , and 3 ) prey contrast does not affect the rate of l e a r n i n g . Figure 3. The effect of switching t rout , conditioned to white prey, to prey with d i f ferent l e v e l s of contrast . The data points indicate the mean react ive distance of each group. The 9b% confidence l i m i t s of each mean on the f i r s t and l a s t day of experience are shown. (A) white prey ( cont ro l ) , (B) l i g h t grey prey, (C) dark grey prey, (D) black prey. 7 0 . - 1 6 0 -5 0 -4 0 -3 0 -2 0 H 10-A T • I T • 1 1 1 I I I I I I I I 2 4 6 8 1 0 7 0 -6 0 -5 0 H 4 0 -3 0 -2 0 -1 0 -T • i T • " I I I I I I I I I I I 2 4 6 8 1 0 7 0 -6 0 -5 0 -4 0 -3 0 -2 0 -1 0 H B i . 1 T '"•I 7 0 - j 6 0 -5 0 -4 0 -3 0 -2 0 -1 0 H I I I I I I I I I I 2 4 6 8 1 0 D T 1 T/ • 1 I I I I I I I I I I 2 4 6 8 1 0 Days of exper ience THE EXTINCTION AND RE-DEVELOPMENT OF REACTIVE DISTANCE To examine the effect of long-term deprivat ion of reinforcement on react ive distance, the 4 f i s h previously conditioned to black prey ( f i g . 3D) were deprived of further experience for 90 days. During th i s period they were fed standard white prey. Upon re-exposure to black prey the i r react ive distance was recorded for several successive days. Figure 4 shows that the group's i n i t i a l response after re-exposure (18 cm) was not s i g n i f i c a n t l y d i f fe rent from t h e i r response when they were naive (20 cm). In add i t ion , they required 4 to 5 days of experience to re-develop a' conditioned response which i s s imi l a r to the 5 to 6 days they o r i g i n a l l y took ( f ig . 3D). Although the effect of short-term deprivat ion of re inforc on react ive distance was not examined in d e t a i l , ene set of experiments indicated trout can maintain a maximum response for up to 14 days without reinforcement. Therefore, some period of depr ivat ion between 14 and 90 days i s su f f i c i en t to reduce the react ive distance back to the o r i g i n a l l e v e l (RDT) when the animal was naive. ATTENTION COMPETITION The experiments described thus far have been concerned with the effect of experience on the distance of reaction when a predator was exposed to one type of prey. If i t were faced Figure 4. The re-development of the react ive distance of 4 t rou t . The prey were 'black' (5 mm in length) . The data points represent the average distance of r eac t ion . The 95% confidence l i m i t s of the means are shown. The curve was f i t t e d by eye. 4 0 - 1 E o 3 0 -CD O C 03 I r in 2 0 -CD > o CTj 1 0 -CD DC IX T •1 T i i — i — i — i — i — m — i — i — i 2 4 6 8 10 12 Days of exper ience with a s i tua t ion in which i t could encounter other f ami l i a r objects then some form of at tent ion .competition, or interference might occur. For example, the distance trout w i l l react to one prey might be somewhat diminished i f they attempted to become general ly responsive to a l ternate forms. To examine th i s p o s s i b i l i t y , 4 trout were conditioned to low contrast (black) prey u n t i l the i r react ive distance s t a b i l i z e d . They were then switched to a s i t u a t i o n in which each time they were released they could encounter e i ther a black, white or dark grey prey, with equal p r o b a b i l i t y . These targets were i d e n t i c a l except for t h e i r contrast . Although the i r most recent experience had been confined to black objects the f i s h had been previously exposed to the a l ternate types and were therefore somewhat f ami l i a r with them. Before they were switched, the trout attacked black prey from an average distance of 32 cm ( n - 15; 1 S .E . +_ 1.0 ); when a l ternates were present, they reacted from a distance of 37 cm ( n a 16; 1 S .E . +_ 2.7 ). This di f ference is not s t a t i s t i c a l l y s i g n i f i c a n t , therefore, i t must be concluded that the presence of a l ternate food did not affect the distance of r eac t ion . DISCUSSION Experience with an unfamil iar prey w i l l a l t e r several components of the feeding behaviour of rainbow t rout : l ) the latency of t h e i r response, 2) the development of a complete attack sequence, and 3) the react ive dis tance. A l l of these aspects have some theore t i ca l importance, the l a t t e r observation however, i s by far the most s i g n i f i c a n t . Beukema (1968) demonstrated that these behavioural a l t e ra t ions also occured when s t icklebacks were exposed to a novel food and led him to in fer that such changes, e spec ia l ly in the distance of react ion, were necessary i f a predator was to develop, a searching image. The present study demonstrates that each time a predator, which i s capable of l ea rn ing , attacks a palatable prey i t w i l l increase i t s react ive distance for that object . One c r i t i c a l condit ion must be met, however, before the attack response can develop' any fur ther . That i s , another prey must be encountered before the new attack distance diminishes back to the o r i g i n a l l e v e l . If the rate of learning i s faster than the rate of response ext inct ion then even a few contacts with a r e l a t i v e l y rare prey could be su f f i c i en t to promote the development of a searching image. The di f ference between these two antagonist ic rates w i l l determine the density of prey that i s required before a predator can form a searching image (maximum react ive d i s t ance ) . Many prey populations tend to be polymorphic with respect to co lor , form, or some other v i sua l q u a l i t y . Croze (1970) showed that i f predators d iscr iminate between morphs then a polymorphic population w i l l be less vulnerable to attack than a monomorphic population of the same dens i ty . This conclusion however, may not be appl icable when prey are extremely abundant ( H o l l i n g , 1965). In any case, over a wide range of dens i t ies the extent to which polymorphism w i l l decrease the r i sk of a prey i s affected by, among other things, the s p e c i f i c i t y of the searching image. The r i sk from predation w i l l be greatest when the morphs are similar' enough to be treated i d e n t i c a l l y by a predator. On the other hand, predation w i l l be minimized when the predator w i l l react to only one form. Although th i s w i l l resul t in maximum protect ion for the populat ion, predation w i l l also be low i f the predator does not react to each morph from a maximum dis tance . In the present study, the response of conditioned trout was shown to be somewhat s p e c i f i c but that they were able to increase t h e i r responsiveness to 'new' prey at a low l e v e l of reinforcement (6 encounters every 48 hours). These c h a r a c t e r i s t i c s imply that they have the potent ia l to learn to discover polymorphic prey that are r e l a t i v e l y scarce. Hol l ing (1965) as mentioned e a r l i e r , developed a general model of the learning process. One of the assumptions of the model i s that pathways of associat ive learning do not interact but are formed independently of ex i s t ing paths. The a l te rna t ive to th i s i s the concept of inter ference , in which the presence of a l ternate s t imul i could affect e i ther the rate of development of new pathways of learning or the performance of already ex i s t ing ones. Interference i s supported by some data for h u m a n s . S h i f f r i n (1970) reported that the probab i l i ty that an ind iv idua l would r e c a l l a p a r t i c u l a r item was inversely re lated to the number of elements associated with the task. Attention competition, which i s but one poss ible form of inter ference , could not be demonstrated for trout nor for s t icklebacks (Beukema, 1968)., Under the experimental condi t ions , however, both animals had to recognize only 2 or 3 types of prey; under more natural condit ions predators w i l l detect both a wider var iety of prey as well as c o n f l i c t i n g s t imul i from the environment. Thus the p o s s i b i l i t y of interference i s increased. The learning model proposed by Hol l ing also assumed that the hunger l eve l of a predator determined whether i t would attack or ignore any prey i t encountered. For an attack to occur, the predator must be hungrier than the attack threshold i t has set for that prey. Learning operates by r a i s ing or lowering the attack threshold from some i n i t i a l general l e v e l . If the prey i s palatable , then the attack threshold i s lowered with each successive encounter; i f i t i s unpalatable, then the threshold i s ra i sed . Since the react ive distance i s postulated to be funct iona l ly dependent upon the predator 's state of hunger the sh i f t in th i s threshold i s overt ly expressed by a change in react ive d i s tance . The re la t ionship between the p a l a t a b i l i t y of prey and the attack threshold of trout was not examined. However, there i s considerable evidence to indicate that the amount of food many f i s h w i l l ingest i s dependent upon i t s p a l a t a b i l i t y . Both Sheperd (1970) and Ishiwata (l968e) have shown that f i sh consume considerably less unpalatable food before they vo lunta r i ly cease feeding. Hence, there i s l i t t l e doubt that the attack threshold of many animals i s related p a l a t a b i l i t y . On the other hand, the re l a t ionsh ip between hunger and react ive distance has not been adequately documented for many predatory species . Even though H o l l i n g ' s model w i l l p r e d i c t . t h a t the distance of reaction should increase as trout acquire more experience with palatable prey there i s some question as to the general i ty of the mechanism through which learning i s proposed to operate. Beukema (1968) presented some evidence which implied that the react ive distance of conditioned s t icklebacks was stable over a wide range of hunger l e v e l s . I also have some preliminary data which suggests that short term changes in the amount of food trout have ingested does not change t h e i r responsiveness (Table 2) . Although neither of these studies are s u f f i c i e n t l y deta i led to allow re jec t ion of the hypothesis that hunger affects the distance from which a l l predators w i l l react, the data do lead one to suggest that the effect of hunger on react ive distance should be care fu l ly examined. If the proposed re la t ionship cannot be demonstrated then the learning model must be modified to incorporate an a l ternate pathway in which experience d i r e c t l y af fects a predator 's responsiveness, rather than i n d i r e c t l y through i t s hunger motivat ion. Irrespect ive of the mechanism of l earn ing , i t was shown that trout can double t h e i r attack distance i f they acquire s u f f i c i e n t experience. Therefore, i f cer ta in condit ions preva i l they have the potent ia l to s e l e c t i v e l y exploi t prey. This inference i s supported by several f i e l d studies (A l l en , 1941; Hamilton, unpublished data; Bryan, personal communication) which found that i n d i v i d u a l salmonids often contain just 2 or 3 main food organisms in t h e i r gut in conjunction with a number of a l ternate prey. This phenomenon could be explained to some degree i f predators r e s t r i c t e d the i r hunting a c t i v i t i e s to s p e c i f i c sectors .o f the environment. However, th i s does not appear to be a complete explanation since animals in the same r e l a t i v e area w i l l often feed on d i f ferent organisms. This i s not supr i s ing , because var ia t ions in the experiences and motivational l eve l s of ind iv idua l predators can be expected to affect the rate they encounter d i f ferent prey as well as the condit ions necessary to promote l ea rn ing . In conclusion, the changes in the feeding behaviour of trout, as they acquire experience,. ref 1 ect many of the c h a r a c t e r i s t i c of the searching image reported by Croze (1970). He showed that before crows sh i f t the i r at tent ion the:re i s a lag phase in which they w i l l not react to an unfamil iar object . This corresponds to the la tent phase for t rou t . However, once they discover new prey both animals display a capacity to learn qu ick ly . Although trout appear to require somewhat more time to become completely responsive th i s i s d i f f i c u l t to determine because they were exposed to a d i f ferent schedule of reinforcement. Another c h a r a c t e r i s t i c of the searching image i s that i t TABLE 2. The e f f e c t of hunger on the reactiv/e d i s t a n c e of 3 t r o u t , c o n d i t i o n e d to 'white' prey (5 mm). The hunger index d e s c r i b e s the p r o p o r t i o n of the stomach that i s empty. An index of 1.0 i n d i c a t e s that there was no food i n the gut. An index of 0.5 i n d i c a t e s that the gut was h a l f f u l l . The f i s h were fed one hour before an experiment. (n) i n d i c a t e s the number of experiments F i s h Hunger n Mean r e a c t i v e d i s t a n c e _+ 1 S.E index (cm) 3 3 5 5 1.0 0.5 1.0 0.5 12 6 12 6 84 + 3.3 94 + 1.5 7 9 ++ 3.6 73 + 4.8 9 9 1.0 0.5 12 6 83 + 3.2 90 + 1.8 i s r e l a t i v e l y s p e c i f i c . This i s true to some extent for trout as w e l l , since the group conditioned to white,\prey did not react immediately when they were switched to black prey. In t h i s case, the la tent period was not very pronounced, possibly because the 'new' object retained many of the c h a r a c t e r i s t i c s of the o r i g i n a l prey. Nevertheless these experiments ind ica te , as Croze pointed out, that the searching image i s spec i f i c but can be transferred i f the o r i g i n a l image i s no longer re in forced . SECTION III PREY ACTIVITY AND VULNERABILITY INTRODUCTION The experiments described in Section I showed that trout only attack exposed prey. Therefore, i t i s es sent ia l to d i s t i n g u i s h between the actual density of a prey population and the vulnerable dens i ty . Cranqonyx richmondensis and Hyale l l a azteca are important prey of the trout population in Marion Lake; both species are burrowing amphipods and. tend to spend much of the time ac tua l ly concealed within the sediment. In'-general, the a c t i v i t y of these animals appears to be confined to the short in te rva l s in which they are exposed at the mud-water i n t e r f a c e . In th i s sec t ion , I w i l l examine the effect of water temperature on the proportion of amphipods that are exposed ( v e r t i c a l a c t i v i t y ) as well as the average amount of time exposed ind iv idua l s spend a c t i v e l y moving over the sediment (horizontal a c t i v i t y ) . The resu l t s of these experiments w i l l form the basis of a prey v u l n e r a b i l i t y submodel that w i l l be integrated with the main attack model in Section IV. METHODS AND MATERIALS F i e l d Studies The v e r t i c a l a c t i v i t y in both the laboratory and of Cranqonyx and Hyale l l a was observed the f i e l d . In the f i e l d studies , cores of sediment were removed from Marion Lake with ,a sampler, described by Hargrave (1970), and were transferred with as l i t t l e disturbance as possible into glass stacking dishes (20 cm diameter) . The natural complement of bottom fauna was not a l tered with the exception that in some dishes, the number of Crangonyx was increased 2 to 5 times above the natural dens i ty . The dishes were then placed back into the lake to maintain them under ambient temperature and i l l u m i n a t i o n . After being t rans ferred , the animals were allowed 24 hours to acclimate before observations were i n i t i a t e d . Each experimental ser ies consisted of 4 r e p l i c a t e cores of sediment. During the study period both the incident rad ia t ion (Bel for t , recording pyroheliograph) and the water temperature were monitor ed. The v e r t i c a l a c t i v i t y of Cranqonyx and Hyalel1 a was examined with respect to the average proportion of each population that was exposed above the mud-water interface during 15 consecutive, 10 second observation periods . Observations were conducted several times throughout the day (0800 to 1700 hrs .P' iS . T . ) and were repeated for for up to 4 consecutive days. Eleven completely independent sets of experiments were conducted during the months of May, June and Ju ly . After the termination of each of these ser ies the number of amphipods in each core was determined by sort ing through the sediment. Prel iminary t r i a l s indicated that this method would produce complete recovery. Laboratory Studies In the laboratory, the v e r t i c a l as well as the a c t i v i t y of exposed amphipods was observed at 4 d i f ferent temperatures that ranged from 5 to 20 C. For each experiment, amphipods were removed from Marion Lake, sorted, and then i so la ted by species into separate containers of sediment. In th i s case the sediment had been screened to remove a l l other macroinver-tebrates . Three rep l i ca te populations of each species were observed at each experimental temperature. Before observations were conducted, the animals were allowed 24 hours to acclimate to the experimental condi t ions . The number of animals in each container was ca re fu l ly contro l l ed (equivalent to 100 to 300 Cranqonyx, or 200 to 800 Hya le l l a per. sq. m.) to insure that i t f e l l within the natural range in density of each species (Appendix I'll). Throughout the experiments, the background i l lumina t ion was maintained at about 10 ft-candles^and the length of day standardized at 10 hours (0900 to 1800 hrs P . S . T . ) . The water temperature was contro l l ed to within 1 C. of the desired test temperature. In order to avoid the p o s s i b i l i t y of oxygen deplet ion or s t r a t i f i c a t i o n the water in each container was slowly c i r c u l a t e d . Hya le l l a i s a deposit feeding species (Hargrave , 1970) and was not fed ( a r t i f i c i a l l y ) ; Cranqonyx, however, i s carnivorous and was fed dead brine shrimp. In t h i s case, the amount of food provided was always in excess of what the populations ,would consume between successive feedings. The morta l i ty of both species during the experiments was less than 5%. Observations on the v e r t i c a l a c t i v i t y of Crangonyx and Hya le l l a were conducted before Cranqonyx was fed and were repeated every 5 minutes for up to one hour. Their hor izonta l a c t i v i t y was expressed in terms of the average proportion of time exposed ind iv idua l s spent moving over the sediment. In t h i s case, observations were conducted at i r r egu la r in terva l s throughout the day. RESULTS THE EFFECT OF WATER TEMPERATURE ON THE VERTICAL ACTIVITY OF CRANGONYX AND HYALELLA Under both natural and laboratory condi t ions , Cranqonyx and Hyale l l a spend much of the day buried below the mud-water i n t e r f a c e . When ind iv idua l s are concealed in th i s fashion they tend to remain inac t ive for some time before they re-expose themselves; th i s c h a r a c t e r i s t i c behaviour was pointed out e a r l i e r (Section I ) . Intensive observations indicated that the instantaneous proportion of animals that was exposed was f a i r l y constant over a short period of time but changed appreciably with the ambient environmental condi t ions . A mult iple regression analys i s of the f i e l d studies showed that the v e r t i c a l a c t i v i t y of amphipods was s i g n i f i c a n t l y corre la ted with several environmental parameters (Table l ) . One of the most s i g n i f i c a n t factors was water temperature. Although there i s a strong cor re l a t ion between the l e v e l of incident rad ia t ion and the time of day observations were conducted, the ambient water temperature was not s i g n i f i c a n t l y affected by d i e l changes in the l e v e l of i l l u m i n a t i o n . Since the temperature of Marion Lake changes considerably throughout the year th i s c h a r a c t e r i s t i c of the environment could induce seasonal changes in the a c t i v i t y patterns of the amphipods. Therefore, the apparent re la t ionsh ip between water temperature and the degree of exposure of both populations was tested under cont ro l l ed condi t ions . Laboratory studies v e r i f i e d the f i e l d observation that the instantaneous proportion of animals that was exposed was temperature dependent. If the f i e l d data i s grouped according to the temperature when observations were conducted both the laboratory and f i e l d resu l t s turn out to be very s imi l a r ( f i g s . 1 and 2) . This i s in tere s t ing as the laboratory studies were extremely a r t i f i c i a l compared to the f i e l d experiments. It was found that the observed re l a t ionsh ip between temperature and the proportion of amphipods at the mud-water inter face could be described by the exponential equation: (1) UP. = e m 3 i ( T ) " m i where (VP^) i s the proportion of species ( i ) exposed at any instant in time, (M3^) and(M4^) are constants and (T) i s the temperature in C. ( f i g . 1 and 2; Table 2A). Although there was some ind ica t ion that the i l lumina t ion and time of day might have influenced the ' v e r t i c a l movements TABLE 1. The v e r t i c a l d i spersa l a c t i v i t y of Cranqonyx and H y a l e l l a , in Marion Lake, with respect to several environmental condi t ions . (n) indicates the number of days in which observations were conducted. Variable Corre la t ion Coef f ic ient Cranqonyx I l luminat ion Temp era tu re Time of day 0 .229 0.511 - 0.102 ( n = 46 ) 1.56 5.10 * 0.70 Hyale l la I l luminat ion Temperature Time of day 0.576 0.372 - 0.384 ( n = .43 ) 4.65 * 3.09* 3.09 * * s i g n i f i c a n t at or less than 0.01 l eve l ure 1. The effect of water temperature on the proportion of Cranqonyx that are exposed at or above the mud-uiater i n t e r f a c e . The s o l i d c i r c l e s indica te the resul t s obtained from laboratory experiments; the open t r i a n g l e s , f i e l d experiments. The 95% confidence in terva l s of each of the means are ind i ca ted . (See Table 2A) Figure 2 . The effect of mater temperature on the proportion of Hya le l l a that are exposed at or above the mud-uuater in te r f ace . The s o l i d c i r c l e s indica te the resu l t s obtained in the laboratory experiments; the open t r i ang le s , f i e l d experiments. The 95$ confidence in te rva l s of each of the means are ind ica ted . (See Table 2A) 0 « o w CO O CL X 0 c .40-o •4—' > 03 Q-.30-o CL 0 .. .20-Hyalel la c o o a o . 10H T — i — r — i — r — i — i — i — i 5 10 15 2 0 25 Water temperature (CO of Cranqonyx and e spec ia l ly Hya le l l a (Table l ) , none of these p o s s i b i l i t i e s were followed up. An expression of the density of amphipods, by s i ze , per sq. m. that are p o t e n t i a l l y vulnerable to trout predation in Marion Lake at d i f ferent times of the year can be derived by coupling equation ( l ) with the actual density of prey (D^, Appendix I I I ) , the s ize composition of each population (P ^ .) and the seasonal temperature pattern (Appendix I I ) . The result i s equation ( 2 ) which can be designated as a prey v u l n e r a b i l i t y submodel. r M3. (T) - IY14 , ( 2 ) UN. . = D. P. . e 1 1 J where (UN^.) i s the number of amphipods of species ( i ) , of s ize (j) that are exposed. The seasonal range in the s ize s tructure of Ccanqonyx and Hyale l l a i s summarized in Appendix I I I . The profound effect that water temperature has on the v e r t i c a l a c t i v i t y of the amphipods in Marion Lake i s i l l u s t r a t e d in Figure 3. As ind ica ted , there i s very l i t t l e re la t ionship between the actual density of e i ther species and the number of animals that are exposed and vulnerable to attack from trout throughout the year. The vulnerable segment of each population was estimated from eguation ( 2 ) . ure 3. The re la t ionsh ip between the actual density of amphipods (white histograms) and t h e i r vulnerable density (black histograms). The vulnerable segment of each population corresponds to the number of animals that are exposed on the sediment as a resul t of the ambient water temperature (Appendix II) (A) Hyale l la (B) Cranqonyx No. amphipods per sq. m. ( X i o o ) THE EFFECT OF TEMPERATURE ON THE ACTIVITY OF EXPOSED AMPHIPODS Cranqonyx tend to move almost continuously when exposed. They do stop p e r i o d i c a l l y , however, to feed or to grasp at pieces of l i t t e r . In contrast , Hyale l la tend to be r e l a t i v e l y inac t ive at a l l t imes. Hargrave (1970) demonstrated that Hya le l l a i s a deposit feeding species and must ingest large quant i t ie s of sediment to meet i t s energetic requirements. Thus, when th i s species i s exposed i t appears to spend most of i t s time ei ther feeding or involved in other a c t i v i t i e s which seem to require l i t t l e movement. In the laboratory, the a c t i v i t y of exposed ind iv idua l s was also affected by temperature. The resul t s ( f i g . 4) suggest that 1 0 ° C . is^the optimum for Crangonyx, above or below t h i s , t h e i r a c t i v i t y decl ined somewhat. Unfortunately, observations o n Hyalel1 a were confined to temperatures above 10 C. At temperatures higher than t h i s , the i r a c t i v i t y also declined which suggests that they may have an optimum temperature which i s • s imi l a r to Cranqonyx. Although the water temperature w i l l a l t e r the movement of both species, Cranqonyx was always more act ive than H y a l e l l a . Assuming that amphipods are most act ive at 1 0 ° C . the effect of water temperature on theirgeneral l e v e l of a c t i v i t y can be described by the parabola: (S) PA. M5. + M6. (T) - M7. (T 2 ) l l v ' l in which case, PA^ i s the proportion of time an ind iv idua l of species ( i ) w i l l spend moving when exposed; M5^, M6^, and M7^ are constants that can be estimated by f i t t i n g the data in Figure 4 by mult iple regression (Table 2B). The resul t s presented above demonstrate that the ambient water temperature w i l l a l t e r the proportion of animals that are vulnerable to attack from trout as well as the a c t i v i t y of exposed amphipods. It was necessary to consider the effect of temperature on the movements of vulnerable animals because trout w i l l react to moving prey from a considerably greater distance than stat ionary objects of the same s i z e . The s i gn i f i cance of th i s observation w i l l become apparent in the next s ec t ion . Figure 4. The re l a t ionsh ip between water temperature and the average amount of time exposed Cranqonyx (A) and Hyale l l a (B) spend moving. The optimum temperature for the a c t i v i t y of both species i s assumed to be 10 C. (See Table 2B) Water temperature C O TABLE 2 A . i • The r e l a t i o n s h i p b e t w e e n t h e a m b i e n t w a t e r t e m p e r a t u r e ( T ) a n d t h e i n s t a n t a n e o u s p r o p o r t i o n o f C r a n q o n y x a n d H y a l e l l a 2 e x p o s e d a t o r a b o v e t h e m u d - w a t e r i n t e r f a c e . ( V P ^ ) . (R ) i s t h e amount o f v a r i a b i l i t y a c c o u n t e d f o r by t h e r e g r e s s i o n a n d ( P ) i s t h e p r o b a b i l i t y t h a t t h e s l o p e i s z e r o (no c o r r e l a t i o n ) . S p e c i e s ( p N3.( ) M 4 ( p R 2 P C r a n q o n y x 0.138 4.63 0.81 0.0027 H y a l e l l a 0.180 5.13 0.84 0.010 The r e l a t i o n s h i p b e t w e e n t h e a m b i e n t w a t e r t e m p e r a t u r e (T ) a n d t h e a v e r a g e p r o p o r t i o n o f t i m e e x p o s e d , a m p h i p o d s s p e n d m o v i n g . (R ) i s t h e a m o u n t o f v a r i a b i l i t y a c c o u n t e d f o r by t h e r e g r e s s i o n . S p e c i e s (^) W5(.) M6(. ) + 1 SE |Y17( . ) l + I S E R 2 C r a n q o n y x H y a l e l l a 0.39 0.01 0.069 0.039 0.069 0.049 0.0033 0.0017 0.0027 0.0019 0.58 0.34 SECTION IV A SIMULATION MODEL OF THE PREDATORY BEHAVIOUR OF TROUT INTRODUCTION One of the most common observations of trophic ecology i s that many animals do not exploi t prey in d i rec t proportion to t h e i r abundance and, therefore, feed s e l e c t i v e l y (Lindstrom, 1955; Ivlev , 1961). Ivlev was well aware of th i s phenomenon and devised the term ' e l e c t i v i t y ' to describe how animals exploi t d i f ferent food organisms. Although ' e l e c t i v i t y ' indices or s imi l a r expressions are useful as descr ip t ive statements, they, provide no insight into, or explanation of, the mechanisms responsible for ' s e l e c t i v e ' predat ion. U n t i l these mechanisms are i d e n t i f i e d , there i s l i t t l e hope of a r r i v i n g at a set of genera l iza t ions to account for th i s phenomenon, which, as Ivlev (1961) pointed out, has tremendous s i gn i f i cance both to evolutionary b i o l o g i s t s ( i . e . Batesian mimicry) and community ecologi s t s ( i . e . energy t r a n s f e r ) . There i s s u f f i c i e n t evidence in the l i t e r a t u r e to suggest that the act ion of any one of three basic components of the feeding process: l ) prey detect ion, 2) prey handling, and 3) learning behaviour, could resul t in ' s e l e c t i v e ' predation; These components operate at d i f ferent stages in the feeding process and some, such as l earn ing , may not be common to a l l animals. Prey detect ion , however, i s a fundamental stage in the feeding of a l l animals, except, perhaps, f i l t e r feeders. Every sensory system i s l imi ted in i t s capacity to receive information and depending upon i t s mode of operation i s biased toward detecting cer ta in types of s ignals ('adequate s t i m u l u s ' ) . For example, the process of v i sua l d i sc r iminat ion i s highly s ens i t ive to both the s ize and contrast of a target (le Grand, 1967; Hester, 1968); while chemoreceptors are sens i t ive to the concentration and nature of the st imulus . Predators w i l l ' s e l e c t i v e l y ' detect some species i f they react d i f f e r e n t l y to the s t imul i emitted by prey. If food se lec t ion can be explained simply on the basis of the process of d i sc r iminat ion i t can be referred to as perceptual s e l e c t i o n . Most predators cannot successful ly pursue (Ivlev, 1961) or capture (Dixon, 1959; H o l l i n g , 1964) every prey they detect . At some point , the a b i l i t y of an animal to capture food i s l i k e l y to be dependent upon prey s i z e . In the animals in which th i s has been looked at there tends to be an optimum sized prey that can be handled most succes s fu l ly . Therefore se lec t ion can also operate at th i s l e v e l . If food se lec t ion can be explained on the basis of d i f f e r e n t i a l capture success i t can be referred to as mechanical s e l e c t i o n . F i n a l l y , many animals have the capacity to learn and thereby a l t e r t h e i r response to prey through experience (Croze, 1970; Beukema, 1968). In the learning process the p a l a t a b i l i t y of a food i s <3f major importance. Most animals w i l l learn to avoid unpalatable objects ( H o l l i n g , 1965; Morre l l and Turner, 1970; Prop, I960) but w i l l increase the i r responsiveness to more palatable prey (Section II ; Beukema, 1968). Therefore, learning could also be responsible for the disproport ionate exp lo i ta t ion of some prey by predators . Select ion at th i s l e v e l can be referred to as behavioural s e l e c t i o n . In th i s sect ion I w i l l examine the process of prey detect ion and recogni t ion , and test the hypothesis that the s e l ec t ive exp lo i ta t ion of several invertebrate prey (espec ia l ly the amphipods) by the trout population in Marion Lake can be explained at the perceptual l e v e l . Since the s imulation model that w i l l be developed to test th i s hypothesis w i l l not consider the effects of hunger or l earn ing , among other things, i t i s not intended to be a complete descr ip t ion of the predatory behaviour of t rout . The model however, was structured so that i t could be eas i ly modified to incorporate these components as more information became a v a i l a b l e . METHODS AND MATERIALS'" Nine rainbow trout which ranged in length from 11 to 14 cm. were obtained from Marion Lake. After the f i sh were transferrer, to the 'laboratory a month of prel iminary experiments were conducted to habituate them to being handled and to condit ion them to respond to a r t i f i c i a l food. The test prey were formed from pieces of chicken l i v e r . Prel iminary experiments indicated that the f i s h considered th i s food to be palatable as they would av id ly consume i t ; they would not respond to other ' l e s s pa la tab le ' foods t h i s r e a d i l y . Throughout the experiments, the form of the prey was standardized (rectangular) The c h a r a c t e r i s t i c s of target s i ze , contrast and motion were a l t e r e d , however, to determine how trout would react to prey with d i f ferent v i sua l proper t ie s . The s ize of the prey was a l tered simply by changing the i r length and holding the i r width constant (3 n.m). The inherent contrast of a target (CQ) can be defined as i the dif ference in luminous flux re f lected by the object (L ) with respect to the background (R) (le Grand, 1967). That i s , R - L * C = R o Two leve l s of contrast were examined. In one case, pieces of l i v e r were stained 'b l ack ' by immersing them in a saturated so lu t ion of Sudan Black B (water insoluble s t a i n ) . For the other l e v e l of contrast the prey retained the natural color of l i v e r ( ' w h i t e ' ) . The inherent contrast of both the 'white ' and 'b lack ' prey was determined with a photometer (Photovolt, model 200) with a neutral density f i l t e r . Under the experimental condit ions the background (tank bottom) re f lec ted 0.3 f t . - cand le s , the inherent contrast of the undyed.prey was found to be 0.67 and that of the 'b l ack 'prey , 0.14. Once the f i s h had been conditioned to respond to a r t i f i c i a l food, the distance from which they would react was observed in a large rectangular tank (180 x 16 x 30 cm), constructed of dear p l e x i g l a s s . At one end, there was a small holding chamber with an opaque, s l i d i n g p a r t i t i o n that was used to i so l a te the f i sh before an experiment. While the f i s h was in the holding area a prey of known s ize and contrast was placed into the tank. The trout was then released and the distance from which i t would react was recorded. After an attack was completed the predator was returned to the holding area while another prey was introduced. The pos i t ion of the food was randomized between successive t r i a l s . Up to 10 successive attacks were recorded for each f i sh during a test per iod . The hunger l e v e l of the f i s h was standardized by adopting a 48 to 72 hour period of food deprivat ion between successive experiments. This was known to be s u f f i c i e n t time to completely c lear a l l the food consumed during the previous meal from the d iges t ive t r a c t . An addi t iona l ser ies of tests demonstrated that the react ive distance was not affected by short term changes in hunger. If the predators were fed up to 50% of the maximum amount of food they could ingest , they reacted from the same distance as they would after a 48 to 72 hour period of food depr iva t ion . Therefore the s l i g h t change in hunger that occurred during an experiment was u n l i k e l y to have affected the distance from which the trout attacked. The effect of prey motion on react ive distance was examined by placing standard 'white ' or 'b lack ' targets on a small platform that moved v e r t i c a l l y to simulate a slow moving animal (3 mm per s e c ) . Control tests demonstrated that the f i s h would respond only i f the platform was supporting food. The experiments dealing with the effect of background d i v e r s i t y on react ive distance and prey recognit ion success were conducted under the same condit ions as those described above. Except in th i s case, the background uuas a l tered ( 'broken') to simulate a diverse substrate. The element of d i v e r s i t y uuas created by scat ter ing small , ' b l ack ' pebbles, the same s ize and contrast as the prey (5 mm; contrast = 0.14), uniformly over the bottom of the tank. The mean distance between adjacent pebbles uuas in the order of 0.5 cm. Before a predator uuas released, a s ing le prey was placed at random into the tank. The distance from the trout attacked both stat ionary and moving targets as well as the i r a b i l i t y to recognize prey under these condit ions was recorded. A f a i l u r e in recognit ion was considered to have occurred i f a f i sh passed by a potent ia l target without attacking*. Recognition success, therefore , was defined as the rat io of the number of attacks that were i n i t i a t e d to the number of opportunit ies the f i s h had to discover prey. These experiments were rep l ica ted over 6 days to determine i f the performance of the trout would improve with experience. RESULTS THE CHARACTERISTICS OF THE VISUAL RESPONSE OF RAINBOW TROUT TO PREY The Relationship Between Prey Size and Contrast Threshold Since the aim of th i s sect ion i s to develop a model to describe the response of a v i sua l predator, such as rainbow trout , to prey i t would be des irable to seek genera l i ty and, therefore , interpret the process of prey detection in terms of a general theory of v i sua l d i s c r i m i n a t i o n . The inherent contrast of an object was defined as the amount of l i g h t i t r e f l e c t s with respect to the background when the distance between i t and an observer i s zero. Due to sca t ter ing and absorption of re f lected l i g h t by the bakcground, th i s contrast w i l l appear to diminish as one moves further away from the ob ject . Duntley (1963) and le Grand (1967) as well as others have found that the re la t ionship between distance and apparent contrast can be described by the negative exponential equation: - 6 (X) C a = C o e where C = the apparent contrast of a target at distance 3 (X), C q = the inherent contrast of a target (X = 0), 6 = the rate of ext inct ion of target contrast . The attenuation coe f f i c i en t (6) should be spec i f ied in terms of the wavelengths of re f lected l i g h t because longer wavelengths are absorbed most rapid ly incwater (Sverdrup et a l , 1942). However, the addi t ion of th i s component would add considerable complexity to the model, therefore , (6) was defined simply in terms of the to ta l attenuation of l i g h t i r re spec t ive of wavelength. The rate of attenuation was determined by measuring the inherent contrast of the prey (photometer with a neutral density f i l t e r ) under the standard experimental condit ions as well as the i r apparent contrast at a distance of 1 meter. Equation ( l ) was then solved to estimate (6). The value of th i s parameter was found to be about 0.50 for both dyed and unstained targets . In the study of v i sua l d i scr iminat ion i t i s conventional to express the s ize of a target in terms of the v i sual angle i t subtends with the ret ina of an observer. This angle ( s ) i s defined as: (2) S = tan 8 = TD/X where (TD) i s the length of diameter of a target and (X) i s the distance between i t and an observer. In the present study the re la t ionsh ip between the v i sua l angle prey subtended with trout and the apparent contrast they presented when they were attacked can be described i f the react ive distance i s expressed as a v i sua l angle. The average distance 4 trout reacted to d i f ferent s izes of prey (inherent contrast , 0.14) i s indicated in Table 1. These data were transformed to a v i sua l angle by subs t i tut ing (RD) for (X) in equation (2) and so lv ing for (S). It was also possible to estimate the apparent contrast of each target , when the f i sh attacked, because the inherent contrast of the prey (CQ) and the attenuation rate (6) were known. The apparent contrast was ca lculated by subs t i tut ing the react ive distance (RD) for (X) in equation ( l ) and solving for (Ca). The resul t s of th i s transformation (CT) are presented TABLE 1. The r e l a t i o n s h i p between prey s i z e (TD) and the average r e a c t i v e d i s t a n c e (RD). (CT) i s the apparent c o n t r a s t of the black prey and (S) i s the v i s u a l angle they subtended when they were a t t a c k e d . The r e s u l t s were obtained from r e p l i c a t e experiments with 4 f i s h , (n) i n d i c a t e s the t o t a l number of o b s e r v a t i o n s . TD (m) n RD (m) + 1 S.E. CT S (min) .0020 20 .20 .017 .127 34 .0035 20 .30 .017 .120 40 .005 20 .32 .016 .119 54 .006 20 .37 .019 .116 56 .009 20 .44 .022 .112 70 .012 20 .49 .018 .109 84 .015 20 .52 .018 .108 100 * inherent c o n t r a s t (C = 0.14) in Table 1. Although the trout responded to large prey from a greater distance than smaller targets the re la t ionsh ip was not l inea r (Table l ) . Another way of s tat ing th i s observation i s that there was an inverse r e l a t i o n between the v i sua l angle that a prey subtended and i t s apparent contrast (CT) when i t was at tacked. This i s d iagramatica l ly i l l u s t r a t e d in Figure 1. There are two l i m i t s to th i s function, however, that are not apparent from the experimental data (Table l ) . The f i r s t c h a r a c t e r i s t i c i s that there i s a minimum l e v e l of contrast (CT . ) that can just be discriminated by a v i sua l animal, min ° 1 If the apparent contrast of a target does not exceed th i s l e v e l then i t cannot be detected ( f i g . l ) . Secondly, there i s also a minimum v i sua l angle (S . ) that must be subtended before mm a target can be d i scr iminated . The performance of any v i sua l system i s r e s t r i c t e d by these two l i m i t s of reso lut ion (Hester, 1968; l e Grand, 1967). If these appropriate l i m i t s are defined, the re la t ionship depicted in Figure 1 can be described by the negative expotentual equation: L - B ( ln(S)) (5) CT = (CIYl)e In which case, (L) and (B) are constants and (ClYl) is the contrast threshold when a prey subtends the minimum visual anqle (S . ) . The r e s t r i c t i o n s of equation 3 were i l l u s t r a t e d y v min' M in f igure 1 and can be summarized as fo l lows : Figure 1. A diagramatic representation of the re la t ionsh ip between the contrast threshold, or the apparent contrast a target must have in order to be d i scr iminated , and the v i sua l angle i t subtends with the eye of an observer. The two l i m i t s to th i s function are (CT . ) and (S . ); (CT , ) 111 JL I I I I I .1* I t 111 JL I t i s the minimum l e v e l of contrast that can be detected, while (S . ) i s the smallest , or minimum v i sua l angle a min 3 target can subtend and s t i l l be d i sc r iminated . The point (Cffl) i s the contrast threshold for a target which subtends the minimum visual angle at a spec i f i ed l e v e l of i l l u m i n a t i o n . Curve (A) simulates the function at a low l e v e l of i l l umina t ion curve (B), at a somewhat higher l e v e l . S min S max V i s u a l a n g l e 1) (S) cannot be less than (S . ), 2) (CT) cannot be less than ( C T m ^ n ) . Some addi t iona l information i s required, however, before the constants of equation ( 3 ) can be estimated. In the present study, neither the minimum vi sua l angle nor the minimum contrast threshold of trout were obtained. Values for each of these parameters, however, have been reported for other animals (Table 2) . Therefore I have assumed that 0.05 i s the minimum contrast that trout can discr iminate (after Hester, 1968) and that the i r minimum v i sua l angle i s in the order of 5 minutes of arc (Nakamura, 1968; Yamanouchi, 1956; Tamura, 1957). Although neither of these parameters weremmeasured d i r e c t l y for t rout , they represent values that have been found for f i sh that have well developed v i sua l systems. If the data in Table 1 are transformed to logarithms and In (CT) i s regressed against In (S), the contrast threshold (Cffl) when a target subtends a v i sual angle of 5 minutes of arc can be estimated, for t rout , by ext rapo la t ion . Once th i s value i s obtained the remaining parameters of equation ( 3 ) can be derived by regressing In (CM/CT) against l n ( s ) . In which case, (8) i s the slope of the l i n e and (l_) i s the Y- intercept . A regression analys i s demonstrated that the data in Table 1 could be adequately described by a s tra ight l i n e ( r = 0.94 ) ; the estimated values of (B) and (L) are presented in Table 5. The Relation Between the Ambient I l luminat ion , Visual Angle and Contrast Threshold It is well documented that the background i l lumina t ion also affects the process of v i sua l d i s c r i m i n a t i o n . This i s i l l u s t r a t e d in Figure 1, where curve (A) simulates the re l a t ions between the v i sual angle and contrast threshold at a low l e v e l of i l l u m i n a t i o n , and curve (B), the function at a somewhat higher l e v e l . Thus, for any angle less than (S ) the contrast max required for d i sc r iminat ion w i l l decrease i f the background i l l u m i n a t i o n i s ra i sed . This means that a v i sua l predator w i l l be able to detect prey from a greater distance at higher l eve l s of i l l u m i n a t i o n . This r e l a t ionsh ip , however, only holds over a spec i f ied range. Before any o p t i c a l system can function the ambient i l lumina t ion must surpass some lower threshold ( ^ m j _ n ) « On the other hand, once the i l lumina t ion reaches some upper l e v e l ( r ' m a x ) the system w i l l perform maximally. Any further increase in the background illumination. : w i l l not improve this performance. Several reported estimates of the upper and lower l e v e l s of i l l u m i n a t i o n that affect the v i sua l acuity of f i sh are presented in Table 2. The influence of i l lumina t ion can be described through i t s effect on the component (CM). Hester (1968) and le Grand (1967) have shown that (CM) diminishes at a decreasing rate i f the background i l l u m i n a t i o n i s raised from the l i m i t of scotopic v i s i o n (R . ) to (R ). This re la t ionsh ip can be approximated v mm' max K by the negative exponential equation: TABLE 2. A comparison of several documented values of the minimum detectable contrast (CT . ) and the minimum visua l angle (S . ) of d i f ferent animals. (R • ) v min' i s the l i m i t of scotopi c v i s ion and (R ) max' i s the lowest leve l of i l luminat ion which produces maximum visual acu i ty . Animal CT . min S . min (a) R XI) R . min (b) Source Human 0.01 0;5 30 3.0 X I O " 7 le Grand, 1967 Goldf i sh 0.05 20.0 10 1.0 X I O " 2 Hester, 1968 Herring 25.0 1.0 X ID" 4 Blaxter, 1968 b Salmon (6sp;) 1 -• 10 1.0 X i o " 4 A l i , 1959 Pla ice 1.0 X i o " 4 Blaxter , 1968 a Marine Teleosts 4.0 - 15.0 Tamura, 1957 Skipjack Tuna 5.0 10 1.0 X I O " 3 Nakamura, 1968 Jack mackerel 12 1.0 X i o " 7 Hunter, 1968 a angle expressed i n minutes of arc . b the l e v e l of i l luminat ion in ft-candles ( 4 ) CM = K1 e " A ( R ) where (K^) i s the contrast a prey must have to be attacked when i t subtends the minimum vi sua l angle and the ambient i l l u m i n a t i o n i s at the predator ' s scotopic l i m i t of v i s ion ( R m ^ n ) ; (A) i s simply a rate constant and (R) i s the l eve l of i l l u m i n a t i o n . The value of (A) was ca lculated from Hester 's (1968) data (Table 5 ) . Although (l^) was not s p e c i f i c a l l y measured for t rout , i t can be estimated by subs t i tut ing the value of (CM) (Table 5 ) at the standard l e v e l of i l l u m i n a t i o n (R = 0.03 ft-candles) into ( 4 ) . Equation ( 4 ) can now be coupled with (3) to obtain a s ing le expression which includes the effect of both the ambient i l lumina t ion and prey s ize on the apparent contrast trout require to d i scr iminate a target . The resul t i s , ( 5 ) CT.., = K e -[A R] + [L - B ( In ( S ) ) ] The inc lus ion of ambient i l lumination,however, imposes two add i t iona l re s t ra in t s on equation ( 5 ) other than those already mentioned in conjunction with (3) : 3) a target cannot be detected i f (R) i s less than (R . ) . m i n ' 4 ) K, reaches a minimum value when (R) = (R ). ' 1 max Now that the re l a t ionsh ip between the i l l u m i n a t i o n , the v i sua l angle and the contrast threshold has been described, i t i s necessary to convert th i s expression into the average distance trout will react to different sizes of prey. This transformation can be accomplished by iteratively solving equations (l) and (5) until a distance (X) is found which satisfies the equality (C = CT). In the model, a computer a program was designed to undertake this operation. Figure 2 presents a comparison between the actual reactive distance (Table l) and the distance that was generated in each case, by iteratively solving (l) and ( 5 ) . The close agreement between the observed and calculated distance of attack indicates that the equations that have been developed do describe the experimental results reasonably well and that l i t t l e accuracy has been lost by transforming the original data. The other point illustrated in Figure 2 is that the distance of reaction is tending towards a maximum. The upper limit to the attack distance will occur when the apparent contrast of the target is equal to the minimum level of contrast the animal can discriminate (CT . ) . min The Effect of Prey Movement on Reactive Distance It has been shown that the distance from which'trout will attack depends upon several characteristics of a prey as well as the environment. The equations that were develqped to describe these effects were based upon the reaction of trout to stationary prey. Therefore an additional series of experiments was conducted to determine i f the predators would respond differently to moving prey. The prey in this case had a high Figure 2. A comparison between the observed react ive distance (data points) of 4 t rout , exposed to d i f ferent s izes of prey, and the calculated react ive distance (curve) . The l a t t e r uuas obtained by solving equations (1) and (5). See text for further explanation. The 95$ confidence l i m i t s of the means are indicated by the v e r t i c a l bars. T T T Target length (mm) Figure 3. The effect of target movement on react ive d i s tance . The open c i r c l e s indicate the average react ive distance (4 f i sh) for s tat ionary prey of d i f ferent s izes (inherent contrast = 0 .67) . The s o l i d c i r c l e s shouu the average react ive distance for moving prey of the same s ize and contras t . The 95% confidence l i m i t s of the means are ind ica ted . CD > -t—• O 03 CD DC 140-120H E o CD £ 100-03 •+-> CO 8 0 -6 0 -4 0 -20« 2 T 2 j . T O T O 1——1 5 10 T a r g e t l e n g t h (mm) l e v e l of contrast (0.67) . In the f i r s t experiment, 4 trout were exposed to stat ionary prey of d i f ferent s i ze s , and the average distance from which they would react was recorded. Once these t r i a l s were completed the f i s h were switched to moving prey with the same l eve l of contras t . The resul t s ( f i g . 3) c l ea r ly show that trout w i l l attack moving prey from a s i g n i f i c a n t l y greater distance than s tat ionary objects with the same visual proper t ie s . At least over the range in s ize that was inves t iga ted , the effect of target motion was addi t ive because the trout would react to moving prey 22 cm further away than they would to stat ionary targets of the same s i z e . If the distance from which trout w i l l attack a s tat ionary object of a given s ize and contrast (j) i s defined as (^j)> then the effect of motion on the distance of reaction can be expressed as: (6) R'. = R . + IY1C J J where, (MC) i s the increment effect of motion. For the purposes of th i s paper, I w i l l assume that the addi t ive effect of motion i s independent of the ve loc i ty of the target and the background i l l u m i n a t i o n . The Relation Between the Background, Reactive Distance and Prey Recognition Success When a prey was the only object in the tank and was contrasted against a f l a t , evenly i l luminated surface ('smooth') the trout mere 100$ successful in recognizing i t regardless of whether i t was s tat ionary or moving. However, i f the background was d i v e r s i f i e d ( 'broken') in the sense that other s imi l a r but non-prey objects were scattered over the surface to break up the uniformity, then the t rou t ' s a b i l i t y to discr iminate or recognize prey might be somewhat impared. A 'broken' surface might also a l t e r the react ive dis tance . Both of these p o s s i b i l i t i e s are worth inves t iga t ing because Cranqonyx and H y a l e l l a , as well as other invertebrates in Marion Lake, are exceedingly c rypt i c and l i v e in associat ion with a very diverse background, the sediment. Table 3 presents the resul t s of some experiments which demonstrate that the presence of other objects can diminish the success trout have in d i scr iminat ing prey. Under these condi t ions , the f i sh were less than 100$ successful in discovering both stat ionary and moving targets , although they were considerably more successful R] + L L- BUn(S))] The distance (X) which s a t i s f i e s the equal i ty (C = CT) can be found by solving ( l j and (5). This distance i s then defined as the react ive distance (Rj) f ° r a s tat ionary .'prey, .of s ize and contrast ( j ) . The effect of prey motion i s given by, (6) R! • R. + MC. J J However, i f the background i s broken, then the distance of react ion i s : (7) HD. = (Rj)E, i f the prey i s s tat ionary or (8; RDj = (Rj'jE, i f the prey i s moving. F i n a l l y , the width of the path of search was expressed as (11) E P i j . = 2 (RD i j - SP 2) 2 THE ATTACK MODEL Hol l ing (1966) has shown that 4 basic components determine the rate predators capture prey, they are: 1) the density of vulnerable prey, 2) the width of the path of search, 3) prey recognit ion and capture success, and 4) the predator ' s v e l o c i t y . The seasonal changes in the density of vulnerable amphipods in Marion Lake (Section III) and the width of the path of search have already been considered; the remaining aspects w i l l be treated in th i s sect ion as the attack model i s synthesized. The ve loc i ty at which trout search for food was determined e a r l i e r (Section I) and was found to average about 4 cm/sec. Although th i s parameter was measured in an a r t i f i c i a l s i t u a t i o n , f i e l d observations v e r i f i e d that trout do search slowly and that a ve loc i ty of 4 cm / sec. i s not an unreasonable estimate. These observations, however, also indicated that trout freguently sh i f t t h e i r v e r t i c a l pos i t ion and do not maintain a spec i f i c pattern of search for any length of time. The resul t of th i s sporadic hunting behaviour i s that in the course of an hour a f i sh may e f f e c t i v e l y search only 10 l i n e a r meters of sediment. Although thi s i s a rather crude estimate of the average ve loc i ty-of trout in the f i e l d i t i s assumed to be reasonable value. Depending upon the type of background and the r e l a t i v e a c t i v i t y of d i f ferent prey, trout w i l l success ful ly discover only a r e l a t i v e l y small proportion of animals that are ac tua l ly within the path they sweep along the sediment. Their average success in recognizing prey (RS^) can be estimated by weighting t h i s component with respect to the proportion of animals that are act ive ( P A ^ ) and the a b i l i t y of trout to recognize both moving and non-moving prey. In which case, (14) RS. = [ ( P A . ) (RS ' ) j + [ ( 1 - P A . ) (RS)] where (RS 1) and (RS) are respect ive ly , the p r o b a b i l i t i e s that a predator w i l l recognize a moving and s tat ionary target (Table 3) . One f i n a l component remains to be considered and that i s capture success. This fragment represents the probab i l i ty that a prey w i l l be success ful ly approached to within s t r i k i n g distance and then captured (Ho l l ing , 1966). Trout are 100% successful in approaching both Hyale l l a and Crangonyx, and on the average 84% of the s t r ikes they attempt capture prey (Section I ) . This i s consistent with the general observation that predators which pursue prey tend to be very successful in subduing r e l a t i v e l y slow moving, or s tat ionary animals ( H o l l i n g , 1966; Messenger, 1968) but not necessar i ly faster moving targets (Dixon, 1959; Braum, 1967). In Marion Lake, most of the benthic-l i v i n g invertebrates are less than 15 mm in length and are f a i r l y slow moving, therefore , the component of capture success can be added to the attack model and treated as a constant. In which case, the rate of prey capture (RC^j) can be derived by combining the prey v u l n e r a b i l i t y (Section III) and react ive distance submodels with the searching ve loc i ty (VY), prey recognit ion (RS^) and capture success (CS) components. The resul t i s , ( 1 2 ) RC. . = PEP. . ( V Y ) U N . . IRS. CS This step completes the attack model. As i t stands, i t i s not a complete descr ip t ion of the predation process because i t does not consider the amount of time trout spend handling food, the effect of hunger motivation, or l e a rn ing . These aspects have been shown to affect the i r feeding behaviour (Section I and I I ) . Nevertheless, the purpose of th i s section i s to test the concept that the se lec t ive exp lo i t a t ion of amphipods,as well as several other invertebrates ,can be explained at the perceptual l e v e l . • This model subsequently accomplishes th i s aim. APPLICATIONS OF THE MODEL The se lec t ive Exp lo i t a t ion of the Amphipods, Odonates, Planorbids and Caddis. A previous examination of f i sh stomachs (Efford and Tsumara, unpublished data) indicated that throughout most of the year trout feed extensively upon benthic invertebrates . Four major groups, the amphipods, the odonates, the caddis ( p r i n c i p a l l y Banksiola c ro tch i ) and the planorbids (Menetus and Helisoma), account for about 60% of the to ta l energy input to the trout populat ion. These four groups can therefore be s ingled out as the most important prey. Some data on the average density and s ize of each of these animals, other than the amphipods, are b r i e f l y summarized for several selected months in Table 6 . Although there i s l i t t l e quant i ta t ive information concerning t h e i r v e r t i c a l a c t i v i t y , there i s some evidence which suggests that the odonates and caddis tend to l i v e at the mud-water in ter face , or can be found in areas of vegetat ion. Neither of these invertebrates appear to burrow into the sediment. The planorbids , on the other hand, w i l l move below the mud-water inter face and may have an a c t i v i t y pattern that i s somewhat s i m i l a r to the amphipods (Delury, personal communication). Guided by these rather l imi ted observations, I have assumed that the planorbids have a v e r t i c a l a c t i v i t y pattern that i s i d e n t i c a l to H y a l e l l a , but that the odonates and caddis always remain exposed. I have also assumed (on the basis of some data) that the caddis spend most of the time moving, while the planorbids and odonates are r e l a t i v e l y i n a c t i v e . The eguations developed above were transcribed into Fortran and a computer s imulation was conducted to predict the rate each of these prey could be captured by t rout . These resul t s were then compared with the actual pattern of exp lo i ta t ion that was observed during the months of February, May, June, August and November. These were the only months in which stomach samples were taken from the trout population (Efford and Tsumara, unpublished data) . TABLE 6 . The population c h a r a c t e r i s t i c s of the odonates, caddis and planorbids for several selected months during an 'average ' year. (MD) i s the i tiean i density (no . / sq . m.), (ML) i s the mean length (mm ) and (PA) i s the instantaneous proportion of prey that are a c t i v e . Prey Group Month Odonates (1) Caddis (2) Planorbids (3) MD ML PA MD ML PA MD ML PA Feb 10 10 0 9 7 1 40 4 * May 4 12 0 4 10 1 63 5 * June 5 10 0 4 10 1 63 5 * Aug 4 8 0 1 10 1 57 5 * Nov 5 11 0 7 10 1 50 4 * * assumed to follow the same a c t i v i t y pattern as H y a l e l l a . Data Sources (1) Pearlstone (pers, com. ); Hamilton (1965). (2) Ulinterbourn (pers, com.) (3) Lee (1967) In the s imulation the water temperature (Appendix II) , density and average s ize of each prey (amphipods, Appendix I II ; remainder, Table 6) were changed to correspond to the average condit ion of each of these var iables during the period in quest ion. The other parameter values required in the model are summarized in Table 5. At each time i n t e r v a l the model simulated the rate trout could capture d i f f e rent prey by searching for one hour. The predicted occurence of food organisms (expressed as a percentage) was ca lculated from these r e s u l t s . Table 7 A presents a comparison between the observed exp lo i t a t ion and an expected d i s t r i b u t i o n that i s based upon the premise that trout capture prey in d i rec t proportion to t h e i r abundance. There i s l i t t l e doubt that there i s an extremely poor c o r r e l a t i o n between the observed and expected d i s t r i b u t i o n s , e spec ia l ly for the amphipods and caddis . Therefore, at least during these months, the trout population was feeding s e l e c t i v e l y . The pattern of exp lo i t a t ion that was predicted by the s imulat ion i s shown in Tablse 7 B. In t h i s case there i s considerably better agreement between the observed and predicted occurence of prey. Part of the explanation why the simulation was more accurate in accounting for the frequency of amphipods and caddis i s that i t took into considerat ion the i r v e r t i c a l d i spersa l behaviour, t h e i r average s ize and a c t i v i t y . TABLE 7 A. A comparison between the expected and observ/ed percentage occurence of d i f ferent prey groups in trout stomachs. The expected (E) d i s t r i b u t i o n assumes that each prey was exploi ted in d i rect proportion to i t s dens i ty . (0) i s the observed d i s t r i b u t i o n of prey. Month Amphipods Caddis Odonates Planorbids 0 E 0 E 0 E 0 E Feb 7 95 47 0.8 7 0.8 38 3 May 49 90 34 5 7 5 9 8 June 70 90 10 0.6 7 0.6 14 7 Aug 38 96 16 0.1 9 0.2 39 3 Nov 20 96 51 0.4 25 0.3 4 3 a 57% a 31% a 10% a 16% Mean deviat ion from observed TABLE 7 B. A comparison between the observed (o) and predicted (P) percentage occurence of d i f f e rent prey groups in trout stomachs. The predicted d i s t r i b u t i o n was generated from the s imulation model. Month Amphipods 0 P 0 Caddis P Odonates 0 P Planorbids 0 P Feb 7 21 47 48 7 30 38 1 May 49 50 34 30 7 17 9 2 June 70 70 10 16, 7 9 14 4 Aug 38 94 16 2 9 2 39 2 2 Nov 20 22 51 56 25 20 4 2 a 14% a 6% a 9% a 18% a Mean deviat ion from observed For example, the amphipods are r e l a t i v e l y smal l , spend a great deal of time concealed and are only moderately act ive when exposed. These c h a r a c t e r i s t i c s w i l l tend to lower the i r v u l n e r a b i l i t y to at tack. On the other hand, the caddis are large , t o t a l l y exposed and tend to spend most of the time ac t ive ly moving; hence, they are more vulnerable than one would expect simply on the basis of the i r dens i ty . The s imulation did not account for the occurence of the planorbids very accurate ly . In th i s case, i t i s probable that the v e r t i c a l and hor izonta l a c t i v i t y of th i s group i s important and shouOia be examined in further d e t a i l . In add i t ion , the p o s s i b i l i t y that trout learn to s e l e c t i v e l y detect planorbids can not be overruled (Section I I ) . Nevertheless, since there i s considerably more information concerning the behaviour of Cranqonyx and Hyale l la the exp lo i ta t ion of these populations can be explored in more d e t a i l . The Size Se lect ive Exp lo i t a t ion of Amphipods There are three s p e c i f i c c h a r a c t e r i s t i c s to the pattern of exp lo i ta t ion of amphipods by the trout population in Marion Lake. The f i r s t , i s that d i f ferent s ize categories of Hya le l l a and Cranqonyx are not consumed in d i rect proportion to the i r abundance. Secondly, even though Hyale l l a i s about 7 times more numerous than Cranqonyx, the l a t t e r i s captured s l i g h t l y more f requent ly . F i n a l l y , the exp lo i ta t ion of both species changes seasonally, becoming more pronounced in the summer and f a l l i n g to a lower lev/el in the spring and la te f a l l . The question i s , can the model account for any of these observations and i f so, what are the major factors involved? In the s imulation conducted for amphipods, an 'average' year was divided into 24 two week i n t e r v a l s . At the beginning of each period the ambient water temperature (Appendix II) , the density of both species (Appendix III) and the s ize composition (considered in s ize classes of 1 mm) were changed to follow the average trend in each of these parameters. The other parameter values for the model are l i s t e d in Table 5. The pos i t ive re l a t ionsh ip between the width of the searching path and prey s ize implies that large prey should be more vulnerable to attack than smaller ind iv idua l s of the same species . In addi t ion , depending upon the average distance trout search from the sediment they may not detect some small s ize classes of amphipods. Therefore, these two c h a r a c t e r i s t i c s suggest that trout should d i sproport ionate ly attack large prey but are not l i k e l y to attack others below a c r i t i c a l s i z e . Figure 6 shows the frequency -"of d i f ferent s ize classes of Hya le l l a captured in the months of June and November. These periods were selected because the greatest number of trout stomachs were taken. In th i s f igure , the expected curves indica te the number of animals that should have been found i f each size class was exploited in proportion to i t s abundance. The observed d i s t r i b u t i o n demonstrates that large Hyale l l a were captured more frequently than one would expect and that animals less than 3 mm were not found at a l l . The predicted frequency curves were ca lculated by mul t ip ly ing the actual number of Hyale l la that were found by the ra t io each s ize c lass was predicted to be captured. In both months ( f i g . 6) there was a s i g n i f i c a n t dif ference between the observed and expected d i s t r i b u t i o n s (X>0.05) which indicates that the trout were s e l e c t i v e l y exp lo i t ing d i f ferent s ize classes of prey.. On the other hand, the s imulation predicted that Hyalel l a less than 3 mm should not have been found in trout stomachs and was able to account for the occurence of other s ize classes of prey to the extent that observed and predicted d i s t rubt ions are not s i g n i f i c a n t l y d i f f e r e n t . A s imi l a r comparison of the observed, predicted and expected s ize composition of Cranqonyx i s presented in Table 8. In th i s case, the f i t of the predicted d i s t r i b u t i o n i s not as close as i t was for Hyale l la but i s s t i l l not s i g n i f i c a n t l y d i f ferent from the observed. Although the simulation predicted that 2 mm was the smallest s ize class of Cranqonyx that trout could detect, th i s supposit ion could not be tested because during the months in which stomach samples were taken, a l l the Cranqonyx in Marion Lake were greater than 2 mm in length. In any case, i t appears as i f the re l a t ionsh ip between prey s ize and react ive distance, as well as the propensity of ure 6. A comparison of the observed ( so l id c i r c l e s expected ( so l id t r i a n g l e s ) , and the predicted (open c i r c l e s ) d i s t r i b u t i o n of d i f ferent s ize classes of Hya le l l a found in trout stomachs at two di f ferent sampling per iods . See text for further explanation TABLE 8. A comparison of the f i t between the expected (E), predicted (P) and actual (0) s ize composition of Cranqonyx found in trout stomachs in the month of November. The expected number i s based upon the assumption that trout were capturing d i f ferent s ize classes in d i rec t proportion to t h e i r respective f i e l d d e n s i t i e s . The predicted number i s based upon the resul t s of a s imulat ion . Size c lass Number Number Number (mm) (0) (P) (E) 4 0 0 0 5 2 1 1 6 2 9 11 7 10 10 11 8 12 8 7 9 9 6 6 10 2 4 3 11 4 2 1 Chi-sguared ( 0 - P ) = 7.4 Chi-squared ( 0 - E ) = 11.2 * S i g n i f i c a n t l y d i f ferent at 0.05 l e v e l trout to maintain a searching pos i t ion are s u f f i c i e n t mechanisms to account for the i r s e l ec t ive exp lo i ta t ion of large amphipods and the existence of a threshold or minimum size of prey they can detect . Seasonal Changes in the Exp lo i t a t ion of Amphipods The number of amphipods a trout could capture, at d i f ferent times of the year, by searching for one hour, i s presented in Figure 7. This s imulat ion shows that the v u l n e r a b i l i t y of both species does not r e f l ec t changes simply in t h e i r seasonal abundance. For example, Cranqonyx reaches a peak density at the end of June (Appendix III ) , and yet i s most vulnerable to attack early in September. The same i s true for H y a l e l l a , i t i s most abundant at the end of August, but apparently i s ""more vulnerable to attack approximately one month l a t e r . Moreover, although Crangonyx i s always less numerous than Hya le l l a , the s imulat ion suggests that both species are just as vulnerable to capture throughout most of the year, except in the la te part of the summer (July and August). At th i s time, Crangonyx i s considerably more suscept ib le . For each month in which samples were ava i l ab le , the stomach contents of the trout were analysed with respect to the average number of Cranqonyx and Hya le l l a that were found in the gut. These data are superimposed over the simulated v u l n e r a b i l i t y curves in Figure 7. With the exception of the A p r i l sample, the observed trend in exp lo i ta t ion c lose ly follows the simulated Figure 7. A comparison of the simulated (curve) and observed (data points) trend in the exp lo i t a t ion of Cranqonyx and H y a l e l l a , by t rout , in Marion Lake. N u m b e r c a p t u r e s per h o u r o c_ i — r ZT C_ > CO o cn JL o JL O J o 1 • i i 1 I i O —* EL CD 13 CD CQ o ID < I—i—i—i— i—r~ No. a m p h i p o d s in t r o u t s t o m a c h s oe [ t rend . A comparison of these resul t s also shows that Cranqonyx was predicted to be just as vulnerable to capture as Hya le l l a which, with one exception, did in fact occur. Therefore on the basis of these data i t seems as i f the s imulation model can account for the seasonal pattern of exp lo i t a t ion as well as the fact that Cranqonyx i s captured just as frequently as H y a l e l l a . Ev ident ly , the c h a r a c t e r i s t i c s of prey, namely, t h e i r s ize and r e l a t i v e a c t i v i t y contribute s i g n i f i c a n t l y to determining the r i sk of predat ion. These c h a r a c t e r i s t i c s can apparently be important enough to override considerable di f ferences in dens i ty . DISCUSSION One of the advantages of developing models of b i o l o g i c a l processes such as predation i s that one can examine the apparent importance of d i f ferent components. Any re su l t ing inferences can then be expressed as hypotheses and tested e i ther experimentally or in the f i e l d . For example, the distance trout can detect prey of a spec i f i c s ize and inherent contrast i s dependent to some extent upon the t u r b i d i t y of the water ( 6 ) . If the ext inct ion coe f f i c ient of the water were increased then the react ive distance would diminish , simply because trout would have to be closer to prey in order to detect the contrast threshold for d i s c r i m i n a t i o n . Subseguently, a l l else being equal, in murky waters v i sua l predators w i l l attack prey at a slower rate than they could in c learer water. The question which ar i ses however, is just how s i g n i f i c a n t is the c l a r i t y of the environment to predation? The simulation model was able to account for the exp lo i t a t ion of amphipods as well as several other prey, therefore, i t may be a reasonable abstract ion of the feeding behaviour of t rout . If th i s i s the case then i t i s worthwhile to examine the s e n s i t i v i t y of the model to some of i t s components. Table 9 presents the change in the v u l n e r a b i l i t y of Crangonyx and Hya le l l a that resulted when several components were a l t e r e d . In conducting these s imulat ions , I a r b i t r a r i l y increased the value of each parameter by 10% of the estimated ' r e a l ' value (Table 5) . The dif ference in the rate of capture each of these attendant changes produced indicates the s e n s i t i v i t y of predation to these parameters. The components of the predation model can be par t i t ioned into 3 categories : l ) environmental c h a r a c t e r i s t i c s , 2) prey c h a r a c t e r i s t i c s , and 3) predator c h a r a c t e r i s t i c s . Environmental Charac te r i s t i c s To answer the guestion raised e a r l i e r , the v u l n e r a b i l i t y of Cranqonyx and Hya le l l a i s r e l a t i v e l y in sens i t i ve to the t u r b i d i t y of the water; a change in (6) was not as important as an increase in e i ther the ambient i l lumina t ion or temperature (Table 9) . Despite the fact that the searching a c t i v i t y of a v i sua l predator w i l l be r e s t r i c t e d by the l e v e l of i l l u m i n a t i o n , even at l eve l s above the l i m i t of v i s ion (R . ) the i l luminat ion could TABLE 9. S e n s i t i v i t y of the attack model to selected parameters. Each parameter uuas increased by 10% of i t s ' r e a l * value (Table 5 ) . In the case of the learning component, the width of the path of search was doubled. The period of s imulat ion was the January 1-15 i n t e r v a l . Parameter % change in the attack rate Cranqonyx Hyale l la A. Environmental Charac te r i s t i c s Temperature (T) 22 33 I l luminat ion (R) 22 20 Turb id i ty ^ ,(6) . - 2 - 5 B. Prey Charac ter i s t i c s Inherent Contrast (C ) 43 128 o V e r t i c a l A c t i v i t y (IYI3) 22 24 Horizontal a c t i v i t y (M7) 17 14 C. Predator Charac te r i s t i c s Learning 50 50 Recognition Success (RS) 22 24 i n d i r e c t l y influence the searching pattern an animal adopts. Hamilton (1965) and Hyatt (personal communication) found that in the l a te summer (August), the trout population in Marion Lake, d isplays a d i s t i n c t d i e l feeding pat tern . In the early morning (0500 to 0900 hrs) most animals appear to be feeding predominantly in the water column on chironomid pupae. Throughout the rest of the day, however, some f i sh w i l l sh i f t t h e i r pos i t ion and hunt sporadica l ly for benthic prey. By the l a te afternoon (2000 hrs) most of the population w i l l have shi f ted back to feeding in the water column u n t i l 2300 or 2400 hrs . After th i s time the population may become r e l a t i v e l y inac t ive since there i s a general decl ine in the amount of food trout contain in t h e i r stomachs. The foraging a c t i v i t y , of t rout , in the water column c lose ly coincides with the d i e l migration of chironomid pupae into th i s region. L i t t l e migration occurs during the day (Hamilton, 1965). As a re su l t , the general sh i f t in the feeding pos i t ion of the trout population could be in response to changes in the a v a i l a b i l i t y of food in d i f ferent sectors of the environment. During periods of low i l lumina t ion (early morning and late evening) trout w i l l not detect prey on the lake bottom as e f f e c t i v e l y as they can in the water column. An object contrasted against the evening sky w i l l present a higher l e v e l of contrast than a s imi l a r target on the sediment (Hester and Taylor , 1965). Therefore, trout should be able to feed on l imnet ic prey at lower l eve l s of i l l u m i n a t i o n than they require to hunt for b e n t h i c - l i v i n g animals. As the lev/el of i l l umina t ion r i se s , however, and the migration of chironomids begins to subside, the population may be able to feed more e f f i c i e n t l y on the lake bottom; in which case, t h e i r a t tent ion could be shi f ted to th i s region. This supposit ion was suggested by e a r l i e r experiments (Section I) which demonstrated that trout w i l l not maintain a searching pattern unless they are reinforced above a c r i t i c a l rate (about 2 captures per minute). Changes in the ambient temperature can also be expected to influence the v u l n e r a b i l i t y of prey such as amphipods. In t h i s case, r i s i n g temperatures w i l l increase the number of animals that are exposed as well as the l e v e l of a c t i v i t y of exposed ind iv idua l s (Section I I I ) . Therefore trout w i l l be more successful in recognizing these prey and w i l l attack from a greater d is tance . The major role that water temperature plays in the in terac t ion between the trout and the amphipods in Marion Lake i s apparent in Figure 7. The v u l n e r a b i l i t y of both Cranqonyx and Hya le l l a i s c lose ly corre lated with the seasonal temperature pattern (Appendix I I ) . In general , seasonal and d i e l changes in the water temperature and the ambient i l lumina t ion may well be two of the most important factors that affect the a c t i v i t y of aquatic animals (Fry, 1947; Thome, 1969) and subsequently the i r v u l n e r a b i l i t y to predat ion . Prey C h a r a c t e r i s t i c s The s e n s i t i v i t y a n a l y s i s (Table 9) i n d i c a t e s that s e v e r a l p h y s i c a l and b e h a v i o u r a l c h a r a c t e r i s t i c s of prey w i l l a l s o a f f e c t the r a t e of c a p t u r e . H y a l e l l a and Cranqonyx spend much of the time concealed e i t h e r i n the sediment or under o b j e c t s at the mud-water i n t e r f a c e . In t h i s r e s p e c t , the behaviour of these s p e c i e s i s s i m i l a r to other a q u a t i c i n v e r t e b r a t e s (Berglund, 1968; S t r a s k r a b a , 1965). As a r e s u l t at any i n s t a n t i n time only a small p r o p o r t i o n of prey w i l l be s u s c e p t i b l e to a t t a c k from v i s u a l p r e d a t o r s . Several authors ( A l l e n , 1941; Huruska, 1961; Grimas, 1963) have suggested that concealment behaviour would g r e a t l y d i m i n i s h the r i s k of a s p e c i e s to a t t a c k . T h i s i n f e r e n c e i s supported by the experimental r e s u l t s and the s i m u l a t i o n . Prey v u l n e r a b i l i t y w i l l vary s e a s o n a l l y , however, and perhaps even during the day depending upon a number of other f a c t o r s such as l ) a l t e r a t i o n s i n a c t i v i t y p a t t e r n s , 2) changes i n p o p u l a t i o r d e n s i t y and, 3) changes i n the s i z e composition of the p o p u l a t i o n . The e f f e c t of a sudden s h i f t i n the s i z e s t r u c t u r e of a prey p o p u l a t i o n on the r a t e of p r e d a t i o n i s i l l u s t r a t e d i n F i g u r e 7. In t h i s example, the simulated v u l n e r a b i l i t y curves f o r Cranqonyx and H y a l e l l a begin to r i s e i n March and A p r i l i n response to the i n c r e a s e i n the ambient lake temperature. T h i s r i s e continues u n t i l the l a t t e r part of May when the v u l n e r a b i l i t y of Cranqonyx suddenly drops. The same phenomenon i s apparent f o r H y a l e l l a except that i t occurs l a t t e r i n June. In both cases the d e c l i n e in the rate of capture i s due to reproduction and the appearance of juveni les in the populat ion. Although the v u l n e r a b i l i t y of Cranqonyx was not depressed for very long, the v u l n e r a b i l i t y of Hya le l l a did not begin to r i se again u n t i l about one month after the onset of breeding. In addit ion to the importance of concealment behaviour and s ize , the inherent contrast of a prey w i l l also affect i t s r i sk of being attacked. In the s e n s i t i v i t y analys i s , a 10$ increase in contrast markedly raised the v u l n e r a b i l i t y of Cranqonyx and H y a l e l l a , but e spec ia l ly the l a t t e r . Although i t i s well known that c ryp t i c animals are less susceptible to being discovered by v i sua l predators than more conspicuous species, the s e n s i t i v i t y of the model to contrast demonstrates just how important th i s component could be to predat ion. In the s e n s i t i v i t y s imulat ion, Hya le l l a proved to be more responsive to a change in almost every parameter that was inves t iga ted . Since most of these a l t e ra t ions changed the react ive distance component, th i s would tend to have a s i g n i f i c a n t effect on a small animal due to the form of the re l a t ionsh ip between react ive distance and prey s ize ( f i g . 2) . In addition, a change in the distance of react ion w i l l become considerably amplif ied because i t w i l l affect the rate trout encounter every s ize c lass of prey. Therefore, seemingly minor changes in some major components can have a s i gn i f i c an t effect on prey r i s k , e spec ia l ly in the case of a small animal. Predator Charac ter i s t i c s One of the most important c h a r a c t e r i s t i c s of the feeding behaviour of trout i s the fact that they maintain a pos i t ion some distance from the sediment when they hunt for food. It was mentioned e a r l i e r , that i f the height of the searching pos i t ion ever, exceeds the distance of react ion then trout w i l l f a i l to discover some small s ize classes of prey. This charac te r i s t i c ; was s u f f i c i e n t to explain why Hyalel1 a less than 3 mm were not captured ( f i g . 6) . In general , the propensity of trout to maintain a searching pos i t ion w i l l tend to favour a small animal,such as H y a l e l l a , because throughout the year there w i l l always be animals in the population that are less than the threshold s i z e . In contrast , larger prey l i k e Crangonyx, wi I I have s ize classes invulnerable to predation only during periods of reproduction. Under cer ta in condit ions trout can increase the i r responsiveness to prey (Section I I ) . The attack model i s formulated on the basis that the predator i t simulates was conditioned to recognize amphipods. Since an animal that i s not conditioned w i l l react from hal f the distance that a conditioned trout w i l l , the effect of learning can be simulated by doubling the width of the path of search. Obviously, i f a predator can increase the area about i t in which i t w i l l respond to p r e y , i t w i l l be able to discover food at a subs tant ia l ly greater rate than one that i s less responsive (Table 9) . Several experimental studies have v e r i f i e d that learning i s an extremely advantageous process through which predators could i n c r e a s e t h e i r r a t e of energy i n t a k e (Beukema, 1968; Croze, 1970). Several c h a r a c t e r i s t i c s of the behaviour of t r o u t , namely, l ) t h e i r v i s u a l response to prey, 2) the dependence of t h e i r p a t t e r n of search on a t h r e s h o l d r a t e of capture, and 3) the f a c t that they can l e a r n to a l t e r t h e i r response to prey, imply that they are admirably adapted to forage i n d i f f e r e n t environments, For example, they can feed on e i t h e r l i m n e t i c or b e n t h i c l i v i n g organisms as well as on the d r i f t i n streams. However, the f a c t that they maintain a sea r c h i n g p o s i t i o n when they hunt over a s u b s t r a t e and respond to prey v i s u a l l y suggests that they are b e t t e r adapted to f e e d i n g i n the water column. Since water e s s e n t i a l l y presents a 'smooth' bafekground, t r o u t w i l l maximize the area of t h e i r r e a c t i v e f i e l d and w i l l be most s u c c e s s f u l i n r e c o g n i z i n g prey when they hunt i n t h i s s e c t o r of the environment. If they feed on b e n t h i c - l i v i n g prey then t h e i r hunting e f f i c i e n c y w i l l be somewhat diminished, l a r g e l y because the background i n t h i s case w i l l be r e l a t i v e l y more 'broken'. Thus they w i l l be l e s s s u c c e s s f u l i n r e c o g n i z i n g food and must be c l o s e r before they a t t a c k . Moreover, s i n c e they maintain a s e a r c h i n g p o s i t i o n they may not d i s c o v e r some small s i z e c l a s s e s of prey. In any case, whether t r o u t feed i n the water column or over a s u b s t r a t e , they are l i k e l y to converge t e m p o r a r i l y i n t o areas i n which food i s r e l a t i v e l y more abundant because t h e i r p a t t e r n of search i s dependent upon a t h r e s h o l d r a t e of capture ( S e c t i o n I ) . In a d d i t i o n , the c h a r a c t e r i s t i c of l e a r n i n g w i l l enable i n d i v i d u a l s to increase t h e i r responsiveness to prey and thereby f u r t h e r improve t h e i r hunting e f f i c i e n c y . In conclusion, the predic t ions generated by the s imulation model do not refute the hypothesis that the se lec t ive exp lo i ta t ion of Cranqonyx and Hyalel1 a, by the trout population in Marion Lake, can be explained by the process of prey recognit ion and de tec t ion . This hypothesis was expressed in the attack model and appears to be s u f f i c i e n t to account for the disproport ionate exp lo i t a t ion of d i f fe rent s ize classes of prey as well as the observation that Cranqonyx i s just as vulnerable to attack as Hyal el 1 a despite a 7 fo ld ' .difference in t h e i r dens i t ies . There are two p r i n c i p a l reasons why Cranqonyx i s imore vulnerable to at tack; in the f i r s t place, i t i s a l arger animal and secondly i t spends considerably more time moving when exposed. Therefore i t w i l l be recognized more success ful ly and attacked from a greater distance than H y a l e l l a . The attack model was also able to rep l i ca te the seasonal pattern in the v u l n e r a b i l i t y of amphipods. The explanation in this- case i s pr imar i ly due to the seasonal change in water temperature. Very few amphipods are exposed and moving over the sediment in the winter months. However, as the water temperature r i ses more animals w i l l be exposed and ; considerably more a c t i v e . Therefore the v u l n e r a b i l i t y of both populations w i l l r i s e since trout w i l l be more successful in recognizing these prey and w i l l discover them from a greater d i s tance . This study, therefore , demonstrates that due to the in te rac t ion between the behaviour of a v i sua l predator and the c h a r a c t e r i s t i c s of i t s prey, predation i s not only affiected by the density of a food organism, but also w i l l be influenced by i t s s i ze , a c t i v i t y and contras t . ' These factors are l i k e l y to be just as i f not considerably more important than densi ty . SUMMARY 1. In order to test the hypothesis that the process of prey detection and recognit ion i s s u f f i c i e n t to explain the se l ec t ive exp lo i t a t ion of prey by t rout , a study was conducted to ident i fy some of the factors that affect the i r v i sua l d i s c r i m i n a t i o n . 2. The distance trout w i l l react was found to be non- l inear ly related to prey s i z e . A general system of equation was developed to describe the process of visual) d i scr iminat ion in terms of the re l a t ionsh ip between the s ize and the apparent contrast a target must have before i t can be detected (attacked). 3. The effect of the background i l l u m i n a t i o n on contrast d i scr iminat ion was not examined. This component was considered, however, on the basis of resu l t s documented in the l i t e r a t u r e . 4. Rainbow trout w i l l react to moving targets from a greater distance than s tat ionary prey. Irrespect ive of the s ize of the t e s t prey, the e f f e c t of t a r g e t motion uuas constant (22 cm). 5. A 'broken' background reduced the a b i l i t y of t r o u t to recognize prey and s i g n i f i c a n t l y d i m i n i s h e d t h e i r r e a c t i v e d i s t a n c e f o r both moving and non-moving t a r g e t s . This e f f e c t was i n c o r p o r a t e d by assuming that the experimental c o n d i t i o n s simulated the d i v e r s i t y of the n a t u r a l l a k e s u b s t r a t e . 6. A general a t t a c k model was developed to simulate the e x p l o i t a t of s e v e r a l i n v e r t e b r a t e groups (odonates, caddis and p l a n o r b i d s ) but p r i n c i p a l l y the amphipods, by the t r o u t p o p u l a t i o n i n Marion Lake. 7. Although the proposed model was reasonably a c c u r a t e i n p r e d i c t i n g the percentage occurence of the amphipods, caddis and odonates i n t r o u t stomachs, i t was unable to account f o r the high occurence of p l a n o r b i d s i n some months. Therefore, i t does not o f f e r a complete e x p l a n a t i o n . 8. An a d d i t i o n a l s i m u l a t i o n adequately p r e d i c t e d the 3 c h a r a c t e r i s t i c s of t r o u t p r e d a t i o n on Cranqonyx and H y a l e l l a ; i ) t h e i r d i s p r o p o r t i o n a t e e x p l o i t a t i o n of d i f f e r e n t s i z e c l a s s e s of amphipods (the model was a l s o able to account f o r the t h r e s h o l d s i z e (3mm) or the s m a l l e s t H y a l e l l a consumed by t r o u t ) , i i ) the f a c t that Cranqonyx i s captured as f r e q u e n t l y as H y a l e l l a even though i t i s 7 times l e s s abundant i i i ) the seasonal pattern to the exp lo i ta t ion of both species . 9. In general , the resul t s of th i s study lead to the conclusion that the se lec t ive pressure exerted by v i sua l predators w i l l tend to favour smal l , c r y p t i c a l l y colored animals, a general reduction in a c t i v i t y and concealment behaviour. BIBLIOGRAPHY A l l , M.A. 1959. The ocular s t ructure , retinomotor and photo-behavioural responses of juveni le p a c i f i c salmon. Can. J . Z o o l . 37:965-996. A l l e n , K.R. 1941. Studies on the biology of the early stages of the salmon (Salmo s a l a r ) . 2. Feeding habi t s . J . Animal E c o l . 10:47-76. Berglung, T. 196B. The influence of predation by brown trout on Asel lus in a pond. Rep. Inst . Freshw. Res. Drottningholm 48:76-101. Beukema, J . J . 1968. Predation by the three-spined st ickleback (Gasterosteus aculeatus L.) : The influence of hunger and experience. Behaviour 31:1-126. Blaxter , J . H . S . and M.P. Jones.. 1967. The development of the re t ina and retinomotor responses in the herr ing . J . Mar. B i o l . Ass. U.K. 47:677-697. Blaxter , J . H . S . 1968a. Light i n t e n s i t y , v i s i o n , and feeding in young p l a i c e . J . Exp. Mar. B i o l . E c o l . 2:293-307. 1968b. Visua l thresholds and spectral s e n s i t i v i t y of herring l a rvae . J . Exp. B i o l . 48:39-53. 1969. Visual thresholds and spectra l s e n s i t i v i t y of f l a t f i s h l a rvae . J . Exp. B i o l . 51:221-230. Braum, E. 1967. The s u r v i v a l of f i s h l a r v a e with r e f e r e n c e to t h e i r f e e d i n g behaviour and the food supply, p 113-131. _Iri S.D. Gerking (ed.) The b i o l o g i c a l b a s i s of freshwater f i s h p r o d u c t i o n . Blackwell S c i e n t i f i c P u b l i c a t i o n s , Oxford. Brawn, V.M. 1969. Feeding behaviour of cod (Gadus morhaa). J. F i s h . Res. Bd. Canada 26:583-596. Brooks, J.L. 1968. The e f f e c t s of prey s i z e s e l e c t i o n by lake p l a n k t i v o r e s . S y s t . Z o o l . 17:272-291. Croze, H. 1970. Searching image i n c a r r i o n crows. Z e i t s c h r i f t f u r t i e r p s y c h o l o g i e , supp. 5, 86 p. Dickman, M.D. MS 1968. The r e l a t i o n of freshwater plankton p r o d u c t i v i t y to s p e c i e s composition during induced s u c c e s s i o n . Ph.D. t h e s i s . The U n i v e r s i t y of B r i t i s h Columbia, Vancouver B.C. 115 p. Dixon, A.F. 1959. An experimental study of the s e a r c h i n g behaviour of the predatory C o c c i n e l l i d b e e t l e A d a l i a decempunctata ( L . ) . J . Animal E c o l . 28: 259-281. Duntley, S.Q. 1963. L i g h t i n the sea. J . opt. Soc. Am. 53: 214-233. E f f o r d , I.E. 1967. Temporal and s p a t i a l d i f f e r e n c e s i n phytoplankton p r o d u c t i v i t y i n Marion Lake, B r i t i s h Columbia. J . F i s h . Res. Bd. Canada 24: 2283-2307. Fry, F . E . 1947. Effects of the environment on animal a c t i v i t y . Univ. Toronto Studies, B i o l . Ser. 55, Publ . Ont. F i s h . Res. Lab. 68: 5-62. Grimas, U. 1963. Ref lect ions on the a v a i l a b i l i t y and u t i l i z a t i o n degree of bottom animals as f i s h food. Zool . B idr . Uppsala. 35: 497-503. Grand, Y. l e , 1967. Form and space v i s i o n . Indiana Univers i ty Press. Bloomington. 367 p. Hamilton, A . L . (VIS 1965. An analys is of a freshwater benthic community with specia l reference to the Chironomidae. Ph. D. thes i s , the Univers i ty of B r i t i s h Columbia, Vancouver, B. C. 74 p. Hargrave , B. T. 1969c. Epibenthic a lga l production and community re sp i ra t ion in the sediments of Marion Lake. J . F i s h . Res. Bd. Canada, 26: 2003-2026 1970. The u t i l i z a t i o n of benthic microf lora by Hyale l la azteca (Amphipoda). J . Animal E c o l . 39: 427-437. F. and J . Taylor , 1965. How tuna see a net. Commercial F i sher ie s Review. 27(3): 11-16. F . J . 1968. Visua l contrast thresholds of the go ldf i sh (Carassius auratus) . V i s ion Res. 8: 1315-1336. Hester, Hester, H o l l i n g , C.S . 1959a. The components of predation as revealed by a study of small-mammal predation of the European pine sawfly. Can. Ent. 91 : 293-320 . 1959b. Some c h a r a c t e r i s t i c s of simple types of predation and paras i t i sm. Can. Ent . 91: 385-398. 1964. The analys i s of complex population processes. Can. Ent. 96: 335-347. 1965. The functional response of predators to prey density and i t s role in mimicry and population regu la t ion . Mem. Ent . Soc. Can. 45: 1-60. 1966. The funct ional response of invertebrate predators to prey dens i ty . Mem. Ent. Soc. Can. 48: 1-86. Horridge, G.A. 1968. Interneurons, t h e i r o r i g i n , ac t ion , s p e c i f i c i t y , grouuth and p l a s t i c i t y . UI.H. Freeman and Co. , San Franci sco . 436 p. Hunter, J .R. 1968. Ef fects of l i gh t on schooling and feeding of Jack Mackerel. J . F i s h . Res. Bd. Canada 25: 393-407. Huruska, V. 1961. An attempt at a d i rec t inves t iga t ion of the influence of the carp stock on the bottom fauna of two ponds. Verh. Internat. Vere in . Limnol. 14: 732-736. Ivlev , U.S. 1961. Experimental ecology of the feeding of f i shes . Yale Univers i ty Press, New Haven. Ishiwata, N. 1968b. Eco log ica l studies of the feeding of f i shes . I I I . Degree of hunger and s a t i a t i o n amount. B u l l . Jap. Soc. S c i . F i s h . 34: 604-607. 1968e. Eco log ica l studies on the feeding of f i shes . UI. External factors a f fect ing s a t i a t i o n amount. B u l l . Jap. Soc. S c i . F i s h . 34: 785-791. Jenkins, T.IY1. 1969. Socia l s t ructure , pos i t ion choice and microd i s t r ibut ion of two trout species (Salmo t rut ta and Salmo qa i rdner i ) resident in mountain streams. Animal Behav. Monog. 2: 57-123. Johannes, R .E . and P .A . L a r k i n . 1961. Competition for food between Redside shiners (Richardsonius balteatus) and Rainbow trout (Salmo qa i rdner i ) in two B r i t i s h Columbia l akes . J . F i s h . Res. Bd. Canada 18 : 203-220. Lee, E .B . MS 1967. Some aspects of the energy budget of the family Planorbidae in Marion Lake, B.C. BSc. Thes i s . The Univers i ty of B r i t i s h Columbia, Uancouver, B.C. 37p. Lindstrom, T. 1955. On the re l a t ion f i sh s ize - food s i z e . Rep. Inst . Freshw. Res. Drottingholm 36: 133-147. Macan, T . T . 1966. The influence of predation on the bottom fauna of a moorland f i s h pond. Arch. Hydrobiol . 61: 432-452. Mathias, J . A . MS 1967. Population energetics of two amphipod species in Marion Lake. M.Sc. t h e s i s . The Univers i ty of B r i t i s h Columbia, Vancouver, B.C. 74p. Messenger, J .B . 1968. The v i sua l attack of the c u t t l e f i s h , Sepia o f f i c i n a l i s . Anim. Behav. 16: 324-357. M o r r e l l , G.M. and J .R. Turner. 1970. Experiments in mimicry: I. The response of wild birds to a r t i f i c i a l prey. Behaviour 36: 116-130. Neish, I. 1970. A comparative analys i s of the feeding behaviour of two salamander populations in Marion Lake, B.C. Ph. D thes i s , the Univers i ty of B r i t i s h Columbia, Vancouver, B.C. 108 p. Nakamura, E . L . 1968. Visual acuity of two tuna Katsuwonus pel amis and Euthynnus a f f i n s . Copeia: 41-49. Parsons, T .R . , R . J . l e Brasseur, J .D . Ful ton , 1967. Some observations on the dependence of zooplankton grazing on the c e l l s ize and concentration of phytoplankton blooms. J . Oceanog. Soc. Japan. 23: 10-17. Polyak, S .L . 1957. The vertebrate v i sua l system. The Univers i ty of Chicago Press, Chicago. Prazdnikova, N.V. 1969. Pecu la r i t i e s of the d i c t i n c t i o n of v i sua l images by f i s h , from Behaviour and reception in f i sh (1967), t rans . R.M. Howland, Bureau of Sports F i sher ie s and W i l d l i f e , Washington, D.C. Pr i t chard , G. 1965. Prey capture by dragonfly larvae (Odonata: Anisoptera) . Can. J . Zoo l . 43: 271-289. Prop, N. 1960. Protect ion against birds and parasites in some species of Tenthredinid l a rvae . Arch, n e e r l . Zoo l . 13: 380-447. R i l l i n g , S., H. Mitte ls taedt and K.D. Roeder. 1959. Prey recognit ion in the preying mantis. Behaviour 14: 164-184. Ruch, T . C . and H.D. Patten. 1965. Physiology and b iophys ics . W.B. Saunders Co. Ph i l ade lph ia . Ruiter , L . de. 1952. Some experiments on the camouflage of s t i ck c a t e r p i l l a r s . Behaviour 4: 222-232. 1966. Feeding behaviour of vertebrates in the natural environment. _I_n Handbook of physiology (Am. P h y s i o l . Soc.) Section 6 ( l ) : 96-116. Schutz, D.C. MS 1969. An experimental study of the feeding behaviour and in te rac t ion of coastal cutthroat trout (Salmo c l a r k i c l a r k i ) and do l ly vardetn (Salvel inus malma). M.Sc. Thesis , the Univers i ty of B r i t i s h Columbia, Vancouver,B. C. Sheperd, B.G. MS 1970. Aspects of the feeding behaviour of the rainbow trout (Salmo qa irdner i ) and kokanee (Onchoryncus nerka) of Marion Lake, B r i t i s h Columbia. B.Sc. Thesis , the Univers i ty of B r i t i s h Columbia, Vancouver, B.C. S h i f f r i n , R.M. 1970. Forge t t ing : trace erosion or r e t r i e v a l f a i lure? Science 168: 1601-1603. Smith, IYI.IAI. 1961. Bottom fauna in a f e r t i l i z e d natural lake and i t s u t i l i z a t i o n by trout (Salvel inus f o n t i n a l i s ) as food. Verh. Int. Vere in . Limnol . 14: 722-726. Straskraba, M. 1965. The effect of f i s h on the number of invertebrates in ponds and streams. M i t t . Int. Vere in . Lminol . 13: 106-127. Sverdrup, H.W., M.w1. Johnson, and R.H. Fleming. 1942. The oceans. P r e n t i c e - H a l l , Englewood c l i f f s . 1087 p. Tamura, T. 1957. A study of v i sua l perception in f i s h , e spec ia l ly on resolving power and accomodation. B u l l . Jap. Soc. S c i . F i s h . 22: 536-557. Thome, M. J . 1969. Behaviour of the caddis f ly larva Potamphylax s t e l l a t u s (Curt i s ) (Trichoptera) . Proc. R. Ent. Soc. Lend.. 44: 91-110. Tinbergen, L. 1960. The dynamics of insect and b i rd populations in pine woods. Arch. neer. Z o o l . 13: 259-472. Yamanouchi, T. 1956. The v i sua l acuity of the coral f i s h , Microcanthus s t r iqatus (Cuvier and Valenciennes) . Pubis. Seto. marine B i o l . Lab. 5:--133-156. Appendix I A LIST OF SYMBOLS FOR THE ATTACK MODEL o CS CT CT D E TP MC PA R min min R max R RA RC RE RD "RS S Inherent contrast of prey Success of the predator in capturing prey i t has attacked. The apparent contrast of a prey that i s required to e l i c i t an attack The minimum l e v e l of contrast the predator can discr iminate Prey density ( no. / sq. m. ) A p ropor t iona l i ty constant descr ibing the effect of the background on react ive dis tance . The width of the encounter path swept along the sediment (m) The motion constant, the addi t ive effect of target movement on the attack distance (m) The proportion of exposed prey moving at any instant in time The l i m i t of scotopic v i s ion ( f t-candles ) The lowest l e v e l of i l l umina t ion which produces maximum visual acuity ( f t-candles ) The l e v e l of ambient i l lumina t ion ( f t-candles ) The rate of attack ( n o . / hr . ) The rate of capture ( n o . / hr . ) The rate of encounter ( n o . / hr . ) The react ive or attack distance, also the radius of the react ive f i e l d (m) Prey recognit ion success The v i sua l angle a prey subtends with the predator ( min. of arc ) The rate of attenuation of target contrast , (the ext inct ion coe f f i c i ent of the water) SP - The distance of the predator from the sediment, i t s searching pos i t ion (m) T - The ambient water temperature ( C ) VD - The vulnerable density of prey (no . / sq. m. ) UP - The proportion of prey exposed at or above the mud-water inter face at any instant in time VY - The predator ' s average searching v e l o c i t y (m/ hr . ) THE PHYSICAL CHARACTERISTICS OF MARION LAKE Efford (1967) and Hargrav/e. (1969c) presented a deta i led descr ip t ion of the basic physical and chemical c h a r a c t e r i s t i c s of Marion Lake. The descr ip t ion which follows w i l l be confined to some of the more important features reported by these authors. Marion Lake i s a smal l , shallow coastal lake with a mean depth of 2.4 meters. The primary production in the water column i s extremely low throughout the year. This i s undoubtly due to the pers is tent f lushing of the lake (Dickman, 1968). As a r e s u l t , the tu rb id i ty of the water (ext inct ion c o e f f i c i e n t , 0.91; H a r g r a v e 1 9 6 9 c ) i s cons i s tent ly high throughout the year due to the amount of par t i cu la te matter that i s washed into the lake and not. because of changes in the abundance of phytopiankton . Although the lake becomes thermally s t r a t i f i e d , a l l regions are subject to warming in the spring and summer. Despite f luctuat ions in the rate of temperature change from year to year the basic pattern i s the same ( f i g . 1) . The sediment in Marion Lake can be b a s i c a l l y characterized as an extremely f loculent ooze (gy t t j a ) . The substrate surface however, i s very diverse with respect to the s ize and shape of the par t i cu la te matter ( s t ick l i t t e r , chironomid l a r v a l cases, and other mater ia l ) . ure 1. The average seasonal in Marion Lake , recorded over region of the lake less than trend in water temperature 5 years (1963-1968) in the 3 meters. 20"= MONTH APPENDIX III THE POPULATION CHARACTERISTICS OF CRANGONYX AND HYALELLA IN MARION LAKE Several studies have been conducted on the amphipods in Marion Lake for a considerable number of years (Hamilton, 1965; Mathias, 1967; Bryan, unpublished data) . These data indicate that Cranqonyx i s , on the average, about 7 times less abundant than Hyale l l a and tends to be f a i r l y evenly d i s t r ibuted throughout the lake . Hya le l l a , on the other hand, i s concentrated in the shallow l i t t o r a l zone. Crangonyx, the larger of the two species, general ly produces a brood of young in the summer about one month before H y a l e l l a . The timing of the reproduction of both species as well as the i r r e l a t i v e f i e l d dens i t ies tend to be very s imi l a r from year to year. Therefore the data from a l l the ava i lab le sources were pooled to e s tab l i sh a general descr ipt ion of the density ( f i g . 2) and the s ize structure (Table l ) of each population over an 'average' year. Figure 2. The relative density of Hyalella (A) and Cranqonyx (B) in Marion Lake. N U M B E R O F A M P H I P O D S — P E R - S Q U A R E M E T E R o o CD o o o o o o o CO o o ro ro o o ro 0) o o TABLE 1. Temporal changes i n the s i z e s t r u c t u r e of amphipods. i n Marion Lake. MONTH T o t a l Body Length (mm) H y a l e l l a Cranqonyx L X H L X H Jan 2.0 4.2 5.5 5.0 7.5 11.0 Feb 2.0 4.2 6.0 6.0 9.4 12.0 Mar 2.2 4.3 6.0 6.5 9.2 12.0 Apr 2.5 4.5 6.0 7.0 9.1 12.0 May 2.5 4.5 6.5 1.0 3.7 11.0 Jun 3.5 4.8 6.0 2.0 4.1 8.0 Ju l 1.5 2.5 6.0 2.0 4.5 9.0 Aug 1 .5 3.1 6.0 3.0 5.7 9.0 Sep 1.5 3.7 6.5 4.0 6.3 10.0 Oct 2.0 4.1 6.5 5.0 7.3 10.0 Nov 2.0 4.1 6.0 5.0 8.1 11.0 Dec 2.0 4.2 6.0 5.0 7.6 12.0 L = lower l i m i t to s i z e range X = mean s i z e H = upper l i m i t to s i z e range