Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Foraging behaviour of the intertidal beetle Thinopinus pictus (Staphylinidae) Richards, Laura Jean 1982

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

Item Metadata

Download

Media
831-UBC_1982_A1 R53.pdf [ 5.13MB ]
Metadata
JSON: 831-1.0095623.json
JSON-LD: 831-1.0095623-ld.json
RDF/XML (Pretty): 831-1.0095623-rdf.xml
RDF/JSON: 831-1.0095623-rdf.json
Turtle: 831-1.0095623-turtle.txt
N-Triples: 831-1.0095623-rdf-ntriples.txt
Original Record: 831-1.0095623-source.json
Full Text
831-1.0095623-fulltext.txt
Citation
831-1.0095623.ris

Full Text

FORAGING BEHAVIOUR OF THE INTERTIDAL BEETLE THINOPINUS PICTUS (STAPHYLINIDAE) by LAURA JEAN RICHARDS B.Sc(Hons.), Dalhousie University, 1976 M.Sc, The University of B r i t i s h Columbia, 1979 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Zoology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA May 1982 (c) Laura Jean Richards, 1982 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements fo r an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. I t i s understood that copying or publication of t h i s thesis f o r f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of The University of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date DE-6 (3/81) i i ABSTRACT O p t i m a l f o r a g i n g m o d e l s g e n e r a l l y a s s u m e t h a t p r e d a t o r s a r e c a p a b l e o f m a k i n g a p p r o p r i a t e f o r a g i n g d e c i s i o n s a n d t h a t t h e s e d e c i s i o n s a f f e c t f i t n e s s . I t e s t e d t h e s e a s s u m p t i o n s i n a s t u d y o f t h e i n t e r t i d a l b e e t l e T h i n o p i n u s p i c t u s L e c o n t e ( S t a p h l y i n i d a e ) . A d u l t b e e t l e s l i v e on s a n d b e a c h e s i n t e m p o r a r y b u r r o w s f r o m w h i c h t h e y e m e r g e a t n i g h t t o p r e y on a m p h i p o d s O r c h e s t o i d e a c a l i f o r n i a n a ( B r a n d t ) . I a l s o p r e s e n t some d a t a f o r i s o p o d s A l l o n i s c u s p e r c o n v e x u s D a n a , a l e s s i m p o r t a n t p r e y s p e c i e s . I m e a s u r e d a m p h i p o d a c t i v i t y p a t t e r n s by p i t f a l l t r a p p i n g , a n d b e e t l e a c t i v i t y p a t t e r n s by d i r e c t c o u n t s o f t h e number o f b e e t l e s a c t i v e on t h e b e a c h i n 1 - h s e a r c h e s . I n g e n e r a l , t h e r e was a g o o d c o r r e s p o n d e n c e b e t w e e n b e e t l e a n d a m p h i p o d t e m p o r a l a n d s p a t i a l a c t i v i t y p a t t e r n s . H o w e v e r , by m a n i p u l a t i n g t h e s p a t i a l d i s t r i b u t i o n o f p r e y , I s h o w e d t h a t b e e t l e s a r r i v e d a t f o r a g i n g s i t e s i n d e p e n d e n t l y o f p r e y a v a i l a b i l i t y . P r e y c a p t u r e r a t e was l o w , w i t h a mean o f 7 5 m i n b e t w e e n c a p t u r e s , s o t h a t b e e t l e s w e r e n o t a l w a y s s u c c e s s f u l i n o b t a i n i n g f o o d d u r i n g a n i g h t . F o o d d e p r i v a t i o n f o r up t o 4 - d i n t e r v a l s d i d n o t a f f e c t b e e t l e s u r v i v a l o r o v i p o s i t i o n r a t e s i n l a b o r a t o r y e x p e r i m e n t s . I c o n s t r u c t e d m o d e l s o f a m p h i p o d s i z e s e l e c t i o n by b e e t l e s , u s i n g t h e s i z e d i s t r i b u t i o n s o f a m p h i p o d s m e a s u r e d on t h e b e a c h , a n d t h e r e s u l t s o f l a b o r a t o r y e x p e r i m e n t s on c a p t u r e s u c c e s s , r e a c t i o n d i s t a n c e a n d f e e d i n g r a t e s . C a p t u r e s u c c e s s d e c r e a s e d a n d t h e p r o b a b i l i t y t h a t a n a m p h i p o d was d e t e c t e d i n c r e a s e d w i t h i n c r e a s i n g a m p h i p o d s i z e . B e e t l e s o b s e r v e d d u r i n g b e a c h searches selected larger sizes of amphipods than predicted from a v a i l a b i l i t y and v u l n e r a b i l i t y of d i f f e r e n t s i z e s . To apply an optimal foraging model, I estimated the p r o f i t a b i l i t y of d i f f e r e n t sizes of amphipods from the number of amphipods of a given size required to satiate a beetle in the laboratory. P r o f i t a b i l i t y was highest for large amphipods and lowest for small amphipods and isopods. However, amphipod abundance on the beach was always below the threshold at which s p e c i a l i z a t i o n on larger sizes was predicted to occur. Male beetles were active longer than female beetles during the night, and fewer male beetles were observed feeding. Male beetles tended to be found higher on the beach and to include more isopods in their diet than female beetles. In laboratory experiments I showed that amphipods were highly preferred over isopods by both sexes of beetles. Male and female beetles were approximately the same size and consumed equal numbers of prey items. I conclude that male foraging behaviour was altered by search for mates. I present an optimal diet model for two prey types, based on the expected foraging time required for a predator to reach s a t i a t i o n . Predictions d i f f e r in some cases from a model based on maximization of the rate of energy intake. Foraging time may be minimized by a predator which begins as a s p e c i a l i s t and then expands i t s diet to include lower value prey when i t i s near s a t i a t i o n . Laboratory experiments on Thinopinus give weak support for these predictions, but I present alternative interpretations of the r e s u l t s . I suggest that most iv invertebrate predators which forage on active prey are limited in their a b i l i t y to assess variations . in prey abundance. Future studies should emphasize how patchiness in prey a v a i l a b i l i t y a f f e c t s foraging behaviour. V TABLE OF CONTENTS ABSTRACT i i LIST OF TABLES v i i i LIST OF FIGURES ix LIST OF SYMBOLS x ACKNOWLEDGEMENTS x i i Chapter 1: General introduction 1 Background to the problem 1 The study animals 4 The study s i t e 6 Chapter 2: Prey patchiness and foraging 7 Introduction •••• • 7 Materials and Methods 8 Predator and Prey A c t i v i t y Patterns 8 Behaviour at a patch 10 Laboratory feeding experiments 12 Survival and oviposition rates 13 Results 14 Temporal changes in amphipod a c t i v i t y 14 Beetle a c t i v i t y patterns 18 Behaviour at a patch 27 Laboratory feeding experiments 38 Survival and oviposition rates 38 Discussion 44 When where and how to forage 44 v i Sex differences 48 Relation to theory 49 Chapter 3: Prey selection 51 Introduction 51 Models 53 Materials and Methods 56 F i e l d data 56 V u l n e r a b i l i t y 57 Feeding experiments 59 Results 60-V u l n e r a b i l i t y 60 Feeding rates 62 F i e l d results 68 A test for prey selection 82 Mechanistic model 82 Frequency-dependent model 84 Optimal diet model 84 Discussion 89 Why did the optimal diet model not work? 92 Isopods versus amphipods 93 Chapter 4: Hunger and Optimal diet 95 Introduction 95 The model 96 A test with Thinopinus 105 Discussion 111 Chapter 5: Concluding remarks 115 Literature c i t e d 117 v i i Appendix A 126 Appendix B 128 v i i i LIST OF TABLES Table I. Time budgets for male and female beetles 2 5 Table II. Results of patch experiments 37 Table I I I . Results of the oviposition experiment 43 Table IV. Capture success 61 Table V. Feeding data 6 5 Table VI. Feeding data for beetles fed to sa t i a t i o n 6 8 Table VII. Parameter values for exact f i t 8 5 Table VIII. Parameters for the optimal diet model 8 6 Table IX. Foraging times generated by the model 8 8 Table X. Values of c 109 LIST OF FIGURES Figure 1. Temporal a c t i v i t y of amphipods and beetles 15 Figure 2. Nightly changes in amphipod size d i s t r i b u t i o n s .. 19 Figure 3. beetle and amphipod abundance 23 Figure 4. Frequency d i s t r i b u t i o n of move durations 29 Figure 5. Proportion of time in active mode 31 Figure 6. Frequency d i s t r i b u t i o n of attacks/min 35 Figure 7. Mean number of amphipods eaten 39 Figure 8. Survivorship curves 41 Figure 9. Reaction distances 63 Figure 10. Mean number of isopods eaten 66 Figure 11. Monthly changes in amphipod size d i s t r i b u t i o n s . 70 Figure 12. Morphological comparisons of male and female beetles 74 Figure 13. Mandible spread 76 Figure 14. Scavengers and amphipod size 78 Figure 15. Predicted and available amphipod d i s t r i b u t i o n s . 80 Figure 16. Expected t o t a l foraging times 100 Figure 17. Expected t o t a l foraging times for the general case 1 03 Figure 18. Mean numbers of amphipods eaten 107 X LIST OF SYMBOLS A pro b a b i l i t y that an encountered prey item i s attacked c c o e f f i c i e n t of preference (Murdoch 1969) c ( i ) probability that prey of size i are captured when detected C in the optimal diet model, the probability that a prey item i s captured, given that i t is attacked d(i) maximum reaction distance for a prey of size i D t o t a l food value required by a predator e(i) food value of a prey of size i f ( i ) r e l a t i v e encounter frequency with prey of size i h(i) handling time for a prey of size i H t o t a l time spent in handling prey items K proportionality constant which depends on the shape of the reactive f i e l d of the predator N t o t a l number of prey items eaten p(i) proportion of prey of size i in the diet of a non-selective forager p ( i ) ' proportion of prey of size i observed in the diet of a non-selective forager r ( i ) proportion of prey of size i observed in the diet R prey encounter rate S t o t a l time spent in prey search v( i ) v u l n e r a b i l i t y of prey of size i to the predator w(i) number of prey of size i eaten in preference experiments xi number of prey of size i presented in preference experiments preference for prey of size i (Chesson 1978) x i i ACKNOWLEDGEMENTS F i r s t and foremost I would l i k e to thank my supervisor, Judy Myers, for making this study possible. Jamie Smith deserves special mention for helpful comments and advice at most stages. I also wish to thank the staff of Bamfield Marine Station and the members of my supervisory committee, Drs. T.H. Carefoot, C.J. Krebs, C. Levings, W.E. N e i l l and J.D. McPhail for help in various ways. I also benefited from discussions with Lee Gass, Sarah Groves, Peter Morrison, Pamela Mace, Richard Palmer, and Jens Roland, and from the mathematical expertise of John Parslow. Elizabeth Boulding and Bruce T i l l a s sisted in the f i e l d in 1979 and 1980 respectively. The Coastal Missions provided the coffee and moral support to keep me awake at night in 1979, and a l l the people with whom I shared accommodation l e t me sleep in the morning. I was supported by a NSERC Science Scholarship and a H.R. MacMillan Family Fellowship. Research costs were met by NSERC and NAHS grants to J.H. Myers. F i n a l l y I wish to express my appreciation to Roy Getman for the song "The Moonlight Beach Bug B a l l " . 1 CHAPTER 1: GENERAL INTRODUCTION Background to the problem Predation is one of the major processes which structure communities (Brooks and Dodson 1965, Connell 1975). The success of a predator in finding suitable prey items w i l l not only af f e c t i t s own fi t n e s s , but w i l l also have consequences on the dynamics of the community in which i t resides. One good example of how t h i s process can operate is a study by Inouye et a l . (1980). In a series of predator removal experiments, they showed that seed predation affected the d i v e r s i t y of annual plants in the Sonoran Desert. The outcome d i f f e r e d however, depending on whether the predators were rodents or ants or both. The o v e r a l l importance of predation has lead to a search for general rules which govern predator behaviour. One such attempt i s known as optimal foraging theory (for reviews see Schoener (1971), Pyke et a l . (1977), and Krebs (1978)). Optimal foraging theory is based on the premise that foraging behaviour has evolved through natural selection. Observed behaviours r e f l e c t the outcome of thi s selection, and should approximate the behaviours which maximize individual f i t n e s s . To apply t h i s approach, the investigator must f i r s t i d e n t i f y the constraints on the animal, such as time or energy l i m i t a t i o n . The optimization procedure i s then applied subject to these constraints. Optimal foraging theory has been severely c r i t i c i z e d for several reasons. One major c r i t i c i s m i s that the hypothesis that 2 animals optimize in some way i s not f a l s i f i a b l e (Maynard Smith 1978). The hope, rather, is that informed modelling of the behaviours can lead to some insight into their evolution. Given th i s r e s t r i c t i o n i t i s surprising, as stressed by Ollason (1980) and Hanski (1980), that in most studies the optimization approach i s followed without consideration of alternate hypotheses. A second major c r i t i c i s m of the theory i s the general lack of f i t between quantitative predictions of the theory and the actual behaviours observed, although the q u a l i t a t i v e predictions may be met. Most of these tests have been conducted on vertebrates, especially birds, under highly controlled conditions. Schluter (1981) has recently reviewed the (lack of) evidence for optimal diets, and concluded that foraging studies of (mainly) vertebrates conducted in the f i e l d do not support predictions of the theory. One reason i s that the necessary decisions are often too complex for the animal to make. Another reason i s that the energy content of a prey item, used by* most investigators to assign prey value, i s not always a s u f f i c i e n t index of i t s value to a predator. I chose to conduct a f i e l d study of foraging in an insect predator, the beetle Thinopinus pictus LeConte (Staphylinidae). My general aims were (1) to assess the a b i l i t y of t h i s beetle to make complex foraging decisions (2) to assess the importance of this a b i l i t y to individual fitness (3) to test predictions of optimal diet models against an alte r n a t i v e model based on d i f f e r e n t i a l prey v u l n e r a b i l i t y . Thinopinus l i v e s in a 3 s t r u c t u r a l l y simple environment, the sand beach. Adults are ambush predators of beach invertebrates (Craig 1970), mainly the amphipod Orchestoidea c a l i f o r n i a n a (Brandt) and the isopod Alloniscus perconvexus Dana. I chose th i s system because of i t s str u c t u r a l s i m p l i c i t y . I could d i r e c t l y observe foraging of the beetles, and I could measure the sizes of prey items of beetles. I could also manipulate amphipod a v a i l a b i l i t y by moving patches of d r i f t seaweed. This thesis i s divided into 3 main sections. In Chapter 2, I describe the behaviour and a c t i v i t y patterns of the beetle and i t s major prey. I test two assumptions of optimal foraging models: (1) that predators can assess and respond to variations in prey a v a i l a b i l i t y , and (2) that short-term foraging success is closely linked to f i t n e s s . In Chapter 3, I consider the problem of diet selection. This aspect of foraging theory has received the most attention, probably because i t s predictions are the simplest to test experimentally. I construct an optimal diet and a simple mechanistic model of prey selection, and test these models against behaviours observed in the f i e l d . In Chapter 4, I present a new version of an optimal diet model. This model i l l u s t r a t e s how the addition of one assumption can a l t e r predictions of the model. I test these predictions in simple laboratory experiments. The l a s t chapter contains a few general remarks on the future of optimal foraging theory. 4 The study animals Thinopinus pictus inhabits exposed sand beaches on the west coast of North America. Adult beetles are active on the sand surface only at night. They spend the day in temporary burrows on the upper part of the beach. After .dark they emerge from these burrows and move down the beach to the high tide l e v e l where they forage. Beetles generally wait within a few cm of d r i f t seaweed and attack prey items moving on or off the weed by lunging and grasping the prey in their mandibles. They feed by inj e c t i n g digestive enzymes and sucking the digested material from their prey. This leaves a carcass which can be i d e n t i f i e d and measured. I frequently observed mating beetles throughout the summer. When a male and female meet, the male lunges forward and attempts to grasp the female in i t s mandibles, as i t would a prey item. The male then mounts the back of the female and link s g e n i t a l i a . The t o t a l time required for copulation is approximately 2 min. Females usually r e s i s t the mating attempt and may continue other a c t i v i t i e s such as burrowing, feeding, or attacking prey items while mating. More than one male may attempt to mate with a female, and males often attempted to mount other males. Sexes of Thinopinus can be distinguished by a c l e f t in the la s t abdominal segment of the male. Female Thinopinus lay their eggs singly in damp sand (Craig 1970). Eggs weighed on average 4.4±0.1 mg (±1SE) (n=25, l i v e weight) and hatched in approximately 3 wks at laboratory temperatures (16-20°C). I was unable to rear larvae in the 5 laboratory to determine the number or duration of in s t a r s . Larvae were not active on the sand surface at night and so were not included in t h i s study. I occasionally observed larvae running across the sand surface in late afternoon, or found them in burrows on the upper part of the beach. The natural history of sand beach amphipods on the C a l i f o r n i a coast has been studied extensively by Bowers (1964) and Craig (1971,1973a,b). The predominant species at my f i e l d s i t e was Orchestoidea c a l i f o r n i a n a . This amphipod digs burrows in the soft sand on the upper part of the beach in which i t spends the day, similar to the pattern shown by Thinopinus. At dusk, Orchestoidea c a l i f o r n i a n a emerges and moves down the beach to the d r i f t l i n e l e f t by the previous high ti d e . It is omnivorous, although i t feeds mainly on d r i f t seaweed. It sometimes forms large feeding aggregations. Juvenile amphipods do not burrow, but remain under d r i f t seaweed during the day. Juveniles are active mainly at dawn and dusk. Isopods, Alloniscus perconvexus, show an a c t i v i t y pattern similar to amphipods. Isopods tend to burrow in drie r sand higher on the beach than amphipods, and to feed on dried seaweed. Apart from Thinopinus, there are several other beetles of the families Staphylinidae, Carabidae, and Curculionidae which comprise the beetle community on the beach. These other beetles often join Thinopinus in feeding on a prey item. The most frequent scavengers are an unidentified staphylinid and Dyschirius obesus L e C , a carabid, both about 2 mm in length. 6 Amphipods and isopods occasionally join in the scavenging as well. Thinopinus reacts to the presence of scavengers by shaking i t s prey or carrying i t away from the s i t e of capture, sometimes as much as several meters. The study s i t e Most f i e l d data were collected between A p r i l and September of 1979 to 1 9 8 1 . The main study s i t e was located at Pachena Beach (48°53'N l a t . , 125°7'W long.) near Bamfield, on the west coast of Vancouver Island, Canada. This i s a wide fine-grained sand beach about 1 km long. Tides in thi s region are mixed semi-diurnal. The two da i l y high tides usually d i f f e r in height, and each leave a l i n e of d r i f t seaweed. I divided the beach into upper and lower sections based on the position of the d r i f t l i n e l e f t by the previous higher high t i d e . The upper beach extended from above this l i n e to the backshore and included a l l burrows of amphipods. The lower beach included the d r i f t l i n e and extended down to the water. For a given night the width of the upper beach remained constant, but that of the lower beach varied as the tide moved in and out. 7 CHAPTER 2: PREY PATCHINESS AND FORAGING Introduction The environment of a foraging animal i s characterized by patchiness in prey d i s t r i b u t i o n (e.g. Wiens 1976, Hassell and Southwood 1978). This patchiness extends over a range of sp a t i a l and temporal scales. The a b i l i t y of a predator to respond to patchiness w i l l determine in part, i t s foraging success. The major attempt to incorporate patchiness into an optimal foraging model was made by Charnov (1976b). His Marginal Value Theorem predicts that a predator w i l l leave a patch when i t s capture rate in the patch decreases to the habitat average This model has been moderately successful in controlled laboratory experiments (Cowie 1977, Cook and Cockrell 1978), but has been found inappropriate in more complex f i e l d situations (Zach and F a l l s 1976, Hanski 1980, Morse and F r i t z 1982). Apart from the argument of whether or not an optimization approach i s correct, there are two simple explanations for the f a i l u r e of t h i s model. F i r s t , there i s always v a r i a b i l i t y among patches. Second, the forager is limited in i t s a b i l i t y to assess th i s v a r i a b i l i t y . The need for predators to sample their environment has been suggested repeatedly to account for deviations between observed and predicted values in tests of optimal foraging theory (Heinrich 1976, Davidson 1978, Krebs et a l . 1978). Yet only a few authors (e.g. Pyke 1978) have considered how foragers might learn about their environment. In t h i s chapter I use Thinopinus  pictus to test the assumption inherent in the model that 8 predators can assess and respond to habitat v a r i a t i o n . I then relate short-term predator success to f i t n e s s . The s p e c i f i c aims of t h i s chapter, then, are to (1) correlate temporal patterns of beetle and amphipod a c t i v i t y , (2) measure the effect of temporal and s p a t i a l d i s t r i b u t i o n s of prey on foraging success, and (3) relate foraging success to predator f i t n e s s in terms of survival and oviposition rates. Materials and Methods Predator and Prey A c t i v i t y Patterns Once each month in 1980 I monitored beetle and amphipod a c t i v i t y over the amphipod a c t i v i t y period. Dry nights were chosen as rain reduced a c t i v i t y of beetles and amphipods. I formed two rows of p i t f a l l traps spaced 3 m apart. One row was set along the d r i f t l i n e l e f t by higher high tide of the day, here written as HW, and the second row along the d r i f t l i n e l e f t by the highest high tide of the month, here written as HHW, about 10 m towards the backshore from HW on the nights chosen. For the May sample there were two rows of 12 traps 5 m apart with the rows spaced 1 m apart. P i t f a l l traps consisted of p l a s t i c cups (8.5 cm diameter, 7.5 cm deep) f i l l e d to one-third with seawater (Craig 1970, 1973a,b, Hayes 1970). The water prevented animals from escaping once they f e l l in the traps. This method sampled the r e l a t i v e abundance of active amphipods which would pass a beetle s i t t i n g motionless on the sand. Alternating traps were placed in position or removed each hour, 9 so that 10 traps in each row were set at any given time. I counted and released the numbers of amphipods (estimated as £10 mm), and the numbers of beetles which were caught each hour. I then counted the number of beetles found between the backshore and the water in a 10 m wide beach section at each end of the row of traps. Each 2 h I c o l l e c t e d and preserved amphipods from additional traps at the end of each row. Amphipod size was measured from the anterior of the head through the t h i r d abdominal segment, allowing for body curvature (Bowers 1963). On 6 July 1980, I compared the sizes of amphipods feeding on weed patches and caught in p i t f a l l traps. Before amphipod a c t i v i t y began, I selected five weed patches and positioned a p i t f a l l trap 15 cm from each patch. These traps were placed after peak amphipod a c t i v i t y had occurred. One hour later I c o l l e c t e d and preserved amphipods from both the traps and weed patches. I c o l l e c t e d additional a c t i v i t y data on beetles in 57 1-h searches on 22 nights between May and July 1981. For each search I wore a headlight and walked systematically in a series of transects. As only a small portion of the beach could be covered in one hour, searches were begun at the same location to minimize l o c a l variation in beetle density. I scored sex, behavior ( s i t t i n g , feeding, mating) and beach position (upper, lower beach) for each beetle found. The s i t t i n g behavior category also included a few beetles which were moving when f i r s t observed. For each search I recorded day of year, temperature, time 10 in hours after sunset, and amphipod abundance. These variables were used in backwards multiple regression analysis (Draper and Smith 1966). Amphipod abundance was estimated from the mean of the number of amphipods caught in 6 to 10 p i t f a l l traps set at the high water l e v e l . Mark-recapture experiments tested i f male and female beetles active on one night were equally l i k e l y to emerge from burrows the following night and i f feeding on one night influenced emergence on succeeding nights. On 7 and 19 July, a l l beetles were collected in two searches early in the night, separated according to sex, and divided into two groups. One group was l e f t undisturbed, and the second group was provided with an abundance of amphipods. At the end of the night, 2-3 h la t e r , beetles were marked with enamel paint on the thorax according to treatment group (food, no food) and released. The following night, three surveys were conducted to search for marked beetles. Behaviour at a patch During 1979 and 1980 I made observations of individual beetles, generally for 10 min periods, for a t o t a l of 62 h. To observe beetles I covered the lamp of the headlight with red cellophane to reduce l i g h t i n t e n s i t y . I was careful to not move or shine the l i g h t d i r e c t l y on the beetles during observations. Using a stopwatch and coding sheet, I obtained a chronology of a l l beetle a c t i v i t i e s . I eliminated records of beetles which were observed for periods of less than 5 min, and of beetles 11 which responded by movement towards my headlight. Beetles selected for observation were located near weed patches at HW. In 1980 I conducted a series of experiments to test which sizes or composition of weed patches attr a c t beetles. I constructed patches with the test c h a r a c t e r i s t i c s from d r i f t seaweed and Phyllospadix collected from the beach. Patches were spaced along HW, alternating patch types. A p i t f a l l trap was set beside each patch to catch beetles. These traps were not f i l l e d with water. Amphipods could escape from the traps, but beetles could not. I counted and released beetles caught in these traps each half hour for a t o t a l of f i v e or six times in the f i r s t part of the night. On 6 August, I tested patches of 20 cm, 40 cm and 60 cm diameter using the most common patch type on the beach, a mixture of mainly Fucus distichus and Phyllospadix  s c o u l e r i . I compared Fucus-Phyllospadix with Egregia menziesi i patches on 15 August and with Nereocystis luetkeana patches on 19 August. These patches were 40 cm in diameter. As a control, on 19 August I constructed 40 cm diameter patches from p l a s t i c garbage bags. I measured the attractiveness of Phyllospadix and these species of seaweed to amphipods in a separate series of experiments. Two species of d r i f t plants were compared at a time. Similar sizes of patches of each species were paired and placed along HW. One to two hours l a t e r I counted the numbers of feeding amphipods. 1 2 Laboratory feeding experiments Feeding experiments tested i f male and female beetles consumed the same numbers of prey items under similar conditions in the laboratory. Twelve beetles of each sex were placed i n d i v i d u a l l y in glass jars (8 cm diameter, 10 cm deep) and covered with a 3 cm layer of damp sand. Each jar also contained 2, 4, 6, 8 or 10 large (16-19 mm) amphipods. Jars were covered to prevent amphipod escape and l e f t overnight for 20-22 h under natural photoperiod at laboratory temperatures. The number of l i v e amphipods in each jar were then counted, and the number eaten determined by inference. There were two treatments. One group was preconditioned by holding for three days without food. The other group had been fed the previous night. Each beetle was only used once in each treatment. The rate of search, a, and the time spent handling prey, h, were estimated by non-linear regression techniques using Rogers' (1972) random predator equation ahE-aT E = N[1-e ] where E i s the number of amphipods eaten, N i s the number of amphipods presented, and T is t o t a l time and i s set to one. Use of this equation enabled comparisons to be made between male and female beetles for search rates and handling times. The equation is similar to the disc equation (Holling 1959) but compensates for removal of consumed prey from those a v a i l a b l e . 1 3 Survival and oviposition rates I c o l l e c t e d beetles for these experiments on 4 July 1980 and 10 May 1981 respectively, from Tapaltos Beach, about 4 km from the main f i e l d s i t e . Prior to beginning the experiment, beetles were l e f t with amphipods for one day to standardize sa t i a t i o n l e v e l s . I then placed beetles in individual jars (8 cm diameter, 10 cm deep) f i l l e d to one-half with damp sand. Beetles were kept under natural photoperiod at laboratory temperatures (16-20°C). I measured the effect of temperature and a regular feeding regime on survival rates with f i v e treatments of eight beetles of each sex per treatment. These were (1) food at 2-d intervals (2) food at 4-d intervals (3) food at 8-d intervals (4) no food (5) no food at low temperature (10-12°C). To feed beetles I placed three 12-15 mm amphipods in each j a r . The following day amphipods were removed and beetle survival was scored. This procedure was followed for 28 days. Beetles deposit their eggs singly in damp sand (Craig 1970). I measured the effect of regular feeding on oviposition rates with three treatments of 20 female beetles each: (1) continuous food (2) food at 3-d intervals (3) food at 6~d int e r v a l s . Amphipods were replaced in treatment 1, and the jars were checked for eggs on the day following feeding in treatment 2. After 30 days, the remaining beetles were dissected to determine the state of egg development in the ovaries. Means for a l l experimental results are shown with one standard error, except where indicated. Proportions used in 1 4 s t a t i s t i c a l tests were f i r s t transformed by arcsine square root. A l l tests were two-tailed with a significance l e v e l of 0.05, except where indicated. Non-parametric tests were used when assumptions of parametric tests were not met and included X 2 , median test, sign test and Spearman rank correlation c o e f f i c i e n t which are described in Siegel (1956). A l l times are given as P a c i f i c Daylight Time. Results Temporal changes in amphipod a c t i v i t y The numbers of amphipods caught in p i t f a l l traps at HW peaked just after dark on most nights (Fig. 1, s o l i d l i n e ) . Abundance then declined gradually with a second peak near dawn on some nights, such as shown in F i g . 1D. The peak was associated with emergence of amphipods from burrows on the upper part of the beach and migration to HW, or with the return to the upper beach from HW. While at HW, amphipods fed on d r i f t seaweed. They sometimes formed large aggregations, especially on kelp patches. These feeding amphipods could not be sampled by the p i t f a l l trap method. Amphipods caught in traps in the middle of the night were moving between patches or returning to the upper beach. The numbers of amphipods trapped at HHW (Fig. 1, dotted line) were lower and less variable than at HW. A few amphipods did feed on patches of dried weed at HHW. However, most amphipods caught in these traps were probably moving between the upper beach and HW. 15 Figure 1. The mean ±1SE numbers of amphipods caught in p i t f a l l traps at HW ( s o l i d line) and HHW (dotted line) at each hour over the night, and the number of beetles observed (open c i r c l e s ) . Beetle data were summed from p i t f a l l trap and transect counts. Arrows indicate the approximate time of high ti d e . The heavy l i n e gives the period of maximum darkness. (A) 6-7 May 1980 (B) 19-20 June 1980 (C) 13-14 July 1980 (D) 18-19 August 1980. Note the scale changes for amphipods in (B) and for beetles in (D). 16 18 Amphipod a c t i v i t y was affected by weather and tide patterns. The small peaks at 0300 in Fig . 1B and D were associated with incoming fog and a 1-2°C ri s e in temperature. Repeated observations indicated that a few amphipods only were active on rainy nights (Craig 1973b). When a high tide occurred early in the night (before about 0100), amphipod a c t i v i t y was delayed u n t i l after high t i d e . When the high tide occurred later in the night, there were a c t i v i t y peaks both before and after the high tide (Bowers 1964, Craig 1973b). F i g . 1C shows an intermediate stage in this t r a n s i t i o n . The size d i s t r i b u t i o n of active beetles varied over the night and with beach position (Fig. 2). Juvenile amphipods were 2 mm in length when released from the female brood pouch, and sexes could be distinguished at 12 mm. Juveniles were active primarily at dusk and dawn. They were trapped in low numbers only over the period of the night when beetles were active. Juveniles did not burrow but remained under weed patches during the day (Craig 1971), so that few juveniles were trapped at HHW. The juvenile peak at HHW at midnight (Fig. 2) probably represented the brood of a female which was released when the female f e l l into the trap. Beetle a c t i v i t y patterns I did not find active beetles each night u n t i l after amphipod a c t i v i t y had begun (Fig. 1, open c i r c l e s ) . Otherwise the o v e r a l l a c t i v i t y patterns of beetles and amphipods were 19 Figure 2. The frequency d i s t r i b u t i o n of the sizes of amphipods caught in p i t f a l l traps on 19-20 June 1980 at d i f f e r e n t times of night at HW and HHW. Mean ±1SE size (n) is given for each time. 20 2 2 0 0 x=4-5±0l(l399) 0 0 0 0 x = 131 ±0-3 (173) 0 2 0 0 x= II I + 0 5 (112) 0 4 0 0 X-II-2+ 0 4 (132) 0545 x=5-2±OI (585) 0 2 4 6 8 10 12 14 16 18 20 22 AMPHIPOD SIZE (mm) 21 02 O H 0-3 0-21 0 1 0-2 >-o UJ 0 | O UJ DC U_ 02 0-02^ 01 J HHW J l 2200 S = IO-3±0-8 (60) 0000 x = 1074 10 (49) l . • i i h 0200 x= 1304 0-4 (69) 4=L 1 0400 x-135+0 2 (98) n 1.,. 0545 x=IO 44 0 8 (49) 0 2 4 6 8 10 12 14 16 18 20 AMPHIPOD SIZE (mm) 22 s i m i l a r . Maximum numbers were counted just after dark, and the number of active beetles decreased gradually over the night. For the data c o l l e c t e d in 1981, I found a c o r r e l a t i o n of 0.273 (Fig. 3, n=57, p<0.05) between the number of beetles counted on the lower part of the beach and the number of amphipods trapped. Part of the variation in t h i s r e l a t i o n can probably be accounted for by differences in recruitment patterns of beetles and amphipods over the summer. I was unable to scale beetle counts to t o t a l size of the beetle population. I also found important fine scale differences between beetle and amphipod a c t i v i t y . For example, beetles did not delay a c t i v i t y when a high tide occurred in the middle of the night (Fig. 1C). Craig (1970) s i m i l a r l y reported peaks in beetle a c t i v i t y both before and after a high t i d e . A more detailed analysis of the beetle a c t i v i t y pattern was obtained from the 1981 search data, which indicated differences between male and female beetles in the amount and location of surface a c t i v i t y . A mean of 58.8±2.2 males and 50.4±2.0 females were counted during these searches (Table I ) . There were no differences in the mean number of males counted on either the upper or lower beach, but s i g n i f i c a n t l y more of the females were counted on the lower beach. Hence, the sex r a t i o was skewed towards males on the upper beach and towards females on the lower beach. This suggested that female beetles fed and then returned to the upper beach to burrow, while males were active longer during the night. A multiple regression on the proportion of males in the upper beach counts included date (r= - 0.425, 23 Figure 3. The number of beetles counted on the lower beach in 1981 as a function of the number of amphipods caught in p i t f a l l traps at the time the beetle count was made (r=0.275, n=57, p<0.05). 24 125 CO LJ HlOOJ LU LU CD O or CD 5 0 25 O O a> o o ° ° o o o o o o o o o 6° CD) O O O O O P O o o o 10 15 20 25 AMPHIPODS/TRAP 30 25 Table I : Time budgets for male and female beetles at different beach positions. Values given are for the mean proportion of time spent s i t t i n g , mating, and feeding, for the mean number of beetles counted, and for the mean proportion of males in 57 searches. A l l male-female comparisons for a given beach position are s i g n i f i c a n t (paired t - t e s t , p<0.0l) as are a l l upper-lower beach comparisons on proportions for males (p<0.05), but not for females (p>0.lO). For upper-lower beach comparisons on l:otal counts, differences are s i g n i f i c a n t for females (p<0.0CM), but not for males (p>0.lO). upper beach lower beach beach t o t a l males females males females males female: s i t t i n g 0.973 0.856 0.934 0.885 0.941 0.865 feeding 0.023 0. 1 33 0.050 0. 103 0.044 0.116 mating 0.004 0.011 0.016 0.012 0.015 0.019 t o t a l 28.6 11.6 30.4 38.7 58.9 50.4 prop, male 0. 735 0. 425 0. 542 n=57, p=0.00l), temperature (r=0.329, n=57, p=0.0l4) and time after sunset (r=0.280, p=0.039) with r2=0.230 (n=57, p=0.003). On the lower beach, the regression included temperature only (r=0.344, n=57, p=0.009). The t o t a l number of beetles counted on the lower beach also increased with temperature (r=0.387, n=57, P<0.01). Proportionately more males were active on warmer nights and l a t e r in the night, and fewer males were active rel a t i v e to females l a t e r in the summer. Males and females also d i f f e r e d in the proportion of times they were observed performing each behaviour. A mean of 11.6% of females were observed feeding in surveys compared to 4.4% of males. Females were observed s i t t i n g , feeding and mating in similar proportions at either beach position. Males, however, 26 spent s i g n i f i c a n t l y less time s i t t i n g / and more time feeding and mating when on the lower beach. Males on the lower beach fed in s i g n i f i c a n t l y lower proportions than did females at the same beach po s i t i o n . In an attempt to determine i f active beetles were foraging, beetles were placed in open buckets on the beach containing amphipods. Rejection of amphipods would suggest that active beetles were not searching for food. There were no differences in response of male beetles from the upper or lower beach, when beetles were coll e c t e d early in the evening, before feeding had begun. Forty-five percent (n=20) of males from the upper beach and 40% from the lower beach (n=20) accepted prey items. When thi s experiment was repeated with males c o l l e c t e d in the middle of the night, after males did have the opportunity to move to the lower beach and feed, 33% (n=12) of males on the upper beach and 95% (n=20) on the lower beach accepted amphipods (X 2 = 11.3, df=1, p<0.0l). Foraging males were more l i k e l y to be on the lower beach, although some males on the upper beach did feed as well. I n s u f f i c i e n t numbers of females were found on the upper beach on these nights to perform comparative t e s t s . Mark-recapture experiments tested i f male and female beetles marked on one night were equally l i k e l y to emerge the following night, and i f feeding affected emergence. In the 7 July experiment, the recapture rate was 14.4% (n=208) with no difference between sexes (X2=0.18, df=1, p>0.l0). Twice as many recaptures were beetles which had not fed the previous night, but t h i s difference was not s i g n i f i c a n t (X2=2.88, df=1, P>0.10). 27 For the 19 July experiment, the recapture rate was 50.3% (n=l95) with 58% of males and 43% of females- recaptured (X2=3.84, df=1, p=0.05). Recapture success did not depend on whether beetles had fed the previous night (X2=0.18, df=1, P>0.10). Differences between the experiments can probably be attributed to weather. On 8 July conditions were less favorable for beetle a c t i v i t y with impending rain and a temperature of 9°C, compared to overcast and 13°C on 20 July. Behaviour at a patch Continuous observations of beetles at HW suggested that beetles alternated between active and ambush foraging modes. I classed beetles in ambush mode when they made moves of no longer than 2 s in duration. These moves were probably a response to prey items detected at a distance and were of three types: (1) forward move - the beetle moved forward 1-3 cm (2) turn - the beetle changed i t s facing d i r e c t i o n , usually by 180° (3) lunge - the beetle appeared to attack, although I did not observe any prey items within s t r i k i n g distance. Because I could not detect whether or not beetles which remained motionless on the sand were in fact foraging, the ambush category also included beetles which may have been engaged in other a c t i v i t i e s , such as mate search or digestive pause. I classed beetles in active mode when at least one move in a sequence was greater than 2 s in duration and pauses were less than 2 s between successive moves. Most moves made by beetles 28 were less than 1 s in duration but moves ranged up to 78 s in duration for beetles which I observed in active mode (Fig. 4). For these longer moves beetles generally maintained a constant forward d i r e c t i o n , either in weed patches, or on the sand along or between weed patches. Beetles in active mode could also be considered as moving between patches, but they attacked amphipods they encountered while moving. Hence, patches could not always be considered as discrete units. Individual beetles spent on average 3.2-4.7% (95% confidence i n t e r v a l , n=362) of the observation time in active mode. The attack rate increased with the proportion of time active (Spearman rank c o r r e l a t i o n c o e f f i c i e n t = 0.168, n=362, P<0.01, F i g . 5). To test whether beetles moved u n t i l they located a good foraging s i t e , or in response to unsuccessful attacks on amphipods, I compared the proportion of time spent in active mode by beetles before and attacks on amphipods. There was a weak tendency for beetles to be more active following an unsuccessful attack (Mann-Whitney U-test, n=248, p=0.08). This suggested that active foraging mode was frequently a response to the detection of amphipods. The duration of individual moves did not d i f f e r before and after an attack (t=1.l6, df=3350, p>0.10). Male and female beetles spent similar proportions of time active (Mann-Whitney U-test, n=141,168, P>0.10). However, only 27% (n=141) of male beetles attacked amphipods compared to 52% (n=168) of females (X2=18.6, df=1, p<0.001). There was no difference between sexes in the proportion of attacks which resulted in capture (X2=0.06, df=1, p>0.!0). I presented 29 Figure 4. The frequency d i s t r i b u t i o n of the duration of individual moves made by beetles observed to attack amphipods. The mean i s 2.5±0 .1 s (n=3389). 30 m CD O m csi — o o g 6 m o b A0N3fl03cJJ CD CVJ CO < oo cr => Q CD > O gure 5. Frequency d i s t r i b u t i o n s of the proportion of time spent in active mode by beetles which did (white bars, n=149) and did not (black bars, n=213) attack amphipods. 0 0 0 1 0-2 0 - 3 0 - 4 0 - 5 PROPORTION OF TIME 33 amphipods d i r e c t l y to beetles in an experiment to test whether differences in attack rates between males and females resulted from behaviour differences or differences in encounter rates with prey. Male and female beetles were col l e c t e d early in the evening before feeding had begun. They were separated according to sex and grouped in open buckets on the beach with an abundance of amphipods. Male and female beetles had an equal opportunity to attack prey items. However, 63% (n=78) of the females captured amphipods compared to 30% (n=76) of the males (X2=15.1, df=1, p<0 . 0 l ) . Because of the differences in a c t i v i t y patterns of di f f e r e n t sizes of amphipods (Fig. 2), the a c t i v i t y and position of a beetle w i l l determine the sizes of amphipods i t encounters. On 6 July 1980 the mean size of amphipods coll e c t e d from weed patches was 6.1±0.1 mm (n=388), in contrast to a mean of 12.6±0.2 mm (n=275) from p i t f a l l traps. There were also s i g n i f i c a n t l o c a l differences in the size d i s t r i b u t i o n s of amphipods within the five samples taken from weed patches (X2=31.6, df=12, p<0 . 0 0 D and p i t f a l l traps (X2=26.8, df=8, P<0.001). Beetles observed on sand, captured amphipods with a mean size of 10.3±0.8 mm (n=32), compared to 6.5±0.8 mm (n=8) for beetles moving over weed patches (t=3.23, df=21, p<0.00l one-tailed). On average beetles attacked 0.147±0.019 (n=362) amphipods/min and captured 9.1% (n=440) of the amphipods attacked. Hence, beetles could expect to wait 6.8 min between attacks and 75 min between captures. The frequency d i s t r i b u t i o n 3 4 of attack rates (Fig. 6) however, was s i g n i f i c a n t l y different from random (X2=430, df=5, p<0.00l for f i t to Poisson). F i f t y - n i n e percent of beetles did not attack during the observation period, and more beetles had high attack rates than expected. This suggested that most foraging s i t e s were of low qua l i t y . Either beetles were r e l a t i v e l y unsuccessful in finding high quality s i t e s , or they did not detect, or did not respond to amphipods at a s i t e . Further, overa l l amphipod abundance did not influence foraging success. I found no relationship between either the number or proportion of beetles observed feeding during surveys in 1981 and mean amphipod abundance (Spearman rank correlation c o e f f i c i e n t = -0.009 and -0.087 respectively, n=57, p>0.10). Beetles were not attracted to foraging s i t e s on the basis of .patch s i z e . There were no differences in the t o t a l number of beetles caught in p i t f a l l traps by patches of 20, 40 or 60 cm diameter (Table I I ) . However, species composition of the patch influenced the number of beetles trapped. S i g n i f i c a n t l y more beetles were caught in traps near Egregia or Nereocystis than near Fucus-Phyllospadix patches. I used p l a s t i c garbage bags as a control for a patch type which did not attract amphipods. Beetles were caught in traps near these patches, but s i g n i f i c a n t l y fewer were caught than near Fucus-Phyllospadix patches. This result suggests that beetles were i n i t i a l l y attracted to a l l patch types, but that they remained longer near, and so were more l i k e l y to be trapped near Egregia and Nereocystis patches than near Fucus-Phyllospadix patches, and 3 5 Figure 6. The frequency d i s t r i b u t i o n of the number of amphipods attacked/min by beetles during observations in 1979 and 1980 (n=362). Dotted l i n e s give predicted values for a Poison d i s t r i b u t i o n with the same mean (0.147 amphipods/min). 6 CO 6 rr-O 6 O 6 ro 6 6 I co o Q-Q-E o LU < cr co 6 6 ro O O o A0N3H03dJ 37 Table I I : The numbers of beetles caught in p i t f a l l traps located near patches of d i f f e r e n t size or species composition, and sample sizes for each type (F-P, Fucus-Phyllospadix; E, Egregia; GB, garbage bag; N, Nereocystis). Probability levels are for ANOVA. Date n No. of beetles counted patch size comparisons 20 cm 40 cm 60 cm 6 Aug. 10 6.1±1.6 5.6±0.8 4.7±1.1 p>0.l0 patch composition comparisons F-P E 15 Aug. 12 8.8±0.9 12.2±1.1 p<0.05 GB F-P N 19 Aug. 12 4.1±0.5 6.6±0.8 10.3±1.0 p<0.000l near Fucus-Phyllospadix patches than near garbage bags. Patches near which I trapped more beetles were also more a t t r a c t i v e to amphipods. In the amphipod experiments, more amphipods were counted on Nereocystis than on Egregia (n=24, P<0.001), on Nereocystis than on Phyllospadix (n=22, p<0.00l), on Egregia than on Phyllospadix (n=14, p<0.05), and on Phyllospadix than on Fucus (n=16, p<0.005, a l l Wilcoxon matched-pairs signed-ranks test, one-tailed). 38 Laboratory feeding experiments Laboratory feeding experiments tested i f there were differences in the number of amphipods eaten at d i f f e r e n t amphipod densities for male and female beetles which had been starved or fed. The number of amphipods eaten by starved beetles increased with the number of amphipods presented (Fig. 7). Beetles which had fed the night prior to the experiment consumed fewer amphipods than those which had been starved. There was a weak tendency for fed females to consume more amphipods than fed males, but t h i s trend was not consistent at a l l densities. There were no differences between male and female beetles either in the estimates of attack rate or handling time for either treatment ( t - t e s t , df=1l7, p>0.10). Survival and oviposition rates In laboratory experiments beetles survived at least 8 days without food at laboratory temperatures, and at least 24 days at the lower temperature which more closely approximated conditions on the beach (Fig. 8). Only one of the beetles which was fed at 2- or 4-d intervals died during the experiment, and there were no differences in percent weight loss of beetles between these two treatments (Mann-Whitney U-test, n=16,15, P>0.10). F i f t y - s i x percent of the beetles fed at 8-d intervals died between days 10 and 28. A t o t a l of nine beetles died or were lost in the oviposition experiment. Oviposition rates were low for the 3 9 Figure 7. The mean ±1SE number of amphipods eaten at each amphipod density for starved (A) and fed (B) beetles during a 20-22 h period. As differences between male and female beetles were not s i g n i f i c a n t , male and female data were combined for curve f i t t i n g . The equations are (A) E = N(1 - exp(0.447E - 1.860)) (B) E = N(1 - exp(0.253E - 1.072)) 40 41 Figure 8. Survivorship curves for beetles according to food treatment: (a) food at 2-d intervals (b) food at 4-d intervals (c) food at 8-d intervals (d) no food (e) no food at low temperature (10-12°C). 43 remaining beetles (Table I I I ) . When I dissected beetles I found Table I I I . Results of the oviposition experiment. The table gives the number of beetles, the proportion which l a i d eggs, the proportion with at least one mature egg in their ovaries, the t o t a l number of eggs l a i d , and the mean egg dry weight for each treatment: ( 1 ) continuous food (2) food at 3-d intervals (3) food at 6-d i n t e r v a l s . Treatment 2 3 n 1 5 20 1 6 prop, l a i d eggs 0.33 0.35 0.19 prop, eggs in ovar. 0.60 0.60 0.06 no. eggs 1 2 18 3 mean egg dry weight 6.2±0.2 5.3±0.2 5.0±0 (mg) that 60% of the beetles which were fed continuously or at 3-d intervals , and one of the beetles fed at 6-d intervals had a mature egg in the common oviduct. This suggested that conditions in the jars were not favourable for oviposition. However, there were no differences in the proportion of beetles which either l a i d eggs, or had at least one egg in their ovaries, for beetles fed continuously or at 3-d in t e r v a l s . I combined these data to test for the effects of feeding at 6-d int e r v a l s . There was no difference in the proportion of beetles which l a i d eggs (X2=0.64, df=1, p>0.10), but fewer beetles which were fed at 6-d intervals had at least one mature egg in their ovaries (X2=20.4, df=1, p<0.00l). 44 Beetles which were fed continuously l a i d eggs of s i g n i f i c a n t l y greater dry weight than beetles fed at 3-d intervals (t=3.03, df=28, p<0.0l). There was no c o r r e l a t i o n between beetle size (measured as head capsule width) and mean egg dry weight (r=-0.023, n=15, p>0.10), between beetle size and the number of eggs l a i d (r=-0.295, n=15, p>0.1u), or between mean egg dry weight and the number of eggs l a i d (r=0.054, n=l5, p>0.10). Discussion When where and how to forage Although the sand beach habitat i s s t r u c t u r a l l y simple, amphipod abundance varied over a range of temporal and s p a t i a l scales. Patterns of abundance could largely be accounted for by weather and tide factors, and the d i s t r i b u t i o n of d i f f e r e n t types of weed patches. In general, a good correspondence occurred between the timing of beetle and amphipod a c t i v i t y , but there were important discrepancies. Beetle abundance did not perfectly track amphipod abundance. In order to apply an optimal foraging model, i t is necessary to assume that the predator can assess or "know" prey abundance both at a patch, and for the habitat average. Foraging beetles could encounter very d i f f e r e n t numbers of amphipods at the same time and at similar s i t e s on consecutive nights. This must complicate any assessment process. One additional feature of this system was that the location and 45 quality of patches changed each night. There was no benefit to a beetle gained by remembering the time or location of a success on the previous night. If beetles and amphipods use d i f f e r e n t cues to i n i t i a t e or maintain surface a c t i v i t y , beetles may be unable to respond to night-to-night differences in the times of amphipod a c t i v i t y . A l t e r n a t i v e l y , i t i s possible that beetles need to forage during periods of both high and low amphipod a v a i l a b i l i t y , to increase the p r o b a b i l i t y of eventual prey capture. Foraging in ambush mode may be s u f f i c i e n t l y inexpensive that the probability of prey capture outweighs any energetic cost. However, beetles foraging before high tide were occasionally h i t and washed away by waves. This may be an important mortality factor. Insects have well-developed chemical responses. I was unable to e l i c i t any response from Thinopinus to seaweed odors in preliminary experiments. Beetles may be using v i s u a l cues such as silhouettes to find patches, as they were attracted to patches moved to at least 20 m below HW. In addition, Craig (1973a) showed that Thinopinus, Orchestoidea c a l i f o r n i a n a and Alloniscus perconvexus a l l moved up-slope when exposed to wet slopes of 5° in the laboratory. These animals may simply move up or down beach according to the slope of the beach and degree of wetness of the sand u n t i l they encounter some object such as a weed patch. None of these mechanisms for finding patches provide information on amphipod abundance. Beetles arrived at a l l types of patches in my experiments, including garbage bags, whether or not these patches were a t t r a c t i v e to amphipods. In general, cues 46 used by insects to find food patches provide l i t t l e information on r e l a t i v e prey a v a i l a b i l i t y at the patch (Hassell and Southwood 1978). Selection of ambush s i t e s by movement from patch-to-patch u n t i l a s i t e with high prey a v a i l a b i l i t y is located, has been suggested for web-building spiders (Turnbull 1964) and damselfly nymphs (Crowley 1979). Although I counted more beetles at patches which were also more a t t r a c t i v e to amphipods, beetles were probably limited in their a b i l i t y to assess amphipod abundance once at a patch. The attack frequency tended to be too low to act as a useful index of amphipod abundance. For example, one attack soon after a r r i v a l at an area of low amphipod abundance could lead to a spurious impression of high amphipod abundance. Scorpions detect prey through substrate vibrations (Brownell and Farley 1979). Thinopinus may use a similar mechanism. Short moves and turns-made by beetles suggested responses to amphipods detected at a distance. However, the number of amphipods on a patch, i f they are not moving, may not be a good indicator of the number available to a beetle in ambush mode. Amphipod abundance also changed over the night independently of beetle a c t i v i t y . I counted more beetles on the lower beach on warmer nights. Mark-recapture experiments suggested that t h i s was because more beetles emerged from burrows on successive nights at higher temperatures. P o l i s (1980) found a similar r e l a t i o n between temperature and surface a c t i v i t y of desert scorpions. Thinopinus probably required more food at higher temperatures as survival during starvation was enhanced at low temperatures in laboratory 47 experiments. For example, dasmselfly larvae (Thompson 1978a) and mites (Everson 1980) increase t h e i r attack rates on prey items at higher temperatures. Temperature probably a f f e c t s a c t i v i t y i n d i r e c t l y through i t s affect on hunger. Hunger has been shown to lead to increased a c t i v i t y for a number of d i f f e r e n t types of predators (Beukema 1968, Griim 1971, Calow 1974). Akre and Johnson (1979) and Crowley (1979) suggested that hungry damselfly nymphs shifted from ambush to active foraging modes at low prey densities. This could also be true for Thinopinus, as the proportion of time spent in active mode tended to increase following an unsuccesful attack. Active foraging mode did have associated disadvantages. As well as the energetic cost, active beetles on seaweed disturbed amphipods, depleting abundance at the patch. This i s what Charnov e_t a_l. (1976) have termed resource depression. Other beetles in ambush mode near the patch could p o t e n t i a l l y benefit from the resulting increase in prey a c t i v i t y . Toft (1980) studied foraging behaviours of several species of t r o p i c a l anurans and found that species which were active foragers took smaller prey items and captured more prey per unit time than species which were ambush foragers. This was also true for Thinopinus. Beetles which were active on seaweed encountered a higher proportion of juvenile amphipods, and captured smaller amphipods on average, than beetles in ambush mode. One beetle I observed by a kelp patch rapidly captured and consumed two 5 mm amphipods in a 15 min observation period. In chapter 3, I show that one large amphipod (>15 mm) i s s u f f i c i e n t to satiate most 48 beetles, while up to six juvenile amphipods are required. Feeding on small amphipods could lead to a s i g n i f i c a n t increase in the t o t a l time required for foraging at low amphipod abundance or in areas of mostly juveniles. Sex differences Sex differences in a c t i v i t y , attack rates, beach feeding experiments and mark-recapture experiments, indicated that either (1) male beetles fed less on the beach or (2) that they spent a greater proportion of their non-foraging time active on the sand surface, or (3) both. I did not find differences in feeding rates of male and female beetles in short-term laboratory experiments to suggest that females had a greater food requirement. However, there may be long-term differences. Females which were fed continuously l a i d heavier eggs than those females fed at 3-d intervals (Table I I I ) . For the cinnabar moth, egg weight is related to hatching success under adverse conditions (Richards and Myers 1980). A similar relationship in Thinopinus may have selected for more frequent feeding in females. As well, male beetles must spend more of their non-foraging time active, and they probably devote t h i s extra time to mate search. I observed mating throughout the summer. The problem of mate search for male beetles is si m i l a r to the problem of searching for amphipods (Parker and Stuart 1976), and they may attempt to perform both behaviours simultaneously. If males are searching for food and/or mates, the reasons 49 for the s p a t i a l differences between male and female a c t i v i t y are not obvious, as both prey items and female beetles were more abundant on the lower beach. Mating may sometimes occur in burrows, so I may have underestimated mating success of males on the upper beach. Even i f the p r o b a b i l i t y of obtaining a mate on the upper beach is lower, s i t t i n g on the upper beach had several advantages. It was energetically cheap, compared to a return t r i p of about 50 m to the lower beach. Females emerged from and return to burrows in the same area. The last male to mate with a female in a night probably f e r t i l i z e s eggs l a i d that night (Schlager 1960, Parker 1970, Smith 1979). Relation to theory To a casual observer, amphipods might appear as an unlimited food resource. Certainly, some beetles were successful in finding patches of high amphipod abundance. I have shown t h i s not to be true in general. Beetles foraged at times and in locations of low amphipod abundance. Mean attack rates and capture success were also low, and beetles were probably frequently unsuccessful in prey capture during a night. Beetles were obviously limited in their a b i l i t y to assess quality of a foraging s i t e , and any assessment process may have been confounded by the v a r i a b i l i t y which could exist on successive nights. As one c r i t i c a l assumption of the optimality approach i s that foragers can assess q u a l i t y of a patch, i t i s not possible to design a test of an optimal patch choice model in t h i s system. Any deviation between the observed and predicted values 50 could be accounted for by either sampling of the beetle or by lack of f i t of the model. Morse and F r i t z (1982) reached similar conclusions in their study of crab spiders on milkweed. They found spiders at good s i t e s more frequently than random, but spiders did not immediately leave poor sites when better s i t e s were provided. They claimed that spiders which moved to new flowers or stems had l i t t l e information available to them other than the number of insect a r r i v a l s at their previous s i t e . A second c r i t i c a l assumption of the optimality approach i s not met with Thinopinus. In most foraging studies the claim i s made that foraging success does affect short-term f i t n e s s . I found no differences in s u r v i v a l , weight changes, or oviposition rates for beetles fed in the laboratory at up to 4-d i n t e r v a l s over a 28-30 d period. In general, low foraging success on a few nights appears unlikely to a f f e c t oviposition or survival rates in this beetle. Hanski (1980) found that movements of dung beetles between cow pats could be accounted for more cl o s e l y by a simple stochastic model than by a model based on maximization of net energy intake. Mechanistic or stochastic models, based on s p e c i f i c systems and which incorporate sensory information, are l i k e l y to be useful in future as tools to predicting foraging behaviours. They should at least be considered as alternate hypotheses. In the next chapter I compare predictions of optimal diet and mechanistic models in prey choice of Thinopinus. 51 CHAPTER 3: PREY SELECTION Introduct ion There i s abundant evidence that animals discriminate among the range of potential prey types, or sizes of a given prey type available to them (e.g. see Pyke et a_l. 1977). Any difference between the d i s t r i b u t i o n of prey types in the diet and a v a i l a b i l i t y in the environment is a measure of prey selection (Eggers 1977). Selection may result d i r e c t l y through active choice by the predator (e.g. Zach 1978) or i n d i r e c t l y through d i f f e r e n t i a l v u l n e r a b i l i t y of prey types (e.g. Pastorok 1981). In t h i s chapter I describe a f i e l d study of predation by the beetle Thinopinus pictus Leconte (Staphylinidae) on di f f e r e n t size classes of amphipods Orchestoidea ca l i f o r n i a n a (Brandt). I also present some data for beetle predation on isopods, Alloniscus perconvexus Dana. I test f i r s t whether beetles select certain sizes or types of prey, and second whether th i s selection results from d i f f e r e n t i a l v u l n e r a b i l i t y of di f f e r e n t prey types or from active choice. Theoretical attempts to predict diet selection f a l l into two general classes: (1) frequency-dependent models (2) optimal diet models. Frequency-dependent models (Murdoch and Oaten 1975, Greenwood and Elton 1979) are based on the hypothesis that the predators feed disproportionately on the most abundant prey items. Abundant items w i l l be over-represented in the diet, and rare items w i l l be under-represented r e l a t i v e to a v a i l a b i l i t y . Optimal diet models (Pyke et a l . 1977, Krebs 1978) f i r s t 52 assume that prey items can be ranked according to some measure of t h e i r p r o f i t a b i l i t y , such as the r a t i o of energy value to handling time. Prey types are added to the diet in their rank order. The optimal diet consists of the subset of prey types which re s u l t s in optimization of some c r i t e r i o n chosen by the investigator, such as the net rate of energy intake. Predictions are that (1) low value prey w i l l be eaten only when more pr o f i t a b l e prey are rare (2) as the abundance of prof i t a b l e prey increases, predators become more specialized (3) prof i t a b l e prey w i l l always be eaten and unprofitable prey never eaten when encountered (Pulliam 1974, Charnov 1976a). F i e l d tests of these models are often d i f f i c u l t due to changing conditions and problems in measurement of prey a v a i l a b i l i t y and d i e t . I chose to measure prey selection in Thinopinus because I could observe t h i s beetle feeding d i r e c t l y in i t s natural habitat, a sand beach. The aims of this chapter are (1) to develop a simple mechanistic model of prey selection based on d i f f e r e n t i a l prey v u l n e r a b i l i t y (2) to develop an optimal diet model based on active choice (3) to compare Thinopinus behaviour in the f i e l d with predictions of the mechanistic, frequency-dependent and optimal diet models. The mechanistic model provides a nu l l hypothesis against which selection in general, and the frequency-dependent and optimal diet models in par t i c u l a r can be tested. 5 3 Models The mechanistic model predicts the frequency d i s t r i b u t i o n of n sizes of a prey type in the diet of a predator, as determined by an observer. For a predator which does not use active choice, the proportion of prey items of size i expected in the diet i s given by p(i) = v ( i ) f (i) E v ( i ) f ( i ) i=1 where v ( i ) is the v u l n e r a b i l i t y of size i prey and f ( i ) i s the r e l a t i v e frequency with which size i prey are encountered by the predator. The term in the denominator ensures that n Ep(i)=1 i = 1 The v u l n e r a b i l i t y of size i prey can be determined from the product of c ( i ) , the pr o b a b i l i t y that a size i prey i s captured given that i t i s detected, and the pr o b a b i l i t y that a size i prey i s detected. The probability that a size i prey i s detected is proportional to the area of reaction of a predator. It i s given by K d ( i ) 2 where K is a constant of proportionality which depends on the shape of the reactive f i e l d of the predator, and d(i) i s the maximum reaction distance for a size i prey, measured d i r e c t l y in front of the predator (Holling 1966). Hence v( i ) = K d ( i ) 2 c ( i ) 54 Because of the manner in which I co l l e c t e d data, I must add another term to the model. This i s the probability that an observer w i l l score a feeding event on a size i prey and is proportional to h ( i ) , the handling time, or the time between prey capture and completion of feeding. Then p ( i ) ' = c ( i ) d ( i ) 2 f ( i ) h ( i ) E c ( i ) d ( i ) 2 f ( i ) h ( i ) i= 1 is the proportion of prey items of size i I expect to observe in the diet of a predator which does not use active choice. A suitable index for measuring deviations from the expected proportions i s the standardized forage r a t i o (c) (Chesson 1978). Let r ( i ) be the actual proportion of prey items of size i observed in the d i e t . Then preference for size i prey can be expressed as o = r ( i ) / p ( i ) ' I [ r ( i ) / p ( i ) ' ] i = 1 This index has the advantage that i t varies between 0 and 1 and is independent of prey a v a i l a b i l i t y (Paloheimo 1979). The optimal diet model I develop here i s based on the t o t a l foraging time T(r) required for a predator to reach s a t i a t i o n , when the diet may include prey with sizes of rank 1 through r only. I define j as the rank of a size i prey item, based on some measure of the value of the prey item to the predator. The optimal diet i s the set of sizes which minimize T ( r ) . A rationale for t h i s approach w i l l be presented l a t e r . T(r) includes both searching and handling times. If prey are randomly 55 di s t r i b u t e d , the expected search time for each prey item can be derived from the mean of the exponential d i s t r i b u t i o n with rate parameter X given by X=R-A-C Here R i s the encounter rate, measured in prey items/time. A i s the p r o b a b i l i t y that the encountered prey i s attacked, that i s , that i t has a size of rank j to be included in the optimal diet, so r A = E d ( j ) 2 f ( j ) J ^ J E d ( j ) 2 f ( j ) j=1 C i s the probability that, the encountered prey item i s captured, given that i t has a size of rank j to be included in the diet, so r C = E c ( j ) d ( j ) 2 f ( j ) 2zl E d ( j ) 2 f ( j ) J = 1 The term in the denominator ensures that the maximum value of C is 1 i f c(i)=1 for a l l i . If a mean of N items is eaten and the t o t a l handling time i s H, then T(r) = N/[R-A-C] + H 56 M a t e r i a l s and Methods F i e l d d a t a I c o l l e c t e d f i e l d d a t a between May and August 1980 and May and J u l y 1981. D u r i n g t h e s e months r e g u l a r s e a r c h e s f o r b e e t l e s were c o n d u c t e d (see a l s o page 9 ) . For each s e a r c h I wore a h e a d l i g h t and walked s y s t e m a t i c a l l y i n a s e r i e s of t r a n s e c t s . I s c o r e d time of n i g h t , beach p o s i t i o n , prey t y p e , prey l e n g t h ( f o r amphipods), b e e t l e sex, and b e e t l e l e n g t h f o r each b e e t l e found f e e d i n g . Amphipod l e n g t h was measured from the a n t e r i o r of the head t h r o u g h the t h i r d abdominal segment, a l l o w i n g f o r body c u r v a t u r e (Bowers 1963). As a check on measurement e r r o r of the l e n g t h s of p a r t i a l l y consumed amphipods e s t i m a t e d on the beach, I f e d 20-22 mm amphipods t o b e e t l e s i n the l a b o r a t o r y and compared l e n g t h s measured b e f o r e and a f t e r b e e t l e f e e d i n g . There was a weak tendency t o u n d e r e s t i m a t e amphipod l e n g t h , w i t h a mean d i f f e r e n c e of -0.8±0.6 mm (n=15). I e s t i m a t e d amphipod abundance from the mean o f the number of amphipods caught i n 1 h i n 6 t o 10 p i t f a l l t r a p s p l a c e d a t the h i g h water l e v e l . P i t f a l l t r a p s c o n s i s t e d of p l a s t i c cups (8.5 cm d i a m e t e r , 11 cm deep) se t i n t o the sand and f i l l e d one t h i r d f u l l of seawater ( C r a i g !973a,b). The water p r e v e n t e d a n i m a l s from e s c a p i n g once they f e l l i n t o the t r a p s . T h i s method sampled the r e l a t i v e abundance of a c t i v e amphipods which would pass a b e e t l e s i t t i n g m o t i o n l e s s on the sand. In 1980, t r a p s were s e t one n i g h t each month. In 1981, t r a p s were se t d u r i n g the p e r i o d when searches were c o n d u c t e d . 57 To measure the sizes of amphipods available to beetles, I col l e c t e d amphipods from two p i t f a l l traps on one night each month. These traps were l e f t throughout the period of beetle surface a c t i v i t y . Samples from the two traps were combined and preserved in 5% formalin. I grouped the amphipods into 4 mm size classes: (1) 4-7 mm (2) 8-11 mm (3) 12-15 mm (4) 16-19 mm <5) 20-22 mm. Juveniles were 2 mm in length when f i r s t released from the female brood chamber. However, 4 mm amphipods were the smallest I could observe when caught by beetles. V u l n e r a b i l i t y I measured the reaction distance of beetles to amphipods of dif f e r e n t sizes, and the capture success for detected amphipods in laboratory experiments. To standardize for hunger l e v e l s , I held beetles at laboratory temperatures without food for two to three days prior to these experiments. Experiments were conducted at night in buckets or trays containing a layer of damp sand. The overhead fluorescent l i g h t s were covered with f i l t e r s to reduce li g h t intensity. To measure reaction distance I formed a grid of 0.6 cm2 blocks in the sand. A l i v e amphipod of known size was t i e d on a thread and dragged in a li n e perpendicular to the head of the beetle at approximately the walking speed of the amphipod. Three presentations at a given distance were made, moving towards the beetle at one block intervals u n t i l a response was obtained. I defined reaction distance as the maximum number of blocks at 5 8 which the beetle responded by movement towards the amphipod. Each beetle was tested once only for each amphipod si z e . For capture success experiments, one beetle and a few amphipods of a given size were placed in buckets. I observed the beetle and counted the number of attacks required to capture an amphipod up to a maximum of 10 attacks. I defined an attack as a forward lunge by the beetle which resulted in contact with the amphipod. This was a conservative d e f i n i t i o n as i t did not include misdirected,lunges. I determined the p r o b a b i l i t y of successful capture from the product of (a) the frequency of captures/attack for successful beetles, and (b) the frequency of success within 10 attacks. I tested for the importance of beetle size for 16-19 mm amphipods only, by recording the number of captures made in 20 successive attacks by each beetle. Amphipods were removed manually between captures. Beetle size was estimated from the width of the head capsule, measured with vernier c a l i p e r s . Beetle size in these experiments ranged from 3.9 to 5.6 mm. Head capsule width (W) was a more consistent measure of size than body length (L) used in beach measurements. They are related by the following equations W = -0.04 + 0.23L for males (p<0.00l, n=40) W = 1.57 + 0.13L for females (p<0.00l, n=40) As a control for the use of overhead l i g h t i n g , I repeated the experiments with blinded beetles, whose eyes had been covered with enamel paint. Beetles were permitted to recover for a few hours before these t r i a l s began. To determine capture 59 success, I used 12-15 mm amphipods only and recorded the number of successful captures made by each beetle in 20 attacks. To determine reaction distance, I used 16-19 mm amphipods only, and followed the procedure described above. Feeding experiments I measured feeding rates for beetles covered with a thin layer of damp sand and placed in individual jars. I weighed beetles to the nearest 0.1 mg and allowed them to recover for about one hour pr i o r to experimentation. They were then presented with amphipods or isopods of known siz e . I used amphipods 4-22 mm in length and isopods 8-11 mm in length. Isopods larger than t h i s could not be held by beetles and were always rejected as food items. I recorded handling times to the nearest minute for the time between prey capture and completion of feeding. Beetles were immediately reweighed and gross consumption was estimated from the difference of the two weights. There were no appreciable weight changes in beetles which did not feed. To determine the number of amphipods of a given size required to satiate a beetle, I presented amphipods continuously to beetles u n t i l they refused further food, at a maximum of 6 amphipods. Beetles were reweighed at the end of the experimental period. I could not measure the number of isopods required to satiate a beetle, as beetles tended to reject isopods presented to them. A beetle would grasp the isopod between i t s mandibles 60 and release i t , although i t would readily accept an amphipod. Isopod experiments tested for differences in preference between male and female beetles for isopods and amphipods. The design was similar to the feeding experiments described in chapter 2 (page 12). Each jar contained one beetle, 10 isopods (under 12 mm in length), and 0, 2, 4, or 6 amphipods. Only starved beetles were used and two sizes of amphipods (8-11 mm and 16-19 mm) were of fered. Results V u l n e r a b i l i t y The mean and the variance of the number of attacks required to capture an amphipod increased with increasing amphipod size (Table IV). For the three smallest size classes, a l l beetles tested captured an amphipod in less than 10 attacks (see Table IV, "prop, of beetles" column). For larger amphipods, some beetles were unsuccessful. The pro b a b i l i t y of prey capture/attack decreased with increasing amphipod size (see Table IV, "prob. of capture" column). Beetle size also affected the frequency with which large amphipods were captured. The correlation between the number of 16-19 mm amphipods captured in 20 attacks and beetle size was 0.447 (n=41, p<0.0l). Such relationships between capture success and predator and prey size are t y p i c a l for insects (Evans 1976, G r i f f i t h s 1980). Small amphipods were captured e a s i l y as the entire body of 61 Table IV. Capture success as a function of amphipod s i z e . The table gives the mean and variance of the number of attacks required for capture by beetles which were successful within 10 attacks, the number of beetles tested, the proportion of beetles which were successful, and the combined probability of prey capture/attack. size (mm) size class mean var. n prop, of beetles prob. of capture 4-7 1 1 .53 0.81 30 1 .000 0.652 8-1 1 2 1 .88 1 .75 40 1 .000 0.533 12-15 3 3.25 6.58 32 1 .000 0.308 16-19 4 3.84 8.81 48 0.646 0. 168 20 + 5 4.14 8.98 23 0.304 0.073 the amphipod f i t between the mandibles of the beetle. For larger amphipods, beetles grasped the amphipod with their mandibles and maintained a position on i t s dorsal surface u n t i l the amphipod was subdued. Amphipods responded to capture by repeated flexing of the uropods. Maximum lengths of such escape responses observed in timed bouts were 40 s for an 11 mm amphipod, 80 s for a 14 mm amphipod, and 96 s for an 18 mm amphipod. In the la s t example the amphipod escaped. Large male amphipods also defended themselves with their second (enlarged) gnathopods and their antennae. These could prevent the beetle from grasping the body of the amphipod. The overhead l i g h t i n g used to conduct these experiments had no ef f e c t on capture success. Beetles which had been blinded captured 6.33±0.32 amphipods (n=l5) in 20 attacks, or had a frequency of 0.317 captures/attack for amphipods in size class 3. This value was not s i g n i f i c a n t l y d i f f e r e n t from the observed 6 2 value of 0.308 for beetles which had not been blinded (t=0.56, df=14, p>0.10). Beetles responded to large amphipods at much greater distances than small amphipods (Fig. 9). The responses of blinded beetles were not s i g n i f i c a n t l y different from those of beetles which had not been blinded (t=1 .48, df = 73, df = 73, p>0.10). Feeding rates There were no differences in gross consumption, handling time or feeding rate between male and female beetles feeding on amphipods or isopods ( t - t e s t , p>0.05). I combined results from males and females for the following analyses. Gross consumption, handling time and feeding rates increased over the f i r s t 3 amphipod size classes and then remained constant for the two largest size classes (Table V ) . Beetles required more than twice the amount of time to feed on a large amphipod as they did on a small one. Feeding rates on isopods were only marginally greater than for the smallest amphipods. One large amphipod was s u f f i c i e n t to satiate most beetles. A l l but one beetle tested on each of amphipod size classes 4 and 5 rejected additional amphipods (Table V I ) . Beetles which were fed on a sequence of the smallest amphipods consumed 3.6 amphipods on average. Handling times were also s i g n i f i c a n t l y greater for the smallest size classes, with an o v e r a l l c o r r e l a t i o n of -0.429 (n=96, p<0.0l) between amphipod length and handling time. Handling times for feeding to satiation did not 63 Figure 9. The maximum presentation distance at which beetles responded to amphipods by movement towards the amphipod. Means are given with 95% confidence intervals (n=50). The value for blinded beetles (n=25) i s given for 16-19 mm amphipods (closed c i r c l e ) . 64 65 Table V. Feeding data for beetles fed one amphipod or isopod. The table gives means, standard errors and sample size s . The superscripts define le v e l s which are not s i g n i f i c a n t l y d i f f e r e n t by Duncan's multiple range test. size (mm) size class gross consumption (mg) handling time (min) feeding rate (mg/min) amphipods 4-7 1 8 . 6±1.2 1 (15) 12.0±1.4 1 (14) 0.70±0.081 8-1 1 2 26.8±2.3 2 (18) 20. 7±1.5 2 (17) 1 .38±0.15 2 12-15 3 43.0±2.5 3 (28) 22.812.02 (28) 2. 17±0.16 3 16-19 4 49.6±3.4 3 (25) 28.8±2.4 3 (25) 2.01±0.20 3 20 + 5 48.8±3.8 3 (21) isopods 27.1±2.5 2 3 (21 ) 1.93±0.153 13.7±1.2 (25) 19.3±1 .4 (25) 0.74±0.06 include the search time between successive prey items. The inclusion of search time would i n f l a t e the difference in t o t a l foraging time between small and large amphipods. There were no differences in t o t a l gross consumption for beetles fed to s a t i a t i o n on d i f f e r e n t size classes. Gross consumption was related to i n i t i a l beetle weight (r=0.505, n=96, P<0.01). The correlation between feeding rate and beetle weight was not s i g n i f i c a n t (r=0.l78, n=96, p>0.05). In the experiments testing for preference of beetles between amphipods and isopods, male beetles consumed more isopods than female beetles when no amphipods were available (Fig. 10, open and closed squares, t=2.l9, df=l8, p<0.05), suggesting that males were either more successful in capturing 6 6 Figure 10. The mean ±SE number of isopods eaten at densities of 10 isopods and 0, 2, 4 or 6 large amphipods (16-19mm, closed c i r c l e s , n=20) and small amphipods (8-11mm, open c i r c l e s , n=13). Male and female beetles d i f f e r e d in the number of isopods eaten in the absence of amphipods only (closed or open squares). For other amphipod dens i t i e s , male and female data were combined. 67 2 - 5 2 0 c " o V> T J a is o (A S 10 -O E 0-5 0 • m a l e s • f e m a l e s • large a m p h i p o d s O s m a l l a m p h i p o d s 0 10 2-10 4=10 6=10 R a t i o of amphipods 1 isopods presented 68 Table VI. Feeding data for beetles fed to s a t i a t i o n on one amphipod size c l a s s . The superscripts define levels which are not s i g n i f i c a n t l y d i f f e r e n t by Duncan's multiple range test. 4-7 1 3 .61±0. 34 1 (18) 39. 6±4 • 2 1 46. 3±4 .41 (14) 0 .98±0. 1 3 8-11 2 2 .42±0. 23 2 (21) 48. 6±3 .61 38. 7±3 .4 1 2 (16) 1 .38±0. 16 12-15 3 1 .30±0. 1 1 3 (20) 51 . 2±3 .81 26. 5±2 .53 (20) 2 .19±0. 19 16-19 4 1 .04±0. 04 3 (25) 50. 6±3 .41 30. 4±2 .8 2 3 (25) 1 .99±0. 21 20 + 5 1 .05±0. 05 3 (21 ) 49. 5±3 .81 27. 4±2 .63 (21 ) 1 .93±0. 1 5 isopods, or were less l i k e l y to reject isopods as prey. In the presence of large or small amphipods, only a few beetles ate any isopods (Fig. 10, open and closed c i r c l e s ) . There were no differences in the number of isopods eaten by each sex. The rapid decrease in isopod consumption and corresponding increase in amphipod consumption with increasing amphipod abundance suggested strong preference for amphipods. Beetles observed in feeding experiments would attack an isopod dropped in front of them, and then reject the isopod, even small isopods that they could capture e a s i l y . F i e l d results In 1980, beetles fed almost exclusively on Orchestoidea  c a l i f o r n i a n a . Of 423 observations, only four were on related amphipod species (Orchestoidea pugettensis and Orchestia  traskiana) present on the beach in low numbers. One beetle only size size no. (mm) class eaten gross consumption (mg) handling t ime: (min) feeding rate (mg/min) 69 was f o u n d f e e d i n g on a n i s o p o d , A l l o n i s c u s p e r c o n v e x u s . The p a t t e r n was d i f f e r e n t i n 1981 a n d t h e r e w e r e a l s o d i f f e r e n c e s b e t w e e n m a l e a n d f e m a l e b e e t l e s . A m p h i p o d s c o m p r i s e d 7 3 . 1 % ( n = l 7 5 ) o f t h e d i e t o f m a l e s , c o m p a r e d t o 9 1 . 4 % ( n=336) f o r f e m a l e s ( X 2 = 2 8 . 8 , d f = 1 , p < 0 . 0 l ) . M a i n l y i s o p o d s c o m p r i s e d t h e r e m a i n d e r o f t h e d i e t . I n a d d i t i o n i n 1 9 8 1 , t h e r e w e r e f o u r o b s e r v a t i o n s o f m a l e s f e e d i n g on D y s c h i r i u s o b e s u s L e C . ( C a r a b i d a e ) , a n d two o b s e r v a t i o n s on E m p h y a s t e s f u c i c o l a M a n n . ( C u r c u l i o n i d a e ) , b o t h s p e c i e s a b o u t 2 mm i n l e n g t h . F o r f e m a l e s , t h e r e w e r e t h r e e o b s e r v a t i o n s o f f e e d i n g on l a r v a l T h i n o p i n u s  p i c t u s . T h e s e w e r e t h e o n l y t h r e e l a r v a e o b s e r v e d on t h e s u r f a c e a t n i g h t i n 1 9 8 1 , a n d t h e y h a d p r o b a b l y b e e n f o r c e d o u t o f t h e i r b u r r o w s i n t h e s a n d n e a r HW by a n i n c o m i n g t i d e . I n 1 9 8 1 , 6 7 % ( n = 1 1 4 ) o f t h e f e e d i n g o b s e r v a t i o n s on a m p h i p o d s a n d i s o p o d s f o r m a l e b e e t l e s , a n d 7 1 % ( n=236) f o r f e m a l e b e e t l e s , w e r e on t h e l o w e r b e a c h ( X 2 = 0 . 7 6 , d f = 1 , P>0.10). H o w e v e r , m a l e s on t h e u p p e r b e a c h w e r e f o u n d f e e d i n g on i s o p o d s 4 2 % ( n=55) o f t h e t i m e , c o m p a r e d t o 1 6 % (n = 1 1 4 ) on t h e l o w e r b e a c h ( X 2 = 1 2 . 3 , d f = 1 , p < 0 . 0 l ) . T h i s was p r o b a b l y b e c a u s e i s o p o d s w e r e r e l a t i v e l y more a b u n d a n t on t h e u p p e r b e a c h . F e m a l e b e e t l e s s h o w e d t h e same t r e n d b u t t h e d i f f e r e n c e , 1 2 % ( n=97) c o m p a r e d t o 6% ( n = 2 3 6 ) , was weak ( X 2 = 3 . 1 2 , d f = 1 , p < 0 . l O ) . The r e l a t i v e a b u n d a n c e o f d i f f e r e n t s i z e s o f a m p h i p o d s c h a n g e d a s t h e summer p r o g r e s s e d ( F i g . 1 1 ).. The l a r g e s t a m p h i p o d s d i s a p p e a r e d a f t e r M a y , a n d j u v e n i l e s w e r e r e c r u i t e d t o t h e p o p u l a t i o n t h r o u g h o u t J u n e a n d J u l y . T h e r e w e r e a l s o y e a r t o y e a r d i f f e r e n c e s . The m o d a l s i z e was l a r g e r i n 1981 t h a n i n 70 Figure 11. Size-frequency d i s t r i b u t i o n s of amphipods caught in p i t f a l l traps for each month in 1980 (A) and 1981 (B) . 71 1 9 8 0 0 15 010 005 020 015 >_ 010 2 005 UJ Z> o or 020 u_ 0 15 010 -l 005 015 010 005 6 MAY n=437 19 JUNE n = 379 13 JULY n= 302 18 AUGUST n = 173 2 4 6 8 10 12 14 16 18 20 22 AMPHIPOD SIZE (mm) 1 9 8 1 16 MAY n = 68 0 2 4 6 8 10 12 14 16 18 20 22 24 26 AMPHIPOD SIZE (mm) 7 3 1 9 8 0 . The mean length of captured amphipods was also greater in 1 9 8 1 ( 1 4 . 0 ± 0 . 1 2 mm) than in 1 9 8 0 ( 1 2 . 3 ± 0 . 2 mm, t = 1 1 . 5 , d f = 8 0 2 , p < 0 . 0 0 1 ) . Beetle length was also greater in 1 9 8 1 ( 2 1 . 9 ± 0 . 1 mm) than in 1 9 8 0 ( 2 1 . 6 ± 0 . 1 mm, . t = 3 . 2 0 , d f = 8 0 2 , p < 0 . 0 l ) . Beetle length and the length of the captured amphipod were s i g n i f i c a n t l y correlated in both years ( r = 0 . 3 5 6 , n = 4 l 9 , p < 0 . 0 l for 1 9 8 0 ; r = 0 . 1 9 0 , n = 3 8 5 , p < 0 . 0 l for 1 9 8 1 ) . Male and female beetles did not d i f f e r in mean body weight ( 0 . 2 3 7 ± 0 . 0 0 9 g, n= 2 5 for males, 0 . 2 4 2 ± 0 . 0 0 8 g, n=27 for females), but males had greater head widths than females of the same weight (Fig. 1 2 ) . As head width was related to how far apart mandibles could be spread (Fig. 1 3 ) , t h i s could affect the maximum size of prey captured. In beach observations, the mean and variance of the size of amphipods on which males and females were observed feeding did not d i f f e r ( t = 0 . 8 , d f = 4 l 7 , p > 0 . l 0 for 1 9 8 0 ; t = 0 . 0 , d f = 3 8 3 , P > 0 . 1 0 for 1 9 8 1 ) . However, beetles found feeding on amphipods were s i g n i f i c a n t l y larger than beetles feeding on isopods for both males ( t = 2 . 2 9 , d f = 1 4 2 , p < 0 . 0 5 ) and females ( t = 2 . 5 0 , d f = 3 0 4 , p < 0 . 0 5 ) . The frequency with which scavengers joined beetles in feeding was related to the size of amphipod captured (Fig. 1 4 , X 2 = 1 3 . 4 , d f = 4 , p < 0 . 0 l ) . Beetles feeding on larger amphipods were more l i k e l y to a t t r a c t various scavengers, mainly an unidentified staphylinid and Dyschirius obesus, a carabid (see page 6 ) , and there was more space for attachment of scavengers on larger carcasses. In 1 9 8 0 , Thinopinus was joined in feeding in 7% of the observations. This figure was greater in 1 9 8 1 with 74 Figure 12. Morphological comparisons of a sample of male and female beetles co l l e c t e d 4 June 1981. Weights were measured 9 h after c o l l e c t i o n . The slopes, but not the intercepts of the regression l i n e s are s i g n i f i c a n t l y d i f f e r e n t (p<0.05). Regressions are y = 3.06 + 8.23x for males, r2=0.853 y = 3.26 + 5.87x for females, r2=0.751 75 0 . 1 6 0 . 2 2 0.2 8 0 . 3 4 W e i g h t (g) 76 Figure 13. Maximum mandible spread for male and female beetles (r=0.937, n=56). Measurements were made on relaxed beetles using vernier c a l i p e r s . 77 9.4-J • m a l e s O f e m a l e s 9.0-1 E E 8.6 0 Q- 7 . 8 T CO CD , — 7.4-O O • • O O O O O O O O O 7.<H CO o o 6.64 O O 6.2 1 ^ • i 3 . 6 4.0 4.4 4.8 5.2 5.6 H e a d w i d t h (mm) 78 Figure 14. The frequency d i s t r i b u t i o n of the sizes of captured amphipods in 1981 when scavengers were present or absent. S C A V E N G E R S P R E S E N T ( n = ! 0 4 ) LU 3 0 - 4 1 Cf LU QC ^ 0 Z \ 1^ S C A V E N G E R S A B S E N T ( n = 2 8 l ) 4 - 7 8 - 1 1 1 2 - 1 5 1 6 - 1 9 AMPHIPOD SIZE (mm) 2 0 + 80 Figure 15. The observed frequency d i s t r i b u t i o n of amphipod size classes in the population obtained by weighted averages of p i t f a l l trap catches from each month (white bars), the predicted d i s t r i b u t i o n of amphipods captured according to the mechanistic model (striped bars), the observed d i s t r i b u t i o n of captures (black bars), and the standardized forage ratios (inset, see page 54) for both years of data. The dotted l i n e gives the expected forage ratios in the absence of preference. 82 scavengers present in 27% of the observations (X2=57, df=1, P<0.001). Only 5% of the beetles feeding on isopods in 1981 were joined by scavengers. A test for prey selection I obtained an average d i s t r i b u t i o n for the sizes of amphipods available to beetles in each year by weighting the frequency of each amphipod size in each month by the number of feeding observations made on beetles in that month (Fig. 15, white bars). The observed d i s t r i b u t i o n s of the sizes of captured amphipods (Fig. 15, black bars) d i f f e r e d from the sizes available for both years (X2=376, df=4 for 1980, X2=128 for 1981 P<0.001). Beetles did not capture sizes of amphipods in proportion to the r e l a t i v e abundance of those sizes. I used the model described above to test whether th i s resulted from differences in the r e l a t i v e a v a i l a b i l i t y of di f f e r e n t sizes to beetles, or by active selection by the beetles. Mechanistic model The mechanistic model was based on p ( i ) ' , the proportion of amphipods of size i , I expected to observe in the diet of a beetle i f the beetle did not use active selection. To predict p ( i ) ' , I combined data on the sizes of amphipods available (Fig. 15, white bars) with laboratory values for v u l n e r a b i l i t y and handling times (Fig. 9, Table IV, Table V). Predicted values of p ( i ) ' d i f f e r e d between the two years (Fig. 15, striped bars). 8 3 The observed d i s t r i b u t i o n s of the sizes of captured amphipods (Fig. 15, black bars) also d i f f e r e d between years (X2=34.8 df=4, P<0.001), but showed similar trends in both years and d i f f e r e d f r o m p ( i ) ' (X2=540, for 1980; X2=85 for 1981, df=4, p<0.00l). The pattern of food selection i s seen most c l e a r l y by examining the standardized forage ratios (Fig. 15, ins e t ) . With five prey size classes, a size class which i s neither selected p r e f e r e n t i a l l y nor avoided has a r a t i o of 0.20. In both years beetles appeared ind i f f e r e n t to the smallest size class, avoided at least one of the middle size classes, and showed strong preference for the largest size c l a s s . The apparent preference for large amphipods could result i f I underestimated capture success for these sizes. To test this p o s s i b i l i t y , I recalculated p ( i ) ' using data on the prob a b i l i t y of capture/attack from beetles which were successful only. This would overestimate the true p r o b a b i l i t y . As the numbers of amphipods captured in the largest two size classes were underestimated by this c a l c u l a t i o n , measurement error was s t i l l i n s u f f i c i e n t to account for the preference observed. The new di s t r i b u t i o n s of p ( i ) ' did however conform more cl o s e l y to the observed d i s t r i b u t i o n s (X2=239 p<0.00l, for 1980; X2=14, p<0.0l for 1981). It i s possible to choose parameter values which force agreement between p ( i ) ' and the observed d i s t r i b u t i o n s . Values which result in an exact f i t for a change in one parameter are given in Table VII. For example, very high values for d(i) or h(i) for size class 5, r e l a t i v e to the other size classes, could 84 improve the f i t of the model. None of the values for parameters l i s t e d in Table VII were within the range of values obtained from experiments. However, i t i s s t i l l possible that errors in different. parameters could combine to produce a spurious but s i g n i f i c a n t difference between p ( i ) ' and the observed d i s t r i b u t i o n s . Frequency-dependent model The frequency-dependent model could also be tested by a comparison o f p ( i ) ' and the observed d i s t r i b u t i o n of the sizes of captured amphipods. According to the predictions of this model, the most abundant size classes should be over-represented in the diet r e l a t i v e to th e i r a v a i l a b i l i t y , and conversely for the least abundant size classes. The data showed the opposite trends. Preference appeared to be strongest for the largest and least abundant size class. Optimal diet model For most optimal diet models, the p r o f i t a b i l i t y of each prey type is defined as the r a t i o of net energy intake to handling time. I approximated th i s r a t i o by the feeding rate for each size class and could not distinguish between the largest 3 size classes. This simple form of the model was c l e a r l y inappropriate, as I measured differences in selection between size classes 3 and 4. An alternative method I chose was to rank prey items based on the number of amphipods of a given size 85 Table VII. Parameter values for f ( i ) , h ( i ) , d(i) and c ( i ) which result in an exact f i t of the observed d i s t r i b u t i o n of the sizes of captured amphipods to p ( i ) ' . These values were calculated for a change in the l i s t e d parameter only. Other parameters used in cal c u l a t i o n of p ( i ) ' were those obtained from the experiments. Values for h ( i ) , d(i) and c ( i ) were standardized to a maximum value of 1. Values are shown for both 1980 and 1981 data. i f (i) h(i) d(i) c (i) 1980 data 1 0.3995 0.1646 0.0333 1 . 0000 2 0.1038 0. 1147 0.0387 0. 3304 3 0. 1764 0.0429 0.0245 0. 0648 4 0.1921 0.2323 0.1741 0. 1510 5 0.1281 1.0000 1 .0000 0. 3009 1 981 data 1 0.1663 0.1715 0.0347 1. 0000 2 0.0870 0.1259 0.0425 0. 3478 3 0.1960 0.2786 0.1581 0. 4009 4 0.3597 0.4965 0.3722 0. 3095 5 0.1910 1.0000 1.0000 0. 2888 class required to satiate a beetle. This was because an increase in the number of items eaten would increase the t o t a l search time required. I could s t i l l not distinguish between size classes 4 and 5 however, and these size classes shared a rank of one. I calculated the expected foraging times, T ( r ) , for beetles feeding to sat i a t i o n on di f f e r e n t sets of size classes as described on page 54. When beetles fed on mixtures of size classes I used the data in Table VI to estimate the mean numbers 86 of items eaten and the handling times (Table VIII). I combined Table VIII. Parameter values for the optimal diet model. The table gives a l l possible combinations of size classes of prey items in the die t , and the mean handling time h ( i ) , and the mean number of items eaten (N) for each combination. Combinations are l i s t e d as the number of the size class in the order of capture for a maximum of four-items eaten. order of capture 2 3 4 handling t ime(min) no. eaten 1 1 1 1-2 42.2 3.6 1 1 3-5 39.4 3.6 1 1 2 42.4 3.0 1 1 3-5 36.6 3.0 1 2 1 -2 42.2 3.0 1 2 3-5 38.0 3.0 1 3-5 32.5 2.0 2 1 1-2 42.2 3.0 2 1 3-5 38.0 3.0 2 2 1-2 42.2 2.4 2 2 2-5 39.4 2.4 2 3-5 33.8 2.0 3 1-2 32.5 1 .3 3 3-5 28.3 1 .3 4-5 28.3 1.0 data on handling times for size classes 1 and 2 and for size classes 3, 4 and 5 where differences were not s i g n i f i c a n t . When feeding occurred on mixtures of size classes with d i s t i n c t 87 handling times, I used an average feeding time, weighted by the proportion of the t o t a l food intake from the d i f f e r e n t size classes. One prediction of optimal diet models, i s that only the subset of sizes which minimize foraging time are included in the optimal d i e t . Using 1980 data, T(r) was minimized (Table IX) when a l l size classes were included in the diet for encounter rates less than 60 amphipods/h, and for encounter rates less than 68 amphipods/h for 1981 data. At higher encounter rates, foraging time was minimized by excluding the smallest two size classes from the di e t . The long foraging times for a beetle which fed on the two largest size classes only, resulted because these size classes were less abundant and required a longer search time. To test predictions from t h i s model with the f i e l d data, I used the numbers of amphipods caught in p i t f a l l traps on the beach as an estimate of the amphipod encounter rate. The size of the trap opening approximated the area of reaction of a beetle. For 1981 data, the maximum encounter rate measured was 30 amphipods/h. According to Table IX, this rate was well below the threshold at which selection should occur. This result had two important implications. F i r s t , I could not dis t i n g u i s h the optimal diet prediction from the n u l l hypothesis. A good f i t to the model could result i f beetles were foraging either optimally or simply according to prey a v a i l a b i l i t y . Second, for the range of encounter rates observed, foraging time differences between most strategies were within one standard error of experimental 88 Table IX. Foraging times (min) generated by the optimal diet model for beetles feeding to s a t i a t i o n on a l l combinations of size classes. (A) the generalist (B) size classes 2-5 only (C) size classes 3-5 only (D) size classes 4-5 only. Encounter rates are in amphipods/h. Calculations are shown based on amphipod size d i s t r i b u t i o n s for 1980 and 1981. * indicates the lowest value. encounter rate A B C D 1980 data 1 258.4* 308.7 358.2 2187.7 10 52.8* 57.4 61 .2 244.2 20 41.4* 43.4 44.8 1 36.2 30 37.6* 38.8 39.3 100.2 40 35.7* 36.4 36.5 82.2 60 33.8 34. 1 33.8* 64.3 1 00 32.3 32.2 31.6* 49.9 1981 1 data 291.1 * 311.7 431 .4 754.6 10 56.4* 58.3 68.6 100.9 20 43.4* 44.2 48.4 64.6 30 39.0* 39.5 41 .7 52.5 40 36.9* 37.2 38.3 46.4 60 34.7* 34.9 35.0 40.4 100 32.9 33.0 32.3* 35.5 values for mean handling time. The only testable prediction from Table IX was that beetles would not specialize on size classes 4 and 5. In fact, the proportion of captured amphipods in size classes 4 and 5 increased with amphipod encounter rate for May and June 1981 (Spearman rank c o r r e l a t i o n c o e f f i c i e n t = 0.410, n=23, p<0.05). This was not true for July 1981 (Spearman rank 8 9 correlation c o e f f i c i e n t = 0.038, n=25, p>0.l0), probably because small amphipods were r e l a t i v e l y more abundant in July. It i s possible that I underestimated encounter rates. This would be especially true i f prey were clumped. However, beetles I observed continuously in 1979 and 1980 had a mean encounter rate of 8.8 amphipods/h (page 33). The feeding rate on isopods was similar to the feeding rate on the smallest size class of amphipods (Table V). By the c r i t e r i a defined above, isopods and size class 1 amphipods share the same rank and isopods should be eaten as encountered whenever the smallest size class of amphipod i s included in the optimal d i e t . For amphipod abundances less than 10 amphipods trapped/h in 1981, 18% (n=135) of the feeding observations were on isopods. Above th i s l e v e l only 6% of the feeding observations were on isopods (X2=10.7, df=1, p<0.0l). Di scussion I have shown that the d i s t r i b u t i o n of sizes of amphipods in the diet of Thinopinus d i f f e r e d s i g n i f i c a n t l y from a n u l l hypothesis model based on d i f f e r e n t i a l v u l n e r a b i l i t y . However, the pattern of prey size selection could not be accounted for by either frequency-dependent or optimal diet models. In order to reject the n u l l hypothesis model, I must assume that the model is both complete and that the parameters were corre c t l y evaluated. In par t i c u l a r I must assume that the parameters measured in the laboratory accurately r e f l e c t behaviour in the f i e l d . A combination of estimation errors in any of the 9 0 parameters could result in an apparent preference or indifference for some size classes. One potential source of error i s in the assessment of the sizes of amphipods ava i l a b l e . This is because the predictions were based on means. I can only state that on average, beetles did not perform as predicted. In chapter 2, I argued that l o c a l s p a t i a l and temporal variations in the sizes of amphipods did occur and were important to the success of foraging beetles. I did not monitor these beetles continuously and I did not know the actual densities or sizes of amphipods encounted by each beetle. In pa r t i c u l a r , the sizes of amphipods caught in p i t f a l l traps did not sample the sizes of amphipods encountered by a beetle in active mode (see page 33). I describe a more precise test of preference between size classes 2 and 4 in laboratory experiments in Chapter 4. I was able to measure weak preference only for size class 4 by starved beetles, and I could not measure any preference of beetles which had not been starved. This suggested that beetles were predominantly opportunistic, and that measured deviations from the mechanistic model were s t a t i s t i c a l l y , but only weakly b i o l o g i c a l l y , s i g n i f i c a n t . Beetles which I observed on the beach did not always attack amphipods near them. This was true for large as well as small amphipods. I did not observe beetles to capture and then reject amphipods of any size. In order for a beetle to a c t i v e l y select certain sizes of prey, i t must have the a b i l i t y to distinguish prey size. The a b i l i t y to distinguish prey size has been well documented in 91 vertebrates (salamanders, Jaeger and Bernard 1981; f i s h , Gardner 1981; birds, Zach 1978, Goss-Custard 1977; shrews, Barnard and Brown 1981). There i s also some evidence for invertebrates (crabs, Elner and Hughes 1978; ants, Davidson 1978). Brownell and Farley (1979) demonstrated that the desert scorpion Paruroctonus mesaensis detects prey through substrate-borne vi b r a t i o n s . Thinopinus may use a similar mechanism. There were no differences in the reaction distance of blind and sighted beetles. Such sand-borne vibrations could also provide information on prey si z e . This mechanism would work most e f f e c t i v e l y i f an amphipod walked toward a beetle. Frequently amphipods jumped. Beetles rapidly attacked amphipods which landed near them. Hence, the potential for error in judging prey size was high. Mistakes in judging prey p r o f i t a b i l i t y are frequently c i t e d as an explanation for the discrepancies observed in tests of optimal diet predictions (e.g. Elner and Hughes 1979, Jaeger and Barnard 1981). There were advantages to feeding on small amphipods which I did not account for in measuring p r o f i t a b i l i t y of d i f f e r e n t sizes of amphipods for the optimal diet model. Small amphipods tended to be clumped and several could be captured in one area in a short time i n t e r v a l . Large amphipods were more costly to capture, as a struggle frequently ensued between the amphipod and beetle. A beetle feeding on a large item also attracted a variety of scavengers. This led to struggles and a loss of a portion of the prey item, with associated time and energetic costs. 9 2 Why did the optimal diet model not work? This is a case in which I measured a l l the necessary values and did everything "correctly". Yet, the optimal diet predictions were i d e n t i c a l to the n u l l hypothesis. Pastorok (1981) reported results similar to those I have presented here. He constructed an optimal diet model for Chaoborus feeding on d i f f e r e n t sizes of Daphnia which was based on d i f f e r e n t i a l v u l n e r a b i l i t y . The model predicted that Chaoborus should include a l l sizes of Daphnia in i t s diet for the range of abundances he measured in the f i e l d . This was, in fact, the pattern observed. In general, attempts to measure size selection of prey by invertebrate predators in the f i e l d have been unsuccessful (Thompson 1978b, G r i f f i t h s 1980, Murtaugh 1981). The range of sizes eaten appears largely to depend on l i m i t s set by the mechanics of prey capture, and on d i f f e r e n t i a l v u l n e r a b i l i t y of d i f f e r e n t sizes. G r i f f i t h s (1981) showed that hungry ant-lions would attack small prey for which the energetic costs of capture exceeded the benefits. For predators such as ant-lions, Chaoborus and Thinopinus, which encounter large variations in prey size and abundance, a generalist foraging strategy may be the best rule. Morse and F r i t z (1982) have suggested that i t i s more important for crab spiders to locate good foraging s i t e s than to s p e c i a l i z e on certain prey while at a s i t e . 93 Isopods versus amphipods Both frequency-dependent and optimal diet models q u a l i t a t i v e l y predicted the types of prey eaten. The feeding rate on isopods was lower than the rate on a l l but the smallest amphipods. Beetles included isopods in their diet at low amphipod abundance only and the relationship with abundance appeared to follow a threshold as predicted by the model. In laboratory experiments I showed that amphipods were highly preferred to isopods. Beetles may be more successful in distinguishing between amphipods and isopods than between di f f e r e n t sizes of amphipods. Beetles would pick up and drop isopods they could easily capture. There were also differences between sexes in the frequency with which isopods were included in the d i e t . Presumably, male beetles were making some kind of "trade-off" between searching for mates and searching for food. E f f e c t s of other trade-offs such as between foraging and t e r r i t o r i a l defence in salamanders (Jaeger et a l . 1981) and birds (Kacelnik et a l . 1981), or foraging and predator avoidance in f i s h ( M i l i n s k i and Heller 1978), have also been shown to reduce attack rates and decrease diet s p e c i a l i z a t i o n . Selander (1966) proposed that sexual differences in foraging evolved to reduce intersexual food competition, and that morphological differences in the food-getting apparatus between males and females re f l e c t e d t h i s competition. Males do have larger mandibles and may be better than females in capturing isopods. However, f i e l d data indicated that i t was the 94 smaller beetles of each sex which were feeding on isopods. The small isopods used in laboratory experiments could easily have been captured by a l l beetles. Morphological differences in mandible size probably resulted from the use of mandibles by male beetles to capture female beetles as well as prey items. Females tended to r e s i s t any mating attempt. More than one male sometimes attempted to mate with a female at a given time, leading to male-male f i g h t s . Larger males may be more successful in such fights as has been shown for the milkweed beetle (McCauley 1982) and for the dung f l y (Borgia 1980). The difference in prey selection between sexes must be primarily a behavioral response on the part of the beetles rather than a morphological one, and could be accounted for by learning. Male beetles spent more time on the upper beach where isopods were more abundant than did females, and would encounter isopods more frequently. Female beetles may not recognize isopods as prey items. I have shown that beetles are selective in the sizes and types of prey included in their d i e t . This selection resulted both from d i f f e r e n t i a l v u l n e r a b i l i t y and from active choice of amphipods over isopods and of larger amphipod sizes. Future studies of foraging should place more emphasis on how invertebrate predators di s t i n g u i s h prey from non-prey items, and how learning and memory affect t h i s process. 95 CHAPTER 4: HUNGER AND OPTIMAL DIET Introduct ion Models of optimal diet (reviewed by Pyke et a_l. (1977)) assume that predators behave so as to maximize their rate of net energy intake. By knowing the energy values e ( i ) , handling times h ( i ) , r e l a t i v e frequencies f ( i ) and overall encounter rate R of i potential prey types, i t is possible to predict which prey should be incorporated into the optimal diet (Pulliam 1974, Charnov 1976a). These prey should always be eaten, and excluded prey never eaten when encountered - the "always or never" rule. Deviations from the always or never rule t y p i c a l l y occur however in experimental tests of the model predictions. They have been explained by the f a i l u r e of the model to account for other constraints on the foragers such as the need for sampling (Krebs et a l . 1978, Davidson 1978, Heinrich 1976), predator avoidance (Milinski and Heller 1978), prey recognition time (Elner and Hughes 1978), nutrient balance (Pulliam 1975, Westoby 1978), and the random nature of encountering prey (Pulliam 1974). The optimal diet model does not allow for changes in the degree of predator hunger. Increasing hunger i s known to a f f e c t the predation process by increasing feeding rate (Ernsting 1977, Beukema 1968, McCleery 1977), reaction distance (Holling 1966), predator a c t i v i t y (Beukema 1968, Calow 1974), prey use (Haynes and Sisojevic 1966, Johnson et a l . 1975), and size range of prey eaten (Heatwole and Heatwole 1968, K i s l a l i o g l u and Gibson 1976). Some of these results may be interpreted in terms of the optimal 9 6 d i e t model i f r e c o g n i t i o n or assessment of food a v a i l i a b i l i t y by pr e d a t o r s i s modified by hunger l e v e l s . Then as predators approach s a t i a t i o n , they should behave as i f prey were abundant and become more s e l e c t i v e (Schoener 1971, P u l l i a m 1974, Charnov 1976a). Here I d e s c r i b e a two prey model which p r e d i c t s the oppo s i t e r e s u l t . I show that under c e r t a i n c o n d i t i o n s , p r e d a t o r s should expand t h e i r d i e t near s a t i a t i o n to i n c l u d e lower value prey. A predator using t h i s r u l e i s what H e l l e r (1980) has termed an expanding s p e c i a l i s t . I then d e s c r i b e an experiment to t e s t whether one predator, the b e e t l e Thinopinus p i c t u s , changes i t s d i e t near s a t i a t i o n i n the d i r e c t i o n p r e d i c t e d by the model. The model Consider a time minimizer (Schoener 1971), a predator which attempts to minimize the t o t a l f o r a g i n g time (T) necessary to o b t a i n a f i x e d food requirement. I make i d e n t i c a l assumptions to the optimal d i e t model with one a d d i t i o n , that the predator can assess the food value D, r e q u i r e d from prey to make up i t s food d e f i c i t . Given the evidence c i t e d above f o r the e f f e c t of hunger on the p r e d a t i o n process, t h i s assumption i s not unreasonable. Maximizing the rate of energy intake over the e n t i r e f o r a g i n g p e r i o d i s e q u i v a l e n t to minimizing T. The problem i s to determine which prey types to i n c l u d e i n the optimal d i e t . For prey types A and B, l e t A be the more p r o f i t a b l e prey, such that e(a)>e(b), and e(a)/h(a) > e ( b ) / h ( b ) . Then there are three reasonable a l t e r n a t i v e s : (1) feed on type A prey only ( s p e c i a l i s t ) 9 7 ( 2 ) feed on type A and B prey as encountered (generalist) ( 3 ) feed on type A prey i f D i s less than a given value, otherwise feed on type A and B prey as encountered (expanding s p e c i a l i s t ) . The model I present here predicts the mean and variance of the t o t a l foraging time for a predator following one of the above rules. F i r s t consider the case where D<e(a). The predator w i l l search for and feed on at most one type A prey, or at most n+1 type B prey. Here n>0 is an integer such that D = ne(b) + z, 0<z<e(b) For the s p e c i a l i s t , expected handling time E(H ) is the S food requirement divided by the intake rate while feeding on the item. I assume that intake rate, and hence handling time, has zero variance. Expected search time E(S ) and i t s variance V(S ) S S can be derived from the exponential d i s t r i b u t i o n with parameter Rf(a). Here R i s the rate at which prey items are encountered (assumed constant), and f(a) is the p r o b a b i l i t y that the encountered item i s a type A prey. Hence f(a) + f(b) = 1 Then expected foraging time i s E(T ) = E(H ) + E(S ) S S S = Dh(a) + 1 iTaT Rf (a) with variance 98 V(T ) = l/[Rf(a)V S For the generalist, the number of items eaten w i l l depend on the order in which type A and B prey are encountered. Expected foraging time i s E(T ) = E(H ) + E(S ) G G G n+1 = Dh(a) + [ h(b)-h(a) ][D - n e(b)]f(b) elaT e l b l iTaT + [ h(b)-h(a) ] e ( b ) [ l - f ( b ) n ] f ( b ) eTBT eTaT ITaT n+1 + [1~f(b) ] Rf (a) Derivations of E(H ) and E(S ) are given in Appendix A and G G B respectively. The f i r s t term is the handling time for the s p e c i a l i s t . The second term i s the extra handling time incurred by the generalist for feeding on an amount z from a type B prey. The t h i r d term i s the extra handling incurred by the generalist for feeding on type B prey before encountering a type A prey. The fourth term is the search time. Expected handling time is greater for the generalist, while expected search time is greater for the s p e c i a l i s t . The generalist w i l l have the lower foraging time when E(T ) < E(T ) G S or when n -n [ h(b)-h(a)]{ z + e ( b ) [ l - f ( b ) ]f(b) } < 1 eTBT eTaT fTaT Rf (a) 9 9 The extra handling incurred by the generalist must be less than the s p e c i a l i s t search time. Foraging time variance for the generalist i s given by V(T ) = V(H ) + V(S ) G G G where V(H ) and V(S ) are derived in Appendix A and B G S respectively. Search time variance i s always greater for the s p e c i a l i s t . Hence the generalist w i l l have a lower foraging time variance i f V(H ) < V(S ) - V(S ) G S G F i g . 16 shows expected foraging times as a function of food requirement. For certain ranges of parameter values, the generalist reaches satiation more quickly than the s p e c i a l i s t . Total foraging time increases with D. Increases are discontinuous at the point where D i s an exact multiple of e(b), that i s , when an additional prey item i s required. The discontinuity represents the time spent in searching between successive prey items when foraging time increases, but food intake does not. An increase in e(b), or a decrease in h(b)/e(b) or f ( a ) , extends the range of values of D for which the generalist has the lower t o t a l foraging time. Foraging time variance i s independent of D for the s p e c i a l i s t and increases with D for the generalist, with small d i s c o n t i n u i t i e s where an additional prey item is required. Next consider the general case in which D can take any value. The mean and variance of foraging time can be derived from the gamma d i s t r i b u t i o n as for the r e s t r i c t e d case. The 100 Figure 16. Expected t o t a l foraging time for the generalist ( s o l i d l i n e ) , and s p e c i a l i s t (dotted line) as a function of the food requirement, D. Error bars represent one standard deviation. Parameter values are R=0.65, e(a)=1 and h(a)/e(a)=20. Other values are given in the f igure. T O T A L FORAGING TIME 5 8 8 P 5 8 8 $ TOT 102 s p e c i a l i s t w i l l search for and feed on at most m+1 type A prey. Here m>0 i s a n integer such that D = me(a) + y, 0<y^e(a) Then E(T ) = m+1 + Dh(a) S RfTa) elaT and V(T ) = (m+1) S T R f T a T T 2 Hence foraging time variance for the s p e c i a l i s t is a step function o f D. The formula for the generalist i s complex, but i t can be shown by simulation that for certain parameter values the generalist w i l l have a lower foraging time than the s p e c i a l i s t . The interesting case however, is for the expanding s p e c i a l i s t , the predator which switches from s p e c i a l i s t to generalist after i t has consumed m type A prey. That i s , at most one type A prey is required for s a t i a t i o n . The mean and variance for t o t a l foraging time for the expanding s p e c i a l i s t c a n be obtained from the sum of the mean and variance o f foraging time for a s p e c i a l i s t feeding on m type A prey and for a generalist feeding on D-me(a) energy units o f type A and B prey. This i s shown in Fi g . 17 which is analogous to F i g . 16b. I f the generalist reaches s a t i a t i o n more quickly than the s p e c i a l i s t when D<e(a), then the expanding s p e c i a l i s t must have a lower foraging time for the same range o f values when D>me(a). 103 Figure 17. Expected t o t a l foraging time for the generalist or expanding s p e c i a l i s t ( s o l i d l i n e s ) and s p e c i a l i s t (dashed line) as a function of the food requirement, D. Error bars represent one standard deviation. Parameter values are R=0.65, e(a)=1, h(a)/e(a)=20, h(b)/e(b)=40, e(a)=1, and e(b)=0.25. 105 A test with Thinopinus F i e l d data on the predator Thinopinus pictus suggested that Thinopinus preferred large sizes of the amphipod Orchestoidea  c a l i f o r n i a n a as. prey items. I designed a laboratory experiment to test whether th i s preference was affected by hunger. Short-term laboratory experiments had shown that on average 2.4 small (8-11 mm) amphipods or 1.0 large (16-19 mm) amphipods were required to satiate a beetle. Mean feeding rates were 1.38 mg/min and 1.99 mg/min on small and large amphipods respectively (Table VII). These values approximately corresponded to those used in F i g . 16a. Qualitative predictions based on the model were (1) starved beetles prefer large over small amphipods (2) beetles near sat i a t i o n show no preference. I used densities of 10 small amphipods and 2, 4, 6, 8 or 10 large amphipods. This p a r t i c u l a r design was chosen so that I could also test the prediction from frequency-dependent and a l l optimal diet models of prey selection that (3) preference for large amphipods increases with the density of large amphipods. To obtain two hunger levels I either held beetles without food for 3 d (starvation treatment) or fed them the night prior to the experiment (s a t i a t i o n treatment). For each treatment, 12 male and 12 female beetles were each placed in a glass jar (8 cm diameter, 10 cm deep) containing a 3 cm layer of damp sand and large and small amphipods. Because of t h i s sand layer, a variable number of amphipods were active on the sand surface at any given time, with the remainder of the amphipods in burrows. 106 B e e t l e s a l s o b u r r o w e d a f t e r f e e d i n g , s o t h e y d i d n o t e n c o u n t e r a m p h i p o d s c o n t i n u o u s l y . J a r s w e r e c o v e r e d t o p r e v e n t a m p h i p o d e s c a p e a n d w e r e l e f t o v e r n i g h t f o r 2 0 - 2 2 h u n d e r n a t u r a l p h o t o p e r i o d a t l a b o r a t o r y t e m p e r a t u r e s ( 1 6 - 1 9 ° C ) . I t h e n c o u n t e d t h e number o f l i v e a m p h i p o d s i n e a c h j a r a n d d e t e r m i n e d t h e number e a t e n by i n f e r e n c e . I n c o n t r o l j a r s w i t h o u t b e e t l e s I r e c o v e r e d a l l l a r g e a m p h i p o d s a n d a mean o f 9 . 3 ± 0 . 3 (n=8) s m a l l a m p h i p o d s . I s e p a r a t e d t h e e f f e c t s o f s e x a n d a m p h i p o d a b u n d a n c e on t h e number o f a m p h i p o d s o f e a c h s i z e e a t e n i n a 2 - w a y ANOVA f o r e a c h a m p h i p o d s i z e a n d t r e a t m e n t . T h e r e was a weak t e n d e n c y f o r m a l e b e e t l e s t o f e e d on m o r e s m a l l a m p h i p o d s i n t h e s t a r v a t i o n t r e a t m e n t ( F = 3 . 5 7 , d f = 1 , 1 l 0 , p = 0 . 0 6 ) , b u t o t h e r m a l e - f e m a l e c o m p a r i s o n s w e r e n o t s i g n i f i c a n t ( F - t e s t s , d f = 1 , 1 l 0 , p > 0 . l 0 ) a n d v a r i a n c e s w e r e h o m o g e n e o u s . T h e r e w e r e no s i g n i f i c a n t d i f f e r e n c e s i n t h e number o f s m a l l a m p h i p o d s e a t e n a t d i f f e r e n t d e n s i t i e s o f l a r g e a m p h i p o d s i n e i t h e r t r e a t m e n t ( F - t e s t s , d f = 4 , 1 1 0 , P>0.10 , F i g . 1 8 , w h i t e b a r s ) , n o r w e r e t h e i n t e r a c t i o n t e r m s w i t h s e x s i g n i f i c a n t ( F - t e s t s , d f = 4 , 1 1 0 , P > 0 . 1 0 ) . The number o f l a r g e a m p h i p o d s e a t e n i n c r e a s e d w i t h t h e number o f l a r g e a m p h i p o d s p r e s e n t e d ( F i g . 1 8 , s h a d e d b a r s ) . T h i s e f f e c t was m o r e p r o n o u n c e d i n t h e s t a r v a t i o n t r e a t m e n t ( F = 6 . 2 2 , d f = 4 , 1 1 0 , p < 0 . 0 0 0 l ) t h a n i n t h e s a t i a t i o n t r e a t m e n t ( F = 2 . 4 2 , d f = 4 , 1 1 0 , p = 0 . 0 5 ) , a l t h o u g h v a r i a n c e s w e r e n o t h o m o g e n e o u s ( B a r t l e t t ' s t e s t , d f = 4 , p < 0 . 0 0 l a n d p = 0 . 0 l r e s p e c t i v e l y ) . A g a i n t h e i n t e r a c t i o n t e r m s w i t h s e x w e r e n o t s i g n i f i c a n t ( F - t e s t s , d f = 4 , 1 1 0 , p > 0 . 1 0 ) . 1 07 Figure 18. The mean ±1SE number (n=24) of large amphipods (shaded bars) and small amphipods (white bars) eaten at densities of 10 small and 2, 4, 6, 8, or 10 large amphipods. (A) starvation treatment and (B) sa t i a t i o n treatment. 108 109 I combined results from male and female beetles to test for differences between hunger treatments at each amphipod density. Beetles from the starvation treatment ate more small amphipods than beetles from the s a t i a t i o n treatment when large amphipods were present at densities of 2 and 10 only ( t - t e s t s , df=46, p<0.05), and ate more large amphipods at a l l densities ( t - t e s t s , df=46, p<0.0!). Table X. Values of c at d i f f e r e n t densities of large amphipods for each hunger treatment. Values given are corrected (corr) or not corrected (ncor) for loss of small amphipods. stavation s a t i a t i o n density corr ncor corr ncor 2 1.99 2.78 1.15 0.86 4 2.26 1.68 1.25 0.89 6 1.43 1.02 0.77 0.59 8 1.72 1.25 0.79 0.55 10 0.94 0.75 0.84 0.54 To test for preference I f i r s t computed the measure o suggested by Chesson (1978) where preference for large amphipods is given by oi = ( w , / x , ) [ ( w 1 / x 1 ) + ( w 2 / x 2 ) ] - 1 where w, and w2 are the numbers of large and small amphipods eaten, and x, and x 2 are the numbers of large and small amphipods presented. Preference for small amphipods is given by a 2 = 1~oi 110 T h i s m e t h o d was c h o s e n b e c a u s e a v a l u e f o r c, a n d o 2 c o u l d be o b t a i n e d when b e e t l e s f e d on o n e s i z e o f a m p h i p o d o n l y . I t h e n u s e d a s i g n t e s t t o c o m p a r e a n d o 2. F o r s t a r v e d b e e t l e s I f o u n d weak p r e f e r e n c e f o r l a r g e a m p h i p o d s a t d e n s i t i e s 2 ( n = 2 2 , p = 0 . 0 5 ) a n d 8 ( n = 2 4 , p = 0 . 0 6 ) , a n d f o r s m a l l a m p h i p o d s a t d e n s i t y 10 ( n = 2 3 , p = 0 . 0 9 ) . F o r b e e t l e s i n t h e s a t i a t i o n t r e a t m e n t , none o f t h e c o m p a r i s o n s w e r e s i g n i f i c a n t ( n = 2 4 , 2 2 , 2 4 , 24 a n d 19 f o r d e n s i t i e s 2 , 4 , 6 , 8 a n d 10 r e s p e c t i v e l y , P>0.10) a s p r e d i c t e d . I a l s o c o m p a r e d c, f o r s t a r v e d b e e t l e s a n d f o r b e e t l e s n e a r s a t i a t i o n u s i n g t h e m e d i a n t e s t . T h e r e was a s i g n i f i c a n t d e c r e a s e i n p r e f e r e n c e f o r l a r g e a m p h i p o d s by b e e t l e s n e a r s a t i a t i o n a t d e n s i t i e s 2 , 6 , a n d 8 ( n = 2 4 , p < 0 . 0 5 ) , a n d d i f f e r e n c e s a t d e n s i t i e s 4 a n d 10 w e r e i n t h e same d i r e c t i o n . To t e s t f o r c h a n g e s i n p r e f e r e n c e w i t h d e n s i t y , I c o m p u t e d c ( M u r d o c h 1 9 6 9 ) w h e r e c = w , x 2 /w 2 x , I u s e d mean v a l u e s f o r w, a n d w 2 f r o m F i g . 18 a n d r e p e a t e d t h e c a l c u l a t i o n s w i t h a n d w i t h o u t a c o r r e c t i o n f o r t h e mean number o f s m a l l a m p h i p o d s l o s t f r o m c o n t r o l j a r s . F o r b o t h t r e a t m e n t s , c e i t h e r s h o w e d no t r e n d o r d e c r e a s e d w i t h i n c r e a s i n g d e n s i t y o f l a r g e a m p h i p o d s ( T a b l e X ) . T h i s was o p p o s i t e t o p r e d i c t i o n s o f b o t h f r e q u e n c y - d e p e n d e n t a n d o p t i m a l d i e t m o d e l s o f p r e y s e l e c t i o n . 111 Discussion The optimal diet model (Pulliam 1974, Charnov 1976a) predicts that predators should spe c i a l i z e i f 1 < e(a)h(b) - h(a) Rf(a) eTBT This condition i s met for the parameter values in F i g . 16. Clearly, i f predators can assess the food value required from prey, the net rate of food intake does not predict when to sp e c i a l i z e . The optimal diet can change in composition such that items of lower value are included as the predator approaches s a t i a t i o n . E s s e n t i a l l y this means that i f only a portion of a high value item is required, a predator can be satiated more quickly i f i t accepts the f i r s t item i t encounters, rather than search f o r a high value prey. A range of D values can also be interpreted as a range of food intake requirements to satiate predators of di f f e r e n t sizes. For Thinopinus pictus for example, I found a cor r e l a t i o n of 0.505 between the weight of food required to satiate a beetle and beetle weight (page 65). Hence, differences in preference are expected for beetles of di f f e r e n t sizes. If the encounter rate with the high value prey is s u f f i c i e n t l y low, the model predicts switches between s p e c i a l i s t and expanding s p e c i a l i s t foraging rules with a continuous increase in predator s i z e . The generalist or expanding s p e c i a l i s t may have a (1) lower mean and variance of foraging time (2) higher mean but lower variance of foraging time (3) higher mean and variance of foraging time r e l a t i v e to 1 12 the s p e c i a l i s t Whether a predator forages as a generalist, expanding^ s p e c i a l i s t or s p e c i a l i s t , w i l l depend in part, on how i t responds to v a r i a t i o n in encounter rates of the prey types. For outcomes (1) and (3), the generalist or expanding s p e c i a l i s t , and s p e c i a l i s t strategies respectively are favored. For example, for D=0.20 mean differences in foraging times for the expanding s p e c i a l i s t and s p e c i a l i s t w i l l be s t a t i s t i c a l l y s i g n i f i c a n t (p<0.05) after about 20 foraging bouts (one-tailed z test) using values from F i g . 16b, or after about 6 foraging bouts using values from F i g . 16c. For predators which feed many times during their l i f e t i m e , even small time savings may become important. For outcome (2), the generalist or expanding s p e c i a l i s t strategy may be favored by a predator foraging in a risk-aversive manner (Caraco 1980, Caraco et a l . 1980). It w i l l have a higher mean foraging time, but w i l l also have a lower prob a b i l i t y of taking a much longer time to achieve s a t i a t i o n . For example, Real (1981) has shown that bees and wasps prefer constant over variable food rewards with equal expectations. Conversely, a forager may favor a s p e c i a l i s t strategy i f i t is foraging in a risk-prone manner. It w i l l require a shorter time on average to achieve s a t i a t i o n , but incurs the risk of taking a much longer time. Rarely w i l l prey be randomly di s t r i b u t e d as t h i s model assumes. Heller (1980) developed a model from which he showed that an expanding s p e c i a l i s t may also have an advantage when prey are clumped. Hunger was not considered in his model, but 113 p r o f i t a b l e prey could be rapidly depleted from a patch and interpatch travel times were long. It i s commonly held that hungry predators accept a wider variety of prey types than do predators near satiation (e.g. Schoener 1971). The best evidence for th i s comes from the work of Ivlev (1961) who showed that carp became increasingly se l e c t i v e as they approached s a t i a t i o n . In contrast, Akre and Johnson (1979) found a decrease in preference near satiation for damselfly naiads feeding on two prey types. Contrary evidence is given by the experiments reported here. Thinopinus pictus showed a s i g n i f i c a n t decrease in preference with s a t i a t i o n for large amphipods at most amphipod densities when starved beetles and beetles near s a t i a t i o n were compared. Although this result was in q u a l i t a t i v e agreement with the model presented here, the quantitative predictions were not met. Beetles in both treatments fed on small amphipods at a l l den s i t i e s . Preference for large amphipods did not increase with density as predicted. The model could account for these deviations from the "always or never rule" i f v a r i a t i o n among beetles was related to size differences, or i f beetles switch from s p e c i a l i s t to generalist foraging rules as they approach s a t i a t i o n during an experimental run. There were alternative interpretations of these results, however. Both beetles and damselfly naiads may change their search behaviour near s a t i a t i o n and a l t e r the r e l a t i v e encounter rates of different prey types. The model presented here is most applicable to predators 1 1 4 near s a t i a t i o n , as other factors l i k e l y affect foraging decisions of a predator which has been starved for a substantial period of time prior to experimentation. Future tests must also allow for changes in predator behaviour near s a t i a t i o n . Prey size should be s u f f i c i e n t l y large r e l a t i v e to the food requirement, so that changes in preference are measurable. Suitable predators for future tests are anthocorid bugs (Evans 1976) or s a l t i c i d spiders (Givens 1978), which require a r e l a t i v e l y few prey items for s a t i a t i o n . 1 1 5 CHAPTER 5: CONCLUDING REMARKS The staphylinid -.beetle Thinopinus pictus l i v e s in a s t r u c t u r a l l y simple environment. Yet, the food supply of Thinopinus varies over a range of s p a t i a l and temporal scales. Thinopinus i s obviously limited^ in i t s a b i l i t y to assess and respond to t h i s temporal and s p a t i a l v a r i a t i o n . I have argued that this result is a general one for invertebrate predators. Most of the support for optimal foraging theory with invertebrates has come from active foragers which search for slow-moving or sedentary prey (e.g. crabs (Elner and Hughes 1978), ants (Davidson 1978), and bumblebees (Pyke 1978)). Huey and Pianka (1981) compare various correlates of active and ambush foraging modes based on studies of desert l i z a r d s . Species which ambush generally eat fewer, more active prey, have lower metabolic costs, and have a limited learning a b i l i t y , r e l a t i v e to species which are active foragers. These results can probably be applied to invertebrates as well. One conclusion which emerges is that there are not simple general rules which govern foraging behaviours of a l l animals. Instead, the rules depend on prey behaviour, on the range of behaviours available to the forager, and on the capacity of the forager to make "correct decisions". A second conclusion i s that the capacities of foragers to respond to prey variation depend on the scale of measurement. Morse and F r i t z (1982) suggest that for ambush predators, predictions from optimal foraging theory based on patch choice are more successful than predictions based on diet choice. This 1 1 6 was not true for Thinopinus, as Thinopinus showed strong preference for the prey type (amphipod) on which i t had the highest feeding rate. However, Thinopinus showed only weak preferences among sizes of amphipods. The difference probably results because a wider variety of cues are available to distinguish among prey types than among sizes of a given prey type. The most successful recent attempts to predict foraging behaviours have been based on simple stochastic rules of behaviour (Hanski 1980), or on assumptions of the memory capacity of the animal (Pyke 1978, Ollason 1980). For example, in a variety of experiments Caraco et a_l. (1980), Waddington e_t a l . (1981) and Real (1981) have shown that food preferences depend not only on the expected outcome of a foraging bout, but also on the variance of that outcome. Hence, as i s so often true in ecological studies, the d i s t r i b u t i o n is more important than the mean. Problems of how foragers learn about and respond to va r i a t i o n in prey a v a i l a b i l i t y w i l l continue to be one exciting and necessary d i r e c t i o n for future research. 1 17 LITERATURE CITED Akre, B.G. and D.M. Johnson. 1979. Switching and sigmoid functional response curves by damselfly naiads with alternate prey available. J. Anim. Ecol. 48:703-720. Barnard, C.J. and Brown, C.A.J. 1981. Prey size selection and competition in the common shrew (Sorex araneus L.). Behav, Ecol. Sociobiol. 8:239-243. Beukema, J. J. 1968. Predation by the three-spined stickleback (Gasterosteus aculeatus): the influence of hunger and experience. Behavior 31:1-126. Borgia, G. 1980. Sexual competition in Scatophaga s t e r c o r i a : s i z e - and density-related changes in male a b i l i t y to capture females. Behavior 75:185-206. Bowers, D.E. 1963. F i e l d i d e n t i f i c a t i o n of five species of Calif o r n i a n beach hoppers (Crustacea: Amphipoda). Pac. Sc i . 17:315-320. Bowers, D.E. 1964. Natural history of two beach hoppers of the genus Orchestoidea (Crustacea: Amphipoda) with reference to their complemental d i s t r i b u t i o n . Ecology 45:677-696. Brooks, J.L. and S. Dodson. 1965. Predation, body size and composition of the plankton. Science 150:28-35. Brownell, P. and R.D. Farley. 1979. Pre y - l o c a l i z i n g behaviour of the nocturnal desert scorpion, Paruroctonus mesaensis: orientation to substrate vibrations. Anim. Behav. 27: 185-193. Calow, P. 1974. Some observations on locomotory strategies and their metabolic effects in two species of freshwater gastropods Ancylus f l u v i a t i l i s Mull, and Planorbis  contortus Linn. Oecologia 16:149-161. 118 Caraco, T. 1980. On foraging time a l l o c a t i o n in a stochastic environment. Ecology 61 :119-128. Caraco, T.,.S. Martindale, and T.S. Whittam. 1980. An empirical demonstration of r i s k - s e n s i t i v e foraging preferences. Anim. Behav. 28:820-830. Charnov, E.L. 1976a. Optimal foraging: Attack strategy of a mantid. Am. Nat. 110:141-151. Charnov, E.L. 1976b. Optimal foraging, the marginal value theorem. Theor. Pop. B i o l . 9:129-136. Charnov, E.L., G.H. Orians and K. Hyatt. 1976. Ecological implications of resource depression. Am. Nat. 110:247-259. Chesson, J. 1978. Measuring preference in selective predation. Ecology 59:211-215. Connell, J.H. 1975. Some mechanisms producing structure in natural communities. In M.L. Cody and J.M. Diamond (eds.) Ecology and evolution of communities, pp.460-490. Belknap Press, Cambridge, Mass. Cook, R.M. and B.J. Cockrell. 1978. Predator ingestion rate and i t s bearing on feeding time and the theory of optimal d i e t s . J. Anim. Ecol. 47:529-547. Cowie, R.J. 1977. Optimal foraging in Great T i t s (Parus major). Nature 268:137-139. Craig, P.C. 1970. The behavior and d i s t r i b u t i o n of the i n t e r t i d a l sand beetle, Thinopinus pictus (Coleoptera: Staphylinidae). Ecology 51:1012-1017. Craig, P.C. 1971. An analysis of the concept of lunar orientation in Orchestoidea corniculata . Anim. Behav. 19: 368-374. 119 Craig, P.C. 1973a. Orientation of the sand-beach amphipod, Orchestoidea corniculata. Anim. Behav. 21:699-706. Craig, P.C. 1973b. Behaviour and d i s t r i b u t i o n of the sand-beach amphipod Orchestoidea corniculata. Marine Biology 23: 101-109. Crowley, P.H. 1979. Behavior of zygopteran nymphs in a simulated weed bed. Odonatologica 8:91-101. Davidson, D.W. 1978. Experimental tests of the optimal diet in two s o c i a l insects. Behav. Ecol. Sociobiol. 4:35-41. Draper, N. and H. Smith. 1966. Applied regression analysis. John Wiley and Sons, Inc., New York Eggers, D.M. 1977. The nature of prey selection by planktivorous f i s h . Ecology 58:46-60. Elner, R.W. and R.N. Hughes. 1978. Energy maximization in the diet of the shore crab, Carcinus maenas. J. Anim. Ecol. 47:103-116. Enders, F. 1975. The influence of hunting manner on prey size, p a r t i c u l a r l y in spiders with long attack distances (Araneidae, Linyphiidae, and S a l t i c i d a e ) . Am. Nat. 109: 737-763. Ernsting, G. 1977. Effects of food deprivation and type of prey on predation by Notiophilus biguttatus F. (Carabidae) on s p r i n g t a i l s (Collembola). Oecologia 31:13-20. Evans, H.F. 1976. The role of predator-prey size in determining the e f f i c i e n c y of capture of Anthocoris nemorum and the escape reactions of i t s prey, Acyrthosiphon pi sum. Ecol. Entomol. 1:85-90. Everson, P. 1980. The r e l a t i v e a c t i v i t y and functional response of Phytoseiulus p e r s i m i l i s (Acarina: Phytoseiidae) and Tetranychus urticae (Acarina: Tetranychidae): The effect of temperature. Can. Entomol. 112:17-24:. 1 20 G a r d n e r , M . B . 1981. M e c h a n i s m s o f s i z e s e l e c t i v i t y by p l a n k t i v o r o u s f i s h : a t e s t o f h y p o t h e s e s . E c o l o g y 62: 571-578. G i v e n s , j R . P . 1978. D i m o r p h i c f e e d i n g s t r a t e g i e s o f a s a l t i c i d s p i d e r ( P h i d i p p u s a u d a x ) . E c o l o g y 59:309-321 . G o s s - C u s t a r d , J . D . 1977. O p t i m a l f o r a g i n g a n d t h e s i z e s e l e c t i o n o f worms by r e d s h a n k ( T r i n g a t o t a n u s ) . A n i m . B e h a v . 25: 10-29. G r e e n w o o d , J . J . D . a n d R . A . E l t o n . 1979. A n a l y z i n g e x p e r i m e n t s on f r e q u e n c y - d e p e n d e n t s e l e c t i o n by p r e d a t o r s . J . A n i m . E c o l . 48:721-737. G r i f f i t h s , D. 1980. The f e e d i n g b i o l o g y o f a n t - l i o n l a r v a e : p r e y c a p t u r e , h a n d l i n g a n d u t i l i z a t i o n . J . A n i m . E c o l . 49: 99-1 2 5 . G r i f f i t h s , D . 1981. S u b - o p t i m a l f o r a g i n g i n t h e a n t - l i o n M a c r o l e o n q u i n q u e m a c u l a t u s . J . A n i m . E c o l . 5 0 : 6 9 7 - 7 0 2 . G r u m , L . 1971. S p a t i a l d i f f e r e n t i a t i o n o f t h e C a r a b u s L . ( C a r a b i d a e , C o l e o p t e r a ) m o b i l i t y . E k . P o l . 19:1-34. H a n s k i , I . 1980. M o v e m e n t p a t t e r n s i n d u n g b e e t l e s a n d i n t h e d u n g f l y . A n i m . B e h a v . 2 8 : 9 5 3 - 9 6 4 . H a s s e l l , M . P . , S o u t h w o o d , T . R . E . 1978. F o r a g i n g s t r a t e g i e s o f i n s e c t s . A n n u . R e v . E c o l . S y s t . 9:75-98. H a y e s , W . B . 1970. The a c c u r a c y o f p i t a l l t r a p p i n g f o r t h e s a n d - b e a c h i s o p o d s T y l o s p u n c t a t u s . E c o l o g y 5 1 : 5 1 4 - 5 1 6 . H a y n e s , D . L . , a n d P . S i s o j e v i c 1966. P r e d a t o r y b e h a v i o r o f P h i l o d r o m u s r u f u s W a l c k e n a e r ( A r a n e a e : T h o m i s i d a e ) . C a n . E n t o m o l . 9 8 : 1 1 3 - 1 3 3 . H e a t w o l e , H . , a n d A . H e a t w o l e 1968. M o t i v a t i o n a l a s p e c t s o f f e e d i n g b e h a v i o r i n t o a d s . C o p e i a 4:692-698. 121 K e i n r i c h , B. 1 9 7 6 . The f o r a g i n g s p e c i a l i z a t i o n s of i n d i v i d u a l bumblebees. E c o l . Monogr. 4 6 : 1 0 5 - 1 2 8 . H e l l e r , R. 1 9 8 0 . On o p t i m a l d i e t i n a patchy environment. Theor. Pop. B i o l . J_7: 2 0 1 -2 1 4 . H o l l i n g , C S . 1 9 5 9 . Some c h a r a c t e r i s t i c s of s i m p l e t y p e s of p r e d a t i o n and p a r a s i t i s m . Can. Entomol. 9 1 : 3 8 5 - 3 9 8 . H o l l i n g , C S . 1 9 6 6 . The f u n c t i o n a l response of i n v e r t e b r a t e p r e d a t o r s t o prey d e n s i t y . Mem. Entomol. Soc. Can. 4 8 : 1 - 8 6 . Huey, R.B. and E.R. P i a n k a . 1 9 8 1 . E c o l o g i c a l consequences of f o r a g i n g mode. E c o l o g y 6 2 : 9 9 1 - 9 9 9 . Hughes, R.N. 1 9 7 9 . O p t i m a l d i e t s under the energy m a x i m i z a t i o n p r e m i s e : the e f f e c t s of r e c o g n i t i o n time and l e a r n i n g . Am. Nat. 1 1 3 : 2 0 9 - 2 2 1 . Inouye, R.S., G.S. Byers and J.H. Brown. 1 9 8 0 . E f f e c t s of p r e d a t i o n and c o m p e t i t i o n on s u r v i v o r s h i p , f e c u n d i t y , and community s t r u c t u r e of d e s e r t a n n u a l s . E c o l o g y 6J_: 1 3 4 4 - 1 3 5 1 . I v l e v , V.S. 1 9 6 1 . E x p e r i m e n t a l f e e d i n g of f i s h e s . Y a l e U n i v e r s i t y P r e s s , New Haven. J a e g e r , R.G. and D.E. B a r n a r d . 1 9 8 1 . F o r a g i n g t a c t i c s of a t e r r i s t i a l salamander: Choice of d i e t i n s t r u c t u r a l l y s i m p l e e n v i r o n m e n t s . Am. Nat. 1 1 7 : 6 3 9 - 6 6 4 . J a e g e r , R.G., K.C.B. N i s h i k a w a , and D.E. B a r n a r d . 1 9 8 1 . F o r a g i n g t a c t i c s of a t e r r e s t i a l salamander: E f f e c t of a c o m p e t i t o r on c h o i c e of d i e t . Am. Nat. i n p r e s s . Johnson, D.M., B.G. A k r e , and P.H. Crowley. 1 9 7 5 . M o d e l l i n g a r t h r o p o d p r e d a t i o n : w a s t e f u l k i l l i n g by d a m s e l f l y n a i d s . E c o l o g y 5 6 : 1 0 8 1 - 1 0 9 3 . 1 22 Kacelnik, A., A.I. Houston, and J.R. Krebs. 1981. Optimal foraging and t e r r i t o r i a l defence in the great t i t (Parus  major). Behav. Ecol. Sociobiol. 8:35-40. K i s l a l i o g l u , M., and R.N. Gibson. 1976. Prey 'handling time' and i t s importance in food selection by the 15-spined stickleback, Spinachia spinachia (L.). J. Exp. Mar. B i o l . Ecol. 25:151-158. Krebs, J.R., A. Kacelnik, and P. Taylor. 1978. Test of optimal sampling by foraging great t i t s . Nature 275:27-31. Krebs, J.R. 1978. Optimal foraging decision rules for predators. Behavioral ecology: an evolutionary approach (Ed. by J.R. Krebs and N.B. Davies), pp. 23-63. Sinauer, Sunderland, Mass. Maynard Smith, J. 1978. Optimization theory in evolution. Annu. Rev. Ecol. Syst. 9:31-56. McCauley, D.E. 1982. The behavioural components of sexual selection in the milkweed beetle Tetraopes  tetraophthalmus. Anim. Behav. 30:23-28. McCleery, R.H. 1977. On sa t i a t i o n curves. Am. Nat. 25:1005-1015. M i l i n s k i , M., and R. Heller 1978. Influence of a predator on the optimal foraging behaviour of sticklebacks (Gasterosteus aculeatus L.). Nature 275:642-644. Morse, D.H., and R.S. F r i t z . 1982. Experimental and observational studies of patch choice at d i f f e r e n t scales by the crab spider Misumena vat i a . Ecology 63:172-182. Murdoch, W.W. 1969. Switching in general predators: Experiments on predator s p e c i f i c i t y and s t a b i l i t y of prey populations. Ecology 39:335-354. Murdoch, W.W. and A. Oaten. 1975. Predation and population s t a b i l i t y . Adv. Ecol. Res. 9:1-131. 1 23 Murtaugh, P. 1981. Size-selective predation on Daphnia by Neomysis mercedis. Ecology 62:894-900. Ollason, J.G. 1980. Learning to forage - optimally? Theor. Pop. B i o l . 18:44-56. Paloheimo, J.E. 1979. Indices of food preference by a predator. J. Fish. Res. Board Can. 36:470-473. Parker, G.A 1970. Sperm competition and i t s evolutionary consequences in the insects. B i o l . Rev. 45:525-567. Parker, G.A. and R.A. Stuart. 1976. Animal behavior as a strategy optimizer: Evolution of resource assessment strategies and optimal emigration thresholds Am. Nat. 110: 1055-1076. Pastorok, R.A. 1981. Prey v u l n e r a b i l i t y and size selection by Chaoborus larvae. Ecology 62:1311-1324. P o l i s , G.A. 1980. Seasonal patterns and age-specific variation in the surface a c t i v i t y of a population of desert scorpions in re l a t i o n to environmental factors. J. Anim. Ecol. 49:1-18. Pulliam, H.R. 1974. On the theory of optimal d i e t s . Am. Nat. 108:59-74. Pulliam, H.R. 1975. Diet optimization with nutrient constraints. Am. Nat. 109:765-768. Pyke, G.H. 1978. Optimal foraging: Movement patterns of bumblebees between inflorescences. Theor. Pop. B i o l . 13: 72-98. Pyke, G.H., H.R. Pulliam, and E.L. Charnov. 1977. Optimal foraging: A selective review of theory and te s t s . Quart. Rev. B i o l . 52:137-154. 124 Real, L.A. 1981. Uncertainty and pollinator-plant interactions: the foraging behavior of bees and wasps on a r t i f i c i a l flowers. Ecology 62:20-26. Richards, L.J. and J.H. Myers. 1980. Maternal influences on size and emergence time of the cinnabar moth. Can. J. Zool. 58: • 1452-1457. Rogers, D. 1972. Random search and insect population models. J. Anim. Ecol. 41:369-383. Schlager, G. 1960. Sperm precedence in the f e r t i l i z a t i o n of eggs in Tribolium castaneum. Ann. Entomol. Soc. Amer. 53: 557-560. Schoener, T.W. 1971. Theory of feeding strategies. Annu. Rev. Ecol. Syst. 2:369-404. Schulter, D. 1981. Does the theory of optimal diets apply in complex environments. Am. Nat. 118:139-147. Selander, R.K. 1966. Sexual dimorphism and d i f f e r e n t i a l niche u t i l i z a t i o n in birds. Condor 63:29-86. Siegel, S. 1956. Nonparametric s t a t i s t i c s for the behavioral sciences. McGraw-Hill Book Co., New York Smith, R.L. 1979. Repeated copulation and sperm precedence: paternity assurance for a male-brooding water bug. Science 205:1029-1031. Thompson, D.J. 1978a. Towards a r e a l i s t i c predator-prey model: the effect of temperature on the functional response and l i f e history of the damselfly, Ischnura elegans• J . Anim. Ecol. 47:757-767. Thompson, D.J. 1978b. Prey, size selection by larvae of the damselfly, Ischnura elegans (Odonata). J. Anim. Ecol. 47: 769-785. 1 25 Toft, C.A. 1980. Feeding ecology of thirteen syntopic species of anurans in a seasonal t r o p i c a l environment. Oecologia 45: 131-141. Turnbull, A.L. 1964. The search for prey by a web-building spider Achaearanea tepidariorum (C.L. Koch) (Araneae, Ther i d i idae). Can. Entomol. 96:568-579. Waddington, K.D., A. Tracy and B. Heinrich. 1981. F l o r a l preferences of bumblebees (Bombus edwardsi i) in r e l a t i o n to intermittent versus continuous rewards. Anim. Behav. 29:779-784. Westoby, M. 1978. What are the b i o l o g i c a l bases of varied diets? Am. Nat. 108:290-304. Wiens, J.A. 1976. Population responses to patchy environmnets. Annu. Rev. Ecol. Syst. 7:81-120. Zach, R. and J.B. F a l l s . 1976. Ovenbird (Aves: Parulidae) hunting behavior in a patchy environment: an experimental study. Can. J. Zool. 54:1863-1879. Zach, R. 1978. Selection and dropping of whelks by northwestern crows. Behavior 67:134-148. 1 26 APPENDIX A Expected handling time for the generalist can be expressed as n i n+1 E(H ) = E[iP+W]f(a)f(b) + Xf(b) G i = 0 where P = h(b)-e(b)h(a)/e(a) W = Dh(a)/e(a) and X = Dh(b)/e(b) The term in square brackets is the time to feed on i type B prey, and then to complete feeding on a type A prey. This term i s multiplied by the p r o b a b i l i t y that such a sequence of events w i l l occur. The second term gives the time to reach s a t i a t i o n for feeding on type B prey only, times the probability that no type A prey are encountered. The mean and variance can be derived from the moment generating function n [ip+w]t i Xt n+1 M (t) = f(a) Ee f(b) + e f(b) H i = 0 Wt (n+l)Pt n+1 Pt -1 = + e f ( a ) [ l - e f(b) ] [1-e f(b) ] Xt n+1 + e f(b) Then E(H ) = M' (0) G H n+1 n = P[nf(b) -(n+1)f(.b) +l]f(b)/f(a) n+1 n+1 + W[l-f(b) ] + Xf(b) and 127 V(H ) = M " ( 0 ) - [E(H ) ] 2 G H G 2 n+1 n+1 = [X-W] [1-f(b) ] f ( b ) n+1 n n+2 - 2P[X-W][nf (b) - ( n + D f ( b ) + l ] f ( b ) / f ( a ) 2 2 n+1 n+1 - P [n d + f ( b ) )+2n]f(b) 2 2n+2 2n+1 n+1 + P [2nf(b) -(2n+1)f(b) +f(b) n 2 - f ( b ) + l ] f ( b ) / f ( a ) 128 A P P E N D I X B The moment g e n e r a t i n g f u n c t i o n o f s e a r c h t i m e f o r t h e g e n e r a l i s t i s g i v e n by . c o . . n 1 r q t - R t i+1 I - 1 M (q ) = Ef ( a ) f (b ) J e R t [ i ! ] d t S i = 0 0 n+1 T q t - R t n+1 n - 1 + f (b ) J e R t [ n ! ] d t 0 The i n t e g r a t i o n t e r m d e s c r i b e s t h e moment g e n e r a t i n g f u n c t i o n o f t h e gamma d i s t r i b u t i o n , t h e d i s t r i b u t i o n o f w a i t i n g t i m e s t o o b t a i n e x a c t l y i+1 p r e y i t e m s . I t i s m u l t i p l i e d by t h e p r o b a b i l i t y t h a t e x a c t l y i+1 p r e y i t e m s a r e e a t e n , a n d summed o v e r a l l v a l u e s o f i . A f t e r p e r f o r m i n g t h e i n t e g r a t i o n s a n d s u m m a t i o n s t h i s b e c o m e s n+1 M (q ) = R f ( a ) ( l - [ R f ( b ) / ( R - q ) ] } S R f ( a ) - q n+1 n+1 + f(b) { R / ( R - q ) } The e x p e c t e d v a l u e o f s e a r c h t i m e i s g i v e n by E ( S ) = M ' ( 0 ) G S n+1 - 1 = [ 1 - f ( b ) ] [ R f ( a ) ] w i t h v a r i a n c e V ( S ) = M " ( 0 ) - [ E ( S ) ] 2 G S G n+1 n+2 2n+2 - 2 = [ 1 - 2 (n+1 ) f ( b ) + 2 ( n + D f ( b ) - f ( b ) ] [ R f ( a ) ] 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0095623/manifest

Comment

Related Items