 

 

 

 

 

 

 

 

_....;.?-.- .

ii":~'-i.‘1‘\£v5

Ill/Il/llll////II///I//////I///I/lllllWlli/W/I

 

 

 

LIBRARY
Michigan State
University l
k ,

 

.aw'

This is to certify that the
thesis entitled
FOHAGING BY CHICKS F R TNO TYPES
OF CRYPTIK PRSY: A TBS T OF THE
SiARCH IMA JE HYPOTH3SIS

presented by

Karen Ruth Cebra

has been accepted towards fulfillment
of the requirements for

Wdegree in Zoology

///( [/5

Major professor

     

/
Date (? iﬂgéi /¢7[/

0-7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
c:\circ\dstedm.pn3-p. 1

FORAGING BY CHICKS FOR TWO TYPES
OF CRYPTIC PREY: A TEST OF THE
SEARCH IMAGE HYPOTHESIS

By

Karen Ruth Cebra

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the Degree of

MASTER OF SCIENCE

Department of Zoology

1990

ABSTRACT

FORAGING BY CHICKS FOR TWO TYPES OF
CRYPTIC PREY: A TEST OF THE SEARCH IMAGE
HYPOTHESIS

By

Karen Ruth Cebra

Support for the existence of perceptual specializations, or search images, for a
given prey type has recently been challenged using the search rate hypothesis. In this
series of experiments domestic chicks (Gallus gallus) were used to test for the property
of interference that is predicted by the search image hypothesis but not by the search
rate hypothesis. Initial experiments determined two prey types (colored rice grains)
which were equally preferred, distinguishable, and of equal crypticity. Experiments
designed to test for interference presented the two prey types simultaneously on a
cryptic background. No evidence for interferencewas found over the course of entire
or successive trials, but some evidence, in the form of runs on a single prey type,

exists for interference within a trial.

ii

To Daniel, without whom this work would never have been done.

iii

Contents

LIST OF TABLES vi
LIST OF FIGURES viii
I INTRODUCTION 1
II METHODS 5
A General Methods ............. . ................ 5
B Preference Experiments - Methods ................... 13
C Discrimination Experiments - Methods ................. 14
D Experiments with Cryptic Prey - Methods ............... 14
IIIRESULTS 16
A Preference and Discrimination Experiments .............. 16
B Experiments with Cryptic Prey ..................... 27
IV DISCUSSION 32
APPENDICES 38
A Data Conversion Program 38

iv

B Asymmetry Value Simulation

C Runs Test Simulation

LIST OF REFERENCES

40

41

43

List of Tables

111.1 The average rate of intake for the various color combinations in the
preference study experiments. The * indicates a signiﬁcant t-test, 0.05

level ..................................... 2 1

111.2 Asymmetry values for Yellow-Red and Orange-Red color combinations.

The * indicates a signiﬁcant Mann-Whitney Utest, 0.05 level. . . . . 22

111.3 Results of the Friedman’s test showing the effect of trial number on
the relative rate at which grains were taken. A * indicates a. signiﬁcant

difference. ................................. 23

111.4 The average t, values in the runs test for the various studies, and the
expected values if there were only random selection of grains. A *

indicates a signiﬁcant t-test, 0.05 level. . . .t .............. 27

111.5 ANOVA table for the experiment with no introduction demonstrating

that trial number had no signiﬁcant effect on the relative rates. . . . 30

111.6 ANOVA table for the experiment with orange introduction demon-

strating that trial number had no signiﬁcant effect on the relative rates. 30

111.7 ANOVA table for the experiment with red introduction demonstrating
that trial number had no signiﬁcant effect on the relative rates. The *

indicates a signiﬁcant difference at the 0.05 level. ........... 30

vi

111.8 Results of t-tests on trial 1 for the three experiments with cryptic prey
showing no signiﬁcant difference between the mean rates at which each

color was chosen. .............................

vii

List of Figures

11.1

11.2

11.3

11.4

11.5

11.6

11.7

11.8

111.1

A photograph of the brooding pen. A heat lamp is provided ...... 6
A photograph of the group cages. Each tier could house up to 12 chicks. 6

Photographs of (a) a female chick and (b) a male chick at 28 days of

age. Note the more pronounced red comb and wattle of the male. . . 7
A photograph of the testing arena used for the experiments with cryptic

prey. ................................... 9
A photograph of the various colors of rice grains. ........... 9

Observations of a trial. The observer records the number of grains

taken of each color both orally and with handheld counters shown above. 11

A photograph of a marked chick. Chicks’ toes were numbered 1 through
6 starting from the left. N o chick had more than two toes marked. This

chick is marked number 1. ........................ 12

A photograph of a pair of chicks foraging. The companion chick is in

the back ................................... 12

A plot of the cumulative grains taken as a function of time for each of
the 10 focal chicks in the green-red preference experiment. The dotted
line corresponds to green grains while the solid line corresponds to red

grains. ................................... 17

viii

111.2 A plot of the cumulative grains taken as a function of time for each
of the 10 focal chicks in the yellow-orange preference experiment. The
dotted line corresponds to yellow grains while the solid line corresponds

to orange grains. .............................

111.3 A plot of the cumulative grains taken as a function of time for each of
the 10 focal chicks in the orange-red preference experiment. The dotted

line corresponds to orange grains while the solid line corresponds to red

111.4 A plot of the cumulative grains taken as a function of time for each of
the 10 focal chicks in the yellow-red preference experiment. The dotted
line corresponds to yellow grains while the solid line corresponds to red

grains. ...................................

111.5 The typical behavior of two focal chicks in discrimination studies over
the course of 5 successive trials. The data presented are the cumulative

number of grains taken as a function of time. .............

111.6 The distribution of t, values (solid histograms) from (a) the preference

experiment, (b) the cryptic prey experiment with no prior introduction

to a prey type, and (c) the cryptic prey experiment with introduction.

111.7 Experiments with cryptic prey - no prior introduction: The typical

behavior of a focal chick over the course of 3 successive trials.

111.8 Experiments with cryptic prey - prior introduction: The typical be-

havior of two focal chicks over the course of 3 successive trials.

1V.1 Predictions of the search image model and the search rate model as a

function of the relative densities of the two cryptic prey species.

ix

18

19

20

24

26

28

29

35

Chapter I

INTRODUCTION

The behavior of a forager affects not only the individual, but also the population
dynamics of the prey species. If the predator nonrandomly selects prey, the effect of
the predation will be a selective pressure upon the prey population (Clarke 1979).
For example, it has been demonstrated that preferential selection of the most com-
mon phenotype can result in apostatic selection (switching) which will contribute to
the maintenance of genetic diversity (Clarke 1962, Allen 1974, Murdoch and Oaten
1975, Cornell 1976, Cooper 1984). One postulated behavioral mechanism that could
produce this effect involves the formation of a ‘search image’ (see review in Krebs
1973) - a term which has often been applied more broadly than originally intended
(Lawrence and Allen 1983). Here it is used, as originally intended by Tinbergen
(1960) and Dawkins (1971a), to refer to a ‘change in the ability of a predator to
detect cryptic familiar prey’ (Lawrence and Allen 1983). More speciﬁcally, ‘search
image’ refers to ‘perceptual changes in the predator that temporarily increase its
ability to detect particular cryptic prey as a result of recent encounters with similar
cryptic prey’ (Guilford and Dawkins 1987). A key property of a search image for a
given prey type is that it interferes with the predator’s ability to detect other prey
types (Croze 1970) which may result in a preference (Krebs 1973). It is this property

that isolates the ‘search image’ mechanism from other types of learning that could

1

account for a preference for a given prey type as postulated by Krebs (1973). As

summarized by Lawrence and Allen (1983) these include learning to:

0 ﬁnd food in a speciﬁc place,
0 search in an appropriate habitat type,
0 adjust the search path to maximize prey encounter probability,

0 prefer or avoid a given prey type over others independent of the predator’s ability

to see the different types (i.e. prey of varying palatability or familiarity),
o utilize speciﬁc hunting techniques, for example, search rate modiﬁcation, and

0 improve the ability to handle prey.

Though all of these types of learning may lead to a preference for a prey type, none
of them will change the predator’s ability to detect one prey type over another as a

search image would.

Many experiments enlisted in support of the search image hypothesis have recorded
improvements in the ability of predators to detect and capture cryptic prey (de
Ruiter 1952, Kettlewell 1955, Clarke 1962, Dawkins 1971a, Pietrewicz and Kamil
1979, Lawrence 1985a, b, 1986, Gendron 1986). Recently, Guilford and Dawkins
(1987) have challenged several studies (Dawkins 1971a, Pietrewicz and Kamil 1979,
Lawrence 1985a, b, 1986, Gendron 1986) employing this type of evidence in support
of perceptual specialization as a foraging tactic because it fails to exclude other hy-
potheses which could also account for improvements in the capture rate as a function
of recent experience, but without perceptual specializations. The same criticisms

put forth by Guilford and Dawkins (1987) can be applied to Croze’s (1970) work

on Carrion Crows, which Krebs (1973) cites as the ‘only detailed attempt so far to

investigate the mechanism of searching-image formation in the ﬁeld.’

As an alternative to search images, Guilford and Dawkins (1987) offer the ‘search
rate hypothesis’, a ‘slightly reinterpreted version’ of Gendron and Staddon’s (1983)
optimal search rate hypothesis. In this scenario, a predator is faced with a trade-off
between foraging speed and detection accuracy - an increase in search speed decreases
the probability of detecting the prey but increases the encounter rate whereas a
decrease in speed increases the probability of detection but decreases the encounter
rate (Gendron and Staddon 1983, Guilford and Dawkins 1987). The optimal search
speed is one in which prey capture is maximized. Clearly, an increased ability to
detect one cryptic or conspicuous prey type should not interfere with the detection of
other equally cryptic or conspicuous prey as it would if search images were employed.
Nevertheless, the search image and search rate hypotheses result in similar foraging
behavior in other ways. Both predict that detection accuracy increases for cryptic
prey as a result of recent encounters with that prey type, and both predict that
frequent encounters with conspicuous prey interfere with the ability to detect cryptic
prey. In addition, both predict that an increased ability to detect conspicuous prey

decreases the ability to detect cryptic prey (Guilford and Dawkins 1987).

The purpose of the present work was to test for the interference predicted by the
search image hypothesis using experiments modeled closely on those in the original
study by Dawkins (1971a); ‘predators’ (domestic chicks) were given ‘prey’ (artiﬁ-
cially colored grains of rice) on backgrounds colored to make the prey appear cryptic
or conspicuous. Domestic chicks were chosen because of the ease with which the ex-
perimental conditions could be controlled. Also, as Dawkins points out, rice grains
were chosen to keep the experiments ‘as natural as possible, by observing the chicks

performing a common part of their behavioral repertoire, i.e. pecking food from the

ground’ (Dawkins 1971a). The key difference between the present work and all pre-
vious experimental studies is that predators were given a simultaneous choice of two
different but equally preferred and equally cryptic prey colors (=types). If a changed
ability to take one prey type automatically interferes with the ability to take the other
type, then the two types should be taken at different rates, as a strict interpretation
of the search image hypothesis would predict. The search rate hypothesis does not
predict interference. A direct test of the search rate hypothesis is not possible in
this study because the individual parameters that make up search rate (for example
handling time and area searched) are not measured, and because the experiment does
not prevent birds from modifying their search rates. Nevertheless this study isolates
the question of whether a change in the ability to take one prey type interferes with

the ability to take another type.

Chapter II

METHODS

A General Methods

As in Dawkins’ (1971a) study, domestic chicks (Gallus gallus) were used as ‘predators’
and colored rice grains were used as ‘prey’. Freshly laid White Leghorn eggs were
obtained from a commercial hatchery and incubated. Upon hatching, chicks were
moved to a brooding pen (Figure 11.1) and maintained on a light:dark cycle standardly
used for commercial production - 24:0 for days 1-3, 23:1 for days 4-6 and then 14:10
from day 7 on. At seven days of age, they were moved to group cages (Figure 11.2) and
raised to the testing age of 28 days, when they could be sexed reliably (Figure 11.3).
During this period chicks were fed commercial chick starter feed and supplied with
water ad Iibitum. Food and water were replenished at regular intervals in order to

establish consistency throughout the experiments.

The behavior of the chicks was observed in an arena measuring 60 X 60 X 45 cm
high (after Dawkins 1971a) and was made of white tempered hardboard (Masonite
Corp.). The ﬂoor was either a 60 X 60 cm piece of white tempered hardboard (pref-
erence and discrimination experiments) or a 60 X 60 cm piece of white tempered
hardboard with two colors of aquarium gravel (Spectrastone by Wil-Marox) glued to

it with clear-drying epoxy (experiments with cryptic prey). Figure 11.4 shows a view

 

Figure 11.2: A photograph of the group cages. Each tier could house up to 12 chicks.

 

  

«_ 5.5,

Figure 11.3: Photographs of (a) a female chick and (b) a male chick at 28 days of age.
Note the more pronounced red comb and wattle of the male.

of the testing arena being prepared for the experiments with cryptic prey. The gravel
was mixed in equal proportions prior to gluing and then scattered over the wet glue
and allowed to dry thoroughly. Careful attention was paid to ensure equal intensity
of the two colors of gravel with respect to each other. Rice grains were dyed with food
coloring (Dec-a-Cake by Durkee Famous Food Inc.) to match as closely as possible

by human eye the colors of the aquarium gravel using the following recipe.

Ingredients :

1. 170 ml rice
2. 170 ml water
3. Drops of food coloring for:

0 Orange: 12 drops yellow, 8 drops red
0 Red: 30 drops
0 Green: 20 drops

0 Yellow: 10 drops

Procedure: Mix rice, water, and food coloring thoroughly; allow to soak 6 hours
and then spread on newspaper to dry; stir occasionally to prevent rice from

sticking.

Green and orange were chosen based on Dawkins’ (1971a) study. Red and yellow
were chosen as possible colors because they could be prepared easily without mixing
colors. Blue and violet hues were avoided because a chicken’s vision is not sensitive
to light of those frequencies (Bowmaker and Knowles 1977). Figure 11.5 shows the

various colors used or considered for these experiments.

Approximately 700 grains (11 grams) of each of the two colors to be used in the

experiment were mixed together and then scattered on the floor of the arena. Chicken

    

A.

 

. 1'.
-- Iv... -. .. g 0 '

Figure 11.4: A photograph of th
prey.

e testing arena used for the experiments with cryptic

 

Figure 11.5: A photograph of the various colors of rice grains.

10

wire covered the arena to prevent chicks from escaping. The observer watched the

chicks from above as shown in Figure 11.6.

Over a 7 day period starting at 21 days of age, chicks were acclimated in groups
to the test arena, to the presence and voice of an observer, and to colored rice grains
as food. Acclimations began about 1 hour after the lights came on and ended when
all the chicks had been exposed to test conditions once. At the start of each day of
acclimation or experimental testing, chicks were marked for individual identiﬁcation
with India ink as shown in Figure 11.7. The food trays were removed from the group
cages just before the lights came on so that the chicks would be hungry for the
acclimation runs. A group of 3 to 5 chicks were placed in the test arena with colored
rice for 15 minutes while an observer made oral comments about their behavior. Over
the 7 days of acclimation, an effort was made to group chicks such that individuals

foraged with as many of the other individuals as possible.

After the seven days of acclimation, the chicks were divided according to sex and
then assigned randomly to be either focal chicks (whose behavior was recorded) or
companion chicks (whose behavior was not recorded). Since chicks become agitated
when alone (Dawkins 1971a) a companion chick had to be placed in the arena with
each focal chick. Neither companion nor focal chicks were used in more than one
experiment, but within an experiment, chicks could be used in multiple trials. Com-
panion chicks could be paired with more than one focal chick in any given experiment.
In no pair were two males placed together since two males tend to ﬁght with one an-
other rather than forage. Figure 11.8 shows a focal—companion chick pair foraging on

a white background.

Each experimental trial began with the introduction of a focal chick and a com-
panion chick into the arena (the companion chick was always put in ﬁrst so that the

focal chick was never alone in the arena) and ended after the focal chick had eaten a

11

 

Figure 11.6: Observations of a trial. The observer records the number of grains taken
of each color both orally and with handheld counters shown above.

 

Figure 11.7: A photograph of a markcd chick. Chicks’ toes were numbered 1 through
6 starting from the left. No chick had more than two toes marked. This chick is
marked number 1.

 

Figure 11.8: A photograph of a pair of chicks foraging. The companion chick is in the
back.

13

speciﬁed number of grains or had foraged for a speciﬁed maximum amount of time,
whichever came ﬁrst (refer to sections B, C, and D in chapter 11 for speciﬁc details).
The observers oral comments about the colors of ‘prey’ selected by the focal chick
were tape recorded and later transferred to computer for analysis. The computer
program ‘TIMER’ (see Appendix A) was used for the transfer. 1t recorded the color
and time for each of the strikes over the course of a trial into a data ﬁle. A plot of
cumulative rice grains of each color against time could be made from these ﬁles so

that the net rates at which each color was selected could be calculated.

B Preference Experiments - Methods

The goal of the preference. experiments was to ﬁnd a pair of artiﬁcial prey colors
which the chicks preferred equally well, thus minimizing the confounding inﬂuence of
color preference in subsequent experiments. Four pairs of colors were tested: green
and red, yellow and orange, orange and red, and yellow and red. Ten focal chicks
were tested on each combination. About 700 grains of each color were scattered on
a white ﬂoor in the arena. The focal chick was observed until it had eaten a total of
75 grains or had foraged for 15 minutes, whichever came ﬁrst. Each focal chick was
tested only once (a total of 40 chicks - green-red: 4 males, 6 females; yellow-orange: 4
males, 6 females; yellow-red: 3 males, 7 females; orange-red: 5 males, 5 females); the

same set of seven companion chicks was used for each of the four color combinations.

Two colors were determined to be equally preferred if the chicks chose them at
equal net rates (as determined in the analysis of the cumulative number of grains of

a particular color chosen versus time) over a trial.

14

C Discrimination Experiments - Methods

To determine whether chicks could distinguish between the two colors found to be
equally preferred (red and orange), rice grains of one of the two colors were made
distasteful. In the ﬁrst experiment, orange was made distasteful by soaking the rice
for 6 hours in a solution of 4 g quinine sulfate and 2 g of powdered mustard per 100
ml of dye solution (Gittleman and Harvey 1980). In the second experiment red was
made distasteful. 1f chicks are able to discriminate between red and orange, then
they should be able to learn to avoid the distasteful color. About 700 grains of each
color were scattered on a white background and the focal chick was observed until
it had eaten 75 grains or until 7 minutes had elapsed, whichever came ﬁrst. During
the preference experiments, it was determined that any pattern would be obvious by
7 minutes. For each of the two experiments, 10 focal chicks (5 males and 5 females)
and 6 companion chicks were presented with one of the colors distasteful. Trials were
repeated four times at hourly intervals, so that each chick was observed for a total of

ﬁve trials.

D Experiments with Cryptic Prey - Methods

In order to test whether an ability to take one prey type interferes with the ability
to take another equally cryptic prey type, a total of 11 male and 11 female focal
chicks were given the opportunity to forage for red and orange rice grains, which
had been determined to be equally preferred and distinguishable by'the chicks in the
prior experiment, on a background of red and orange aquarium gravel. The gravel
was glued to the floor such that chicks were unable to dislodge it either by pecking
or scratching. In addition, chicks were able to learn quickly the difference between

gravel and ‘prey’ and thus did not waste time pecking at gravel.

15

Two experiments were performed. In the ﬁrst, performed with 10 focal chicks (5
male and 5 female) and 2 companion chicks, a focal chick and companion chick pair
was placed into the arena with about 700 grains each of red and orange rice grains
and observed until the focal chick had eaten a total of 75 grains or had foraged for 7
minutes, whichever came ﬁrst. Trials were repeated two times at hourly intervals, so

that each chick was observed for a total of three trials.

The second experiment, with 12 focal chicks (6 male and 6 female) and 4 com-
panion chicks was similar to the ﬁrst except that each chick was exposed prior to the
trial to an introductory arena where approximately 60 grains of one of the two colors
was spread on a white background. 3 males and 3 females, with companions, were
introduced to red, and 3 males and 3 females were introduced to orange. Chicks were
transferred from the introductory arena to the arena with cryptic prey after the focal
chick had eaten 10 grains (Dawkins 1971b) or had foraged for 30 seconds, whichever
came ﬁrst. The goal was to determine whether prior experience with a single prey
type would affect later choice behavior when chicks were subsequently presented with

two prey types.

Chapter III

RESULTS

A Preference and Discrimination Experiments

Two simultaneously presented colors were considered to be equally preferred if the
following two conditions were met. First, there should not be a signiﬁcant differ-
ence between the rates, as measured by the cumulative number of grains taken over
time averaged for a sample of chicks, at which each color was taken. Figures 111.1
through 111.4 show the behavior of individual chicks for the preference experiments
plotted as the cumulative number of grains selected against elapsed time. The aver-
age rate at which each color was selected is shown in Table 111.1. For chicks tested on
two of the combinations, green-red and yellow-orange, clear preferences were observed
(t-test, 0.05 level; and refer to Figures 111.1 and 111.2, and Table 111.1). For chicks
tested on orange-red and on yellow-red, the t-test showed no statistically signiﬁcant

preference (refer to Figures 111.3 and 111.4, and Table 111.1).

Second, there should be no signiﬁcant tendency for individual chicks to specialize
on one color or the other irrespective of whether the intake rates averaged across
chicks were the same. This condition would be violated if some chicks were to take
one color predominantly and others were to take the other color, resulting in equal

average intake rates. This tendency was measured by calculating an ‘asymmetry

16

l7

£53m we“ 3 miaoamouuoo 0:: «cm—om 2: £23 339% ~8on 3
2:533:23 2.: 10:3. 2;. 4.851338 ova—Daemon; BEA—poem 2: 5 3—3—3 Ego.“ 3
2: Ho :38 8H 2:: mo A5322: e we :81: 953w 9:31:56 2: mo .63 < "Hg: ordwmh

Ampcooomv DE; commedm
Se com com 2: 0 com SN 03 o

 

2 x25 . m ono .8
. . _ . _ _ . _ o

 

 

. .ov

a ono . a. x25 .8

_ . _ _ . a . . . . o

. .om

. .ov

. m x030 .oo
. t . _ o

4 TON

. .ov

b x330 . N xodﬁo '00

_ . i . . . . .. o

. ............ .om

 

 

a x25

— L n h

 

 

 

 

a 53:8 . A c328 5e

 

 

Hamel 81119.19 eAnemurnQ

18

353w @988 3 meeommeboo on: «53. 3: BE? 233w 30:9» 3 mweommotoo
2:— .ooﬁow 2;. .ﬁacEmuoaS 8—88.33 owseuoéozom 2: E 93ch 180.. 3 2: mo
:33 ac.“ cE: Ho :osoeé e we :33 v.58» cage—=55 3: Ho .33 < "NA: enema

Ameaooomv DEE. DoerE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

oov com com cog o com com 003 o
r
03 x030 T m ono
a 328 a a. ono
w x030 . m x030
. rom
r row
a. #050 . m x030 .om
. p T iki.l|\|il\i»ii.\i O
. .om
o ono . H ono ﬂow

USHQL 81119.13 aAnemrunQ

19

633w we 3 meeoameuoo 2.: mice 0.3 223 23.9% 0988 3
accoamvtoo 3.: “.3335 v.3. 3:25.238 8:333; $26953 2: E 8.3.? .30.. 3
e... .3 :35 8.. 2:: .1. .532... e we :31: 2:3» o>_..e_..:.=o 2... Ho 33—: < 3...:— 3.3.3

Ampcooomv DEE. Dommoﬁm

 

 

 

 

 

  
 

 

 

 

 

 

 

 

 

 

 

 

 

 

8.. com com S. o com com oo. oo
iilii\n\i\ll\ll\ill E. ................ ..
is..-.-.......\i ..... ......... r
o. 5.25 . a D350
.
a 3.2.6 . a 3.38
a 35 . . m v.25
a. 35 ...... .. . N 3.2.8
F . _ h p h p b o
. . .om
‘ . ‘ .3.
o 3.8 . . 5.25 .oo

 

 

USHQL suieif) aAnemurnQ

20

352w ace 3 3:59.828 8:: 3—8 8.: 8—23 v.58» 30:8» 3
meaoamvuuoo 8:: @333. 8.3. .aeoEtcmxv 8.88.83 68.7328» 8.: E SEED 183 3
8.: no :98 .8.“ 2:: ho .5522; a 3 8x3 38% 8:33:85 8.: Ho 8E < 3;: enema

Ampaooomv 855. comma-.5

 

 

 

 

 

 

 

00.3V com com cod 0 com. com OOH
OH x350 . m x320
m x050 . a. x030

F b _ h pi p p F

 

m 3.26 . m v.25

— l—i » P p »

 

a. .330 . m 5.2.8

_i r b p — b

 

o 5.2.8 ,_ . . v.25

 

 

 

 

1135191, $1119.15) azxnemurng

21

Table 111.1: The average rate of intake for the various color combinations in the
preference study experiments. The * indicates a signiﬁcant t-test, 0.05 level.

 

Color Mean

Pair Grains / Sec Signiﬁcance

Green .04 :i: .07
vs.

Red .51 i .37

 

 

4:

 

Yellow .18 :i: .38
vs.

Orange .81 i .59

 

Orange .36 :1: .27
vs. N.S.
Red .28 :l: .22

 

Yellow .33 i .46
vs. N.S.
Red .49 :i: .46

 

 

 

 

 

22

value’ (A. V.) for each chick as follows:

A V _ (rateA — rateB/(rateA) if rateA > rateB
’ ’ _ (rateA — rateB/(rateg) if rateA < rates

The asymmetry value ranges between 0 (no preference, total symmetry), and 1 (spe-
ciﬁc preference, total asymmetry). Table 111.2 lists the asymmetry values for the

yellow-red and orange-red combinations. 1f individual chicks choose randomly be-

Table 111.2: Asymmetry values for Yellow-Red and Orange-Red color combinations.
The * indicates a signiﬁcant Mann-Whitney U test, 0.05 level.

 

 

 

 

I Pair I Yellow-Redl Orange-Red J
1 0.88 .47
2 0.87 .05
3 1.00 .07
4 1.00 .06
5 0.47 .26
6 1.00 .02
7 0.79 .40
8 0.59 .23
9 0.82 .51
10 0.56 .59
Signiﬁcance * N.S.

 

 

 

 

 

tween colors, then large degrees of asymmetry are expected to be rare. Because
asymmetry values can not take on negative values, a Poisson distribution corresponds
to random selection of colors. The standard deviation represents the dispersion of
individual responses. In order to generate an expected distribution to compare with
the experimental distribution of asymmetry values, a computer simulation was run
(refer to Appendix B). The simulation produced a Poisson distribution with standard
deviation 0.2. The magnitude of the standard deviation was based on the deviations
calculated from the experimental distributions. The experimental distributions were

compared to the expected distributions using a Mann-Whitney U test. For yellow-red

23

the comparison yields signiﬁcance and for orange-red no signiﬁcance (Mann-Whitney
U test, 0.05 level). This comparison. shows that individual chicks demonstrate speciﬁc
preferences for either yellow or for red when they are presented together. However,
there is no such asymmetry for the orange-red combination. In addition, a t-test (0.05
level) showed no signiﬁcant difference between the behavior of males and females in

the orange-red study.

Of the pairs of colors tested, yellow-red met the ﬁrst condition but not the second
(Tables 111.1 and 111.2). Only orange-red met both conditions (Tables 111.1 and 111.2).

Therefore, orange and red were chosen for use in the subsequent experiments.

Discrimination experiments showed that birds could distinguish orange and red
grains. Figure 111.5 illustrates the behavior of typical focal chicks over ﬁve successive
trials in each of the two reciprocal discrimination experiments. “Each Chick’s ﬁrst
trial confronted it with learning that one of the colors which had been previously
palatable during the acclimation was suddenly distasteful. During the middle trials,
the chicks showed a reluctance to eat anything, but began to choose proportionately
more of the tasty color. By the ﬁnal trials, the chicks had learned which of the
two colors was tasty and chose that color almost exclusively. A Friedman’s test
(because of an overwhelming block effect, AN OVA’s were not used) calculated for the
reciprocal discrimination experiments (Table 111.3) demonstrated that trial number
had a signiﬁcant effect on the relative rate at which grains were taken (0.05 level). The

Table 111.3: Results of the Friedman’s test showing the effect of trial number on the
relative rate at which grains were taken. A * indicates a signiﬁcant difference.

 

I I Orange Distasteful I Red Distasteful I

Calculated 20.72 * 1 1.2 *
Table 9.48 9.48

 

 

 

 

 

 

 

Cumulative Grains Taken

 

24

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(a) ( bl
Trial 1 ‘ Trial 1
Trial 2 Trial 2
Trial 3 Trial 3
f -___
60" Trial 4 Trial 4
40‘ “m _
20‘ i "’7‘
0 . a
601 Trial 5 i Trial 5
40 II ....-....._....._....__..,... _. .....
20i """
O 100 200 300 O 100 200 300 400

Elapsed Time (seconds)

Figure 111.5: The typical behavior of two focal chicks in discrimination studies over
the course of 5 successive trials. The data presented are the cumulative number of
grains taken as a function of time. The dotted line corresponds to orange grains
while the solid line corresponds to red. (a) orange grains distasteful; (b) red grains

distasteful.

25

speciﬁc quantities that were compared were trial number and (ratewange - ratend).
A Wilcoxon’s signed ranks test revealed a signiﬁcant difference between trials 2 (the
middle trials) and 5 (the ﬁnal trials) for the quantity (rate-"mg, - ratend) in both
discrimination experiments (0.05 level). The ability of the chicks to learn which color
was distasteful and their associated switch to foraging on the alternate color (as

demonstrated statistically) provides evidence that the two colors are distinguishable

by the chicks.

A runs test performed on data from the original preference experiments indepen-
dently suggests that chicks can discriminate orange and red grains. A run is deﬁned
as a set of one or more grains of a given color taken in an unbroken sequence. For
each focal chick, a t, value is calculated (Sokal and Rohlf 1981) from the number of
observed runs (nmm), and the total number of grains taken of each of the two colors

(n, and n2).

1'. ___ nruns — (2111112)/(I11 + 112) — 1
s (/2n1n2(2n1n2 — n1 — n2)/(n1 + n2)2(n1 + n2 — 1)

The value of t, is positive if there is an anti-correlation between successive choices
and negative if there is a tendency for the grains to be selected in runs. The average
calculated from the distribution of t, values from the 10 focal chicks was compared us-
ing a t-test to an average calculated from an expected distribution assuming random
selection of the two colors (Figure 111.6a and Table 111.4). This expected distribution
was generated by a computer simulation (refer to Appendix C). There is a signiﬁcant
tendency for the grains to be taken in runs (t-test, 0.05 level). This non-random selec-
tion of colors provides further evidence that the chicks are capable of distinguishing

between red and orange.

26

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

i
41 V (a)
3 .
2 2.
o I——‘ . ........
9 1 ‘ I
f. 0 . ........... a
it 6‘ i , (b)
O 4 . “—7 . .............. I
O 2 . ﬂ
E O ................... I """""""""
E 8 r < ............ ﬂ!
. C)
:3 .
Z 6
4 T .............. E
O -------------------- . ﬂ ............................ ﬂ
_ 4 _ 2 O 2 4
ts

Figure 111.6: The distribution of t, values (solid histograms) from (a) the preference
experiment, (b) the cryptic prey experiment with no prior introduction to a prey
type, and (c) the cryptic prey experiment with introduction. All experimental distri-
butions are compared to an expected distribution (dotted histograms) generated by
a computer simulation. The arrows indicate the average values of the experimental
distributions. The mean of the expected distribution is always zero.

27

Table 111.4: The average t, values in the runs test for the various studies, and the ex-
pected values if there were only random selection of grains. A * indicates a signiﬁcant
t-test, 0.05 level.

 

 

 

 

 

 

LTest I t. I :i: 5.1). l
preference -1.6 1.3 *
experiment
cryptic prey -1.2 1.2 *
no introduction
cryptic prey -1.9 1.3 *
with introduction
expected 0.00 0.01

 

 

 

 

(random selection)

 

B Experiments with Cryptic Prey

The experiments with cryptic prey test whether a change in an ability to ﬁnd one
cryptic prey type will interfere with the ability to ﬁnd another equally cryptic type.
Figure 111.7 displays the behavior of a typical focal chick without a prior introduction
to a single prey type, and Figure 111.8 displays the behavior of one chick introduced
to red grains and one chick introduced to orange grains. Three ANOVA tests
(Tables 111.5, 111.6, and 111.7), one each for no introduction, red introduction, and
orange introduction, demonstrated statistically that trial number had no signiﬁcant
effect on the relative rate (ratewgort —— ratecozorg) at which grains were taken (0.05
level). Therefore, only the ﬁrst trial was considered in the three subsequent t-tests
to compare the mean rates at which each color was chosen (Table 111.8). With or
without an introduction, the mean rates are not signiﬁcantly different (t-test, 0.05
level). Thus over the course of a single trial and over three successive trials, there is

no indication that specialization develops, hence no indication of interference between

prey types.

28

50 . Trial 1

4O . ______ _.

 

 

60 , Trial 2

40“ ............... . *
20 i

 

BOI Trial 3

Cumulative Grains Taken

 

 

 

o 100 260 360 460
Elapsed Time (seconds)

Figure 111.7: Experiments with cryptic prey - no prior introduction: The typical
behavior of a focal chick over the course of 3 successive trials. The data presented are
the cumulative number of grains taken as a function of time. Orange and red grains
were presented against a cryptic background. The dotted line corresponds to orange

grains while the solid line corresponds to red.

29

(61) ~ (13)

50. Trial 1 . Trial 1
4OI
20* _
0 5 I.-- 1,-
60‘
40*
20‘

O ' . . . -
50. Trial 3 . Trial 8

 

 

 

 

Trial 2

 

 

 

Cumulative Grains Taken

I 4’-
4 q — I d ‘

20-
/

Of " . . . . . . .
O 100 200 300 400 O 100 200 300 400 500

Elapsed Time (seconds)

 

 

 

 

Figure 111.8: Experiments with cryptic prey - prior introduction: The typical behavior
of a two focal chicks over the course of 3 successive trials. Chick (a) was introduced
to orange grains while chick (b) was introduced to red. The dotted line corresponds
to orange grains while the solid line corresponds to red. The spike near time zero

corresponds to the chick picking up introductory grains.

 

30

Table 111.5: ANOVA table for the experiment with no introduction demonstrating
that trial number had no signiﬁcant effect on the relative rates.

 

 

I Source I dF I SS IMS = SS/dF I

 

 

 

 

 

 

 

F 0.: 0.59
Blocks 9 .027 .003 F“ " 1 28

_ treat '
Treatment 2 .013 .007 Ftabzebtock (.05,9,18) 2.46 n.s.
Error 18 ~093 '005 F 5,, ( 05 2 18) 3 55 in s
Total 29 .133 .005 ‘° ""‘°‘ ' ’ ’ ' ° '

 

 

 

 

 

 

Table 111.6: ANOVA table for the experiment with orange introduction demonstrating
that trial number had no signiﬁcant effect on the relative rates.

 

 

ISource IdFI ss IMS=SS/dFI

 

 

 

 

 

 

F x 1.41
Blocks 5 .063 .013 u k
T 2 030 015 Ftreat 1'70
reatment ’ ’ Ftable,blocle (0575310) 3'33 11.5.
Error 10 ~089 ”09 F ( 05 2 10) 4 96 n 8
Total 17 182 011 table,treat ' 7 7 ' ' '

 

 

 

 

 

 

 

Table 111.7: AN OVA table for the experiment with red introduction demonstrating
that trial number had no signiﬁcant effect on the relative rates. The * indicates a
signiﬁcant difference at the 0.05 level.

 

 

I Source IdF I SS I MS = SS/dﬂ

 

 

 

 

 

 

F 0,, 5.31
Blocks 5 .406 .081 F:" ’: 2 93
Treatment 2 .089 .045 Ftable,blocle 0055,10) 3.33 *
Error 10 .153 .015 F ( 05 2 10) 4 96 n 5
Total 17 .648 .038 ‘“"‘""°°‘ ' ’ ’ ' ' ‘

 

 

 

 

 

 

 

31

Table III.8: Results of t—tests on trial 1 for the three experiments with cryptic prey
showing no signiﬁcant difference between the mean rates at which each color was

 

 

 

 

 

chosen.
Test Mean Red Rate Mean Orange Rate tmau tmu,
Trial 1 Trial 1

Cryptic Prey .07 :1: .04 .12 :1: .08 . 1.77 2.101 (.05,18)
N 0 Introduction

Cryptic Prey .12 :I: .13 .07 i .05 1.09 2.228 (.05,10)
Red Introduction

Cryptic Prey .12 i .09 .08 :1: .04 1.32 2.228 (.05,10)

Orange Introduction

 

 

 

 

 

 

 

Runs tests, described previously, provide evidence that there may be interference
between prey type over short time spans within trials. The distribution of t, values
calculated for the ten focal chicks in the experiment with no introduction and the
twelve focal chicks in the experiment with a prior introduction are compared to the
expected distributions assuming random selection of grains (refer to Appendix C) in
Figure [11.61) and III.6c. Table III.4 gives the average values for the experimental
and expected distributions which were compared using a t-test. In both experiments,
there is a signiﬁcant trend for the grains to be selected in runs (t-test, 0.05 level).
Since both grains and gravel were well—mixed prior to their distribution into the arena
and were mixed further as they were scattered, it was regarded as unlikely that the
runs resulted from heterogeneity in either prey distribution or microhabitat. No
signiﬁcant differences between the behavior of males and females were found in any

of the experiments with cryptic prey (t-test, 0.05 level).

Chapter IV

DISCUSSION

Allen (1989) cites two ways to distinguish the search rate and search image hypothe-
ses. The ﬁrst, measuring actual search rates during the course of a trial, would show
an increase in search rate in the case of the search image hypothesis and a decrease
in the case of the search rate hypothesis. This approach is complicated by the dif-
ﬁculty of separating the parameters that make up search rate. In my experiments,
overall rates between experiments cannot be compared either. Although the rate at
which chicks foraged was depressed when colored grains were presented on a cryptic
background as compared to the rates for presentation on a conspicuous background,
and this depression would appear to be evidence for search images, both the search
rate and search image hypotheses predict this result. With grains more difﬁcult to
ﬁnd on a cryptic background, the development of a search image would be prolonged
(search image hypothesis) or there would be a decrease in search rate (search rate

hypothesis).

The goal of this series of experiments was to test for evidence of interference as
predicted by the search image hypothesis (Guilford and Dawkins 1987), the second
approach cited by Allen (1989). When chicks were presented with orange and red
grains on a background colored to make both cryptic, they took the two colors in

equal proportions. This result suggests that a chick’s ability to detect and capture
32

33

one form of cryptic prey does not automatically interfere with its ability to detect
another equally cryptic prey type over the course of an entire trial. The same result
was obtained when chicks were given prior experience with one of the two prey just
prior to experiencing both types together. Interference, therefore, does not appear to
be a necessary consequence of learning to forage for cryptic prey over the course of
a trial or over consecutive trials, at least when the prey are offered in nearly equal

proportions.

The fact that no interference was found over the course of an entire trial does not
exclude the possibility that the interference implied by the search image hypothesis
would be manifested under different conditions. At least four modiﬁcations in my
experiments might have revealed evidence of interference. First, the prey types I
offered were similar in all respects except color; perhaps some other visual feature,
such as size or shape, is more relevant in the development of perceptual specializations.
Croze (1970), for example, found some evidence for search image based on shape when
presenting crows with cockles and mussels, though it is difficult to separate his results

from a simple preference for mussels.

Second, the length of the prior exposure of chicks to a single prey type may need
to be longer or the introductory prey may need to be cryptic instead of conspicuous.
Den Boer (1971), in a study with a Coal Tit given lengthy prior exposure to the green
caterpillar larvae which occurs cryptically in its natural habitat, found that the tit
preyed upon the green cryptic larvae almost exclusively when presented with cryptic
green and conspicuous yellow prey. However, since only one choice was given in a
trial, the aversion for yellow could result from the novelty of the prey item. On the
other hand, Croze (1970), cited by Krebs (1973) as the most complete evidence for
search images, found that Carrion Crows given lengthy prior experience on ‘standard

red’ mussels chose equal numbers of ‘standard red’ and ‘red-red’ mussels when they

34

were presented together on a cryptic background in equal proportions. Presumably,
the crows could tell the difference between the two types based on the fact that they
always chose ‘standard red’ ﬁrst when a pair of mussels were presented to them.
Though no speciﬁc mention is made that the two colors were of equal crypticity,
their reﬂectance spectra (Croze 1970, page 46) are quite close. Though Den Boer’s
study may suggest that lengthy prior exposure could result in search image formation,

other studies suggest that a search image can be formed quickly (Croze 1970, Dawkins

1971a,b, Murton 1971).

Third, the chicks I tested encountered the alternative prey types in equal pro-
portions. In this case, the search image hypothesis predicts that a chick could not
gain enough prior experience on (and thus an enhanced ability to detect) one color
to interfere with its ability to ﬁnd another. On the other hand, a study by Murton
(1971) in support of search images showed that Wood Pigeons offered equal propor-
tions of two seed types at approximately equal crypticity with respect to each other
specialized on one seed type or the other. The search rate hypothesis predicts no
interference in a Chick’s ability to choose one color over another no matter what the
proportions and thus predicts that the proportions of each color taken should reﬂect

the proportion at which each color was offered (Figure IV.1).

Perhaps, though, chicks would only develop perceptual specializations favoring an
extremely common type over an extremely rare type. Indeed, in some formulations
of the search image hypothesis (Den Boer 1971, Murton 1971), differences in relative
abundance, hence in relative rates of encounter, of alternate prey types are assumed
to play a crucial role in the formation of search images. For example, Den Boer
(1971) presented Great Tits and Coal Tits with different proportions of green and
yellow larvae. The interpretation of the results, however, is confounded by the unequal

crypticity of the prey types. The observed ratiOs of the captured prey can be explained

35

 

 

 

100 —-

: -------- Search Rate /
I — Search Image

‘3 80 — ’

av

.24 _

cu _

E >—

H 60 —

>§ _

m :—

1., _

0.. _

.1.) 4O "—

a .—

Q) .—

0 .—

s-« _

d‘j 20 —-
L

O L

O 20 4O 6O 80 100

Percent Prey 1 Available

Figure IV.1: Predictions of the search image model and the search rate model as a
function of the relative densities of the two cryptic prey species.

36

entirely based on differences in detection efficiency. Murton (1971) presented Wood
Pigeons with different proportions of supposedly equally cryptic tic beans and maple
peas and found that the birds had a signiﬁcant tendency to prey almost exclusively
on tic beans. Once again, the interpretation of the results is confounded by the failure

to control for equal preference and equal crypticity for the two prey types.

Finally, this experimental approach might have yielded evidence of perceptual
specialization if applied in a more naturalistic context to a species more certain to

have been under strong natural selection to exploit cryptic prey.

Although the experiments revealed no evidence of interference over the duration
of entire foraging bouts, evidence for a temporary form of interference is suggested
by the results of the runs tests, which showed that the choice of grain color has some
dependence on previous choices. This result can not be fully explained by the search

rate hypothesis, and may indicate the formation of a short-lived search image.

These short-term specializations expressed as runs on a speciﬁc prey type must
make a local impact on the relative abundance of different prey types in order to
contribute to apostatic selection. For example, if the predator runs only on one prey
type or for a signiﬁcantly longer period of time on one prey type during the time

spent in a patch, the overall effect would be selection against that type.

It is important to stress the value of this approach for further studies of the
search image hypothesis. Of previous laboratory studies on search images, only that
of Pietrewicz and Kamil (1979, 1981) has presented predators with two cryptic prey
types and is, therefore, directly comparable to this study. However, in their study, the
two prey types used were not of equal crypticity. In addition, the prey were presented
as photographic images on successive slides and not as food items that could be eaten

on being recognized as such. This complicates the relationship between performance

37

accuracy and response time to a stimulus since the birds must also learn about the
experimental setup. These latter two points have enabled Guilford and Dawkins

(1987) to reinterpret the data using the search rate hypothesis.

Certain of the more naturalistic studies reviewed by Krebs (1973) did incorporate
the approach of using more than one prey type, though these studies also have their
shortcomings. Not only can Croze’s work (1970) be explained using the search image
hypothesis, but the search rate hypothesis can be applied as well. For example, when
crows were presented with red and black mussels, the birds ﬁrst took all the red
mussels and then all the black mussels (Croze 1971, page 40). Croze interprets this as
evidence for search images. But, since red mussels are more conspicuous then black
ones, the search rate hypothesis would predict that the search rate would be set by
the red mussels. Black mussels would be overlooked. Krebs (1973) criticizes Murton’s
work (1971) on the basis that individual birds could have speciﬁc seed preferences.
At the same time, he states that since independent ﬂocks of pigeons did not exhibit
exactly the same behavior, individual seed preferences are unlikely. Den Boer’s study
(1971) has the two problems previously mentioned. In the experiments without prior
exposure, there is unequal crypticity of the prey types, and in the experiments with

prior exposure, there is the issue of novelty.

Controlled experiments similar to those outlined in the present work, but per-
formed at proportions of prey types other than 50:50, would isolate the question
of interference from the search rate hypothesis and would lead to a better under-
standing of the role of interference in predator behavior. In particular, controls for
preference and crypticity would eliminate many of the problems found in previous
studies. Finally, similarly controlled experiments should be extended to naturalistic
systems in order to reduce the problems associated with the artiﬁciality of behavior

in a laboratory system.

10

20

30

90

Appendix A

Data Conversion Program

this
run

program timer
program will create a time file for each experimental

real timel,time2

integer status,icount

character ctime*32,tit1e*80,filename*80,input*1
type *,’ Please input a title for this run.’
accept lO,title

format(a)

encode(20,20,fi1ename)

format(a14,’.dat’)
open(10,fi1e=title.form=’formatted’,status=’new’)
type *,’ Hit <CR> to start to run, and then hit <CR>
l for every strike.’

type *,’ To finish the run, type E’
accept lO,input

type *,’ The run has been started.’

call libSinit_timer

time1=secnds(0.)

icount=0

accept lO,input

if(input.eq.’E’.or.input.eq.'e’)then
time2=secnds(0.)

type *,’ Test ended at ’,(time2-timel),’ seconds’
write(10,*)’ Test ended at ’,(time2-timel),’ seconds’
go to 90

end if

time2=secnds(0.)
icount=icount+1

type *,’ Event Number
1 ,(timeZ-timel)
write(10,40)icount,(timeZ-timel)
format(i6.f10.4)

call 1ib$show_timer(,2)

go to 30

end

,icount,’ Elapsed Time

38

9

ll

39

subroutine title_read(tit1e,argument)
character title*l

integer argument

dimension title(32)

type 11,title

format(a)

type *,’ Argument=’,argument

return

end

=

 

Appendix B

Asymmetry Value Simulation

C
C
C

10

program CoinFlips

This program randomly flips a coin a given
number of times, and
then determines the asymmetry parameter.

implicit none
integer iseed,i,j,k,Mat(10),Ntrials,Nf1ips
real Nh,Nt,AP,rnd

type *,’ Please input a random seed.’
accept *,iseed
type *,’ Input Number of trials’
accept *,Ntrials ‘
type *,’ Input Number of flips per trial’
accept *,Nflips
do i=l,Ntrials
Nt=0
Nh=0
do j=l,Nflips
rnd=ran(iseed)
if(rnd.gt.0.5)then
Nh=Nh+1
else
Nt=Nt+1
end if
end do
Ap=0
if(Nh.gt.Nt)AP=(Nh-Nt)lNh
if(Nh.lt.Nt)AP=(Nt-Nh)th
k=int(10.0*Ap)+1 ‘
if(k.gt.10)k=10
mat(k)=mat(k)+1
end do
do k=l,lO
type *,mat(k)
mat(k)=0
end do
go to 10
end

40

Appendix C

Runs Test Simulation

program CoinFlipsZ
This program randomly flips a coin a given number
of times, and then determine the number of runs
in the sample. The probability of getting a .true.
is not equal to the probability of getting a
.false.
implicit none
integer iseed,i,j,k.Mat(-20:20),Ntrials,Nf1ips
real Nh,Nt,AP,rnd,runs,H,T,Ts,prob,Num,Dem
logical list(100)

("30000

10 type *,’ Please input a random seed.’
accept *,iseed
type *,’ Input Number of trials’
accept *,Ntrials
if(Ntrials.eq.O)go to 99
type *,’ Input Number of Heads and number of Tails.’
accept *,H,T
Nflips=H+T
Prob=H/float(Nflips)
do i=l,Ntrials
Nt=0
Nh=0
do j=1,Nflips
rnd=ran(iseed)
if(rnd.gt.Prob)then
Nh=Nh+1
list(j)=.true.
else
Nt=Nt+1
list(j)=.false.
end if
end do
runs=l
do j=2,Nf1ips
if(List(j-1).ne.List(j))Runs=Runs+l
end do
c type *,’ N1 ’,Nh,’ N2 ’,Nt,’ Runs ’,Runs
Num=(runs-((2*Nt*Nh)/(Nt+Nh))-1)
Dem=sqrt(2*Nt*Nh*(2*Nt*Nh-Nt-Nh)/
l ((Nt+Nh)**2*(Nt+Nh-l)))

41

99

42

Ts=NumlDem
k=nint(l0.0*Ts)
if(k.gt.20)k=20
if(k.1t.-20)k=-20
mat(k)=mat(k)+1

end do

do k=-20,20
type *,float(k)/10,mat(k)

end do

go to 10

do k=-20,20
write(99,*)float(k)/10,mat(k)

end do

end

43

LIST OF REFERENCES

Allen, J .A. 1974. Further evidence for apostatic selection by wild passerine birds: train-

ing experiments. Heredity, 33, 361-372.
Allen, J.A. 1989. Searching for search image. Trends in Ecol. and Evol, 4, 361.

Bowmaker, J .K. and Knowles, A. 1977. The visual pigments and oil droplets of the

chicken retina. Vision Research, 17, 755-764.

Clarke, B.C. 1962. Balanced polymorphism and the diversity of sympatric species. In:
Taxonomy and Geography (Ed. by D. Nichols), pp. 47-70. Oxford: Systematics

Association.

Clarke, B.C. 1979. The evolution of genetic diversity. Proc. Roy. Soc. Lond. B., 205,

453-474.

Cooper, J.M. 1984. Apostatic selection on prey that match the background. Biol. J.

Linn. Soc., 23, 221-228.

Cornell, H. 1976. Search strategies and the adaptive significance of switching in some

general predators. Am. Nat, 110, 317- 320.
Croze, H..I. 1970. Searching image in carrion crows. Z. Tierpsychol., 5, 1—85.

Dawkins, M. 1971a. Perceptual changes in chicks: another look at the ‘search image’

concept. Anim. Behav., 19,556-574.

Dawkins, M. 1971b. Shifts of ‘attention’ in chicks during feeding. Anim. Behav., 19,

575—582.

Den Boer, M.H. 1971. A color polymorphism in caterpillars of Bupalus piniarius (L.)

(Lepidoptera: Geometridae). Netherlands J. Zool., 21, 61-116.

Gendron, RP. 1986. Searching for cryptic prey, evidence for optimal search rates and

the formation of search images in quail. Anim. Behav., 34, 898—912.

44

Gendron, RF. and Staddon J .E.R. 1983. Searching for cryptic prey: the effect of search

rate. Am. Nat, 121, 172-186.

Gittleman, J.L. and Harvey, P.H. 1980. The evolution of conspicuous coloration, some

experiments in bad taste. Anim. Behav., 28, 897-899.

Guilford, T. and Dawkins, M.S. 1987. Search images not proven: a reappraisal of recent

evidence. Anim. Behav., 35, 1838-1845.

Kettlewell, H.B.D. 1955. Selection experiments on industrial melanism in the Lepi-

doptera. Heredity, 9, 323-342.

Krebs, J .R. 1973. Behavioral aspects of predation. In: Perspectives in Ethology. Vol. I
(Ed. by P.P.G. Bateson & P.H. Klopfer), pp. 73-111. London: Plenum Press.

Lawrence, E.S. 1985a. Evidence for search image in blackbirds (Turdus merula L.),

short-term learning. Anim. Behav., 33, 929-937.

Lawrence, E.S. 1985b. Evidence for search image in blackbirds Turdus merula L., long-

term learning. Anim. Behav., 33, 1301-1309.

Lawrence, ES. 1986. Can great tits (Paras major) acquire search images? Oikos, 47,

3-12.
Lawrence, ES. and Allen J.A. 1983. On the term ‘search image’. Oikos, 40, 313-314.

Murdoch, W.W. and Oaten, A. 1975. Predation and population stability. Adv. Ecol.

Res, 9, 1-131.

Murton, R.K. 1971. The signiﬁcance of a speciﬁc search image in the behaviour of the

wood pigeon. Behaviour, 40, 10-42.

Pietrewicz, A.T. and Kamil, A.C. 1979. Search image formation in the bluejay (Cyanocitta

cristata). Science, 204, 1332-1333.

45

Pietrewicz, A.T. and Kamil, A.C. 1981. Search images and the detection of cryptic prey:
an operant approach. In: Foraging Behavior (Ed. by A.C. Kamil & T.D. Sargent),
pp. 311-331. New York and London: Garland STPM Press.

Ruiter, L. de 1952. Some experiments on the camouﬂage of stick caterpillars. Behaviour,

4, 222-232.
Sokal, RR. and Rohlf, J.J. 1981. Biometry. 2nd edn. San Francisco, W.H. Freeman.

Tinbergen, L. 1960. The natural control of insects in pine woods. I. Factors influencing

the intensity of predation by songbirds. Arch. Ne’erl. Zool., 13, 265-343.

 

RN STATE UNIV.

51311er! rmmrunlllllﬂf

2930.790