_ mmmlmlL lllmm 7 5111! m L-

                  

 

LIBRARY

Michigan State
University m

 

This is to certify that the

thesis entitled
SEQUENCE EFFECTS
IN THE ANTIPREDATOR BEHAVIOR
OF THE HOUSE CRICKET (ACHETA DOMESTICUS L.)

 

presented by

Thomas E. Hagaman

has been accepted towards fulfillment
of the requirements for

EDA. D degree 1n PSfiLAoIDQ?

Major professor

Date )7 A/av. H77

0-7639

 

HAT ? :3 1.991
1 1h?

 

SEQUENCE EFFECTS
IN THE ANTIPREDATOR BEHAVIOR

OF THE HOUSE CRICKET (ACHETA DOMESTICUS L.)

 

BY

Thomas E. Hagaman

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Psychology

1977

ABSTRACT

SEQUENCE EFFECTS
IN THE ANTIPREDATOR BEHAVIOR
OF THE HOUSE CRICKET (ACHETA DOMESTICUS L.)

 

BY

Thomas E. Hagaman

A series of experiments was conducted on anti-

predator behavior in the house cricket (Acheta domes-

 

Eiggg L.). First, cricket predation by rats was observed.
These insects responded with autotomy (reflex separation
of a metathoracic leg) or thigmotaxis, freezing, hopping,
struggling, freezing after capture, and immobility.

Then the eliciting stimuli for each of these behaviors
were experimentally investigated. Finally, eliciting
stimuli for immobility and autotomy were presented at
different points in the normal sequence of behavior with
the result that the eliciting stimuli were maximally
effective only at their appropriate positions in the
antipredator sequence. This provided evidence for a

hierarchy of antipredator behavior in A. domesticus.

DEDICATION

This dissertation is dedicated to the memory of

Dr. Stanley C. Ratner who directed this research until

his death in December, 1975.

ii

ACKNOWLEDGMENTS

The author is pleased to acknowledge the advice

and encouragement of Dr. L. I. O'Kelly, Dr. J. Zacks,

Dr. R. T. Cardé, and Dr. N. Schmitt.

iii

TABLE OF CONTENTS

Page
LIST OF TABLES . . . . . . . . .. . . . V
LIST OF FIGURES O O O O O O O O O O O 0 Vi
Chapter
I. INTRODUCTION . . . . . . . . . . 1
II. EXPERIMENTS . . . . . . . . . . . 9
Experiment 1: Behaviors Elicited by
Actual Predation . . . . . . . . 9
Experiment 2: Early Components of
Antipredator Behavior . . . . . . . 15
Experiment 3: Freezing in Response to
Simulated Predation . . . . . . . 18
Experiment 4: Escape in Response to
Simulated Predation . . . . . . . 23
Experiment 5: Struggle and Captive Freeze
in Response to Simulated Predation . . 31
Experiment 6: Immobility in Response
to Simulated Predation . . . . . . 33
Experiment 7: Sequence Position as a
Factor in the Effectiveness of an
Eliciting Stimulus . . . . . . . . 37
Experiment 8: Cricket Size and mmo-
bility Duration . . . . . . . . . 45
Experiment 9: Autotomy . . . . . . . 47
SUMMARY.............. 53
BIBLIOGRAPHY . . . . . . . . . . . . . 65

iv

Table

LIST OF TABLES

Behavior Classification . . . . . .

Behavior at Three Data Collection
POintS O O O O O O O O O O 0

Procedure Outline for Experiment 4 . .
Escape Behavior; Raw Data . . . . .
Escape Behavior; Percentage Distribution
Procedure Outline for Experiment 3 . .
Results of Experiment 7 . . . . . .
Procedure Outline for Experiment 9 . .

Separation Frequency by Group and Type .

Page
12

20
27
28
29
41
44
50
51

LIST OF FIGURES

Figure Page
1. Antipredator Behaviors . . . . . . . 57
2. Thigmotaxis . . . . . . . . . . . 58
3. Freezing . . . . . . . . . . . . 59
4. Freezing in the Presence of a Rat . . . . 50
5. Escape Behavior . . . . . . . . . . 61
6. Captive Freeze . . . . . . . . . . 52
7. Immobility . . . . . . . . . . . 63
8. Autotomy . . . . . . . . . . . . 64

vi

CHAPTER I

INTRODUCTION

An expanding knowledge of the variety of
behavioral strategies employed by animals in meeting the
problems of survival as well as the principles by which
these behaviors are organized is important to the con-
tinued development of a truly comparative psychology.
Insects provide both a fascinating variety of behavior
for study and many Opportunities to investigate the
manner in which components of behavior are assembled.

One type, or class, of behavior in which insects
show this variety is antipredator behavior. For example,
the aphid is usually considered to be a behaviorally
uninteresting insect. Yet Dixon (1958) reported five
components of antipredator behavior in response to
coccinellid adults and larvae (ladybird beetles). The
components were walking, dropping from the plant, kicking,
dropping a leg, and "waxing." In waxing behavior, an
oily liquid is secreted from the rear of the abdomen.
This substance quickly solidifies to wax and if it comes
into contact with the head of the predator its mouth-

Parts are sealed in wax. This is a novel and effective

defense. Dixon reported that it took the coccinellids

from 42 to 94 minutes to remove the wax and free their

mouthparts, thus allowing the aphids ample time to escape.
The study of behaviors like the above has

delighted naturalists from many disciplines and is worth-

while as an end in itself. In addition, the presence of

complex behaviors in such "simple" animals can provide

us with powerful tools for the analysis of behavior.

Thus the present study had two purposes. First, it

explored the behaviors used by a cricket, Acheta domes—

 

Eigug, in response to a predator. Second, it investi-
gated some relationships among these behaviors and the
manner in which they are ordered.

The complex, and apparently ordered, chains of
behavior in animals, especially insects, have long
fascinated students of behavior. These chains have been
labeled, and too often explained, by using the term
instinct. J. Henri Fabre, a contemporary and corres-
pondent of Charles Darwin, devoted his life to this
study. Fabre's observations led him to believe that,
once activated, an instinctive sequence of behavior must
be followed to its conclusion with very little variation.
An example may be found in Teale (1949):

Because one thing has been done, a second thing
must inevitably be done to complete the first
or to prepare the way for its completion, and

the two acts depend so closely upon each other
that the performing of the first entails that

of the second, even when, owing to casual
circumstances, the second has become not only
inOpportune but sometimes actually opposed to
the insect's interests. (p. 61)

Fabre's work is full of examples of what he
called the "ignorance of instinct." This refers to
situations in which normally adaptive behavior becomes
grotesquely maladaptive due to the insects' inability to
depart from set patterns of behavior. He described how
a hunting wasp captures and feeds honey bees to its
larvae. After capturing a bee, the wasp typically eats
the pollen and nectar carried by the foraging bee before
taking it to her larvae. If the wasp herself is captured
during the consummatory stage of her feeding behavior she
is unable to respond appropriately to the new situation.
Again, Teale (1949) provides an example from Fabre:

I have seen the wasp, with her prey, seized by
the Mantis: the bandit was rifled by another
bandit. And here is an awful detail: (while
the Mantis held her transfixed under the point
of the double saw and was already munching her
belly, the WaSp continued to lick the honey of
her Bee, unable to relinquish the delicious
food even amid the terrors of death. (p. 57)

Fabre chronicled inumerable examples like the
above, but he did not engage in the experimental study
of possible mechanisms underlying the behaviors that he
observed. He assumed, on the basis of his observations,
that there was an actual sequence or hierarchy involved

in.the behaviors that he studied. It is very easy to

Observe an animal performing a set of behaviors having

a common end function and conclude that an ordered
sequence of behavior exists. However, even if behavior
B always follows behavior A no causal relationship is
necessarily implied. A third factor or a time delay
function could be involved.

Recently, an increasing number of researchers
have been attempting to go beyond simple description of
the behavioral responses of animals to given stimuli.
Descriptions of behavior are becoming more quantitative
and, where apparently ordered sequences of behavior occur,
transition probabilities between all behavioral compo—
nents in the sequence are studied. Once a researcher has
determined the components in such a sequence, questions
can be asked about the manner in which the components are
assembled. Baerends (1976) reviewed some of the issues
involved in the study of the hierarchical and functional
organization of behavior and pointed out the need for
more research in this area. Sibley and McCleery (1976)
carried the analysis one step farther by moving from the
ordering of components in a single class of behavior such
as feeding to a consideration of how two classes of
behavior (in this case feeding and drinking) interact.
McFarland (1974) and his associates have reported progress
in the develOpment of SOphisticated techniques for the
analysis of behavior based on systems analysis and

decision theory. However, before these techniques can be

fully utilized much basic research on behavior in a
variety of species needs to be conducted in order to
identify appropriate research preparations for closer
analysis.

The present study represents an attempt to ana-
lyze a selected class of behavior in a single species
and to obtain information about how behavior in this
selected case is organized. The rationale for the
present series of experiments was to identify the beha—
viors emitted by an animal in a given class of behavior
and the proximal stimuli that are at least occasionally
sufficient to elicit these behaviors. In the next stage
of the study these stimuli were presented at their normal
points in the apparent sequence of behavior and at
earlier or later points. If these proximal stimuli
always elicit responses with the same probability and
topography no evidence is obtained for the existence of
an order effect or a hierarchy of behavior. If, on the
other hand, a stimulus is maximally effective only at
its appropriate position in the sequence, evidence is
Iprivided for an organization of behavior in which pre—
Kriously emitted behaviors combine with incoming proximal
sstimuli to determine an animalfs next response.

Acheta domesticus, the house or domestic cricket

 

(formerly called Gryllus domesticus), was chosen as a

sgeneral research animal after a literature review and a

series of pilot studies found that this species has a
wide variety of interesting behaviors and is easy to

obtain and maintain. A. domesticus is an Old World

 

insect that has been introduced into the United States
and Canada. An apparent sibling Species, until recently

thought to be a strain of A. Domesticus, occurs in

 

Pakistan. However, Ghouri and McFarlane (1957) were
unable to mate individuals from Pakistan with individuals

of A. domesticus from Canada. They found no morphological

 

differences between the two populations although the
Canadian crickets were generally larger than crickets of
the Pakistani strain. They did find sufficient behavioral
differences between the two strains to suggest that

they are reproductively isolated. Females did not
respond to males from the other locality except that
Canadian females "often devoured the smaller males of

the Pakistani strain." Ghouri and McFarlane found that
reproductive isolation in this case evidently consists in
a failure of insemination but the precise mechanisms
involved are unknown. Alexander (1962, 1964) has studied
the behavioral isolation mechanisms in Gryllidae includ-

ing Gryllus veletis and g. pennsylvanicus.

 

 

A. domesticus is commonly found in basements and

 

refuse piles. Bate (1969a, 1969b, 1970, 1971) has
studied this cricket both in the laboratory and in its

natural habitat in British dumps. Bate's ecological

research suggests that a laboratory environment would not
be a tremendous distortion of the natural habitat of this
species.

Khalifa (1949) documented the sexual behavior of
the house cricket. The male produces a calling song and
if a female approaches when the male has a spermatophore
he changes to a mating song which, compared to the cal-
ling song, is softer but higher in frequency and is
accompanied by side to side rocking movements of the
body. Next the female mounts the male and/or the male.
pushes himself underneath the female and a spermatophore
is transferred to the female. After c0pu1ation the male
"watches" the female for a period that may last for over
an hour.

The female oviposits in moist soil. The eggs
hatch after 14 days at 30°C. and the crickets are
sexually mature after 30 to 60 days. There is no care
of the young behavior. The entire life cycle may be
easily studied in the laboratory.

This insect is basically a scavenger and is
occasionally cannibalistic. Resting, grooming, and
possibly exploration occur but these behaviors have
not been systematically studied.

A, domesticus was selected for research on

 

antipredator behavior after pilot studies found that

these insects reSpond to human predation with a defensive

repertoire that includes walking, freezing, dropping
appendages, hopping, struggling, and immobility.

There is little information in the literature on
the defensive behavior of A. domesticus. Brousse-Gaury
(1958) reported the occurrence of autotomy, the reflex
separation of a metathoracic leg, in this insect. This
report included a detailed description of the structures,
muscles, and nerve impulses involved in autotomy in
A. domesticus. The trochanter of the metathoracic leg
is reduced to a thin ring that is inserted into the
interior of the coxa. The trochanter-femur joint is
immobile and together these two segments form a monocon-
dyle joint with the coxa. Three muscles pass through the
coxa and attach to the trochanter. Contraction of these
muscles results in a rupture of the leg at the weak
trochanter-femur joint. Brousse-Gaury found autotomy
to be "capricious and irregular" and was unable to make

any statements about possible eliciting stimuli.

 

CHAPTER II
EXPERIMENTS

Experiment 1:
Behaviors Elicited by
Actual Predation

 

The purpose of this experiment was to obtain .
descriptive data on the topography, frequency, order,

and possible function of each behavior associated with

 

predation.

Background Observations

 

A series of pilot studies was conducted in order
to identify a predator species for use in this experiment.
A total of several hundred A. domesticus was exposed to
predation by rats, mice, cats, ferrets, quail, a parrot,
frogs, toads, snakes, and humans. The hooded rat was
selected for more controlled laboratory study because
rats are ecologically valid predators and these animals
were the most reliable eliciters of the greatest variety

of antipredator behaviors from the crickets.

Method
Subjects.--Two hundred mature Acheta domesticus

from a larger group of crickets purchased from a

10

commercial supplier were used in this experiment. The
insects were maintained in glass terraria (25/50/30 cm.
deep) containing ground corncob bedding and cardboard

egg case inserts that provided shelter and resting sites.
Ground chicken mash and water were always available.
Adult crickets typically live for two to four weeks

under these conditions. These experiments used crickets
that were a few days to three weeks old. The laboratory
temperature was set at 20°C. and never varied more than
4°. A 12:12 light cycle (50 lux/O lux) was maintained.
Six hooded rats were obtained from a laboratory colony at
Michigan State University. The rats were housed indivi-
dually in glass terraria similar to those used.for crickets.

The two sets of terraria were never interchanged.

Apparatus.——The apparatus for a given test con—
sisted of a rat terrarium, a cricket terrarium, and a
removable passageway that allowed a rat to climb into
the cricket terrarium. This passageway was constructed
of Habitrail tubing sections. When the passageway was
placed between the terraria, a rat could climb a 45 cm.
vertical section, cross a 45 cm. horizontal section, and
descend into the cricket area through another vertical

section. The crickets could not climb through the tubes.

Procedure.—-A test trial consisted of a rat

having access to a cricket terrarium containing 10

ll

crickets for a period of 15 min. Four tests were con-
ducted on each of five successive days. The following
general procedure was used for each of the 20 tests.

Ten crickets (five males and five females) were
haphazardly selected from the colony and placed with
food, water, and fresh bedding in a terrarium and allowed
to acclimate for 20 to 24 hours. Then a cricket terra-
rium was placed on a table next to a rat terrarium for
another acclimatization period that lasted 15 min. Then
the passageway was placed connecting the two terraria.

The test was terminated after 15 min.

Measurement.——A tape recorder was used in an
attempt to record as many observations as possible.
Attention was focused on the behavior of the crickets

nearest to the rat.

Results

Table 1 presents the behaviors observed in
response to a rat. One hundred and five crickets were
captured and 10 of these were captured twice so 115
captures were observed. Other behaviors such as stridu-
lation, ovipositing, and grooming were common before the
rat entered the cricket terrarium but all such behaviors
ceased as soon as a rat entered the cricket terrarium.

Autotomy was never observed in the absence of

the rat. Four individuals may have shown this response

 

12

Table l.--Behavior Classification.

 

 

Behavior Definition

Autotomy Reflex separation of a metathoracic
leg.

Walk Coordinated locomotion.

Freeze Upright posture with no movement.

Hop A jump of more than 1 cm.

Struggle General wiggling and pushing with
legs.

Captive freeze Cessation of movement while being
held by a rat.

Immobility Ventral side up with no movement

for at least 3 sec.

 

but it was never possible to be absolutely sure that the
rat had not torn the leg off. However, each of these
cases involved a metathoracic leg and there were no
signs of avulsion.

Walking was common both in the presence and
absence of a rat. There were no obvious differences
between walking in the presence and absence of a rat.

Freezing was also observed both beforeand after
the arrival of a rat. There was some indication that
freezing may be distinguishable from resting behavior
on the basis of antennal position. Crickets freezing
in response to a rat were frequently observed with their
antennae in a raised position.

Hopping was observed in only one individual

before the rat entered the cricket terrarium. Hopping

13

was a very common response to the close proximity of a
rat. Hopping seemed to be closely related to freezing
in that most h0ps started from a freeze posture and after
one or more hops the freeze posture was reassumed for at
least a few seconds before walking occurred.

Struggling was observed in 13 of the captures
and was always of short duration (one to three sec.).
One of these crickets escaped after struggling but it
was recaptured a few seconds later. Captive freeze, or
the cessation of struggling, was observed in 12 of the
above captures. Crickets remained in this posture until
they were killed or released at which time they either
moved away or showed immobility.

Immobility was never observed during the absence
of the rat but did occur in 17 individuals. This beha-
vior was only observed after contact with the rat. In
16 cases the cricket was captured in the rat's mouth or
held down on the bedding by a paw, released ventral side
up, and remained immobile. In one case an immobility
response was elicited when a rat sat on a cricket. Four
of the immobility reactions were terminated by the
cricket righting itself and walking away. The durations
of these four reactions were 13, 15, 24, and 31 sec. One
immobility reaction was terminated by contact with the
experimenter after 400 sec. of immobility. The rest of

the immobile crickets were picked up by a rat after

l4

4 to 266 sec. In each case struggling was resumed for
a few seconds. The rat eventually ate these eleven
crickets.

Table 1 lists freezing, captive freezing, and
immobility as separate behaviors. Yet the classification
at this point in the research was partially based on
structural considerations and the concurrent behavior of
a rat. For example, without recourse to physiological
research there was no way to rule out the possibility of
upright immobility being confused with freezing, or even
resting. However, even at this point, there were some
indications that the behaviors were different. No
instances of freezing were observed in which a rat con—
tacted a cricket and the cricket failed to respond,
almost always with hopping. Similarly, no instances of
immobility were observed in which a cricket hopped
although a cricket would occasionally show some movement.
A distinction was also made between captive freezing and
immobility. It may be that these two behaviors are best
considered as points on a continuum. Some, but not all
instances of captive freezing were followed by immobility.
Photographs of the behaviors described here may be

found in the Summary section.

15

Experiment 2:
Early Components of
Antipredator Behavior

 

 

 

The results of Experiment 1 clearly identified
several consummatory behaviors that are commonly emitted
in response to predation. The early components of anti-
predator behavior remained unclear. The results of
Experiment 1 were used to formulate the following hypo—
theses about these early responses to a predator.

1. An early component of the defensive sequence
is the cessation of behavior in other classes.

2. The approach of a predator increases the
frequency of freezing.

3. The approach of a predator increases the
probability that a cricket will orient
towards thigmotactic stimulation.

Method

Subjects.--Twenty adult A. domesticus (10 males

 

and 10 females) were used in this experiment along with
the female rat that was the most reliable predator in
Experiment 1. The crickets were maintained as described

in Experiment 1.

Apparatus.--The apparatus from Experiment 1 was

 

used in this experiment. A cardboard screen (57/70 cm.
with a 5/5 cm. hole in the center) was placed between

the experimenter and the crickets in order to minimize

16

the amount of experimenter movement visible to the

cricket in the test terrarium.

Procedure.--The same procedure was used as was
reported in Experiment 1 with the following exceptions.
Only 1 cricket was used for each trial. On 10 tests (5
males, 5 females each) the rat could enter the cricket
-errarium and on 10 tests the tube was placed in the
cricket terrarium but the rat terrarium was not in the
room. Running order was counterbalanced. A time samp-
ling method was used such that behavior and location a

were recorded at 5 sec. intervals.

Measurement.-—Cricket location was recorded in
terms of position in the terrarium (wall, corner, or
center) and position relative to the tube (between the
tube and the wall, in the tube, on top of the tube, or
next to the tube). Cricket behavior was recorded with
the classification system presented in Table l with one
addition. Thigmotaxis was scored if the cricket was in

a corner, next to a wall, or behind or next to the tube.

Results.--Each experimental cricket was matched
with a same sex control cricket such that the number of
minutes observed (determined by how long it took for the
rat to enter the cricket terrarium) was the same for

each member of the pair. Thus the analysis focuses on

17

those preconsummatory behaviors emitted before
actually entered the cricket terrarium. These
could be in response to mechanical, visual, or
cues.

The time sample data were converted to

of the observation period (the absolute period

the rat
behaviors

chemical

percentages

duration

differed for each pair) in order to allow the use of the

Kruskal—Wallis one-way analysis of variance (Siegel,

1956) to compare the rat present condition with the

control condition.

The rat entered the cricket terrarium and cap—

tured the cricket in each of the rat trials.

The time

for the rat to enter the cricket terrarium ranged from

1 to 5 min. with a mean of 2.18 min.

The results of this experiment did not

permit

the evaluation of the first hypothesis concerning the

cessation of behaviors in other classes.

One male

cricket stridulated briefly before the rat entered the

passageway in the rat condition.
instances of behaviors from other classes.

The
of freezing
group. The
distant but
of the time

the control condition.

There were no other

second hypothesis predicted a greater amount
in the rat condition than in the control
crickets in the rat group responded to the
approaching rat by spending a mean of 76%
freezing as compared to a mean of 33% in

Application of the Kruskal-Wallis

18

test found that this difference was significant at the
.025 level using a one—tailed test.

The third hypothesis predicted a greater amount
of thigmotaxis in the rat condition. The crickets
responded to the approaching rat by spending a mean of
83% of the time in locations defined earlier as demon-
strating thigmotaxis as compared with a mean of 67% in
the control group. This difference was significant at

the .05 level using a one-tailed test.

Discussion

 

This experiment does not establish the actual
point at which the cricket detects the cues that signal
the approach of a predator. The results do Show that
preconsummatory antipredator behaviors occur considerably
before the predator actually reaches the immediate
vicinity of the cricket. Given that this response occurs
one may ask questions about the necessary stimuli for
these behaviors. Experiment 3 studied freezing behavior
and the roles of visual and mechanical stimuli in the
elicitation of freezing.

Experiment 3:

Freezing in Response to
Simulated Predation

 

 

 

Experiment 2 demonstrated that crickets are

capable of responding to the approach of an actual

l9

predator from a distance. This experiment marks a shift
from actual predation to more controlled administration
of test stimuli by the experimenter. The present experi—
ment was designed to obtain quantitative data on the
tepography of freezing. Some information was also
obtained about the roles of visual and mechanical stimuli

in the elicitation of freezing.
Method

Subjects.—-Fifty crickets were haphazardly
selected from a larger group in the colony terrarium.

The colony was maintained as described in Experiment 1.

Apparatus.--Twentyefive clear plastic cups (5.4

 

cm. diameter at the base, 7.3 cm. diameter at the top,
and 9.8 cm. high) were used. Each cup was used twice.
A small metal weight (1.2 g.) was attached to a string
and allowed to swing through an arc of 6 cm. to strike
against the side of the cup to provide a reasonably

constant amount of vibration.

Procedure.--One cricket was placed in each cup

 

and all crickets were allowed to adapt for 15-45 min.
The same procedure was used for all of the subjects.
The experimenter approached and stopped at a distance
of 2 m. from the cricket and noted the behavior of the

insect and the position of its antennae. The

20

experimenter then approached to within 10 cm. of the cup
and again noted the cricket's behavior and antennal
position. Then the impact stimulus as described above
was administered and again the insect's behavior and
antennal position were noted. If at this point the
cricket was in a freezing posture, the duration of this
behavior was timed with an arbitrary termination maximum

set at 150 sec.

Results.--The results may be divided into two
sections. First, the changes in behavior observed
during approach and impact, and second, the relationship
between antennal position and freezing.

Table 2 presents raw data for the frequency of
three behaviors at the three data collection points that
represent a gradually lessening distance between the
predator (experimenter) and the cricket. Eighty—eight
percent of the crickets were in a freeze posture after
impact. Freeze duration ranged from one sec. to the

limit of 150 sec. (n=3) with a median of seven sec.

Table 2.—-Behavior at Three Data Collection Points.

 

 

Distance Walk and Climb Groom Freeze
2 m. 22 6 22
10 cm. 16 6 28

Impact. 5 l 44

 

21

The cells for walking-climbing and grooming
were combined and a Cochran Q test (Siegel, 1956) for k
related samples was performed. There was a significant
overall increase in freezing as the distance to the
cricket decreased (p < .001). There was an increase in
freezing at the 10 cm. point over the 2 m. point, but
this difference was not found to be significant as a
McNamar test for the significance of change (Siegel, 1956)
yielded a Chi2 value of only 3.13. The same test found
a highly significant increase in freezing at impact as
compared to 10 cm. (p < .001).

The second part of the analysis focused on the
topography of freezing, eSpecially changes in antennal
position. Antennal position was recorded as horizontal,
crossed, raised, or moving. The antenna were most
likely to be moving if the insect was walking or climbing.
The antennae were moving in 84% of active crickets, 46%
of grooming crickets, and 3% Of freezing crickets.

The horizontal and crossed antennal data were
combined and a Cochran Q test was conducted on the data
for the 22 crickets that were in a freeze posture at
all three data collection points. These results show
that at 2 m. 23% of these crickets had their antennae in
a raised position. At the 10 cm. point, 26% had their
antennae in a raised position, and after impact 59%

had their antennae in the raised position. The overall

22

change was significant at the .05 level. There were 12
crickets in which a progression was made from freeze
(antennae down, crossed, or moving) to freeze with
antennae in the raised position over the 2 m., 10 cm.,
impact observation series. In all of these cases the
antennae remained raised after impact. The percentage
given for those insects that raised their antennae

after impact does not include seven individuals in which
the antennae were already raised at the time of impact '
but that responded by a brief startle response (body A
and antennal jerk) while keeping the antennae in the

raised position.

Discussion

The results of this experiment are compatible
with the results of Experiment 2 in that both demonstrate
that freezing is an early component of antipredator
behavior. The present experiment did not find any sub—
stantial evidence for a response to an approaching
potential predator over a distance. The results on the
topography of freezing are interesting because they pro-
vide some basis for distinguishing between resting beha-
vior and antipredator freezing. The position of the
antennae of a resting cricket is most likely to be
horizontal or moving. The antennae of a freezing

cricket are most likely to be in a raised position.

23

It is interesting that no hopping was elicited
during this experiment. This is consistent with the
observations made in Experiment 1. HOpping did not
occur in response to the cues associated with the
approaching rat (visual and mechanical) and in fact only
occurred when the rat was literally on top of and pre-
paring to capture a cricket.

Experiment 4:

Escape in Response to
Simulated Predation

 

One of the puzzling observations made in Experi-
ment 1 was that after the crickets moved into a corner
and froze they would allow the rat to approach to within
less than 1 cm. before hOpping. Typically, the crickets
were oriented such that they were approached from behind.
It is possible that the stimulus that elicits hopping
is some component (or combination of components) of the
ratS' breath. Pilot studies testing the response of
these crickets to moving objects and the results of
Experiment 3 on response to visual and mechanical stimuli,
found that this stimulation may elicit freezing but
does not elicit escape behavior such as hOpping or
running. However, breathing gently into a terrarium
containing crickets always elicits a substantial amount

.of hopping.

24

It seems unlikely that a system would evolve in
which air alone would elicit hopping because then any
stray puff of air would elicit components of the anti-
-predator sequence. (But crickets in their natural
habitat of basements and dumps would not be exposed to
very many stray puffs of air.) It also seems unlikely
that crickets would evolve a set of responses to speci-
fic predator odors. The most likley combination of
stimuli that would signal the approach of most predators
would be warm moist air with a high carbon dioxide
content. Other insects have evolved detection systems
for these substances. Gillet (1972) described such a
system in the mosquito.

Testing an insect's response to carbon dioxide
is complicated by the fact that this substance functions
as an anesthetic at high concentrations and may have a
variety of other behavioral and physiological effects.
Edwards and Patton (1965) and Patton and Edwards (1968)
studied what may be called the side effects of the
widespread use of carbon dioxide as an anesthetic. They
reported that anesthesia with carbon dioxide levels
above 40% had significant effects upon general physio-
logical variables such as growth and maturation, longe-

vity, and heartbeat in A. domesticus.

 

Klopfer and Quist (1955) studied carbon dioxide

avoidance in the mealworm, honeybee, and cockroach in

 

25

an air/carbon dioxide choice situation. This research
represented an attempt to demonstrate avoidance at
ecologically meaningful concentrations. The authors
attempted to explain species differences in avoidance
behavior in terms of the species' habitats. This work
was greatly weakened by a total lack of carbon dioxide
measurements in the habitats discussed. Klopfer and
Quist did find that of the three insects studied, the
cockroach was the most sensitive to carbon dioxide and
reliably avoided a 10% concentration but not concentra-
tions of 25%, 50%, or 100%. Cockroaches responded to
the 10% concentration with hyperactivity. There were no
instances of anesthesia below the 25% concentration level.
It seems unlikely that a brief exposure to a concentra—
tion below 10% would produce anesthesia in a cricket.

A pilot study was conducted to determine the
range of levels of carbon dioxide produced by potential
predators. A Beckman Medical Gas Analyzer (model LB-2)
was used to measure the carbon dioxide present at a
point 1 cm. away from the noses of three mammalian
predators--a rat, a gerbil, and a psychologist. Concen-
trations ranged from 1% to 4% as compared to an atmos-

pheric baseline of .04% by volume.

26
Method

Subjects.--Sixty male and sixty female adult

A. domesticus were used. Maintenance procedures were

 

as described in Experiment 1.

- Apparatus.--The air streams were obtained by

 

valving the laboratory air supply and carbon dioxide

from a tank through filters and flowmeters and combining
the gases in a Y tube. The system was set up to provide
air at 5 l/min., air + 4% CO2 at 5 l/min., or nothing
(tube presentation alone). A glass terrarium (25/50/30
cm. deep) was used as a test area. Teflon tubing (3 mm.
inside diameter) was used to minimize extraneous chemical
stimuli coming from the apparatus. Clear plastic cups

were used for temporary confinement and adaptation.

Procedure.--Crickets were gently captured in

 

test tubes and transferred to individual plastic cups.
Confinement lasted for a minimum of 15 min. and a
maximum of 30 min. A trial consisted of gently sliding
a cricket out of its plastic cup into the test area and
then presenting one of six stimulus conditions after a
l min. adaptation period. The conditions were chosen
at random and an assistant adjusted the gas apparatus
and the experiment was run blind relative to the gases

in the tube. Table 3 outlines the six conditions.

27

Table 3.--Procedure Outline for Experiment 4.

 

 

Group Condition Stimuli Presented
1 Control None
2 Tube Tube Alone
3 Air Tube + Air Stream
4 Air + CO2 Tube + Air + 4% C02
5 Freeze + Air Vibration + Tube + Air
6 Freeze + Air + CO2 Vibration + Tube + Air
+ 4% CO2

 

Behavior after stimulus presentation was
recorded in terms of six categories; freeze, dig, climb,
walk, run, and hop. In the control condition behavior
was recorded 64 sec. after release. In the rest of the
conditions the tube was presented from above and its end
was moved to within 1 cm. of the cricket's cerci and
then withdrawn. In the freeze conditions (groups 5 and
6) vibration produced by tapping the side of the terra-
rium was used to elicit freezing (no body movement,
antennae raised). Crickets were kept in individual cups
for 24 hr. after the experiment. Data were discarded

for any cricket that died during this period.

Results.--Table 4 presents the frequencies of
observed behaviors and the number of surviving crickets
in each of the conditions. Table 5 gives the percentage
distribution of normal behavior (freeze, walk, climb, and

dig), escape behavior (run and hop), and freezing

 

28

considered by itself. A series of Chi2 tests was con-
ducted within the data in Table 4 with the significance
criterion set at the .01 level. i

The control data provide a baseline of normal
behavior. It was not practical to distinguish between
freezing as antipredator behavior and simply not moving
as resting behavior. It is important to note that escape
behavior (running and hopping) was totally absent from
the baseline data. Presentation of the tube alone was
associated with significantly more freezing than in the
control group or in any other group. It seems reasonable
to assume that the increase is due to an increase in

antipredator freezing.

Table 4.--Escape Bahavior; Raw Data.

 

Behavior Control Tube Air Air + CO F + Air F + Air + CO2

 

n=l9 n=21 n=21 n=20 n=l4 n=20
Freeze 7 l7 7 6 3 3
Climb 4 0 2 0 l 1
Walk 6 3 6 5 4 2 I
Dig 2 l 0 0 0 0
Run 0 0 2 5 3 6
Hop 0 O 4 4 3 8

 

All of the Air and CO2 groups showed significantly
more escape behavior than the control group. There were
no significant increases in the individual CO2 conditions

compared to the corresponding Air conditions. If the

29

two CO2 groups are combined and compared with the com-
bined (Air groups, there was an increase in escape behavior
that was significant at the .05 level. Although there

was no substantial difference in amount of escape behavior
between the Air and Air + CO2 conditions, there was a
significant increase in escape behavior in each of the

Air and Air + CO2 conditions compared to the presentation

of the Tube alone.

Table 5.--Escape Behavior; Percentage Distributions.

 

Behavior Control Tube Air Air + CO F + Air F + Air + CO

 

 

2 2
% Normal 100 100 71 55 57 30
% Escape 0 0 29 45 43 70
% Freeze 37 81 33 ‘ 30 21 15
Discussion

 

The results of this experiment do not support the
hypothesis that carbon dioxide plays a substantial role

in antipredator behavior in A. domesticus. It is inter-

 

esting that the tube alone elicited freezing but the
tube and any airstream elicited escape behavior. This
is consistent with the results of Experiment 1 and the
opinion that cerci frequently carry sensory organs
sensitive to air currents. It is possible that the
crickets in Experiment 1 were responding to the rat's

breath.

30

The results of this experiment provide some
information about the relationship between stimuli
eliciting certain behaviors and the point in the beha-
vioral sequence at which the stimuli are presented.
However, the theory that a given eliciting stimulus will
be most effective only at its appropriate point in a
sequence is not supported here. Given that an air stream
elicits hopping and that hopping usually follows freezing,
this theory predicts an increased amount of escape
behavior elicited by an air stream in the vibration-
freeze groups (5 and 6) relative to the escape behavior
shown in the groups that had no vibration (3 and 4).

The data show a trend in the predicted direction. The
vibration-freeze groups showed 70% escape behavior as
compared with 45% escape behavior in the no vibration
groups but this difference was not Statistically signifi—
cant.

Given the trends in the data, it is possible
that larger sample sizes would allow finding a statis-
tically significant response to CO2 and some evidence
for a hierarchy of behaviors. Yet it is unlikely that
these effects play a major functional role in the anti-
predator behavior of this insect. It was for this
reason that the approach used in this experiment was
abandoned and the search for a more powerful preparation

for the study of order effects was continued.

31

Experiment 5:
Struggle and Captive Freeze
in Response to
Simulated Predation

 

 

 

 

The purpose of this experiment was to obtain
quantitative data on the behaviors of struggling and

captive freeze observed in Experiment 1.

Method

Fifty crickets (25 males and 25 females) were
used in this study. Each cricket was captured with
forceps from the colony terrarium. Each cricket was
grasped and held by one metathoracic leg in position
10 cm. above the bottom of the terrarium. The duration
and topography of struggle and the duration of the cap-
tive freeze were recorded and then the cricket was

released into a separate terrarium.

Results.-- Forty—one crickets responded to
capture with struggle. The mean duration was 4 sec.
(SD = 6 sec.). The behavior consisted of vigorous
limb movements and general exploration of the forceps
with the mouthparts. This behavior terminated abruptly
in 39 individuals that next went into a rigid captive
freeze (§ = 15 sec., SD = 14 sec.). After self-
termination of the captive freeze the crickets again

made limb movements and explored the forceps with their

32

mouthparts but these movements seemed distinctly less
vigorous than in the initial struggling response to
capture. Struggling duration ranged from 1 to 39
sec. with a median duration of 2 sec. The captive freeze
duration ranged from 2 to 81 sec. with a median of 10 sec.
There were several exceptions to the usual capture-
struggle-captive freeze order. Two individuals reSponded
to capture with slow convulsive movements. Two indivi-
duals responded with a modified freeze in which the body
remained still but the antennae and mouthparts moved, in
one case vigorously. Eight cases of autotomy were
observed, four immediately after capture, three after
2-5 sec. of struggle or jerking convulsions, and one after
1 sec. struggle followed by a 10 sec. captive freeze. All

of the breaks occurred suddenly and without avulsion.

Discussion

 

This experiment demonstrated that the normal
sequence (If antipredator behavior includes a brief
struggle after capture followed by a captive freeze
that may or may not be similar to immobility. The
results match the observations made of actual captures
of crickets by the rat in Experiment 1. The occurence
of autotomy in this experiment is interesting in that it
demonstrates that capture is at least occasionally a

stimulus that can elicit autotomy.

33

Experiment 6:
Immobility in Response to
Simulated Predation

 

 

 

Several cases of apparent immobility were observed
in Experiment 1 and the results of this experiment sug-
gested that chase, contact, and inversion might be
involved as eliciting stimuli for this behavior. Compara-
tive data (Ratner, 1977) suggests that capture, contact,
and inversion are stimuli that commonly elicit immobility
in a variety of invertebrates. The purpose of this
experiment was to investigate the eliciting stimuli for,

and topography of, this behavior.

Method

Subjects.—-Forty A. domesticus were used, (20 in

 

Treatment 1, and 20 in Treatment 2).

Apparatus.--A terrarium (50/24/30 cm. deep)

 

filled to a depth of 2 cm. with corncob bedding was

used as a test enclosure.

Procedure.--In Treatment 1, each cricket was

 

captured, held by the metathoracic legs, and positioned
ventral side up resting on the bedding at the bottom of
the test enclosure. This position was maintained for

15 sec. at which time the cricket was released.

 

34

In Treatment 2, each cricket was gently removed
from the colony terrarium and released into the test
enclosure. Chase was simulated by flicking movements
with four fingers of one hand such that intermittent
contact was made with the cricket and the cricket was
'occasionally flipped upside down. The number of flicks
required to elicit immobility, the duration of the chase,
and the elapsed time until the cricket righted itself
were recorded. A 3 sec. minimum duration was required

for nonmovement to be scored as immobility.

Results.-—Treatment 1 never elicited immobility
but it did elicit struggling in 17 out of 18 crickets
followed by captive freezing in 16 out of 18 crickets.
The median struggle duration was 2 sec. after the cricket
had been positioned on the substratum. All of the
crickets immediately righted themselves upon release
and walked away.

Treatment 2 elicited immobility in all of the 20
surviving crickets in this group. The mean chase time
to elicit immobility was 26 sec. (SD = 15 sec.). The
mean duration of immobility was 32 sec. (SD = 25 sec.).
There were no significant sex differences in either chase
time or in the duration of the response. The mean number
of flicks per second was 1.35 (SD = .21). The range

of immobility durations was 4 to 96 sec.

35

Discussion

 

The results of Treatment 1 match the results of
Experiments 1 and 5 in regard to the topographies and
durations of struggling and captive freeze. The failure
of this procedure to elicit immobility was as clear as
it was surprising as the stimuli presented are sufficient
to elicit immobility in several other invertebrates.

The results of Treatment 2 clearly indicated
that immobility could be reliably elicited by a com-
bination of chase, contact, and inversion. Immobility
was never elicited by chase and contact alone in this
experiment.

A possible interpretation of immobility is that
contact with a predator in the form of a sudden blow
causes temporary injury and "stuns" the cricket. However,
these insects can be dropped six feet onto a hard floor
and they immediately right themselves and move away.

It therefore seems unlikely that immobility is a result
of a sudden blow delivered by a rat.

This experiment also demonstrated that immobility
does not occur as a simple response to loss of tarsal
contact with the substrate. The flicking procedure
used here was an effective method to produce a loss of
tarsal contact. Yet this occurred an average of 35

times before immobility occurred.

36

It is apprOpriate to add some data on the range
of immobility durations from other experiments using the
same procedure as outlined here. A total of several
hundred immobility reactions have been elicited over a
course of 3 years. The durations of immobility have
ranged from the criterion of 3 sec. to 12.5 min. using
this procedure. No cricket was ever encountered in
which this procedure failed to elicit an immobility
response. The procedure was exceptionally reliable in
eliciting immobility but it was difficult to quantify and
control the eliciting stimuli. Another variable that
could not be controlled using this procedure was the
amount that a cricket in immobility was covered with
bedding. The flicking procedure resulted in varying
amounts of this material being deposited on a cricket
when it entered immobility. This provided a varying
amount of extra and constant contact stimulation. Some
evidence was obtained that the duration of immobility
was positively related to the amount of bedding on top
of the immobile cricket.

It seemed reasonable at this point that chase,
or the Opportunity to run off preconsummatory components
of antipredator behavior, might be crucially important

in eliciting immobility.

37

Experiment 7:
Sequence Position as a
Factor in the Effectiveness
offian eliciting Stimulus

 

 

 

 

Experiment 1 demonstrated that immobility is a
consummatory component of antipredator behavior in the
domestic cricket. Experiment 6 found that contact and
inversion will elicit immobility only if a period of
chase preceded the presentation of these stimuli in which
the cricket ran off the preconsummatory components of
freezing, running, and hopping. However, it was necessary
to design a more controlled method of presenting measurable
amounts of the stimuli that elicit these preconsummatory
components before experimentation could begin on their
role in ordering the antipredator sequence.

Experiment 4 found that a small amount of vibra-
tion elicits freezing. Pilot studies found that larger
amounts of vibration would elicit running and hOpping.
Pilot studies also demonstrated that immobility could
be elicited by placing a cricket in a horizontal test
tube and suddenly inverting the cricket by rotating the
tube. A tube with a diameter of 2.5 cm. was chosen
because larger tubes apparently did not provide suffi-
cient contact and immobility was never obtained in tubes
larger than 3 cm. in diameter. Immobility was easily

elicited by inversion in somewhat smaller tubes but in

38

these cases immobility frequently lasted for over 30 min.
and in some cases for several hr. Such long durations
were considered impractical for study.

The rationale of this experiment was to hold
contact constant while presenting the eliciting stimulus
for immobility (inversion) after varying amounts of
vibration elicited antipredator behavior. The major
hypothesis was that immobility is more likely to occur
and has a longer duration if the cricket has been stressed
or allowed to run off preconsummatory components of the
antipredator sequence before being exposed to the elicit-
ing stimulus for immobility. A more general statement
of the same hypothesis would be that the eliciting stimu-
lus for a given behavior is maximally effective only
when presented at its appropriate point in the sequence

of behavior.
Method

Subjects.--Four hundred A. domesticus adults
(200 males and 200 females) were used. They were main-

'tained under the same conditions as in Experiment 1.

Apparatus.-- Two hundred clear glass test tubes

 

(20 cm. long and 2.5 cm. in diameter) were used. A
vibrating platform driven by an electric motor with an

eccentrically weighted shaft was used to elicit

39

preconsummatory components of antipredator behavior.
The platform was adjusted to swing a test tube containing
a cricket through an arc of 1.5 cm. at a frequency of 7

cycles per sec.

Procedure.--Eight treatment conditions were used

 

but they all shared the following common general proce-
dure. Each cricket was captured in the laboratory
colony by gently chasing it into a test tube. The open-
ing of the tube was closed with a square of netting
secured by a rubber band. The tube was then placed
horizontally behind a screen on a table and the cricket
was allowed to adapt for 1 hr. At this point a tube
containing a cricket was very gently removed from the
adaptation table and placed on the vibration platform
that was set up behind a second screen on another table.
A l min. adaptation period was allowed before the pre-
sentation of one of 8 treatment conditions.

Immobility was defined as remaining ventral side
up with no movement for 3 sec. or more. If the first
inversion failed to produce immobility the stimulus was
repeated up to a total of 10 times and the number of
inversions and the duration of immobility recorded.

Each cricket was then placed in a separate plastic cup
for 24 hr. If a cricket died during this period the

data for that individual were discarded. Running order

40

was counterbalanced by sex and by treatment and 25 males
and 25 females were used in each treatment condition.
Each test tube was used twice. After the first use each
tube was scrubbed in tap water and stored for 10 days
before being used again. Half of the subjects in each
condition were run in used tubes.

Table 6 presents an outline of the treatment
conditions. Treatment 1 consisted of using one hand to
invert the cricket by lifting the Open end of the tube
through an arc of 180° above the table surface while the
other end of the tube remained at a point on that surface.
Treatment 2 consisted of using one hand to roll a tube
through 180° across a table top in order to invert the
cricket without lifting the tube from the table. Treat-
ment 3 consisted of placing the cricket tube on the
vibrating platform for 5 sec. and then inverting the
tube on that platform as in Treatment 2. Treatment 4
was identical to Treatment 3 except that each cricket..
received 15 sec. of vibration. Treatment 5 was identical
to Treatment 3 except that each cricket received 60 sec.
of vibration. Treatment 6 was identical to Treatment 5
except that a probe stimulus of 2 sec. of vibration was
presented to each immobile cricket after 5 sec. of
immobility. Treatment 7 was a control condition.
Neither vibration nor inversion were presented. Treat-

ment 8 was a control for possible damaging effects of

41

the vibration. In this condition each cricket received

60 sec. of vibration but no inversion.

Table 6.-—Procedure Outline for Experiment 7.

 

 

Treatment Abbreviation Vibration Inversion
1 F None Flip
2 R None Roll
3 SVR 5 sec. Roll
4 15VR 15 sec. Roll
5 60VR 60 sec. Roll
6 60VRP2 60 sec. Roll
2 sec. probe
C None None
60VC 60 sec. None

 

Results.--A routine check for sex differences in
immobility duration across all of the roll treatments
found that the mean female immobility duration (105 sec.)
was significantly longer than the male duration (x = 61
sec.) based on a Mann-Whitney U test (p < .003). This
was contrary to the results of Experiment 6 and it was
hypothesized that this difference was due to the larger
average size of the females which would result in more
contact with the test tube. An additional experiment
(8) was conducted using Treatment 1 which found, as will
be seen, that size accounted for the obtained differences

in immobility duration. The data for males and females

42

in this experiment were therefore combined for purposes
of further analysis.

Table 7 presents the major results of this
experiment including the percentage of individuals in
each treatment in which an immobility response was
obtained, the mean number of inversions required to
elicit immobility in these cases, and the mean log
durations of obtained immobility for each condition.
This table also gives the number of insects within each
condition that died during the first 24 hr. following
the experiment.

A test for homogeneity of variance was performed
on the log duration data with the result that these data
did not meet the assumptions required for the use of the
analysis of variance (Winer, 1971). Therefore, nonpara—
metric tests were used for the analysis of these results.

A Kruskal-Wallis test was performed on the log
duration data and these results were significant at the
.001 level. A series of Mann-Whitney U. tests found
that Treatments 2, 3, and 4 all elicited significantly
shorter durations than Treatment 5 (p < .03 in each
instance). Similarly, in the % immobility obtained
data, Treatments 2, 3, and 4 elicited significantly
fewer immobility responses than treatment 5 (p < .01 in

each case). Treatments 2 and 3 required significantly

43

more inversions to obtain immobility than did Treatment 5
(p < .02 in both cases).

In condition 1 (Flip), only 11 immobility reactions
were obtained (22%) as compared with 69% in the R011
condition. This difference was significant at the .001
level. In condition 6 in which a 2 sec. vibration probe
stimulus was presented after each cricket was in immo-
bility, only 2 out of 41 insects remained in immobility.
The rest responded to the stimulus by righting them-
selves.

There were no significant differences in mor-
tality across the treatment conditions even if the data
were combined to compare the mortality of crickets that
received vibration with the mortality of crickets that

received no vibration.

Discussion

 

These results provide some support for the
hypothesis that, in this insect and this class of beha-
vior, there is an order effect and that previously
emitted behaviors (or amount of "stress") combine with
proximal stimuli to determine the next response and the
topography of that response. When the condition in
which 60 sec. of vibration is presented is compared with
the three roll conditions with less vibration and less

preconsummatory behavior, we find that less stimulation

 

44

 

 

III IIII IIII IIII o o HOHuaOU >om m
III IIII IIII III! 0 o HOHDGOU m.
In: 1111 mh.m Hm.m am e Nmm>oo 0
mm. hh.a ma.~ oo.m mm a m>oo m
mm. om.H b>.m mm.m an m m>ma v
mm. om.a mn.~ vH.v on N m>m M
5m. wv.H No.m oa.v mm m Haom m
as. Hm.a mm.~ Hm.¢ mm o sham H
om M . om m.
sowumuso mom m:0fimum>sH # MDHHHQOEEH w apaamumoz pcmfiumoue

 

.w psmafiuomxm mo muHSmmmll.m magma

45

(number of inversions) is required to elicit immobility.
We also find that the greater amount of vibration results
in increased immobility duration.

The initial analysis of the results of this
experiment found a sex difference in duration. Experiment

8 investigated this finding.

Experiment 8:

Cricket Size and
Immobility Duration

 

 

The purpose of this experiment was to determine
if the sex difference in immobility duration found in
Experiment 7 could be accounted for by the larger size of
the female crickets. This larger size would result in
more contact with the test tube used in inversion

elicited immobility.
Method

Subjects.--Twenty male and 20 female crickets
were used. The crickets were selected for size such

that large males and small females were chosen.

Apparatus.--The apparatus used in Experiment 7

 

was used here. A pair of calipers was used to measure
the crickets and a Fisher Gram-atic balance was used to

weigh them.

46

Procedure.——The Roll procedure (Condition 2) used

 

in Experiment 7 was duplicated here. The length and

width of each subject was measured after the experiment.
The subjects were kept in individual cups and allowed to
die and desiccate for six months at 20°C. and a humidity

of 60% before they were weighed.

Results.--The crickets were measured and the
mean width and length for males was 6.4 mm. and 20.8 mm.
as compared to 6.7 mm. and 19.9 mm. for females. A
width X length score was calculated for each individual
and a Kruskal-Wallis one-way analysis of variance was
performed comparing scores of males and females. There
were no significant size differences due to sex in this
experiment. In the male group 18 immobility responses
were obtained with a mean log duration of 1.87 sec. The
female group had 16 immobility responses with a mean
log duration of 1.14 sec. This difference was not

significant.

Discussion

 

We may conclude that size can account for the
sex differences in immobility duration found in the
previous experiment. This is consistent with the idea
that some form of contact is necessary before inversion
can elicit immobility. These results are also consis-

tent with the results of the pilot studies that were

47

done to determine the size of the test tubes used in
these experiments. Smaller diameter tubes commonly
elicited durations of up to several hours making them
impractical for use. Immobility was never elicited in
tubes that exceeded 3 cm. in diameter, probably due to

a lack of contact while the crickets were being inverted.

Experiment 9:
Autotomy

 

During the course of these experiments approxi-
mately 30 cases of apparent autotomy were observed.
Several pilot studies were conducted under the hypothesis
that there would be a positive correlation between amount
of stress (vibration, etc.) and the probability of auto-
tomy. No such relationship was observed and these
experiments were uniformly uninformative. The haphazard
occurrence of autotomy and its unknown position in the
antipredator sequence indicated that some additional
research was necessary.

Experiments 1-8 outlined a sequence of anti-
predator behavior and provided some support for a model
of behavior based on an ordered chain of stimuli and
responses. Experiment 9 was conducted to determine the
probability of autotomy at each point in the known

order of antipredator behavior. The basic procedure was

48

based on the finding in Experiment 5 that sudden capture
at least occasionally elicits autotomy.

Two levels of autotomy were postulated in this
experiment. The first level, or actual autotomy, required
that the insect drOp a metathoracic leg without any
mechanical stress being placed on it beyond the weight
of the cricket. The second level, or potential autotomy,
involved placing a mechanical stress on the metathoracic
legs and required a clean separation to occur between
' the femur and coxa at the ring formed by the trochanter
with no torn muscle tissue visible. Pilot studies
indicated that the usual result of such stress was an
avulsion in which the rupture occurred between the coxa
and the metathorax with a substantial amount of torn
muscle tissue visible and an irregular break in the

sternum.
Method

Subjects.--Forty-eight male and 48 female A.

domesticus were used in this experiment.

 

Apparatus.--Twenty-four plastic cups were used

 

for individual cricket adaptation and temporary confine-
ment. One glass terrarium (25/50/30 cm. high) with a
3 cm. deep layer of ground corncob bedding was used as
a test enclosure. The gas apparatus used in Experiment

4 was also used here.

49

Procedure.--Table 8 outlines the procedure used.

 

The rationale was that the probability of autotomy might
vary with the position in the antipredator sequence
occupied by the individual cricket. Therefore each
group was taken further into the sequence than the pre-
ceding group. In Table 8, stimuli are capitalized
(Capture, C02) and responses are in small case (freeze,
immobility). For Group 1 a typical capture was accom-
plished by suddenly reaching into the colony terrarium
and grasping a cricket by one metathoracic leg. The
colony was subjected to minimal disturbance and at least
five min. elapsed between capture attempts.

The same procedure (Pull) was used to test for
"autotomy potential" in each group. One metathoracic
leg wasgrasped between the thumb and forefinger of each
hand with the cricket held upright with its dorsal side
towards the experimenter. The legs were then pulled
apart until a rupture occurred. For Group 1 this pro-
cedure was used within 4 sec. of capture.

Groups 2, 3, and 4 were tested for autotomy at
successively later points in the antipredator sequence.
In each group crickets were captured in the colony
terrarium by gently pushing them into large test tubes.
Crickets readily enter such tubes by themselves. They
were then transferred to individual plastic cups for a

10 min. adaptation period. A test began when a cricket

50

was tipped out of its cup onto the floor of the test
enclosure. Chasing consisted of repeatedly touching the
cricket with a finger for 10 sec. Immobility was elicited
in Groups 3 and 4 by chasing the cricket with one hand

and repeatedly flipping the cricket over onto its back

as in Experiment 6. Twelve males and 12 females were

used in each group. The groups were divided in two and

an A B C D D C B A design was used to control for effects

associated with running order.

Table 8.--Procedure Outline for Experiment 9.

 

Group . Procedure

 

Quick Capture, Pull.

2 Capture, Adaptation, Release, walk, Vibration,
freeze, C02+Air, run or hop, Pull.

3 Capture, Adaptation, Release, walk, Vibration,
freeze, C02+Air, run or hOp, Chase, hOp or run,
Inversion, immobility, Pull (after 5 sec. of
immobility).

4 Same as Group 3 but Pull after self termination
of immobility.

 

Results.--Table 9 shows the frequency of the
three types of leg separation for the four groups. Three
cricketsthat showed autotomy in Group 1 actually
escaped and had to be recaptured. The distinction
between autotomy and potential autotomy may be arti-
ficial. This experiment required a leg to drop off with

no mechanical stress from the experimenter to be counted

51

as autotomy. Several cases of potential autotomy were
observed in which both legs were being held prior to
"Pulling" and one leg came off without any tension being
applied. There was a clear distinction between poten-
tial autotomy and avulsion and there did not seem to be

varying degrees of damage caused in cases of avulsion.

Table 9.--Separation Frequency by Group and Type.

 

 

 

Group Autotomy Potential Autotomy Avulsion
l 12 ll 25
2 0 5 43
3 O 8 40
4 0 4 44
A Chi2 test was performed after combining the

cells for autotomy and potential autotomy. The results
were significant at the .001 level. The overall results
for potential autotomy alone were significant at the .05
level. Inspection of the data indicates that the primary
effect was obtained in Group 1. It is important to note
that actual autotomy never occurred after the insect

began the regular sequence of antipredator behavior.

Discussion

 

Contrary to expectations, autotomy seems to be
a very early component of antipredator behavior instead

of a late component. In this experiment actual autotomy

52

was only elicited if capture was very sudden. It also
seemed that the probability of autotomy or potential
autotomy was increased if the cricket was engaged in
behaviors characteristic of another class such as
feeding or grooming when captured. Having "run off" the
first preconsummatory components of the normal sequence
of antipredator behavior totally inhibited actual auto-
tomy. Some evidence was found for a lowered probability
of potential autotomy as the insect moved further into
the normal antipredator sequence.

The data for actual autotomy, and to a certain
extent, the data for potential autotomy, add additional
support to the hypothesis that a behavioral sequence or
hierarchy does exist for antipredator behavior in A,

domesticus.

 

Dethier, in To Know a Fly wrote:

 

Although small children have taboos against step-
ping on ants because such actions are said to
bring on rain, there has never seemed to be a
taboo against pulling off the legs or wings of
flies. Most children eventually outgrow this
behavior. Those that do not either come to a
bad end or become biologists." (p. 2)

In a certain sense, this research has discovered
a third alternative outcome for such children. They can
grow up to by psychologists and manipulate stimulation

such that the "fly" tears its own legs off.

CHAPTER III

SUMMARY

This study set out to discover the elements of
antipredator behavior found in the house cricket and
to obtain some information about how these elements are
ordered. The project began with basic observation of
semi-natural predator-prey interactions and moved to
controlled laboratory simulations to identify the
stimulus-response relationships within the antipredator
repertoire. Sufficient eliciting stimuli for each of
these behaviors were identified.

Figure 1 outlines the sequence of antipredator beha-

vior observed in Acheta domesticus during the course of

 

these experiments. The gradually closing distance fromtflua
predator to the prey is a reasonable and convenient
hypothetical variable around which to arrange these
behaviors. If a predator is detected while still a con-
siderable distance away, a cricket will move into a

crack or similar sheltered area. If the predator con-
tinues to approach, or if detection is considerably
delayed, the cricket will freeze, typically with both

antennae raised. Figure 2 shows crickets in a situation

53

54

similar to Experiment 1 that have moved into a corner
in apparent response to cues from an approaching rat.
Figure 3 shows a cricket in the typical raised antennae
freeze posture. Figure 4 illustrates continued freezing
in the presence of the rat. Figure 5 illustrates typical
behavior a fraction of a second after that illustrated
in Figure 4 in which a rat is shown beginning to open its
mouth. Suddenly, hopping occurs. Figure 5 shows two
crickets in mid air and one cricket already in immobility,
possibly after being rolled over and stepped on by the
rat. If a cricket is captured it usually struggles for
a few seconds and then all movements cease. Figure 6
shows a cricket in a captive freeze or immobility being
carried back through the tubing to the rat's nest and
Figure 7 illustrates a typical instance of immobility.

It is interesting that during the course of some
preliminary observations one rat carried 30 crickets
back to her nest and piled them in a corner. Many of
these were in immobility and righted themselves and
moved about the rat terrarium after the rat had left.
When the rat returned with another cricket all of the
"escaped" crickets were recaptured (eliciting immobility)
and returned to the hoard. This process was repeated
four times for some individuals. This suggests that

immobility may be an effective antipredator behavior.

55

In a natural setting the crickets could have easily
escaped from a rat's nest.

The above outline covers what may be called the
"normal" sequence of antipredator behavior. If, however,
the predator-prey distance is effectively zero when the
cricket detects the predator, an alternate sequence may
occur. Figure 8 shows the results of autotomy. The
metathoracic leg on the left was severed by autotomy.

The leg on the right, showing a severe avulsion, illus-
trates what happens if a leg is removed by force after
a cricket has begun its normal sequence of behavior. A
cricket may return to the normal sequence after it has
responded to a predator with autotomy.

An order effect or hierarchy was found to be
operating in antipredator behavior in this insect.
Inversion, the proximal eliciting stimulus for immo-
bility, was found to be maximally effective only after
the cridket had an opportunity to run off preconsummatory
components of the sequence. Here is a situation where
previous responses combine with incoming proximal stimuli
to determine the next response.

An even more interesting order effect was found
in the study of autotomy. Here, sudden capture, the
proximal eliciting stimulus for autotomy was found to be

effective only if the cricket had not had an opportunity

56

to run off preconsummatory components of the "normal"
sequence.

It seems that in this case, we have an example
of a branching hierarchy. At the point at which the
first stimuli associated with a predator are detected
the cricket either begins the "normal" sequence, or, if
the predator—prey distance is already 0, autotomy may
be elicited. The single element autotomy sequence is
apparently very unlikely to be entered once the normal
sequence has begun. However, after autotomy a cricket
may reenter the normal sequence.

It is unlikely that all, or even many, behavior
classes are organized in this fashion. But the pro-
cedures used in this study; the combination of a real
predator, laboratory simulation, and sequence analysis,
offer a methodology for studying the manner in which

behavior is organized.

 

Predator
Detection

Hide Autotomy

 

 

 

Freeze Walk/Hop

 

 

 

 

 

 

Captive
Freeze

(Immobility

Figure l.--Antipredator Behaviors.

(
«lipl
I

 

 

\JQQQKJ

58

 

Figure 2.——Thigmotaxis.

Figure 3.-—Freezing.

59

 

60

.uwm m mo moswmwum was :H msflmmmumll.w wusmflm

 

61

 

Figure 5.-—Escape Behavior.

62

.mumwum w>Hsme|I.o madman

 

63

 

——Immobility.

Figure 7.

64

.weouou5¢1|.m onsmflm

 

BIBLIOGRAPHY

65

BIBLIOGRAPHY

Alexander, R.D. The role of behavioral study in cricket
classification. Systematic Zoology, 1962, Al,
53-72.

 

Alexander, R. D. The evolution of mating behavior in
arthropods. In: Highnam, K. C. (Ed.), Insect
Reproduction, gymp, Roy. ent. Soc. Lond., 1964,
2, 78-94.

 

Baerends, G. P. The functional organization of beha-
vior. Animal Behaviour. 1976, A, 726-738.

 

Bate, J. On some aspects of the biology of Acheta
domesticus (Insecta, Orthoptera) with reference
to the infestation of refuse tips. Pedobiologia.
1969a, 2, 300-322.

Bate, J. Flight activity in a population of Acheta
domesticus (Insecta, Orthoptera, Gryllidae).
Pedobiologia. 1969b, A, 426-431.

 

 

Bate, J. Activity in adults of Acheta domesticus
(Insecta, Orthoptera, Gryllidae). Pedobiologia.
1970, A9, 45-51.

 

 

 

 

Bate, J. Life history of Acheta domesticus (Insecta,
OrthOptera, Gryllidae). Pedbbiologia. 1971, 1;,
159-172.

Brousse-Gaury, P. Contribution a L'etude de L'autotomie

chez Acheta domestica L. Bull. Biol. {3. Belg.
1958, 23,55-85:

 

 

 

 

Ikethier, V. G. To Know a Fly. San Francisco, Holden-
Day, 1962.

 

Dixon, A. F. G. The escape responses shown by certain
aphids to the presence of the coccinellid
Adalia decempunctata L. Trans. Roy. ent Soc.

 

66

 

 

 

 

 

67

Edwards, L. J., and Patton, R. L. Effects of carbon
dioxide anesthesia on the house cricket., Acheta
domesticus (Orthoptera: Gryllidae). Annals of
the Entomological Society of America. 1965, 58,
828-8321 -_

 

 

Ghouri, A. S. K., and McFarlane, J. E. Reproductive
isolation in the house cricket (Orthoptera:
Gryllidae). Psyche J. Entomol., 1957, AA,
30-36.

 

Gillett, J. D. The Mosquito. Garden City, N. Y.:
Doubleday, 1972.

 

Khalifa, A. Sexual behaviour in Gryllus domesticus L.
Behaviour, 1949, A, 264-274.

 

 

Klopfer, F. D., and Quist, J. A. Reactions of the meal—
worm, honeybee, and cockroach to some carbon
dioxide concentrations. Journal of Comparative
and Physiological Psychology. 1955, AA, 69472.

 

 

McFarland, D. J., (Ed.) Motivational Control Systems
Analysis. N.Y.: Academic Press, 1974.

 

 

Patton, R. L., Edwards, L. J., and Gilmore, S. K.
Delivering safe levels of C02 for insect
anesthesia. Annals of the Entomological Society
of America. 1968, EA, 1046-1047.

 

Ratner, S. C. Immobility of invertebrates: What can
we learn? Psychological Record. 1977, A, 1-13.

 

Sibley, R. M., and McCleery, R. H. The dominance
boundary method of determining motivational
state. Animal Behaviour, 1976, AA, 108-124.

 

Siegel, S. Nonparametric Statistics for the Behavioral
Sciences. N.Y.: McGraw-Hill, 1956.

Teale, E. W., (Ed.). The Insect World of J. Henri Fabre.
N. Y.: Dodd, Mead, and Co., 1949.

Tinbergen, N. The Study of Instinct. Oxford, England:
Clarendon Press, 1951.

 

Winer, B. J. Statistical Principles in Experimental
Design (2nd Ed.). N.Y.: McGraw-Hill, 1971.

 

    

IIIIIIIIIIIIIIIIIIIIIIIII

llllll1lllll2lllllllllllllllllllllslllzlllllllllsllllllllll