APPROACH LEAMENG AND
EXTINCTION AS A FUNCTION
OF AVOEDING PREDATORY ATTACKS

Thesis far ”to Dwm of Hi. D.
MECBEGAN STATE UNIVERSITY
Joseph Weldon Jennings, Jr.
1985

   
     

TH ES‘S

  

L I B R A R Y
Michigan Saw
University

APPROACH Lilo-1531.340 A33.) EXTIEUTIOS
AS A FURCTIOH OF A‘JOIDIHG
PREDAE’ORY ATTACKS

by Josoph Holden Jonnings, Jr.

Thin thooio prooonto o nodol tor opproaoh loarning
and oxtinction in pro: organiono ao o function 01 ropootod
prodotony ottooto, o nothod to: tooting tho letl. one data
thot boar on tho nodol.

Tho loorning,lodo1 involvod tho osounption that tho
haste toot oontronting tho proy otganisn to to loorn to
o)prooeh o rosion of tho ontironnont to obtain noodod rein-
!oroo-ont (orator) and than tloo this am rogion to avoid
tho diroot phyoieal attack of an approaching prodator. Tho
rooponoo of approaching the goal. it wao arguod. 1o condi—
tions! to tho otilnlt accompanying the onotional rooponnoo
of too: and trnotrotion.

root is olioitod by the operation of tho nrOGotor.
Attor a sufficient nunbor of opproacheo to tho goal rogion,
toorwprodunoé stimuli oorvo as partial aiacrininotivo
ottnnli olioiting tho opproooh responoo.

Frustration to olioitod by tho tormination of pooitivo
rotatoroonont noooooitotod by tho prey loaving tho souroo

a! roiltoroulont tn ordor to avoid tho predator's attack. An

Joseph Weldon Jennings, Jr.

with fear, after a sufficient number of approaches to the
goal region, frustrationuproduoed stimuli serve as partial
discriminative stimuli eliciting the approach roeponee.

Two hypotheses were derived. Both concerned reeietanco
to extinction of the approach reeponee. The first hypothesis
predicted that prey without prior experience in approoching
the reinforced region under threat of predatory attack would
chow lees resistance to extinction than had they had such
prior experience. The eecond hypothesis predicted that prey
allowed less time in the goal region (3 eeconde) prior
to the onset of predatory attack would chow more resistance
to extinction of the approach response compared to prey
having more time in tho goal region (11 seconds) prior
to predatory attack. Both hypotheeee were supported by the
reeults.

The apparatus ueed in the three experimente involved
o safe box in which the prey (female, hooded rat) was safe
from any attack. This cote box was connected with the goal
rosion by e 3 foot onelooed alloy. The goal region was
actually a three feet continuation of the alloy with water
available Just within the region. At the farthest end ct
tho goal region wee the predator. The predator was a 4 inch
aluninun dieo carrying 11,000 volt. .001 ampere charge on
its curtace. The electrical charge wee produced by an

automobile ignition coil. This mechanical predator was

Joseph Weldon Jennings, Jr.

prepelled the length of the goal region and alley in pursuit
of the prey. I: the prey allowed the mechanical predator
within 1/8 inch of itself it received a shock. The prey
would continue to be shocked until it moved faster than the
predator on the way towards the safe box.

Each subject (prey) was run individually once a day
for ten minutes. During a session. various measures of the
subject's behavior were taken. These measures included:
number of traversale between the safe box and the goal
region; the tics spent in the safe box. alley and goal
region during a session; the number of shocks received; and
slount or water consumed.

Suggestions were made for the extension of this work
to further laboratory analysis or prey~predator situations
and to field work in this area.

‘i? /7 7»
_ ’ /4' ", /(\
Approvefi i:,,- met“: ( ( gm“

dormittee Chairman /

Date & ; egg/51!. [75, ”
7 r_
I

 

a
I

0

V

x

\
-' \.
p U“ -mo..~r~;.-*M‘--~M‘.. ‘v‘ ’h“m.‘
' - 'I :0 0 ’v ‘ .“f '
. my; c" i - ’ 1,.1‘.

«hm muafiu

..~- E. NQZ‘ gm‘& .- w .

APPROACH LEARSIKG AED
EXTIBCTION AS A FUHCTIOH

OF AVOIDIEG PfiLDATOHY ATTACKS

By

Joaeph Weldon Jennings, Jr.

A THESIS
Submitted to
Michigan State University
in partial tuliillmen of the requirements
for the degree of
DOCTOR OF PHILOS §HY
Department of Psychology

1965

In more ways than I can recount, this thesis was
made posaiblo by a: wife's unatinting love. Carolyn

10 the author of my happiness and achievements.

ii

AthOuLhflGamfith

The author acknowledges the great assistance given
bun by his major professor. ML Ray Dannyy His wise counsel
and encouragement helped make this study pooaibla. I also
owe a debt of thanks to the other members of my guidance
committoes Doctors Stanley C. Ratnor and William Stellwagon
of the Department of Psychology and Doctor John A. King of
the Department of Zoology.

I also wioh to thank Mr. Royal Olson for his suggestions
and help in building the apparatua and Miss Jean Mcflartin

for her many patient hours spent in transcribing data.

111

TABLE OF COfiTEfiTS

ACKEOHLEDGEHSfiTS . . . . 0
LIST OF TABLiS . . . . . .
LIST OF FIGURES . . . . . .
IEQROD CTICE o I o o a o o

FTUStrfltiOno o o o a 3
Fear .0...

0....
Fear and Frustration Campos

Predator 0 o I o o o o

Reoiotauco to Extinction

General Considerations
Hypotheses o c o o o o

EXPERIHEKT I o o o o o o o

mathOd o o 6 u o o o 0
Results and Liaouooion

EXPELLIFELLQT II o a o o o a. o

Methadooooooao

0

Results and Discussion .

EXPERIEEgT III. 0 o o o I o
Hethofl Q a a o~o o o 0

Results and Biscuoaion .

CQEJCLUSIOIES A313 SUM‘iAh‘Y . .

“V“?DWH‘V”
hufiunénc&u . o o o n o o o

AP‘EEDICEfi o o o o o o o 0

1v

0

0..

3d.

O
o
O
O

Page
iii

vii

9
12
15
18
20

21

Q1
59

55

56
61

73

73
76

‘0'
K."

LC:
U1

Tohlca
I. Illumination, in foot candlco, of the
£1003 Of the a???‘rfltg8 7,. ,. o o o I ‘C O 0:
II. Boise lovelo, in decibolo,.recorded with
the Boise Level Meter microphone beside
the drinking tUbG . o o o o 0}. 0‘. 0
III. Kciee levolo, in decibels, recorded with
30156 Level Meter microPhone in the
miédle Of thfi safe box I o o o o o o o 0
IV. Humericel key to the various-componente
of the experimental apparatus shown in
Figure I. o o o o o o o o o o o o o o 0
V. Analysis of Variance of response rates
aorooe sessions of Phase 2 versus Phase
4 (Experiment I). . . . . . . . . . . .
VI. Mean water consumption (ml.) per session
of rhaoea 1 and 3 of Experiment I. . . .
VII. Frequency of shocks received per session
of the four phases of Experiment I. . .
VIII. Group composition in termo of the mean
reoponoo rate across Phase 1 (Experiment
II). C D O C O O O 0 O O U 0 0'. 0
II. Analysis of Variance of Group A veroue
Group B response rate across Phase 2
(EIperimont II). o o I o o o o o o o o o
X. Mean water consumption (ml.) by $3 of
Groups A and B during Phaeee I and 3 of
EXPeriment II. 0 O o o o o o o o o o o a
II. Frequency of shocks received by fig of
Groups A and B during the four Phases of
Experiment 11. o o o o o o o o o o o o 0
III. Group composition in terms of mean

mac or me To

roe once rate across Phase 1 (Experiment
111?.

O O O O I O O O O O O O O I I

33

39

54

54

58

61

71

72

75

LIST OF TABLES (cont.)

rail!
XIII.

XIV.

moan water ooneumption by g3 of Group
C and D during Plaoe l and 3 of
EXperlmOntIIIoboocoooooooo
Frequency of shocks received by.§g
of Groupe C and D during the four Phoeee
OffiPB-‘fi'mentlllootoot-ace...

vi

as

86

Figure
I.
II.

III.

1V3.

IVb.

V.

VI.

VIIa.

VIIb.

VIII.

IX.

I.

II.

LIST 0? FIGURES

Isometric drawing of the apparatus . . . . 32

leomotrio drawing of the mechanical
predator o o o o o 0

mean response rate per ooeeion of the
four pheeee of EXPeriment I. . o . . . . . 50

Mean total time spent in the safe box,
alley, and goal region per session of
Phaeea l and 2 of Experiment I. . . . . . 51

Mean total time opent in the safe box,
alley, and goal region per session of
Phases 3 and 4 of Experiment I. o . . . . 52

mean time in the goal region per response
per eeeeion of the four phases of
Ekperiment It .0 o a o o o o o o o o o o o 5)

Heen response rate per session of the
four phases of EXperiment II. . . . . . . 67

Mean total time spent in the safe box,
alley, and goal region per eeesion or
Phases 1 and 2 of Experiment II. . . . . 68

Mean total time spent in the safe box,
alley, and goal region per eeeeion of
Pheeea 3 and 4 of Experiment II. . . . . 69

Mean time in the goal region per response
per session of the four phases of
Experiment II. o o o o o o o o o o o o o 0 7O

Keen reoponee rate per oeeeion or the
four phases of Experiment III. . . . . . 82

mean total time spent in the safe box, alley,
and goal region per eeeeion of the four
phases of Experiment III. . . . . . . . . 83

Kean time in the goal region per responee
per eeeaion of the four pheeee of
EXPGIiment III. o o o o o o o o o o o o o 84

vii

r

This roooaroh deals with possible acquired ohongoo in
the behavior of prey organisms under repeated oxpoouro to
attack by a predator. A body or literature exists on the
lubjoot of prey reactions to predators but it has little
relationship to acquired changoo, particularly under con-
trolled conditions. There is a lack of data for oovoral
roaoonl.‘ On. in that most observations which have any
boating on the topic havo boon made in the field. that 19,
undo: rolativoly uncontrolled conditions (Seton. 1355;
Rudiger. 19503 Elton, 1953). A130, thooo studios have otton
involved opooioa-apooifio response pattopno (Simmons, 1935:
Molzach, 1961:.Bindo, 1961: Curti, 1935). Further, in many
of those studioo tho past history of the organism was
unknown and coulo not Q. controlled (Griffith, 1360} Richard-
son. 1942; Joalln, 1964): What 13 wanted 19 a situation in
which 1) tho onvironnontol and sequential oopooto of the
situation one known and controllable; 2) the oomo animal
can to obaorvod nepoatodly.

Since the ncoo £13k control is paramount, an apparotuo
was dovelupod that was assumed to be on anologuo of a
oituatlon in which a predator ooulfi confront a prey. The
pilot opporatua as first devoIOpod involved, as does the

inal apparatus, throo operationally dlotinct orooo. Theoo

I

were 1) the “safe box', analogous to the prey'e lair, which
consisted of a 5' x 12' high box. The prey (hooded rat)-

had eeey exit and entrance to this ”safe box' via a 3' hole
near the floor. 2) This hole led directly into e 6 foot long
alloy 4' wide and 12' high (the ”unsafe region"). 3) In

the middle of the alloy (3 ft. from either end) there was

a dish containing approximately one ounce of water (the goal
region). Thee the rat had to traverse three feet of alloy

to reach the goal region. .At the farthest point from the
'eefe box", there was a 3%” hollow steel sphere oerrying a
high tension (10,000 volts), very low amperage charge on its
surface. This sphere was suspended by on are from a movable
platform outside the alley which could be propelled the
length of the alley up to i inch of the "safe box“ door. This
sphere. its high tension charge, and ite movement were

eegggpg to be a nechenioel analogue of a predator.

This apparatus defines the task for the water deprived
rat, which was to learn to obtain water by moving from the
“safe box“ to the water dish and at the same time elude the
nechsnicel predator. After a brief delay following gig entry
into the goal. the mechanical predator begins to move down the
elley. I: the rat fails to return to the ”safe box” in sat:-
101¢at time and allows the mechanical predator to come l/B
inch it receives I punishing shook. It then continues to

be pursued and possibly shocked until it reenters the

'eete box”.

One might Justifisbly ask why the presentation of
apparatus details at this point. It is necessary for the
reader to appreciate these details in order to understand
the nature of the problem. The pilot work demonstrated that
rats learned to cape with the conditions imposed by the
apparatus as evidenced by an incremental increase in response
rate (trevereele between the “safe box“ and the goal) while
at the same thus swiftly achieving an active avoidance of the
mechanical predator at a probability in excess of 98%.

After an asymptotic response rate was achieved, the
primary reinforcement (water) was removed in order to observe
the rate of extinction. This was done for more than
curiosity's sake. There was reason to believe that the
animal might show greater resistance to extinction than a
group of animals which had equal access to the water but
had not learned to cope with the predator. Pilot work
suggested that this was a valid expectation under the

conditions imposed by the apparatus.

l;§g1;§£$gn. The role of frustration and frustration pro-
duced stimuli in resistance to extinction will be considered
first. under the conditions imposed by the apparatus, trus-

tretion is present during both acquisition and extinction.

During the acquisition the rat must leave the goal
to escape or avoid the mechanical predator. Leaving the
goal region results in the termination or withdrawal of rein-
forcement. feroter (1957. 1958) found that stimuli signaling
I"time out“ from positive reinforcement acquire aversive
properties. He also showed that the withdrawal of a positive
conditioned reinforcer had the functional prepertiee of a
negative reinforcer, 1.0., suppressing the rate of respond-
ing. Therefore. the rat is assumed to be frustrated each time
it must leave the goal at the approach of the mechanical
predator. '

On each trial, then. the rat will respond by being
frustrated. We might. therefore, expect that stimuli of
the goal region will become cues capable of eliciting frus-
tration response in the rat as it approaches the goal. Thus,
after repeated trials, the rat anydevelop a conditioned
anticipatory frustrations responses to the approaching
goal cues. But. through repeated approach responding. the
stimuli produced by the anticipatory frustration response
should become partial discriminative stimuli for the goal
approach response. In other words, at least late in acquisi-
tion training, the stimuli generated by anticipatory frustra-
tion will become cues eliciting continued approach to the goal.

Therefore. when primary reinforcement is discontinued

in extinction. rats which have learned to approach the goal

in reeponee to frustration produced ones will be acre
rueietent to extinction than to rate which have not been
frustrated in their ettenpte to obtain water.

Another he: to eay ell this ie no follcues Honreinforce-
lent of the approach response during.extinctiou elicite
frustration reeponee, but, it frustration-produced etimuli
here already become conditioned elicitore of the approach
response. then the frustration engendered by noureinforcement
is leee effective in eliciting reepoueee incompatible with
continued approach. In other words, because frustration
produced internal stimuli occur in both acquisition and
extinction. the two conditione are less diecriuinable to the
rat than had it not orpericnced frustration during acquisi-
tiou.

Such arguments as the preceding have been put forward
by imael (1958. 1962) and Spence (1360) to explain how
partial reinforcement producee result in an increased resis-
tance to extinction in instrumental reward situations. These
workers have pointed out that the intensity of frustration
elicited by noureinforcement should increase gradually during
acquisition ea anticipatory reward increases. Under such
conditions, g; are loco likely to show a passive avoidance
response to alley and goal cues during extinction compared to
§giwhich have not had such training. Therefore, the termin-

ation of reinforcement which results as the animal avoids

the lwchhnical predator is viewed as having the some effect
as partial reinforcement. That ie, the intensity of true-
traticn elicited by withdrawal of reinforcement should increase
gradually during acquieition ac anticipatory reward increases.
Assuming that chaining back also occura, then §§.will be
trained such that anticipatory frustration will elicit
reeponoeo which are compatible with the instrumental roe-
pence. Thus. the internal emotional response of frustration
elicited during extinction should result in greater reeie-
tauoe to extinction than would be seen in animale not
originally trained to approach while at the eaue time being

frultrated.

£325. The mechanical predator repreeente a source of punieh—
rent it not successfully eluded. Therefore, any persistent
reeponding on the part of the rat during extinction appears
to be ueaochietic, as its behavior would eeeu to court
disaster. handler (1964) hoe reviewed the problem of naeo~
ohien and notes an.uopubliohed observation by HowrerClQSO).
A rat was trained to run an alley to escape an electrified
alley; otter training, the check was aduinietered g£;;_in a
small area Just prior to the eecape box. the reeulte
indicated that keeping this email region electrified resulted
in increased reeietanoe to extinction even though the rat
could have paeeively avoided any further ehoek by not
running the alley. Evian (1949) found that rate trained

like Movror'e would run the alley taster and display greater

resistance to extinction than control rats that were not
shocked during extinction.

Brown, Martin and Morrow (1964) were able to accentuate
the effects reported by Howrer and Gwinn. In two experi-
ments, rats were trained to escape shocks in a starting box
and alloy by running down the alley into an uncharged gocl
box. During extinction, shock was no longer administered in
the start box, but some groups received shock in part or
all of the alley. Control rats were not shocked during
extinction. In the first eXperiment attempted, the control
and experimental groups performed alike, that is the experi-
mental animals were not more resistent to extinction for
being shocked during extinction. In the second experiment
they changed the procedure so that the magnitude of extinction-
shock and the number of escape trials were reduced. This
was done to make the transition from acquisition to extinction
less discriminsble. Under these conditions the animals
shocked during extinction took longer to extinguish than non-
shocked control animals.

Azrin and Hols (Asrin, 1959. 1960; Azrin and H012,

19613 Hols and Asrin, 1961) did a series of studies which

have even greater significance for the present research.

Their procedure was such that a positively reinforced response
engendered punishment. The method involved the pigeon's

upecking response in a key peck Skinner box. fiftcr the birds

reaponéiug was shaped and food reinforced on various cohe-
dulea, shock of varying intensities and durations was
administered as a second and acditional contingency to key
pecking. The data from Azrin (1360) indicates that pigeons
will deliver checks to themselves several hundred times to
receive intermittent food rewards. Although several varia—
tions combine to produce this behavior, the results can be
interprcted ac inflicating that shook under these conciticns
does little to interfere with typical intermittently re;n-
forced rcaponéing.

H012 and Azrin (1961) report stronger evidence that
pigeons may increase their respcnce rates as a function of
punishment alone. The subjectc were run under two daily
conditions. In one, the response contingency was a VI food
reinforcement and CR! chock. The second involved extinction
without the reintcrcement~chock contingency.

Assume that punishment or cues accocicted with punish—
ment elicit an internal rccyonse. The label for this
internal reagonse will be 'fcar'. The internal response of
fear is assumed to be involved in increased resistance to
extinction of the approach recponce by the prey organism for
the some reasons as were true for frustration. Namely.

1.) During acquisition, roar produced ctimuli become ccn~
diticncd clicitorc for the approach response.

2.) Fear exists in both the acquisition and extinction

conditions, as do its attendent stimuli.

3.) Because of the presence of fear produced stimuli in
both acquisition and extinction conditions, the two
conditions are less discriminable for the prey animal
than for an animal which has not eXperienced fear during

acquisition.

gear and Frustration Cemented.~ A study by Brown and Wagner
(1964) provides support for the functional similarity of
”fear“ and frustration. Three groups of rats were trained
in‘a simple runuay. During acquisition. Group l was exposed
to nonreintorcement on a 50% reward schedule. Group 2 was
exposed to gradually increasing punishment along with con-
eistent food reward. Group 3 was never punished or non-
reintorced. Half of each group was then tested for the
decremental effecte of either consistent nonreiniorcement or
consistent punishment. Group 1 and 2 fig were more resistant
than Group 3 fig not only to the decremental variable which
they had been trained on, i.e., punishment or nonreinforcement,
but also to the alternate toot variable. These results were
interpreted in support of a commonality between the emotional
consequences of punishment and nonreinfcroement. The results
showed that responding for food and shock was greater than
for neither. More importantly, the behavior was maintained
even when the food reinforcement wee withdrawn from the

first condition. In fact, delivering shock alone during the

10

second or extinction condition resulted in an increased rate
of responding.

To quote Hole and Asrin (1961), ”These experiments
demonstrate that a relatively severe punishment can increase
responding...Thie procedure o1 selectively pairing a
stimulus with a reinforcer is the usual procedure for estab-
lishing a discrimination. This discriminative property that
the punishment acquired produced the apparent anomaly.
Indeed. the discriminative property came to exert an even
greater effect on responding than did the aversive property."
(p. 231)

a quote tron.3rown and eagner (1964) serves to clarify
the cannon effects of "fear“ and frustration on resistance
to extinction. ”If there is more than a conceptual simil-
arity between the emotional responses of fear and antici—
patory frustration, it would be reasonable to expect some
degree of transfer of behaviors learned in the presence of
one to occasions when the other ie aroused. Thus it might
be expected that §g_whioh have learned to approach in the
presence at anticipatory frustration would also persist in
approaching in the presence of rear. Likewise §§_trained
to approach in the presence or fear might be expected to
continue to approach in the presence of anticipatory trus-
tration. In this context, the present findings of a partial

tranator between the learned resistances to punishment and

11

extinction would argue for a degree of commonality between
the two emotional responses." (p. 507)

The reasons for a I'partial transfer” rather than a
complete transfer observed by Brown and Wagner can be
clarified by Carlsmith‘s (1961, reviewed by Church 1963)
study. Carlsmith found that the mean number of trials to
a criterion of extinction was uninfluenced by the conditions
of punishment (shock or loud sound) but there was a large and
significant interaction. If the same aversive stimulus was
used as a punishment that was used as s U03 for avoidance
training, resistance to extinction was much greater than if
the other aversive stimulus was used as punishment. This
was interpreted as supporting a discrimination hypothesis.
Thus facilitation may occur in cases of punishmnnt of negative
instrumental acts because of a reinstatement of specific
stimuli present earlier in training.

Therefore. we may View the effects of the type of
situation proposed for study as involving the simultaneous
conditioning of the internal responses of fear and frustra~
tion. Both internal responses appear to work in concert to
facilitate resistance to extinction since the animal has
been trained such that the anticipatory responses of fear
and frustration elicit responses compatahle to continued
approach to the goal. Also, since the anticipatory responses

should be developing a pace under the some external stimulus

12

conditions, anticipatory responses of fear and frustration
should be elicited by the some external stimuli.

Thus the situation confronting the rat or other enihel
can be viewed as involving fear and fruotration as internal
responeee. After sufficient eXperience stimuli concomitant
with these responses come to elicit responeee compatable with
approach to the goal region. Therefore. resistance to
extinction, should be greater in animal. which have had the
experience at learning to cope with the repeated atteoke of
a predator compared to those which have not had to learn to

cope with a predator.

Predator. To say that the prey muet learn to ccpe with

the predator, is to say that the prey learns to deal effect-
ively with the actions of the predator. The predator'l
action towards the prey ie labelled "threat".

The predator can be conceived of as representing two
types of threat to the prey. One is potential, the other
actual. ihet is, the animal can behave differentially
depending upon behavior of the predator. The animal must move
towards the mechanical predator ot obtain reinforcements. It
is therefore exposing itself to potential threat. The
potential aspect refers to the fact that the mechanical
preaator is not waiting beoide the water source to spring
instantly upon the rat. It io at a distance from the rat

and begins to approach the rat only after the rat hue arrived

13

at the water source. Once the mechanical predator is in
motion the threat is termed “actual" rathsr_than potential.

The above View of the dual mode of predator action
finds support in Keehn's (1959) work. Keehn's study involved
an avoidance situation (rats were held in an electrifiablc
activity wheel) where interval responses served to postpone
the onset of the next trial. Animals were trained with and
without a warning signal. Those animals which received the
warning signal were free to postpone the onset of the signal
as well as the shock. What Keehn found was that the animals
supplied with a warning signal behaved so as to postpone the
shock but not the warning signal. He argued that the signal
was not a secondary negative clicitor. Rather, the signal
had tho properties of a discriminative stimulus because it
marked the time in which the appropriate avoidance response
would be reinforced.

Keehn's work resembles Sidman and Baron's (1957) although
Sidnan and Baron's was not as well controlled. They gave
their animals considerable avoidance training before the
signal was introduced. Sidman and Boron interpreted their
results such as Keshn did by suggesting that the discriminated
avoidance situation may be considered a multiple schedule in
which one avoidance contingency prevails in the presence of
the warning stimulus and another in its absence. Thus in
the situation where a specific stimulus precedes a noxious

event. such as the sights and sounds of the mechanical

 

l4

predator as it ceases to be a potential threat and becomes
an actual threat to the rat. this stimulus comes to die«
tinguieh between occasions when avoidance reeponaee will be
reinforced and when they will not. For the analogous prey-
prodator situation, an avoidance response prior to the
onset of actual threat should be punished by too early a
termination of reinforcement. That is, the animal should
learn to etay at the source of primary reinforcement at
least until the warning signal of an approaching predator
has been perceived.

Additionally. a study by Kelvin and Brown (1964) may
polaibly indicate that even the onset or the predator's
approach might not be sufficiently aversive to immediately
Ilicit the avoidance response. In this study a noxious
bright light preceded food delivery to rate, for 20, 40, or
80 pairings. After this training, the light was then used
alone in an escape learning test. It was found that light-
food pairings diminished the light'e aversivenese, the
effect increasing with frequency of pairings. Loss of
aversivonees was attributed to the light'a having acquired
tendencies to elicit food seeking which competed with eecepe
responses. is an example of a competing response incompet-
iblo with escape responses, Melvin and Brown site a loco-
motor movement toward the former location of the food cup
at the back or the apparatus; in other words movement in

a direction opposite that necessary to escape.

15

figgietsncetg_§;tinotion. By invoking a competing response
explanation of resistance to extinction, it is possible

to understand why the organism may persist at the source of
the primary positive elicitor at least up to and possibly
slightly beyond the onset at the actual threat from the
approaching predator. Therefore,it seems reasonable to
suppose that the effects of frustration and tear compete
with the develOpment of a passive avoidance response to the
one: of the alley and goal regions.

The terms, “frustration" and “tear" are considered
internal responses elicited in the organien by external
stimuli that may be labeled negative elicitore, i.e..
potential and actual threat and the termination or a hrinary
positive elicitor. The effect of the frustration and tear
an overt behavior may be comparable. as flelvin and Brown
(1964) have indicated. Therefore, in attest. fear and
frustration may be olassed together as internal response
normally antagonistic to the internal response of relaxation
es hypothesised by Denny and associates (Denny and Adelssn
(1956): Denny one Heisman (1964) ). This View at “frustration-
fear“ is in accord with Amsel’o (1962) proposition that
extinction is an sotiVe not a passive process. It is on
active process neéiated via disorisinoble internal stimuli.

Also, it is maintained that the internal response

produced stiluli, generated within an animal by external

16

stimuli which have fcorful and frustrating associations
con to discriuinitivc stimuli eliciting continued approach
towards tho situation associated with potential threat of
punishment and withdrawal of rain: rccmcnt. Therefore, in
the prey—predator analogue, the n nreinforccuont of approach
responses in the face of threatening cuss should result in
greater resistance to extinction in animals which have
previously learned to approach threatening stimuli than in
animals which were previously able to obtain primary rein-
forcement without tearful or threatening consequences.
Further, on internal response of fcsr~£ruotration
is viewed as a continuous variable. That is to say, the
fear—frustration response may vary in magnituoc. While a
certain magnitude or Icor~frustrsticn may become a conditioned
segment of the approach response, an increase in tho fear-
fruotrstion response over and above the conditioned level
should provide responses which compete with continued approach.
Bronco. the increased magnitude of the competing recr-
frustration rooponco over~aud~sbove the previously conditioned
level should require some time to become ccntigiouoly
associated with the external stimulus cooplox, especially
the stimuli of tho goal rcgion. That is, extinction takes
some time to occur as it torso time for tho competing roo-
ponsoa to build up. Possible support for this latter

suggestion is found in a study by Lonny (1959). In the first

17

of two experiments, two groups of rats were trained to press
a her one trial a day for 10 or 50 trials. Both groups were
then given 75 extinction trials, one per day. A control
group, with no prior training, was given 25 unrcwsrded trials,
also one per day. The bar was always removed from the box

as soon as § had depressed it.

During extinction, the latency of the bar-pressing
response greatly increased for the control animals, but
reached a low, stable, level in the two groups which had
previously received either 10 or SO reinforced trials. There
was no evidence of extinction in these two latter groups.

In the second experiment, rats were trsinod to prose
a her five trials 3 day until oach,§;g_latency was at least
cqual to the group in the first experiment which had 10 rein-
forced trials. The: were then given extinction trials until
a Salinute no-rolponco criterion was attained. But, the
III in this case was 5 minutes not one day. Under these
conditions, all a; extinguished in less then so trials.

Denny interpreted the results from this study as follows:
1.) bar—retraction immediately after discrete bar-proooing
docs not itself prevent extinction; and 2.) one trial a day,
or at least highly spaced trials, is essential to the
Virtual prevention of extinction. "The importance of bar—
roctriction use definitely suggested, however, since E
observed that approach to the ever present food trey extin-

guished during the non-rewarded trials even her responding

18

did not.” ”One compelling observation by §.in Experiment
11 was that on the trial or two just before § extinguished,
all §;§_hegen to make vigorous attempts to escape from the
box. An implication here is that frustration effects may
accumulate with a 5-minute intertrial interval and become
sufficiently strong to instigate competing responses.” (p.85)
I The implications of Denny's study for the present study
lies in the similarity in conditions. Removal of the bar,
in the Denny study. in viewed as similar to the animal's
withdrawal from the goal region compelled by the "actual
threat“ of the approaching mechanical predator. During
extinction, the conditioning of competing frustration
responses to the goal stimuli £111 be impeded if the
animal must quickly avoid the oncoming mechanical predator.
0n the other hand, if the prey organism is permitted to
spend increased tine in the goal region the cues of the
goal region would become elicitors o: the frustration
response and thereby hasten extinction. I

general aneigegationg, To summarize the intent of
this study is to impose experimental control on an assumed
prey-predator situation where the temporal and environmental
aspects of the situation are known. or equal importance
is the fact that the behavior of the prey animal can be
repeatedly observed and recorded. Through manipulation of
factors controlling reeietence to extinction, an attempt

will be made to underetand how the prey organism learns to

- .‘-v

 

19

to cepe with a situation involving potential and actual
threat of predation.

The experimental apparatus regardleoe of claims to an
analogy to the naturally occur*in5 eituetion, could actually
be considered a conglomerate of familiar luberotory
apparatus. Indeed, the experimental apparatus involveo
nopecto very similar to those found in a shuttle box, a
straight alley, and perhaps an obstruction box.

Con idered in another light, it should be remembered
that many animals have leirs, burrows, or noete. They must
forage to stay alive for rarely are the needs of the organ~
ion met without leaving the neet. By foraging, the animal
expoeee itself to predatore. Alec, in the competition with
other members of its own species, they (the prey organiem) muet
take maximal advantage of the limited sources of ouetonn“ce
available to it (Wynne-Edwards, 1963; Calhoun, 1362). The
complexity of the naturally occurring situation dictotce the
use of a laboratory situation paralleling the natural
in ito essentials.

Because of the temporal and apatinl eecuoncee of events,
the prey-predator situation can also be characterieed on
involving ”approach-then-uvoid behavior.” The study of
avoidance behavior, from the outeet, has been viewed by
some workere as having a direct bearing on the prey-predator
oitueticn. A3 Pavlov (1927) has noted, "The strong carni-
vorous animal preys on weaker animals, and if they waited

to defend themselves until the teeth of the foe were in

:p
i)

their flesh, would speedily be exterminatod. The oooe
takes on a different aspect when the defense roflcx is
called into play by the sights era sounds of the enemies
approach. Khan the prey hoe a chance to cave itself by
hiding or by flight' (p.14). Hull (1934) also referred
directly to this tapic when he wrote, ”In the violent
otruggle for existence pictured by organic evolution....
those animals which responded by flight and other defense
reactions in advance of injury would be far more likely
to ooonpe who horco woulé have immensely greater choncee
of survival and ultimate reproduction than would arimala

which did not possess such a tendency.” (p.454).

error-.1 13:12:;
From the foregoing considerations, two hypothesee
more formulated even though tlifi otuéy io Quito explanatory
in format.

1. In the prey-predator analogue, the prey
which had previously learned to approach
threatening stimuli in order to obtain positive
reinforcement shows greater reeiotanco to
extinction than an animal which had obtained the
some positive reinforcement without fearful

or threatening consequences and/or without
being forced to leave the goal region preme-
turcly (because of threat).

2. In he prey-predator analogue, the prey
which hfid previously learned to approach
threatening stimuli in order to obtain posi-
tive reinforcement shows greater resistance
to extinction if forced to leave the goal
region sooner (because of threat) then a prey
which is not force& to leave the goal regioo
a! 90011.

 

Suhigotq. The gg'wero S naive female hooaod rats from

 

the colony maintained by the Papartment of Psychology of
Michigan State University. All 23 were 110 days old at the
start or the experiment.

Beginning two weeks prior to the start of the experi-
ment, each §Iwas handloa for five minuteo a day. Beginning
one week prior to the start or the experiment, §3.were
placed on a water deprivation oohedulc of ten minutes of
access to water in the individual homo cages every 24 hours.
Food waa constantly available.

fipogggtgg, An isometric drawing of the apparatus is
presented in Figure I, togother with labels for the major
apparatus components. The major component of the apparatna
consisted of a long, rectangular box measuring 8 feet long
by 5 inches wide and 2 feet high. This box wao closed at
both endo; at one end by a fixed end wall, at the other
by a hinged door which swung outward and downward from the
top. All four sides of this box were corotructed of 1/8
inch Masonite with the finished surface facing inward.

Thin flasonite box was firmly mounted to the top and
to one side of a very rigid box platform. This platform
measured 8 feet long by 1 foot wiée by 5% inches high and
was constructed by 3/4 inch plywood for the toy (which also

21

P)
l“ '2

served as the bottom of the hasonitc box) and 5/4 inch
finished pine planking on four aides.

Two rectangular frames were rigidly attached to the
platform. One frame was attached to the platform 22-3/4
inches from the end with the hinged door. The frame con»
sisted of two 5/4 inch by 5 3/4 inch pine uprighte, one on
either side of the platform, rising 29% inches. The tope
of these uprighte were connected by a 13 5/8 inch, 3/4 by
5 3/4 inch piece of pine. it the OppOSite end of the plat-
form was the other an& identically constructed frame
although it was not inset from its end of the platform.

The Opening formed by this frame and the end of the platforh
was covered by a aheet of % inch plywood. This made the

frame especially rigid and formed a vertical mounting

surface for an electric motor. The motor was a 115 volt D.C.,
.60 ampere, 1725 rpm, 1/20 h.p., ESE-33, reversable Bodine.

It was mounted so that its shaft was 20% inches from the

top of ita plywood base with its shaft lying horizontally.

A 5/8 inch pally wee attached to tho motor's shaft. The
current for the motor was supplied by a varies to control
motor epoch and a full-wave bridge rectifier.

The varioue euperstructures of the apparatus were
attached to the horizontal, overhead members of the two
upright frames. This included a pulley and belt system,

microswitcaes, the track for the mechanical predator and

r~ ‘9
\21

a mirror. The overhead par of the mechanical prceator
depenoed from a 6 foot track consisting of a sliding

.5.

closet door track of extruded alum‘num (Soars Catalogue Io.

Under this arrangement, two feet of the Hasonito
rectangle was not between the two upright frames. Thio
aegment of the larger box was the "safe box” and was phy-
sically separated from the remainder by a 22 inch high
partition with a 2% inch diameter hole with the bottom
edge of the hole 2 inches from the bottom of the partition.
This hole communicated to he remaining 6 foot segment of
the box. Cn the other end of the safe box was the previously
mentioned hingefi door.

The mechanical predator was drawn to and fro along the
overheea track via a long loop of high tensile strength
cotton string attached by screw eyes to either end of the
overhead component of the mechanical predator. This 100p
of string ran immediately under the track and was guided
by pulleys around the ends of the horizontal segments of
the uprights and over the tap of the trace. At the end
of the rue on which the elecuric motor was mounted, the
string loop was driven by a e inch diameter pulley. This
pulley wet connected by a short shaft to a 1 1/8 inch pulley
which wee ériren through a 21 inch rubber belt from the
pulley moulted on the motor shaft.

The movement of the mechanical predator up and down

24

the 6 feet of the alley was controlled by two microswitches
(Micro V4—l4). Each switch was suspended over the inner
edge of the upright frames by angel brackets. They were
actuated by rotational force, so from the shaft of each
switch a finger of flexible piano wire hung down into the
path of the overhead component of the mechanical predator.
The resting position of the mechanical predator was at
the end of the alley opposite the safe box. To place it
in operation, fihmomentarily depressed a hand held switch
which looked closed a relay supplying current to the
electric motor. On reaching the safe box and of the track
the mechanical predator deflected the finger actuating the
microswitch at that end. This action unlocked the first
relay and locked on a second which fed a reversed current
to the armature or the electric motor. The mechanica
predator, therefore. reversed direction and returned towards
its start ng position. Just before reaching this end of
its track, the mechanical predator deflected the finger
from the other microswitch which unlocked the second relay
and opened the circuit to the electric motor, thus stopping
the mechanical predator until its cycle was begun again by
E.

The mechanical predator consisted of two connected
components, an overhead apparatus and another at floor

level. The latter component was what the gs had direct

interaction with. A drawing of the complete unit is to
be had in Figure II. The overhead component consisted of
the four, wheeled track runners supplied with the closet
door track. Two runners were hung in the two parallel
J-troughe of the track. The two runners in each track
were separated on 5 inch centers and ran in Oppoeeed pairs
front and back. Mounted between the track runners with
its bottom surface flush with the bottom edges of the
runners was a 7 inch long, 1 7/16 inch wide and 1 5/8
inch high block of pine. mounted flush to the bottom of
the block was a piece of i” Masonite measuring 7 inches .
long by 3 7/8 inches wide in the horizontal plane. Sne-
ponded below this horizontal Masonite platform by encir-
cling Haeonito mounts was a 6 volt, Allstate ignition coil
(Sears catalogue no. 28A8240). The 0011 was mounted so
that its long axis was aprallel to the floor of the Illey,
its electrical terminals projected towards the safe box
and of the alley, and ite bottom was flush with the back
edge of the Mhsonito platform from which it hung.
Projecting downward from the horizontal Masonite
platform.wee a 17 inch, 1/8 inch diameter bronze rod. The
rod hung from a point 3/3 inch from the front edge of the
Masonite platform and 2 3/8 inches from either side wall
of the alley. This placed the rod directly in front of
the high tension electrode of the ignition coil and was

connected to this electrode by a coil spring assuring good

.1,
I.

ootrioal condutlon and a fat iguo Ir eo linkage Thin bronze
rod thus cozduo od high voltage currc 2L dovrnard to the

lower component of the mechanical predator.

)

The lower co W; est consistei o; a 4 inch, 12 gauge
alminmn dioc 1.131121;- sprayed with no t Lute p.112; 1: for
inoroo: cd also ri; lnability by Q, Tho dioo Lao bolted to an
additional threaded iILCh of the bronze rod bent at a right

m'le to tho vertical segment and pointing fo wards towards
the safe box and of the alley. The mounting hole Of the
disc was in its exact cantor. Thus the flat ourface of
too disc was mrpondicular to the fl oor anfi facing tho sofa
box end of the alloy..

Tho 6 volt current was supplied to the ignition coil
from an Allstate battery charger (Home l 30. 608.L600) first
being fed through a Motorola automobile radio vibrator
(Type 485522000). The current was fed to the overhead
compor ent of tllo mechanical predator from a fine gouge,
inoulatoo d, twin lead hanging from the top of a 2 inch wide,

inch tlli 0k, 43 inch high poot fastened 63 inches to center
from the 333 ebox end of the platform. The twin lead wire
was 40 inohoo in longth and was c‘nnootcd to the mechanical
predator 1y 0 vertical otand o; f with an arm projecting
over the oide of the alley. This 338 ”tn aallowod positive

aleotric supply to the coil without twisting or stretching

27

the wire or allowing it to foul the mechanical operation
of the predator.

This system was capable of delivering a 11,600 volt,
.001 ampere shock to the gg’if they ailowed the disc witnia
1/8 inch of any pirt of their bodies as valtuge at this
pressure was quite capable of arcing through the fig hair.
This shock gunisnad §,for failure to avoid in aurficiont
time. The shock was assumed to be roughly analogoua to
the effects of tho teeth and claws of a predator on the
EL; body without the necessity of actually producing
lasiona of the skin.

The floor or the apparatus on which §.actually saved
consistea of § inch hardware cloth. In the 6 foot alley
this floor was formed by benéing down the edges of a 6 foot
run of-hardwara cloth 90 that the surfacg was raised 1 1/3
inches above the wooden bottom of tha alley. This provided
a self-cleaning surface besides being an electrical ground
When g received a shock.

The floor of the safe box was also formefi o: i inch
hardware cloth with an inch of additional material folded
downward and slightly under for aaditional rigiaity. this
was dong as this floor moved up and down by pivoting on
fulcruma immediately out$ide of and to one side of-the safe
box.

The lever aupparting the flour 2 inches above the

wooden subfloor was formed from a 3/16 inch bronze rod. This

rod baa bent into a Q aidci lcC'Lxrglc thin¢ two 7 irch
abort aides and a it; irch side. Three inches of each of
the short c;des 13 Qj€(.ti2d past the fulcrums through the
wall 5 the sa¢s box to be atts ched to and suprort th

hardware cloth floor. The remainin g 4 inches of the chart

{9

idc and thc counecting 163 inches of rod were outside the

safe buy. Lith thc fulcrum; outside the safe box, §_could

ti,“inr the £100: downwaré. A minimum of 50 grams wa,
B u

necessary to tip the fluor. The lever was counter balanced

a
W .1

a;c 1;g t.u9 163 inches arm by a brags weight.

Resting in the middle of the 16§~1nch arm was a
pivoting mercury switch (an thus cilant). The mercury
switch was pivoted on a separate 2 inch arm. Thus whcn_§
wag in the eafe ho x the 16:: nch extc-dal arm rais ed anfi it
tilted the mercury switch upwards. When this occured, the
mercury snitch 0p ad a normally closed circuit to a
recording pen on a four pen Gc*brands recorder (Model 30.

24). This tilting floor and mercury switch system formed
an automatic recoréing detector for the duration of the QE
presence in or absence frcm the 8”f@ box duri1g an experi-
mental Be 2‘1 on.

Another pen of the recorcer automatically recorded the

operation of t1 3 mechanical predator. A third pen was

uaed to record the time spent in the "goal region”. The

29

goal region was that port of

332:;
the

inches from the safe box

The beginning of

«a
alley.

b1

for

m

orange cardboard lying on the wooaen subfloor.

had about the some brightness, to 3, as

wooden floor. By means of a

through the pen recorder, the time

the 6 foot alley beginning
and extending the romainder of

this goal region was marked

'3 convenience by the front edge of a nouoro of light

This orange
the surrounding
hand held switch, 3 recorded,

3 spent in the goal

region. g depressed and held the button when ever and as
long as any part of §_wa3 in the goal region. By moans of
those recordings, a permanent, accurate record was formed
of tho rumber of sofa ho" and goal region entrance and
exits and the amount of time spent in each and in the

inleyway connecting the two.

A stainless steel firinli

7‘1”“.‘3

‘0'- V

n; tube 'ectod into the

goal region 2& inches from the ease of the goal region

or ;6 inches from the safe box.

inches in o he

(22‘

ware cloth floor.

0
.

the side wall but througa a roceosoo

translucent white

"fial With its éna l

The tube did not

plastic 23

The tube projocteé 1%
inch above the hard-

projoct directly through

O
inches in

inch 603;. The drinkir; tabs #J‘ ouspenfio: on the outaifie
from a rulher stoppered graduated cylinder. Through this
means, é could measure the amount of water av‘fi consumed
during a daily session.

30

In spite of the high walls of the app aratus and over-
head aurora tructures, §;could cle: zrly ob: 3rv Ea behavior
in the apparatus through overhead mirrors. 033 mirror was
mounted above the safe box, aszothor over the alloy.

The laboratory was W3“ 11 lllomL33t3d my dif fu 336
fluroeso.nt lamps, but due to the 53033133 of the apparatus,
the ill 3Lu3 3’33 12* 3id3 the apgara.tu3 was r3 3*uo ed an
further refiuced the low albedo of £33 interior walls.
Table I gives the ill 32ira‘ion in foot candles along the

floor of *Exc apgaratus.

r} ’1. -1-9 I“; I
“.-JJ‘¢.¢‘
*m

Tllumination, in fact c3;dles, of the floor of the
apparatus measured at one foot intervals from the mats
box end.

 

 

. sgf- :x r- . 3;, 3
oistance in :3 9 hr alloy 153; rebign
33me Q 11 2* "' 4 6 a
lOOt Cunu1032.03.8.81.5 108W 2: 8 .4- 30 2. .
*readinga of 1.6 and 1.8 were made immediately on either
side of €33 33f box wall, 33103 333 located at ox3otly

two feet.

The entire apparatus was raised 20% inches above the
laboratory floor for fiLg convenience in taking fig in an&
out of the apparatus and for observing §§ through the
mirrors.

The zuan’ng 3pp3r3tus was in one room while all con—
trol circuitrj. power aupplies, and the pen recorders were

in an adjacent room. The wiring connecting the two was fed

through a hole in the wall. This was 3333 to isolate £3
from as much extraneous noise as poooiblc; especially
circuitry cm*ttor ac3033ohyiru the recording of 3g.various
activities. Pilot work ficmonotratcd dcmonot'otod that
abrupt noises tended to inhibit the g3,

The operation of the mechanical procatcr produced
noise. $313 noise was considerco part oi the stimulus
complex oicoallin~ actual throat to the 3. Below are two
tables lioting noise levels within the apparatus during
the operation of the mochonicol predator.

‘1'“. A "." I “i ‘
1.73..)Hul I T
mu...-

.5

Noise level 3, in decibels, recorded with the fioioe
Level “333? rlorco onc brzioo the (: aging tuoc. floodings

were taken at one foot into-V313 as tho mes clm onioul predator
move 3 toxcrdc tPc oofo box. All roooingo in?lucc 3.; ambient

noise 19.31 of 72 db.

 

,- .“1 ._,..- _ . ~~
{30-213. I’C‘L‘XLOH alloy 1
fie toncc in Iccmt G 1 E 1 i 5

 

‘001501l 78.0 79.0 77.0 77.5 78.0 73.0 78.5

 

Hoisc lovclo, in dcoibcls, rccorccc 3:33 tlc Loiae
Level Motor microphone in the middle of the oofo box. cadinus
1': CI"? ta! Kié'; got 0210 f0 “CD it IQlV'lS 3.“ :11; 143,\LL~1_...\.(...L. lxredatcl‘
.movcd towards the safe box. All readings incluoo an ambient
noise level of 72 db.

 

 

 

 

(«‘3 7 154' ,- 3; new: r .3 '1 .q-vr
-7 41:):’:m fi‘Jh- -, «A.
1 ‘u «- '1‘ ‘
a 3 once 1ng t 0 l 2 i» 4 5 6
; I~.:~g..:--.;Q~~- «v-0..- org-2:" «Tibia-y’all-J" —'—..--r~v '- 3| II. "F -. ‘i‘r.’ :um: ' 1 ,: 7'P‘ ‘—
Lccibolr ; .3} 3.5 (5.3 (int; {3.) I ‘.3 .0
l

Figure 1. Isometric drawing of the
The key to the numbered ‘

following

appara .. .
9

componen.s

 

page.

 

 

 

 

vmh4adfzs

  

l...’

2'32; .1“?! ca

ajporoivq

m ‘ ‘pr 7* IV
4.. '13.)..3 \-
~0- mm

1’- .-O 4. b- V - a». .'--v.. ..- ,~- . . ,. "V. 1'7 -s.- :- '._. .~ 1-. . 5—. 1 41".
.fiingt.’ U3 t £ ~t'o‘l -:-CL\;‘ LEA lfi‘Z'UJAbzab‘J 0:3. Liar—7 ('ufik‘fil‘lfllun V0.01

o]*L.w~ L;3; orfor.

5&2

1 -
2 -
3 -
W i

(1‘ “J 0. U1
1 I I 3

t
a...

‘5
.1. 0'.
1 f.
.C. .’
q
.l. -
I
.5} -
C: -
4'
w .
as b "'

‘ . 3.. .. "‘ rw. WV .53.. ‘1'
«my .3 i-;*io i. :33 u'MpP3ougo are wistod in
'1

Corroroni

‘¢.|-p. ‘- .

» W' 4...... -.
almighiuia £4901.

C4
1
C:
‘0
C.
S...
c"
(72‘
f-
C»

a;
t—J
r-
{.7
5 .
(D
h
E
C
{2,
*3
5‘

king tube
hgrdwaro 3"o.1
io her 303.33336 of moohani«: i predator
Elli 3.111.} ' U".¥'...'t011

miorosyiioh

31'1018

urgflgo unrubouru Sguare

‘3

ovorLe3d compononi ofn ohaziioal predator

 

 

 

 

 

 

 

 

 

Figure 11. Icometric drawin5

of the mechanical predator. L
Below: Numerical key to the ‘

alphabetized componerte. l

:R
o
e

\)03 -o:hu\¢wanahl I

Qomgonentg

aluminum diec
aluminum track
automobile ignition coil
brace

bronze rod

high voltage electrode
insulator connecting
brace and bronze rod
Masonite platform
pine block

standoff

string

track runner

award
(OI-'0

 

wMaJ t2 5

35

£53ng, A within subjects or eteaay state design was used
to provide maximal control of individual differencee and
unknown varieblee. The four eonéitione or phases of the
experiment were: 1) Free Access; 2) Extinction; 3) Requi-
sition; 4) he-extinction.

Each phase lasted 15 Gaye. The experiment, therefore,
required 60 days to complete. Regardless of the phase,
each § was run for one, on minute session per d3y.

In Ihaae l or Free Access g baa unlimited access to the
goal region throughout, without fear of pursuit from the
predator. In Yhase 2, §_waa exposed for the first time to
conditions where goal entries were both nonreinforced and
subject to potential and actual threat from the prefiator.

In ?hase 3, goal entries were again reinforceé as in Phase
1, but such approach wee followed by attack from the
predator. Thus, § in this phase were subject to both fear
and frustration. In Phase 4, g was re—extinguiohod with the
expectation that extinction would take longer than in

Phase 2.

firgeedure. The four experimental phases are described in
detail below. Phase 1 consisted of fifteen days of free
access to water in the experimental apparatus. Buring

this hase, entrance into the goal area or region éid not
result in the approach of the mechanical predator. Through-

out this pause. the mechanical predator was present and

36

electrically charged though always stationary at the end
of the alley opposite the safe box. In Phase 2. the water
reinforcement was not available although the drinking tube
was still in place. Also, any entry by any part of §,into
the goal reb ion resulted in the approach of the mechanical
predator after a fixed delay of 7 seconds. For Phase 3.
water reinforcement was reinstated. In all other respects
it was identical to Phaeo 2. During this phase, 3 learned
to approach the goal stimuli when these stimuli were
associated with frustration and threat or punishment. The
last phone, Phase 4. was procedurally identical to Phase 2.
the goal entry response was the major dependent variable.
This response was defined as any traversal trom.the safe
box to the entry of guy part otI§ into the goal region.
Ercept for the differential treatment: given during
H Phase 1 and 3, the remainder of the running procedures
were uniform throughout the four phases. ‘All §§|wero
continuously meintained on the water deprivation schedule
to which they care adapted prior to the beginning of the
experiment. This schedule was ten minutes of access to
water in the individual hone cages once a day not sooner

than thirty minutes after an.§_had been run for that day.
The running order of the animals. for each day, was eyetemu

atioally varied so that no S was consistently run in the

same position within a day as on the proceeding day. §g.

37
were brought, individually, from the colony room into
the tasting room, run, and then returned to the colony
room. This was done to prevent any g3 exposure to
apparatus coundc while not being run.

An equally uniform running procedurc was used across
all phases. A g was placed in the safe box; g then turned
on the switch controlling the paper feed for the remote
pen recorfior. ‘E then started a stopwatch while at the
ammo timo lifting a panel from in front of the safe box
hole. The otcpwatchtaa used to takc the session length
and the 7 second delay between §LQ entry into the goal
region and operation of the mechanical predator. The panel
prevented §g_leaving the safe box until the stapwatch woo
started. §_thon cat in a position whore he could observe
§ via the overhoad mirrcro on the apparatuc. In this pos-
1tion E also Operated the switches which controlled the
remote recording of the time spent by §.in the goal area on
each entry and the switch which turned on the motor driving
the mechanical predator. .§ also cht notes on g g behavior
during tho session.

Throughout Phases 2, 3, and 4. the mechanical predator
was actuated only after a constant delay of 7 seconds after
the entrance of.§ into tho goal region. A 7 second aolcy
was chosen on the basic of pilot work. Dolayo shorter
than 7 seconds were found to retard the acquisition of

tho approach response which would have prolonged Phase 3.

In Phases 2, 3, and 4, the mechanical predator was
prepellcd to and fro at a conotant velocity of 1 foot per
second. This velocity was chosen so that firhad to retreat
from the goal and into the safe box at minimally a fast
walk or slow run to avoid contact with he advancing disc.
This was done to reduce variability in flight from the
goal. The advantages or disadvantages of compelling §_to
run as swiftly as possible were unknown.

The response by glupon which actuation of the mechan-
ical predator was contingent woo the same in Phases 2. 3,

and 4. S was in no way forced to remain in the goal

region after entering.

RESULTS AED DISCUSSION
§§ati§tical Agalzsio. The data are presented in graphic and
tabular form at the and of this section and in Appendix I.

(5

Figure III gives the rossonso rate (number at times 3_cntors
tho goal region in s 10 minute session) across sessions for
the four experimental phases. A analysis of variance response
rates during the two extinction phases (Phases 2 and 4) showed
that there is a significant difference in the absolute
nagnitudoa of response rates. The response rate is aigni~
Iicantly higher in Phase 4 than in Phase 2. This is clear
evidence that the learning during Phase 3 had the prodicted
effect of conditioning the approach reoponao to the goal

region under threat of attack from the predator.

TABLE V

Analysis of variance (Edwards. 1960) or response rates
across sessions of Phase 2 versus Phase 4.

 

 

64?;Vfflean gggagea_ F

 

Treatment 1267.30 1 1267.30 7.44“
Error 1363.62 8 170.45
Scasious 169.87 14 12.13 1.72
Treatment X Sessions 140.80 14 10.06 1.43
Error zoggza ;;2 7.04

Total 3730.37 149

* significant at the .05 level
The treatment means and standard deviations for Phase
2 are 3 v1.84, SD =- .717. For Phase 4 they are I «7.65, so
al.981.
39

40

The nonsignificant Sessions effect indicetee that when
the treatment effect is averaged ecross sessions there
remains no effect peculiar to sessions alone. The nonsigni-
ficaet interaction of treatments X sessions indicates that the
extinction process was of the some form in Phase 2 and 4.
differing only in magnitude.

The low response rate recorded during Phase 1 must be
interpreted in terms of £3; task. During Posse 1, fig were
never forced to leave the goal region. Observations of fig;
behavior during Phase 1 indicated that most of the time
accumulated in the goal region was spent drinking (see Table
VI for the mean water consumption across sessions of Phases
1 and 3). The remainder of the time during a Phase 1 session
was spent exploring the aposratue. This orplcrstory behavior
probably accounts for the feet that there was a reoponoe
rate greater than 1 response per session.: Figures liaend
IVb show the mean total time spent in the safe box, alley ondp
goal across sessions during the four phases. Figures IVs and
IVb show that during Phase 1, the gg spent most of each 600
second or 10 minute session in the goal region. This was
not the case during the remaining phases where most of the
time was spent in the safe box.

A further means of comparing the resistance to extinction
of £2 responding was through the difference between regression
coefficients of the curves from Phaeee 2 and 4. Inspection

of the response rates during Phase : ineioatea thst the effect

41

of the extinction sessions was completed by the eighth
session. Therefore, the slopes of both curves were derived
for the first eight aessiona. For those 2, tho regression
coefficient is bxy” -.285. For Phase 4, the regression
coefficient is bxy' -.480. The poolea error term 13 33"447’
This yields a t~rotio of .436, d.f. = 12, uhich is not
significant at the .05 level. This teat shows that when

the differences in absolute magnitude are eliminated, there
is no significant difference in'the rate of extinction in
Phases 2 and 4.

The test of regression coefficients inuicateo that the
tasks prior to extinction is or the utmost importance. glg
rate of reoponding will be greater during extinction, in
terms of the absolute number of responses, after they have
learned to cope with predatory attacks to obtoin reinforce-
ment, a was learned in Phase 3, compared to conditions which
did not equip the §g to cope with predatory attacks (Phase 1).
Once extinction has begun, though, either type of prior
experience results in the same relative rate of extinction.

Another way to analyze fig,behnvior is through the mean
time spent in the 3031 per response. This information is
available in Figure V. Inspection of Figure V and response
ratio given in Figure III suggest a relationship between
roaponae rate and tho time spent in the goal region per rea-
ponao. A Pearson product moment correlation between response

rate and mean time in goal per response computed across

h
H

Beaalana in Phage 1 yieldafi an E n - .fils (Hignificmflt at
the .0035 level). For th33 2, tha correlation has £.2 + .521
(significant at the .D;S level). Fer Phaaa 5, the correlation
was + .713 (significant at the .SQS level). Emile fer rhaae
4. the correlation was glm * .644 (eignificfint at the .335
level).

A3 far as Phase 1 in concerneé. thare 13 a strcng inverse
relationship between reeponae rate and tima in anal per rea-
ponae. Zba fact that this rclatiomshifi is not § 1.00 Can moat
probably be attributed to expleratcry behfivier. For the
remaining 3 phases the ralatienanip seems to be a direct one.
Euring the last 3 phaaea, the reayonae rate and time spefit
in the gnal per renpnnaa can be considered fume ioaa of the
aa&e ftctor. Thia factor appears to be the extent to which
fear aafi frustration prcduced stimuli are congitioxed elicitvra
of the goal reaponat. this aaeumea that both refipanfia rate
and time in £031 per reaponae may bath be indicative of tna
degree to which canditioning hug takan place. Inayectian
of Figure V shows that during Innae 2 the mean time Eiéflt
1n tfie @931 pa? reaporma wag alwtya leaa thlfl the 7 aacond fie-
lay periwd. This may indicate thét fitaying in tha gfial
region was an noxious ua eateriag thfi goal regian. in rhasea
3 and 4, though, time npemt in tag goal par IQBPOHfie egualed
or exceafieé the 7 aecenu delay period axcapt in Phase
3 and late in Phase 4: E rly in Phuae 3 little conditioning
had takan pléca, late.in Phase A, extincticn wag well
afivunced. the resultm with respect to time in the anal were

not specifically predicted althaugh the poaaihility uaa

43

discussed in the introduction under the section labeled
Predator.

Phases 2, 3. and 4 were compared statistically on mean
time in goal region par response by means of matched-pairs
E-taats. The comparisons were made across the 15 session by
pairing sessions in order. The results showed that £3 spent
significantly more time in the goal per response during
Phanu 3 compared with Phase 2 (Td a 7.12, é.f. I 14, signifi-
cant n the .001 level). The gig bah-Avior am m differ
oigniricantly botwcen Phase 3 and 4 (Td a 1.84. d.f. = 14,
not significant at the .05 level). It appears that once the
conditioning to the posited fear and frustration cues had
taken place more time was spent in the goal region than
prior to conditioning. Also, once conditioning toox place,
tho behavior producing the increased time was very resistant
to extinction.

ObgngationaL_Bata. While §ngere free to explore the
apparatus during Phase 1, they aoon learned to avoid an area
approximately 2 inchea in front of the disc. The disc was
charged at all tin... After fig had received a shock apiece
during the firat session (oneug took 2 shocks in a row), they
continued to show interest in the disc but only from a dis-
tanoe. In the remaining sessions of Phase 1, three different
fig ventured clone enough to be shocked more than once (see
Table VII for the frequency of shocks across sessions of the

four phases).

45

region. Also, no Phase 2 progreoeod, the fig behavior in the
safe box changed. During the early oeooiono, §§ frequently
appearcfi very eggitotcd. The; paced about the safe box and
frequently crouched facing the safe box hole. In later
sessions, the g3 spent increasing amounts of time oitting
quietly in one or another corner of the safe box. Also, by
later in Phase 2, most Eg,had entirely abandoned any attempts
to drink. What time was spent in the goal was need to
crouch or nervously econ the surroundings.

Table vII chews that during ?haec 3, the‘flg received
more chock. than our ng any other phase. Three fig’who
received shocks during the first session or this phase because
once they had begun drink :3, they were very loath to
discontinue. Either they did not stop drinking until shocked
or moved towards the cafe box too slowly and hecitnntly as
if it were equally noxious to leave the water and to remain
and face the approaching predator. This hesitant, conflict
like retreat from the goal region wan the usual reason for
receiving shocks during Phase 3. Quite frequently during
the first 3 sessions of ?hase 3, the £3 won a flee the reel
beforo'tho 7 second delay had elapsed. This was not the case
during the letter sessions of Phase 3 as can be eeen in
Figure V.

Observations made on the §g behavior during the latter
part of Phase 3 ouggeeted that the frequent hesitant retreats

from the goal region were the results of partial extinction

I 46

of the avoidance response. ieouming that the argumente
put forth in the Introduction are correct, a partial
extinction of the avoidance response could take place. ibis
is especially true in the letter oeeeicnn of Phase 3 es the
stimuli which ehould elicit avoidance becomes inetead con-
ditioned elicitore for approach and entry into the goal
region. As was previously mentioned, the heeitrnt retreat
iron the goal was the ueuel reason. §g were ehocked during
the latter part of Phase 3. Also, during the latter Part of
this phase, besides an increasing amount of time spent in
the goal region perreeponee meny fig would, with increasing
frequency, emerge from the safe box and begin approaching
the goal reginn soon after the mechenical predator had
reached its closest approach to the safe box and woe then
on its return to the opposite end. Occasionally an §_would
approach and enter the goal region by following on the very
heels of the withdrawing meohenicel predator. Also on the
inoreone at the end of this phase were burote of eggiteted
appearing behavior in the safe box. The So would suddenly
begin bounding about the safe box in a very vigorous way.
These bursts of behavior were usually accompanied by very
swift dashes to the goal region. Finally, in the letter
sessions the usual approach to the goal region was a swift
run from the safe box door without oboerveble signs of
heeitancy or warineee.

Table VII shows that during Phase 4, the fig received

as many shock: as in Phase 2. In Phase 2, though. 8 of the

47

10 shocks ware rcceiveé during the firnt gezsinnn while in

Phase 5 th~ 10 shanks were Shattered throughout fihn phase.

(3

Obanrvat1~nn of thn‘gg’ bahavior during Phana 4 nuggeatsfl that
the reasons for raceivfn: a shock were different than during
Phnsn ?. In Phfise A it nppnnred that an tho ntrnngth of the
approach decranned 30 did the avoidance of the mechanical
predator. Perhaps as tha stimuli elicitiv; Approach infit
their power to do so, these etinuli Rlno lont the pnwnr to
elicit avoidance. Thin might be a tannhie suppnnitinn in
terms of the hypntheaia.

In the early aesgionm of Phase A, the 33 on naveral
ocoaainnn woulfl, on entering the gnal raginn, cnmfilnfiely
ignora the drinking tube and continua moving tnunrda the
statinnnry disc. The closer they approached the dine the
more wary their behavior aphenred to bannme. No fi’qu avar
seen tn come closer than 6 inchaa from the disc. A9 seamians
prngreanefl, the drinkifig tube wag mare and more ignored but
not in favor of approaching the disc, but in favor of can-
tiouwly scanning tha nurrounfla crou?hin3 by the drinking tube
facing in tha directian of the 6130.

During the early massinns of Phase 4, on a number or
occasiona fig'wnre seen to chew and tug vigoroualy at the
tube. Also during the early sessions, the §g_would on
occasion bound and leap about the safe box eapecially after
having attemptad to drink. They would sometimes chew the

edge of the safe box hole and the hardware cloth floor. The

48

incidence of standing on the hind legs increased during this
phase together with vigorous sniffing. Occasionally an‘g
would attempt to jump straight up the sides of the safe box
walla. hone succeeded in reaching the top though. Perhaps
these behaviors were indicative of an increased magnitude of
frustration.

Gggcrhl Considerations:

Becouoe of the orploratory naturo of this experiment, the
decision was made to use the some §g_in both extinction phases
(Phases 2 and 4). The assumption was made that. in rats,
ro-oxtinotion is not retarded and, in Iaot,may be foscilitatod
(Rorth and marten, 1962). Any fascilitation of ro-oxtinction
would work against the hypothesis. In the future, though,
it is recommended that independent groups be used so that
r0~oxtinotion will be unneceooary and thus making assumptions
unnecessary.

The model used in this eXperimont to predict differential
resistance to extinction represents a combination of exist-
ing theories such as competing response theory, frustration
theory, and secondary reinforcement theory. The theory which
best fits the general noool is Elicitation Theory (Denny and
Adelman, 1956).

Thoorico of the drive reduction, inhibition, and
generalization decrement were not viewed as being copablo of
predicting the present results. It is hard to conceive of
any way in which drive reduction or inhibition theories could

predict those reoulto. Generalization docremont theories

49

would stand a better chance by noting that the stimulus
conditions between Phases 1 ané 2 were change to a much
greater extant than between Phases 3 and 4. It is difficult
to see how a generalization decrement position would com-
pletely account for the results in that all fig received
chock during Phase 1 so that the advent of punishment in
Phase 2 was not an entirely novel state of affairs.

In any case, only the present model could account for

the results in the following experiments.

50

 

 

 

 

 

 

 

 

 

 

4.0 TIIIIYTTITTTIFW
PHASEZ

3.0-

2.0_.

I.O__

COLLllLllJllllllll

0123456789l01ll2|5|4|5

 

 

 

 

 

SESSION

 

 

 

‘

 

 

I5.OTIIITWTITIITIr
PHASE4
I0.0_
5.0-
011111111L1_1411
Ol23456789l0|ll2l3l4l5

SESSION

6.0TIIITITIITTIII
5.0..
u
'—
é
MILOL
«a
2
o
(L
m
3
23.0.. PHASE'
<
u
2
2.04111i1111111111
Ol23456789|on|2|3l4l5
SESSION
30.0 ITWITTTFITrITr
u, PHASE 3
p.
(20.0
a r-
u
U)
2
o
c.
3m
a ' ‘
z
<
u
‘2
O LllljllLJlJllL
Ol23456789l0llltl3l4ls
SESSION
Figure 111.

Experiment I.
graph to graph.

Mean response rate per session of the four phases of
Take note of the change in ordinate units from

MEAN TOTAL TIME (Ssc.)

MEAN TOTAL TIME (5“)

51

 

 

 

 

 

 

 

 

 

 

 

600 I I I I T I I I I I I I I I I
500,.
PHASE I
400,
300 _ GOAL
200;
ALLEY
IooL.
SAFE. sex

0 J L I L 41 l L L L 1 l I J L I

o I 2 3 4 5 6 7 8 9 I0 II I2 I5 l4 I5

SESSION

600 __:_..
500_ ALLEY
400;
300— SAFE BOX
zooL PHASE 2
IooL
o L I L J I i l 41’ L L_ L l J_ l L I

C) I z 3 ‘4 5 6 7 8 9 Io II I2 I3 I4- IS

SESSION
Figure IVs. Mean total time spent in the safe box, alley, and

goal region per session of Phases 1 and 2 of Experiment I.

 

 

MEAN TOTAL TIME (5“)

MEAN TOTAL TIME (Scc.)

52

 

 

 

 

 

 

 

 

 

coo
500,

400

r ALLEY
soo_
200.. SAFE BOX
PHASE 3

I00.

0 J I l l l l i l 1 1 L l l l l
o I 2 545 678 9|OIIIZI5I4IS
SESSION
600 r I I I T T I I I I I I I

500
F ALLEY
400_.
300% SAFE BOX
200_
PHASE 4
IooL
o I I lo I I I L I L, LL_ L L I I I
OI 25456783IOIII2I5I4I5
SESSION

Figure IVb. Mean total time spent in the safe box, alley, and
goal region per session of Phases 3 and h of Experiment I.

 

 

MEAN TIME IN GOAL PER RESPONSE (Sec)

MEAN TIME IN GOAL PER RESPONSE (See)

I?
0

5

I00

8

O
O

4
O

O
0

0‘
0

U‘

0

53

 

 

 

 

 

 

 

 

I l I I I I I I I I I I I If
‘ PHASE I "
r— —-
P— —I
I— d
F ‘

I L LJ 1 l l l I l 1 L1 1 L

OI23456789IO|II2|5I4I5
SESSION
I I I I I I I I I I T FT T I
PHASE 3

 

 

I I III I I I I I I I I I I I
OI 2 3456759IOIIIZI3I4I5
SESSION

Figure V.

I5

IO

0

I5

 

O

 

IIIjIIITIIITTII

PHASE 2

 

 

 

I IIJ I III I I I III I I I L
o I 2 345 6 7 89IOII IZISMIS
SESSION

 

 

 

 

 

IIIIIIIIIIIIII
23456789IOIIIZI3I4I5
SESSION

 

I
O I

Mean time in the goal region per response per session

of the four phases of Experiment I.
at the 7 second mark indicates the time of onset of the mechanical
predator from the time of an 3's entry into the goal region.

The broken, horizontal line

54

 

 

 

o A n o a a o o o H o w a H o v 33A
m N N m u w o n a A m H o o m n 32$
0.. o o o o o o o c m o o o A a m 3.5
o o o A o o o o n A o o o H w a 25E
hazeuvcuu noonm
NA 2 W m a w m! 44 In a H 542$

 

.H wnoawumnum Ho Gunman know one we moanmon hag co>fiouon exoonn we hosenrehm

 

v.n p.w wnn ¢.¢ v.m u.n ch mow o.w w. w. m. do A. a. H0¢Q3.No .dl
ma wa nfi‘ NA NA CHII MW n 1N o m «1 INI N fl

 

 

0.5 0.6 0.5 +.o m.> o.m 0.0 mow N.> ¢.b v.» mos mob o.> w.n have) no
nfiJaflfidSaprIstlswufignew

 

 

.H aucldnoanm no m use a seesaw we moan-ea new A.Hlv neuvgnshsoe Ream) use:
g

ue

v.

I II:

‘0. III

II-‘ I

gargantuan II

This experiment constituted the test of the second
hypothesis. Given the confirmation of the first hypothesis
obtained in Experiment I, it was assumed that the internal
stimuli from fear-frustration continued to serve as discrim-
inative stimuli. Additional support for the present analysis
through a procedure which allowed for the manipulation of
the opportunity to make competing responses in association
with the discriminative stimuli. I

It was assumed that extinction is an active process
involving the conditioning of frustration responses to the
goal stimuli so that the goal stimuli elicit responses in
competition with continued goal entry. But. since fear-
fruetretion stimuli are assumed to be capable of eliciting
goal entry because of the conditioning which took place
during acquisition, it would appear that an increased inten-
sity of the internal stimuli would be necessary to elicit the
occupating responses. Presumably. this should occur when
reinforcement is removed.

To demonstrate a predictable differential, it was further
assumed that the longer gg’were allowed to remain in the
immediate vicinity of the goal stimuli during non-reinforce-
ment the greater the likelihood that frustration responses
would occur and compete with goal entry. A group allowed to

55

56

spend a longer time in the goal region should display
lees resistance to extinction than a group spending less

time in the goal region.

fietggg
Subjegjg. The §3_vere 10 naive female hooded rats from the
colony maintained by the Department of Psychology of
Kiehigan State university. All garnere from 110 to 120
days old at the start of the experiment.

Beginning two weeks prior to the start of the experi-
ment. each §_wss handled for five minutes a day. Beginning
one week prior'to the start of the experiment, §3.were placed
on a water deprivation schedule of ten minutes of sccess to
water in the individual home cages every 24-hours. Food was
continuously available. Also, during the last week prior
to the beginning of this experiment, gg,were. individually,
allowed ten minutes per day to explore the experimental

upperetus and drink in the apparatus.

Apparafing. The apparatus used in Experiment I was used in
this experiment without modification.

‘zzgggggggp the procedure used called for a combination within
and between subjects design. The four conditions or phases
of the experiment were: 1.) Acquisition; 2.) Extinction;
3.) Reecquieition; 4.) Be-extinction.

than. 1 lasted fifteen days as did Phase 2. Phase 3
looted ten days. Phase 4 lasted five days. Regardless of

h

57

the phase, each g‘wee run for one, ten minute session per
day.

During Phase 1 or Acquisition, the gs learned to approach
the goal. They were water reinforced, but. they entered the
5081 under conditions of threat from the mechanical predator
which in turn resulted in termination of reinforcement on each
trial. This phase was prooedurally identical to Phase 3 of
Experiment I. Any entry by any part of an §_into the goal
resulted in the approach of the mechanical Predator after
a fixed delay of 7 seconds.

At the end or this phase. the 10 §§.were divided into
two groups. Each group had approximately the some mean
response rate across the sessions of Phase 1. The groups
were eeeembled from pairs of §§, more or less matched for
mean response rate eoroee sessions (see Table VIII below).
The response was defined as a traversal from the safe box
into the goal region. All the measures taken in Experiment
I_were taken in Experiment II.

The groups were randomly assigned to the eXperimentel
conditions of Ehase 2.

In Phone 2 ell gs, regardless of group, underwent
extinction. Ibis is, that the water reinforcement was no
longer available upon reaching the goal. The independent
verieble wee the enount of time the §g|were allowed in the
goal prior to the approach of the mechanical predator. There

was two conditions of the independent variable, 3 seconds

 

Group composition in terms of the mean response rate aorooo
Phaoo 1.

 

 

 

2.2222 A W W o H w :
agpjgct geag {espouse gate gugjggt goon response rats

1. 17.5 6. 17.6

2. 16.1 7. 15.9

3._ 12.6 8. 12.3

4. 11.9 9. 10.1

5. ___fi&§_ , 10. _~_§&;~_

63.9 64.0
2:12.78 2:12.80

and 11 seconds. Group A received the 3 second condition;
Group B the 11 second condition.

Three seconds and 11 seconds are both 4 seconds away from
the 7 second delay usad in Phase 1. These values were chosen
for two reasons: 1) pilot work indicated that a 3 second
delay strongly retarded initial acquisition of the approach
response. Because the second hypothesis predicts greater
resistance to extinction for Group A (3 seconds) compared
to Group B (11 coconda) making the short delay a time known
to retard responding during acquisition strengthens the
design it differences develop in the predicted direction.

2) Both 3 and 11 seconds represent equal changes in time
from 7 seconds and should constitute approximately an equal

amount of generalization decrement, if any. Bidirectional

59

‘gencralization gradients are typically symetrical.

It must be understood that regardless of the time used,
either 3 or 11 seconds, the §g|were not forced to remain in
the goal for any length of time. They were simply allowed
more or lees time in which to remain in the presence of
the goal cues.

Phase 3 was identical to Phase 1 with the sole exception
that Phase 3 was ten days long. fhat is, each snimel was
given ten daily sessions, one session per day. Phase 3 with
its return to e 7 second predator delay was used to allow
the groups to return to their response rates achieved towards
the end of Phase 1. ihis was necessary too, as Yhase 4 or
Reoxtinction constituted a test of the reliability of the
results obtained in Phase 2.

In Phase 4, reinforcement was not available. Group A
which had received a 5 second delay during fihase 2 now
received an 11 second delay. Group B-which had received an
11 second delay in Phase 2, now received e 3 second delay.

By reversing the conditions for the two experimental
groups from those imposed during Phase 2, two things were
accomplished. first, this procedure was a stringent test
of the hypothesis. If the results were again in the pre-
dicted direction, this would be interpreted as strong support
for the hypothesis and eliminate the possibility that
differences found in Ehsse 2 were due to some uncontrolled

differences between the groups. Secondly, the reversal of

60

conditions made pooeible a within-subgects analysis.
Excegt for the differencee.in procedure diecueoed above,
all other procedures need in this experiment were identical

to thoee used in Experiment I.

F .USULTS fC DIS CU'TSIOH‘V

Thad data are prasented in gr9phlo and tabular form
at the and of this section 911d in Appendix II. Figure VI
51999 the r99p0n9e rate for Groups A and B aoroas aeaaions of
the four experimental phases. Analysiu of variance of the
responae rates of Group A versus Group B fiurin 9 Ph1se 2 shows
that the two groups did not differ significantly in resistance
to extinction (see the analysis of variance Table IX). By

this te9t, the second hypothesis was not supported.

TABLE 9'

   

analysis of variance (Efiwa.rda, 1960) of Group A versua
Group 3 r9. 990999 rata across Phase 2.

 

 

frcatment 198.80 1 128.60 .40
Error 2559.22 8 319.53
39991099 1036. 97 14 79.16 7.999
Treatment X Sesaiona 145 70 14 15.39 1.44
Error 112 11.L

Total 49

 

‘aignifioant at the .01 level

The treatment mean and 9t9nd9rd deviation for Group 9
are fi’u 8.04, 99-” 3.417. For Group B they are :1 . .19,

CD a 2.693.

The significant 999910n9 effect indicates that when tha
treatmen t effect 19 aver9ged acro99 sessions there in an
effect peculiar to sessions alone. Inspection of tha Fhaae a
graph in Figure VI suggestea a negatively accelerated function

usually associated with extinction.

61

62

Regardless of the fact that for 11 of the 15 sessions
Group A.maaho were higher than Group 3 means, the treatment
effects ware insufficient to produce a significant difference.
Inspection of the graph of Phase 2 in Figure VI shows that
the greatast absolute difference between group means occured
during the first session. A T-tost of thcec two meana'gavo
a value of T=1.405 with 8 dofo which is not significant
at the .05 level. Thus. for the first session of Phase 2,
the results were in the predicted direction but not to a
significant extent. ‘

Turning to Phase 4, Figure VI shows that for at least
‘tho first session, the results are again in the predicted
direction. The data. also, took a form very much like the
first fivc sessions of Phase 2. Inspection of the data indi-
outed once more that if there was any significant difference
betweon the group means it would be in the first session.

A T-toat of these two neaoo gave a value of T=.846 with

8 d.f. which was not even significant at the .20 level.
Thus. for the first session of 11131 9, the results were in
the prefiictcd but to no significant extent.

A further analysis of the response rate data was made
by using each g as its own control. That is, each 99} rea-
ponac rate under the 3 second delay condition was matched
against its resyonoe rate for the 11 second condition. A
matched-pairs T-test was used to analyze. This gave a TC 9

2.04 with 9 d.£. which is significant at the .10 level but

63

not at the .05 level. hith individual differences controlled,
the orpected trend is more apparent.

The lack of clearly significant results not withetenéing,
the response rate data euggeet that predicted effecte are
real. The effect may be only of short duration. The effect
is replicable if Phase 4 is consioered e replication of Phase
2. The difficulty appears to lie in an ineufficient number
of subjects, as the main effects are obscured by a high degree
of voriebility.

Inspection of the re3ponec rate data from rheee l inci—
cates a great similarity to the results obtained in rheae 3
of Experiment I. This suggests that there was no savings
for the fig in Experiment I at the beginning of Phase 3. This
was not the case for 23 in Phase 3 of Exreriment II. The
combined mean reeponee rate of Groups i and B for the first
session of Phase 1 of Experiment II is ? =-3.9. The combined
mean response rate of Groups A and B to: the first o'coion
of Phase 3 of Experiment II is ‘ a 16.7. This gave a eevinge
score of 303% in terme of response rate. A matched-pairs T-
teet of the mean response rates of the first sessions of
Phases 1 and 3 gave a Td a 4.870 with 9 d.f. A T-retio of
this line ie significant at the .CCl level. Inspection of

hence 2 end 4

hi

the response rates for the first eeeeiore of
also Buggected a savings. The mean combined rccronee rate

for Groups A and B for session 1 of Phase 2 is _ x 15.9. For
.li'

64

the first cccoiou of Phase 4, the moon is x a 8.9. Skis
represents a éfifi savings for the first session. A matched-
paira T-toat of the near rccponcc rates gave a Td a 2.609.
With a 9 d.f. the T-valuc is significant at the .05 level.
These savings represent a significant amount of positive
transfer.

The following analysis is of particular importanccibr
the major hypothesis. In Figures VIIa and VIIb, the graphs
for Ihaoe 2 and 4 show a difference in the moan total time
spent in the goal region between the group receiving the 3
second and the group receiving the 11 second delay concition.
This difference is seen more clearly in Figure VIII. iho
graphs of the mean time in the goal per response for Fhaaeo
2 and 4 Show that when a group is allowed more time in the
goal region, moré time is spent there. Also, the graphs
for Phases 2 and 4 of Figure VIII show that the two curves,
session for session. arc about equally displaced from the
7 second point. On the ave~c5e, a group under the 11 second
condition spent less than the full delay interval in the gocl
region per response. while a group under the 3 second condition
Spent moro than the delay interval. In line with the argument
put forward in the discussion, it 13 suggested that comething
less than 11 seconcs was ourficient for the maguitudo of
the frustration response to increase erough beyond previously

conditioned lcvols so that the §§ spontaneourly fled the

65

goal region. Under the 3 second delay condition, on any

one trial. the magnituae of the frustration responae éid not
increase sufficiently to result in the §§_spontaneou31y
fleeing the goal region.

This brings up the queotion of why there woo no decline
of time in the goal region per reayonse in Phase 2 (Group
A). bxya + .270; (Grouy B), bxyg +.230. In Phase 2, the
responoo of remaining in‘the goal upon entering may be more
resistant to extinction than the entry reoponoe. lhis would
seem reesona le assuming that the alley cues decline in
approach eliciting power in relation to their dis once from
the @031. Therefore, as extinction progresood, the ayproach
gradient would collapse towards tho goal. The negative
slopes seen in the Phage 4 graph of Figure VIII (Group l,

b = ~.230; Group B, b -.280) may be the result of

xy xyu
repeated extinction making the functional reinforcing pro-
perties of the conditioned elicitors generally more
vunerahle to the extinction process. This is all very hypo—
thetical, but worth conaidering for future research.
In the later sesoiona of Phase 1 and throughout Those 3
there was a decided increase in the mean time in the goal

‘

region per response beyond the 7 second delay period. Tnis
13 the some effect as seen in Phase 3 of Exporiment I and
probably occuro for the same reasons previously prepoood.

Table 1 gives the freouoncy of shocks by soosion for

Groups A and B. Inspection of tho shook frequencies for

66

Phases 2 and 4 show that any éecreased zesista co to extinn

W

ction shown by the 35 under the 11 second delay conéition

1'“)

cannot be attributed to a higher ahock frequency. Ehe _3
under the 3 second delay conditionr Vlvays received tho are eat~
est amount of ehocks. Obsezmx tiona of the So behavior indicates
that a hesitancy to retreat from the goal region or healtanoy
to enter the Qbre box was the usual reason on g receivoé a
shook.l This behavior was menife: to d more freq.uen tLy i.n the
group unfier the 3 second conoition comcared to the group
under the 11 second condition. This differential amount of
hesitancy may have a common cause with the rasponoo of re“
maini ng in he goal region upon entry. The reasons for this
were proviouoly éiscusaed.

Beyond the differentm 13 in observed behavior patterns
reported abovo, the behavior of both groups during Phases
1 and 3 closely resembled the behavior observed for tho §§_
in Phaée 3 of Experiment I. There was an ezual 01 Hit?

(II-

between observed behavior of the go qu i

'3

.5 Fhuae d of

Experiment I ano the behavior of the fig in Experiment II

during Phage 2 axtd 4. With the exception of the di forential

gi-

tendency to hcs1tant retreats ano trnder:: oy to (toy in la

1"!

3031 region, war no'; able to (18cc1n up; reliable difference

(
-
‘

in the manifestation of fear or frustration between Groups

A and B.

MEAN RESPONSE RATE

MEAN RESPONSE RATE

67

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

.300 I I I I I I I I I I I I I I r
/\/
2°”? PHASEI ‘
Io.o_. _
\J
0 11111141111111
OI23456789IOIII2I5I4IS
SESSION
.300 I I I I I I I I II
20.0 /\ _
pj/ \, /\
loo _ s
PHASE3
O LIIIIIllll
0|23456759I0
SEsyoN
GrogpA'
Figure VI.

of EXpsrimont II.

 

 

 

 

3&0 ’I’I I I I I T I I I I I I T"T
W- PHASE 2
IQO _
o I I l I I I I I I I I I I I
O I 2 3 4 S 6 7 8 9IOIII2I3I4I5
SESSON
300 I I I I I
20.0 _ .I
PHASE4
I0.0 _\ ..
\
o I I I I L
O I 2 3 4 S
SESSION
Group 8 -—-~—-—

Mean response rate per session of the four phases

 

MEAN TOTAL TIME (Sec)

MEAN TOTAL TIME (Sec)

600

500

400

300

200

I00

68

 

 

PHASE I

 

 

 

O I Z 3 4 5 6 7 8 9 IO ll I2 I3 l4 I5

300

£00

IOO

 

 

SAFE 50X

 

 

I J l I l l l I l l l l
2 3 4 5 6 7 ‘e 9 I0 II I2 I3 I4- I5
SESSION

 

Group A ——- Group B __

Figure VIIa. Mean total time spent in the safe box, alley, and
goal region per session of Phases 1 and 2 of Experiment II.

69

 

600

500 ._

a
O
O

I

(A
O
O
l

 

MEAN TOTAL TIME (Sec.)
'8
O
I

 

SAFE BOX

PHASE 3

 

 

 

 

 

 

 

 

I00 _. ..
a l J I I I I l l l I
O I Z 3 4 5 G 7 8 9 Io
SESSION
600 I I ,2.___'.——:=_""I‘
W
500_. ./ _
’1‘
3400 *' SAFE sax -
{’3
u300
2 PHASE 4
I-
_J 200 __ _
<
’—
O
I»- ma _ _+
z
(
“J
2 O L l l 1 1
O I 2 3 4 S
SESSION
Group A ———- Group B __

Figure VIIb. Mean total time spent in the safe box, alley, and
goal region per session of Phases 3 and h of EXperiment II.

MEAN TIME IN GOAL PER RESPONSE (scc)

MEAN TIME IN GOAL PER RESPONSE (8“)

I 5.0

I0.0

70'

 

 

 

 

 

 

 

 

 

 

 

OLIIIIllllllllll
OI234sc7eemnmwum
sesmou
ISO IIFTFITTI
PHASES
Iw_ .
W
LAvJ _________________
wL _
o I I I I I I I I I I
OI25456789m
SESSION

Group A

I50

II
I0.0

 

TTIIIIrIIIITII

PHASE 2

 

_----———————.———_——n_--.—.————-——_--.--

11111141111111

 

 

 

OI 254 5 6789IOIIIZI3I4I5

I50

II
10.0

50

0

SESSION

 

IFIWT

PHASE 4

-—--———-————q

 

 

1L111

O I 2 3 4 5
SESSION

 

Group B -——--——-

Figure VIII. Mean time in the goal region per response per
session of the four phases of EXperiment II. The broken,
horizontal lines at the 3, 7, and 11 second marks indicate
the time of onset or the mechanical predator from the time
of an gig entry into the goal region, depending on the
group and the phaselsee the text).

71

0.6 $6 0.6 won wow mo+ o.» N.m new ecn weasozoo .Hfl Mum 93.5

fin. «a we we to in ad 3. 3. a: 3...:28 4s m... .325
Ilqm. m .19 H m m eIINIIMIIIwI .3339“
wig

 

 

.6 ex. e.» e4 0A n4 mg m.“ mam m.” or.“ In a. e. a. message Jam .m muons
.. e o.» «A. In O; 0.6 e. n ed ON Ga 04 .3 a. a. e. 3538 Jan .4 goes

 

 

.HH 383.:th «O
n 2.3 H mesefim magma m 9.3 < smack... no ab :23" a: nowumaaeace have) seem

 

"I

.2 IV

.‘vl.

.I\

72‘

 

 

 

A o H H H ”enough.“ Neonm ”m M593
0 o H o 0 Omega.» MOO d. :93
m e n M a mmowmamm
«Iummmm
N a o a e w m a v w hasaaeosu neon“ .m geese
v m u H n m u HI o H hoseuasmmImmmmm .a apnea
Ids. m m r m m w n m a 0 on
”Iummmm
O o o o o o o o M a o H o a o nos-seoua use.» .n means
. , : homage .4 98.3
MIumuaM

a m o o a o a m. H a m H. n . n «a nesuavsnu gamma .n anonu

 

muummmqu

VIYII. . I 4.... . I! ,I I~

ah

LIBI:T III

 

Due to the failure to obtain statistically eignifice-t
results in Experiment II, Experiment II was replicated in

this experiment. The number of subjects was increased.

Subjegtg, The §g_were l4 naive female hooded rate from the
colony maintained by the Departmentof Psychology of Michigan
State University. All §§,were from 110 to 120 days Old at
the start of the experiment.

Beginning two weeks prior to the start of the experi-
ment. each.§,uas handled for five minutes a day. Beginning
one week prior to the start of the experiment, §g,were placed
on a water deprivation schedule of ten minutes of access to
water in the individual home cages every ZI-houre. Food was
continuously available. Also, during the lent week prior to
the beginning of the eXperiment, g; were, individually allowed
ten minutes per day to explore the exyerimental apparatus

and drink in the apparatus.

A arat 9. ins apparatus used in Experiments I and II was
M

used in this exPeriment without modification.

Procedure. The procedure for this experiment wee identical
to that need in Experiment II with but one exception. In
the present OXperiment, Phases 2, 3 and 4 (Extinction, Ho-
acquietion, and Re-extinction respectively) were all reduced

73

74

in length to five days each. That is. five daily sessions
per each g, one session per day. This reduction in phase
length was justified by the results from Experiment II.
Exporbment II results showed that if significant differences
were to occur in Phases 2 and 4, they would occur during the
first few days. EIperiment II also showed that no more than
five days were needed for Phase 3 to return ggbto their
steady response rates obtained at the end of Phase 1. It was
apparent from Experiment II that Phase 1 or Acquisition had
to remain fifteen days long.

As in Experiment II. at the end of Phase 1. the 14 fig
were divided into two groups. Each group had approximately
the name mean response rate corona oeeeione. Again, the
groups were assembled from.peirs of §§,more or less matched
for mean response rate across sessions (see Table III below
and Figure in the results section). A responee was de-

fined as a traverse from the safe box into the goal region.

75

TABLE 11;

Group composition in terms of the mean response rate across
Phase 1. ‘

 

 

 

 

W? 1; w: W... 5”... 1.1219 D 1 __ :3.
§EEJ£££.:§F33 r9399fifli_£§£2_.._ »1§221§££ Roan geaponag rate
1 "" 17080 P: '- 21047
2 - 14.47 9 - 14.40
3 - 12.73 10 - 13.20.
4 - 11.93 11 - 9.53
5 ” 9.00 12 ¢ 9.53
6 "" 8080 13 "' 600°
7 - .2322 14 - 5,60
‘ 77093 ' 79073
I a 11.13 I =- 11.39

Each group was then randomly assigned to the experimental
conditions in Phase 2.

EESULTS goo DISCUfiSIGH

The date are presented in graphic and tabular form at
the end of this oeeoion and in Appendix III. Figure IX
.giveo the response rate for Groups C and D eoroae eeoeione
or the four experimental phases. As in lxyerimeet II, inspec-
tion of the graph of Phase 2 in Figure II shows that the
greatest éifference between group means occured, no might be
expected during the first session. A Tnteet of these two means
ave a value 1:1.703 with 12 d.£., thio difference is not signi-
ficant at the .05 level. A likely reason for this nonsignificant
result lies in the response rote of one § of Group C. Thio a
made only 1 reoponoe during the first eeeoion. The response
ratee for the remaining §§_of Group I in the first session
ranged from 13 to 24. The response rate of the one g
badly skewed the data. khan the date from g with the lowest
responne rate was discarded from Eggg Groups 0 and D, a T-
teet of the reoulting two means (Group C, Y'2 17.8}: Group B
X‘z 10.83) gave a T32.79 which with 10 d.f. is significaot
at the .02 level. Thus. for the first session of Phase 2.
the rooulto wore in the predicted direction and were eigni-
rioant when the data of a highly Coviont g were excluded.
Turning to Phase 4, Figure IX shows that the results
are again in the predicted direction. Once again, the
greatest @ifference between group means occured during the

first seeeion. i T-teet of theae two moane gave a value

76

77

of Ta 3.20 which with 12 d.f. is significant at the .01
level. To be consistent with the analyeie in EIperiment II,
the data from the §,with the lowest reoponae rate for this
session was discarded from.§g§h Groups 0 and D. The two
lowest performing §g_1n Phase 4 were not the some two lowest
performing §g,in Phase 2. After the removal of the data for
the two poorest performing §ghin session 1 of Phase 4, the
mean for Group 6 become 7.33 versus the previoue mean of
6.71. For Group D, tho mean became 15.6? versus the previoue
mean of 14.71. A intact or these reeulting tub memos gave a
in 3.10 with.10 d.t.'uhioh is significant at the .02 level.
Thus. for the first session of Phase 4. the results were
clearly in the predicted direction. The second hypothesis
was supported.

A further anelyeia or the response rate data was made
by using each.§ as its own control (No data were excluded).
That is, eachwfiflg response rate under the 3 second delay
condition was matched against its response rate for the 11
second condition. A natched~paire T-teet was need to analyze
the reeulte. This gave a Tia 3.64 with 13 d.f. which is
significant at the .01 level. This matched-paire analysis
clearly supports the predicted effect.

Inspection of the response rate data from Phase 1
indicateo a great similarity to the reeulte obtained in
Phone 3 of Experiment I and Phase 1 of Experiment II. In

this orperiment as in Experiment II, there is some savings

78

in roaoquioition. The combineo mean rosponee rate of
Groups C an& D for the first session of Phase 1 of EXperimcnt
III is Kin 3.93. The combineo mean response rate or Groupe
C and D for the first eeeeicn of Phase 3 of Experiment III
is 3'2 10.76. This represents a aavinge of 288% in terms of
reopenee rate. A matched—pairs Tuteat of the mean response
ratee of the first oeoaione or Phaeee l and 3 (using each
3 an its own control) gave 3 Ti 3 3.205 with 13 d.f. which is
significant at the .01 level. ‘Inepeotion of the response rates
for the first eeeeione of Pheeee 2 and 4 revealed only a 14%
savings. Compare this savings with the 44% savinge round
in Experiment 11. A matchedwpaire Tnteat between Phases 2
and 4 of Experiment III gave a Ta u .675. d.f. a 13. The
14% savings represents a nonoignificent savings in reeponfiing.
Probably the shortening of Phase 2 (or extinction) from
Ezporiment II to Experiment III accounts for the appreciable
lose in savings. The five ecseione of Phase 2 in Experiment
111 may have been an insufficient number of sessions to allow
the doveloyment of an appreciable amount of positive trane~
fer.

A3 in Exyeriment II, the nelyeie of tire measures
it of particular importance for the major hypothesis. ¥igure
1 presents the mean total time spent in the goal region
for the group receiving the 3 second delay ccnd tion and the
group receiving the 11 eeccnd dele‘. The difference noted

here we: also econ in Experieent II. Thie filiference can be

79

better appreciated in Figure XI. The graphs of the mean

time in the goal per response for Phases 2 and 4 show, as
they did in Experiment II, that when a group is allowed more
time in the goal region, more time is spent there. In Figure
II. the two curves are not quite so equally displaced from
the 7 aocond point as was the case in EXporimont II. However,
the effects were very similar. The 11 second group again
spent less than the full delay interval in the goal region not
rooponao while the group under the 3 second condition Spent
more than the delay interval. As in Experiment II, the

eons explanation could apply: Hamely, that on ll second delay
is sufficient timo for the frustration response to increase
beyond the previously conditioned levels, while a 3 second
delay is not.,

One of the moat interesting aspects of the extinction
phones for those data is that the time in goal par rooponoo
tends to inoreaoa rather than decrease. This in especially
true in Phase 4 for the 11 oecond delay group (Group C). The
inoreaae in goal time in Group C in Phase 4 is possible be-
cauao the predator does not otart moving until after an
ll second ficlay. Why goal time per response increases when
there in no water available can only be explained, in part
at least, by tho assumption that staying in tho goal is
atill the strongest or propotont response in this stimulus
situation.

The slope constants for tho curves soon in the graph

of Phase 2 in Figure KI are bx a + .110 for Group C and

y

 

80.

bxy a + .08: for orcpp r. These'alight positive slopes
might suggest that there was little it any margin for
further confiitioning at the outset of Phase 2. Tho slight
negative 510pe computed for the Group D curve of time in
tho goal region per response in Fhaso 4 suggests that since
this group started out spending a deal of time in the goal
per response compared to Group C in Phaso 2, Group D'o
performance may have been near maximal in the first session
nné thus had no place to go but Gown.

As in Exporimont II, in the final four sessions of
Phase 1 and throughout Phase 3, their is a reliable increase
in the mevt time spent in the goal region per response beyond
the 7 second delay period. Once again, this is the some
effect as oeon in ?hase 3 of prerinent I. Reasons for this
were proposed in Experiment I.

Table XII gives the frequency of shocks for each session
anfi each phone for Groups 0 and D. As in Experiment II,
inspection of the shock frequencies for Phones 2 and 4 again
showa that decreased resistrfico to extinction shown by the
_ §§,un&cr the ll aocond delay condition cannot be attributed
to a higher shook frequency. Tho‘fig under the 3 oec.nd delay
c ndition always received the greatest number of shocks during
the firot oeooicn and too entire phase. A3 in Experiment II,
the observations of £9 behavior again inéicated that hesitoncy

in retreating rem the goal region and entering the safe box

81

were major raascna for an 83 receiving a shock. Again, tha

obmerved differential in amount of hesituncy was seen.

Finally, g was not able to discern any reliable difference

in the manifestation of fear or frustration between Groups

C and D.

£32

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

zo.mmum zo_mmwm zo.mmmm
n v n N . o m ¢ n w 0 ¢ n m _ o
_ q ‘ fl ‘d 0 4w 4 _ _ o 1‘ _ _ _ o
q\anHHHH”//\\\\ 11/
u //// .1o. I o. _\ .10.
m um<Ia
//
.v um<1a
: .loN I \\\\ o~ .loN
\ m um<1a
_ _ _ _ _ on _ _ _ _ om _ _ _ F on
zo_mmum
o
0
.ll a 990.5
I lo.
0 asouw
.HHH acoaduonxm
no nooana pack I I
on» «o :oamuom om
hon ouch ouOQnou
5o: .5 83.3
b _ _ \k _ uh! _ p p p r on

 

 

 

31va 39NOdSBd NVBW

3 .LVd BSNOd $38 NVHW

83

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n .v n u o m V n N _ o m ¢ n o
_ _ 4 . o q 1 lfi 4 . o _ _ _ o
I 02 I 4 00. I co.
m M. mm<IQ DON cow
I 00 I I I
In um<Ia N um<IQ
x00 uh?» F
I 8» fi I 8m 8m
x3 3% /\\ .6. us:
I \ T 54.3 I I
Be ./ 09. co.
/
I (E< com I 439 I can I 8.»
[I‘ll‘ll‘llu
_ b _ _ 00¢ _ _ F b _ 00% 00m
zo_mmum
m. I m. ~— _. o. m s u u .v n o
. _ _ _ d % d _ d _ _ q 0
III a @598
oo.
o 96.5 _ mm<I¢
o u
xon Pin o
1.23 £83
Iauonkm no 03an com
hack on» go scan
Icon you scam?"
Huom cad .noanu ooe
:25 32. on» 5
«can. 3.3 H33
53: .x Rama 2;

 

 

 

 

 

00w

(ms) zwu 1V.LO.L Iwaw

(we) awn. ‘IVlOl nvaw

84

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20—mmum zo.mmmm 2030mm
n .v n m _ o m e n m . o n .v n N _ o
d _ _ _ _ o _ _ . T q o _ d _ _ a o
d w
............................... n --I\\-n um
I! I III
< a
I I m I L n I I m a H.
d W
mm 3
9
.l. ,I l //\>/\\ M510
I0. 0. fl JO. 3 W
................................ : I-------I--------.-II----:----I : C
.v wm<Ia M mm<Ia .m wn<Ia
_ _ _ _ L n. L _ _ _ p m. _ _ _ _ p m.
20.nmum
m. .2 n. N. : o. a o b u 0 ¢ m N . o
q A _ fi fi _ 4 _ q i _ _ a . 4 0
III a 95.5 h w
a w.
U N
0 950.8 m 3 I.
s .w.
w 3
.HHH 8.83 r m I
Iguana no conga 3 N
.33 on» ma do? 9%
Icon hog cocoa r I a V
loo." hag no.3?“ 0. P1
Hdom on» :a and»
.302 J? ohamdb _ wm<Inu
_ _ _ _ _ L _ _ _ _ F L _ _ F m.

 

 

 

85

 

 

 

 

m.p w.w m.o m.m o.m coasmzco .Ha_m
a.> c.m v.m >.¢ m.~ moasmmao .Ha_w
W VII m. m H rmowmvoc
nIcmwxm
m.m m.m o.m $.é o.m m.m n.m n.m w.m ¢.m m.m m.“ H. o. o.H moasmsoo .Ha.w
o.m m.m ~.m o.m 0.5 m.m ~.m m.¢ h.m N.m m. o.~ m. 0.0 H.H neawmnoo .fla m
ma ¢fiIIIwH «a: HflIIIoH m w p w w v n m a rccflmme
H mmmdrm

n ass a ommnm maauzc a ans 0

.HHH pgmafihwmxm mo
QSOHO M0.MW hp aoHS ch fioavmfifimmoo Hmwma nabm

.m among
.0 macho

.Q Queue
.0 nacho

91’. I...

I.o”

.‘u

.2.

86

 

 

 

VN'?‘\O
3" N

3 r"!

H

("I
"\J
H
H
H
F"
vi

 

“vac: «\nn4 anca

wflmcs

“(V r“! O

{\J 1-! Cd

{\3 (“Jr-4

IIIIIIII

(JIM N

houmswwgu MocAw
mumomrmum Moon;

II!!! is.

mxcwwnou

HON

o hoiwsammm Manw
o hoxwfiwmmm goon;
H muowmmmm

hommnwmuw Moon:
mommfivmuw Xuosw
euomeum

HHO

mahocmsawmw ”conm
wfihusmfimmmu Moozm
H anofimmmm

.HHH pcmafiuomum mo momwam

axon 98» unfladv Q and o masons no mMWhn no>woomu whocmm mo hummuwoum

bHH mam¢w

Q

vi:

mfioku
Adena

macaw
macho

macaw
awoke

muonu
nachu

1 'f.

.
u . ‘I c.
,. . I I
.
. . I. u.
~V'
. .
111‘ 1....I it :0 I1 u.
.
. Q
r
0‘
. A .
.
.1 ’I ni“ III! '
I .
.
t. ..
a
.
If .. . P

QOIQLUSIOfiS RED SUfiHiRY

In the introduction to this study, a model was presented
for approach learning and extinction in prcy argoniame co
3 function of repeated predatory attacks. The learning model
involved the assumption that the booic task confronting the
prey organism is to learn to approach a region of the environ~
ment to obtain needed reinforcement and then flee this came
region to avoid the direct physical attack of on approaching
predator. The roeponoe of approaching the goal, it was
argued, is conditioned to the ctimuli accompanying the
emotional responoee of fear and frustration.

Fear is elicited by the operation of the predator.
After a sufficient number of approaches to the goal region,
fear-produced etimuli serve as partial discriminative stimuli
eliciting the approach response.

Fructration is elicited by the termination of positive
reinforcement necessitated by the prey leaving the source
of reinforcement in orécr to avoid the predator’s attack. As
with fear, after a sufficient number of approaches to the
goal. fruetraticn-produced stimuli serve as partial diocrimw
inative stimuli eliciting the approach stimuli reoponoo.

Two predictiono were mace concerning resistance to
extinction. The preaiction was made that prey without prior
experience in approaching the reinforced region under threat

of preéntory attack would chow leoo reciotnnce to extinction

87

n b
80

than had they have such prior experience. Alec. it was
predicted, from the model, that prey allowed less time in
the goal region prior to the onset of predatory attack would
show more resistance to extinction of the approach response
compared to prey having more time in the goal region prior
to predatory attack. These predictions were supported by
the results.

The nodal makes use of apparently justified assumption
that under appropriate learning conditions. normally noxiouo
and cvereive stimuli can come to serve no stimuli helping
to increase resistance to extinction. Convincing evidence
for this assertion in to be had by noting thst in Experiments
II and III the groups which show the greatest resistance to
extinction are also the groups which receive the greatest
number of checks.

The results supporting the two hypotheses and the
other results which case to light during this study must
be generalized with great caution. Direct application of
the present findings to the field must be guite tenetive.
However, two conclusions seen telly warranted. First, this
study has set forth e means of eyetemstically studying a
couple: problem. The problem is to find a way to evaluate
the type and form of learning which a prey organism displays
as it learns to cope with its predators. It is felt that
this study represents a step in the right direction. The
basic took confronting the subjects in the present study

may bo limited in scope in terms of all possible confrontations
of prey and predator, but the situation is readily modifiable
along several dimensions. These dimensions are, for instance:
The special relationship of cafe region, goal, and intorvon mg
Space; the initial locus of the predator; the detectability
of the predator before and after the beginning of its
approach towards the prey orginioms the stimuli characteriz¢
ing the predator such as size, speed of a promch, typo of
stimuli signaling opproach of the predator: the possibility
of varying the type of and number of alternative paths
leading to and from the goal region; variation in the type
of reinforcement and deprivation conditions of the prey;
and of course. the type of prey organiom used.
The second point is that an attempt has been made to
formulate the problem. ibis formulation or theorizing gives
a structured framework for future research in the fluid.
Tho present study could serve to focus field efforts towards
the gathering of detailed information on the continuing
behavior of individual prey organisms. This study could
servo to equip the observer with a set of expectations on
the learning of prey organisms. Theoc expectations could
then be supported, modified, or dioconfirmed by careful
obacrvationo in the field. By thin neana a reciprocal ro-
lationohip between laboratory and field work might be developed.
Shear common sense would dictate that the conditions

confronting the prey orgauiom in the field are much less

c: O

stable than in the laboratory. Lvenoo, some tentative
suggestions can be put forth from the results of the present
'hyporotable' study. It in suggeotod that increased
resistance to extinction of opproach to a region previously
associated with reinforcement might prove to be adaptive

for the prey organism. This might be so in that the con-
ditions which make reinforcement available in the first place
may repeat themselves. Thio would make it unnecessary for
the organism to search new and wider arena for other sources
of reinforcement. This effect might help explain territor-
iality in orgenioms which have fairly demarooted foraging
ranges.

Also, increased resistance to extinction under conditions
of short delay in predatory attack compared to long delays
might prove adaptive. More returns to a previously reinforced
region could disclose other eourceo of reinforcement close
to the original region.

Also, through the conditioning of approach response to
fear-frustration stimuli it 16 possible that a prey organism
may be more resistant to abandoning new, potential sources
of food or water (reinforcement) if it is attacked in this
region while being reinforced.

These suggestions are put forth to show further avenues
of study in the area of modification of prey behavior through
the medium of predator encountera.

Finally, it is hoped that the present study will alert

91

other workers to the need to View behavior modification in
light of its adaptive significance for the organism display-
ing the behavior change. In the last analysis, learning
should be viewed as one way an organism deals with a hostile

environment.

nag-13 pjytwnyn?q
.~ '4! .‘f;..'..,.;. "v _.L,j

‘Amsel, A. Tkie role of frustrative nonreward in noncontinuous
reward situations. Pczchol. Bu11., 1J38, £2, 102-119.

.Amsel, A. Frustrative nonreward in partial reinforcement
and discrimination learning: Some recent history and a
ghgoreticel extension. Fsvchol. ReV., 1962, fig, 306-
_t_. *'*“**"‘“*'

Azrin, N.H. Punishment and recovery fluring fixed-ratio
performance. 1, E312.Anc1. I? er sv.. 1359, L, 301~335.

Azrin, 3.3. Effects of punishment intenaity during variable-

interval reinforcement. 1. exz. gag}, §g§g1.. 1:60. 2.
123-142.

Azrin, L.H. and H012, w.c. Punishment during fixed-interval
reinforcement. i. exp. Anal. EQhéng, 1961, 1, 345-347.

Brown, J.. flsrtiu, R.C.. and Morrow, Mu. colt-punitive
behavior in the rat: Facilitative effects of punishmezt

on resistance to extinction. i, 20mg. pgxeiol. £3 1chol..
1&6“. 21' 147’133.

Brown, R.T. and tagner, 1.3. Resistance to punishment and
extinction following training with shock or nonreinforce-

lent. Q, exg, Esxcho;.. 1964, gg, 503-507.

Calhoun, J.B. ‘ e ” ~ of the Forwey Rat.
weshingtonz rubl'o flea th service Pablication so. luofi,
1‘64“

   

Church. R.M. The varied effects of punishment on behavior.

22mm” 1963. 12. 369-462.

Curti. R.W. Native tear responses of white rats in the
presence of sets. gszchol. Monog.. 1955. 5Q, 78-98.

Denny, M.R. One bar-press per day: Acquisition and
extinction. g, exg. Ana . behev., 1959,Ԥ. 61-85.

 

Denny, H.R. and Adelmen, H.H. Elicitation Theory II: The
formal theory and its application to instrumental escape
and avoidance conéitioning. unpublished theoretical
paper, nichigen State University, 1956.

Denny, h.R. and Aeismrn, h.G. Avoidance behavior so a
function of lenrth of nonshock confinement. 1, come.
lb hgelol. Pszclol., 1964, fig, ESP-257.

. .
.
‘
. 6
‘,
I
.
I U
a 0

V

  

Edwards. AOL.
Row York:

      

Ritehart and 00., 1960.

Elton, 0.3. The use of cats in farm rat control. Brit. 1,
“Rig. B8h3Vo. 1955. 13 151-155.

 

Forster, 0.3. Withdrawal of positive reinforcement as punish-

Ferster, 0.8. Control of behavior in chimpanzees and
pigeons by time out from positive reinforcement. Pczchcl.
gram-2., 1958, 12; (8, whole to. 461.)

Griffith, C.R. The behavior of white rats in the presence
of cats. Psxchobiol., 1920, g, 19-28 (abstract)

Guinn, G.T. The effects of punishment on acts motivated
by fear. i2 exo. Psvohol., 1949, 23, Eco—£69.

fiedeger, 3. Wild Animals in Captivitz, London: Butterworth,
1950.

Hinds, R.A. Factors governing the changes in strength of
a partially inborn response, as shown by the mobbing
behavior of the chaffinch (Fringills coelebs) III. The
interaction of short and long term incrementsl and
decremental effects. £239. £21. Soc. §., 1961, 153,
3id“4£00

H012, W.C. and Arrin, H.H. Discriminative preperties of
punishment. i, cry. Anal. Behav., 1961, i, 2:5-232.

Eull, C.L., Learning: 11 The factor of the confiitioned
reflex. In C. Murchison (ed.). Handbook of genera;
Er eri ante szcgolggz. Worcester; Clerk Univer. rress,

      

Joslin, 5., Fletcher, H., and Emlen, J. A comparison of
the response to snakes of lob~uni wild-reared Rhesus
monkeys. Animal Bohev., 1964, lg, 346-356.

Keehn, J.D. The effect of a warning signal on unrestricted
avoidance behavior. Brit. 1, Psvchol., 1959, 22.
1¢5*1550

Melee k, K. On the survival of mallard ducks after
'hebituation' to the hawksheped figure. Behavior, 1961,
$1: 9’16.

34

Melvin, K.B. eni Brown, J.S. fieutrwllzation of on aversive
light stimulus as a function of number of paired
presentations with food. g, 0033. h-eiol.

 
     

Pevczcl.,

Earth, A.J. and Morton, m.L. Succeeeive acquisition and
extinctions of an instrumental response. i. 0333.
physiel. Psychol.. 1962. 55. fl74-377.

 

Pavlov, I.P. genditioned Reflexes. London: Oxford Univer~
airy Preee., 19:7.

Richerdeon, W.B. Reaction towards snakes as shown by the
Wood Rat (Restore elbignia) g, coma. Fovcho;.. 1942,

34, 1-10.

Sandler, J. Moeochiem: in empirical analysis. Psxcicl.
31111.; 1954, fig. 197-3304.

Seton, E.T. Lives 9! GameewAnimele, 3 vol. newton Center,
Masc.8 Chirlee T. bronford, 1353.

 

Sideen, M. and Boron, J.J. A comparison of two types of
warning stimulus in on avoidance situation. i, 2332.
ghycicl. Fojchg;., 1957, 2;, 282~287.

Simmons. K.E.L. Ehe nature of the predator~reectione of
waders towards humans; with Special reference to the
role of the aggressive, escape and brooding drives.
Bohgvior, 1955. Q, 150-173.

   
   

Spence, K.W. Behav or The and Learning. Englewood

Cliffs, N. .a 'rentice~ all, 1930.

Wynne-Edwarie, v.0. Animal Diepereicn in Relation to Social
Behovior. Eflinburgh: Oliver and Boyd, 1962.

 

. .Q.‘ ,V’r‘ 0. oh.
vu W Cu , a4” ‘ O“ .H O c

O 7D
b-
O

W

n ’n
O

J

 

 

 

 

 

a
.H . . “J. .. . u L. H. u. . c .5. a... ”no 7.. ..
4‘ - l , 4 ,A. II I. a ‘
v a a .. n . . . w ; u n o D." . o \ o u Q
1 - 1 1 . a . V g
. .
r J.“ g E31, 9. .2 .. X ,n . ..,J.f..;r
- q a - é..- . 3 v 1:593.» J. ‘ m.
i. p a V : {ng
n .. o .. , a..¢ 1 6‘ mt}? a in ‘
a»: .W .V. » . t f :1 A.‘ H1
4". \ I 0V .. \< . J x I u
a o o . ~ ( f. a . Mn . a 3%.. .4 - 3...... 2‘ r...“
v .. ..
L .l . w. .’ ,W
.. 4 1..
O . . . . O . ) ...
.y u a~ i. ,LD , 1 ‘ u 4 .ow .1 n . J
01”.; A I 05W: A (al.\ . \ .‘4 V c u‘ ‘4 . 44,4 ~ u. .\ (1 a u 4 .41.“ .jH _ .\2 .

M."

ed

 

 

Hog
H «no-«nomnm we nunuam

.33 on». no 330.2. ”on :33 033:... 1; u.- unouauuuou vangua- and

H ”Hagan.

once:

 

‘.u.\,

.bt’...‘..

S
. t . . I. :1 . .01 t. . C .3. a .. \
Q
0 a. 4
u... . .I 1 1.!
8 ¢ . I .
o . \ . ‘ T » A .. ‘ I. . . l 3!, ..
v
. o 9 '
Lu . 1 . .k‘ .
L D O . I 9
. 1 A . V ‘ l .. .ul \IO nan. A .
. n . 1
O
F 4 I l .
o o I o
I I I '

 

 

 

 

 

 

 

 

 

 

 

 

n .33» _
J ‘ J ._ q a e. . . m. .- .Kiwawl..rnfl+tl;fli
wlle... ... m . I4.m'.«fl.l.mi!i. ll..f.l.ii.mti. l. 3 in. H
.n mm: .3 .. NI'H¢.AI'B'~t' (aid at Eryiflatfl
. u . O4. , . .0 ad
dd 4% 2 Na d

«D!

. . .rflofiq «4. . .... ..
, . .lr. 1,0? o . q
04.1 m a

. .. L... ~32
N - 9..-.-m. wn.w._mnw_.w:.swflwsr .
m 3a .

 

Agliléi ..« w , . atlagbrfl~cfivfin .
. vag «ggnggvgv 5K...» Expat! Kufitafl. Ea: ,
OMKQJWI'AKImK an“?! KJfiimKvaitr E.. EHJLHKHL an
a a.“ n... c . o .1 v. . u we} .
mm m.” a mu .3 Wm fl

“I!
a w w m c m m. a %
Human."

HH 335$ no 3333 55% 3». Ho
nod-non “on n 28 4 33.8 «a seam." amino.“ 3», no uaowadatu 235$- 23 Bio!

HH gnu;

 

cl IL

,L H 5.

.r’ .

“. .\.

.l.‘

 

 

 

« ,. v .. .. Lu MU_. ”n... we Wu“. w...
Uop‘ maxim}: ., \Jifl": a. 2L. 4

 

m 4w m , ‘w‘ H amowmn

w @wfinm

 

 

 

 

 

 

n magma

 

 

 

 

 

 

a.
mu W M N H ”fiat—”mm,

 

 

 

 

 

 

 

 

 

 

 

 

JANA-Q V . .
mmdw w mm. meu MH.¢ Qw
.L x V. \F U. 11.3N0’Jl’. .. O H
w:.: a r. cc 4 mm m tw.n mack
uni \ “.1 )JU ‘ RN10 1. 1.”, M
\r.« ‘ ,;V H “I” . Np a «5;
w« c a Mud; .hmcr M .q :9
PER punt Jr. a“ ‘ .
H m m .H a much
\ 11
w m w mmufimmm

 

H manna

HHH aumawhegxm mo aewmna know an» no
mowmamu 969 Q vcm o mmfionu Ho moawu @mzoammn an» we acoaawfib¢v vhmcamum was «swam

HHH NHQWflmmd