THE QUANTIFICATION OF DRIVE H.
TWO METHODS OF FOOD PRIVATION

Thesis for the Degree of Ph. D,‘
MICHIGAN STATE COLLEGE

Orville Auverne Smii‘h, Jr.
1953

 

 

 

0-163

This In to certify that the
thesis entitled
The Quantification of Drive II.
Two Methods 01‘ 1'00? Privation

presented by

Orville A. Smith, Jr.

has been accepted towards fulfillment
of the requirements for

%&

 

Major @{ofemr

mew

 

 

 

V n........-.u.~.ouu-. ..... . .

 

 

 

THE QUANTIFICATION OF DRIVE II.
TWO METHODS OF FOOD PRIVATION

BY

Orville Auverne Smith, Jr.

A THESIS

Submitted to the School of Graduate Studies of Michigan
State College of Agriculture and Applied Science
in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Psychology
Year 1953

ACKNOWLEDGEMENT

Grateful acknowledgement is made to

Doctor M. Ray Denny for his advice and

assistance which made possible the

Asuccessful completion of this research.

INTRODUCTION . .

STATEMENT OF‘THE

SUBJECTS e e e e

APPARATUS. . e e

PROCEDURE. 0 e e

PREDICTIONS. e 0

RESULTS O O O O O

DISCUSSION e e 0

TABLE

PROBLEM e

SUMMARY AND CONCLUSIONS. 0

REFERENCES e e e

OF CONTENTS

Page

12

13

11L

18

27

66

85

88

LIST OF TABLES
Table Page

1 Analysis of variance of latency scores as a
function of hours deprivation. . . . . . . . . 31

2 "F" ratios of latency variances for all
% Fed SUbgPOUpSe e e e e e e e e e e e e e e e 32

3 Analysis of variance of amplitude scores
as a function of hours deprivation . . . . . . 36

h Analysis of variance of amplitude scores
- as a function of % Fed (LW subgroup) . . . . . 39

5 Analysis of variance of amplitude scores
as a function of % Fed (SW subgroup) . . . . . no

6 Analysis of variance of amplitude scores

as a function of % Fed (LD subgroup) . . . . . Al
7 Analysis of variance of activity scores as

a function of hours deprivation. . . . . . . . 43

8 Analysis of variance of activity scores as a
function of % Fed (Combined SW, LD and SD
SUbgPOUPS) e e e e e e e e e e e e e e e e e 0 us

9 Analysis of variance of activity scores as a
function of % Fed.(tW subgroup). . . . . . . . A6

10 Chi-square test of linearity for amplitude
data of hours deprivation group. . . . . . . . 50

ll Chi-square test of linearity for amplitude
data of LW subgroup. . . . . . . . . . . . . . 51

12 Chi-square test of linearity for activity
data of the SD, LD, and SW subgroups combined. 52

13 Correlation coefficient between latencies of
hours deprivation group and SD subgroup
equated as to activity with a fluctuation
Of 2 p01nt8. e e e e e e e e e e e e e e e e e S“

14 Correlation coefficient between latencies of
hours deprivation group and SD subgroup
equated as to activity with a fluctuation
Of 20 pOlntS e e e e e e e e e e e e e e e e e 55

LIST OF TABLES Continued
Table Page

15 Correlation coefficient between latencies
of hours deprivation group and SD subgroup
equated as to activity with a fluctuation
Of 5 pOlDUSe e e e e e e e e e e e e e e e e e e 57

16 Correlation coefficient between amplitude
scores of hours deprivation group and LD
subgroup equated as to activity with a
fluctuation of 5 points. . . . . . . . . . . . . 58

17 "F" test on variances of the number of
trials to extinction for high and low
drive SUbJOOtB Of % Fed GPOUPe e e e e e e e e e 60

18 't' test of the number of trials to
extinction of the LW subgroup vs. the s
SD subgroup. e e e e e e e e e e e e e e e e e e 61

19 Correlation coefficients between the
deprivation levels of both hours
deprivation group and % Fed group with
EMP11DUd6 SOOPOS e e e e e e e e e e e e e e e e 62

20 Standard deviations from comparable points
of amplitude scores of the hours
deprivation and % Fed groups . . . . . . . . . . 6h

21 Correlation coefficients utilizing
deprivation level to predict amplitude
compared to activity to predict amplitude. . . . 65

LIST OF FIGURES
Figure Page
1 Apparatus e e e e e e e e e e e e e e e e e e e 15

2 Mean activity scores for two groups of animals
on 13 consecutive days. . . . . . . . . . . . 28

3 Latency as a function of number of hours
fOOd deprivation. e e e e e e e e e e e e e e 29

h Latency as a function of deprivation level
with four different reward situations . . . . 33

5 Standard deviation as a function of depri-
vation level with four reward situations. . . 3h

6 AMplitude of response in degrees turn of a
wheel as a function of hours food
deprivation e e e e e e e e e e e e e e e e e 35

7 Amplitude of a response in degrees turn of
a wheel as a function of deprivation
level and three types of reward . . . . . . . 38

8 Activity level as a function of the number
or hours fOOd deprivation e e e e e e e e e e #2

9 Activity'as a function of deprivation
level and four types of reward. . . . . . . . an

10 Mean activity scores of three combined
groups (SW, LW, SD) as a function of
deprivation level e e e e e e e e e e e e e e AB

11 Activity, latency and amplitude as a
function of the number of hours food
deprivation e e e e e e e e e e e e e e e e e #9

INTRODUCTION

”Among the more elusive variables in the behavior of
the intact organism, drive stands preeminent.” This quo-
tation from Skinner (17) is probably an understatement,
principally because there remains the question of whether
drive is actually a primary behavioral variable or whether
the behavioral concomitants attributed to it may be deriv-
able in terms of other principles.

Behaviorists, in reaction to the instinct theories,
brought forth the term 'primary drive' with the connotation
of "greatly diminished reliance upon hypothetical innate
sensoriemotor connections . . .' (8). However, the behav-
iorists have had little success in agreeing as to Just what
drive is and.what may be attributed to it.

Referring to drive or motivation, Holt (9) has said,
"In short, all of the appetite drives are persistent afferent
impulses coming from organs situated within the body, and
producing . . . at the outset, that is, previously to trial
and error learning, merely random movements of general
restlessness.I This is the only way he uses the term
'drive' - which in essence is as an internal stimulus and
nothing more. Holt's usage does not in the least admit of
any 'directive preperties' of motivation current in present

day psychology.

{5

I.

r.

N

Guthrie (7) holds a similar viewpoint. Drive or
motivation for him is the action of 'maintaining stimuli'.
The maintaning stimuli may come from within, e.g.. hunger,
thirst, or may come from.without, e.g., electric shock,.
intense auditory stimuli. These stimuli 'drive' the
organimm.

Tolman.(l9) removes the concept from the equation with
raw stimulus and gives it status as a physiological 'state',
saying, ”The ultimate motivators of all behavior . . . are,
we assume, certain innately provided appetites and aversions.
These consist in ultimate demands to get to physiological
quiescences or from final physiological disturbance.” The
demands in this case are 'internal initiating physiological
states' which emerge as one of his four independent causes
of behavior. This conceptualization obviously retains
instinctual and teleological connotations.

Skinner's system (17) contains a flat rejection of
drive as a stimulus or as having any teleological impli-
cations. He says, "In.measuring the strength of a drive
we are in reality only measuring strength of behavior. A
complete account of the latter is to be obtained from an
examination of the Operations that are found to affect it.
The 'drive' is a hypothetical state interpolated between
Operation.and behavior and is not actually required in a
descriptive system.” And elsewhere he states, "The con-

ception of drive as a state rather than as a stimulus is

valuable in avoiding arguments about purpose." In these
writings are found one of the first formal considerations
of 'drive' as an unobservable theoretical construct.

Keller and Schoenfeld (13) more recently have again
supported the Skinnerian outlook by saying that drive is
‘322 stimulus, 222 a reaponse, and 222 a physiological concept.
For them drive is the name for the fact that certain oper-
ations can be performed on an organism.that have an effect
upon behavior which is different from other operations.

In the above, drive has been reduced to a set of
operations which have an effect on behavior. This treatment
gives drive the essential requisites of a theoretical con-
struct. However, even these conceptualizations have not
resulted in the strict theoretical formulation employed by
Bull, whose theory is founded on the concepts of drive mud
drive reduction. Hull's formulation is the most explicit
available at the present time and, consequently, the most

testable.
The Hullian Concept of Drive

Hull (10, 11) has attempted to formalize drive by
naming OXplloit the experimentally manipulated antecedent
conditions and the associated observable behavioral conse-
quences. In order for a concept to obtain status as a

construct, it must meet several requirements.

/V

The requirements for a construct as put forth by
iHaatsch and Behan (1h) are:

1. Operational definition which contains “a description
of manipulations which are performed in the labor-
atory, and a description of the effects of these
manipulations upon the behavior of the experimental
subjects."

2. The concept "must bear explicitly stated relation-
ships to other constructs in the theory of which
it is a part."

3. It must "vary unidimensionally and continuously,
and must affect at least one abstracted aspect of
behavior such that the behavior will vary unidimen-
sionally and continuously."

Hull (10) operationally defines drive (D) by indicating
the manipulations in the eXperimental situation (antecedent
conditions) which is the number of hours elapsed since the
last satiation period for the particular primary drive in
question. For the hunger drive, the conditions would be
the number of hours since the last food intake. The conse-
quent event or behavioral concomitant is "the amount of
energy which will be expended by the organism." If energy
is expended by an organism and we ignore the magnitude of
energy expenditure as a result of neural activity alone
without concomitant muscular activity; or that involved in

‘maintaining muscle tonus, then the energy transformation

from potential into kinetic will result chiefly from gross
movements of the organism's musculature. In other words,
the organism will be more active, 1.6., the random move-
ments of the organism will increase.

We assume from this that with an increase in the number
of hours of deprivation there would occur an increase in
the activity of the organism. Hull (10) uses the now
classical work of Richter (l6),‘Wada (20), and others to
illustrate Just this point.

As to the second requirement, Hull (11) relates D to

the other constructs in his system by the formula

sEr: erXV XJxKXD
where sEr is reaction potential, er is habit strength,
V is stimulus intensity dynamism, J is the delay in rein-
forcement and K is the amount of reward. sEr or performance
is measured in terms of latency of response, amplitude of
response, number of trials to extinction and the probability
of reaction evocation.

Regarding the 3rd requirement, Hull's theory must say
that with all other factors held constant, any variation in
D will result in variation of performance (sEr) as measired
by above methods, and the variation in performance will
occur for 21351 variation in drive.

According to Hull (10), ". . . when a condition arises
for which action on the part of the organism is a prerequisite
to Optimum probability of survival of either the individual

or the species, a state of need is said to exist." A need,
then, produces a primary drive and every individual drive
summates physiologically with every other drive to produce
3 generalized M 9.39.152. in the organism. It is this

 

generalized drive which is formulated as the construct (D).
The Scientific Usefulness of Drive (D)

The problems involved in using this concept in a strict
mathematical theory of behavior are not overlooked by Bull.
He recognizes that there must be two equations: one express-
ing the degree of drive as a function of the amount of
deprivation, another expressing the vigor of organismic
action as a function of the degree of drive. One of the
major problems is that of finding a measurement parameter
with a specified extensive unit with which to express the
degree of drive on the behavioral side. If (D) in conjunc-
tion with the other factors V, K, J, and er, bears a

constant relationship to reaction potential regardless g;

 

£23 dominant typg'gf privation and if this fundamental
relationship can be quantitatively determined, then Drive
may be used as a genuine theoreticalconstruct. If, on

the other hand, a set of equations must be derived for each
primary drive constituting the major component of general-
ized Drive, there will be no reason to consider generalized
Drive as a useful scientific construct. Inasmuch as Drive
(D) is one of the most fundamental concepts in Hull's theory
validating the concept is essential..

 

”-1

The Problem of Quantification

A curve representing drive as a function of hours of
deprivation may be derived easily by using latency, ampli-
tude, or trials to extinction as'Yamaguchi (21) has done.
However, as Hull (10) points out, this ignores the whole
problem of getting at the intervening variable of Drive
itself. Hull specifically mentions the necessity of defin-
ing a unit independent of the performance measurements with
which to represent Drive. Inasmuch as the activity level
of the organism is assumed to be directly associated with
the drive level of the animal and yet is not considered to
be a function of habit, it may be that the answer to the
problem can be found in the measurement of activity. It
seems reasonable that if Drive determines the energy output
of the organism that possibly the activity level of the
organism.may be taken as the measurement parameter and
provide the unit of measurement for Drive. In other words,
Drive may be quantifiable in terms of activity level. If
activity level is to serve as the unit of measurement, then
a situation.must be devised in which amount of deprivation
may be functionally related to activity level, and activity
level may be functionally related to latency, amplitude and
number of trials to extinction.

With this situation the experimenter could manipulate

drive conditions in terms of type of deprivation as well as

m]

OB

amount. If activity level can be used as a measure of (D)
then regardless of type of deprivation, activity levels

could be equated and the comparisons for the two conditions

of drive should result in very similar measurements of
latency, amplitude and trialstn extinction, 6.3., a corre-
lation coefficient between latencies obtained from different
animals under different types of deprivation but equated as
to activity level should prove to be very highly significant.
The use of a correlation coefficient is a much weaker require-
ment than demanding equal latencies for equal activity levels,
but it may be that different types of deprivation may

result in different absolute magnitudes of response and the
use of the correlation technique would allow for this differ-
ence. Again, then, if different types of privation. or
different methods of inducing drive with but one type of
privation, result in different absolute values of sEr,

still the shapes of the curves should be the same and a
factor should be ascertainable which would enable one to
convert the scores of the first condition into the scores

of the other condition.
Program of Validation

A complete investigation into the possibility of quan-
tifying drive in terms of activity level would involve two
major investigations:

1. One investigation would have to be made involving

I-

I.

a cross comparison of different types of deprivation. For
example, a group of animals trained under thirst deprivation
would be compared to a group under hunger conditions.
Activity levels for the two groups would be equated and
latency and amplitude would be compared. This would serve
to indicate whether or not activity level could be used as

a measurement parameter for various 21333 of deprivation.

2. The second investigation would involve using a
different method of setting up drive conditions but with the
scum type of deprivation, In other words a different set
of Operations which should result in the same physiological
effect for the animal should be performed. The consequences
based upon equating of activity levels could then be compared
to the results derived from the usual 'hours deprivation'

method.
Problems Involved in Carrying Out Program

Before this program could be carried out there would
necessarily be a great deal of preliminary investigation not
only in the line of practical difficulties but also regarding
the rationale for size of reward in the water and food
cross comparison (1 above).

Specifically, what size food pellet equals what size
water reward as an equal reinforcing value? This work has
been carried out by Davis (A). The actual work on problem 1

remains to be completed. The experiment to be reported.here

it)

is concerned with the second investigation mentioned.

The alternative method of setting up drive with the
sump type deprivation (in this case, food) was chosen with
the object not only of finding a method which would yield
a change in physiological state similar changes brought
about in the hours deprivation.method, but also to determine
whether or not a different method might yield more homo-
geneous results, or, better said, yield more precise results.

The {percentage fed' method was chosen for this purpose.
With this method the amount of food normally consumed by the
animal which results in satiation is measured over a period
of time. When the day to day amount consumed becomes
asymptotic the average amount on the asymptote-days is
taken as the basal amount to satiation. In varying the
drive level for the various organisms, a specified percentage
of the basal amount is fed to the animal a determined time
before testing.

This method should result in a differential physiological
state for the organism just as the hours deprivation method
does. ‘We assume that this method will consequently alter
the drive of the organism in a uniform fashion.

This specific technique has to the writer's knowledge,
never been used, but it is similar to methods used by
Skinner (l7) and Bruce (2). Skinner fed his animals 0, 2,

h, or 6 grams of food before measuring the rate of response

in the bar-pushing apparatus. The resulting curves, however,

l1

,2...»
H

were not clear out; there was considerable overlap, but
the general result was more rapid responding with the
lesser amounts Of food consumed. The results of Bruce (2)
seemed to show an effect on performance brought about by
prefeeding or prewatering immediately before testing. This
may possibly be avoided by increasing the time between
feeding and testing.

A complete study of the major drive variables could
be made in this situation by altering reward size and reward

type as well as deprivation level.

fl‘ ‘lh
I
-._ 6.!

STATEMENT OF THE PROBLEM

The problem here undertaken is to determine whether
or not activity level may be used as a means of quantifying
the Hullian construct of Drive (D) when Drive is induced
by two different methods of food deprivation. The study
will also include variations in size and type of reward
in order to gain information as to possible interaction
effects between Drive (D) and the variable incentive

reinforcement (K).

i 3

SUBJECTS

The subjects used in the present experiment were h8
experimentally naive male albino rats from the colony
maintained by the Department of Psychology of Michigan
State College. The ages ranged from 90 to 120 days.

{as
:4:-

Apparatus

In order to investigate the present problem an
apparatus was constructed in which it was possible to
measure activity level, response latency, response ampli-
tude or force, and the number of responses to extinction.
It consisted of a rectangular box constructed of %" plywood
with overall dimensions of 20" x 16" x 11".

Figure 1 presents a cross section of the apparatus.
At the bottom of the activity chamber there was a false
floor which was supported by 3 springs and a rubber ball

at its exact center. At the four corners of this false
floor, small attached, rubber balls served as stOps, pre-
venting the floor from tipping any more than &' in any
direction.

A guillotine door at one end of the activity chamber,
when raised, gave access to a hinged, h" x 2" panel. This
panel was constructed of a thin rectangular piece Of wood,
1/16" diameter; at the upper end of the panel was a small
piece of %" plywood 2%" long, which formed a base for the
hinge and brass strips (See Figures 1, 2, and h).

The flat gray interior was illuminated by a 7-watt
bulb covered by a piece of Opal-flashed glass situated at
the end of the box Opposite the guillotine door.

Entrance to the box from the tOp was gained through a
hinged door, in the center of which was placed a large clear-

glass observation window.

15

338a

H8...
soars

owe» ma confluence usueaeon; .a easwam

 

 

 

 
 

   

 

 

 

Hench

nee-5
3 condo;

flee—5
caved—E

ewes.—

 

 

 

 

 

 

 

 

 

 

Saved—".30

 

 

s\\\\\\\\\\x\~\§\\\\\\\\\§\K\\\\N\\\ \\.\\\\\\\\\\\\\\N\\\\\\\\\\\\\\\\\\\\\.
m sea?» bio!
I “ one?" no

_$S55§§SS\\\\\\\\§\\\§\“\\ gamma» — \\\\\\\\\\\\.m
“ OOH-h

....... \ “

a

a

w m

r". m

\

A .1. m

“ w

\

\ m
\ “ m .33
\ a 583......

ll.“

16

Activity level was measured by a device consisting of
a GE 2-36KRL mercury switch, suspended vertically beneath
the false floor and connected in series to a Gorrell and
Gorrell 115 volt electric counter. The mercury switch was
situated beneath the floor in such a manner that movement
by the animal tilted the floor, causing the liquid in the
tube to move, momentarily making and breaking the circuit
in accordance with the strength and number of movements
(See Figure 1).

A thin metal rod, hinged at the tOp of the panel was
twisted so as to extend to the back of the panel in one
direction and to the tOp of the box in the other. The rod
was so designed that the lower half of it "rode" back on
the panel as it was pushed Open, and the upper half came
.forward toward the activity chamber. By means of this rod,
the force applied to the door was transmitted to a slender
stick of wood which was attached to a light, plastic wheel,
mounted on a plastic axle. The force of the response, which
was applied to the door, was thus transmitted into the
movement of the wheel. The distance which this wheel moved
was measured in degrees. Because the wheel Offered very
little resistance to the metal rod, the initial movement of
the push-panel invariable caused it to turn out of range of
further movements of the metal rod. That is, the degrees
which the wheel was displaced depended upon the £9522 with
which the door was figgt struck, and not merely upon the

distance through which the door was moved.

.-

I»)

Response latency was measured by a Standard Electric
Timer, connected in series through two switches. The first
switch consisted of two brass strips, one being placed
along the tOp Of the push-panel, and the other being attached
to the upper push-panel jamb, directly above the strip on
the push-panel. A second switch was attached to the tOp
of the guillotine door. When the guillotine door was raised,
this switch was closed. Thus, when the push-panel was
closed and the guillotine door Opened, the circuit was
closed, starting the Timer. When the panel was Opened by
approximately 1/32", the switch was Opened, stOpping the
Timer. Thus the timer started.when the guillotine door
was raised and stepped as the panel was being pushed Open
by an animal.

Single reward pellets were placed on a tray which was
located approximately fi” below the lower panel jamb. Metal
walls were built up on either side of this tray to dis—
courage exploratory behavior. The corners of these walls
were bent towards the door, forming stOps to prevent the
animals from forcing the door and breaking it. This also
reduced exploratory behavior. On the tray itself a small
wall of solder was constructed to hold the food pellets in

place.

PROCEDURE

The portion of the procedure which was common to all
groups will be presented first, and the differences among

the groups will be specified subsequently.
Habituation

All animals were handled for a period of 20 minutes
on each of three days previous to being placed in the
apparatus. The handling consisted of stroking the animals,
picking them up, setting them down and allowing them to
run freely on a table tOp on which several boxes were
placed. On the fourth day the S's were placed in the appar-
atus for a period of six.minutes during which.time activity
level was recorded. These six minute habituation trials in
the apparatus continued for five days or up to the 9th day

of the experimental series.
Training

On the 9th day the_animals were given the regular six
minute habituation trial and at the end of this trial the
guillotine door was raised, allowing the animal access to
the hinged panel. The panel door was Open to its fullest
extent (approximately two inches) making the reward pellet
on the food receptable easily accessible. After the animal

had taken the reward the panel door was slowly closed and
the guillotine door was carefully lowered. As soon as the
animals had finished eating, indicated by cessation of
chewing movements, the guillotine door was raised again,
presenting the S with the Opened panel and the reward
pellet on the receptacle. This procedure continued for
8 trials. On the 9th trial the panel was Opened only about
} inch or about i'of the full Open position. This forced
the animal to nose the panel slightly in order to obtain
the pellet. Another trial was given in the same manner,
making a total of ten reinforced responses for the day.
On the next day the animals received two trials with the
door in the full Open position and eight trials with the
door in the % Open position. The following day the panel
was presented in the % Open position for two trials and was
fully closed for the remaining eight trials. The animals,
then, received 10 trials per day with the door fully closed
for a period of five more days. At the end of training
each animal had received 80 rewarded trials in the apparatus.

The six minute activity period.before the first trial
of each day was recorded.

Latency and amplitude measures for each trial were

recorded on the last three days of training.
Testing

On the 17th day of the series the test trials began.

m
C:

The animals were randomly assigned to the particular
deprivation level under which they would be tested on that
day. The apprOpriate deprivation manipulations were carried
out (details to be specified below) and the animals were
introduced into the apparatus. The six minute activity
levels were recorded and the Sis were given four rewarded
trials in exactly the same manner as during the last five
training days. Latency and amplitude were recorded for
each Of the four test trials. If an S refused to respond
for a period of six minutes it was removed from the apparatus
and a 'NO Response' was recorded. Each animal was tested

at each deprivation level under consideration, however the
order Of levels was systematically randomized for each S.
There were six deprivation levels studied; hence, there

were six testing days, bringing the experimental series to

a total Of 22 days.
Extinction

On the 23rd day Of the series the animals were ran-
domly assigned to a deprivation level. They were again
given the six minute activity level period in the apparatus
and the panel pushing response was then extinguished to a
three minute no response criterion. The number of responses
to criterion, latency and amplitude were recorded.

A summary Of the procedure common to all groups is as

follows:

Days 1 - 3 Handling - 20 minutes per day.

Days A - 8 Habituation in apparatus - 6 minutes -
activity recorded.

Days 9 - 16 Training - 10 trials per day - activity
recorded on all days, latency and amplitude
recorded on days lh, 15, and 16 only.

Days 16 - 22 Testing - h trials per day under appro-
priate drive condition - activity recorded -
latency, amplitude measured on each trial.

Day 23 Extinction - to a three minute no response
criterion - activity recorded - latency and
amplitude recorded on all trials.

All animals were weighed at the beginning and end Of
the 23 day series and at least one other time during the
series. '

The aspects of the procedure which differentiated the

groups were as follows:

Hours Deprivation Group - These animals were fed ad lLE until
day 8 Of the series when they were put on a 2k hour depriva-
tion schedule. They continued on this schedule during the
training days 9-16. 22% hours after their last feeding

the animals were placed in the apparatus for the activity
level period and the 10 training trials. The reward pellet
was approximately 0.08 grams Of dry Calf Manna. After the
conclusion of the trials the animals were placed in a feeding
cage for a period of at least six minutes, after which time
they received the same type pellets used as reward pellets
for a period Of 15 minutes. They then received a wet mash
composed Of a ho% Purina dog chow and 60% water combination

for a period Of 25 minutes. After this time they were

removed tO the home cages until the following day.

The procedure during the testing days involved feeding
the animals with the above mentioned satiation procedure a
specified number Of hours before they were to be run. The
periods were 1, 2, 6, 12, 2h and 48 hours measured from
the time the animals were removed from the feeding cage and
replaced in the home cage. ‘Water was available in both

‘home and feeding cages.

Large'Wet Mash, % Fed Group - hereafter to be designated.Lw.
The S's in this category were handled in the same
manner as those above with the exceptions that:
1. They were fed wet mash only.
2. The amount of mash consumed by each animal was
recorded on the basis of four hours eating time
the first day, l B/h hours the second day, 1 hour
the third day, i hour the remaining days.

3. They were on a 2h hour deprivation schedule from
the beginning of the habituation days (Day h).

h. During testing they were fed 0, 10, 25, SO, 75 or
90 percent of the average amount Of mash they had
eaten on the four days previous to the test days.

5. All animals were run 3 hours after they had consumed
the above mentioned percentage.

6. The reward pellet was 0.2 gm of wet mash.
7. After testing they were fed enough mash, after a
six minute delay in the feeding cage, to equal

100% Of the previously determined amount consumed
to produce satiation.

Small Dry Pellet, % Fed Group - hereafter to be designated SD.

These S's were handled as the Lw group with the following

p
‘.—v—-.-—_—_-—.A b—v—_-~‘-’—--—‘————n
' v
i
U
C
0‘.
.
~ I
'.
___._._..- _ ._.__- m- ._.__i

exceptions:

N)
C»:

1. They were fed 10 minutes on dry pellets (Calf
Manna) and 30 minutes on wet mash: the total
volume Of wet food consumed was computed each day.

2. Reward pellet was 0.08 gm of dry pellet.

3. Fed only mash before running - fed pellets and mash
after testing to equal 100% of satiation amount.

Large Dry Pellet, % Fed Group - hereafter to be designated.LD.

The only difference between this group and the SD group

was the reward pellet size, which was 0.2 gm for this group.

Small Wet Mash, % Fed Group - hereafter to be designated SW.

This group differed from the LW group regarding size Of

reward which was 0.08 gm Of wet mash and time fed during

training trials, which was hO minutes.

A summary Of the critical differences among groups is

as follows:

Hours Deprivation
N’= ll

LW-%Fed
N=1L|.

sw-%Fed
N’= 6

LD-%Fed
N=7

SD -

"ha
'121
o
n.

S's satiated a specified number Of hours
before running. Reward was 0.08 gm dry
Calf Manna.

S's run 3 hours after receiving specified
percent; Of amount required to produce
satiation. Reward was 0.2 gm wet mash.

S's run 3 hours after receiving Specified
percent of amount required to produce
satiation. Reward was 0.08 gm wet mash.

S's run 3 hours after receiving Specified
percent Of amount required to produce
satiation. Reward was 0.2 gm dry Calf
Manna.

S's run 3 hours after receiving specified
percent Of amount required to produce
satiation. Reward was 0.08 gm dry calf
Manna.

PREDICTIONS

In order to facilitate the presentation Of the results
most Of the specific predictions to be tested by the present
study are given below. Predictions 1 through 12 are
deductions from Hullian theory and the assumptions seemingly
implicit in his system with particular regard to the
assumption that activity level will measure Drive (D)
directly.

1. From the fourth to the ninth day Of the series
(corresponding to the first 5 days on Fig. 2) the average
activity of all animals will gradually decrease because Of
adaptation or what may be termed a decrease in 'eXploratory
drive'.

2. The decrease in activity level will be greater
for the animals satiated during this period than for those
under 22% hours or food deprivation.

3. When the satiated animals are put under 22% hours
deprivation activity level will rise abruptly to the level
Of the regularly deprived group and remain constant at that
level as long as the deprivation schedule is maintained.

h. There will be differences between deprivation
levels regarding activity and amplitude which will vary
positively with deprivation level and latency which will

vary negatively with deprivation level.

5. The function for amplitude will be a straight line.

6. The curve for reciprocals Of latency will be
negatively accelerated.

7. The curve for activity should approximate Yama-
guchi's (21) curve which represents Drive (D) as a function
Of hours deprivation. Yamaguchi's curve is described as,
". . . from h = 0 to about 3 hours drive rises in an
approximately linear manner until the function abruptly
shifts to a near horizontal, then to a concave-upward
course, gradually changing to a convex-upward curve reaching
a maximum Of 12.30- at about h = 59, after which it grad-
ually falls to the reaction threshold (er) at around
h.= 100."

8. If animals from the Hours Deprivation Group are
matched with animals from the % Fed Group on the basis of
activity level, the correlation coefficient between their
corresponding latencies and amplitudes Of response will be
highly significant and positive.

9. Groups receiving large reward will perform better
than those receiving small reward, i.e., lower latencies,
higher amplitudes.

10. There will be no significant differences in
activity level as a function Of size or type of reward.

11. There will be more trials to extinction for animals

{\T'
(3')

extinguiShed under higher deprivation conditions than those
extinguished under low deprivation conditions.

12. There will be differences between groups in
extinction based upon size Of reward, i.e., the larger the
reward, the more trials to extinction.

13. If 9% Fed method delimits level of deprivation
better than the number Of hours since satiation, then a
higher correlation.will result from deprivation levels Of
the % Fed Group with latency and amplitude than with levels
Of the Hours Deprivation Group and latency and amplitude.

1h. Also, if % Fed method delimits level Of depri-
vation better than the number of hours since satiation,
the standard deviation at each level for % Fed animals will
be less than for approximately corresponding points for the
Hours Deprivation Group.

15. If activity measures Drive more reliably than
hours deprivation, then there should be a higher corre-
lation coefficient between activity and latency, and
activity and amplitude than for hours deprivation corre-

lated with latency and amplitude.

RESULTS

Figure 2 shows fluctuations in activity level Of two
groups of animals measured over a six minute period fOr
thirteen consecutive days. The activity levels of the
animals decreased during the first five days in the box
(habituation period) as predicted. Prediction 1 was con-
firmed.

The animals in the Deprived Group were under 22% hour
hunger conditions while the animals in the Satiated Group
had food available at all times. Despite this difference
in the groups there were no differences in activity level.
Inspection reveals complete overlap Of the two curves.
Prediction 2 was not confirmed. Upon putting animals of
the Satiated Group under 22% hours food deprivation on the
sixth day, there was no sudden increase in activity level,
and no separation Of the two curves. The increase in
activity level was gradual and was shown by both groups of
animals from the sixth to the thirteenth days, 1.6., during
the training period in which time food was presented to
the animals via panel pushing just subsequent to the six
minute activity period. Prediction 3 was not confirmed.

Latency as a function Of deprivation level in the
Hours Deprivation Group is shown in Fig. 3. Inasmuch as

the variances for these animals proved to be homogeneous

whee cipfioomdoo ma do manage mo
9 N, 2 O, , O
omZmdwo \\\ \\\\
owvffim \\
I
I
I
I ~
I
I ~71! K
II s IL
. a
s
I s
a a
z ~

~-
-u---‘

I
(-
--
~-
~-
----
-e-
~--
Q-..

CO,

OON

con

soapmbasoop poo.“ mason .HO pong: .HO £0323.“ m we hodopwq .m enswam
20752530 930:
fiuu N. m N.
a _ q _ _

 

 

Q"

'5335 NI

ID

AONBiV'l/Ol

an analysis Of variance was performed on the data, which
is given in Table 1. From the insignificant "F" it is
Obvious that the differences in means of the various levels
could have occurred by chance. Figure h shows latency as
a function Of deprivation for the % Fed Groups, representing
the four curves on a single graph to indicate the differ-
ences due to size and type of reward. Table 2 presents
the results of ”F" tests performed on the variances of the
% Fed Group to determine whether or not the homogeneity of
variance requirement was fulfilled. The "F" test indicates
that the variances within the groups; that is, between drive
levels, are so divergent as to preclude the possibility Of
performing any valid analysis of variance or “t” test. An
attempt to normalize the data by eliminating all scores
beyond Liq- of the mean failed to make the 'F's' non-signifi-
cant.

Regarding latency data prediction number n is not con-
firmed by the results Of the Hours Deprivation Group and
no statement can be made concerning the results Of the % Fed
Group.

Figure 5 shows the changes in standard deviations Of
latencies with changes in deprivation level.

Figure 6 presents the changes in amplitude Of response
as a result Of change in deprivation for the Hours Depriva-

tion Group. Table 3 presents the analysis Of variance

3 1

TABLE 1

ANALYSIS OF VARIANCE OF LATENCY SCORES AS A
FUNCTION OF HOURS DEPRIVATION

 

 

 

 

Source Of variation Sum of squares d.f. Mean square
Levels 137.85 5 27.5?
Subjects 172.10 10 17.21
Interaction 606.hS uh. 13.785

No significant differences

 

TABLE 2
"F" RATIOS OF LATENCY VARIANCES FOR ALL 96 FED SUBGROUPS

 

 

 

 

 

 

 

0 10 25 50 75 90
m 0.2 .0h2 .030 .096 .067 .259 .267
25 - + ‘+
75 -
90
LD 0.2 .096 .255 .109 .21.? 1.795 10.038
0 - - - + +
10 - - + +
25 - + +
50 + +
75 +
90
SD 0.?- .056 .188 .118 .373 1.932 5.76
0 + - + + +
10 .. . .. + +
25 + + +
75 +
90
sw 0.2 .0001; .0021 .0096 .011 .036 .616
0 + + - + +
10 - + + +
25 - - +
50 II +
75 +
90

 

+ indicates significance at 5% level or less.

msogmdpflm panama pdmaowgo .50.“ £33 HpbOH doflpmuflpmop mo mogoqdm m we hummus 3: 09mg

 

 

om... .\.
o 0. mm on on cm
T w J 1 d
\h h
J‘

 

Anna #03 Hamam Tlll

mama no: omnmq Ill... . III

erased use Hausa IIIII
wedded mac omamH

   
 

.nu

 

nfim“

m.N

.n

SONOOBS

 

 

 

page .3 men»... asoa pom amped deapmaaaaeo mo aoapoasa m we 32$me neapmaaop campampm .m Oadwah
on... ex.
0 0. mm on ma om
T W at q . q
‘ ‘\
h, e e I e O I O I e l e
’I’ .II‘ ‘ One 1 I I, e I,
I‘ ‘e I'll I e l e no
' I. ’ I l' e I .
II
I
I
’
I
I
10..
I
III
I .
I i m _
I
I
I
I
// .1 O.N
I
I
I
,i
mesa no: :55 T1 “N
name no: emamH Illil
HOHHOQ mac Hanam Illl
wedded haw OmadH CM
.0.»

 

VWOIS

Goapmbaadop
mm poem mason a0 soapoada m as Hoes: a mo dawn meoawop Ga Omaommoa no opspaaa5< .o mammam
zo_._.<>_mdwo mmaoz
mv «N N. m N.

«N

1
“D
N
$338930

1
00
(V

 

3%;

TABLE 3

ANALYSIS OF VARIANCE OF AMPLITUDE SCORES AS A
FUNCTION OF HOURS DEPRIVATION

 

 

Source Of variation Sum of squares d.f. Mean square
Levels 612.512 5 122.50**
Subjects 737.199 10 73.72
Interaction 606.705 h3 lu.109

 

**Significant at 1% level Of confidence

>

.o—.

C53)

performed on these data. The results support the respect-
ive part Of prediction h, namely that amplitude varies
directly with increase in deprivation level.

In Fig. 7 the changes in amplitude as a function Of
deprivation level are given for the LW, SW, and.LD sub-
groups Of the % Fed Group. Apparatus failure prevented
Obtaining accurate measurement for the entire SD Group,
hence the dataare not included. Also there was a slight
difference in the setting Of the amplitude measuring
instrument for the LW Group, making comparisons with this
group somewhat suspect. However, a mean curve for the
three groups was determined and plotted. This mean curve
is also represented in Fig. 7 and seems to lend support to
prediction 5 concerning the linearity Of the amplitude
function.

Tables h, 5, and 6 present the results Of analyses Of
variance for the LW, SW and LD Groups respectively on the
amplitude scores. Using the % Fed method prediction h is
again supported.

The increase in activity with increase in deprivation
for the Hours Deprivation Group is represented in Fig. 8
and the analysis Of variance indicating significant differ-
ences is found in Table 7. The related graph for the % Fed
Groups is Fig. 9 and the analyses Of variance are shown in
Tables 8 and 9. It was found that the Lw Group differed

significantly from the other three groups regarding both

vamzea no woman mean» one Hoboa soapsbaaaop
.ao moaned?“ a we Hoonz e no Gaza moeawov a.“ ounOdmoa m HO ovfipaaaad .w sad—Mam

38

own #

O. D N 00 m# 00
n d “N

 

10
d
.J

[on

 

at! . \\\

 

 

Low

3..

 

N801 $338930

3- 9

TABLE h

ANALYSIS OF VARIANCE OF AMPLITUDE SCORES AS A
FUNCTION OF % FED (Lw SUBGROUP)

 

 

Source Of variation Sum Of squares d.f. Mean square
Levels 2,713 5 5h2.6
Subjects 8,587 13 660.5
Interaction 11,300 65 173.85

 

F = 3.12 - significant at 3% level of confidence

I.

TABLE 5

ANALYSIS OF VARIANCE OF AMPLITUDE SCORES AS A
FUNCTION OF % FED (Sw SUBGROUP)

 

r
.—.-:

 

Source Of variation Sum Of squares d.f. Mean square
Levels 2226.2 5 hh5.2h**
Subjects 2373.99 5 h7ho79
Interaction 15h8.35 25 61.93

 

**Significant at 1% level of confidence

4.- 3.

TABLE 6

ANALYSIS OF VARIANCE 0F AMPLITUDE SCORES AS A
FUNCTION OF % FED (LD SUBGROUP)

 

 

—
-

 

Source of variation Sum of squares d.f. , Mean square
Levels 3h5.7 5 69.1h
Subjects 2hhl.97 6 uOé.99
Interaction 98h.67 30 32.82

 

F = 2.11 - less than 10% level of confidence

42-

noapmaaaaop pooa mason mo areas: on» no doaponsa a me HO>oH haa>apo<

zo.._.<>_mn_wo manor

Axum
_

 

.m oasmam

N .

. ,
is.

 

OON

loom

CYV

100'

A ILAIJ.

Icon

1000

 

43

TABLE 7

ANALYSIS OF VARIANCE OF ACTIVITY SCORES AS A
FUNCTION OF HOURS DEPRIVATION

 

 

 

Source Of variation Sum Of squares d.f. Mean square
Levels 1,096,9h9 5 219.390**
Subjects 68h,369 10 68,h37
Interaction 912,536 50 -18,251

 

**Significant at 1% level of confidence

.pamama a0 momma asoa new amped seapmaaammp mo qoapodsm m we mpabapo< .m madmam

 

 

 

out. fl
6. nu on 2. on
a a A d n
nhN
can
nmv
l 0.00
1 as. n
venom haw omamn I
9032” mam Hiram IIIII
name no: HHSMI.‘ L000
gems no? omaQH

 

 

AilAllOV

TABLE 8

ANALYSIS OF VARIANCE OF ACTIVITY SCORES AS A
FUNCTION OF % FED (COMBINED SW, LD AND
SD SUBGROUPS)

 

 

 

Source Of variation Sum Of squares d.f. Mean square
%%
Levels 395.666 5 79.133
Subjects 1,302,h62 22 59,203
Interaction 927,908 110 8,h36

 

**
Significant at 1% level Of confidence

 

,. v» -— -. _. _ 7--- n .,,-,. a~ -.— . -- ~-.7- _.- _--~_‘._7. -_—-.——.—-

46

TABLE 9

ANALYSIS OF VARIANCE 0F ACTIVITY SCORES As A
FUNCTION OF % FED (LW SUBGROUP)

 

 

 

Source of variation Sum of squares d.f. Mean square
Levels 720,u99 5 lhh,100**
Subjects 2.258.717 13 173.7u7
Interaction 1,257,629 65 19,3h8

 

**Significant at 1% level of confidence

,fl

11

means and variances so that a comparison with the other
three groups could not be made; however, the LD, SD, and
SW subgroups were found to have comparable means and
variances, hence were combined into a single group (Table
8). Figure 10 shows the mean curve for these three sub-
groups. All the results using the activity measure clearly
support prediction h.

Curves representing amplitude, latency, and activity
as functions of hours of deprivation are combined on Fig. 11
tO indicate the similarity Of shapes, particularly regarding
the drOp in all curves at the two-hour level.

The Chi square test for linearity as given in Table 10
shows that the deviations from linearity for the amplitude
data in the Hours Deprivation Group is not significant,
indicating the rectilinearity is the best assumption for
the relationship.

Using the amplitude data from the LN subgroup of the
% Fed Group, which appeared when plotted to be somewhat
curvilinear, it was found that there was still no evidence
to indicate any significant deviations from rectilinearity.
This Chi square test is found in Table 11.

Due to the extreme variability Of the latency data,
nothing can be said concerning the true shape of the curve.
The activity data for the SD, SN, and LD subgroups

combined were analyzed for linearity in Table 12 and the

{9

1‘

Hm>ma doapwbaaemp a0 cOapocsa m we AQm.2H.3mV mmdoam pecanEOo moanp mo mmaOom mpa>apom dmoz .OH madmaw

 

Guam»
c. on on he on
— d a _ a

on

Is.
no...

1 ohm

[10¢

1 out

 

ALIAIAOV

coausaaaaop pooa macros .HO aoofiza on» .HO soapoga m we epspaadfiw was honour.” .mpaaapos.

zo.h<>_mdmo manor
0? «Sum N. m

Awm3w<m2 MQDHSQI<

 

 

VN! mo... mmwmowo z. m._<om DZ<I FIE”:
mourns—241
>QZMP<J\ O. .Illl.
OO_\>:>_._.0< Ill. \
\
\
\
\
\
mu-
\\
O, 0
[III \
....I. x
A ’74 \
I, \
II x
I]! \
III \
II \
0 [III \
N”. ”’O\

 

.aa enemas

N .

 

BEDSVBH BSNOdSBCI

 

50

TABLE 10

CHI-SQUARE TEST OF LINEARITY FOR AMPLITUDE DATA
OF HOURS DEPRIVATION GROUP

 

 

Pearson 'r' .557
Eta coefficient .527

Chi-square 2.h96 - not significant

 

 

TABLE 11

CHI-SQUARE TEST OF LINEARITY FOR AMPLITUDE DATA
OF LW SUBGROUP

Q“
1

 

Pearson 'r' .3h6
Eta coefficient .365

Chi-square 1.22 - not significant

 

CE
It

TABLE 12

CHI-SQUARE TEST OF LINEARITY FOR ACTIVITY DATA
OF THE SD, LD, AND SW SUBGROUPS COMBINED

 

 

Pearson 'r' .351
Eta coefficient .388

Chi-square 4.36 - not significant

 

A
t
03.-

deviations were found not to be significant; hence, activity
as a function Of deprivation level in the % Fed animals is
most safely considered as rectilinear. However, the
analysis of activity in the Hours Deprivation Group resulted
in a Chi square of 9.3. A Chi square of 9.49 is needed for
significance at the 5% level Of confidence with h d.f.;
however, one would still not reject the possibility that
the data are curvilinear. InSpection Of these data indicates
that negative acceleration would probably be the best
assumption if the drOp from 1 hour to 2 hours is ignored.
However, neither rectilinearity nor negative acceleration
could be considered similar to the function found by Yama-
guohi (21). Therefore no support was found for prediction 7.
With reSpect to prediction 8 which is concerned with
the possibility Of using activity as a measurement para-
meter Of the construct Drive, three different analyses were
made. Each analysis used a different amount Of variation
in the matching of the activity levels. The first Pearson
'r' (Table 13), using subgroup SD Of the % Fed Group to
equate for size and type of reward with the Hours Depri-
vation Group, was computed allowing a plus or minus functu-
ation of 2 activity points. The N was limited to 16 with
this procedure and the derived 'r' Of 0.29 between latencies
was not significant. Allowing a 20 point deviation (Table 1h)
an N of 23 was Obtained from the same groups as above. In

this comparison the last 3 training days were used as a

En
gs

TABLE 13

CORRELATION COEFFICIENT BETWEEN LATENCIES OF
HOURS DEPRIVATION GROUP AND SD SUBGROUP
EQUATED AS TO ACTIVITY WITH A
FLUCTUATION OF 2 POINTS

 

 

35.75

2X = 1111.92 ZY

2x2 = 271.7722 2Y2 200.1505

ZXY = 138.671;

'1" = .29 + not significant

 

CI 1
0'“

TABLE 1H

CORRELATION COEFFICIENT BETWEEN LATENCIES OF
HOURS DEPRIVATION GROUP AND SD SUBGROUP
EQUATED AS TO ACTIVITY WITH A
FLUCTUATION OF 20 POINTS

T‘T
'— i

A
d

32.9u

EX = 14-5017 ZY

- 130.3515 2Y2 = 85.2280

NI

>4
m
I

'r' = .211 - not significant

 

‘3".

source instead Of the testing days. The resulting 'r' Of
0.2h was not significant. Table 15 shows the 'r' which
was based on a 5 point deviation using the testing days
Of the SD Group as‘a source. This increased the N to 3h,
but the Obtained 'r' Of 0.26 was still not significant.

Because of the previously mentioned apparatus failure
there was no accurate amplitude data for the SD subgroup;
however, the LD subgroup had the same range Of amplitude
score as the Hours Deprivation Group and was selected for
the amplitude analysis. The correlation between amplitude
scores of the LD subgroup and the Hours Deprivation Group,
equated as tO activity plus or minus 5 points was found to
be -O.27 (Table 16). Prediction 8 clearly receives no
support.

Prediction 9, concerning size Of reward, was definitely
not confirmed at least as far as latency data are concerned.
This is Obvious by inSpection Of Fig. 9. The small reward
groups performed better than the large reward which is
Opposite to the prediction; therefore, no statistical test
was employed. Regarding the amplitude data very little can
be said because of the change in setting of the measuring
device for the LW Group and the confounding Of reward type
in the other two groups. The crossing Of the SW and LW
curves (Fig. 6) seems to indicate, however, that at least
all differences were not in favor Of the large reward group.

It may be concluded that in this eXperiment size Of reward

5:}?

TABLE 15

CORRELATION COETTTCIENT BETWEEN LATENCIES 0F
HOURS DEPRIVATION GROUP AND SD SUBGROUP
EQUATED AS TO ACTIVITY WITH A
FLUCTUATION OF 5 POINTS

 

 

EX = 65.5 215 = 126.11
2 2
EXY = 311.7.77

'r' = .26 - not significant

 

58

TABLE 16

CORRELATION COEFFICIENT BETWEEN AMPLITUDE SCORES OF
HOURS DEPRIVATION GROUP AND LD SUBGROUP
EQUATED AS TO ACTIVITY WITH A
FLUCTUATION 0F 5 POINTS

 

 

2x = 799 EY == 656

2X2 = 28.955 2Y2 18,668

ZXY = 21,1187

'r' = ~.27 - not significant

 

with the values employed was not positively correlated
with performance.

Reference to Fig. 7 indicates that there were no
differences among the SW, LD, and SD subgroups regarding
activity which would tend to support prediction 10. It was
found that the LW Group had lost an average Of 17 grams
body weight from the beginning Of the experiment while on
the average all other groups gained weight. Inasmuch as
there were no differences between LD, SD, and SW subgroups,
the indications are that the increased activity Of the LW
Group was due to the loss of body weight instead of to size
or type Of reward.

With reSpeOt to the extinction data there is support
for prediction 11 as to absolute numbers. However, as
indicated in Table 17, the variances are much too disparate
to allow any valid 't' test to be performed. The only valid
analysis which could be performed was between the LW and SD
subgroups as given in Table 18. The resulting significant
difference, however, gives no indication of the particular
variable bringing about the difference because of the con-
founding Of size and type of reward variables. It may
possibly indicate an interaction effect between the two.
Consequently no statement is made concerning prediction 12
which posited an influence of size Of reward on the number
Of trials to extinction.

Table 19 gives the correlation coefficients between

TABLE 17

"F" TEST 0N VARIANCES OF THE NUMBER OF TRIALS T0
EXTINCTION FOR HIGH AND LOW DRIVE SUBJECTS
0F % FED GROUP

 

 

 

High Low
0-2 676. 82.
Mean2 811e2 9Se66

 

F = 8.h9 - significant at 1% level of confidence

TABLE 18

6 1

't' TEST OF THE NUMBER OF TRIALS TO EXTINCTION OF

THE LW SUBGROUP VS.

THE SD SUBGROUP

 

M

 

 

 

LW SD
0’ 9.3 8.h8
I 20.6 10.3
N 9. 10.
01h 3.29 2.82
0711:. = “‘33
t -.-. M. .-.- 2.38 - significant at the 5% level of
“.33 confidence

TABLE 19

CORRELATION COEFFICIENTS BETWEEN THE DEPRIVATION
LEVELS OF BOTH HOURS DEPRIVATION GROUP AND
% FED GROUP WITH AMPLITUDE SCORES

 

Group Correlation
'1"
Hours Deprivation .53 -- significant at 1% level

% Fed .27 -- not significant

 

on
or

deprivation level and amplitude for both % Fed and Hours
Deprivation Groups. The 'r' Of 0.53 for the Hours Depri-
vation Group as compared to the 'r' Of 0.27 for the % Fed
Group (utilizing the data Of the LD subgroup as it is the
most comparable group to the Hours Deprivation Group)
indicates a trend toward the superiority Of the Hours
Deprivation Group because Of the significance Of the 0.53
correlation and lack Of significance in the 'r' Of 0.27.
The difference between the coefficients was not statistically
significant. These results do not support prediction 13.

Prediction 1h receives no support from the data given
in Table 20. The standard deviations for the % Fed Group
are consistently larger than for the Hours Deprivation
Group, indicating that the procedure used for the latter
group served to delimit deprivation more precisely.

The slight superiority of deprivation level in pre-
dicting amplitude over activity predicting amplitude is
not of sufficient magnitude to either support or reject
prediction 15. The assumption of no difference is the
safest. Data are given in Table 21.

($4

TABLE 20

STANDARD DEVIATIONS FROM COMPARABLE POINTS OF
AMPLITUDE SCORES OF THE HOURS DEPRIVATION
AND % FED GROUPS

 

 

 

 

r
GrOUP At 0% or At 90% or
22% hours 2 hours
deprivation deprivation
Hours Deprivation 5.31 3.09
% Fed -- LN 17.61 9.25
LD 11-9h 5.31

SN lu.1u 6.58

 

TABLE 21

CORRELATION COEFFICIENTS UTILIZING DEPRIVATION LEVEL
TO PREDICT AMPLITUDE COMPARED TO ACTIVITY
TO PREDICT AMPLITUDE

CH

 

 

Deprivation level correlated with amplitude

Hours Deprivation Group 'r' = .53

% Fed Group 'r' = .3h

Activity correlated with amplitude

 

Hours Deprivation Group 'r' 2 .50

% Fed Group 'r' = .23

 

68

DISCUSSION

As is the case in many experiments, the peripheral
data, analyzed simply because they are available, have pro-
duced illuminating implications. The first indication
that activity level does not reflect Drive, EEEHEE’ is found
in the gradual decrease in activity Of animals under 22%
hours deprivation (Fig. 2). The decline for these animals
actually exceeded in some degree the decline in activity Of
the satiated animals. The lack Of any difference between
the groups presents evidence contrary to the implications
assumed by Hull (10) to be present in the work of Wade (20),
Richter (l6) and others; namely, that the general activity
Of the organism is a function Of deprivation level. The
present data indicate that with the gradual decline in
activity an asymptote is reached and after introduction of
food into the apparatus there is an increase in activity
which shows the common negative acceleration of a 'normal
learning curve'. This apparently represents activity as a
function of learning rather than Of Drive or deprivation.
If this be the case, then the consequent side Of Hull's
Operational definition Of Drive has been removed and Drive
has been left hanging by its antecedent conditions. As
further support for the thesis that the bulk Of the activity

measure does not reflect Drive but Habit, is the lack of

f.

correlation between the performance (sEr) measures of the
two groups (Hours Deprivation and % Fed) when theyvvere
equated as to activity level (prediction 8). Besides
showing that activity cannot be used as a basic measurement
parameter for the quantification Of Drive, the lack Of
correlation indicates that activity is peculiar to the
situation and to the past experience Of the organism rather
than tO a 'state' of the organism which would be common to
all members Of the Species when they were put under the same
deprivation conditions.

The question.may then arise as to why activity varies
positively with deprivation level as was indicated by the
analyses Of variance and the correlation coefficients. And
even more disturbing is the question Of why amplitude varies
with deprivation level if neither Drive nor habit is invoked
to explain it.

The answers to these questions must necessarily involve
some reformulation of current Hullian theory. It is believed
that utilizing our knowledge Of physiology, in conjunction
with a strict mechanistic approach to the study of the overt
behavior Of the organism, we may arrive at a plausible
answer to the question, 'Is organismic activity a function
Of generalized Drive or is it a function Of learning?‘

In answering this question.we must first recognize

that a certain amount Of activity is due to the structural

68

connections between afferent and efferent nerve fibers.
Modern neurOphysiOlogy provides a comparatively thorough
list of the various innate responses Of the organism to
external and internal stimuli and also lists the intercon-
nections from sense organs to the various parts of the
nervous system which could result in innate responses.

These reflexes are not limited to minute pupillary dilations
or knee jerks, but may involve complete bodily orientation.
The ramifications and interconnections of the Olfactory
system are particularly important and peculiarly ignored by
the comparative psychologists. As an indication of the
numerous possibilities for reflex action originating with
Olfactory stimuli we find the Olfactory bulb connected to
the cortex in the anterior perforated Space, the lamina
terminalis and the gyrus subcallosus; In the hypothalamus

it connects with the mamillary bodiesaand the ventro-
medial hypothalamic neucleus. The habenula Of the epi-
thalamus is particularly important in that it is an Olfacto-
somatic correlation center. Connections also go to the
dorsal tegmental nucleus, the interpeduncular nucleus, the
nuclei of 5, 7, the superior and inferior salivatory

nuclei, the dorsal efferent nucleus and nucleus ambiguous.
Most psychologists are aware of the autonomic connections
with the Olfactory bulb, but unaware Of the connections with
the skeletal muscle system via the habenula. This latter

fact will prove more important when we consider the amplitude

data. Besides the routing Of the neural pathways we also
have some knowledge as to nerve fiber action in conjunction
with spread Of effect. We know, for instance, that the
greater the afferent stimulation the greater the efferent
resultant - especially concerning innate responses, for
example it is obvious that an auditory stimulus Of above
threshold intensity may result in the reflex action of neck
and head orientatiO; in the direction of the stimulus,
whereas increased intensity will result in increased bodily
involvement, culminating in the well known startle response.
It should also be acknowledged that not all efferent result-
ants are measureable or even observable as they may be
visceral as well as skeletal. .

It is here being suggested that the major activity
component, however, is due to learning, and not a resuhsant,
at least entirely, Of innate receptor-effector connections
qualified by the action Of "generalized Drive".

Using an elicitation theory of reinforcement” which
asserts that S-R connectiOns are set up by contigual
association Of a new stimulus with an inherent S-R bond
or reflex, it is readily visible that if we can assert an
innate tendency for the animal to approach food, that approach
would involve locomotion or activity. It also follows that

with Optimal bodily conditions the approach response would

 

*Current Michigan State College theory. Papers soon to
be published on it by Dr. M. Ray Denny and Jack L. Maatsch.

{I

 

 

70

be enhanced and a Simple conditioning pattern appears as
follows:

Bodily conditions -- Food -- approach or locomotion

smell
The conditioning pattern above says that bodily con-
ditions will bring about activity, but the same principle
holds in this situation as holds in any conditioning sit-
uation, that is the stimulus complex must remain constant,
any change in stimulus conditions will result in decreased
response strength. Also, before a locomotion reSponse will
occur in a particular situation there must be a series of
reinforced trials in which the constant stimuli of environ-
mental surroundings are associated with the innate Food ---
locomotion bond. This then presents a complex stimulus
situation in.which bodily conditions are merely a small
portion of the total complex, hence its effects are minimal
and may be superseded by almost any additional extraneous
stimulus, until learning has occurred in.the situation.
Following well substantiated conditioning theory we know
that if the organism is put under given deprivation con-
ditions so as to keep the internal stimuli constant and is
placed in a given apparatus, and a reSponse is learned
under these conditions, then any deviation from these con-
ditions results in generalization. It is asserted here
that the reSponse learned in the present situation was not
only the panel pushing response but that the locomotion

reSponse (not new to the animal, but new to the animal in

on
H

the particudar'situation) is also learned.

From this typeAanalyses it follows that in a new
situation the organism must learn to eat or make the associ-
ation between a stimulus complex and the response to food.
It is seen then that the maximal activity Of the animal
will occur when bodily conditions are Optimal and the
reSponse of eating in a particular situation is asymptotic.
If either of the above conditions does not hold, activity
declines. The decrease in activity of the animals in all
groups away from the original 22% hour training deprivation
level is then asserted to be a generalization effect instead
of a decreased Drive effect. The Obvious deduction from
this approach would be that if the animal were fed in a
new situation under given bodily conditions (1.0., depri-
vation level) that shifting the animal to a different
deprivation level would result in generalization from the
learned reSponse and a consequent decrease in activity.

Now, it does not seem reasonable that if an animal were
trained under 10 hours deprivation that when it was put

under 2h hours deprivation that a decrease in activity would
occur. Yet, this result is what this analysis demands;
however, it is very possible that there is an interaction
effect between the innate reflexes and the learned reSponses
such that at low levels the innate reflexes might predominate,
so that if an actual decrease from 10 to 24 hours did not

occur, the increase would be very slight, not at all

/§

s It

comparable to the increase between those approximate levels
in the present study.

Actually in the present study there is some evidence
for this stimulus generalization effect in that there was
a decline in activity from the 22% hour level to the AB
hour level shown in Fig. 8. The reduction in activity is
definitely not significant, but even if it is considered
to be a straight line, this still does not compare with
the curve derived by Yamaguchi (21) who stated that Drive
is maximal at 59 hours deprivation. In fact. the function
Yamaguchi derived shows the increase between 22% and AB
hours to be very steep. That the absence Of increase is
not due to the approach to the upper limit Of the apparatus
in recording activity is seen in the fact that at times
several of the animals scored over 1000 on the activity
scale and the mean for the 22% hours group equals only u25,
which leaves ample room for increase as deprivation increased.

Theoretically, to a certain extent amplitude and
activity are correlated. In the free situation in which the
organism responds directly to a bit Of food the locomotion
of the organism will be highly correlated with the amplitude
Of the reSponse; in fact, they may be identical.‘ In the
panel pushing device, the difference between the two is
increased principally by forcing the animal to make a pushing
reSponse instead of a simple approach reSponse which would

be entirely determined by locomotion. The difference between

.3
C!)

the two measures allows the possibility for Operation of

a new factor in the amplitude measures. The additional
factor is the Olfactory stimulus. If we can assume that
the deviation of the LW Group (Fig. 9) from the others is
due to loss of body weight, it may be readily assumed that
the activity levels do not deviate as a function of type

or amount of reward. In the Hours Deprivation Group (Fig.
8) the range is between 150 and 500 which compares closely
to the ranges of the % Fed animals (Fig. 9). However, in
making the available amplitude comparisons we can see that
the animals who were rewarded with wet mash were definitely
superior at the high drive levels; whereas, the groups were
essentially the same at the low levels. Figure 6 shows
this quite clearly regarding the LD Group; although the
data for the SD Group is not plotted, the highest mean
amplitude recorded for this group was 36 which coincides
with the high point for the LD Group. Also, if four points
are added to the amplitude scores Of the Hours Deprivation
Group to make the low point coincide with the low point mean
Of the % Fed Groups, we find that the maximum amplitude
score attained is about 36. These three groups just men-
tioned were rewarded with the dry pellets. A glance at
Fig. 6 again shows that the high points for the wet mash
rewarded groups were above AB in both cases. The differences

between the groups cannot be validly tested, due to differences

«3
543::

in settings of the instrumentemd size of reward but the
weight of the evidence indicates that wet mash rewarded
groups reSpond more forcibly at the high deprivation levels
than do the dry pellet rewarded groups. These data probably
tie in with the food preference studies carried out by
YOung (23); however, if one continues to adhere to the
strict S-R outlook, one would be forced into asserting the
Operation of an additional stimulus to account for the
preference. In this case, it is believed to be the Operation
of an Olfactory stimulus. To the human receptors, at any
rate, there is a decided difference in the intensity Of the
Olfactory sensation between the wet mash and the dry pellet.
It does not seem to be overly anthrOpomorphic to suggest
that there is definitely a discriminable difference between
the two types of reward for the rat.

It may be noticed that there is a definite increase
from 22% hours to A8 hours in the amplitude data Of the
Hours Deprivation Group (Fig. 6). Although the increase
is not statistically significant, it still indicates a trend
away from the generalization gradient from the 22% hour
training conditions asserted to hold for activity. This
may be an interaction effect between the drive stimuli and
the olfactory stimuli. We know from Tinbergen's results (18)
that the internal conditions of an organism must be Optimal
before the innate reSponse to a stimulus will be Optimal.

The maximal amplitude measure may reflect the existence Of

re

’Q
U‘.

the Optimal internal conditions prerequisite for elicitation
of approach response to food by the external Olfactory
stimulus. Also, from Tinbergen (18) we know that the ex-
ternal stimulus which elicits a reSponse must also be Of
Optimal quality and intensity. It is suggested that the

wet mash more closely approximates the Optimal conditions
than does the dry pellet. No measurements were taken
beyond the training level for the % Fed Groups; hence,
evidence for this is not available.

The preceding analysis asserts that the amplitude Of
reSponse is a function largely of the eliciting Olfactory
stimulus; in other words, if no Olfactory cue were avail-
able, there should be very little change in amplitude as a
function of change in internal conditions (i.e., deprivation
level). This, then would remove amplitude as a possible
means Of quantifying "Drive". In fact, if activity is a
function Of learning and amplitude is a function of the
eliciting stimulus, there is little basis for asserting the
action Of the construct Drive unless a case for "Drive"
may be built upon the fluctuations of latency with fluctu-
ations in deprivation level.

With regard to latency the present data could be dis-
missed as apparently confounded by the action Of unknown
variables, changing, so tO Speak, the pOpulations from which
the various samples were drawn. From the statistical view-

point, this might be the thing to do, but it is the author's

76

contention that the effect of variables on the variance Of
groups has been totally ignored by most investigators and
that the variances may be the source Of important information
unattainable by reference to group means or correlation
coefficients. Reference to Fig. 5 will Show the increase
of the standard deviations Of the various groups as a function
of decrease in deprivation level. Comparison of this figure
to Fig. A shows the high degree of correlation between the
latency measures and the variances Of the groups. One fact
is particularly outstanding. At the training level, under
zero % Fed conditions, the latencies, and the standard
deviations for that matter, are very similar. In other
words, the animals were apparently reSponding with approxi-
mately the same latencies, and it was not until the stimulus
conditions (internal stimuli) were changed that the means
and variances began tO become diSparate.

In all ththe % Fed Groups the 'F's' were significant
but the variances were also very highly Significant. In
the Hours Deprivation Group however, the variances were
homogeneous and the means were not Significantly different
from one another. This presumably indicates that the
differences in means is only a reflection Of the difference
in variances which shows that in this case the effect of
varying deprivation level is to increase variability Of

reSponse, not to affect the Speed Of the reSponse directly.

When considering the differences between subgroups,
the type of reward rather than Size Of reward seems to be
the meortant variable. The means and variances Of the wet
mash group are consistently below those of the dry pellet
reward group. This represents an interaction effect between
type of reward and drive level. This could be an indication
of different gradients Of generalization as a function Of
different types Of reward. This again may be explained
by referring to the interaction between internal conditions
and external stimuli emphasized by Tinbergen. It may be
that an external stimulus Of lesser intensity will exert
progressively less effect as bodily conditions are altered
than will a stimulus Of stronger intensity.

The indications are that latency does not reflect the
same principle as either amplitude or activity. There are
indications from the work Of Cotton (3) that the major
portion of latency is due to the intervention of competing
reSponses. Gotten found that after eliminating trials
during which reSponses incompatible with performing the
learned response occurred, the difference in latencies
between deprivation levels from 0 to 22 hours was only 0.h
of a second, measured in an eight foot long runway. This
component is so small that the effects of deprivation level
on latency Of response can be considered negligible as far

as representing a relevant behavioral variable.

Cotten's findings are supported in some degree by the
results Of the present study, when the range of latency
scores within a group for all levels of deprivation are con-
sidered. The smallest range was 0.2 second occurring in
the LW subgroup, the greatest was h.7 seconds in the LD
subgroup. The other ranges were 2.h seconds for the SD
subgroup and 0.9 second for the SW subgroup. The greatest
range was more than 23 times that of the smallest range.

The difference appears to be a function of the type of
reward interacting with varying internal conditions. With
near Optimal reward, in this case large wet mash, bodily
conditions are for all practical purposes behaviorally in-
consequential with reSpect to latency. The type of reward
seems to be a much more important variable in determining
mean latency than deprivation level.

The charge that latency Of reSponse is not a reflection
Of sEr is particularly important in view of the fact that
sEr was quantified by Hull and his associates on the basis
of the latency Of a bar pressing reSponse (5, 6, 12, 22).
Drive was held constant at 23 hours deprivation and the number
Of reinforcements were varied. The construct sEr was then
quantified on the basis Of the derived latencies. In the
body Of one article (6) it is explicitly stated that the
validity of the quantification is dependent upon the equation

sEr = er X D

and the appeal to the validity of the Drive construct was

'79

in reference to Perin's data (15), in which reinforcements
were held constant at 16 and the deprivation level was
varied, in a situation similar to the one employed in the
sEr quantification series. Perin pointed out in this study
that the latency data were not amenable to any sort of
analysis, so he presented most Of his results in terms of
the number of trials to extinction. It was the extinction
data which Hull used as support for his sEr = er X D
equation. In addition when Yamaguchi (21) quantified Drive
more fully he also used the number of trials to extinction
as his measure. It would seem that if latency yields data
adequate for quantifying sEr it would also yield data
adequate for quantifying Drive which supposedly is reflected
in sEr. Why then, was the utilization of two different
measures necessary? The results Of the present study seem
to indicate that the reason for this is that variation Of
internal bodily stimuli introduces extreme variability in
latency scores. If latency is considered a function Of the
number Of competing reSponses in action at the time, this
would mean that latency would probably not be a measure Of
momentary effective reaction potential and therefore not

a valid basis for quantifying sEr. It should also be pointed
out that the standard deviations for latencies at the
various reinforcement levels were not publialed in the
quantification series, making it impossible to determine

whether or not even those variances were homogeneous.

8'0

The extinction data in the present study were of small
value because Of the small N in each group and the extreme
variability of the data. There was an increase in the
mean number Of trials to extinction with increase in depri-
vation in every group. However even this measure cannot
be used as evidence for greater strength Of reaction potential
mainly because Of the lack of any Operant level data and
the possibility that much of the panel pushing after the
first few extinction trials may be merely a concomitant of
activity. Such possibilities may be easily tested by
allowing untrained animals access to the panel after an
initial activity recording period and comparing the number
Of panel activations with the recorded activity.

In the Hours Deprivation Group there is one phenomenon
which is particularly interesting. This is the dip in
every curve at the two hour level. Every measure showed a
decrease in reSponse strength from one to two hours. None
of the decreases reached the 5% level of significance.
However, the coincidence of the drOp in all three curves
is has hardly attributable to chance, particularly since
there were no consistent reversals Of the % Fed Group curves.
The dip, probably more than anything, reflects the period
Of time required for complete assimilation of food into the
blood stream. This supports the work Of individuals who

claim that the neural basis for hunger is not localized

8’1

solely in the stomach, but is more essentially dependent
upon central neural centers which reSpond to changes in
blood chemistry, i.e., chemo-receptors. The present data
indicate that there is probably a constant decrease in the
organism's reSponsiveness from complete satiation to a
point somewhere near two hours after satiation. Reference
to minimal effective internal conditions should probably
be in terms of this two hour level instead Of zero hours
Of deprivation from time Of satiation. This observation
is consistent with the foregoing drive stimulus analysis
and does not support the "Drive" construct formulation.
The drive stimulus point of view would postulate that any
decreased firing of receptors would, innately, reduce the
magnitude Of certain aspects of behavior. AlsO, decreased
receptor firing would result in internal stimulus conditions
further out on the gradient of generalization from the
originally conditioned stimulus complex and would serve to
reduce performance of the learned aspects Of behavior.

The two hour dip may be one Of the causes for the curvi-
linearity of the activity data for the Hours Deprivation
Group, which is in contrast to the same data for the % Fed
Group which was rather definitely a rectilinear function.
Another possible difference was the fact that the Hours
Deprivation Group was tested at A8 hours which is past the
original training point, and the decrease rather than a con-

tinued increase would tend to make the function appear to be

822

curved instead of rectilinear. The % Fed Group was never
extended beyond the point of original training hence only
one side of the gradient was involved.

Although the results of this study very definitely
show that activity cannot be used as a means Of quantifying
Drive and that the whole concept of a generalized Drive is
extremely tenuous it does not present exactly incontrovert-
ible evidence against Drive. The Drive theorist might say
that every time a reference in this paper was made to
internal conditions or internal stimuli or drive stimuli
that the word 'Drive' could have been substituted. However,
this would at any rate point up the confusion existing
between the concepts of Drive and drive stimulus. If the
Drive theorist did not make this assertion then it should
be clear that the behavioral results attributed to Drive
may be explained more parsimoniously with more primary
physiological and behavioral concepts. The value of doing
this lies chiefly in the fact that modern research methods
are enabling the physiologist and ethologist to investigate
more thoroughly the various aSpects of stimuli and their
interaction with an organism. We have shown in the present
paper that Drive cannot be tied down on the consequent side
Of the Operational definition by measuring energy expendi-
ture in the form of activity. This means that Drive, then,
must be accepted as a matter of faith and not scientifically
testable. It must be acknowkadged that the present analysis

00
Q5:

consists in part of 22 222 explanations; however the point
is that the explanations were made by appealing to well
known or highly substantiated ethological and physiological
data which lead to deductions which may be experimentally
investigated.

Obviously if one denies the concept of Drive, then
something must be said concerning an even.more important
concept - that Of reinforcement, which for the Hullians is
drive reduction. If one accepts the notion Of reduction
Of drive stimulation instead of "Drive" reduction then the
essence Of the principle may possibly be retained. However,
then again, of what use in this connection is the concept
of generalized "Drive"? Although various individuals have
been interpreting drive reduction as reduction of drive
stimulation, even this tack becomes difficult in the light
Of some evidence from the present study. Three of four
activity curves and one of three Of the amplitude curves
in the % Fed Groups showed an elevation of the 10% Fed
level over the 0% which might indicate that feeding a
slight amount of food would actually increase the intensity
of the stimulation instead of acting to reduce it. This
ties in with the results from the experiments on pro-feeding
(1, 2) which show that pro-feeding seems to aid performance,
increases activity, etc. If this is the case that pre-
feeding increases activity (shown in three of the four groups

Of the present study), then a bit of food.would neither

QC
a:

reduce drive stimuli nor generalized "Drive" and thusly
the possibility Of using these principles as explanations
of reinforcement is effectively eliminated.

The value Of the present study lies not so much in
the absolute evidence which it p'esents for or against a
given theory but in the Opening up of new methodological
approaches to clearing up several pertinent problems in
modern learning theory. The technique utilized in this study
measures practically every aSpect Of measurable behavior I
Of the organism, enabling the author to make inter-behavioral
comparisons which heretofore could not be made. The failure
to do this in the past shows up OSpecially in the formu-
lation of what were supposed to be reflections of reaction
potential (sEr). As this discussion has attempted to point
out, these measurements may be only slightly related. The
major portions Of each seem to be a result of the Operation
of very different principles, with particular reference to

habit, innate reflexes and competing reSponses.

oo-
Q)".

SUMMARY AND CONCLUSIONS

The present investigation is concerned with the
possibility of utilizing the activity level of an organism
as a direct reflection Of Drive when two different methods
Of inducing hunger conditions are employed. The results of
such a study should present evidence as to whether or not
activity level may be used as a measurement parameter which
would make possible the quantification of the construct
Drive independent of the commonly used reaction potential
(sEr) measures, e.g., latency, amplitude, trials to extinction.
Further, the size and type of reward are varied to give an
indication Of the effect of these variables on the behavior
of organisms while performing under varying drive levels.

A combined activity box and panel pushing device was
employed in the study. The panel pushing device was so
arranged as to provide measurements of latency of reSponse,
amplitude Of reSponse and the number Of trials to extinction.

Forty-eight male albino rats were used in the study,
and were divided into two major groups:

1. Hours Deprivation Group N = 11

2. % Fed Group N = 37

The % Fed Group was divided into four subgroups on the

basis Of size and type Of reward as follows:

(1

85

l. LW subgroup (Large wet mash - 0.2 gm) N
2. SW subgroup (Small wet mash - 0.08 gm) N
3. LD subgroup (Large dry pellet - 0.2 gm) N
M. SD subgroup (Small dry pellet - 0.08 gm) N

1%

10

All animals were given habituation in the apparatus for
a period of five days, and received training on the panel
pushing device 10 trials per day for the next 8 consecutive
days. All animals were under 22% hours food deprivation.

0n the testing days the animals in the Hours Deprivation
Group were satiated and then given four trials in the appar-
atus either 1, 2, 6, 12, 2A, or A8 hours later. The % Fed
animals were fed either 0, 10, 25, 50, 75, or 90 percent Of
the total amount of food required to produce satiation,
computed on the basis of the amount of food each animal had
eaten to satiation on the last four days Of the training
procedure. They were then given four trials in the apparatus
three hours after they had consumed the stipulated percentage.
All animals performed under each deprivation level in their
reSpective groups, being assigned to each level in a ran-
domized order. Activity level, latency and amplitude were
recorded on each trial. At the conclusion of the testing
days all animals were extinguished under various deprivation
levels.

The findings were as follows:

1. Activity is not a reflection of Drive pg; 32, but
is dependent upon learning in a given situation.

2. Activity cannot be used aS a measurement parameter

for Drive, inasmuch as the correlations between reSponse

,fi

L7".

measure Of different animals equated as to activity Show
no significance.

3. There is definite interaction effect between type
of reward and deprivation level.

A. Latency as a measure Of reaction potential is
extremely suSpect.

5. The results do not support the notion Of general-
ized Drive as a hypothetical construct.

The possibility that the various measures Of reaction
potential were essentially measures Of very different factors
was discussed and evidence was Offered for this point of

view.

l.

2.

3.

A.

5.

6.

7.

8.

9.

10.

ll.

12.

86"

REFERENCES

Anderson, E. E., The externalization of drive IV.
The effect of prefeeding on the maze performance
Of hungry non-rewarded rats. ‘g. comp. Psychol.,
l9hl, 31s 3M9 ' 352e

Bruce, R. H., The effect of lessening the drive upon
performance bg white rats in a maze. ‘g. comp.
O

Psychol., 193 25, 225 - 2H8.

Cotton, J. W., The relationship between running time
in a straight runway and length Of the period of
food deprivation. To be published in 1. exp. Psychol.

Davis, R. H., The quantification of drive I. Incentive
values Of food and water. Ph.D. thesis on file in
the Michigan State College Library, 1953.

Felsinger, J. M., Gladstone, A. I., Yamaguchi, H. G.,
and Hull, C. L., Reaction latency as a function of
the number of reinforcements (N). g, exB. Psychol.,
19h7. 37. 21h - 228.

Gladstone, A. 1., Yamaguchi, H. G., Hull, C. L., and
Felsinger, J. M., Some functional relationships of
reaction potential (sEr) and related phenomena.

ie 3.2. PSzChOIe, 19(47, 37, 510 II 526.

Guthrie, E. H., The s cholo Of learni . New York:
Harper and BrOEEers FuBIISEerST I935.

Helson, H. (Ed.), Theoretical foundations g; psycholqu.
New York: Van Nostrand, 1951.

 

Holt, E. B., Animal drive and the lea nin process.
New York: Henry HOIE and Company, .

Hull, C. L., Principles 2f behavior. New York:
D. Appleton-Century 00., Inc., I9u3.

, Behavior postulates and corollaries -

 

lguge FEEOhOI. ROVee 19509 57’ 173 ' 1800
, Felsinger, J. M., Gladstone, A. I., and

 

Yamaguchi, H. G., A prOposed quantification of habit
strength. Psychol. Rev., 19h7, SH, 237 - 25h.

Q
m.— on...
r O I
a
- ~— . _
- k,
u I O

I
I
e
.l.
»m
I

“P

 

 

O O '
. h.
.—
fl
6 1
( a e
i Q l '
t I O
‘ .
I
t I O
I
.
t C 0 '
{f 0 O t
. a... '-
. -
Q 0 C ' ‘o
c. ' ' i
s 0 '
e . e
\
r O 1
H.,—M.“
C O
h
\
Q e
.K -
.
(:
C 0
~—.-—_ .—.——_I...— --..__.
f.
l . . r
. .
__ -_.‘..____._l.. ._..._.

13.

1k.

15.

16.

17.

18.

19.

20.

21.

22.

23.

Keller, F. W. and Schoenfeld, w. N., Princi les g£
psycholoE%. New York: Appleton-CenEury-Crofts,
rice, 0

Maatsch, J. L. and Behan, R. A., A more rigorous
theoretical language. Psychol. Rev., 1953, 60,
189 - 196.

Perin, T. 0., Behavior potentiality as a joint function
of the amount of training and the degree of hunger
at thz time of extinction. l. exE. Psychol., 19k2,
30. 9 .

Richter, C. P., Animal behavior and internal drives.

guar. Egg. Biology, 1927, 2, 307 - 343.

Skinner, B. F., The behavior 2: organisms. New York:
D. Appleton-Century Eompany, 1938.

 

Tinbergen, N., The stud 22 instinct. Oxford:
Clarendon Press, IGSI.

Tolman, E. C., Purposive behavior in animals and men.
New York: The Gentury Co., 1932.

 

Wada, T., An eXperimental study of hunger in its
relation to activity. Arch. Psychol., 1922, No. 57.

Yamaguchi, H. G., Drive (D) as a function of hours of
hunger (h). ‘g. 6X2. Psychol.,1951, u2, 108 - 117.

, Hull, C. L., Felsinger, J. M., and
Gladstone, K.—I., Characteristics of dispersions '
based on the pooled momentary reaction potentials

(sEr) of a group. Psychol. 531., 19u8, SS, 216- 238.

 

Youngé P. T., Motivation of behavior. New York: Wiley,
193 . '—'

py <——.—

I,

K O

 

MUM—+962 a

 

 

WW

9

H8
Illa]
Ill-0
S I."
R “‘79
E H
V I1