Instinctive drift in nondomesticated rodents
Bulletin of the Psychonomic Society
Instinctive drift in nondomesticated rodents
ROBERT w. POWELL 0
MICHAEL CURLEY 0
0 University ofSO'Uth FWrida , Tampa, Florida 33620 , USA
In the first experiment, Mongolian gerbils were studied under gradually increasing and decreasing fixed·ratio schedules of food reinforcement. Most of the animals showed a permanent displacement of the leverpress response by more genetically based behaviors, such as scratching and biting, as the ratio requirement increased. This type of phenomenon has been referred to as instinctive drift. The change in response topography resulted in high response rates and brief postreinforcement pauses, which remained relatively constant despite changes in the ratio. The few animals that maintained the leverpress response, which had initially been shaped, showed the usual direct relationship between ratio size and duration of the postreinforcement pause. The above effects were independent of the type of manipulandum, which included an omnidirectional lever, a conventional rectangular lever, and a metal chain suspended from the ceiling of the test chamber. In the second experiment, a group of nondomesticated cotton rats was studied under conditions similar to those employed in Experiment 1, except that higher ratios were included. All of these animals quickly developed instinctive drift, with the leverpress being displaced by scratching and biting. Response topographies of the rats were quite similar to those displayed earlier by the majority of gerbils. The duration of postreinforcement pauses remained relatively constant as the ratio increased, except for a marked increase in the mean pause at the highest ratio (90) studied. The overall results suggest that a dual system of control may operate during appetitive schedules, with responding being influenced by both elicitation and reinforcement mechanisms.
Breland and Breland (1961
) coined the term "instinc
tive drift" to refer to the displacement of arbitrary con
ditioned responses by more genetically based behaviors,
which are associated with food getting in the natural en
vironment of the animal. In their attempts to train a
variety of species, they had frequently observed this
phenomenon. As an example, they cite the case of a rac
~oon which was easily trained to pick up a coin, carry it
to a metal box, and deposit it ia a slot for food rein
forcement. However, when the animal was required to
do the same thing with two coins, it perSisted in rubbing
them together and dipping the coins into the box, but
frequently would not let go of them. This resulted in
[lOnreinforcement, but, despite this fact, the rubbing and
dipping occurred more frequently over time. These be
rraviors appeared similar to responses associated with
feeding in the natural environment of the raccoon.
With the recent resurgence of interest in the biology
of behavior among experimental psychologists, the con
~ept of instinctive drift has received considerable atten
tion, particularly in regard to its theoretical implications
(Bolles, 1972; Staddon & Simmelhag, 1971)
other than the Breland's report, which was entirely anec
dotal, there is little empirical evidence bearing on the
validity of this concept.
It was not our initial purpose to study instinctive
drift. Rather, we sought to extend the generality of the
direct relationship which has been found to exist be
tween the size of a fIxed-ratio (FR) schedule of food
reinforcement and the duration of the pauses which
follow food delivery
(Felton & Lyon, 1966; Ferster &
Skinner, 1957; Powell, 1968, 1972)
. When we observed
~hanges in the topography of the responses in our fIrst
group of subjects that were correlated with a relatively
constant postreinforcement pause (PRP) regardless of
the FR size, we decided to replicate our study with
additional subjects and different manipulanda. We
suspected that the manipulandum used with our fIrst
group, an omnidirectional lever, might have interacted
idiosyncratically with the behavior of the gerbil to foster
apparent instinctive drift.
Subjects. The subjects were eight female and six male
Mongolian gerbils, between 1 and 2 years of age, maintained at
approximately 80% of their normal weights during the experi
Apparatus. A conventional operant conditioning test
chamber, 30 x 20 x 23 cm, housed within a sound-attenuating
chamber, was used. The test chamber contained a foodcup
mounted on one Wall, .5 cm above the grid floor. Noyes pellets
(45 mg) served as the reinforcer. One of three manipulanda was
present during successive phases of the experiment. The manipu
landa used were a round plastic omnidirectional lever, a rec
tangular metal lever, and a metal chain which was suspended
from the ceiling of the test chamber. Each of the levers was
mounted on the same wall as the foodcup, and 4.0 em from it.
Illumination was provided by a 7-W yellow houselight. Schedule
programming, delivery of reinforcers, and recording of responses
were accomplished by conventional electromechanical equip
ment. Direct observations of response topography were made
during each experimental session.
Procedure. The experimental program was a systematic
replication of an earlier experiment in our laboratory which had
. When evidence of instinctive
drift was found in the fIrst group of six subjects, studied with an
omnidirectional lever, we extended the study to include two ad
ditional groups of six and five subjects each, studied with a
~-~ RECTANGULAR LEVER
_ INSTINCTIVE DAIFT
<>---0 SHAPED RESPONSE
tangular lever and a metal chain, respectively. The latter group
included three gerbils that had been studied first with the rectan
gular lever. With this exception, all of the gerbils were experi
mentally naive prior to FR training. The gerbils were shaped
first by successive approximation to make the required operant
response. All subjects were exposed to a schedule of increasing
and decreasing FR requirements, the highest of which was
FR-40. Each ratio requirement remained in effect for 5 sessions
for all groups, except that FR-40 continued for 10 sessions for
subjects making the chain-pull response. This was done to allow
additional time for responding to stabilize. The duration of a
session was based on the time it took each gerbil to obtain 34
Although all of the gerbils were initially shaped to
press downward with two paws upon the levers, or to
pull the chain downward with two paws, 10 of the 14
gerbils studied showed a substantial change in response
topography during the experiment. For these animals.
the operant response was replaced by behaviors which
appear to be part of the animal's genetically based
response repertoire. This typically occurred at FR-IO 0
FR-15. Once the change in response topograph~
occurred (instinctive drift), the original operant respons
did not reappear, even when the FR decreased. With th
appearance of instinctive drift, PRPs showed little
change as FR size increased, and then decreased. Thus
the PRP curves for the three groups are relatively flat, a:
shown in the upper portion of Figure 1.
Five of the six gerbils studied with the omnidirec
tionallever showed instinctive drift. These animals devel
oped rapid scratching and biting upon the lever, behavio
which appeared topographically similar to burrowing
Four of the six gerbils studied with the rectangular leve
also displayed instinctive drift, with their response
being quite similar to the behavior described above. Thl
five gerbils trained to respond by chain-pulling include(
three animals that had displayed instinctive drift whel
studied with the rectangular lever. The gerbils in thi
group were shaped to grasp the chain with two paws an(
pull it downward a minimum of 1.0 cm. The respons,
topography of all animals, including the three gerbil
studied earlier, appeared similar following shaping. Fou
of the five gerbils in this group subsequently displayee
instinctive drift. The predominant form of respons,
which developed was for the gerbil to grasp the chail
with both paws, tilt its head to a 90-deg angle, and gnav
the chain. The gerbils also rotated the chain as the~
gnawed upon it. The three gerbils from the rectangula
lever group displayed instinctive drift here also .
The lower portion of Figure I presents a breakdowl
of the PRP data based upon the predominant responSt
topography displayed by each animal. The four gerbil
that maintained the shaped response showed systemati,
increases and decreases in PRP duration as a function 0
changes in the FR requirement. These results are con
gruent with the findings for other species
Skinner, 1957; Powell, 1972)
. On the other hand, the II
animals displaying instinctive drift showed little chang'
in pause duration as a function of changes in the re
sponse requirement. Vanderweele, Abelson, and Tellisl
(1973) reported a similar finding with gerbils, but the~
did not present a quantitative analysis of their results
Figure 2 presents an analysis of the response rate cal
culated independently of the PRP (number of response
divided by session time minus PRP time). Tltis is some
times referred to as the running rate. The upper portiol
of the figure shows that the response rate generally in
creased for the animals in each group as the Ff
increased to 40. However, the rate continued to increas,
as the FR decreased to 25 and 15, but then declinee
with further reductions in the ratio requirement. Thi
apparent "overshooting" may have occurred because 0
the relatively brief time each ratio was in effect (flv
sessions), so that sensitivity to schedule changes wa
delayed. The lower portion of the figure again presents
breakdown of the aninlals into two groups based OJ
their predominant response topography. The 10 animal
that displayed instinctive drift maintained a substantiall:
This experiment was intended to assess the generality
of our earlier results using a completely nondomesti
cated species, the cotton rat. It was reasoned that the
displacement of operant behavior by more genetically
based responses should be even more prominent in a
species which had not been exposed to domestication.
Mongolian gerbils have been bred in captivity for
approximately 20 years
, so it may be
claimed that the species is partially domesticated.
Cotton rats (Sigmodon hispidus) and gerbils (Meriones
unguiculatus) are rather closely related, both belonging
to the same order (Rodentia), suborder (Myomorpha),
and family (Cricetidae).
Subjects. Five cotton rats, two males and three females, were
used. The rats were the fust generation offspring of field-trapped
adults. They were between 6 and 11 months of age at the start
of the experiment and their weights ranged from 80 to 112 g. All
rats were experimentally naive and were maintained at approxi
mately 80% of their normal weights during the experiment.
Apparatus. The apparatus was the same as in Experiment I,
except that only the rectangular lever was employed.
Procedure. The cotton rats were fust shaped to press the
lever with two paws, and all rats successfully acquired the
response. Then the rats were exposed to the following series of
increasing FR requirements: 5, 10,15,25,40,60,90. Two of
the rats were then studied as the FR decreased to 60 and then
30, while the remaining three rats were studied at FR40 only
during the decreasing series. Each ratio requirement remained in
effect for a particular rat until responding stabilized. The stabil
ity criterion required that response rate remain within a range of
± 10% of the mean rate over three consecutive sessions. The dura
tion of a session was based on the time it took each cotton rat
to obtain 34 reinforcers.
There was clear evidence of instinctive drift for each
cotton rat by the time the FR increased to 15. That is,
the shaped leverpress was permanently replaced by a re
sponse topography in which scratching and biting of the
lever was prominent. In addition, four of the rats fre
quently grasped the lever in their teeth and vigorously
shook their heads in the vertical plane.
The mean postreinforcement pause for the group was
relatively constant from FR-5 to FR-60, but then
increased substantially at FR-90, as shown in Figure 3.
When the FR requirement decreased, the mean pause
durations decreased also, to values which approximated
those obtained during the increasing series. The reversi
bility shown here demonstrates that ratio size did in
fluence pause duration, but oPly when the value of the
former was rather large.
Figure 3 shows also that the mean response rate for
the cotton rats, which was calculated as described
earlier, increased gradually as the ratio requirement in
creased. The curve for the group appears to be a nega
tively accelerated increasing function. The response rates
obtained with decreasing ratio requirements closely
approximated those obtained during the increasing
series, so there is further evidence of functional control.
The most significant finding here was the occurrence of
instinctive drift in almost all animals studied, with exposure to
FR schedules of food reinforcement. Our results appear similar
to the observations of
Breland and Breland (1961
), in that an
arbitrary operant response was displaced by responses which
appear to be genetically based and are more prepotent in the
animal's behavioral repertoire. While the appearance of instinc
tive drift was associated with increases in the number of re
sponses required to obtain food, in both our study and the
prevailing contingencies of reinforcement -a point stressed
heavily by the Breland's. However, ~0me sensitivity does remain,
as shown by the marked increase in the pausing of the cotton
rats at FR-90, and the subsequent decrease which occurred when
the FR decreased. This result seems indicative of a process
involving dual control by genetic and environmental factors, in
which control may be shared or one factor may override the
other, depending upon their relative potency.
Analysis of response rates shows that, in general, rate in
creased with increase in FR size, at least up to FR-40. The con
tinued increase in rate when FR decreased, which was found in
Experiment J, probably is attributable to the small number of
sessions each FR remained in effect. When responding was
studied until it stabilized at each FR, as in Experiment II, this
effect did not occur. While the direct relationship found here be
tween response rate and FR size is not unique
, an inverse relationship has been reported more
(Felton & Lyon, 1966; Powell, 1968, 1970, 1972)
The present research shows that a phenomenon identified as
instinctive drift may intrude into experiments involving appeti
tive reinforcement. This may be more likely to occur when one
is studying nondomesticated animals. When instinctive drift does
occur, it seems to alter substantially the functional relationships
that have typically been observed under similar conditions. While
our research does not identify the causes and boundary con
ditions of this phenomenon, these appear as worthy goals to pur
sue in future investigations.
BAROFSKY. I.. & HURWITZ. D. Within -ratio responding during tixed-ratio performance . Psychonomic Science . 1968 . 11 . 263 - 264 .
BOLLES. R. C. Reinforcement. expectancy. and learning . Psychological Review . 1972 . 79 . 394 - 409 .
BRELAND. K .. & BRELAND , M. The misbehavior of organisms . American Psychologist . 1961 . 16 . 681 - 684 .
FELTON. M.. & LYON. D. O. The post-reinforcement pause . Journal of the Experimental Analysis of Behavior . 1966 . 9 . 131 - 134 .
FERSTER. C. B .. & SKINNER. B. F . Schedules of reinforcement. New York: Appleton-Century-Crofts. 1957 .
MONROE. B. N. Gerbils. Jersey City. N.J: T. F. H. Publications . 1967 .
POWELL. R. W. The effect of small sequential changes in fixed-ratio size upon the post-reinforcement pause . Journal of the Experimental Analysis of Behavior , 1968 . 11 . 589 - 593 .
POWELL. R. W. Probability of reinforcement and flxedratio responding: A model . Psychological Record . 1970 . 20 , 73 - 80 .
POWELL. R. W. Responding under basic schedules of reinforcement in the crow . Journal of Comparative and Physiological Psychology , 1972 , 79 , 156 - 164 .
STADDON , J. E. R. , & SIMMELHAG , V. L. The "superstition" experiment: A reexamination of its implications for the principles of adaptive behavior . Psychological Review . 1971 , 78 . 3 - 43 .
VANDERWEELE , D. A.• ABELSON , R. M. • & TELLISH. J. A. A comparison of ratio behavior in the gerbil and white rat . Bulletin of the Psychonomic Society , 1973 . 1 . 62 - 64 .