Fundamental Motor Patterns of the Mammalian Fetus
Scott R. Robinson * and William P. Smotherman
Laboratory of Perinatal Neuroethology, Center for Developmental Psychobiology, Department of
Psychology, Binghamton University, Binghamton, New York 13902-6000
SUMMARY
Techniques that permit direct observation of fetuses in
vivo recently have expanded our understanding of prena-
tal behavioral development in mammals. Although fetal
motor activity seems to lack the dynamic, goal-directed
character of postnatal behavior, the dimensions that de-
fine behavioral organization after birth are applicable to
the movements expressed by fetuses. Fetal activity ex-
hibits temporal, sequential, and spatial organization that
emerges between the inception of movement and term.
Fetal rodents, for example, exhibit coordinated motor
patterns antecedent to postnatal righting, locomotion,
suckling, maternal-infant communication and grooming
HISTORIC OVERVIEW OF FETAL
RESEARCH
Every pregnant mother knows that the fetus she
carries moves before birth. The fact of prenatal
movement has been known to science at least since
the writings of Aristotle. Hieronymus Fabricius
( 1604; cited in Carmichael, 1954), William Har-
vey ( 165 1 ) , Karl Ernst von Baer ( 1828), and other
early pioneers of biology remarked on the move-
ments of vertebrate embryos or drew inferences
about the prenatal formation of “the sensitive
soul,” to phrase the question in Aristotelian terms.
Bichat ( 1827), who persisted in the peripatetic tra-
dition in the 19th century, concluded from his fetal
observations that the fetus subsists in a vegetative
existence devoid of sense experience or function.
This conclusion was echoed a half-century later in
Received August 12, 1992;accepted August 12, 1992
Journal of Neurobiology, Vol. 23, No. 10, pp. 1574-1600
(1992)
0 1992 John Wiley & Sons, Inc.
CCC 0022-3034/92/101514-27
* To whom correspondence should be addressed.
1574
behavior, while other action patterns appear to be func-
tional adaptations to the intrauterine niche. Fetuses also
are behaviorally responsive to sensory stimulation and
changes in environmental conditions in utero. Expres-
sion of these behavioral properties emphasizes continu-
ity between prenatal and postnatal life while implying an
adaptive role for hehavior before birth. o 1992 John Wiiey
& Sons, Inc.
Key words: prenatal behavior, rodent fetus, cyclicity,
chemosensation, action pattern, ontogenetic adaptation,
heterochrony .
William James’s famous allusion to the “bloom-
ing, buzzing confusion” of the human newborn
(James, 1890) and continues to be repeated in
modern biology and psychology textbooks. In-
deed, a comparison of passages penned by Aristotle
and James more than 22 centuries apart illustrates
the unity of scientific viewpoint in addressing the
issues of behavior and sensation before birth:
As soon then as the offspring of all animals are born,
especially those born imperfect, they are in the habit of
sleeping, because they continue sleeping also within the
mother when they first acquire sensation. . . . But never-
theless they are found to wake even in the womb (this is
clear in dissections and ovipara), and then they immedi-
ately fall into a sleep again. This is why after birth also
they spend most of their time in sleep. (Aristotle, De
Generatione Anmalium, p. 32 1 )
Prior to all impressions on sense-organs the brain is
plunged in deep sleep and consciousness is practically
non-existent. Even the first weeks after birth are passed
in almost unbroken sleep by human infants. It takes a
strong message from the sense-organs to break this
slumber. (James, 1890, p. 456)
Despite this early history (or perhaps because of
the negative conclusions), questions regarding the

movement and habits of the fetus have received
scant attention outside of folklore and popular be-
lief until modern times. The field of behavioral em-
bryology has remained to the present somewhat
apart from the mainstream of both developmental
and behavioral science.
The first modern attempts to investigate the pre-
natal development of behavior are in general attrib-
uted to Wilhelm Preyer, who summarized his fetal
research in an extensive monograph published just
over a century ago ( 1885). Preyer’s novel ques-
tions and observations encouraged a later genera-
tion of neuroanatomists to develop rudimentary
techniques for observing fetuses in vivo (Brown,
1915; Swenson, 1926; Avery, 1928). The field of
behavioral embryology flourished in the 1920s and
30s in the laboratories of Coghill (Swenson, 1926;
Coghill, 1929; Angulo y Gonzalez, 1932), Windie
( Windle and Griffin, 1931 ; Windle, O’Donnell,
and Glasshagler, 1933; Windle, Minear, Austin,
and Orr, 1935), Carmichael (Coronios, 1933; Car-
michael, 1934;Carmichael and Smith, 1939), and
Barcroft ( Barcroft, Barron, and Windle, 1936;Bar-
croft and Barron, 1939), among others. This flurry
of research was motivated by an interest in the ori-
gins of reflexes and neuromuscular development.
Prenatal reflexology was seen principally as a use-
ful method for identifying stages of neural develop-
ment (see reviews by Hooker, 1952; Carmichael,
1954).
The early emphasis on fetal reflexes as a diag-
nostic tool probably reinforced the impression that
fetal motility lacks intrinsic coordination or spatio-
temporal organization. This impression, stressed
in many of the classic studies, fostered a traditional
viewpoint that prenatal motor activity is random,
directionless, and without purpose; fetal move-
ment is a mere epiphenomenon of the early devel-
opment of the nervous system. This perspective,
which continues to be expressed in current develop-
mental research (Stelzner, 1986), is well illustrated
by the following quotation by Hamburger ( 1973,
p. 53):
The phenomenon of uncoordinated motility which,
moreover, does not involve sensory input to any extent,
may not seem to be of much interest to developmental
psychobiologists. . . . It has been objected that our
claim that the combinations of movements in Type I
motility are uncoordinated has not been subjected to a
rigorous statistical analysis. It would be desirable, in-
deed, to make the recording method more objective and
to do a correlation analysis of sequence of movements,
perhaps using slow-motion moving pictures. However,
even if such analysis would reveal a low degree of corre-
lation between those particular types of movements that
Mammalian Fetal Motor Activity 1575
become integrated later, this would not necessitate a re-
vision of my basic position . . .
For reasons of technical simplicity, this neuro-
logical perspective became increasingly focussed
on avian models of prenatal behavior (see reviews
by Hamburger, 1963; Kuo, 1967; Bekoff, 1981;
Oppenheim, 1981, 1982) to the virtual exclusion
of fetuses: Between 1945 and 1980, the behavior of
mammalian fetuses received little research atten-
tion (cf. Narayanan, Fox, and Hamburger, 1971 ) .
In addition, the focus on evoked reflexes and neuro-
anatomy diverted attention from the ontogeny of
nonevoked fetal activity and functional behavior.
The viewpoint that the fetus is a passive pas-
senger during its residence in utero thus has had a
long and rich history. Today, however, this view-
point is undergoing rapid change due in part to
renewed interest in fetal research within a diversity
of disciplines ranging from psychobiology and
ethology to teratology and pediatric medicine and
in part to the influence of improved technical, ana-
lytic, and conceptual tools with which to study be-
havior. Within the past decade, research concerned
with the patterning of prenatal behavior has discov-
ered temporal rhythms in fetal activity (Robert-
son, 1985), the emergence of motor coordination
before birth (Bekoff and Lau, 1980; Robinson and
Smotherman, 1987, 1988), and the prenatal devel-
opment of species-typicalaction patterns ( Smoth-
erman and Robinson, 1987a). Research on sen-
sory development in human and nonhuman fe-
tuses has demonstrated prenatal chemosensation
(Pedersen, Greer, and Shepherd, 1986; Schaal,
1988; Smotherman and Robinson, 1990a, 1990b)
and audition in utero (Birnholz and Benecerraf,
1983;Geubelle, 1984; Fifer and Moon, 1988). Re-
search on the importance of experience during pre-
natal life has documented the ability of fetuses to
respond to changes in the intrauterine environ-
ment (Smotherman and Robinson, 1986), learn
by association of stimuli (Smotherman and Robin-
son, 1985, 1987b), and retain prenatal experiences
into postnatal life (Smotherman, 1982; Stickrod,
Kimble, and Smotherman, 1982; Hepper, 1988).
It has become manifest that the behavior of mam-
mals has its roots in the prenatal period and a com-
plete understanding of behavioral development
must include investigation of the behavioral capa-
bilities of the fetus.
METHODS OF FETAL RESEARCH
Much of the recent growth of interest in the fetus
stems from new or improved technologies for gain-
1576 Robinson and Smothermun
ing access to fetal behavior. Two important technol-
ogies have come to be routinely used in clinical
applications in the past 10-20 years. External fetal
monitoring, which involves little more than the
placement of a sensory transducer around the ab-
domen of a pregnant woman, enables the record-
ing of fetal heart rate and detection of gross fetal
body movements. More precise information about
the kinds of movements exhibited by the fetus is
provided by real-time ultrasonography, in which
the echoes of high-frequency sound pulses are used
to define contours and surfaces within the body
and thereby produce a video image of the fetus as it
moves in utero (Birnholz, 1988). These two inno-
vations have created a window on human fetal de-
velopment that has replaced earlier approaches,
such as examination of aborted fetuses (Hum-
phrey, 1953). However, ethical considerations
limit the application of true experimental designs
to the study of human fetuses. Moreover, the
current generation of noninvasive imaging technol-
ogy does not provide sufficiently detailed informa-
tion about the form and patterning of fetal move-
ments. Creation of continuous, fine-grained re-
cords of fetal behavior continue to require direct
observation of the fetus. Therefore, comparative
study of nonhuman animals remains centrally im-
portant to understanding the prenatal develop-
ment of behavior.
In our laboratory, we employ surgical and exper-
imental procedures that permit direct observation
of rodent fetuses from the inception of movement
(the earliest age at which motor activity is ex-
pressed; E 16 in the rat) through term (the last day
of gestation; E2 1 in the rat). Briefly, these proce-
dures involve ( 1 ) chemical blockade of the spinal
cord that desensitizes the pregnant female without
rendering fetuses immobile through general anes-
thesia, ( 2 ) surgical externalization of the uterus
into a benign fluid medium to bring fetuses into
direct view, and ( 3 ) maintenance of the umbilical
cord and placental connection to the uterus to en-
sure the viability of fetal subjects throughout the
period of observation. Additional techniques are
employed to provide controlled manipulation of
the fetal central nervous system (CNS) or of the
physical and sensory environment of the fetus.
A technical ability to view the fetus is incom-
plete, however, without a correspondingly precise
method of description and quantification of both
nonevoked activity (motor behavior that occurs
without explicit stimulation by the experimenter)
and evoked responses (behavior that occurs in asso-
ciation with the presentation of sensory stimuli).
In most experiments, we employ a system for
quantifying fetal movements that defines discrete
categories of behavior in five basic body regions:
forelimb (F), rearlimb ( R ) , head ( H ) , trunk ( T ) ,
and mouth ( M ). These categories can occur in iso-
lation (simple movements) or in combination ( syn-
chronous or concurrent movements). Videotape
records of observation sessions augment this re-
cording scheme and provide permanent records
containing more detailed information about the to-
pography and patterning of fetal behavior (Robin-
son and Smotherman, 1991a). The result is a reli-
able, continuous, fine-grained behavioral record
that preserves information about the distribution
and abundance of fetal movements in time, se-
quence, and space. Detailed descriptions of these
technical and observational procedures may be
found in Smotherman and Robinson ( 199 1 ) .
Because these technical and observational pro-
cedures provide the experimenter with direct physi-
cal access to the fetus, it is possible to measure fetal
responses to controlled sensory stimulation in che-
mosensory or somatosensory modalities. Chemo-
sensation encompasses a number of sensory sys-
tems that develop during the perinatal period, in-
cluding olfactory epithelia associated with both
main and accessory (vomeronasal) olfactory sys-
tems (Pedersen et al., 1986), lingual taste buds
(Mistretta and Bradley, 1986), pharyngeal and
epiglottal taste buds (Travers, Travers, and Nor-
gren, 1987), and the trigeminal system (Silver and
Moulton, 1982). Chemical cues play an important
role in directing suckling behavior, maternal-in-
fant interaction, and kin recognition during the
neonatal period, so it is reasonable to expect fetal
rats to be responsive to a variety of chemical stim-
uli in utero. Research on fetal chemosensation dur-
ing the past decade represents a distinct shift of
emphasis from earlier studies, which focused on
reflexive responses to somatosensory stimulation
(Carmichael, I954; Humphrey, 1969; Narayanan
et al., 197 1 ) . Chemosensory stimulation has
proven more effective than tactile cues in evoking
organized patterns of fetal response. For instance,
we identified different classes of chemical solutions
that reliably elicit species-typical behavior in the
rat fetus (e.g., the facial wiping response to strong
odorants and the stretch response to milk). The
combined use of quantitative analytic techniques
and chemosensory stimulation has promoted the
growing appreciation of the complexity and im-

portance of behavior during prenatal development
( Smotherman and Robinson, 1 9 9 0 ~ ) .
TEMPORAL PATTERNING OF FETAL
MOTOR ACTIVITY
An important aspect of the organization of behav-
ior is its distribution in time. Early studies of fetal
behavior included subjective comments about the
tendency for fetal movements to wax and wane
with time ( Angulo y Gonzalez, 1932; Carmichael,
1934). Qualitative characterization of periodic
motor activity in chick embryos and in the pattern-
ing of spontaneous neural activity within embry-
onic spinal cord segments also has provided impor-
tant clues to the neurogenic nature of prenatal
movement (Hamburger, Winger, and Oppenheim,
1966; Provine, 1986). However, a crucial aspect of
objectively characterizing temporal patterns in
nonevoked behavior is distinguishing apparent
patterns from random variations in fetal activity.
Chance is a recurrent theme in both classic and
modern discussions of behavioral embryology,but
only recently have subjective notions of random-
ness been explicitly evaluated by quantitative
means (Robertson, 1985; Robinson and Smother-
man, 1988).
Cyclic Motor Organization
Spontaneously generated rhythmic or cyclic pat-
terns of motor activity are fundamental to virtually
all animals ( Aschoff, I98 1 ) . Many behavioral
rhythms correspond to naturally occurring envi-
ronmental cycles, such as solar, lunar, or seasonal
cycles. Ultradian rhythms (with periods less than
24 h ) from another class of temporal organization
that has been widely documented in the non-
evoked activity of animals. Cyclic organization
with a period of a few minutes has been reported in
the motor activity of human infants during both
active sleep and nonsleep states (Robertson,
1987). Nearly identical patterns of cyclic motility
are evident in human fetuses during the last half of
gestation, suggestingcontinuity in the temporal or-
ganization of general motor behavior before and
after birth (Robertson, 1985, 1988).
Recent application of quantitative methods for
assessing cyclicity have confirmed the existence of
cyclic motor activity in the E20 rat fetus (Smother-
man, Robertson, and Robinson, 1988). Spontane-
ous oscillation is evident not only in overall fetal
Mammalian Fetal Motor Activity 1577
activity but also in the movements of individual
parts of the body. After communication between
rostral and caudal portions of the spinal cord is
interrupted by midthoracic transection, rearlimbs
remain active and exhibit cyclic organization
nearly identical to that of intact fetuses. Body parts
rostral to the transection also continue to exhibit
cyclic activity, but the rate of oscillation decreases
relative to intact fetuses (Robertson and Smother-
man, 1990). These findings demonstrate that iso-
lated fragments of the fetal CNS can support cyclic
motor activity. Therefore, this form of temporal
patterning of fetal movement is not limited to a
single neural source (cf. circadian cyclicity regu-
lated by the suprachiasmatic nucleus in the fetus,
Reppert and Weaver, 1988) and may be a property
that emerges from the interaction of components
within the motor system (Robertson, 1990). Sec-
ond, these data suggest that the interaction that oc-
curs between different segments of the nervous sys-
tem in the intact fetus involves dominance of more
caudal elements over rostral elements.
Motor Synchrony
The first movements of the fetus involve simple
movements of individual parts of the body, such as
the head, limbs, or trunk (Narayanan et al., 1971;
Smotherman and Robinson, 1986). Fetuses also
move different parts at the same time; concurrent
or simultaneous movement of two or more parts of
the body (also referred to as “motor synchrony”) is
one of the earliest forms of coordination to appear
during motor ontogeny (Provine, 1980; Robinson
and Smotherman, 1987).
The simultaneous movement of multiple body
parts seems to imply the central coordination of
motor behavior. It is possible, however, for two in-
dependent movements to occur at the same time
by coincidence. Especially in immature animals
that exhibit imperfectly coordinated behavior, the
null hypothesis that motor synchrony arises
through the chance association of independent
events must be considered. To objectively deter-
mine whether an observed pattern results from the
chance association of independent processes or
from central coordination, it is necessary to com-
pare the occurrence of the pattern to an explicit
standard. An example of this research approach is
provided by Provine ( 1980), who documented the
emergence of interning coordination in chick em-
bryos by comparing the incidence of synchronous
movements involving both wings with a presumed
1578 Robinson and Smotherman
random process, the incidence of synchronous
movements involving one wing and one leg. Early
during incubation, wing-wing movements are no
more frequent than wing-leg movements, suggest-
ing that synchronized limb movements in young
chick embryos is accidental. Later, wing-wing
movements increase in frequency as wing-leg
movements become relatively uncommon, indi-
cating the development of central coordination of
bilateral wing behavior.
Formal stochastic models also have been used to
generate testable predictions about the incidence of
motor synchrony in rat fetuses (Robinson and
Smotherman, 1987, 1988). Predictions regarding
the rate of concurrent movement are obtained by
considering each instance of fetal movement in-
volving a single body part as an independent point
event generated by a Poisson process. Because
most fetal movements have a brief but measurable
duration (less than 1 s), independent movements
generated by single or dual Poisson processes that
are initiated at nearly the same instant would over-
lap and appear synchronized. The parameters of
the stochastic processes, estimated from the overall
frequency of fetal movement and the number of
1-s time samples in the observation session, are
used to generate the expected frequency of appar-
ently coincident movement events (Robinson,
1989).
Stochastic models that incorporate single or
dual Poisson processes are quite accurate in pre-
dicting the overall incidence of motor synchrony
during the first 2 days after inception of movement
in the fetal rat (EL6-E17), but indicate that syn-
chronous movements occur more frequently than
should be expected by chance association during
the last 3-4 days of gestation (E 18-E2 1 ) . Related
models have addressed the more detailed question
of whether specific patterns of motor synchrony,
such as the concurrent movement of forelimbs and
rearlimbs, occur more often than expected on the
basis of random association. In the rat fetus, most
of the possible combinations of simple movements
occur at rates that do not differ from predictions
generated by stochastic models. The occurrence of
certain categories of motor synchrony, however,
cannot be accounted by the chance coincidence of
independent events. Concurrent movements in-
volving the forelimbs and head (FH) are the earli-
est to exceed the level of chance occurrence during
prenatal development. FH synchrony is evident as
early as E 17 (only 1 day after inception of move-
ment) and continues to be expressed through term
(E2 1) . Concurrent movement of forelimbs and
rearlimbs (FR) becomes apparent later during ges-
tation. Although the head and rearlimbs are in-
volved in synchronous activity (FH and FR), they
are associated together (HR) much less frequently
than should occur by chance. These findings sug-
gest that FH and FR coordination emerges during
prenatal development at the same time HR
synchrony is actively suppressed.
Experiments employing spinally transected ani-
mals have confirmed that intersegmental neural
communication is necessary for the production of
concurrent head and limb movements. Although
most studies concerned with the development and
coordination of behavior focused on the role of the
brain, complex motor patterns can be supported
by spinal cord circuitry alone (Berkinblit, Feld-
man, and Fukson, 1986). Spinal transection pro-
cedures have been used to probe the neural control
of motor synchrony in rat fetuses from E l 8
through term. After transection of the spinal cord
at the midthoracic level, rat fetuses continue to ex-
hibit activity in forelimbs (rostra1 to the cut) and
rearlimbs (caudal to the cut). Concurrent move-
ments involving the forelimbs and head ( F H ) also
continue to be expressed. But, concurrent move-
ments involving the forelimbs and rearlimbs ( F R ) ,
ordinarily one of the most common categories of
motor synchrony on E20, are virtually eliminated.
Conversely, fetuses with cervical spinal transec-
tions exhibit forelimb and rearlimb activity and
FR synchrony, but concurrent FH movements are
abolished. These results permit two general con-
clusions about the control of motor synchrony in
rat fetuses. First, the sparing of motor activity cau-
dal to the site of spinal transection confirms other
reports (such as the experiments on cyclic motor
organization, described above) that neural cir-
cuitry within the spinal cord alone is sufficient to
initiate fetal motor activity. Second, the near elimi-
nation of FH synchrony after cervical transection
and FR synchrony after thoracic transection, and
the sparing of FR synchrony after cervical transec-
tion, indicates that intersegmental communication
within the spinal cord is necessary to support the
occurrence of motor synchrony in the fetus.
Similar patterns in the development of motor
synchrony are expressed by cotton rat fetuses. Un-
like most other members of the rodent family Mur-
idae, which give birth to altricial offspring that pos-
sess limited sensory and motor abilities at the time
of birth, cotton rats (Sigmodon hispidus)give birth
to precocial young that are furred, open their eyes
within 24 h, and exhibit relatively adultlike motor
abilities. Because cotton rats are precocial, the pro-

duction of FH and FR synchrony at levels above
chance occurs relatively earlier during gestation,
but FH movements still develop slightly in ad-
vance of FR movements. Concurrent movements
involving the head and rearlimbs (HR) also be-
come suppressed at a relatively early prenatal age
(E2 1 of the 27-day gestation) and rarely occur for
the remainder of gestation. In precocial fetuses,
more complex patterns of motor synchrony also
appear during the prenatal period, including the
simultaneous activity of forelimbs, head, and rear-
limbs (FHR). Initially, FHR movements occur at
chance levels, then diminish in frequency in paral-
lel with the reduction in HR activity. By E23, FHR
movements occur significantly less often than pre-
dicted by random association. Over the next 2 days
of gestation, however, FHR activity increases
sharply, exceeding rates generated by the stochastic
model. This indicates that coordination of more
complicated patterns of movement involving three
or more body parts becomes evident during the
prenatal period in precocial rodents, an ontoge-
netic pattern that has been described during the
neonatal period in altricial species such as the rat
(Blanck, Hard, and Larsson, 1967).
Despite the prevalence of attempts to establish
rigorous time courses for the ontogenetic emer-
gence of gross motor abilities (e.g., Tilney, 1933;
Bolles and Woods, 1964; Altman and Sudarshan,
1975), the likelihood for young animals to express
complex motor patterns can be heavily influenced
by environmental conditions present at the time of
behavioral testing. For instance, when separated
from the mother and placed in an open field infant
rats tend to exhibit simple forms of activity involv-
ing unilateral or bilateral movements of the head.
The tendency for pups to exhibit more complex
patterns involving simultaneous movement of
head and limbs gradually increases over the first
postnatal week under these conditions (Eilam and
Golani, 1988). However, if exposed to ammonia
odor while in an open field newborn rats engage in
high levels of activity involving synchronized
movements of the head, forelimbs, and rearlimbs
(Gard, Hard, Larsson, and Peterson, 1967). Envi-
ronmental stimuli can modify the expression of
motor behavior in young animals and thereby alter
the timetable of motor development inferred from
experimental data. The concept of a timetable of
motor development therefore has limited utility
for any age where subjects are responsive to
changes in environmental conditions.
The tendency of fetuses to express motor
synchrony also is sensitive to features of the prena-
Mammalian Fetal Motor Activity 1579
tal environment (Smotherman and Robinson,
1986;Robinson and Smotherman, 1987).Concur-
rent movements by the rat fetus increase in relative
abundance during the first days after inception of
movement. Motor synchrony remains high
through term in fetuses observed outside the uterus
and embryonic membranes (exutero) . But, fetuses
observed in utero through the transparent wall of
the uterus and embryonic membranes exhibit a
sharp decrease in the incidence of concurrent
movement late in gestation (Fig. 1 ). This reduc-
tion in motor synchrony coincides with the reduc-
tion in amniotic fluid volume that occurs before
parturition in the rat and other rodents (Marsh,
King, and Becker, 1963). Coupled with the rapid
growth of the fetus during late gestation, the conse-
quence of reduced amniotic fluid is a dramatic re-
duction in free space within the uterus. Under such
conditions of increased physical restraint in utero,
complex motor behavior is suppressed (Smother-
man and Robinson, 1988a).
The influence of environmental constraint on
motor synchrony is a general feature of prenatal
behavior, as demonstrated in comparative studies
of different rodent species that vary in patterns of
development (Robinson, 1989). Cotton rat fetuses
show a decline in amniotic fluid volume immedi-
ately before parturition. They also exhibit a higher
incidence of motor synchrony ex utero than in
utero. But, differencesin their expression of concur-
rent movements inside and outside the uterus are
evident as early as 1 week before birth (Fig. 1).
These comparative findings suggest that precocial
fetuses can express responsiveness to environmen-
tal conditions in utero well in advance of the disap-
pearance of amniotic fluid.
ACTION PATTERNS OF THE FETUS
The development of cyclic motor activity and mo-
tor synchrony are just two examples of prenatal
behavioral organization. Fetal behavior also ex-
hibits a tendency to occur in discrete bouts and
predictable sequences during the latter portion of
gestation. Basic patterns of fetal motor activity
may well provide the foundation upon which more
complicated patterns of behavior are constructed
( Robinson and Smotherman, 1988). This concep-
tion of behavioral development is consistent with
the views expressed by Gottlieb and Kuo ( 1965, p.
187) that “the ontogeny of behavior is gradual and
continuous and that sequences of action occurring
in embryo are significant precursors to the ‘pre-
1580 Robinson and Smotherman

40
Norway Rat
+ In Utero
-0- Ex Utero 7-
20

10

El7 E l8 El9 E20 E21

40

30

20

10

I I I I I I

E l 8 E l 9 E20 E21 E22 E23 E24 E25 E26 E27

Gestational Age
Figure 1 Relative abundance of concurrent movements expressed by rat fetuses (top) and
cotton rat fetuses (bottom) observed in utero or ex utero. Relative concurrent movement is
calculated as the frequency of events comprising two or more movement categories (e.g., F and
R ) divided by the total number of movement events in the observation session. Points repre-
sent means; vertical lines depict SEM.

adapted' quality of action patterns in postnatal
life." Indeed, a growing list of postnatal action pat-
terns has been identified in the nonevoked or
evoked activity of rodent fetuses, including compo-
nents of contact righting, stepping, punting, suck-
ling, facilitation of maternal licking, and grooming
behavior.
Righting
The righting reflex is a fundamental behavior of
tetrapods that involves actively returning the body
to a prone posture after displacement. The ability
to perform righting and the latency to return to a
prone posture are commonly used as general in-
dexes of the relative maturity of the motor system
(Tihey, 1933; Volokhov, 1968; Altman and Su-
darshan, 1975;Almli and Fisher, 1977). However,
ontogenetic changes in the form and neural control
of righting suggest that turning to a prone posture
may embrace a number of functionally related yet
distinct patterns of motor behavior (Pellis, Pellis,
and Tetelbaum, 1991 ). The earliest form of right-
ing to appear during development is contact right-
ing (as distinct from air righting, which requires a
vestibular sense), in which the body is returned to
a prone posture while in contact with a substrate.
Experiments involving transection at various levels
of the neuraxis indicate that contact righting can be
performed by newborn rats with only the lower

medulla and spinal cord intact, but 7-8 days after
birth rats require midbrain structures for righting
to occur (Bignall, 1974). The development of con-
tact righting has long been appreciated to have
roots in the prenatal period (Windle and Fish,
1932), in particular among species that give birth
to precocial offspring (Cannichael, 1934, 1954).
A principal component of the righting response
of neonatal rodents is ventriflexion or torsion of
the body trunk around its long axis (Altman and
Sudarshan, 1975; Pellis et al., 1991). Slow-motion
playback of videotape records of fetal activity has
revealed that rodent fetuses exhibit trunk torsion
that closely resembles the movements involved in
contact righting. Trunk torsion occurs uncom-
monly during nonevoked activity in rat fetuses as
early as E20. It is expressed much more commonly
in cotton rat fetuses and appears relatively earlier
during gestation, about E2 1. Trunk torsion rarely
results in successfully turning to a prone posture in
rat fetuses, which are observed under conditions in
which the fetus is not supported by a solid sub-
strate. However, turning to a prone position is evi-
dent in fetal cotton rats observed under similar
conditions as early as E22 and is expressed by a
majority of subjects by E24 of the 27-day gestation.
These observations suggest a perinatal emergence
of the rudimentary motor patterns involved in
contact righting in altricial species and a prenatal
emergence of this behavior in precocial species.
Stepping
Another pattern of behavior expressed to varying
degrees by young mammals is the cycle of stepping
movements involved in walking or running loco-
motion. Mature walking requires coordination of
stepping movements by all four limbs in a charac-
teristic pattern of foot contact (Halbertsma, 1983).
In young animals, interlimb coordination is first
apparent within a girdle: The two opposing limbs
move in regular alternation, with each shifted 50%
in phase relative to the other limb. Thus, the left
limb remains in contact with a substrate (the
stance phase) as the right limb is flexed and moved
forward (the swing phase). Adult animals that re-
ceive brainstem or spinal cord transection retain
many aspects of normal walking locomotion (Shik
and Orlovsky, 1976; Grillner, 198 1 ) . For example,
cats with a brainstem transection can be supported
in a sling to reduce the gravitational load on the
limbs and eliminate postural and balance demands
on the motor system. Under these experimental
conditions, tonic electrical stimulation of reticulo-
Mammalian Fetal Motor Activity 1581
spinal tracts posterior to the transection elicits
stepping movements of forelimbs and rearlimbs
strikingly similar to walking movements by intact
cats. Similar results may be obtained by injection
of drugs, such as L-dopa, that stimulate dopaminer-
gic and noradrenergic receptor systems within the
spinal cord (Kellogg and Lundborg, 1972;
Grillner, 1981). Cats with a low thoracic spinal
transection also are capable of producing rearlimb
stepping movements on a treadmill or while sus-
pended in the air (air stepping). Spontaneous air
stepping occurs rarely in adult animals but is exhib-
ited by kittens that receive low-spinal transection
shortly after birth (Bradley and Smith, 1988).
These findings indicate that newborn mammals
possess sufficient neural circuitry within the spinal
cord alone to initiate stepping movements and
maintain intra- and interlimb coordination closely
resembling the form of intact walking movements.
The alternating movement of left and right
limbs within a girdle that characterizes the synergy
of the step cycle was originally reported in the rat
fetus by Brown ( 1915) and more recently has been
objectively described on E20 (Bekoff and Lau,
1980). Alternating stepping is not often seen dur-
ing nonevoked activity in fetal or neonatal rats.
However, air stepping that involves coordinated
movement of all four limbs can be evoked reliably
in rat pups soon after birth by administration of
L-dopa (Van Hartesveldt, Sickles, Porter, and Ste-
houwer, 199 1 ). Preliminary experiments in our
laboratory suggest that L-dopa injection also stimu-
lates coordinated stepping movements by rat fe-
tuses that are externalized from the uterus (ex
utero) but remain submerged within the saline
bath. In contrast to the altricial fetuses of rats, pre-
cocial fetuses such as cotton rats often engage in
alternating stepping movements involving two
limbs within a girdle (pectoral or pelvic) or, less
commonly, all four limbs. Walking and running
movements may be expressed during periods of
nonevoked activity or in response to sensory stimu-
lation in cotton rats during the last week of gesta-
tion (E24-E27).
Punting
Quadrupedal stepping associated with crawling
and walking is not the earliest form of locomotion
normally expressed by neonatal rats ( Altman and
Sudarshan, 1975). During the first week after
birth, most locomotor activity takes the form of
pivoting, in which the rear part of the pup is an-
chored to the substrate and the front part moves in
1582 Robinson and Smotherman
a circular path around this pivot point (Eilam and
Golani, 1988) . Early pivoting involves a distinc-
tive motor pattern, called punting, that does not
resemble the alternated limb coordination of the
step cycle. Punting is accomplished in a recogniz-
able sequence of movements, in which ( 1) the
head is turned and placed over the ipsiversive fore-
limb, in the direction of pivoting, (2) the ipsiver-
sive foreleg is withdrawn from under the head and
replaced in the direction of the pivot, and ( 3 ) as the
head and upper torso continue to turn the contra-
versive forelimb is extended and swept backward
and to the side, pushing the pup through the pivot
(Smotherman and Robinson, 1 9 9 0 ~ )Throughout
. followed by brief disengagement from the nipple.
this sequence, the rearlimbs remain stationary or
make small positional adjustments as the pup
turns. In this way, the rearlimbs probably help the
pup maintain a prone position throughout the
movements of punting.
The forelimb and head synergy involved in
punting is performed during nonevoked activity in
rat fetuses. On E 19-E2 1, fetuses exhibit occasional
bouts of activity in which the head is turned lat-
erally over the ipsiversive forelimb as the contra-
versive forelimb is swept to the side and backward.
Although this motor pattern is not performed in
relation to a substrate, and therefore does not result
in a change in fetal position, it is isomorphic to the
pattern of head and forelimb activity expressed
during neonatal punting. It is interesting to note
that close inspection of videotape records has failed
to identify this pattern of movement at any point
during gestation in cotton rat fetuses (Robinson
and Smotherman, 1990) . Because punting is appar-
ently absent during the prenatal period in cotton
rats, and is not expressed by the precocial offspring
ofthis species, it may not be a necessary stage in the
ontogeny of mature quadrupedal locomotion (cf.,
Eilam and Golani, 1988). Punting therefore may
represent a behavioral adaptation specially suited
to the needs and motor abilities of infant rats and
other altricial rodents.
Stretch Response and Behavior
Associated with Suckling
After birth, the newborn rat pup initiates a se-
quence of activities that culminates in ingestion of
milk (Drewett, Statham, and Wakerley, 1974;
Blass and Teicher, 1980). This sequence begins
with orientation to the lactating female’s ventral
surface, side-to-side searching movements of the
head (rooting), and attachment to the nipple. A
period of organized, rhythmic sucking behavior,
associated with other pup behavior such as paw-
treading, triggers a milk let-down response in the
female. During let-down, milk is ejected from the
nipple into the pharynx of the pup and swallowed.
Rat pups consistently exhibit behavioral respon-
siveness to milk ejection, including mouthing, lo-
comotion, and general behavioral activation. Fur-
ther, pups express a stereotypic action pattern that
occurs only in the context of milk ejection at the
nipple-the stretch response (Hall, 1979; Lau and
Henning, 1985). During the stretch, the pup ex-
tends all limbs, the back becomes rigid, and the jaw
is opened. Performance of the stretch response is
Performance of the stretch response by newborn
rats during their first suckling episode indicates
that the mechanisms governing this behavior are
intact at the time of birth. Delivery of a single infu-
sion of milk into the mouth of the E20 rat fetus
reliably elicits a modest increase in overall motor
activity. An important component of the fetal re-
sponse to milk is the expression of a pattern of be-
havior that appears isomorphic with the neonatal
stretch response (Robinson and Smotherman,
I992a). The occurrence of the stretch is delayed by
2-4 min following infusion.
In addition to the stretch response, fetal rats ex-
hibit other patterns of behavior associated with
suckling during the neonatal period. Lateral root-
ing movements of the head and forepaw move-
ments resembling paw-treading are occasionally
seen during nonevoked activity or following milk
infusion. Rooting activity is relatively common
during late gestation in precocial cotton rat fetuses,
in particular among fetuses in a prone posture on
the submerged surface of the uterus or maternal
abdomen. The identification of various compo-
nents of neonatal suckling behavior during the pre-
natal period provides a good example of the devel-
opment of behavior in anticipation of its expres-
sion in a functional context.
Leg Extension Response to Anogenital
Stimulation
Mother-infant interactions in laboratory rats are
symbiotic and involve the bidirectional exchange
of chemosensory, tactile and thermal information,
and material resources (such as body fluids) (Al-
berts and Cramer, 1988). Much of this material
and informational exchange takes place during
bouts of maternal licking. During licking bouts di-
rected to the anogenital region, pups exhibit a leg
extension response (LER), which involves eleva-

tion of the hindquarters and straightening of the
rearlimbs (Moore and Chadwick-Dias, 1986). The
LER probably functions to facilitate access by the
mother to the perineum, necessary to stimulate
micturition and defecation by the neonate. In rat
fetuses on E20 or E2 1 of gestation, the LER can be
elicited by repeatedly stroking the anogenital re-
gion with a soft camel-hair brush (Smotherman
and Robinson, 1988b). This response is spared in
fetuses following spinal transection at the midtho-
racic level, indicating that the LER is controlled by
neural circuitry located entirely within the caudal
spinal cord. Like the behavior associated with
suckling, the LER provides an example of prenatal
behavioral development in anticipation of neona-
tal function.
Response to Umbilical Cord Occlusion
One of the likely reasons that fetal behavioral adap-
tations have received little attention is the uncer-
tain relation of the fetus to its environment. Ma-
ture animals must seek food, shelter, and protec-
tion from predators and mates. But, the fetus
seems to have all its needs satisfied through a life-
support system consisting of the umbilical cord
and placenta. Maintenance of this life-support sys-
tem is crucial to fetal survival and growth. There-
fore, one of the most likely places to find behav-
ioral adaptations during the prenatal period is in
association with maintenance of a healthy umbili-
cal connection to the placenta.
Indeed, fetuses are at risk of transient occlusion
of blood circulation within the umbilical cord dur-
ing gestation. Acute fetal hypoxia, induced by um-
bilical cord compression, has been implicated as an
important cause of brain damage in human fetuses
(Mann, 1986) and can occur during unremarkable
pregnancies ( Itskovitz, LaGamma, and Rudolph,
1987). Accidental cord occlusion can occur by
twisting of the cord as a result of fetal activity or by
pinching the cord between the fetus and a hard part
of maternal anatomy (such as the pelvis) or, in
species that bear multiple offspring, against an ad-
jacent sibling in utero.
Previous investigators have reported observa-
tions that suggested the existence of a behavioral
response of fetuses to umbilical cord occlusion.
Many of the early studies of fetal motor develop-
ment conducted during the 1920s and 30s were
criticized by Windle ( 1944) as lacking in proper
controls to avoid the effects of fetal hypoxia; to
support his claim, Windle described a predictable
sequence of behavioral changes that occur when
Mammalian Fetul Motor Activity 1583
fetuses become anoxic. More recently, descriptions
of a brief period of fetal hyperactivity following
transient hypoxia have been reported in rats
(Becker, King, Marsh, and Wyrick, 1964) and
dogs ( Arshavsky, Arshavaskya, and Praznikov,
1976).
Research conducted in our laboratory con-
firmed and extended these previous reports in a
series of quantitative experimental investigations
of the response of rat fetuses to transient occlusion
of the umbilical cord (Smotherman and Robinson,
I987c, 1 9 8 8 ~ )Fetuses
. exhibit a three-phase behav-
ioral response that follows placement of a microvas-
cular clamp on the umbilical cord, which com-
pletely occludes umbilical circulation. This re-
sponse consists of ( 1 ) initial suppression of fetal
activity, ( 2 ) an intermediate phase of fetal hyperac-
tivity, in which fetal movements increase four- to
fivefold over baseline levels, and ( 3 ) a final phase
of behavioral suppression in which fetal activity
drops to near zero levels until removal ofthe umbil-
ical cord clamp. The movements that comprise
this hyperactivity are stereotyped and consist pri-
marily of lateral flexion or bending of the body
trunk, often in association with rostra1 extensions
of the head ( “head-tossing” ) and bilateral exten-
sion of the forelimbs. Rat fetuses also exhibit an
immediate and pronounced bradycardia following
umbilical cord occlusion, which occurs during the
initial suppression of motor activity. However, the
critical feature of the clamp response is the second
phase-hyperactivity-which distinguishes it as
an active behavioral response to cord occlusion
and not an incidental effect of reduced oxygen
availability to tissues. The expression of a stereo-
typed behavioral response to umbilical cord occlu-
sion in other rodent species suggests that this is a
general feature of prenatal behavior (Robinson
and Smotherman, 199 1b ) .
Facial Wiping and Other Grooming
Behavior
Movements associated with grooming behavior are
common to most mammals that retain a general-
ized limb anatomy. Marsupials, insectivores, ro-
dents, lagomorphs, and carnivores all employ the
forelimbs in grooming movements directed toward
the head and face (Eisenberg, 198 1 ). In most spe-
cies, the forepaws engage in sweeping overhead
strokes that sweep in contact over the ears and
sides of the face. Finer movements of the forepaws
are involved in paw-licking. These forelimb move-
ments are integrated with other grooming behav-
1584 Robinson and Smotherman
ior, such as nibbling, biting, or licking the fur and
scratching the head or body trunk with the rear-
paw, in complex, stereotyped behavioral sequences
(Richmond and Sachs, 1980; Fentress, 1988).
Components of adult grooming and aversion se-
quences have been identified in the behavior of fe-
tal rodents. Facial wiping is a stereotyped, species-
typical action pattern that involves forelimb move-
ment that closely resembles the overhead strokes
employed by adult rats during grooming bouts.
The form of facial wiping involves simultaneous
placement of the two forelimbs in contact with the
side of the head and sliding the limbs toward the
nose in contact with the face (see Fig. 1 in Smoth-
erman and Robinson, 1987a). Facial wiping can
be evoked consistently in rat fetuses on E20 and
E21 of gestation by infusing a small volume of a
novel chemosensory solution, such as lemon ex-
tract, into the mouth ofthe fetus. The wiping move-
ments that ensue occur a few seconds after infusion
in a brief flurry of 5- 15 strokes.
The facial wiping response of the fetal rat is a
clear example of coordinated behavior. However,
motion analysis of videotape records has revealed
that the high degree of organization evident in this
behavior is the culmination of organizational
changes that occur before the first wiping stroke, at
the initiation of the wiping bout, and as the bout of
wiping proceeds (Robinson and Smotherman
199 1a). Within seconds after infusion, E20 fetuses
exhibit an increase in overall motor activity char-
acterized by an increase in forelimb movements
and a selective suppression in rearlimb activity. As
facial wiping begins to be expressed, nondirected
forelimb movements are transformed into orga-
nized wiping strokes that contact the face. At the
same time, head movements are suppressed, with
the head stabilized at the onset of each wiping
stroke. Initial wiping strokes are brief and ran-
domly coordinated between right and left sides. As
more strokes are performed, paw-face contact in-
creases in duration and bilateral strokes become
more highly synchronized. The initiation of facial
wiping thus appears to be neither the beginning nor
the end of organizational changes in fetal behavior
evoked by chemosensory infusion. This finding
contrasts with conventional views of “central pat-
tern generation” of action patterns (Grillner,
1985), and suggests that facial wiping behavior
emerges from the coupling of simpler movement
components in real time (Robinson and Smother-
man, 1991a).
The wiping response to chemosensory infusion
is influenced by olfactory qualities of the stimulus;
fetuses exhibit the response following presentation
of stimuli in the gas phase ( Smotherman and Rob-
inson, 1990a) but are less likely to perform wiping
movements following surgical isolation ofthe olfac-
tory bulbs (Smotherman and Robinson, 1990b).
The wiping response also can be elicited by appli-
cation of a punctate tactile stimulus to the perioral
area of the fetus (Smotherman and Robinson,
1990b), an observation noted in guinea pig fetuses
more than 50 years ago (Carmichael and Smith,
1939). Together, these facts suggest that facial wip-
ing is a general behavioral response to novel or
aversive stimulation; olfactory and tactile stimuli
are sufficient to elicit wiping but other sensory sys-
tems, such as the trigeminal system, also are likely
to play a role in the control of the wiping response
(Berridge and Fentress, 1986).
Other components of adult grooming behavior
have not been reported in rat fetuses. Newborn rats
can perform scratching movements by the rearpaw
in response to tactile stimulation applied to the
body. This response is spared by midthoracic spi-
nal transection (Stelzner, 1986). In the fetal sheep,
which gives birth to precocial offspring after a ges-
tation of 150 days, the rearlimb scratching re-
sponse can be evoked as early as E70 ( Barcroft and
Barron, 1939). During nonevoked motor activity,
precocial fetuses of cotton rats on occasion express
rearlimb scratching movements, as well as fine fore-
limb behavior such as paw-licking. But, these mo-
tor patterns apparently occur in isolation and are
not integrated with other grooming movements,
such as facial wiping, in more complex sequences.
DETERMINANTS OF FACIAL WIPING:
A DISAPPEARING ACT
Developmental Expression
Facial wiping is an action pattern expressed by ro-
dent fetuses that has been extensively investigated
in our laboratory. Although the appearance of the
wiping response in the fetus is strikingly similar to
the forelimb-head strokes that occur in the context
of adult grooming behavior, facial wiping is not
expressed continuously from the time of its prena-
tal emergence through adulthood. Our initial stud-
ies described a sudden ontogenetic appearance of
this behavior pattern in rat fetuses; no fetuses per-
formed wiping on E l 9 but virtually all exhibited
wiping immediately after lemon infusion on E20
(Fig. 2; Smotherman and Robinson, 1987a). Wip-

0 3 100
.-S
a 80
60
40
3 20
cn
-
E l 9 E20 E21 PO
Figure 2 Expression of the facial wiping response to chemosensory stimulation during the
perinatal period in the laboratory rat.
ing remains a consistent response of term fetuses
(E2 1 ) , but caesarean-derived pups tested within 1
h of birth (PO), and therefore the same postcon-
ceptional age as E21 fetuses, are significantly less
likely to exhibit facial wiping (Smotherman and
Robinson, 1989). The expression of facial wiping
continues to decline and virtually disappears from
the neonatal rat’s response to lemon infusion after
the first postnatal day (PI; Fig. 2). Toward the end
of the second week after birth ( P 1 1 ) , facial wiping
reappears as a consistent response to chemosen-
sory stimulation ( Smotherman and Robinson,
1989; Johanson and Shapiro, 1986). The develop-
ment of the facial wiping response thus assumes a
pattern of prenatal appearance, neonatal disappear-
ance, and later postnatal reappearance.
The temporary disappearance of the wiping re-
sponse during the neonatal period is not a unique
example of reversible changes in motor develop-
ment. For instance, human neonates exhibit a
stepping reflex when the infant’s body is held and
the feet are placed in contact with a solid substrate.
The relationship of neonatal stepping to later walk-
ing has remained a point of controversy in develop-
mental psychology because the stepping reflex dis-
appears 1-2 months after birth and infants begin to
walk some 9-12 months later. Thelen and col-
leagues demonstrated that the disappearance of the
stepping reflex is due to environmental conditions
in which infants are tested, not to a programmed
schedule of motor development. Examination of
electromyogram (EMG) patterns during spontane-
ous limb movements by infants reveals an underly-
ing continuity in the organization of motor behav-
Mammalian Frtul Motor Activity 1585
PI P3 P5 P7 P9 P I 1
Age
ior during neonatal stepping, the kicking that oc-
curs when an infant is in a supine posture, and later
crawling and walking (Thelen and Fisher, 1982).
The apparent reason behind the disappearance of
neonatal stepping lies in the rapid postnatal in-
crease in leg mass, which surpassesthe infant’s mus-
cular strength in the legs during the first 2 months.
This hypothesis has been confirmed by experi-
ments in which infants are immersed to waist
depth in water, providing a buoyant medium that
reduces the gravitational load on the legs. Under
these conditions, 3-month-old infants continue to
exhibit stepping movements, even though they do
not exhibit stepping in a terrestrial environment
(Thelen, Fisher, and Ridley-Johnson, 1984).
The past few years have seen numerous addi-
tional examples that demonstrate the role of envi-
ronmental context in the expression of motor be-
havior. For example, the development of mono-
manual reaching by human infants ( Rochat and
Stacy, 1989) and terrestrial locomotion in anuran
amphibians (Stehouwer and Farel, 1984) can be
accelerated by manipulating environmental con-
text at the time of testing. Similarly, reinstating en-
vironmental features characteristic of early ontog-
eny can prolong the expression of perinatal behav-
ior patterns such as hatching movements by adult
chickens (Bekoff and Kauer, 1984) and suckling at
the nipple by adult rats (Pfister, Cramer, and Blass,
1986;Lichtman and Cramer, 1989). Experimental
manipulations such as these demonstrate the per-
vasive role of environmental context in the expres-
sion of species-typicalbehavior ( Smotherman and
Robinson, 1 9 9 0 ~ ) .
1586 Robinson and Smotherman
100
80
60
40
20
0
Bath
Figure 3 Facial wiping by rat fetuses on El9 following lemon infusion delivered by blunted
needle inserted into the mouth or by intraoral cannula.
Appearance of the Wiping Response
Environmental context also plays an important
role in the ontogenetic trajectory of facial wiping in
rodents. As mentioned above, our initial experi-
ments indicated an abrupt emergence of facial wip-
ing during prenatal development. Although deliv-
ery of a chemosensory solution (such as lemon ex-
tract) into the mouth of the fetus reliably elicits
facial wiping on E20 and E2 1, younger fetuses do
not exhibit this action pattern as part of their re-
sponse to infusion ( Smotherman and Robinson,
1987a). These findings have been replicated sev-
eral times in experiments involving externalization
of the fetal subject from the uterus and embryonic
membranes prior to stimulation. Many other
aspects of behavioral organization are promoted by
observing fetuses ex utero (see examples above).
However, it now appears that testing fetuses under
these environmental conditions may affect their
ability to express the facial wiping response.
The possibility that manipulation of environ-
mental context can promote the precocial expres-
sion of the wiping response has been confirmed by
experiments in which fetuses are tested after exter-
nalization from the uterus but within the embry-
onic membranes ( i n amnion). To permit delivery
of the lemon solution to a fetal subject in amnion, a
blunted hypodermic needle is inserted into the
mouth of the fetus. Under the moderate physical
restraint provided by the amnion and uterus at this
age, fetuses exhibit the facial wiping response to
infusion. Fetuses tested ex utero that receive a
lemon infusion by this method exhibit behavioral
activation but fail to express the wiping response
(Fig. 3 ) . More recently, we developed methods for
Amnion
cannulating fetuses within the confines of the amni-
otic sac. Nearly all El9 fetuses tested in amnion
that receive a lemon infusion through an intraoral
cannula perform the wiping response ( Robinson
and Smotherman, I99 la). These findings demon-
strate that the age at which the facial wiping re-
sponse appears during ontogeny is contingent
upon the environmental conditions present at the
time of testing.
Fetuses that remain within the amnion appear
to differ in several potentially important ways from
fetuses externalized from the uterus and mem-
branes. First, the fetus in amnion is maintained in
a posture that conforms to the curvature of the am-
niotic sac. The trunk is flexed ventrally and the
head tucked nearer to the chest, compared to fe-
tuses ex utero, which exhibit a straighter trunk and
slightly elevated head. Thus, in a resting posture in
amnion the forelimbs are positioned in proximity
to the head and mouth. The posture of the fetus in
amnion may facilitate forelimb-head contact, nec-
essary for the expression of facial wiping. Second,
the elastic properties of the chorion and uterine
wall tend to resist fetal movements directed out-
ward and may help to stabilize the head as forelimb
strokes are performed during facial wiping. When
E l 9 fetuses an: tested ex utero, they exhibit in-
creased activity in response to infusion, including
forelimb activity, but abundant movements of the
head from side to side appear to prevent forelimb
strokes from contacting the face. Within the am-
nion, however, lateral head movements are selec-
tively suppressed, thereby promoting paw-face
contact (Robinson and Smotherman, 199la).
These findings argue that manipulation of environ-

mental context at the time of stimulus delivery
promotes the coordination of movement trajecto-
ries involving the head and forelimbs into the inte-
grated action pattern that constitutes facial wiping.
Disappearance of the Wiping Response
Nearly all fetuses tested at term exhibit the wiping
response, compared to only half the pups tested on
the day of birth. Over the next 3 postnatal days, the
tendency to exhibit facial wiping after chemosen-
sory infusion disappears almost completely. On
PO, when the neonatal tendency to perform facial
wiping is greatest, infusion of a novel olfactory so-
lution produces an increase in overall behavioral
activity. When facial wiping is expressed, it occurs
during this period of activity after the pup has
rolled to its side or back. Newborn pups are rela-
tively incapable of maintaining a prone posture or
returning to this posture after displacement, indica-
tive of the immature development of a righting re-
sponse at this age (Altman and Sudarshan, 1975).
Behavioral activation evoked by stimulation dis-
places the pup and thereby breaks the contact of
the forelimbs with the substrate. Once freed from
their involvement in postural maintenance, the
forelimbs are available for facial wiping.
The disappearance of facial wiping in the neo-
nate therefore may be due to postural constraint
induced by features of the postnatal environment.
Specifically, as pups become better able to main-
tain a prone posture, and therefore to maintain
forelimb-substrate contact, the forelimbs are less
available for facial wiping and the wiping response
disappears. Especially in young animals, chemo-
sensory infusion evokes a variety of behavior pat-
terns in addition to facial wiping, some of which
also involve the forelimbs and therefore may be
physically incompatible with the simultaneous oc-
currence of facial wiping. Thus, facial wiping may
be suppressed by the activation of other behavioral
responses that compete for expression. In rats, the
onset of competing responses coincides with the
change from a prenatal to postnatal environment.
An ideal test to distinguish the environmental con-
straint and competing response hypotheses would
separate the effects of environment, determined by
the timing of birth, from the relative maturity of
the motor system (Robinson and Smotherman,
1992b).
This ideal is difficult to achieve in studies lim-
ited to a single species. However, comparative
study of precocial cotton rats provides access to
fetuses that exhibit patterns of motor behavior ordi-
narily restricted to the postnatal period in altricial
Mammalian Fetal Motor Activity 1587
species. Fetal cotton rats begin to exhibit non-
evoked activity on E 17 of the 27-day gestation. By
E20, facial wiping can be reliably elicited by in-
traoral infusion of a chemosensory stimulus. If the
neonatal disappearance of the wiping response in
rats was due to features of the postnatal environ-
ment, such as the greater effect of gravity and the
presence of a hard substrate, then cotton rats
should continue to express facial wiping in re-
sponse to infusion through term. However, the in-
cidence of facial wiping declines on E23 and disap-
pears entirely from E24 through birth. The prena-
tal disappearance of facial wiping coincides with
the appearance of other patterns of motor behavior
elicited by chemosensory stimulation, including a
righting response (actively turning to a prone pos-
ture) and the expression of quadrupedal stepping
movements (Fig. 4). The observation that facial
wiping exhibits a temporary disappearance during
the late prenatal period in this precocial species ar-
gues that competition with other motor responses,
rather than the postnatal environment per se, sup-
presses the wiping response. The preferential ex-
pression of righting behavior and locomotor move-
ments instead of facial wiping by older cotton rat
fetuses further suggests the prenatal development
of hierarchical organization in motor control
(Dawkins, 1976; Fentress, 1983).
If the prenatal disappearance of facial wiping in
precocial rodents appears to correspond to the
postnatal disappearance of this behavior in the rat,
then competing motor responses also should be ex-
pressed by the neonatal rat. This hypothesis has
been confirmed by experiments in which the ex-
pression of facial wiping and other motor behavior
is manipulated by varying environmental condi-
tions. Certain environmental contexts minimize
alternative responses to infusion and thereby pro-
mote the postnatal expression of the wiping re-
sponse by pups. For instance, pups suspended in a
buoyant fluid medium are more likely to perform
facial wiping in response to chemosensory stimula-
tion. The facilitatory effect of suspension in water
is most pronounced on P 1. By P3, an effect of sus-
pension is still evident but the majority of pups do
not exhibit the wiping response (Fig. 5 ) .
In the preceding experiment, the pup was sus-
pended in water in a supine posture, holding the
head above the water surface. By P3, supine pups
respond to infusion with vigorous activity that ap-
pears to include elements of the righting response.
It thus appears that righting behavior and other
motor activity interferes with the expression of fa-
cial wiping, especially in older neonates. Because
righting occurs when the pup is displaced from a
1588 Robinson and Smotherman

100

80

8 20

0
E l 8 E l 9 E20 E21 E22 E23 E24 E25 E26 E27

Gestational Age
Figure 4 Behavioral responses of cotton rat fetuses to chemosensory stimulation. The expres-
sion of facial wiping (Wiping), turning to a prone posture (Righting), and performance of
alternated walking or running limb movements (Stepping) is shown from the inception of
movement through term. Note that the wiping response is not expressed by older fetuses.

prone posture, pups in a prone posture should not
express this competing response. This idea was
tested by comparing the response of pups to an
intraoral lemon infusion while suspended in a fluid
medium in either a prone or supine posture. Facial
wiping was expressed by virtually all pups, regard-
less of posture, on PI. On P3 and P5, wiping con-
tinued to be expressed by nearly all pups in a prone
posture but the incidence of wiping declined
among pups in a supine posture (Fig. 6 ) . These
findings suggested that promotion of facial wiping
in age-atypical environments depends upon main-
tenance of the pup in a posture that minimizes the
expression of motor responses that compete with
facial wiping for expression.
The occurrence of motor behavior associated
with the righting response apparently interferes
with the expression of facial wiping. However,

loo] Terrestrial
Immersed
Immersed + Substrate

1 PO P1 P3
Figure 5 Facial wiping by rat pups in response to chemosensory stimulation in three environ-
mental contexts: Tested in a warm incubator on a solid surface (Terrestrial), tested while
immersed to chest depth in a warm water bath (Immersed), or tested on a solid substrate under
the water surface while immersed in a warm bath (Immersed + Substrate).

100
40
20
Figure 6 Facial wiping by rat pups tested while immersed in a warm water bath in either a
supine or prone posture.
pups that remain in a prone posture on a substrate
do not exhibit righting yet do not express facial
wiping in response to infusion. The presence of a
solid substrate plays a role in the pup’s mainte-
nance of a prone posture and apparently inhibits
the wiping response during the first 5 days after
birth. If a pup is supported in a prone posture on a
substrate that provides contact only to the torso,
leaving the forelimbs free, then facial wiping can be
reliably elicited ( 80-90% of subjects). However, if
a pup is supported in a prone posture on a substrate
that provides contact to both the torso and the fore-
paws the incidence of facial wiping is sharply re-
duced (0-20% of subjects). The presence of a solid
substrate also is effective in inhibiting the wiping
response among pups that are immersed in a
buoyant fluid environment: Less than 15% of pups
on PO-P3 exhibit wiping when stimulated while
forepaws are in contact with a submerged substrate
(Fig. 5 ; Smotherman and Robinson, 1989).
Although altricial fetuses consistently exhibit fa-
cial wiping in response to intraoral infusion, their
responsiveness to environmental context also de-
velops during the prenatal period. For instance, fe-
tal rats on E20 of gestation when tested on a sub-
merged substrate as described above nearly always
perform facial wiping. On E2 1, however, only half
the fetuses tested on a substrate express the wiping
response. The ontogenetic appearance of a sub-
strate effect coincides with environmental changes
that ordinarily occur during the last few days of
gestation, most notably the decrease in volume of
amniotic fluid. The disappearance of a fluid buffer
promotes contact between the fetus and the elastic
Mammalian Fetal Motor Activity 1589
!a
P3 P5
wall of the uterus (Smotherman and Robinson,
I988a). The contact between the fetus and uterus
is physically analogous to the interaction between a
pup and the substrate to which it is gravitationally
bound. It is interesting that fetuses that receive an
intraoral infusion in utero on E2 1 consistently fail
to exhibit facial wiping. The correspondence be-
tween the environments immediately before and
after birth suggests that fetal experience with the
substrate-like conditions in utero on the last days of
gestation may influence the development of respon-
siveness to a solid substrate.
Reappearance of the Wiping Response
When tested in an age-typical terrestrial environ-
ment, pups do not exhibit facial wiping in response
to chemosensory stimulation until the second post-
natal week of life. Lemon infusion continues to
evoke other kinds of behavioral responses-older
pups show high levels of motor activity after infu-
sion-but facial wiping is seldom expressed (Jo-
hanson and Shapiro, 1986; Smotherman and Rob-
inson, 1989). Throughout this period of develop-
ment, rat pups are not capable of maintaining a
hunched or erect posture in which the body is sup-
ported solely by the rearlegs. However, juvenile
and adult rats ordinarily balance on the rearlegs
during facial grooming behavior (Richmond and
Sachs, 1980). The postnatal reappearance of the
facial wiping response appears to depend upon the
development of postures that free the forelimbs
from a support role.
We examined the reappearance of the wiping
1590 Robinson and Smothermun

response in preweanling rats by videotaping ses-

sions in which pups receive an intraoral lemon in-
fusion. Adult-like facial wiping involves forelimb
strokes that contact both sides of the face in a syn-
chronized pattern (bilateral wiping). Although this
form of facial wiping is evident in fetuses and be-
gins to reappear during the postnatal period
around PI I (Fig. 4), close inspection of videotape
records indicates that younger rat pups exhibit
components of the adult grooming sequence that
include forelimb-face contact. On P7 and P9, prior
to the development of erect postural support, pups
utilize provisional and highly variable postures to
perform forelimb strokes that contact the face. One
of the more common postures observed involves
tripedal support of the body, freeing one forelimb
to engage in unilateral strokes directed at the face.
Some pups employ only one forelimb in these uni-
lateral strokes; others alternate support and strok-
ing by left and right forelimbs. In another common
posture, the pup supports the forebody on its el-
bows and lowers its head to the substrate, between
the two forepaws. The forelimbs then engage in
limited movement, with both paws contacting the
face in a synchronized pattern. Other postures that
permitted forelimb-face contact, such as bracing
the head, forebody, or rump against a block or ver-
tical surface, were observed less often, further sug-
gesting the possibility of individual variation
(Clark and Ehlinger, 1987) in the ontogeny of fa- Figure 7 Expression of forepaw-facial strokes by pre-
cial wiping (Fig. 7). It remains for future experi- weanling rat pups 7-9 days after birth. Provisional pos-
tures are used to provide forebody support with one fore-
ments to determine whether young rats exhibit in-
limb (top) or both elbows (middle), freeing one or both
dividual preferences for particular postures that fa- paws for facial strokes. On occasion, both forelimbs free
cilitate forelimb-facial contact. when an upright posture is achieved by bracing the body
against a vertical surface (bottom).
Implications for the Conduct of
Developmental Research
2. Because behavior patterns can temporarily
The experimental results discussed in the foregoing disappear during development, it may not be
sections demonstrate that features of the environ- fruitful to employ research strategies in
ment at the time of behavioral assessment and which the expression of a focal behavior is
other aspects of the motor behavior of the subject traced to successivelyyounger animals in the
interact to influence the expression of species-typi- attempt to identify the developmental origin
cal motor behavior, as exemplified by the facial of the pattern.
wiping response. The research strategy employed 3. It may be a general feature of motor develop-
in the experiments described in this section calls ment that contextual manipulations can pro-
attention to some caveats in the study of behav- mote the precocial expression of behavior in
ioral development ( Smotherman and Robinson, young animals, extend the range of ages over
1988d). which a pattern of behavior is performed, or
facilitate the expression of infantile behavior
1. First, environmental context can influence in adult subjects.
not only the form of the response but also 4. Given that behavior is inextricably linked to
whether the response is even performed at a the conditions present at the time of testing,
particular age. and that different manipulations of context

can either accelerate, prolong, or retard the
expression of age-typical behavior, it may
not be meaningful to refer to age-specific
milestones in the development of behavior.
5. Considerable variability exists in the expres-
sion of facial wiping in young rats, but adult
rodents exhibit stereotypy in both the con-
text and form ofthis behavior. Because equiv-
alent developmental endpoints can be
reached by markedly different ontogenetic
trajectories, one may not assume that similar
measures of performance at one age are the
result of similar developmental histories or
predict similar developmental outcomes.
Further, the existence of individual variation
in the course of development, which should
be more common at some ages and less com-
mon at others, necessitates experimental de-
signs that trace developmental change in indi-
vidual subjects.
6. Finally, the need to follow individual sub-
jects during ontogeny and the conflicting
need to test subjects without prior experience
in an experimental situation demand the use
of experimental designs that combine aspects
of longitudinal and cross-sectional methodol-
ogies (Schaie, 1965). Such designs permit
conclusions about age-related differences in
behavior that are not confounded with either
the effects of repeated testing or the age at
which testing begins (cohort effects).
GENERAL PRINCIPLES OF
BEHAVIORAL DEVELOPMENT
IN THE FETUS
Ontogenetic Antecedents and
Ontogenetic Adaptations
On a subjective basis, the motor activity of fetuses
does not rival the dynamic, complex character of
postnatal behavior. However, the dimensions that
define behavior during the postnatal period extent
to movements before birth. Fetal movements ex-
hibit temporal, sequential, and spatial organiza-
tion that emerges between the inception of move-
ment and term. Further, fetuses are behaviorally
responsive to immediate environmental condi-
tions and sensory stimulation. Expression of these
behavioral properties by the fetus implies continu-
ity between pre- and postnatal life (Bekoff, 1988;
Smotherman and Robinson, 1988e).
Indeed, a growing list of behavioral attributes,
many of which have been discussed in the present
Mammalian Fetal Motor Activity 1591
article, appear to be antecedents of postnatal behav-
ior. This list includes: ( 1 ) the patterning of activity
in objectively defined temporal bouts (Slater,
1974; Machlis, 1977; Robinson and Smotherman,
1988) ;( 2 )the production of synchronously coordi-
nated movements (Bekoff and Lau, 1980; Provine,
1980); ( 3 ) the sequential organization of motor
behavior ( Briem, 1986; Fentress and McLeod,
1986); ( 4 ) the activational response to sensory
stimulation, such as intraoral infusion of novel
odorants (Grill and Norgren, 1978; Ganchrow,
Steiner, and Canetto, 1986;Johanson and Shapiro,
1986); ( 5) and the expression of species-typical ac-
tion patterns. In this last category, for instance, fa-
cial wiping and paw-licking of fetuses resembles
postnatal grooming behavior (Golani and Fen-
tress, 1985; Smotherman and Robinson, 1989),
the stretch response to milk appears identical to the
behavior of neonatal rats at the nipple (Hall and
Rosenblatt, 1977; Lau and Henning, 1985),the leg
extension response corresponds to neonatal behav-
ior expressed during episodes of maternal licking
(Moore and Chadwick-Dias, 1986), contact right-
ing behavior anticipates postnatal patterns of dy-
namic postural maintenance (Bignall, 1974; Alt-
man and Sudarshan, 1975; Pellis et al., 1991), and
synergies of limb coordination resemble the punt-
ing behavior of rodent neonates ( Altman and Su-
darshan, 1975) and quadrupedal stepping (Bekoff
and Trainer, 1979; Bekoff, 1986). With the possi-
ble exception of alternated stepping movements,
which may be used by the fetus to change orienta-
tion within the uterus (Suzuki and Yamamuro,
1985), these behavioral attributes develop prior to
and in anticipation of their postnatal expression in
a functional context.
Other documented patterns of fetal behavior are
more difficult to relate directly to postnatal behav-
ior. For example, the fetal response to umbilical
cord occlusion appears to have no clear counter-
part in the postnatal period. Neonatal rats under
conditions of hypoxia exhibit reduced levels of ac-
tivity and cardiac acceleration (Eden and Hanson,
1987). Yet, the response of fetal rats to hypoxia
induced by occlusion of the umbilical cord is an
organized three-phase modulation of motor activ-
ity coupled with transient bradycardia (Smother-
man and Robinson, 1988~).The vigorous, out-
wardly directed movements of the fetus during the
hyperactivity phase of the clamp response may
well increase the probability of alleviatingacciden-
tal cord compression and thereby promote the sur-
vival of the fetus. Because accidental cord com-
pression can occur during otherwise unremarkable
1592 Robinson arid Smotherman
pregnancies, the fetal response to umbilical cord
occlusion appears to present a good case as an on-
togenetic adaptation involving prenatal behavior
(Oppenheim, 1984; Alberts and Cramer, 1988).
Other patterns of fetal behavior may be inter-
preted as possible behavioral adaptations to the in-
trauterine environment. Chief among these is the
fetal response to physical restraint in utero. One of
the most robust behavioral effects we documented
in many independent experiments involves the
condition of fetal observation. The behavior of fe-
tuses observed in utero is qualitatively and quantita-
tively different than that observed ex utero. In gen-
eral, fetuses ex utero exhibit more activity, greater
behavioral diversity, and more pronounced tem-
poral patterning of motor activity than fetuses in
utero. Further, the expression of organized action
patterns during nonevoked activity is much greater
when fetuses are observed outside the uterus.
Several general explanations may be proposed
to account for this effect of environmental context
on fetal behavior. First, fetuses delivered outside
the uterus for observation may be physiologically
compromised and exhibit aberrant responses. How-
ever, this explanation is contravened by available
evidence. Improved methodologies developed for
fetal research (Smotherman and Robinson, 1991 )
ensure that the health of fetuses is not impaired by
observation ex utero. The behavior of fetuses ob-
served ex utero is distinctly different from the activ-
ity of moribund fetuses or fetuses under hypoxic
conditions (Windle, 1944; Smotherman and Rob-
inson, 1 9 8 7 ~ )Late
. in gestation, it is possible to
deliver fetuses, even after extended observation
sessions, that may be cross-fostered to and reared
by newly parturient females (Smotherman, Robin-
son, and Miller, 1986). Further, it seems improba-
ble that perturbation of a complex system, such as
a developing organism, would result in the false
impression of greater organization in many inde-
pendent variables. Yet, fetuses ex utero express be-
havioral organization more, not less, similar to
postnatal behavior.
If the condition effect is not an artifact, then it
probably reflects the behavioral responsiveness of
fetuses to increased restraint in utero. The space
available for movement is much less in utero than
ex utero, and this space diminishes as amniotic
fluid disappears and the fetus continues to grow
near term. Two possible mechanisms are apparent
that could account for reduced behavioral organiza-
tion under conditions of reduced free space. First,
fetal movements may be physically restrained by
the elastic wall of the uterus, in effect filtering all
but the strongest and most vigorous of fetal move-
ments from overt expression. By this hypothesis,
motor commands continue to be generated at the
level of the spinal cord or brain, but immature mo-
tor units lack the strength to distend the uterine
wall and execute movements. This view is compa-
rable to the explanation proposed by Thelen and
colleagues (Thelen and Fisher, 1982; Thelen,
1988) regarding the postnatal disappearance of
neonatal stepping movements. Because the leg
mass of human infants grows faster than leg muscle
strength, infants experience a period of develop-
ment when the legs can no longer be controlled in a
gravitational environment. Eliminating the gravita-
tional restraint by immersing the infant’s legs in
water (analogous to observation of term fetuses ex
utero) is effective in reinstating early stepping be-
havior (Thelen et al., 1984).
The second possible explanation for the condi-
tion effect assumes that any reduction in behav-
ioral organization is due to sensory responsiveness
of the fetus to its surrounding environment, not to
passive environmental masking of fetal move-
ments. Fetuses are responsive to external sensory
stimuli; even altricial fetuses are capable of soma-
tosensation and proprioception that would permit
them to detect restraint and potentially adjust to
diminishing free space in utero (Gottlieb, 1971;
Naryanan et al., 1971). Such sensory feedback
may in turn result in modification of the central
commands governing motor activity. Thus, the al-
tered motor organization observed in utero may
reflect true behavioral responsiveness of the fetus
to environmental conditions. Further, reduction of
abundant or vigorous fetal movements in utero
may reduce the energy expended to distend the
uterus during activity. If this hypothesis proves
correct, then fetal responsivenessto intrauterine re-
straint may be interpreted as another example of
ontogenetic adaptation.
Are data available to choose between these com-
peting hypotheses? Direct evidence would require
some means of measuring motor commands, such
as by EMG recordings. On the one hand, Bekoff
( 1976) demonstrated that patterned motor com-
mands are generated by the spinal cord of the chick
embryo that are not expressed as overt move-
ments. On the other hand, rodent fetuses are de-
monstrably capable of inhibiting the expression of
certain forms of synchronous movement (HR,
HFR) and complex action patterns (facial wiping)
in different contexts. It would seem no less plausi-
ble that fetuses within the confines of the uterus are

capable of comparable behavioral sensitivity to en-
vironmental conditions.
Scaffolding in the Developmental
Construction of Behavior
As originally formulated, the concept of ontoge-
netic adaptation refers to anatomic, physiological,
or behavioral attributes that promote the immedi-
ate survival or well being of an immature organism
in an ontogenetic niche (Oppenheim, 1982). How-
ever, another class of transient ontogenetic attri-
butes may be postulated that exist only to provide a
substrate for further development. Anatomic exam-
ples of such transient structures abound in the em-
bryological literature. Some, like the mesonephric
ducts that give rise to components of the mature
reproductive system in mammals, are phylogenetic
palimpsests of ancient functional systems. Others,
like the gastrula of early embryogenesis, have no
link to past adaptation and serve only as a means of
organizing living matter. Structures of the latter
sort may be as important in early development as
scaffolding is in the construction of an arch. The
individual bricks of the arch cannot be set in place
without external supports, but once the arch is
complete the support system can be removed.
The metaphor of scaffolding may be useful for
understanding certain processes involved in behav-
ioral development (Robinson and Smotherman,
1988, 1991a). A possible example involves pat-
terns of motor synchrony. Synchronous move-
ments by very young fetuses are generated essen-
tially by random association; the incidence of mo-
tor synchrony can be predicted from overall fetal
activity and the relative abundance of motor com-
ponents. Later in gestation, several categories of
synchronous movement are expressed at frequen-
cies significantly greater or less than should be ex-
pected by random association, suggestingthat their
occurrence is the result of central generation or co-
ordination of motor commands.
The sequence in which nonrandom synchro-
nous movements emerge during gestation is in par-
ticular interesting. The first category of synchrony
to emerge from a background of random associa-
tion is forelimb-head movements (FH), followed
almost immediately ( precocial species) or after a
period of days ( altricial species) by forelimb-rear-
limb movements (FR). As the category FR in-
creases in frequency, the category comprising syn-
chronous head-rearlimb movements (HR) de-
creases below chance levels. This reduction in HR
is the earliest behavioral evidence of central inhibi-
Mammalian Fetal Motor Activity 1593
tion of motor activity (cf. Barcroft and Barron,
1939). In precocial cotton rats, the decrease in HR
is accompanied by a corresponding reduction in
the triplet category FHR. Note that at this time
different painvise combinations of the components
F, H, and R are differentially expressed: FH and
FR remain elevated, while HR remains sup-
pressed. FHR, however, reverses its early pattern of
expression and eventually exceeds all other syn-
chronous categories in absolute frequency late in
gestation in precocial species.
This chronology of movement synchrony sug-
gests the following developmental interpretation.
Originally, synchronous movements are produced
by random association of simple motor events.
Later, as central coordinative linkages between
components are established, certain patterns in-
crease (activation) or decrease (inhibition) in ab-
solute and relative frequency. The early coincident
reduction of HR and FHR suggests that these two
patterns may initially be governed by the same in-
hibitory process. Only at a later stage of develop-
ment, exhibited prenatally only by precocial fe-
tuses, is FHR differentiated from HR.
The relationship ofthese abstract movement cat-
egories to behavioral development is important
and fundamental. The coordination of synchro-
nous forelimb and head movements is crucial to
the development of mature grooming behavior
(Golsni and Fentress, 1985) and suckling behavior
of neonatal rodents ( MacFarlane, Pedersen, Cor-
nell, and Blass, 1983). FR synchrony is equally
important in interlimb coordination involved in
locomotion (Bekoff and Lau, 1980; Grillner,
1981 ) . Indeed, control of head position and onen-
tation is also important in the maintenance of bal-
ance during locomotion, so FHR synchrony proba-
bly is related to locomotor development as well.
However, the establishment ofearly interlimb coor-
dination (e.g., FR) may be facilitated without the
added complication of simultaneous head move-
ments. Thus, the synchronous category FR, so evi-
dent in the behavior of altricial fetuses near term
and precocial fetuses at intermediate ages, may be
a transient structure that gives way to or is sub-
sumed by the category FHR at later ages.
A similar interpretation recently has been pro-
posed to explain developmental patterns of early
facial grooming in rodents (Golani and Fentress,
1985). Young mice initially exhibit grooming
movements involving bouts of forelimb strokes
with relatively little involvement of head or trunk.
As development proceeds, temporal clustering of
strokes disappears and is replaced by individual,
1594 Robinson and Smotherman
spatially restricted strokes as head movements are
integrated into the overall sequence. Finally, bout
structure of forelimb strokes reemerges as smooth
coordination of movements involving limbs, head,
and trunk becomes apparent. The second phase of
development described in this sequence involves
retrogression in some aspects of overall organiza-
tion (see also Golani, Bronchti, Moualem, and
Teitelbaum, 1981; Oppenheim, 1981 ), yet it is ap-
parently important in establishingcoordination be-
tween limbs and head. As such, it may be closely
analogous to the FR synchrony exhibited by fe-
tuses. Both transient structures may be examples of
behavioral scaffolding.
The provisional postures expressed by young rat
pups during forelimb strokes that contact the face
provide a further example of transient patterns that
may facilitate behavioral development. The facial
wiping response of the rat fetus disappears during
the first week of postnatal life and does not reap-
pear in adult-like form until the pup is able to sup-
port its body in an erect posture. However, before
pups are able to balance on the rearlimbs the fore-
body may be supported with one forelimb or on
the elbows of both forelimbs, permitting the ex-
pression of incipient movements that involve com-
ponents of the complete facial wiping response.
These provisional postures are not expressed dur-
ing grooming behavior in adulthood and therefore
are functional only for a limited period during the
development of motor behavior.
Heterochrony in the Evolution of Fetal
Behavior
The foregoing discussion emphasized common
patterns in the prenatal development of behavior
in a variety of species. But, important behavioral
differences also should be expected between spe-
cies. In particular, if behavioral development con-
sists of multiple ontogenetic pathways the natural
variation inherent in divergent evolution should
provide independent descriptions of developmen-
tal patterns. This is one of the principal strengths of
comparative analysis: to use carefully selected spe-
cies to disassociate seemingly related processes
(Ridley, 1983).
The potential for evolutionary change to break
up old associations during development is the
source of the prediction of behavioral hetero-
chrony. Heterochrony involves more than a simple
acceleration or retardation of overall development;
it involves a shift in the relative timing of different
developmental processes (McKinney, 1988). In
the evolution of' precocial patterns of development
from an altricial ancestral condition, heterochrony
is predicted in the relative times of emergence of
prenatal ontogenetic adaptations and antecedents
of postnatal behavior. Specifically, behavior pat-
terns necessary for the well being of the fetus in
utero should be retained until birth in all species
and therefore should appear developmentally re-
tarded relative to the more rapidly developing ante-
cedents of postnatal behavior in precocial species.
Examples of prenatal heterochrony may be
gleaned from the recent literature on neural and
physiological development in Norway rats and
spiny mice (Acornys cahzrinus) . Brunjes described
patterns of brain development in spiny mice, some
of which appear accelerated and others retarded
with respect to rats. For instance, many indices of
neural maturation, such as the degree of myelina-
tion of the CNS, indicate that spiny mice are more
mature at birth than Norway rats (Tessitore and
Brunjes, 1988). On the other hand, olfactory bulb
growth is most rapid in both species during the pe-
riod immediately after birth (Brunjes, 1983; Leon,
Coopersmith, U libarri, Porter, and Powers, 1984),
suggesting a heterochronic shift in the timing of
different aspects of neural development. A similar
conclusion is suggested by the developmental pat-
terns of enzyme systems in different organs. En-
zymes associated with neural function are pro-
duced much earlier in development, relative to the
time of birth, in spiny mice than rats. But, enzymes
associated with the lung, pancreas, and small intes-
tine appear to be more strongly correlated with
changes occurring around the time of birth in both
species. Because altricial and precocial species both
begin to breathe and ingest food after birth, the
maturation of these enzyme systems appears to be
related to the oriset of function (Lamers, Mooren,
de Graaf, Markiewicz, and Charles, 1985).
Behavioral processes provide an independent
test of the heterochrony hypothesis. The clearest
comparison may be made between two coordi-
nated action patterns that emerge during the prena-
tal period: facial wiping, an antecedent of postnatal
grooming behavior, and the behavioral response to
umbilical cord compression, a putative ontoge-
netic adaptation of the fetus. Facial wiping is
clearly expressed relatively later during gestation in
altricial species than in precocial species. However,
the expression of hyperactivity in umbilical cord
compression is absent during the early portion of
gestation in cotton rats, resulting in a developmen-
tal retardation of the clamp response relative to
Norway rats. The evident shift in the relative tim-

ing of development of these two action patterns is
consistent with the predicted occurrence of behav-
ioral heterochrony.
The existence of prenatal behavioral hetero-
chrony implies that the ontogeny of behavior, even
during its earliest expression before birth, should
not be characterized as a unitary process of motor
maturation. Instead, it appears to be more appro-
priate to view behavioral development as encom-
passing multiple, independent ontogenetic path-
ways that sometimes exist in close association and
other times result in behavioral competition for ex-
pression. This view of behavioral ontogeny is decid-
edly interactionist (Oyama, 1985). An important
task for future research will be to elaborate and
further test a multipartite conception of behavioral
ontogeny, which may prove to be a point of com-
mon ground between traditional ethological
(Hinde, 1970; Bateson, 1987), psychobiological
(Kuo, 1967; Hofer, 1981), and systems theory
(Kugler and Turvey, 1987; Thelen, 1989) ap-
proaches to understanding behavioral develop-
ment.
This research was supported by Grant HD 16102-08
and Research Career Development Award HD 007 19-05
from the National Institute of Child Health and Human
Development (NIH), a grant from the NATO Collabora-
tive Grants Research Programme 055 1 /88 to W.P.S.,
and Grant HD 2823 1 (NIH) to W.P.S. and S.R.R.
REFERENCES
ALBERTS,J. R. and CRAMER, C. P. ( 1988). Ecology and
experience: Sources of means and meaning of develop-
mental change. In: HandbookofBehavioral Neurobiol-
ogy, vol. 9, Developmental Psychobiology and Behav-
ioral Ecology. E. M. Blass, Ed. Plenum Press, New
York, pp. 1-39.
ALMLI,C. R. and FISHER,R. S. ( 1977). Infant rats: Sen-
sorimotor ontogeny and effects of substantia nigra de-
struction. Brain Res. Bull. 2:425-459.
ALTMAN, J. and SUDARSHAN, K. ( 1975). Postnatal de-
velopment of locomotion in the laboratory rat. Anim.
Behav. 23:896-920.
ANGULO Y GONZALEZ, A. W. ( 1932). The prenatal de-
velopment of behavior in the albino rat. J. Comp.
Neurol. 55395-442.
ARISTOTLE. De GenerationeAnimalium (“On the gener-
ation of animals,” trans. A. Platt, 1952). Encyclopae-
dia Brittanica Press, Chicago, 1L.
ARSHAVSKY, I. A., ARSHAVSKAYA, E. I., and PRAZNI-
KOV, V. P. ( 1976). Motor reactions during the antena-
tal period correlated with the periodic change in the
Mammalian Fetul Motor Activity 1595
activity of the cardiovascular system. Dev. Psychobiol.
9:343-352.
ASCHOFF,J., Ed. ( 1981 ). Handbook of Behavioral Neu-
robiology, vol. 4, Biological Rhythms. Plenum, New
York.
AVERY,G. T. ( 1928). Responses of foetal guinea pigs
prematurely delivered. Gen. Psychol. Monogr. 3:245-
331.
BAER,K. E. VON. (1828). Entwicklungsgaeschichte der
Thiere: Beobachtung und Refloxion. Borntrager, Kon-
igsberg.
BARCROFT, J. and BARRON,D. H. ( 1939). The develop-
ment of behavior in foetal sheep. J. Comp. Neurol.
70:477-502.
BARCROFT,J., BARRON,D. H., and WINDLE,W. F.
(1936). Some observations on genesis of somatic
movements in sheep embryos. J. Physiol. 87:73-78.
BATESON, P. P. G. (1987). Biological approaches to the
study of behavioural development. Int. J. Behav. Dev.
1O:l-22.
BECKER,R. F., KING,J. E., MARSH,R. H., and WYRICK,
A. D. ( 1964). Intrauterine respiration in the rat fetus.
1. Direct observations-comparison with the guinea
pig. Am. J. Obstet. Gynecol. 90:236-246.
BEKOFF,A. ( 1976). Ontogeny of leg motor output in the
chick embryo: A neural analysis. Brain Res. 106:271-
291.
BEKOFF, A. (1981). Embryonic development of the
neural circuitry underlying motor coordination. In:
Studies in Developmental Neurobiology: Essays in
Honor of Viktor Hamburger. W. M. Cowan, Ed. Ox-
ford University Press, New York, pp. 134- 170.
BEKOFF, A. ( 1986). Ontogeny of chicken motor behav-
iors, evidence for multiuse limb pattern generating cir-
cuitry. In: Neurobiology of VertebrateLocomotion. S.
Grillner, P. S. G. Stein, D. G. Stuart, H. Forssberg, and
R. M. Herman, Eds. MacMillian Press, Hampshire,
UK, pp. 433-454.
BEKOFF,A. ( 1988). Embryonic motor output and mo-
tor patterns: Relationship to postnatal behavior. In:
Behavior of the Fetus. W. P. Smotherman and S. R.
Robinson, Eds. Telford Press, Caldwell, NJ, pp. 19 1-
206.
BEKOFF,A. and KAUER,J. A. ( 1984). Neural control of
hatching: Fate of the pattern generator for the leg
movements of hatching in post-hatching chicks. J.
Neurosci. 11:2659-2666.
BEKOFF,A. and LAU,B. ( 1980). Interlimb coordination
in 20-day-old rat fetuses. J. Exp. Zool. 214: 173-175.
BEKOFF,A. and TRAINER, W. ( 1979). The development
of interlimb co-ordination during swimming in post-
natal rats. J. Exp. Bid. 83: 1- 1 1.
BERKINBLIT, M. B., FELDMAN, A. G., and FUKSON,0.I.
(1986). Adaptability of innate motor patterns and
motor control mechanisms. Behav. Brain Sci. 9:585-
638.
BERRIDGE,K. C. and FENTRESS, J. C. ( 1986). Contex-
1596 Robinson and Smotherman
tual control of trigeminal sensorimotor function. J.
Neurosci. 6:325-330.
BICHAT,M. F. X. (1827). Physiological Researches
upon Lye and Death (2nd American ed., trans. F.
Gold). Richardson and Lord, Boston, MA.
BIGNALL, K. E. ( 1974). Ontogeny of levels of neural
organization: The righting reflex as a model. Exp.
Neurol. 42566-5 73.
BIRNHOLZ, J. C. ( 1988). On observing the human fetus.
In: Behavior ofthe Fetus. W. P. Smotherman and S. R.
Robinson, Eds. Telford Press, Caldwell, NJ, pp. 47-
60.
BIRNHOLZ, J. C. and BENECERRAF, B. R. (1983). The
development of human fetal hearing. Science
222~516-518.
BLANCK,A., HARD,E., and LARSSON, K. ( 1967). Onto-
genetic development of orienting behavior in the rat.
J. Comp. Physiol. Psychol. 63:32 7-328,
BLASS,E. M. and TEICHER,M. H. ( 1980). Suckling,
Science 210: 15-22.
BOLLES,R. C. and WOODS,P. J. ( 1964). The ontogeny
ofbehaviour in the albino rat. Anim. Behuv. 12:427-
441.
BRADLEY, N. S. and SMITH,J. L. ( 1988). Neuromuscu-
lar patterns of stereotypic hindlimb behaviors in the
first two postnatal months. 11. Stepping in spinal kit-
tens. Dev. Brain Res. 3853-67.
BRIEM,V. (1986). Temporal organisation in the be-
haviour of newborn infants in active sleep. Behavior
99: 189-220.
BROWN, T. G. (1915). On the activities of the central
nervous system of the unborn foetus of the cat; with a
discussion of the question whether progression (walk-
ing, etc.) is a ‘‘learnt’’ complex. J. Physiol. 49:208-
215.
BRUNJES,P. C. ( 1983). Olfactory bulb maturation in
Acomys cahirinus: Is neural growth similar in preco-
cia1 and altricial Murids? Dev. Brain Res. 8:335-34 1.
CARMICHAEL, L. (1934). An experimental study in the
praenatal guinea-pig of the origin and development of
reflexes and patterns ofbehavior in relation to the stim-
ulation of specific receptor areas during the period of
active fetal life. Gen. Psychof. Monogr. 16:337-49 1.
CARMICHAEL, L. ( 1954). The onset and early develop-
ment of behavior. In: Manual qf Child Psychology,
2nd ed. L. Carmichael, Ed. John Wiley, New York,
pp. 60-185.
CARMICHAEL, L. and SMITH,M. F. ( 1939). Quantified
pressure stimulation and the specificity and generality
of response in fetal life. J. Gen. Psychol. 54425-434.
CLARK,A. B. and EHLINGER, T. J. (1987). Pattern and
adaptation in individual behavioral differences. In:
Perspectives in Ethology, vol. 7. P. P. G. Bateson and
P. H. Klopfer, Eds. Plenum, New York, pp. 1-47.
COGHILL,G. E. (1929). Anatomy and the Problem of
Behavior. Cambridge University Press, Cambridge,
UK.
CORONIOS,J. D. ( 1933). Development of behavior in
the fetal cat. Gen. Psychol. Monugr. 14283-386.
DAWKINS,R. ( 1976). Hierarchical organisation: A can-
didate principle for ethology. In: Growing Points in
Elhology. P. P. (3. Bateson and R. A. Hinde, Eds.
Cambridge University Press, Cambridge, UK, pp.
7-54.
DREWETT,R. F., STATHAM,C., and WAKERLEY, J . B.
(1974). A quantitative analysis of the feeding be-
haviour of suckling rats. Anim. Behav. 22:907-9 13.
EDEN,G. J. and HANSON,M. A. ( 1987). Maturation of
the respiratory response to acute hypoxia in the new-
born rat. J. Physiol. 392:l-9.
EILAM,D. and GOLANI,I. ( 1988). The ontogeny of ex-
ploratory behavior in the house rat (Rattus rattus):
The mobility gradient. Dev. Psychobiol. 21:679-7 10.
EISENBERG, J. F. ( I98 1 ). The Mammalian Radiations:
An Analysis of Trends in Evolution, Adaptation, and
Behavior. University of Chicago Press, Chicago, IL.
FENTRESS, J. C. ( 1983). Ethological models ofhierarchy
and patterning of species-specific behavior. In: Hand-
book of Behavioral Neurobiology, vol. 6, Motivation.
E. Satinoff and P. Teitelbaum, Eds. Plenum, New
York, pp. 185-234.
FENTRESS, J. C. ( 1988). Expressive contexts, fine struc-
ture, and central mediation of rodent grooming. Ann.
NY Acad. Sci. 525: 18-26.
FENTRESS, J. C. and MCLEOD,P. J. (1986). Motor pat-
terns in development. In: Handbook of Behavioral
Neurobiology, vol. 8, Developmental Psychobiology
and Developmental Neurobiology. E. M. Blass, Ed. Ple-
num, New York, pp, 35-97.
FIFER,W. P. and MOON,C. (1988). Auditory experi-
ence in the fetus. In: Behavior of the Fetus. W. P.
Smotherman and S. R. Robinson, Eds. Telford Press,
Caldwell, NJ, pp. 175-190.
GANCHROW,J. R., STEINER,J. E., and CANETTO,S.
( 1986). Behavioral displays to gustatory stimuli in
newborn rat pups. Dev. Psychobiol. 19~163-174.
GARD,C., HARD, E., LARSSON,K., and PETERSSON,
V. A. ( 1967).The relationship between sensory stimu-
lation and gross motor behaviour during the postnatal
development in the rat. Anim. Behav. 15563-567.
GEUBELLE, F. ( 1984). Perception ofenvironmental con-
ditions by the fetus in utero. In: Progress in Reproduc-
tive Biology and Medicine, vol. 1 I . P. 0. Hubinont,
Ed. S. Karger, Basel, pp. 1 10-1 19.
GOLANI,I., BRONCHTI, G., MOUALEM, D., and TEITEL-
BAUM, P. ( 198 I ) . “Warm-up” along dimensions of
movement in the ontogeny of exploration in rats and
other infant mammals. Proc. Natl. Acad. Sci. USA
18~7226-7229.
GOLANI,I. and FENTRESS, J. C. (1985). Early ontogeny
of face grooming in mice. Dev. Psychobiol. 18529-
544.
GOTTLIEB,G. ( 1971 ). Ontogenesis of sensory function
in birds and mammals. In: The Biopsychology ofDevel-

opment. E. Tobach, L. R. Aronson, and E. Shaw, Eds.
Academic Press, New York, pp. 67-128.
GOTTLIEB,G. and Kuo, Z. Y. ( 1965). Development of
behavior in the duck embryo. J. Comp. Physiol. Psy-
cho/. 59:183-188.
GRILL,H. J. and NORGREN,R. ( 1978).The taste reactiv-
ity test. I. Mimetic responses to gustatory stimuli in
neurologically normal rats. Brain Res. 143:263-279.
GRILLNER,S. ( 198 I ) . Control of locomotion in bipeds,
tetrapods, and fish. In: Handbook of Physiology, sec-
tion 1, The Nervous System, vol. 11, Motor Control,
part 2. Williams & Wilkins, Baltimore, MD, pp.
1179-1236.
GRILLNER,S. (1985). Neurobiological bases of rhyth-
mic motor acts in vertebrates. Science 228: 143- 149.
HALBERTSMA, J. M. ( 1983). The stride cycle of the cat:
The modelling of locomotion by computerized analy-
sis of automatic recordings. Acta Physiol. Scand.
S ~ p p l52
. 1:1-75.
HALL,W. G. ( 1979). The ontogeny of feeding in rats: I.
Ingestive and behavioral responses to oral infusions. J.
Comp. Physiol. Psychol. 93:977- 1000.
HALL,W. G. and ROSENBLATT, J. S. (1977). Suckling
behavior and intake control in the developing rat pup.
J. Comp. Physiol. Psychol. 91: 1232- 1247.
HAMBURGER, V. ( 1963). Some aspects of the embryol-
ogy of behavior. Q. Rev. Biol. 38:342-365.
HAMBURGER, V. ( 1973). Anatomical and physiological
basis of embryonic motility in birds and mammals. In:
Behavioral Embryology. G. Gottlieb, Ed. Academic
Press, New York, pp. 5 1-76.
HAMBURGER, V., WENGER,E., and OPPENHEIM, R. W.
( 1966). Motility in the chick embryo in the absence of
sensory input. J. Exp. Zool. 162: 133- 160.
HARVEY,W. ( I65 1 ). AnatomicalExercises on the Gener-
ation of Animals (reprint, 1952). Encyclopaedia Bri-
tannica Press, Chicago, IL.
HEPPER,P. G. ( 1988). Adaptive fetal learning: Prenatal
exposure to garlic affects postnatal preferences. Anim.
Behav. 36:935-936.
HINDE,R. A. ( 1970). Animal Behaviour: A Synthesis of
Ethology and Comparative Psychology. McGraw-Hill,
New York.
HOFER,M. A. ( 1981 ). The Roots of Human Behavior:
An Introduction to the Psychobiology ofEarly Develop-
ment. W. H. Freeman, New York.
HOOKER,D. ( 1952). The Prenatal Origin of Behavior.
18th Porter Lecture Series. University of Kansas Press,
Lawrence, KS.
HUMPHREY, T. ( 1953). The relation of oxygen depriva-
tion to fetal reflex arcs and the development of fetal
behavior. J. Psychol. 353-43.
HUMPHREY,T. (1969). Postnatal repetition of human
prenatal activity sequences with some suggestions of
their neuroanatomical basis. In: Brain and Early Be-
haviour: Development in the Fetus and Infant. R. J.
Robinson, Ed. Academic Press, New York, pp. 43-84.
Mammalian Fetal Motor Activity 1597
ITSKOVITZ,J., LAGAMMA, E. F., and RUDOLPH,A. M.
(1987). Effects of cord compression on fetal blood
flow distribution and O2 delivery. Am. J. Physiol.
252:H 100-H 109.
JAMES,W. ( 1890). Principles of Psychology (reprint,
1952). Encyclopaedia Bntannica Press, Chicago, IL.
JOHANSON, I. B. and SHAPIRO,E. G. ( 1986). Intake and
behavioral responsiveness to taste stimuli in infant
rats from 1 to 15 days ofage. Dev. Psychobiol. 19593-
606.
KELLOGG,C. and LUNDBORG,P. ( 1972). Ontogenic
variations in responses to L-DOPA and monoamine
receptor-stimulating agents. Psychopharmacologia
23: 187-200.
KUGLER,P. N. and TURVEY,M. T. (1987). Informa-
tion, Natural Law, and the Self-Assembly ofRhythmic
Movement, Lawrence Erlbaum, Hillsdale, NJ.
Kuo, Z. Y. ( 1967). The Dynamics ofBehavior Develop-
ment: An Epigenetic View. Random House, New
York.
LAMERS,W. H., MOOREN,P. G., DEGRAAF,A., MAR-
KIEWICZ, A., and CHARLES,R. (1985). Perinatal or-
gan development in rat and spiny mouse: Its relation
to altncial and precocial timing of birth. In: The Physi-
ological Development of the Fetus and Newborn. c.T.
Jones and P. W. Nathanielsz, Eds. Academic Press,
London, UK, pp. 41-45.
LAU,C. and HENNING,S. J. ( 1985). Investigation ofthe
nature ofthe “stretch response” in suckling rats. Phys-
iol. Behav. 34:649-65 1.
LEON,M., COOPERSMITH, R., ULIBARRI,C., PORTER,
R. H., and POWERS,J. B. ( 1984). Development of ol-
factory bulb organization in precocial and altricial ro-
dents. Dev. Brain Res. 12:45-53.
LICHTMAN, A. H. and CRAMER,C. P. (1989). Relative
importance ofexperience, social facilitation, and avail-
ability of milk in weaning rats. Dev. Psychobiol.
22~347-356.
MACFARLANE, B. A., PEDERSEN, P. E., CORNELL,C. E.,
and BLASS,E. M. ( 1983). Sensory control of suckling-
associated behaviours in the domestic Norway rat,
Rattus norvegicus. Anim. Behav. 31:462-47 1.
MACHLIS,L. ( 1977). An analysis of the temporal pat-
terning of pecking in chicks. Behavior 63: 1-70.
MA”, L. I. ( 1986). Pregnancy events and brain dam-
age. Am. J. Obstet. Gynecol. 1556-9.
MARSH,R. H., KING,J. E., and BECKER,R. F. ( 1963).
Volume and viscosity of amniotic fluid in rat and
guinea pig fetuses near term. Am. J. Obstet. Gynecol.
85~487-492.
MCKINNEY,M. L., Ed. ( 1988). Heterochrony in Evolu-
tion. Plenum, New York.
MISTRETTA,C. M. and BRADLEY, R. M. ( 1986). Devel-
opment of the sense of taste. In: Handbook of Behav-
ioral Neurobiology, vol. 8, Developmental Psychobiol-
ogy and Developmental Neurobiology. E. M. Blass, Ed.
Plenum, New York, pp. 205-236.
1598 Robinson and Smotherman
MOORE,C. L. and CHADWICK-DIAS, A. M. (1986). Be-
havioral responses of infant rats to maternal licking:
Variations with age and sex. Dev. Psychobiol. 19:427-
438.
NARAYANAN, C. H., Fox, M. W., and HAMBURGER, V.
( 197 1 ). Prenatal development of spontaneous and
evoked activity in the rat. Behavior 40: 100- 134.
OPPENHEIM,R. W. ( 198 1 ). Ontogenetic adaptations
and retrogressive processes in the development of the
nervous system and behaviour: A neuroembryological
perspective. In: Maturation and Development: Biologi-
cal and Psychological Perspectives. K. J. Connolly and
H. F. R. Prechtl, Eds. Lippincott, Philadelphia, PA,
pp. 73-109.
OPPENHEIM,R. W. (1982). The neuroembryological
study of behavior: Progress, problems, perspectives.
In: Current Topics in Developmental Biology, vol. 17,
Neural Development, part 3. R. K. Hunt, Ed. Aca-
demic Press, New York, pp. 257-309.
OPPENHEIM, R. W. (1984). Ontogenetic adaptations in
neural development: Toward a more ‘ecological’devel-
opmental psychobiology. In: Continuity of Neural
Functions from Prenatal to Postnatal Lge. H. F. R.
Prechtl, Ed. Lippincott, New York, pp. 16-30.
OYAMA,S. ( 1985). The Ontogeny oflnformation. Cam-
bridge University Press, Cambridge, UK.
PEDERSEN, P. E., GREER,C. A., and SHEPHERD, G. M.
( 1986). Early development of olfactory function. In:
Handbook of Behavioral Neurobiology, vol. 8, Devel-
opmental Psychobiology and Developmental Neurobi-
ology. E. M. Blass, Ed. Plenum Press, New York, pp.
163-203.
PELLIS,V. C., PELLIS,S. M., and TEITELBAUM, P.
( 1991 ) .A descriptive analysis ofthe postnatal develop-
ment of contact-righting in rats (Rattus norvegicus).
Dev. Psychobiol. 24~237-263.
PFTSTER, J . F., CRAMER,C. P., and BLASS,E. M. ( 1986).
Suckling in rats extended by continuous living with
dams and their preweanling litters. Anim. Behav.
34:415-420.
PREYER,W. ( 1885). Specielle Physiologie des Embryo.
Vntersuchungen uber die Lebensercheinungen vor der
Geburt. Grieben, Leipzig.
PROVINE,R. R. ( 1980). Development of between-limb
movement synchronization in the chick embryo. Dev.
Psychobiol. 13:15 1- 163.
PROVINE,R. R. ( 1986). Behavioral neuroembryology:
Motor perspectives. In: Developmental NeuroPsycho-
biology. W. T. Greenough and J. M. Juraska, Eds. Aca-
demic Press, New York, pp. 2 13-239.
REPPERT,S. M. and WEAVER,D. R. (1988). Maternal
transduction of light-dark information for the fetus.
In: Behavior of the Fetus. W. P. Smotherman and S. R.
Robinson, Eds. Telford Press, Caldwell, NJ, pp. 119-
139.
RICHMOND, G. and SACHS,B. D. ( 1980). Grooming in
Norway rats: The development and adult expression
of a complex motor pattern. Behavior 7582-96.
RIDLEY,M. ( 1983). The Explanation of Organic Diver-
sity: The Comparative Method and Adaptations for
Mating. Clarendon Press, Oxford, UK.
ROBERTSON, S. S. (1985). Cyclic motor activity in the
human fetus after midgestation. Dev. Psychobiol.
18:411-419.
ROBERTSON, S . S. (1987). Human cyclic motility: Fe-
tal-newborn continuities and newborn state differ-
ences. Dev. Ps-vchobiol. 20:425-442.
ROBERTSON, S. S. ( 1988). Mechanism and function of
cyclicity in spontaneous movement. In: Behavior of
the Fetus. W. P. Smotherman and S. R. Robinson,
Eds. Telford Press, Caldwell, NJ, pp. 77-94.
ROBERTSON, S. S. ( 1990). Temporal organization in fe-
tal and newborn movement. In: Sensory-Motor Orga-
nization and Development in Infancy and Early Child-
hood. H. Bloch and B. Bertenthal, Eds. Kluwer Aca-
demic Publishers, Dordrecht, The Netherlands, to
appear.
ROBERTSON, S. S. and SMOTHERMAN, W. P. (1990).
The neural control of cyclic activity in the fetal rat.
Physiol. Behav. 47: 12 1- 126.
ROBINSON, S. R. ( 1989).A ComparativeStudyofPrena-
tal Behavioral Ontogeny in Altricial and Precocial
Murid Rodents. Unpublished doctoral dissertation,
Oregon State University, Corvallis, OR.
ROBINSON, S. R. and SMOTHERMAN, W. P. ( 1987). Envi-
ronmental determinants of behaviour in the rat fetus.
11. The emergence of synchronous movement. Anim.
Behav. 351652-1662.
ROBINSON,S. R. and SMOTHERMAN, W. P. (1988).
Chance and chunks in the ontogeny of fetal behavior.
In: Behavior of the Fetus. W. P. Smotherman and S. R.
Robinson, Eds. TeIford Press, Caldwell, NJ, pp. 95-
115.
ROBINSON, S. R. ,and SMOTHERMAN, W. P. ( 1990). The
prenatal expression of punting locomotor behavior in
altricial and precocial fetuses. Paper presented at the
annual meeting of the International Society for Devel-
opmental Psychobiology, Cambridge, UK.
ROBINSON,S. R. and SMOTHERMAN, W. P. (1991a).
The amniotic sac as scaffolding: Prenatal ontogeny of
an action pattern. Dev. Psychobiol. 24463-485.
ROBINSON,S. R. and SMOTHERMAN, W. P. ( 1991b).
Behavioral response of altricial and precocial rodent
fetuses to acute umbilical cord compression. Behav.
Neural Biol. 57:93-102.
ROBINSON, S. R. and SMOTHERMAN, W. P. ( 1992a). Or-
ganization of the stretch response to milk in the rat
fetus. Dev. Psychobiol. 2533-49.
ROBINSON,S. R. and SMOTHERMAN, W. P. (1992b).
Motor competition in the prenatal ontogeny of spe-
cies-typical behavior. Anim. Behav., to appear.
ROCHAT, P. and STACEY,M. (1989). Reaching in
various postures by 6 and 8 month-old infants: The
development of the mono-manual grasp. Poster pre-
sented at the SRCD meeting, Kansas City, KS.
SCHAAL,B. (1988). Olfaction in infants and children:

Developmental and functional perspectives. Chem.
Senses 13:145- 190.
SCHAIE,K. W. ( 1965). A general model for the study of
developmental problems. Psychol. Bull. 64:92- 107.
SHIK,M. L. and ORLOVSKY, G. N. ( 1976). Neurophysi-
ology of locomotor automatism. Physiol. Rev.
56:465-501.
SILVER,W. L. and MOULTON,D. G. ( 1982). Chemosen-
sitivity of rat nasal trigeminal receptors. Physiol. Be-
hav. 28:927-93 I .
SLATER,P. J. B. ( 1974). The temporal patterning of
feeding in the zebra finch. Anim. Behav. 22506-5 15.
SMOTHERMAN, W. P. ( 1982).Odor aversion learning by
the rat fetus. Physiol. Behav. 29:769-77 I .
SMOTHERMAN, W. P. and ROBINSON, S. R. (1985). The
rat fetus in its environment: Behavioral adjustments
to novel, familiar, aversive and conditioned stimuli
presented in utero. Behav. Neurosci. 99:521-530.
SMOTHERMAN, W. P. and ROBINSON, S. R. ( 1986). Envi-
ronmental determinants of behaviour in the rat fetus.
Anim. Behav. 341859-1873.
SMOTHERMAN, W. P. and ROBINSON, S. R. ( 1987a). Pre-
natal expression of species-typical action patterns in
the rat fetus (Rattus nowegicus). J . Comp. Psychol.
101:190- 196.
SMOTHERMAN, W. P. and ROBINSON,S. R. (1987b).
Psychobiology of fetal experience in the rat. In: Perina-
tul Development: A Psychobiological Perspective.
N. A. Krasnegor, E. M. Blass, M. A. Hofer, and W. P.
Smotherman, Eds. Academic Press, Orlando, FL, pp.
39-60.
SMOTHERMAN, W. P. and ROBINSON, S. R. ( 1 9 8 7 ~ )Ste-
. STICKROD,
reotypic behavioral response ofrat fetuses to acute hyp-
oxia is altered by maternal alcohol consumption. Am.
J. Obstet. Gynecol. 157:982-986.
SMOTHERMAN, W. P. and ROBINSON,S. R. (1988a).
The uterusas environment: The ecology of fetal experi-
ence. In: Handbook of Behavioral Neurobiology, vol.
9, Developmental Psychobiology and Behavioral Ecol-
ogy. E. M. Blass, Ed. Plenum Press, New York, pp.
149-1 96.
SMOTHERMAN, W. P. and ROBINSON,S. R. (1988b).
Fetal expression of the leg extension response to ano-
genital stimulation. Physiol. Behav. 43:243-244.
SMOTHERMAN, W. P. and ROBINSON, S. R. ( 1 9 8 8 ~ )Re-
. THELEN,E. ( 1988). On the nature of developing motor
sponse of the rat fetus to acute umbilical cord occlu-
sion: An ontogenetic adaptation? Physiol. Behuv.
44:13 1-135.
SMOTHERMAN, W. P. and ROBINSON,S. R. (1988d).
Behavior of the Fetus. Telford Press, Caldwell, NJ.
SMOTHERMAN, W. P. and ROBINSON, S. R. ( 1988e). Di-
mensions of fetal investigation. In: Behavior of the
Fetus. W. P. Smotherman and S. R. Robinson, Eds.
Telford Press, Caldwell, NJ, pp. 19-34.
SMOTHERMAN, W. P. and ROBINSON, S. R., Eds. ( 1989).
Cryptopsychobiology: The appearance, disappearance
and reappearance of a species-typical action pattern
Mammalian Fetal Motor Activity 1599
during early development. Behav. Neurosci. 103:246-
253.
SMOTHERMAN, W. P. and ROBINSON, S. R. ( 1990a). Rat
fetuses respond to chemical stimuli in gas phase. Phys-
iol. Behav. 47:863-868.
SMOTHERMAN, W. P. and ROBINSON,S. R. ( 1990b).
Olfactory bulb transection alters fetal behavior after
chemosensory but not tactile stimulation. Dev. Bruin
Res. 57: 175- 180.
SMOTHERMAN, W. P. and ROBINSON,S. R. ( 1 9 9 0 ~ ) .
The prenatal origins of behavioral organization. Psy-
chol. Sci. 1:97-106.
SMOTHERMAN, W. P. and ROBINSON, S. R. ( 199 1 ). Ac-
cess to the rat fetus for psychobiological investigation.
In: Developmental Psychobiology: New Methods and
Changing Concepts. H. Shair, G. A. Barr, and M. A.
Hofer, Eds. Oxford University Press, Cambridge, UK,
pp. 148-164.
SMOTHERMAN, W. P., ROBINSON,S. R., and MILLER,
B. J. ( 1986). A reversible preparation for observing
the behavior of fetal rats in utero: Spinal anesthesia
with lidocaine. Physiol. Behav. 3757-60.
SMOTHERMAN, W. P., ROBINSON,S. R., and ROBERT-
SON, S. S. ( 1988). Cyclic motor activity in the fetal rat
(Rattus nowegicus) . J. Comp. Psychol. 102:78-82.
STEHOUWER, D. J. and FAREL,P. B. (1984). Develop-
ment of hindlimb locomotor behavior in the frog. Dev.
Psychobiol. 17:2 17-232.
STELZNER,D. J. ( 1986). Ontogeny of the encephaliza-
tion process. In: Developmental Neuropsychobiology.
W. T. Greenough and J. M. Juraska, Eds. Academic
Press, Orlando, FL, pp. 24 1-270.
G., KIMBLE, D. P., and SMOTHERMAN, W. P.
( 1982). In utero taste/odor aversion conditioning in
the rat. Physiol. Behav. 27:557-558.
SUZUKI,S. and YAMAMURO, T. (1985). Fetal move-
ment and fetal presentation. Early Human Dev.
11:255-263.
SWENSON,E. A. (1926). The Development of Move-
ment of the Albino Rat before Birth. Unpublished
doctoral dissertation, University of Kansas, Lawrence,
KS.
TESSITORE, C. and BRUNJES,P. C. ( 1988). A compara-
tive study of myelination in precocial and altricial
murid rodents. Dev. Brain Res. 43:139-147.
systems and the transition from prenatal to postnatal
life. In: Behavior of the Fetus. W. P. Smotherman and
S. R. Robinson, Eds. Telford Press, Caldwell, NJ, pp.
207-224.
THELEN,E. ( 1989). Self-organization in developmental
processes: Can systems approaches work? In: Systems
and Development, Minnesola Symposia on Child Psy-
chology, vol. 22. M. R. Gunnar and E. Thelen, Eds.
Lawrence Earlbaum, New York, pp. 77-1 17.
THELEN, E. and FISHER,D. M. (1982). Newborn step-
ping: An explanation for a “disappearing reflex.” Dev.
Psychol. 18:760-775.
1600 Robinson and Smotherman
THELEN,E., FISHER, D., and RIDLEY-JOHNSON, R.
( 1984). The relationship between physical growth and
a newborn reflex. Infant Behav. Dev. 7:479-493.
TILNEY,F. ( 1933). Behavior in its relation to the devel-
opment of the brain. Part 11. Bull. Neurol. Inst. N Y
3:252-358.
TRAVERS,J. B., TRAVERS,S. P., and NORGREN,R.
( 1987). Gustatory neural processing in the hindbrain.
Annu. Rev. Neurosci. 10:595-632.
VAN HARTESVELDT, C., SICKLES,A. E., PORTER,J. D.,
and STEHOUWER, D. J. ( 1991 ) . L-DOPA-induced air-
stepping in developing rats. Dev. Brain Res. 58:25 1-
255.
VOLOKHOV, A. A. ( 1968). Comparative studies of the
functional development of analyzer systems in ani-
mals in the process of ontogenesis. Prog. Brain Res.
22~521-540.
WINDLE,W. F. ( 1944). Genesis of somatic motor func-
tion in mammalian embryos: A synthesizing article.
Physiol. Zool. 17:247-26 1.
WINDLE,W. F. and FISH,M. W. (1932). The develop-
ment of the vestibular righting reflex in the cat. J.
Comp. Neurol. 54535-96.
WINDLE,W. F. and GRIFFIN,A. M. ( 1931 ). Observa-
tions on embryonic and fetal movements ofthe cat. J.
Comp. Neurol. 52: 149-1 88.
WINDLE,W. F., MINEAR,W. L., AUSTIN, M. F., and
ORR, D. W. (1035). The origin and early develop-
ment of somatic; behavior in the albino rat. Physiol.
ZOO/.8:156- 185.
WINDLE,W. F., O'DONNELL,J. E., and GLASSHAGLE,
E. E. ( 1933). The early development of spontaneous
and reflex behavior in cat embryos and fetuses. Phys-
101. Zool. 6 5 2 1-54 1.