American Journal of Primatology 17:ll-25 (1989) Early Development of Locomotor Behavior in Vervet Monkeys JOEL A. VILENSKY AND EVA GANKIEWICZ Department ofdnatomy, Indiana University School of Medicine, Fort Wayne, Indiana The locomotor development of three vervet infants across approximately the first 2 months of life is described. Fairly normal-looking walking movements (as compared to adults) were seen in all the animals by approximately 1 month of age and galloping was observed by 2 months. Early locomotor footfall patterns were often aberrant and bounding-type gaits were sometimes exhibited. Most of the symmetrical gaits observed were classifiable as lateral sequence. Across the 2-month period the animals showed decreased three- and four-foot support and improvements in joint angular displacement patterns. From their earliest locomotor movements the infants showed significant linear relationship between both cycle duration and swing and stance durations of the limbs. We suggest that locomotor control mechanisms are probably fairly mature at birth but that weight support and postural control problems explain the initial locomotor difficulties exhibited by these infants. Key words: primates, locomotor development, gaits, footfall patterns INTRODUCTION Primates are not able to locomote a t birth. But at what age do various taxa achieve the ability to walk, run, and gallop? Surprisingly, there are few data providing such information for most nonhuman primate species. For vervets (Cercopithecus aethiops) Hurov [19821 provides data for three individuals; one at age 15 days, one a t ages 56-75 days, and one at ages 81-105 days. However, since Chalmers [19721 states that vervets achieve “steady” locomotion by a mean age of 36 days, Hurov’s description is primarily based upon animals that had already achieved fairly mature locomotor skills. For the present study we investigated the locomotor development of three infant vervets from their birth date (or next day) until they were approximately 2 months of age. These animals were filmed weekly and, later, every 2 weeks, as they locomoted or attempted to locomote on a platform. METHODS The three infant monkeys used in this study were all female. Animal No. 1 was filmed on days 1,8,15,22,29,43,and 58 of life (day 0 = day of birth); animal No. Received for publication July 8, 1988; revision accepted September 22, 1988. Address reprint requests to Dr. Joel A. Vilensky, Fort Wayne Center for Medical Education, 2101 Coliseum Blvd., East, Ft. Wayne, IN 46805. 0 1989 Alan R. Liss, Inc. 12 I Vilensky and Gankiewicz TABLE I. Number of Strides Available for Each Animal for Each Age (Days) No. 1 Age 8 15 22 29 43 58 No. 2 No. 3 Strides Age Strides Age Strides 3 6 28 42 10 13 21 28 2 2 2 4 6 40 56 14 5 9 64 7 9 a 2 on days 0, 7,13,21,28,40, and 56 of life; and, animal No. 3 on days 0 , 7 , 1 3 , 21, 28,42, and 64 of life. For these filming sessions the infants were taken from their mothers and placed on a Plexiglas-enclosed elevated surface (1.17 x 0.43 m). Their locomotor attempts and behavior were then filmed by using a 16-mm camera set to run at approximately 64 framesls for animal Nos. 1 and 2 and 100 framesls for animal No. 3. Additionally, each animal was filmed while air-stepping, i.e., walkinglike movements exhibited by infants while suspended [cf. Vilensky et al., 19891. Each filming session lasted about 15 minutes. Exact filming speeds for each trial were determinable from timing markers placed on the film a t 0.01-s intervals. Following the completion of each filming episode the animals were returned to their respective mothers. Animal No. 1 was housed with her mother in a singleanimal cage. The other two infants were housed with their respective mothers and other animals in large group cages. The films, once processed, were initially viewed (using an analytical projector) with the goal of making qualitative evaluations as to how “normal” (i.e., adultlike) the locomotor movements appeared and of making basic gait determinations. These qualitative evaluations were based upon our extensive observations of adult vervet locomotion [Vilensky et al., 19881. All of the available film was used in this analysis. The filmed locomotor sequences were then divided into actual “strides.” Strides were defined as the interval between a hind-foot touchdown to the subsequent touchdown of the same foot. It is important to note that not all of the data were definable as strides because of the erratic nature of some of the movements. Furthermore, in many cases, the strides filmed were abnormal in that the footfall order or timings were not those typically observed in a more mature animal. Table I shows the total number of strides available for each animal a t each age. It is clear from the table that there were only a limited number of strides available for animal No. 3. All of the defined strides were initially classified as symmetrical or nonsymmetrical based upon the footfalls of the hind limbs. Specifically, if a hind-foot touchdown occurred between 40 and 60%of the stride defined by the contralateral hind-foot’s touchdown, we considered it to be a symmetrical stride [cf. Hildebrand, 19661. Furthermore, the symmetrical strides were classified as being either diagonal sequence (DS) or lateral sequence (LS). In a DS gait, hind-limb touchdown is followed by the touchdown of the contralateral forelimb (e.g., left hind limb, right forelimb). In a LS gait, hind-limb touchdown is followed by that of the ipsilateral forelimb (e.g., left hind limb, left forelimb). The DS and LS strides were then subdivided into single foot (SF),lateral couplets (LC), and diagonal couplets (DC) types depending on the precise timings of the footfalls. A SF gait occurs if the footfalls of all the limbs are evenly spaced in time. A LC gait occurs when the footfalls on the same side of the body are coupled in time, and a DC gait occurs if Locomotor Development in Vervets / 13 there is coupling between the diagonally opposite fore- and hind limbs [Hildebrand, 1966, 19801. Finally, if diagonal limbs move approximately synchronously the gait is classified as a trot. The nonsymmetrical strides were divided into typical asymmetrical gaits such as gallops and bounds, in which the touchdowns of the fore- and hind-limb pairs are each coupled in time (e.g., left hind-limb touchdown followed by right hind-limb touchdown), and strides are coupled with abnormal footfall patterns. These were strides in which alternate hind- and forelimb footfalls occurred, but the hind-limb touchdowns were not within the 40-60% phase interval of each other. For these we only determined the limb movement sequence (i.e., DS or LS) while we classified the galloping strides as transverse or rotary and the bounds as complete bounds or half-bounds. A transverse gallop is characterized by both the fore- and hind limbs having similar leading limbs (e.g., the left hind limb striking the ground before the right, and the left forelimb striking the ground before the right) while in a rotary gallop the leading limbs are different. In a complete bound, the hind limbs contact the ground in unison as do the forelimbs. A half-bound is characterized by only the hind limbs contacting the ground in unison [Hildebrand, 19771. In order t o ascertain if the number or organization of limbs supporting the animals during locomotion changed as they aged, the types of limb support the animals used (e.g., unipodal, diagonal, lateral, etc.) were determined for each stride. These data were then averaged for each animal for each date. The timings of the stance and swing durations of the limbs were evaluated for each animal. That is, within the strides, each complete locomotor cycle for a particular limb was divided into its two component parts (stance and swing). We then examined how each component changed relative t o differing durations of the entire cycle both within a date and across time. Finally, hind-limb joint angular displacement patterns were analyzed. This was accomplished by digitizing the estimated locations of the metatarsophalangeal, ankle, knee, and hip joints and by using a computer to calculate the corresponding angles. (The hip angle was measured to the horizontal.) This procedure was not uniformly done across animals and dates. Depending on the quality and types of strides exhibited by each animal on each date as well as whether the movements were roughly perpendicular to the axis of the camera lens, we chose to digitize between zero and three strides per animal per date. Each of the raw joint displacement patterns from these strides was smoothed by using a digital filtering procedure [Winter, 19791 with cutoff values of either 5 or 6 . RESULTS Gaits Animal No. 1. Figure 1 illustrates some of the gaits used by this animal during the study period. On day 1of life this animal generally toppled over when placed on the platform. She did, however, exhibit one bounding-type stride in which some forward movement was achieved. At 8 days, animal No. 1 showed a clear ability to support her weight for brief periods during locomotor movements (Fig. 1). She exhibited bounding strides and nonsymmetrical gaits which were all LS. At 15 days of age, animal No. 1 exhibited some normal-looking LS strides as well as bounds and half-bounds. Three of the LS strides were symmetrical. Of these, two were classified as SF gaits and one as DC. At 22 days of age, animal No. 1 exhibited normal-looking LS and bounding strides. One of the LS strides was classifiable as a SF gait. At 29 days of age, the same three types of basic gaits seen earlier were observed 14 I Vilensky and Gankiewicz 8 days 22 days 29 days 58 days Fig. 1. Tracings from film of animal No. 1 locomoting at the noted ages. Gaits illustrated are: 8 days, LS, SF; 15 days, LS, SF; 22 days, LS, SF; 29 days, LS, SF; and 58 days, transverse gallop. (LS, half-bounds, and bounds). Only one LS stride was symmetrical; it was of the SF type. Note in Figure 1the tail posture during this week and also the elevated position of the hindquarters reflecting increased joint extension (cf. below). At 43 days of age, LS and bounding-type gaits were again observed. None of the LS sequences were sufficiently symmetrical to be defined. At 58 days, LS and galloping gaits were observed as well as two occasions of DS limb movements. The five symmetrical LS gaits were all of the DC type. The galloping sequences were transverse. Animal No. 2. Figure 2 illustrates some of the gaits used by this animal during the study period. On her day of birth this animal was able to sit but immediately toppled over when she tried to move. At 7 days of age, the animal was able to move across the platform but, with her ventral surface contacting the platform for much of the “stride” (Fig. 2). At 13 days of age, this animal still had some difficulty supporting her weight (Fig. 2) and many of her strides were aberrant in appearance. Of two symmetrical strides, one was a DS, DC type and one was a trot. At 21 days of age, bounding-type gaits as well as fairly normal-looking symmetrical gaits were observed. Within the latter, both LS and DS sequences were observed with one classifiable as DS, DC. Note the elevated body position compared to the earlier dates (Fig. 2). At 28 days of age, four symmetrical LS gaits were recorded; three were DC and one was a SF. Additionally, a half-bound sequence was observed. At 40 days, this animal only exhibited highly coordinated trotting-type gaits. On day 56, animal No. 2 exhibited trots and both rotary and transverse galloping sequences. Animal No. 3. This animal was uncooperative during the initial filmings and did not exhibit any usable locomotor movements until 28 days of age. At this age Locomotor Development in Vervets / 15 7 days 13 days 21 days .- 28 days 56 days Fig. 2. Tracings from film of animal No. 2 locomoting at the noted ages. Gaits illustrated are 7 days, not classifiable; 13 days, DS, DC; 21 days, DS, DC; 28 days, LS,DC; and 56 days, rotary gallop. she used normal-looking LS walking gaits. Two strides were symmetrical and classifiable as LC. At 42 days, this animal also showed LS gaits; one sequence was classifiable as DC and the other as a trot. At 64 days, both transverse and rotary-type galloping sequences were observed. One symmetrical sequence was classified LS, DC. Support Patterns Table I1 presents the limb support patterns as a percent for each of the animals for each date. For animal No. 1,three-legged combinations (triplets) supplied the greatest percentage of body support during all weeks, providing nearly 50%a t 8 days of age. The only clear trends evident over time for this animal are an increase in lateral support across all but the last date and a decrease in quadrupedal and three-legged support during the later dates compared to the earlier dates. For animal No. 2, three-legged combinations supplied the bulk of support during the early dates but were replaced by diagonal support at the final two ages. Thus, there is evidence of an overall trend to increase diagonal support over time with a corresponding trend toward decreased three-legged support. Table I1 also indicates a general decrease in four-foot support. Animal No. 3 showed very similar results to animal No. 2 with a decrease in triplet and quadrupedal support and an increase in diagonal support. This animal also had notable (30%)unipodal support at 64 days of age. Temporal Characteristics Figure 3A presents the hind-limb stance durations plotted against cycle duration for all animals across all weeks. The appropriate regression lines are depicted for each animal. The overall r value is .95 with a standard error of .04. The r values for the individual monkeys range from .92 to .99 with standard errors between .02 and .05. Clearly, the slopes are similar for all the animals. When the hind-limb stance/cycle duration data are examined over time for animals Nos. 1 and 2 (there were insufficient data for animal No. 3) no consistent trends in slope, intercepts, or correlation coefficients are apparent. 16 I Vilensky and Gankiewicz TABLE 11. Limb Support Patterns (%) for Each Animal at Each Age* Quad. Triplet Diag. Lh/Rh Lf/Rf Lat. Unipodal None l(8) l(15) l(22) l(29) l(43) l(58) 9.3 12.1 9.9 1 3.3 2.5 48 39.9 46.4 39 29.3 33.8 30.7 24.4 17.2 10.5 13.7 44 6.7 13.6 9.1 24 21 0.8 0 1.4 4.7 6.5 14.3 14.3 5.3 8 12.4 19 16.7 4.8 0 0.6 0.4 0 1.7 0 No. 2 (13) No. 2 (21) No. 2 (28) No. 2 (40) No. 2 (56) 29.6 13.2 8.6 18 10 47.6 58.8 35.7 23.4 23.5 22 20 31.3 58.6 40.8 0.4 8 9.7 0 17 0 0 7.1 0 2.8 0.4 0 6.6 0 5.5 0 0 0.6 0 0 No. 3 (28) No. 3 (42) No. 3 (64) 16 0 0 82 41 9 0 23 28 0 0 20 Animal & age (days) No. No. No. No. No. No. 2 26 13 0 10 30 0 0 0 *Quad. = quadrupedal support; triplet = support by any three-legged combination; Diag. = support by either pair of diagonal limbs; LMRh, Lf/Rf= support by either the left and right hind limbs or the left and right forelimbs; Lat.= support by either pair of ipsilateral limbs; Unipodal= support by any single limb; None= flight phase. Figure 3B depicts the hind-limb swing durations plotted against cycle duration for the three animals across all ages. For all the animals across all dates the r value equals .68 with a standard error of .04. The regression equation for animal No. 3 has a very high overall r value (.96) compared to the other two (.66 and .63), but that may simply reflect the relatively small number of strides available for this animal (cf. Table I). The regression lines for animals Nos. 1and 2 are similar while that for animal No. 3 is divergent. There are no consistent changes in slopes, intercepts, or correlation coefficients across the dates for animals Nos. 2 and 3. Figure 4A depicts plots of forelimb stance duration vs. cycle duration for animals Nos. 1 and 2 (insufficient data were available for animal No. 3). Remarkably, the composite slope, intercept, standard error, and r values for these data are identical to the hind-limb stance composite data (.77, - .05, .04, .95, respectively). The values for the individual animals (across all dates) are similar to the composite values. As with the hind-limb stance data there are no consistent changes in coefficients across time. Figure 4B depicts the forelimb swing durations for animals Nos. 1 and 2 plotted against cycle duration. As would be expected based on the stance data, the composite slope, intercept, standard error, and r values for forelimb swing are virtually identical to those for the hind limb. Interestingly, animal No. 1 has a notably higher r value across all the weeks (.75) than animal No. 2 (.48). There are no consistent changes in any of the forelimb swing regression values across time for either animal. Joint Angular Displacement Figure 5 depicts the hip displacement patterns for the individual monkeys for each date. On each plot the shaded area represents the range of values observed for an adult vervet over a wide assortment of speeds. Animal No. 1, beginning a t days 8 and 15 (Fig. 5A), shows somewhat normal-looking hip displacement patterns with distinct periods of flexion and extension which are generally within the adult Locomotor Development in Vervets / 17 I:: 0.30 I I I 0.00 0.20 0.40 0.60 0.80 1 .oo 0.30 0.20 0.1 0 4 0.00 0.00 I I I I 0.20 0.40 0.60 0.80 I I 1.oo CYCLE DURATION ( S ) Fig. 3. Plots of hind-limb cycle duration vs. hind-limb stance duration (A) and hind-limb swing durations (B) for all the animals across all ages with regression lines for each animal. Circles and solid line= animal No. 1; squares and dashed line= animal No. 2; triangles and dashed-dotted line = animal No. 3. range. During subsequent weeks this animal's hip displacement curves are generally normal in appearance and stay within the adult range (Fig. 5B,C). Animal No. 2's hip displacement curves are grossly abnormal a t 13 days (Fig. 5A). Even at 21 days, one of the depicted curves (Fig. 5B) shows very little flexion-extension oscillation while the other is more normal in appearance. At 28 days, one of the curves shows more flexion than is normal while the other is somewhat typical. During the final two filmings, the hip curves for this animal appear fairly typical (Fig. 5C). A hip displacement curve is only presented for animal No. 3 at 42 days and is mature looking. Adult monkeys show a quadriphasic knee displacement pattern (Fig. 6). For 18 I Vilensky and Gankiewicz 0.80 i 0.70 -0.60 -- 0.50 -- 0.40-0.30-0.20 -- 0.1 0 0.10-- 0.00 t 0.00 I 0.20 I 0.60 0.40 0.80 1 10 0.30 . ... 0.20 .. s . 0.1 0 0.00 0 0 I I I 0.20 0.40 , I 0.60 0.80 1.oo CYCLE DURATION (S) Fig. 4. Plots of forelimb cycle duration vs. forelimb stance duration (A) and forelimb swing duration (B) for animals Nos. 1 and 2 across all ages with regression lines for each animal. Symbols as in Figure 3. animal No. 1 a t 8 and 15 days, the knee displacement curves (Fig. 6A) are not consistent with each other and generally are not adultlike in appearance. Similarly, at 22 and 29 days, animal No. 1’sknee displacement patterns extend beyond the adult range and do not show much initial flexion (Fig. 6B). The patterns for this animal during the final two ages are more normal-looking in appearance although they still tend to show increased flexion (Fig. 6C). The curves for animal No. 2 at 13 days of age show grossly abnormal knee displacement patterns characterized by little oscillation and extreme flexion (Fig. 6A). At 21 and 28 days of life, the knee curves for animal No. 2 (Fig. 6B) are less abnormal but exhibit little initial flexure and show periods of flexion with values outside the normal range. During the final two filmings, the knee displacement Locomotor Development in Vewets / 19 130 120J A 110 100 90 80 70 60 50 40 30 20 10 -20 -30 -50 -40 110 100 n 90 w 80 70 cc 2z 60 50 40 30 20 10 -20 -30 -40t -50 0 I I I , 25 50 75 100 PERCENT OF CYCLE Fig. 5. A Plot of some hip displacement curves for animal No. 1 at 8 days (open circles) and 15 days (filled circles), and animal No. 2 at 13 days (filled squares). B Similar plot for animal No. 1 at 22 days (open circles) and 29 days (closed circles) and animal No. 2 at 21 days (open squares) and 28 days (filled squares). C: Similar plot for animal No. 1at 43 days (open circles) and 58 days (closed circles),animal No. 2 at 40 days (open squares) and 56 days (closed squares), and animal No. 3 at 42 days (open triangles). The shaded area in each plot represents the range of values exhibited by an adult animal. 20 I Vilensky and Gankiewicz 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 40-30-- Fig. 6. Plot of some knee displacement curves for the animals. A-C: Symbols and shaded area a s in Figure 5. Locomotor Development in Vervets I 21 patterns for animal No. 2 tend to show more normal initial flexion periods, but the secondary periods show greater than normal maximum flexion values. The knee displacement curve for animal No. 3 a t 42 days shows a normallooking pattern except for greater maximum flexion. The ankle joint’s displacement pattern in adult animals is similar to that of the knee (Fig. 7). Animal No. 1 a t 8 days of age shows an ankle displacement pattern which is highly abnormal and characterized by greater than normal flexion values during much of the cycle. At 15 days of age, the displayed patterns for this animal vary, with one being almost normal in appearance and range and the other two showing periods of abnormal flexion and atypical patterns. At 21 days of age, the curves for animal No. 1 reveal two fairly normal-looking patterns (although exhibiting a high degree of flexion) while the pattern displayed for day 29 is atypical (Fig. 7B). Two of the ankle displacement patterns shown for animal No. 1 on days 40 and 56 (Fig. 7C) are fairly typical while one (day 56) is not. The ankle displacement patterns shown for animal No. 2 at 13 days are grossly abnormal and display excessive flexion (Fig. 7A). For 21 and 28 days of age, abnormal patterns are again evident for this animal although the flexion is not as exaggerated (Fig. 7B). The patterns exhibited for the final two filmings are quite variable with one of them (day 40) looking fairly normal. For animal No. 3, the ankle displacement pattern exhibited in Figure 7C (42 days of age) is fairly normal in appearance although some flexion values exceed the adult range. DISCUSSION Gait Patterns By approximately 1 month of age all three infants exhibited relatively normal-looking symmetrical gaits as well as gaits with nonsymmetrical footfall patterns that were not characteristic gallops. Additionally, bounds and halfbounds were common at younger ages. One of the animals “walked as early as 8 days while another still had some difficulty supporting her weight at 13days. Most of the symmetrical gaits recorded were LS. Toward the later dates there was a trend toward gaits that emphasized diagonal limb coordination (DC or trots). True galloping patterns were first observed in the animals at about 2 months of age. Thus, vervets appear to achieve the mature complement of gaits by this age. Hurov [19821 states that a 15-day-old vervet made locomotor attempts which involved LS limb movements. The report also notes that an older animal (56-75 days) used both LS and DS gaits as well as an atypical gait. Hurov concludes that vervets reach locomotor maturity by a t least 81 days. This is based on the fact that one animal consistently used DS gaits by this age. However, since some adult animals readily use LS as well as DS gaits [Vilensky et al., 19881this is not a valid test for determining locomotor maturity. Nevertheless, the paucity of DS gaits exhibited by the infants in our study does suggest that further modifications of locomotor control mechanisms do occur after 2 months of age which result in a higher frequency of DS gaits. The only additional data on vervet locomotor development are provided by Chalmers [19721. This report states that five captive vervets achieved “steady” locomotion a t ages ranging from 31 to 49 days, with a mean of 36 days. Detailed studies on the locomotor development of monkey species other than vervets are not abundant except for the rhesus monkey. Unfortunately, much of the data are inconsistent for this species. Specifically, for Macaca mulatta, studies report walking as first appearing at 3 days [Castell & Sackett, 19731, 13 days 22 I Vilensky and Gankiewicz 1754 165 155 145 135 125 115 105 95 85 75 65 55 ! v, w LLI [y. ?I I 165 155 145 135 125 115 105 95 a5 I I I I 25 50 I 55 , 175 165 155 145 135 125 115 105 95 85 q _- 55 0 ---+-____( 75 100 PERCENT OF CYCLE Fig. 7. Plot of some ankle displacement curves for the animals. A-C: Symbols and shaded area as in Figure 5. Locomotor Development in Vervets / 23 [Foley, 19341, 1-5 days [Tinklepaugh & Hartman, 19321, 12 days [Lashley & Watson, 19131, 2 weeks [Hildebrand, 19671, and 3 weeks [Lawrence & Hopkins, 19761. Hines [19421 also provides detailed data on the development of gait in rhesus monkeys, reporting that the earliest type of diagonal progression is observed on days 1 and 2 of life. This variation in the noted ages of first walking probably reflects differing methodologies (mother-raised vs. isolates), different definitions of “walking,” and variability among rhesus neonates as in vervets (cf. above; [Taylor et al., 19801). Additional data on the early development of gait in other monkey species may be found in Rose  for baboons, Negayama et al.  for Japanese macaques, Rollinson and Martin [19811 for talapoin monkeys and mangabeys, and Chalmers  for Syke’s and DeBrazza’s monkeys and for mangabeys. Support The only clear trend evident from all of the support data is that three- and four-foot support decreases across time. Clearly, this demonstrates a decreased need for stability as the animals aged. Hurov  reports that for a vervet infant across days 56-75 of life, triplets supported the animal up to 74% of a locomotor cycle and quadrupedal support was used for up to 22% of a cycle. Our data suggest that vervets of that age do not normally require such stable support patterns. Accordingly, Peters [ 19801 states that although l-week-old kittens use only threeand four-foot limb support, by 4 weeks of age gaits that maximize support on diagonals are used. Our animals Nos. 2 and 3 also showed a trend toward increasing diagonal support as they aged. Temporal Parameters In adult vervets, stance durations decrease linearly with cycle duration while swing durations tend to remain constant for all cycle durations [cf. Vilensky et al., 19881. For the infant vervets, both stance and swing durations were significantly correlated with cycle duration although the latter was poorly correlated. This may reflect some immaturity in the animals’ locomotor control apparatus; however, since even more mature animals can show significant relationships between swing and cycle durations [Vilensky et al., 19881our findings are not conclusive evidence of locomotor control immaturity. Furthermore, the general similarity in regression lines and coefficients among the animals and between the fore- and hind limbs suggests that similar “hard-wired” locomotor control networks are operative in these animals at or before birth. Such a conclusion has also been reached regarding kittens as well as frogs, chickens, and rats [cf. Bradley & Smith, 19881. Accordingly, Taub  suggests that the motor programs for locomotion are probably already established within the primate CNS by the end of the second trimester of pregnancy. Finally, because there were no consistent changes in the slopes, intercepts, or correlation coefficients of the computed regression equations over time, we attribute the variation we did observe to the aberrant nature of some of the strides and low weekly sample sizes. Angles Animal No. 1exhibited much more normal-looking hip displacement curves at the earlier ages than did animal No. 2. By 28 days, however, this animal was also showing a fairly adultlike hip displacement pattern. The knee displacement curves were grossly abnormal for animals Nos. 1and 2 during the earliest weeks. During these weeks and during the middle weeks excessive flexion was often exhibited. By the final filmings, the patterns were 24 1 Vilensky and Gankiewicz much more normal in appearance, although exaggerated flexion was still apparent. The results for the ankle joint are basically similar to those for the knee except that the patterns appear to be somewhat more abnormal at every age, and even those from the final filming are typically not adultlike. Accordingly, human data [Vilensky et al., 19871 and cat data [Peters, 19801 suggest that the ankle may be the most labile hind-limb joint during locomotion, easily modifying its movement as locomotor demands change. In accord with our results, exaggerated flexion of the joints during early forms of locomotion has been reported for rhesus macaques [Hines, 19421, cats [Peters, 19801, and humans [Wickstrom, 19771. CONCLUSIONS 1. Based upon the temporal data reported here as well as a related study on air-stepping, the basic neural mechanisms for intra- and interlimb coordination in vervets are apparently operative at birth although gaits, support, and angular displacement patterns are not fully normal for many weeks after birth. 2. Muscle weakness andlor immaturity in postural control mechanisms probably creates weight support problems for neonates which generally prevent the intrinsic locomotor control circuitry from exhibiting mature-looking locomotion. As muscle strength andlor postural systems mature, the extrinsic appearance of locomotor movements becomes progressively more adultlike so t h a t by 6 weeks of age there are often only minor differences in limb angular displacement patterns compared to adults. 3. Very young infant vervets may use bounding gaits to achieve movement. This probably represents a n attempt by the CNS to overcome muscle weakness. Nevertheless, some animals are able to exhibit relatively normal-looking symmetrical gaits at a very young age (e.g., 8 days). 4. Subsequent to 6 weeks of age, vervets develop true galloping patterns, and in most animals, a preference for DS symmetrical gaits. ACKNOWLEDGMENTS We are grateful to Ms. Penny Wilson for assistance with the animals, to Mr. Gregory Duncan for assistance during filming, to Ms. Diana Andrews for typing the manuscript, and to Ms. Roberta Shadle for help with the illustrations. We also thank M.D. Rose and two anonymous reviewers for their constructive comments on a n earlier version of this paper. Support for this study was provided by the Indiana University School of Medicine. REFERENCES Bradley, N.S.; Smith, J.L. Neuromuscular patterns of sterotypic hindlimb behaviors in the first two postnatal months: I. Stepping in normal kittens. DEVELOPMENTAL BRAIN RESEARCH 38:37-52, 1988. Castell, R.; Sackett, G. Motor behaviors of neonatal rhesus monkeys: Measurement techniques and early development. DEVELOPMENTAL PSYCHOBIOLOGY 6: 191-202, 1973. Chalmers, N.R. Comparative aspects of early infant development in some captive cercopithecines. Pp. 63-82 in PRIMATE SOCIALIZATION. F.E. Poirier, ed. New York, Random House, 1972. Foley, J.P., Jr. First year development of a rhesus monkey reared in isolation. JOURNAL OF GENETIC PSYCHOLOGY 45:39105, 1934. Hildebrand, M. Analysis of the symmetrical gaits of tetrapods. FOLIA BIOTHEORETICA 6:l-22, 1966. Hildebrand, M. Symmetrical gaits of primates. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 26:119-130, 1967. Hildebrand, M. Analysis of asymmetrical Locomotor Development in Vervets / 25 gaits. JOURNAL OF MAMMALOGY 58: 131-156,1977. Hildebrand, M. The adaptive significance of tetrapod gait selection. AMERICAN ZOOLOGIST 20:255-267,1980. Hines, M.D. The development and regression of reflexes. postures. and Droaession in the young macaque. CARNEGIRINSTITUTE OF WASHINGTON PUBLICATION NO. 541. CONTRIBUTIONS TO EMBRYOLOGY' NO. 196, pp. 155-209, 1942. Hurov, J.R. Diagonal walking in infant vervet monkeys. AMERICAN JOURNAL OF PRIMATOLOGY 2:211-2 13,1982. Lashley, K.S.; Watson, J.B. Notes on the development of a young monkey. JOURNAL OF ANIMAL BEHAVIOR 3:114139,1913. Lawrence, D.G.; Hopkins, D.A. The development of motor control in the rhesus monkey: Evidence concerning the role of corticomotoneuronal connections. BRAIN 99: 235-254,1976. Negayama, K.; Kondo, K.; Itoigawa, N. Development of locomotor behavior in infant Japanese macaques (Macaca fuscata). ANNALES DES SCIENCES NATURELLES, ZOOLOGIE. PARIS 5:169-180. 1983. Peters, S.E. THE ONTOGENY OF LOCOMOTION: A BEHAVIOR AND MORPHOLOGICAL ANALYSIS OF THE DEVELOPING LOCOMOTOR SYSTEM OF THE DOMESTIC CAT. Doctoral Thesis, University of California, Davis, 1980. Rollinson, J.; Martin, R.D. Comparative aspects of primate locomotion, with special reference to arboreal cercopithecines. SYMPOSIUM ZOOLOGICAL SOCIETY OF LONDON, NO. 48, pp. 377-427, 1981. Rose, M.D. Positional behavior of olive baboons (Papio anubis) and its relationship to maintenance and social activities. PRIMATES 18:59-116.1977. Taub, E. Motor behavior following deafferentation in the developing and motorically mature monkey. Pp. 675-705 in NEURAL CONTROL OF LOCOMOTION. R.B. Herman; S. Grillner; P.S.G. Stein; D.G. Stuart, eds. New York, Plenum, 1976. Taylor, E.M.; Sutton, D.W.; Lindeman, R.C. Somatic reflex development in infant macaques. JOURNAL OF MEDICAL PRIMATOLOGY 9:205-210,1980. Tinklepaugh, O.L.; Hartman, C.G. Behavior and maternal care of the newborn monkey (Macaca mulatta-"M. rhesus"). JOURNAL OF GENETIC PSYCHOLOGY 40: 257-286, 1932. Vilensky, J.A.; Gankiewicz, E.; Gehlsen, G. A kinematic comparison of backward and forward walking-in humans. JOURNAL OF HUMAN MOVEMENT STUDIES 13: 29-50, 1987. Vilensky, J.A.; Gankiewicz, E.; Townsend, D.W. Effects of size on vervet (Cercopithecus aethiops) gait parameters: A crosssectional approach. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 76463-480, 1988. Vilensky, J.A.; Wilson, P.E.; Gankiewicz, E. An analysis of air-stepping in normal infant vervet monkeys. JOURNAL OF MOTOR BEHAVIOR (in press), 1989. Wickstrom, R.L. FUNDAMENTAL MOTOR PATTERNS. Philadelphia, Lea and Febiger, 1977. Winter, D.A. BIOMECHANICS OF HUMAN MOVEMENT. New York, Wiley, 1979.