The Relationship between Speed, Contact Time and Peak Plantar Pressure in Terrestrial Walking of Bonobos

Transcription

1 Folia Primatol 2004;75: DOI: / The Relationship between Speed, Contact Time and Peak Plantar Pressure in Terrestrial Walking of Bonobos Evie Vereecke a Kristiaan D Août a Dirk De Clercq b Linda Van Elsacker c Peter Aerts a a Laboratory for Functional Morphology, University of Antwerp, Wilrijk-Antwerp, b Laboratory for Movement and Sport Sciences, University of Ghent, Ghent, and c Royal Zoological Society of Antwerp, Antwerp, Belgium Key Words Plantar pressure ` Speed ` Primate gait ` Bonobo locomotion ` Vertical force ` Contact time Abstract The aim of this paper is twofold. Firstly, we investigate whether contact times, as recorded by pedobarographic systems during quadrupedal and bipedal walking of bonobos, can be used to reliably calculate actual velocities, by applying formulae based on lateral-view video recordings. Secondly, we investigate the effect of speed on peak plantar pressures during bipedal and quadrupedal walking of the bonobo. Data were obtained from 4 individuals from a group of bonobos at the Animal Park Planckendael. From our study, we can conclude that both walking speeds calculated from contact times and lower leg length or simply from recorded contact times are good estimators for walking speed, when direct observation of the latter is impossible. Further, it was found that effects of speed on peak plantar pressures and vertical forces are absent or at least subtle in comparison to a large variation in pressure patterns. In bonobos, the same pressure patterns are used at all walking speeds, and, in consequence, we do not expect major changes in foot function. Copyright 2004 S. Karger AG, Basel Introduction The idea that velocity has an effect on gait parameters has long been established. For instance in humans, walking velocity influences descriptive kinematic variables such as duty factor and step and stride frequencies [Vilensky and Gehl- Fax karger@karger.ch S. Karger AG, Basel /04/ $21.00/0 Accessible online at: E. Vereecke, Lab. for Functional Morphology, Dept. of Biology, University of Antwerp, Universiteitsplein 1 BE 2610 Wilrijk-Antwerp (Belgium) Tel , Fax evie.vereecke@ua.ac.be

2 Table 1. Subject data Individual Sex Age (years) Body mass (kg) Lower leg length (m) Speed range (m/s) Number of sequences bipedal quadrupedal Re M Un F Ko F Ho F sen, 1984] (a treadmill study) and dynamic variables such as vertical peak ground reaction forces and plantar pressures [Kimura et al., 1983; Hughes et al., 1991; Keller et al., 1996; Wunderlich, 1999; Perttunen and Komi, 2001]. In non-human primates, studies dealing with the effect of speed on gait are scarce [Kimura et al., 1983; Vilensky and Gehlsen, 1984; Reynolds, 1985; Kimura, 1990; Tuttle et al., 1992; Demes et al., 1994; Aerts et al., 2000], and there is no congruity as to the nature and quantitative importance of these effects. Kinematic and kinetic recording of non-human primate gait is less readily performed than that of human gait, since the subjects typically live either in their natural habitat or in a semi-natural zoo environment. The recent availability of highresolution pedobarographic systems may, however, enable researchers to obtain information on plantar pressures and vertical ground reaction forces under seminatural circumstances. Apart from plantar pressures, these plates also measure contact times. The aim of this study is twofold. Firstly, we will establish whether or not contact times, measured by a pedobarographic system, can be used as an alternative for directly measured walking speed in bonobos (Pan paniscus), one of the three African ape species which is especially interesting in relation to other apes and (early) hominids [McHenry, 1984; Zihlman, 1984; Kano, 1992]. Previous studies in humans [Vilensky and Gehlsen, 1984] and some non-human primates [Kimura et al., 1983; Demes et al., 1994] found a strong correlation between contact time and walking speed, and we first establish if this is also true for bonobos. We will see if these contact times can be used to reliably calculate actual velocities. Aerts et al. [2000] set down formulae based on lateral-view video recordings that allow for the calculation of walking speed from contact time and lower leg length. It will be examined if these formulae also apply when contact times, as measured with a pedobarographic system, are used as input. Secondly, and most importantly, we will analyse plantar pressure patterns of walking bonobos and relate them to corresponding walking speeds. Material and Methods For this study, we used data of 4 bonobos with a walking speed ranging from 0.4 to 2.5 m/s (n = 71; table 1). The data were collected during 2000 and 2001 in the Animal Park Planckendael (Belgium). Plantar pressures (Footscan, RSscan International) and threedimensional ground reaction forces (AMTI) were registered, and the lateral and frontal views were simultaneously recorded with 2 S-VHS cameras at 50 Hz (for more details, see Walking Speed in Bonobos Folia Primatol 2004;75:

3 D Août et al. [2001] and Vereecke et al. [2003]). All variables were standardised, using body weight or lower leg length as the characteristic dimension [Kimura, 1990; Aerts et al., 2000]. This measurement (knee to heel distance) was obtained for each animal from digitised lateral video images since no direct morphometric measurements were available and because this method is easily applicable in the field. All correlations and linear regressions were performed on log 10 -transformed data using SPSS 10.0 software. Prediction of Walking Speeds Based on lateral-view video recordings of bipedal and quadrupedal walking bonobos in the Planckendael Animal Park (Belgium), Aerts et al. [2000] established a set of power functions that describe all relationships between (scaled) spatiotemporal gait characteristics and (dimensionless) speed. Below, we transform these formulae, so that contact time and lower leg length are the sole input variables. This will allow us to calculate walking speeds from pressure-based contact times. The formulae of Aerts et al. [2000]: Bipedal walking: DSF = DS 0.38 duty factor = DS 0.15, Quadrupedal walking: DSF = DS 0.38 duty factor = DS 0.19 with DSF = dimensionless stride frequency = stride frequency (L/g) 1/2 ; DS = dimensionless speed = square root of Froude number = (v²/gl) 1/2 ; L = lower leg length; v = speed; g = gravitational acceleration, 9.81 m/s² The relationship between contact time and duty factor is defined as: duty factor = contact time (s) stride frequency (1/s) (1) If we multiply equation 1 with (L/g) 1/2, we can replace stride frequency by dimensionless stride frequency, since stride frequency (L/g) 1/2 is DSF: duty factor (L/g) 1/2 = contact time (s) DSF (2) In equation 2, we can replace duty factor and DSF by their expressions in relation to dimensionless speed (DS) as formulated by Aerts et al. [2000]: Bipedal walking: ( DS 0.15 ) (L/g) 1/2 = contact time ( DS 0.38 ) Quadrupedal walking: ( DS 0.19 ) (L/g) 1/2 = contact time ( DS 0.38 ) Since dimensionless speed (DS) is the square root of the Froude number (i.e. v 2 /gl), we finally derive a formula that allows us to calculate walking speed from lower leg length and contact time for bipedal and quadrupedal walking of bonobos: Bipedal walking: v = ( L /contact time g ) Quadrupedal walking: v = ( L /contact time g ) Measuring Walking Speeds In order to compare walking speed calculated from contact time and lower leg length (using the above formulae), i.e. the calculated walking speed, with the directly observed speed, i.e. the measured walking speed, we measured progression velocities on lateral video images of sequences which corresponded with the recorded plantar pressure data. We used a NAC XY co-ordinate recorder and our own custom-written software to calculate the average speed between two digitised frames. For calibration, we used a reference wall [D Août et al., 2001] on which we digitised 2 points, as far distant as possible, and gave the corresponding (horizontal) separation. Then we digitised 1 point (e.g. the posterior end of the bonobo, near the coccyx) on the first image of the sequence and the same point on the last image of the sequence. This gave us the measured walking speed of the sequence. It should be stressed that gait parameters in bonobos are highly variable [D Août et al., 2001; Vereecke et al., 2003] and that contact time, and hence calculated walking speed, reflect instantaneous velocity rather than average velocity. The velocity referred to here is the average walking speed over a selected bout, of multiple (4 5) strides. 268 Folia Primatol 2004;75: Vereecke/D Août/De Clercq/Van Elsacker/Aerts

4 Fig. 1. Assignment of the 6 foot regions on a plantar pressure image (Footscan, RSscan International). Comparison of Calculated and Measured Walking Speeds We analysed the relationship between measured and calculated walking speed for bipedal (n = 19) and quadrupedal (n = 52) sequences, to establish if walking speeds calculated from contact times and lower leg length (using the above-mentioned formulae) are a valid alternative to speeds measured directly. Seven outliers were detected (walking speeds >2.1 m/s), but comparison of correlations with and without outliers revealed no clear differences, so the outliers were not excluded from the data set. Measured and calculated walking speeds were log transformed, because both variables lack normality (Shapiro-Wilks parameter <0.9). A trend line was plotted, and a correlation coefficient was calculated for bipedalism and quadrupedalism separately to quantify their relationship. To illustrate the relationship between recorded contact times and measured walking speeds, we also give a scatter plot of contact time versus measured walking speed with the concomitant trend line and correlation coefficient. The Effect of Walking Speed on the Peak Plantar Pressures To investigate the correlation between the peak plantar pressure and the measured and calculated walking speeds, respectively, a scatter plot was made for each morphological region of the foot (i.e. heel, lateral and medial midfoot, the four lateral metatarsal heads, the four lateral toes and the hallux; Footscan Multistep software; n = 71; fig. 1). The peak plantar pressure is the maximal plantar pressure under a specific foot region during the stance phase of that foot. A correlation coefficient was calculated for each foot region to quantify the relationship between peak plantar pressure and walking speed. Following the same method, we looked at the effect of speed on peak vertical forces. The peak forces were obtained by multiplying the plantar pressure by the surface of the specific foot area (heel: 6.75 cm 2 ; lateral midfoot: cm 2 ; medial midfoot: cm 2 ; metatarsals: 13.5 cm 2 ; toes II V: 13.5 cm 2 ; hallux: 6.75 cm 2 ). To combine data for 4 different-sized animals, all variables were standardised using body weight or lower leg length as the characteristic dimension. Peak plantar pressures and vertical forces were scaled to body weight, hereafter referred to as SPPP (scaled peak plantar pressure) and SPVF (scaled peak vertical Walking Speed in Bonobos Folia Primatol 2004;75:

5 force), and the square root of the Froude number was taken as dimensionless speed, for comparison of the different sequences. Subsequently, all (dimensionless) variables were log transformed [log(variable + 1) to avoid problems with zero values], since they all lacked normality (due to the presence of some outliers; Shapiro-Wilks <0.9). Results Calculated Walking Speed versus Measured Walking Speed Logically, and in agreement with previous studies in humans [Vilensky and Gehlsen, 1984] and non-human primates [Kimura et al., 1983; Demes et al., 1994], contact time and walking speed are negatively correlated during bipedal (n = 19; Pearson correlation coefficient = 0.54; linear regression: y = 0.495x ; R² = 0.291; t = 2.643; d.f. = 17; p < 0.05; fig. 2a) and quadrupedal (n = 52; Pearson correlation coefficient = 0.42; linear regression: y = 0.514x 0.082; R² = 0.180; t = 3.315; d.f. = 50; p < 0.01; fig. 2a) walking of bonobos. During bipedalism, there is an equally strong correlation between measured walking speed and pressurebased contact times and the calculated walking speed, respectively (from contact time and lower leg length; n = 19; Pearson correlation coefficient = 0.55; linear regression: y = 0.263x , R 2 = 0.300; t = 2.698; d.f. = 17; p < 0.05; fig. 2b), whereas during quadrupedalism, we found a slightly weaker, but still significant, correlation with the measured walking speed (n = 52; Pearson correlation coefficient = 0.34; linear regression: y = 0.265x 0.032, R 2 = 0.114; t = 2.539; d.f. = 50; p < 0.05; fig. 2b). Effect of Walking Speed We did not find a significant correlation coefficient (p > 0.05; table 2) between measured walking speed and SPPP for any of the foot regions during terrestrial walking (both bipedal and quadrupedal) of bonobos (fig. 3 and table 2). Taking the calculated or measured walking speed as the independent variable yields the same result. We also investigated the relationship between speed and SPVF under the 6 foot regions, since most existing publications have looked at the effect of speed on vertical forces instead of on plantar pressures, but, again, no significant correlations were found (p > 0.05; table 3), except for the toe region where the peak vertical force increases with increasing (measured) walking speed (n = 71; Pearson correlation coefficient = 0.23; linear regression: y = 0.74x ; R² = 0.06; t = 2.0; d.f. = 69; p < 0.05). Fig. 2. a Scatter plot of log measured walking speed versus log contact time for bipedal (b; n = 19) and quadrupedal ( ; n = 52) sequences. A linear regression line with a 95% mean prediction interval is drawn for bipedal (solid line) and quadrupedal (dotted line) walking. b Scatter plot of log measured versus log calculated walking speed for bipedal (b; n = 19) and quadrupedal ( ; n = 52) sequences. A linear regression line with a 95% mean prediction interval is drawn for bipedal (solid line) and quadrupedal (dotted line) walking. Regression equations and R² values are drawn on the graph. 270 Folia Primatol 2004;75: Vereecke/D Août/De Clercq/Van Elsacker/Aerts

6 Walking Speed in Bonobos Folia Primatol 2004;75:

7 3 272 Folia Primatol 2004;75: Vereecke/D Août/De Clercq/Van Elsacker/Aerts

8 Fig. 4. Scatter plots of log SPPP (times body weight) in function of log measured dimensionless speed (DS1) for the heel (a) and lateral midfoot (b) region, respectively. A linear regression line is only drawn for quadrupedal walking (b; solid line; n = 52; p > 0.05), since the correlation between log SPPP and log DS1 was not significant for bipedal walking (H; n = 19). Additionally, the effects of individual (Re, Un, Ko, Ho) and type (bipedalism, quadrupedalism) on SPPP and SPVF were analysed using a multiple-way ANOVA (SAS 8.01 for Windows). The effect of type on SPPP was not significant, except for the heel (p < 0.05) and the lateral midfoot region (p < 0.05). To make a more profound analysis of this type effect, we performed separate regressions for bipedal and quadrupedal walking (fig. 4). During quadrupedal walking, we found a significant correlation between measured walking speed and SPPP under the heel (n = 52; Pearson correlation coefficient = 0.29; linear regression: y = 0.74x ; R² = 0.08; t = 2.1; d.f. = 50; p < 0.05) and under the lateral midfoot region (n = 52; Pearson correlation coefficient = 0.40; linear regression: y = 0.77x ; R² = 0.16; t = 3.1; d.f. = 50; p < 0.01), whereas no significant correlation was found during bipedalism (p > 0.05) or during quadrupedalism and bipedalism jointly (p > 0.05). A significant effect of individual was found for the SPPP of the heel (p < 0.05), metatarsal (p < 0.05) and hallux regions (p < 0.05), meaning that some individuals experience a significantly different loading of these regions compared to others. The variance explained by intersubject ( individual ) Fig. 3. Scatter plots of log SPPP (times body weight) in function of log measured (DS1) and log calculated (DS2) dimensionless speed, respectively. For none of the foot regions were the correlations statistically significant (p > 0.05). All graphs represent bipedal and quadrupedal walking sequences jointly. Walking Speed in Bonobos Folia Primatol 2004;75:

9 Table 2. Statistical analysis of the correlation between the scaled peak plantar pressure and the measured (DS1) and calculated (DS2) dimensionless walking speed, respectively, under 6 different foot regions, for bipedal and quadrupedal walking (n = 71, d.f. = 69) R² r p DS1 DS2 DS1 DS2 DS1 DS2 Heel Lateral midfoot Medial midfoot Metatarsal heads Toes II V Hallux All variables were log transformed [log(y + 1)]. r = Pearson correlation coefficient; p = probability (significance level = 0.05). Table 3. Statistical analysis of the correlation between the scaled peak force and the measured (DS1) and calculated (DS2) dimensionless walking speed, respectively, under 6 different foot regions, for bipedal and quadrupedal walking (n = 71, d.f. = 69). R² r p DS1 DS2 DS1 DS2 DS1 DS2 Heel Lateral midfoot Medial midfoot Metatarsal heads Toes II V * Hallux All variables were log transformed [log(y + 1)]. r = Pearson correlation coefficient; p = probability (significance level = 0.05). differences was 13.4% in the heel, 50.1% in the metatarsal and 24.7% in the hallucial region. There was also a significant interaction effect between measured walking speed and type for the lateral midfoot, pointing to a difference in regression slope, between measured walking speed and SPPP, during bipedalism and quadrupedalism (p < 0.05). Taking SPVF as the dependent variable in the multipleway ANOVA yields the same results. Discussion If we compare the correlation between contact time and measured walking speed during bonobo walking with that observed during human walking (r > 0.9, slope ~ 1 in Kimura et al. [1983]; the treadmill study of Vilensky and Gehlsen 274 Folia Primatol 2004;75: Vereecke/D Août/De Clercq/Van Elsacker/Aerts

10 [1984]), monkey (r ~ >0.9, slope ~ 1 [Kimura et al., 1983; Vilensky and Gehlsen, 1984]) and ape walking (r > 0.7, slope > 0.8 [Kimura et al., 1983; Demes et al., 1994]), we observe that our correlation coefficients (r = 0.42 and r = 0.54) and slopes (~ 0.5) are strikingly lower. The weaker correlation between contact time and walking speed may be caused by the unsteady gait of the bonobos [D Août et al., 2001], but also by using measured walking speed, which is the average speed over multiple steps, to represent progression velocity. Furthermore, regarding the shallower slopes during bonobo locomotion, contact times (and hence duty factors) decrease less rapidly with increasing speed during bonobo walking than during human walking (see also Reynolds [1985]). It was also noticed that at a given walking speed, contact times are longer during bipedalism than during quadrupedalism of bonobos. This could be a way to enhance stability during bipedal locomotion. For bipedal walking, the correlation between measured walking speed and contact time (r = 0.54) is similar to that between measured and calculated walking speeds (r = 0.55). For quadrupedal walking, however, the correlation between measured walking speed and contact time (r = 0.42) is higher than that between measured walking speed and calculated walking speed (r = 0.34). Indeed, correlations for quadrupedal walking were generally lower than for bipedal walking, no doubt reflecting the larger sample size (4 rather than 1 subject) and hence higher variability of the former. Nevertheless, in view of the significant, if weak, correlations that did exist, we regard both calculations of speed from contact time, and raw contact time, as reliable estimators of walking speed, when direct observation of the latter is impossible and general patterns over a wide range of speeds are studied. It was found that the SPPP and/or SPVF increase markedly during human walking with increasing speed [Soames and Richardson, 1978; Kimura et al., 1983; Hughes et al., 1991; Keller et al., 1996; Li et al., 1996; Wunderlich, 1999] or walking cadence [Winter, 1990; Zhu et al., 1995]. But not all regions under the human foot are equally affected by walking speed: whilst the loading under most regions increases with speed, loading under the lateral and central forefoot tend to decrease with increasing walking speed [Andriacchi et al., 1977; Jansen and Jansen, 1978; Rosenbaum et al., 1994; Perttunen and Komi, 2001]. It has also been reported that peak vertical forces increase with speed during walking of common chimpanzees [Reynolds, 1985; Demes et al., 1994]. However, when we looked at the effect of speed on SPPP and SPVF during walking of bonobos, we found that both SPPP and SPVF are generally independent of walking speed, which is not in agreement with previous reports on human and non-human primate walking. We only found a significant increase in SPVF with speed for the toe region (bipedal and quadrupedal sequences taken together), and only during quadrupedal walking of bonobos was there a significant effect of walking speed on the SPPP and SPVF under the heel and lateral midfoot region. This increase in SPPP and SPVF with speed under the heel and lateral midfoot region can be ascribed to the higher impact forces at higher walking speed, since initial foot contact is made, more or less simultaneously, with the heel and lateral midfoot. Accordingly, an increase in the SPVF under the toe region with speed might be due to the higher propulsive forces occurring under the lateral toes at higher speeds, since propulsion is mainly generated by the lateral toes in bonobos [Vereecke et al., 2003]. But how can we explain the absence of a speed effect under the other foot regions during quadrupedalism and under all foot regions during bipedalism? Walking Speed in Bonobos Folia Primatol 2004;75:

11 In theory, the sum of the impulses (i.e. the integral of the vertical force over contact time) of all limbs in a cycle, divided by the cycle duration, equals the body weight (the impulse-momentum approach [Kimura, 1990]). When the shape of the force curve is invariable, but duty factors decrease, as they do when speed increases (see Aerts et. al. [2000] for bonobos), the peak forces will increase. In practice, this relationship is only manifest in uniform locomotion and if a large number of successive steps/strides are incorporated. Usually, this is not the case (in particular when studying apes) and steps or strides are examined separately, resulting in an integral that is somewhat higher or lower than the body weight, as a consequence of the variation in steps. In view of the large interstep variation during bonobo walking, and especially during bipedalism [D Août et al., 2001; Vereecke et al., 2003], it is thus not surprising that we only found an effect of speed during quadrupedalism, and solely under 2 foot regions. It is possible that the vertical peak forces do not increase with increasing walking speed as a consequence of the high interindividual and/or interstep variability of bonobo walking. Besides, Reynolds [1985] found that for common chimpanzees (Pan troglodytes) the increase in peak forces with speed is slower than for other primates and humans and suggested that this might be due to a slower decrease in duty factor with increasing speed. This is in agreement with the relationship between contact time and walking speed in our study (see above) and might thus partly explain the subtle or absent speed effect on peak forces. Schmitt [1999] found also that limb compliance affects contact times in Old World monkeys. Limb compliance is characterised by longer contact times at equal speeds, hereby lowering the vertical peak reaction forces. It has not yet been investigated if apes also adopt a compliant walking gait to reduce and reorient peak ground reaction forces, but if this is the case, then limb compliance might explain the slower rise in contact times with speed in chimpanzees and bonobos. If we look, however, at the results presented by Kimura et al. [1983], it may be asked how it is possible that vertical forces decrease with increasing speed during bipedal locomotion of common chimpanzees (P. troglodytes). In theory, this means that duty factors should increase with increasing speed, which is very improbable. The observed decrease in vertical force could be due to a large gait variability, but since only one chimpanzee was studied (n = 24) gait variability would solely be due to interstep variation. Another possibility is that the shape of the force curves changes drastically with speed, resembling a square wave rather than assuming a parabolic shape. The second explanation seems the most plausible, and our data also suggest that the shape of the vertical force curve will change with speed and hence compensate for the decrease in duty factor. Further research should reveal if this is actually the case. Conclusions During terrestrial walking of bonobos, there is a negative correlation between contact time and walking speed, and speeds calculated from contact times and lower leg length are positively correlated with walking speeds as measured on lateral-aspect video. Therefore, we believe that calculated walking speeds, or simply recorded contact times, can be good estimators of true walking speed. Hence, the 276 Folia Primatol 2004;75: Vereecke/D Août/De Clercq/Van Elsacker/Aerts

12 acquisition of data using only a pressure mat can be justified if the relationship between speed and contact time has been previously established. In bonobos, the same pressure patterns are used at all walking speeds and, in consequence, we do not expect big changes in foot function with speed. Therefore, pressure data from different speeds can be pooled in further analysis. We stress, however, that we only examined the effect of walking speed on peak plantar pressures and vertical forces and that no clear effect was found. There might, however, be an effect of speed on other foot-related parameters, such as the timing of plantar (peak) pressures, foot position, roll-off pattern etc. which are yet unexplored. Acknowledgments This study was supported by a grant from the FWO-Flanders, and structural support was given by the Flemish Government through the Centre for Research and Conservation (Royal Zoological Society of Antwerp). References Aerts P, Van Damme R, Van Elsacker L, Duchêne V (2000). Spatio-temporal gait characteristics of the hind-limb cycles during voluntary bipedal and quadrupedal walking in bonobos (Pan paniscus). American Journal of Physical Anthropology 111: Andriacchi TP, Ogle JA, Galante JO (1977). Walking speed as a basis for normal and abnormal gait measurements. Journal of Biomechanics 10: D Août K, Aerts P, De Clercq D, Schoonaert K, Vereecke E, Van Elsacker L (2001). Studying bonobo (Pan paniscus) locomotion using an integrated setup in a zoo environment: Preliminary results. Primatologie 4: Demes B, Larson SG, Stern JT, Jungers WL, Biknevicius AR, Schmitt D (1994). The kinetics of primate quadrupedalism: Hindlimb drive reconsidered. Journal of Human Evolution 26: Hughes J, Pratt L, Linge K, Clark P, Klenerman (1991). Reliability of pressure measurements The EMED F system. Clinical Biomechanics 6: Jansen EC, Jansen KF (1978). Vis-velocitas-via: Alteration of foot-to-ground reaction forces during increasing speed of gait. In Biomechanics VI-A (Asmussen E, Jorgensen K, eds.), pp Baltimore, University Park Press. Kano T (1992). The Last Ape: Pygmy Chimpanzee Behavior and Ecology. Stanford, Stanford University Press. Keller TS, Weisberger AM, Ray JL, Hasan SS, Shiavi RG, Spengler DM (1996). Relationship between vertical ground reaction force and speed during walking, slow jogging, and running. Clinical Biomechanics 11: Kimura T, Okada M, Yamazaki N, Ishida H (1983). Speed of the bipedal gaits of man and nonhuman primates. Annales des Sciences Naturelles (Zoologie, Paris), 13e série 5: Kimura T (1990). Voluntary bipedal walking of infant chimpanzees. In Gravity, Posture and Locomotion (Jouffroy FK, Stack MH, Niemitz C, eds.), pp Firenze, Il Sedicesimo. Li Y, Crompton RH, Alexander RMcN, Günther MM, Wang WJ (1996). Characteristics of ground reaction forces in normal and chimpanzee-like bipedal walking by humans. Folia Primatologica 66: McHenry HM (1984). The common ancestor: A study of the postcranium of Pan panisucs, Australopithecus and other hominids. In The Pygmy Chimpanzee: Evolutionary Biology and Behavior (Susman RL, ed.), pp New York, Plenum Press. Perttunen JR, Komi PV (2001). Effects of walking speed on foot loading patterns. Journal of Human Movement 40: Reynolds TS (1985). Stresses on the limbs of quadrupedal primates. American Journal of Physical Anthropology 67: Rosenbaum D, Hautmann S, Gold M, Claes L (1994). Effects of walking speed on plantar pressure patterns and hindfoot angular motion. Gait and Posture 2: Schmitt D (1999). Compliant walking in primates. Journal of Zoology 248: Soames RW, Richardson RPS (1978). Stride length and cadences: Their influence on ground reaction forces during gait. In Biomechanics IX-A (Winter DA, Norman RW, Wells RP, Hayes KC, Patla AE, eds.), pp Champaign, Human Kinetics Publishers. Walking Speed in Bonobos Folia Primatol 2004;75:

13 Tuttle RS, Webb DM, Tuttle NI, Baskh M (1992). Footprints and gaits of bipedal apes, bears and barefoot people: Perspectives on Pliocene tracks. In Topics in Primatology (Matano S, Tuttle RH, Ishida H and Goodman M, eds.), vol 3: Evolutionary Biology, Reproductive Endocrinology and Virology, pp Tokyo, University of Tokyo Press. Vaughan CL, Damiano DL, Abel MF (1997). Gait of normal children and those with cerebral palsy. In Three-Dimensional Analysis of Human Locomotion (Allard P, Cappozzo A, Lundberg A, Vaughan CL, eds.), pp Chichester, Wiley & Sons. Vereecke E, D Août K, De Clercq D, Van Elsacker L, Aerts P (2003). Dynamic plantar pressure distribution during terrestrial locomotion of bonobos (Pan paniscus). American Journal of Physical Anthropology 120: Vilensky JA, Gehlsen G (1984). Temporal gait parameters in humans and quadrupeds: How do they change with speed? Journal of Human Movement Studies 10: Winter DA (1990). The Biomechanics and Motor Control of Human Gait: Normal, Elderly and Pathological, 2nd ed. Waterloo, University of Waterloo Press. Wunderlich R (1999). Pedal Form and Plantar Pressure Distribution in Anthropoid Primates. PhD dissertation, State University of New York, Stony Brook. Zihlman AL (1984). Body build and tissue composition in Pan paniscus and Pan troglodytes, with comparison to other hominoids. In The Pygmy Chimpanzee: Evolutionary Biology and Behavior (Susan R, ed.), pp New York, Plenum Press. Zhu H, Wertsch JJ, Harris GF, Alba HM (1995). Walking cadence effect on plantar pressures. Archives of Physical Medicine and Rehabilitation 76: Folia Primatol 2004;75: Vereecke/D Août/De Clercq/Van Elsacker/Aerts