Human encephalization and developmental timing

Journal of Human Evolution 49 (2005) 762e776

Human encephalization and developmental timing* Lucio Vinicius* Leverhulme Centre for Human Evolutionary Studies, University of Cambridge, Downing Street, Cambridge CB2 3DZ, United Kingdom Received 6 August 2004; accepted 11 August 2005

Abstract Human evolution is frequently analyzed in the light of changes in developmental timing. Encephalization in particular has been frequently linked to the slow pace of development in Homo sapiens. The ‘‘brain allometry extension’’ theory postulates that the progressive extension of a conserved primate brain allometry into postnatal life was the basis for brain enlargement in the human lineage. This study shows that published primate and human growth data do not corroborate this model. Instead, the unique encephalization of H. sapiens is alternatively described as the result of evolutionary changes in three aspects of developmental timing. The first is a moderate extension in the duration of brain growth relative to our closest extant relatives, contrary to the view that human brain growth is drastically prolonged into postnatal life. Second, humans evolved a derived brain allometry in comparison with chimpanzees and early hominins. Third, humans (and other anthropoid primates to a lesser degree) display a significant retardation in early postnatal body growth in comparison with other mammals, which directly affects adult encephalization in our species. The rejection of the ‘‘brain allometry extension’’ model may require a reevaluation of the adaptive scenarios proposed to explain how human encephalization evolved. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: Homo sapiens; Primates; Encephalization; Brain allometry; Growth

Introduction Changes in developmental timing played a significant role in human evolution (Smith and *

The data used for this study are available on the author’s academic website: http://www.human-evol.cam.ac.uk/Members/ Vinicius/Publications. * Tel.: C44 1223 335 461; fax: C44 1223 335 460. E-mail address: [email protected]

Tompkins, 1995; Minugh-Purvis and McNamara, 2002). Prolonged development is associated with traits such as large body size, increased encephalization, and extended lifespan in Homo sapiens and primates in general. For example, Smith (1989) showed that primate brain size is strongly correlated with duration of dental development and with some aspects of life history timing (gestation length, age at weaning, age at first reproduction).

0047-2484/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.jhevol.2005.08.001

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

Various studies of brain and body growth in primates (Count, 1947; Holt et al., 1975; Vrba, 1998; Leigh, 2001, 2004; Rice, 2002) have offered further evidence for a link between developmental retardation in H. sapiens and the evolution of large brain size. Despite the general consensus that humans grow and live slowly, arguments for developmental retardation are sometimes contradictory. For example, it has been argued that postnatal brain growth in humans is extremely prolonged and proceeds at ‘‘fetal’’ rates for nearly a year, in comparison with a few weeks in chimpanzees (Holt et al., 1975; Gould, 1977; Martin, 1983; Coqueugniot et al., 2004). In direct contrast, Vrba (1998) and Rice (2002) showed that human brain growth prolongation relative to chimpanzees amounts to less than 30%, while Leigh (2004) failed to identify any prolongation. Additionally, while many authors (Count, 1947; Martin, 1983; Smith and Tompkins, 1995) have argued that the developmental extension of a conserved primate brain growth allometry was the key to human encephalization, it now seems clear that the changes in timing that characterize human brain growth (Vrba, 1998; Rice, 2002; Leigh, 2004) do not coincide with the changes in timing that characterize human body growth (Laird, 1967; Bogin, 1999; Leigh, 2001). Moreover, diverging patterns of brain and body growth evolution suggest changes in brain growth allometry, rather than its conservation. Finally, while many authors have asserted that the prolongation of ‘‘fetal’’ brain growth rates into postnatal life in H. sapiens has led to a decrease in neonatal brain size (as a proportion of adult brain size) from 40% or 50% in chimpanzees to only 25% in humans (Schultz, 1940, 1941; Holt et al., 1975; Passingham, 1982; Martin, 1983; Dienske, 1986; Foley and Lee, 1991; Smith and Tompkins, 1995; Hrdy, 1999; Coqueugniot et al., 2004), it is possible to find a recent reference to a similar ratio of around 33% for both species (Kennedy, 2005). In this paper, I argue that such contradictions point to weaknesses in the mainstream theory of human encephalization, namely Count’s ‘‘brain allometry extension’’ model. The purpose of this paper is to redefine the role of developmental

763

retardation in the evolution of human encephalization by reanalyzing available data on brain and body growth in humans and other primates. Timing of brain and body growth in humans Vrba (1998) proposed a link between developmental timing and brain size by fitting brain growth in humans and chimpanzees (the closest extant model of early hominin brain growth) to a common four-phase polynomial curve. In linear (double logarithmic) form, human brain growth is represented by Yh ZA0 Ck1 ðT1 ÞCk2 ðT2
T1 Þ Ck3 ðT3
T2 ÞCk4 ðT4
T3 Þ where Yh is adult human brain weight in grams, A0 is log(intercept), ki is the growth coefficient for phase i, and Ti is log(conception age in months at end of phase i). Calculating brain size at a given phase excludes contributions from subsequent phases (e.g., calculating brain size at age T with T2 ! T ! T3 excludes the factor k4(T4
T3) from the computation). Using the parameters estimated for H. sapiens, Vrba (1998) obtained Yh Z
1:39C4:74ð0:64ÞC3ð0:97
0:64Þ C1:05ð1:26
0:97Þ C0:24ð2:06
1:26ÞZ3:13Z1349 g The curve describes the known decrease in brain growth rates during ontogeny (since k1 O k2 O k3 O k4), and suggests the existence of two prenatal phases ending at 4.4 months and 9.3 months from conception, respectively. The latter closely approximates gestation length in humans. The first postnatal phase (the third from conception) lasts for 8.9 months, and is followed by a fourth and final phase of slower brain growth ending at 115 months (or 9.6 years) from conception. By assuming that all of the curve parameters (number of phases, intercept, growth coefficients) are the same in humans and chimpanzees, except for the duration of the phases (the values of Ti), Vrba (1998) proposed the hypothesis of ‘‘proportional growth prolongation’’ (or PGP). According to

764

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

the PGP hypothesis, the difference in brain size between chimpanzees and humans can be explained solely by proportional (i.e., equal) changes in the duration of all four brain growth phases. The statistical comparison between human and chimpanzee data showed that the duration of brain growth is about 29% longer in H. sapiens. In other words, each stage and phase transition Ti in the chimpanzee curve is shorter by a factor
log(1.29) Z
0.11. Adult chimpanzee brain size (Yc) is therefore given by the curve Yc Z
1:39C4:74ð0:53ÞC3ð0:86
0:53Þ C1:05ð1:15
0:86Þ C0:24ð1:95
1:15ÞZ2:61Z406 g As in humans, the end of the second brain growth phase in chimpanzees (about 7.2 months from conception) also approximates the pregnancy term. The predicted duration of the first postnatal phase is 6.9 months, and brain growth ends at 7.4 years from conception. Despite the good fit to data in both species, doubts can be raised as to how accurately the parameters of Vrba’s multiphasic growth model (number and duration of brain growth phases, and growth rates) can be estimated from cross-sectional data (this issue is further discussed below). However, in favor of Vrba’s approach, Rice (2002) tested the PGP hypothesis without using a brain growth equation and reached similar conclusions. The advantage of Vrba’s (1998) formulation of an explicit brain growth curve is that her model is easier to manipulate analytically. For example, the effect of the ‘‘proportional growth prolongation’’ of a baseline chimpanzee-like brain growth curve on a putative hominin descendant (YD) is given by YD Z YA C k1log(a), where YA is the logarithm of the reference chimpanzee brain size, k1 is the growth coefficient for phase one from Vrba (1998), and a is the prolongation factor (Vinicius and Lahr, 2003). The formula leads to a similar estimate of a Z 1.29 in the comparison between modern humans and chimpanzees. In summary, the evidence suggests that the brain size increase in the H. sapiens lineage resulted from the proportional prolongation of a brain growth pattern shared by chimpanzees, early hominins, and modern humans.

A more complex picture is derived from studies of the evolution and timing of human body growth. The existence of a general sigmoidal growth curve characteristic of mammals and birds, to which primates would be the most remarkable exception, was postulated decades ago (Laird, 1967). Primates would deviate from the general pattern due to the evolution of new life history stages (such as the juvenile phase), which extend the duration of growth, a process that reaches the extreme in slowly developing humans (Bogin and Smith, 1996; Bogin, 1997, 1999). Charnov and Berrigan (1993) estimated that primate postnatal body growth rates were only 40% of the expected value for a similarly sized mammal. Humans grow more slowly, even by primate standards, displaying growth rates below 30% of the mammalian average (Hill and Hurtado, 1996). However, such reduction in body growth rates is not uniform. Leigh (2001) showed that, while human body growth rates fall within the expected range of variation in primates during pregnancy and the adolescent growth spurt, low growth rates are distributed over a prolonged period between birth and the subadult spurt (Leigh, 2001). Thus, a reduction in rates of early postnatal growth may have engendered an overall prolongation of the growth process in H. sapiens. Encephalization and the ‘‘brain allometry extension’’ model Changes in developmental timing have also been implicated as a key factor in human encephalization, a measure of brain size relative to body size (Holt et al., 1975; Gould, 1977; Martin, 1983; Smith and Tompkins, 1995). Possibly without exception, those claims rely on the classic model proposed by Count (1947), which consists of three postulates: (1) ontogenetic brain allometry (the rate of brain to body growth) is constant during development; (2) ontogenetic brain allometry is conserved across primates, and displays a marked upward shift (i.e., a change in intercept) in comparison with other mammals; and (3) encephalization of humans resulted from an extreme prolongation of the common primate brain allometry into later developmental stages. As summarized by Gould

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

(1977: 372), humans would ‘‘have preserved this basic [primate brain allometry] pattern with one crucial quantitative modification: we have prolonged the high fetal slope well into postnatal life and achieved thereby our remarkable encephalization.’’ Estimates of such prolongation of fetal allometry vary, but references to a dramatic difference between humans and other primates are common. Holt et al. (1975: 28) claimed that the transition from high fetal rates to residual growth occurs ‘‘just after birth for chimpanzees, and about 2 years of age postnatally for the human.’’ According to Martin (1983: 27), fetal brain growth allometry is still maintained ‘‘until at least 12 months after birth’’ but most authors seem to believe that humans ‘‘manage to sustain high fetal rates for a full year after birth; chimpanzees and macaques may do so only for a month or two’’ (Smith and Tompkins, 1995: 270). A noticeable effect of the increased role of postnatal brain growth in humans would be a significantly smaller ratio of neonatal to adult brain size in our species, decreasing from around 50% in chimps and apes to only 25% in H. sapiens (Schultz, 1940, 1941; Holt et al., 1975; Passingham, 1982; Martin, 1983; Dienske, 1986; Foley and Lee, 1991; Smith and Tompkins, 1995; Hrdy, 1999; Coqueugniot et al., 2004). For this reason, studies in the evolution of hominin life history are, as a rule, accounts of the behavioral, social, and physiological mechanisms evolved to sustain the extension of fetal brain allometry rates into postnatal life and to cope with the energetic challenge of growing 75% of our brain after birth (Martin, 1983; Foley and Lee, 1991; Aiello et al., 2001). Difficulties with the ‘‘brain allometry extension’’ theory The ‘‘brain allometry extension’’ scenario, originally proposed by Count (1947), has played a major role in debates on human encephalization. However, recent studies in brain and body growth reviewed above (Vrba, 1998; Leigh, 2001, 2004; Rice, 2002) directly contradict some central features of the model. Three major difficulties with the ‘‘brain allometry extension’’ theory can be identified:

765

(1) Is brain allometry conserved across primates? Count (1947) postulated the conservation of brain growth allometry (intercept and allometric coefficient) in primates, including humans. However, the postulate seems to clash with current views on human brain and body growth. As pointed out above, the combination of magnified brain growth (Vrba, 1998; Rice, 2002; Leigh, 2004) and slower body growth (Leigh, 2001) in humans seems to imply changes in ontogenetic brain allometry relative to other primates. Unfortunately, no statistical test of differences in brain allometry trajectories across primates has ever been reported, although brain allometry regressions of H. sapiens and Macaca mulatta were shown to be very similar (Holt et al., 1975). Count (1947) presented brain allometry data for a sample of chimpanzees, which consisted of individuals over three years of age (excepting four premature births), when brain growth is nearly finished, and for Trachypithecus cristatus [referred to as Semnopithecus maurus by Count (1947)]. Holt et al. (1975) introduced data for M. mulatta, and more recently, a curve for Saimiri sciureus was published by Manocha (1979). Since then, no systematic study of primate brain allometry from conception has been published, and the hypothesis of a shared primate brain growth allometry remains virtually untested. In the context of the possible paradox involving brain growth, body growth, and brain allometry pointed out above, one could also consider the possibility that human and primate body growth rates are not exceptionally slow but result from a hidden scaling effect. This was recently hinted at by West et al.’s (2001) ‘‘general’’ model for animal growth, which covers a wide range of animal taxa from shrimp to cows. According to this model, the energy budget for a growing organism (its basal metabolic rate, B) is partitioned into cell metabolism (body maintenance) and a surplus available for cell division (body growth); while B grows in proportion to m3/4 (where m is body mass), maintenance would scale with cell number and m. For this reason, the surplus channeled to growth necessarily decreases with body size. Animal growth is therefore necessarily asymptotic, as expressed by the formula r Z 1
e
t, where r and t are dimensionless size and age variables

766

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

correcting for species-specific scaling effects (see Materials and methods). Dimensionless size and age variables are supposed to reveal ‘‘invariant’’ biological properties (Charnov, 1993), which might, in principle, relate to Count’s idea of a conserved brain allometric trajectory in primates. Unfortunately, primates were not included in West et al.’s (2001) study, despite the availability of body growth data for various species. The postulate of a shared primate brain allometry therefore remains controversial. (2) How prolonged is human brain growth? According to the ‘‘brain allometry extension’’ model, brain growth in humans is considerably extended. However, Vrba (1998) showed that human brain growth is prolonged by a factor of only 29% compared to chimpanzees. According to her ‘‘proportional growth prolongation’’ hypothesis, the same value of 29% applies to the first postnatal brain growth phase, implying a prolongation from about 7 months in chimpanzees to 9 months in humans. Rice (2002) proposed an even smaller extension of 21% in human brain growth duration. Finally, Leigh (2004) did not find any difference in brain growth duration between the two species, and attributed the larger brain of H. sapiens to higher brain growth rates. Therefore, none of the recent studies in primate brain growth corroborates the hypothesis of a dramatic extension of postnatal brain growth in humans relative to chimpanzees. (3) Is postnatal brain growth more important in humans than in other primates? Count’s (1947) model predicts an amplified role for postnatal brain growth in humans. Ratios of 70% of neonate to adult brain size in macaques, 50% in chimpanzees, and only 25% in humans are frequently cited as evidence for the ‘‘brain allometry extension’’ model. Such ratios are, in principle, easy to estimate, but primate brain growth data are scarce. For this reason, most comparisons of postnatal brain growth in humans and chimpanzees (Holt et al., 1975; Passingham, 1982; Martin, 1983; Dienske, 1986; Smith and Tompkins, 1995; Hrdy, 1999; Coqueugniot et al., 2004) do not present new data for the two species and rely on a few

classic studies (Schultz, 1940, 1941; Blinkov and Glezer, 1968). The most common reference is Schultz’s (1941) comparative analysis of hominoid development. Schultz’s sample included a single male newborn chimpanzee (with a cranial capacity of 128 cc), two adult males (average cranial capacity of 420.5 cc), and five adult females (average cranial capacity of 353 cc). Based on the male specimens, the ratio of neonatal to adult brain size is only 30.4%, while the total adult sample of seven specimens implies a ratio of 34.4%. Nonetheless, Schultz (1941: 273) stated that, in chimpanzees, ‘‘cranial capacity has reached at birth already 46% of its final size in adults.’’ The latter value was derived from a previous publication (Schultz, 1940) in which a cranial capacity of 171 cc was presented as the ‘‘newborn’’ value in chimpanzees (in fact, corresponding to 46% of chimpanzee adult brain size). However, the ‘‘newborn’’ sample in Schultz’s (1940) earlier study consisted of a single unsexed individual aged 74 days (or 2.5 months). Published chimpanzee brain growth curves (Vrba, 1998; Rice, 2002; Leigh, 2004) show that not less than 10% of total brain growth (around 40 g) occurs during the first postnatal 2.5 months. Therefore, the value of 46% in Schultz’s (1940) earlier study is clearly an overestimate and does not represent the ratio at birth for chimpanzees. A more recent study, based on a sexed and larger sample (Herndon et al., 1999), reported a chimpanzee ‘‘newborn’’ brain weight of 142.5 g (n Z 9), and an adult value of 389.9 g (n Z 13). Although Herndon et al. (1999) also included older specimens (up to one year of age) in the ‘‘newborn’’ group, the ratio of neonatal to adult brain size obtained was only 36.5%. More recently, Coqueugniot et al. (2004) analyzed early brain growth in Homo erectus by comparing the Pening 1 subadult specimen (the ‘‘Mojokerto child’’) to a large growth series of chimpanzee and human skulls. The authors stated that chimpanzee brain size at birth corresponds to 40% of the adult value, compared to 25% in humans. However, these values were not calculated from their own sample, but were from Schultz [1941; via Leutenegger (1972) and Gould (1977)]. Coqueugniot et al.’s (2004) own data (see their Figure 3) point to ratios of neonate to adult brain

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

size of around 30% in humans and 45% in chimpanzees. Unfortunately, these estimates cannot be verified due to the fact that the authors assigned all specimens to classes of 10 percent of adult values [e.g., cranial capacity of the first chimpanzee is given as its ‘‘percentage of the adult value’’ (Coqueugniot et al., 2004: 300)], but neither individual nor average adult cranial capacities were reported (see supplementary Tables 1e4 accompanying Coqueugniot et al., 2004). It is also impossible to predict values at birth, since the equations for the plotted nonlinear brain growth curves (described only as ‘‘regressions’’) were not given. Finally, nearly 40% of the adult chimpanzee sample (about 80 of 200 skulls) consisted of Pan paniscus specimens, while the seven ‘‘newborn’’ chimpanzees (aged up to 2 years) were exclusively Pan troglodytes. The ratio of neonatal brain size to adult brain size may be distorted or overestimated if adult brain size in Pan paniscus is smaller than in Pan troglodytes, but this possibility could not be further investigated since absolute values of cranial capacity are not given in Coqueugniot et al. (2004). In conclusion, the few studies based on first-hand (and replicable) estimates of chimpanzee brain size point to values between 30% and 36.5% of total brain size achieved at birth. Despite being so frequently cited, values of 50% or even 40% can be criticized for being either overestimates or impossible to verify and replicate. A ratio of 25% of neonatal to adult brain size in humans is equally questionable. Based on direct measurements, Schultz (1941) reported a ratio of 23.7% in H. sapiens, while Blackfan (1933) presented values of around 25%, but those figures seem to represent a lower limit in humans. Blinkov and Glezer (1968) compiled data from various sources, which indicate ratios of 26.3% in men and 31.3% in women (Muhlmann, 1957); 29.4% in men and 24.5% in women (Blinkov and Glezer, 1968); and for unsexed samples, 28% (Ellis, 1920) and 26.9% (Glezer, 1961). More recently, Harvey et al. (1987) reported a value of 30.7%, while the brain growth curves calculated by Cabana et al. (1993) predicted a ratio of 31.1% in men and 40% in women (assuming a pregnancy of 40 weeks). Coqueugniot et al. (2004) suggested a ratio of around 30%, but the exact value derived from

767

their sample cannot be inferred. Therefore, the ratio of neonatal to adult brain size in humans seems to range between 24% and 31% [the value of 40% in women calculated by Cabana et al. (1993) seems to be an overestimate]. Since the studies reviewed above were based on different methods, sample sizes, and skeletal collections, variation in estimates of neonatal relative brain size is not surprising. What is relevant is that the ranges of variation they delimit (around 24% to 31% in humans, and 30% to 37% in chimpanzees) do not point to dramatic differences between humans and chimpanzees, and even imply possible overlap. Therefore, acceptance of an extremely hypertrophied postnatal brain growth in humans is questionable on empirical grounds. This represents a problem for the ‘‘brain allometry extension’’ theory, which predicts and depends on a more prominent role of postnatal brain growth in humans. An alternative scenario for human encephalization Count’s (1947) theory of human encephalization, characterized by an emphasis on brain allometry extension, fails to recognize the role of other equally relevant evolutionary changes in the timing of human development. In three steps, I present formal evidence against the ‘‘brain allometry extension’’ model and for an alternative view on how changes in timing relate to human encephalization. First, I compare the few known primate brain allometry curves and identify significant differences among species (in particular, between humans and chimpanzees). Second, I show that the moderate prolongation of human brain growth proposed by Vrba (1998) and Rice (2002) does not imply a reduction in the ratio of neonatal to adult brain size in humans, which is compatible with the evidence for partial overlap with chimpanzees; I also propose an explanation for why the average ratio is not strictly similar in the two species. Third, by demonstrating that humans and other anthropoid primates are an exception to the ‘‘universal’’ growth pattern proposed by West et al. (2001), I argue that a reduction in postnatal body growth rates (Leigh, 2001) was a crucial factor for the evolution of encephalization in H. sapiens.

768

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

A comment on the limitations of the methods and data used here is necessary. Comparative growth studies should ideally be based on longitudinal and sexed samples assessed from conception (using non-intrusive techniques such as magnetic resonance imaging). Such high-quality data are not presently available. Prenatal brain growth data in primates are scarce, and virtually unknown for chimpanzees. For this reason, studies addressing the issue of human encephalization have relied on a few sources of cross-sectional data. While mean growth rates can be derived from crosssectional data, individual variation in growth rates and other parameters cannot (Boas, 1892; Eveleth and Tanner, 1992). Moreover, since most studies pool males and females, sexual variation in growth is also commonly neglected. Data are often digitized from scanned plots, introducing another source of error. For example, Vrba (1998) and Rice (2002) could be criticized for drawing inferences about brain growth trajectories from a dataset characterized by these problems (Leigh, 2004), and the same criticism applies to the results presented in this article. However, analyses based on the currently available primate dataset, despite its evident drawbacks, can be valid and informative. Crosssectional data can satisfactorily detect interspecific variation in growth patterns and serve as a guideline to longitudinal studies (Leigh, 2004). Finally, it should be stressed that the ‘‘brain allometry extension theory’’ also draws inferences about brain growth duration, growth rates, and phase transitions from the same cross-sectional, second-hand, and unsexed primate data used for this project. Without denying the importance of additional brain and body growth studies in primates, it is reasonable to argue that the same data and methods underlying the ‘‘brain allometry extension’’ scenario can be legitimately used to support an alternative view of human encephalization.

Materials and methods Human brain allometry A cross-sectional sample of 568 individual allometric data points (brain size vs. body size) was

obtained from published human growth studies (Count, 1947; Larroche, 1967; Dobbing and Sands, 1973; Burn et al., 1975). When tables were not available, data points and axes were scanned from figures; their coordinates were then digitized and transformed into the original measurements. Double-log transformation of the latter was followed by Model II regression analysis (reduced major axis), excluding points past brain growth termination. In order to minimize the effects of individual variation, I also ranked points by body weight and defined groups of ten points, starting from the lowest value. For example, the first group included the smallest to the tenth smallest body weights in the sample, the second group included the eleventh to the twentieth smallest body weights in the sample, and so on. Average brain weight and body weight were then calculated for each of the 47 ten-point groups, which were used in a second regression presented in the main text. Comparisons between primate brain allometries Brain allometry comparisons required primate fetal data, available only for Trachypithecus cristatus [Count (1947) who referred to this taxon as Semnopithecus maurus], Macaca mulatta (Kerr et al., 1969; Cheek, 1975), and Saimiri sciureus (Manocha, 1979). Cross-sectional data from tables or digitized figures were analyzed using Model II regression (reduced major axis) after double-log transformation. Pairwise comparisons between Model I (instead of Model II) regressions of log(brain weight) on log(body weight) were used to assess differences in slope (t-test) and intercept (ANCOVA) between the human and each of the other three primate regressions. ‘‘Universal’’ and primate body growth The general model for animal growth (West et al., 2001) is given by the equation r Z 1
e
t, where r Z (m/M)1/4 is a dimensionless measure of body size and t Z at/4M1/4
ln(1
(m0/M)1/4 is a dimensionless measure of age from birth (m is weight, M is adult weight, a is a cell metabolism parameter, t is age from birth, and m0 is birth weight).

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

Data on primate weight (m, in grams) were converted into the dimensionless variable r using the formula r Z (m/M)1/4, and data on age (t, in days) were converted into the variable t Z at/4M1/4
ln(1
(m0/M)1/4, and an empirical curve was then plotted for each primate. Growth data for primate species were obtained from published studies (Grether and Yerkes, 1940; Long and Cooper, 1968; Saxton and Lotz, 1990; Eveleth and Tanner, 1992; Leigh and Shea, 1996; Garber and Leigh, 1997). Fetal body growth data for humans and rhesus monkeys are from Brenner et al. (1976) and Kerr et al. (1969). When tables were not available, data were digitized from growth curves.

Results Humans and primates do not share a common ontogenetic brain allometry According to Count’s (1947) theory, a common brain growth allometry is shared across primates, and its differential extension is the main explanation for interspecific variation in brain size and encephalization. In humans, the correlation between brain and body size before cessation of brain growth is very strong (r Z 0.98, p ! 0.0005, n Z 470, after curve linearization by doublelog transformation). The resulting brain growth allometry equation is Yh Z
0:70C0:94Xh where Yh is log(brain weight) and Xh is log(body weight). Based on ten-point averages (see Materials and methods), an improved regression is obtained (Fig. 1A; r Z 0.998, p ! 0.0005, n Z 47): Yh Z
0:66C0:93Xh The constancy of the allometric coefficient during ontogeny (Fig. 1A) is remarkable, both in the light of the four orders of body size magnitude covered by the regression and of the intricate character of human multiphasic brain growth (Vrba, 1998). Allometric curves can be calculated for only a few other primates (T. cristatus, S. sciureus,

769

and M. mulatta). As in humans, regression analysis after double-log transformation shows that brain allometry is also approximately linear in T. cristatus (r Z 0.98, p ! 0.001, n Z 14), S. sciureus (r Z 0.98, p Z 0.001, n Z 6), and M. mulatta (r Z 0.99, p ! 0.001, n Z 26). The slope values (allometric coefficients) and intercepts suggest differences among species (Table 1). Statistical testing of differences was based on comparisons between the Model I regressions of brain weight on body weight. In a comparison with humans (Fig. 1), both slope (t Z 5.61, p ! 0.001) and intercept (F Z 7.8, p ! 0.01) are different in S. sciureus, while the intercept is different in T. cristatus (F Z 90.8, p ! 0.001) but not the slope (t Z 1.66, p Z 0.09). Finally, the intercepts (F Z 6.1, p ! 0.05) but not slopes (t Z 1.71, p Z 0.09) of M. mulatta and H. sapiens cannot be distinguished. The question is whether the revealed statistical differences are relevant for explaining variation in brain size and encephalization among primate species. Of special interest is the comparison between humans and chimpanzees, but the chimpanzee brain allometry regression cannot be calculated due to the poor quality of the data (prenatal data are not available). For this reason, the hypothesis of a common primate brain allometry needs to be tested indirectly. The rationale is that, if Count’s (1947) model is correct and primates share a common brain allometry, human brain allometry should accurately predict brain size for a given body weight not only in humans, but in other primate species as well. This can be tested at birth. For a birth weight of 3320 g (Potter and Craig, 1975), the equation predicts a brain weight of 412 g in humans, which is in close agreement with the data (Blinkov and Glezer, 1968; Potter and Craig, 1975; Vrba, 1998). However, for reported chimpanzee birth weights of 1756 g (Harvey et. al., 1987) and 1885 g (Grether and Yerkes, 1940), the equation predicts neonatal brain sizes of 227 g and 243 g, respectively, against reported values ranging from 128 g (Schultz, 1941) to 142.5 g (Herndon et al., 1999). In other words, if one assumes that chimpanzee brain allometry is similar to that of humans, one obtains an overestimate of 60% to 90% in chimpanzee neonatal brain size.

770

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

Fig. 1. Brain allometry differs among humans and other primates. A) Brain allometry in H. sapiens is nearly constant during development. Regression of log(brain size) on log(body size), excluding points postdating brain growth cessation, is shown as a solid line. Each solid circle represents an average of ten individuals (see Materials and methods). B) Brain allometries in M. mulatta (dashed line; data points shown as open circles) and humans (solid line) differ in intercept but not in slope. C) Brain allometries in T. cristatus and humans also differ in intercept but not in slope. D) Brain allometries in S. sciureus and humans differ both in slope and intercept.

Chimpanzee neonatal brain size can also be estimated from the brain allometries calculated for the other three primate species. Based on birth weights of 1756 g (Harvey et. al., 1987) and 1885 g (Grether and Yerkes, 1940), the Macaca model drastically overestimates chimpanzee neonatal brain size (236 g and 253 g), while the Saimiri model drastically underestimates it (74 g and 76 g). Trachypithecus brain allometry satisfactorily Table 1 Brain allometry in H. sapiens, M. mulatta, T. cristatus, and S. sciureus

H. sapiens T. cristatus M. mulatta S. sciureus

Model I

Model II

y Z
0.66 C 0.93x y Z
0.7 C 0.87x y Z
0.80 C 0.97x y Z
0.03 C 0.57x

y Z
0.66 C 0.93x y Z
0.74 C 0.89x y Z
0.84 C 0.99x y Z
0.01 C 0.58x

Double-log transformation of brain size and body size data was followed by Model I and Model II regression analysis (reduced major axis). The human regression is based on ten-point averages (see Materials and methods). Regression statistics are shown in the main text.

predicts chimpanzee neonatal brain size (140 g and 150 g). Postnatal brain growth does not explain encephalization in H. sapiens The ‘‘brain allometry extension’’ theory postulates a magnified role for postnatal brain growth and a much lower ratio of neonatal to adult brain size in humans. As reviewed in the introduction, available data do not corroborate this view. In the following, it is shown on theoretical grounds that the contribution of postnatal brain growth in humans and chimpanzees is, in fact, expected to be fairly similar. A relationship between brain growth duration and gestation length was indirectly suggested by Vrba (1998). In both chimpanzees and humans, the estimated transition between second and third phases of brain growth closely coincides with birth. While the transition is estimated to be 225 days and 284 days after conception for

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

chimpanzees and humans, respectively, gestation length is from 228 to 240 days in chimpanzees (Harvey et al., 1987; Kappeler and Pereira, 2003) and from 267 to 282 days in humans (Potter and Craig, 1975; Kappeler and Pereira, 2003). Thus, if it is assumed, as a first approximation, that birth occurs at the end of the second brain growth phase (T2 in Vrba’s model), neonatal (Nc) and adult (Yc) brain size in chimpanzees can be written respectively in log form as Nc ZA0 Ck1 ðT1 ÞCk2 ðT2
T1 Þ Yc ZA0 Ck1 ðT1 ÞCk2 ðT2
T1 Þ Ck3 ðT3
T2 ÞCk4 ðT4
T3 Þ The ratio of neonatal to adult brain size in chimpanzees is therefore nc/yc, or Nc
Yc in log form. Vrba (1998) showed that, relative to the chimpanzee-like pattern above, all human brain growth phases are extended by a factor a, so that each phase transition ti becomes ati, or log(a) C Ti in log form (Vinicius and Lahr, 2003). Human neonatal brain size (Nh) is therefore obtained by adding log(a) to each phase transition Ti: Nh ZA0 Ck1 ½logðaÞCT1  Ck2 ½logðaÞCT2
logðaÞ
T1 Zk1 logðaÞCNc Since adult brain size in humans is Yh Z k1log(a) C Yc (Vinicius and Lahr, 2003), the ratio of neonatal to adult brain size in humans (nh/yh, or Nh
Yh in log form) is Nh
Yh Z½k1 logðaÞCNc 
½k1 logðaÞCYc  ZNc
Yc Therefore, the predicted ratio of neonatal to adult brain size at the end of ‘‘fetal’’ brain growth in Vrba’s (1998) model is the same in humans and chimpanzees (Nh
Yh Z Nc
Yc). The ratio, derived from the values of A0, ki, and Ti (Vrba, 1998), is 0.32, or 32%, implying a contribution of 68% of postnatal growth to adult brain size in both species. A ratio of 32% is close to the intersection of the observed intervals in humans (from 24% to 31%) and chimpanzees (from 30% to 37%).

771

However, the observed intervals also show that the ratio of neonatal to adult brain size is not strictly similar in the two species. A possible explanation is that birth does not correspond exactly to the predicted transition from the second to the third brain growth phase in Vrba’s (1998) model. Thus, for a predicted brain growth transition at 284 days, gestation length in humans ranges from 267 to 282 (Potter and Craig, 1975; Harvey et al., 1987; Kappeler and Pereira, 2003). Using Vrba’s multiphasic equation, the ratio is 26.5% at 267 days, and 31.3% at 282 days. On the other hand, for a predicted brain growth transition at 225 days in chimpanzees, gestation length is from 228 to 240 days (Harvey et al., 1987; Kappeler and Pereira, 2003), implying ratios of 33.1% (at 228 days) and 34.9% (at 240 days). Those values are within the empirical intervals observed in both species. Postnatal body growth retardation in humans and primates The fact that the role of postnatal brain growth is nearly similar in chimpanzees and humans is strong evidence against Count’s (1947) ‘‘brain allometry extension’’ theory. As argued below, postnatal development remains a crucial factor for human encephalization, but for reasons related to body growth instead of brain growth. The ‘‘general’’ growth model (West et al., 2001) postulates a ‘‘universal’’ curve underlying body growth of diverse animal taxa, from crustaceans to birds and mammals. However, the original study by West et al. did not include primates. As shown in Fig. 2, when the ‘‘general’’ model is applied to data for anthropoid primates, a marked developmental delay is observed in the group, particularly in humans. West et al. (2001) argued that fluctuations around the predicted growth trajectory might occur due to growth spurts in body weight. Primate growth spurts are not universally distributed, and are clearly absent in some small-bodied New World monkeys, such as Cebuella pygmaea and Callimico goeldi (Leigh, 1996). However, these two species also deviate from the general curve (Fig. 2). A further reason why growth spurts cannot be an explanation for

772

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

Fig. 2. Postnatal body growth in anthropoid primates is slow and departs from West et al.’s (2001) general animal growth pattern (dashed line). From left to right, curves shown are: cow (A; plotted as controls), Goeldi’s monkey (6), pygmy marmoset (:), rhesus monkey (,), gorilla (-), chimpanzee (B), and human (C). The mass ratio r Z (m/M)1/4 is a dimensionless measure of body size, and t Z at/4M1/4
ln(1
(m0/ M)1/4 is a dimensionless measure of age (m is weight, M is adult weight, a is a cell metabolism parameter, t is age from birth, and m0 is birth weight).

the relative retardation of primate growth is that mice and cows also exhibit pubertal growth spurts in body weight (Bogin, 1999), but still conform to the general model. The use of the ‘‘general’’ growth formula also sheds light on the contrast between prenatal and postnatal body growth in humans revealed by Leigh (2001). While West et al. (2001) based their analyses on data from birth only, Fig. 3 shows dimensionless curves from conception in H. sapiens and M. mulatta. Fetal body growth in the two species is very similar, not only to each other, but also to the general animal pattern. However, a clear inflection in both trajectories is observed soon after birth, and is more intense in humans. This result agrees with the analysis of Leigh (2001), who argued that the prolonged growth period in humans is caused by a reduction in early body growth rates (i.e., between birth and the adolescent growth spurt), while prenatal body growth does not significantly deviate from the general ape pattern.

Discussion Evolutionary accounts of how large brains evolved in humans often appeal to changes in

Fig. 3. Primate body growth retardation is limited to the postnatal period. Plot shows dimensionless body growth curves from conception for Macaca mulatta (,) and H. sapiens (C). Birth in both species (vertical line) occurs when t is approximately one. West et al.’s (2001) general model is represented by the dashed line. The equation r Z (m/M) 1/4 is a dimensionless measure of body size, and t Z at/4M1/4
ln(1
(m0/M)1/4 is a dimensionless measure of age (m is weight, M is adult weight, a is a cell metabolism parameter, t is age from birth, and m0 is birth weight).

developmental timing. Together, the postulates of a conserved brain growth allometry in primates, its differential extension across species, and an exceptionally large fraction of brain growth occurring after birth in H. sapiens define the ‘‘brain allometry extension’’ theory of human encephalization. I have argued that the same evidence that inspired the ‘‘brain allometry extension’’ model, complemented by recent studies in brain and body growth, suggests that the model is incorrect. As pointed out by Deacon (2000), human encephalization is not a simple extrapolation of a general primate trend: it is a complex combination of changes in brain growth and body growth unique to the hominin lineage. From the point of view of changes in developmental timing, the evolution of human encephalization can be better understood as the result of three equally important factors. First, a modest prolongation of less than 30% in brain growth and brain allometry can account for a human brain more than three times the size of a chimpanzee or australopith brain (Vrba, 1998; Rice, 2002; Vinicius and Lahr, 2003). In particular, postnatal brain growth is not dramatically extended, and does not play a magnified role in

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

H. sapiens. In both chimpanzees and humans, postnatal brain growth is responsible for about 65%e70% of total brain growth. There is a slight deviation from an equal ratio of neonatal to adult brain size in the two species (calculated as 32%), probably because chimpanzee birth seems to occur a few days later, and human birth a few days earlier than predicted from differences in brain growth duration in chimpanzees and humans. Second, a distinctive brain allometry seems to have evolved in humans. Humans and chimpanzees cannot share a common brain growth allometry, as the latter predicts a chimpanzee brain size at birth nearly twice the observed value. Brain allometries in humans and other primates display statistical differences in intercept and slope. The similar slopes in humans and M. mulatta, or the satisfactory prediction of neonatal brain size and body size in chimpanzees using the T. cristatus brain allometry equation, seem to be examples of convergence, rather than plesiomorphies. Therefore, instead of being highly conserved, brain allometry may be a very labile character. Studies in other primate species are required to confirm this conclusion. Finally, human encephalization involved a substantial reduction in postnatal body growth rates relative to an already slow primate pattern, a difference that is apparent even when dimensionless weight and time variables (West et al., 2001) are employed. Rather than revealing universality in animal growth, West et al.’s model could be seen as a new method for representing and understanding the distinction between a ‘‘general’’ (or ‘‘infraprimate,’’ after Laird, 1967) growth pattern on the one hand, and a primate growth pattern on the other. Observed levels of retardation among anthropoid primates seem to reproduce grade relationships (Fleagle, 1998; Lewin and Foley, 2004), as it ranges from modest deviation in New World monkeys, to intermediate levels in Old World monkeys, to extreme retardation in apes and especially humans. Despite being widely acknowledged, slow postnatal body growth has rarely been implicated as a developmental factor in human encephalization. Many authors have argued that postnatal development contributes to human encephalization

773

through increased brain growth, instead of reduced body growth (Gould, 1977; Martin, 1983; Bogin and Smith, 1996). Others have identified correlations between slow postnatal growth and large brain size, but focused on its adaptive reasons (Janson and van Schaik, 1993; Ross, 2004). It is not generally recognized that the reduction in human postnatal body growth rates has a direct bearing on human encephalization, as it implies higher ratios of brain to body size in adult humans in comparison to chimpanzees and, presumably, earlier hominins. Postnatally, brain size multiplies by the same factor (from 3 to 4 times) in both humans and chimpanzees. In contrast, while chimpanzee body size increases by a factor of around 22 after birth, human body size increases about 17 times (see appendix in Kappeler and Pereira, 2003), despite the fact that human body growth is prolonged. In sum, postnatal development is crucial to human encephalization because of the reduction (relative to ape standards) in human body growth rates, not because of hypertrophied brain growth. While the analyses and conclusions of this study were presented in allometric terms (changes in rates and duration of absolute and relative growth), human encephalization has also been debated within the framework of heterochrony (Gould, 1977; Shea, 1983, 2002; McKinney and McNamara, 1991; Vrba, 1998; Rice, 2002; Vinicius and Lahr, 2003). Translated into the language of heterochrony, the ‘‘brain allometry extension’’ model is an example of ‘‘ontogenetic scaling’’ (Shea, 1983), or differential extension of conserved allometries in evolutionary descendants. In other words, the ‘‘brain allometry extension’’ model implies the evolution of peramorphic (‘‘overdeveloped’’) brains in H. sapiens through hypermorphosis (terminal extension of brain allometry). However, since brain allometry is not conserved in humans, it follows that ontogenetic scaling cannot account for the evolution of human encephalization. Although the evolution of a derived brain allometry in humans could still be described in heterochronic terms as a hypothetical combination of hypermorphosis (terminal extension), acceleration (increase in allometric coefficient), and predisplacement (higher intercept), the evolution of

774

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

human brain and body growth cannot be described by heterochronic categories. McKinney and McNamara (1991), Vrba (1998), and Rice (2002) showed that human brain growth evolved through sequential hypermorphosis or proportional prolongation of brain growth phases. However, sequential hypermorphosis generates derived ontogenetic trajectories, and since the conservation of shape trajectories (and their extension or truncation in descendants) is the condition for heterochrony (Shea, 1989; Vinicius and Lahr, 2003), sequential hypermorphosis does not generate brain/body proportions that can be classified as paedomorphic or peramorphic (i.e., heterochronic). Finally, Leigh (2001) showed that the timing of prenatal and postnatal body growth evolved differently in humans, leading to a result that cannot be described by either sequential hypermorphosis or classic heterochronic processes (such as hypermorphosis or neoteny). Thus, while the evolution of human encephalization can be described in allometric terms as the combination of proportional prolongation of brain growth, distinctive brain allometry, and reduced rates of postnatal body growth, it seems that ‘‘there is no central component of heterochronic transformation that predominantly accounts for the bulk of morphogenetic and evolutionary transitions in human morphology’’ (Shea, 2002: 95).

Conclusion Changes in developmental timing were a crucial factor in the evolution of human encephalization. The mainstream ‘‘brain allometry extension’’ model, which by definition is characterized by an emphasis on a single aspect of developmental timing, provides an incomplete account of human encephalization. From the point of view of evolutionary changes in developmental timing, human encephalization should be understood as a combination of three components: extended brain growth, retarded postnatal body growth, and a derived brain growth allometry. Adaptive scenarios should be formulated in order to help elucidate how and why those features could have jointly evolved in the hominin lineage leading to H. sapiens. The lack of brain

allometry data remains an important challenge to analyses of primate and human encephalization, and further empirical studies are needed. Acknowledgements This research was funded by a fellowship awarded by the Leverhulme Centre for Human Evolutionary Studies (LCHES), Cambridge. I thank Marta Mirazon Lahr, Rob Foley, and Andrea Migliano for their critical input. I am also grateful to William Kimbel, and the two anonymous referees for their valuable suggestions.

References Aiello, L.C., Bates, N., Joffe, T., 2001. In defense of the expensive tissue hypothesis. In: Falk, D., Gibson, K.R. (Eds.), Evolutionary Anatomy of the Primate Cerebral Cortex. Cambridge University Press, Cambridge, pp. 57e78. Blackfan, K.D., 1933. Growth and Development of the Child: Part II, Anatomy and Physiology. The Century Co., New York. Blinkov, S.M., Glezer, I.I., 1968. The Human Brain in Figures and Tables. Plenum Press, New York. Bogin, B., 1997. Evolutionary hypotheses for human childhood. Yearb. Phys. Anthropol. 40, 63e89. Bogin, B., 1999. Evolutionary perspectives on human growth. Annu. Rev. Anthropol. 28, 109e153. Bogin, B., Smith, B.H., 1996. Evolution of the human life cycle. Am. J. Hum. Biol. 8, 703e716. Brenner, W.E., Edelman, D.A., Hendricks, C.H., 1976. A standard of fetal growth for the United States of America. Am. J. Obstet. Gynecol. 126, 555e564. Boas, F., 1892. The growth of children. Science 19, 256e257. Burn, J., Birkbeck, J.A., Roberts, D.F., 1975. Early fetal brain growth. Hum. Biol. 47, 511e522. Cabana, T., Jolicoeur, P., Michaud, J., 1993. Prenatal and postnatal growth and allometry of stature, head circumference, and brain weight in Quebec children. Am. J. Hum. Biol. 5, 93e99. Charnov, E.L., 1993. Life History Invariants: Some Explorations of Symmetry in Evolutionary Ecology. Oxford University Press, Oxford. Charnov, E.L., Berrigan, D., 1993. Why do female primates have such long lifespans and so few babies? Or life in the slow lane. Evol. Anthropol. 1, 191e193. Cheek, D.B., 1975. The fetus. In: Cheeks, D.B. (Ed.), Fetal and Postnatal Cellular Growth. Wiley, New York, pp. 3e22. Coqueugniot, H., Hublin, J.-J., Veillon, F., Houet, F., Jacob, T., 2004. Early brain growth in Homo erectus and implications for cognitive ability. Nature 231, 299e302.

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776 Count, E.W., 1947. Brain and body weight in man: their antecedents in growth and evolution. Ann. N. Y. Acad. Sci. 46, 993e1122. Deacon, T.W., 2000. Heterochrony in brain evolution: cellular vs. morphological analyses. In: Parker, S.T., Langer, J., McKinney, M.L. (Eds.), Biology, Brains, and Behavior: The Evolution of Human Development. SAR Press, Santa Fe, pp. 41e88. Dienske, H., 1986. A comparative approach to the question of why human infants develop so slowly. In: Else, J.G., Lee, P.C. (Eds.), Primate Ontogeny, Cognition and Social Behaviour. Cambridge University Press, Cambridge, pp. 145e154. Dobbing, J., Sands, J., 1973. Quantitative growth and development of the human brain. Arch. Dis. Child. 48, 757e767. Ellis, R.S., 1920. Norms for some structural changes in the human cerebellum from birth to old age. J. Comp. Neurol. 32, 1e35. Eveleth, P.B., Tanner, J.M., 1992. Worldwide Variation in Human Growth. Cambridge University Press, Cambridge. Fleagle, J.G., 1998. Primate Adaptation and Evolution. Academic Press, New York. Foley, R.A., Lee, P.C., 1991. Ecology and energetics of encephalization in hominid evolution. Philos. Trans. R. Soc. Lond. B 334, 223e232. Garber, P.A., Leigh, S.R., 1997. Ontogenetic variation in smallbodied New World primates: implications for patterns of reproduction and infant care. Folia Primatol. 68, 1e22. Glezer, I.I., 1961. The principles of growth of the pyramidal cells of the frontal cortex in human postnatal ontogenesis. Vopr. Antrop. 6, 121. Gould, S.J., 1977. Ontogeny and Phylogeny. Harvard University Press, Cambridge, MA. Grether, W.F., Yerkes, R.M., 1940. Weight norms and relations for chimpanzee. Am. J. Phys. Anthropol.181e198. Harvey, P.H., Martin, R.D., Clutton-Brock, T.H., 1987. Life histories in a comparative perspective. In: Smuts, B.B., Cheney, D.L., Seyfarth, R.M., Wrangham, R.W., Struhsaker, T.T. (Eds.), Primate Societies. University of Chicago Press, Chicago, pp. 181e196. Herndon, J.G., Tigges, J., Anderson, D.C., Klumpp, S.A., McClure, H.M., 1999. Brain weight throughout the life span of the chimpanzee. J. Comp. Neurol. 409, 567e572. Hill, K., Hurtado, A.M., 1996. Ache Life History: The Ecology and Demography of a Foraging People. Aldine de Gruyter, New York. Holt, A.B., Cheek, D.B., Mellits, D., Hill, D.E., 1975. Brain size and the relation of the primate to the nonprimate. In: Cheeks, D.B. (Ed.), Fetal and Postnatal Cellular Growth. Wiley, New York, pp. 23e44. Hrdy, S.B., 1999. Mother Nature: A History of Mothers, Infants, and Natural Selection. Chatto & Windus, London. Janson, C.H., van Schaik, C.P., 1993. Ecological risk aversion in juvenile primates: slow and steady wins the race. In: Pereira, M.E., Fairbanks, L.A. (Eds.), Juvenile Primates: Life History, Development and Behaviour. Oxford University Press, Oxford, pp. 57e74. Kappeler, P.M., Pereira, M.E., 2003. Primate Life Histories and Socioecology. The University of Chicago Press, Chicago.

775

Kennedy, G.E., 2005. From the ape’s dilemma to the weanling’s dilemma: early weaning and its evolutionary context. J. Hum. Evol. 48, 123e145. Kerr, G.R., Kennan, A.L., Waisman, H.A., Allen, J.R., 1969. Growth and development of the fetal rhesus monkey. I. Physical growth. Growth 33, 201e213. Laird, A.K., 1967. Evolution of the human growth curve. Growth 31, 345e355. Larroche, J.-C., 1967. Maturation morphologique du systeme nerveux central: ses rapports avec le developpement ponderal du foetus et son age gestationnel. In: Minkowski, A. (Ed.), Regional Development of the Brain in Early Life. F.A. Davis, Philadelphia, pp. 247e256. Leigh, S.R., 1996. Evolution of human growth spurts. Am. J. Phys. Anthropol. 101, 455e474. Leigh, S.R., 2001. Evolution of human growth. Evol. Anthropol. 10, 223e236. Leigh, S.R., 2004. Brain growth, life history, and cognition in primate and human evolution. Am. J. Primatol. 62, 139e164. Leigh, S.R., Shea, B.T., 1996. The ontogeny of body size variation in African apes. Am. J. Phys. Anthropol. 99, 43e65. Lewin, R.A., Foley, R.A., 2004. Principles of Human Evolution. Blackwell, Cambridge, MA. Leutenegger, W., 1972. Newborn size and pelvic dimensions of Australopithecus. Nature 240, 568e569. Long, J.O., Cooper, R.W., 1968. Physical growth and dental eruption in captive-bred squirrel monkeys. In: Rosenblum, L.A., Cooper, R.W. (Eds.), The Squirrel Monkey. Academic Press, New York, pp. 193e205. Manocha, S.L., 1979. Physical growth and brain development of captive-bred male and female squirrel monkeys, Saimiri sciureus. Experientia 35, 96e98. Martin, R.D., 1983. Human Brain Evolution in an Ecological Context. Fifty-second James Arthur Lecture on the Evolution of the Human Brain. American Museum of Natural History, New York. McKinney, M.L., McNamara, K.J., 1991. Heterochrony: The Evolution of Ontogeny. Plenum Press, New York. Minugh-Purvis, N., McNamara, K.J. (Eds.), 2002. Human Evolution Through Developmental Change. The Johns Hopkins University Press, Baltimore. Muhlmann, L., 1957. Die Abha¨ngigkeit des Hirngewichtes von Ko¨rpergewicht, Ko¨rperla¨nge und Ko¨rperbautypen, Munich. Passingham, R., 1982. The Human Primate. Freeman, San Francisco. Potter, E.L., Craig, J.M., 1975. Pathology of the Fetus and the Infant. Lloyd-Luke, London. Rice, S.H., 2002. The role of heterochrony in primate brain evolution. In: Minugh-Purvis, N., McNamara, K.J. (Eds.), Human Evolution Through Developmental Change. The Johns Hopkins University Press, Baltimore, pp. 154e170. Ross, C., 2004. Life histories and the evolution of large brain size in great apes. In: Russon, A.E., Begun, D.R. (Eds.), The Evolution of Thought: Evolutionary Origins of Great Ape Intelligence. Cambridge University Press, Cambridge, pp. 122e139. Saxton, J.L., Lotz, W.G., 1990. Growth of rhesus monkey during the first 54 months of life. J. Med. Primatol. 19, 119e136.

776

L. Vinicius / Journal of Human Evolution 49 (2005) 762e776

Schultz, A.H., 1940. Growth and development of the chimpanzee. Contrib. Embryol. 518, 1e63. Schultz, A.H., 1941. The relative size of the cranial capacity in primates. Am. J. Phys. Anthropol. 28, 273e287. Shea, B.T., 1983. Allometry and heterochrony in the African apes. Am. J. Phys. Anthropol. 62, 275e289. Shea, B.T., 1989. Heterochrony in human evolution: the case for neoteny reconsidered. Yearb. Phys. Anthropol. 32, 69e101. Shea, B.T., 2002. Are some heterochronic transformations likelier than others? In: Minugh-Purvis, N., McNamara, K.J. (Eds.), Human Evolution Through Developmental Change The Johns Hopkins University Press, Baltimore, pp. 79e101.

Smith, B.H., 1989. Dental development as a measure of life history in primates. Evolution 43, 683e688. Smith, B.H., Tompkins, R.L., 1995. Toward a life history of the Hominidae. Annu. Rev. Anthropol. 24, 257e279. Vinicius, L., Lahr, M.M., 2003. Morphometric heterochrony and the evolution of growth. Evolution 57, 2459e2468. Vrba, E.S., 1998. Multiphasic growth models and the evolution of prolonged growth exemplified by human brain evolution. J. Theor. Biol. 190, 227e239. West, G.B., Brown, J.H., Enquist, B.J., 2001. A general model for ontogenetic growth. Nature 413, 628e631.