Age at First Molar Emergence in Early Miocene Afropithecus Turkanensis

Journal of Human Evolution 44 (2003) 307–329

Age at ﬁrst molar emergence in early Miocene Afropithecus turkanensis and life-history evolution in the Hominoidea Jay Kelley1*, Tanya M. Smith2 1 Department of Oral Biology (m/c 690), College of Dentistry, University of Illinois at Chicago, 801 S. Paulina, Chicago, IL 60612, USA 2 Interdepartmental Doctoral Program in Anthropological Sciences, Stony Brook University, Stony Brook, NY 11794, USA Received 12 July 2002; accepted 7 January 2003

Abstract Among primates, age at first molar emergence is correlated with a variety of life history traits. Age at first molar emergence can therefore be used to broadly infer the life histories of fossil primate species. One method of determining age at first molar emergence is to determine the age at death of fossil individuals that were in the process of erupting their first molars. This was done for an infant partial mandible of Afropithecus turkanensis (KNM-MO 26) from the w17.5 Ma site of Moruorot in Kenya. A range of estimates of age at death was calculated for this individual using the permanent lateral incisor germ preserved in its crypt, by combining the number and periodicity of lateral enamel perikymata with estimates of the duration of cuspal enamel formation and the duration of the postnatal delay in the inception of crown mineralization. Perikymata periodicity was determined using daily cross striations between adjacent Retzius lines in thin sections of two A. turkanensis molars from the nearby site of Kalodirr. Based on the position of the KNM-MO 26 M1 in relation to the mandibular alveolar margin, it had not yet undergone gingival emergence. The projected time to gingival emergence was estimated based on radiographic studies of M1 eruption in extant baboons and chimpanzees. The estimates of age at M1 emergence in KNM-MO 26 range from 28.2 to 43.5 months, using minimum and average values from extant great apes and humans for the estimated growth parameters. Even the absolute minimum value is well outside the ranges of extant large Old World monkeys for which there are data (12.5 to <25 months), but is within the range of chimpanzees (25.7 to 48.0 months). It is inferred, therefore, that A. turkanensis had a life history profile broadly like that of Pan. This is additional evidence to that provided by Sivapithecus parvada (Function, Phylogeny, and Fossils: Miocene Hominoid Evolution and Adaptations, 1997, 173) that the prolonged life histories characteristic of extant apes were achieved early in the evolutionary history of the group. However, it is unclear at present whether life-history prolongation in apes represents the primitive catarrhine pace of life history extended through phyletic increase in body mass, or whether it is derived with respect to a primitive, size-adjusted life history that was broadly intermediate between those of extant hominoids and cercopithecoids. Life history evolution in primates as a whole may have occurred largely through a series of grade-shifts, with the establishment of fundamental life-history profiles early in the histories of major higher taxa. These may have included shifts that were largely body mass dependent, as well as those that occurred in the absence of significant changes in body mass. 2003 Elsevier Science Ltd. All rights reserved.

Keywords: Miocene hominoid; dentition; dental eruption; enamel microstructure; primate life history; primate evolution

* Corresponding author. Tel.: +1-312-996-6054; fax: +1-312-996-6044 E-mail addresses: [email protected] (J. Kelley), [email protected] (T.M. Smith).

Introduction

Life history is one of the most fundamental attributes of a species’ biology. The term ‘life history’ encompasses a host of specific traits, but is most commonly conceptualized in terms of a series of growth and maturational phases ultimately related to the scheduling of reproduction and lifetime reproductive output. These include gesta- tion period, age at weaning, age at sexual maturity and first breeding, interbirth interval, and lon- gevity. Given the importance of life history, it is not surprising that it has become an important issue in primate paleobiology. To date, most of the effort to reconstruct the life histories of extinct species has been focused on the human lineage. Fig. 1. Least squares regression of age at first breeding (months) However, attempts to reconstruct aspects of the against average body mass (kg), both log-transformed, in the life histories of extinct non-human primates are following extant primate higher taxa (numbers of included species in parentheses): Al, Alouattini (2); At, Atelini (2); Ca, becoming increasingly common (Lee and Foley, Callitrichinae (3); Cb, Cebinae (3); Ce, Cercopithecinae (6); Co, 1993; Kelley, 1997, 2002; Kelley et al., 2001; Colobinae (8); Ho, Hominidae (3); Hy, Hylobatidae (2); In, Godfrey et al., 2002; Schwartz et al., 2002). The Indriidae (3); Le, Lemuridae (7). Data on age at first breeding from Godfrey et al. (2001) and from K. Strier, personal evolution of primate life histories, and the role of communication, for Brachyteles arachnoides (Atelini); body life history in the adaptive radiations of major mass data from Smith and Jungers (1997). Body masses are primate groups, are also beginning to receive averages of male and female means for the included species. Results are unchanged using female mass rather than average increasing attention (Charnov and Berrigan, 1993; mass. Kelley, 1997, 2002; Ross, 1998; Godfrey et al., 2001; Macho, 2001). Among catarrhines, extant apes and Old World inferred from the slowed life histories of gibbons, monkeys can be characterized as having under- which probably diverged from the great apes in the gone life-history divergence; apes have relatively early Miocene or early middle Miocene (Caccone slow life histories for their body mass whereas and Powell, 1989), but ultimately this hypothesis monkeys appear to have relatively fast life histories can only be tested in the fossil record. for their mass (Fig. 1; see also Harvey and Importantly, the above hypothesis presumes Clutton-Brock, 1985; Watts, 1990; Kelley, 1997). that life histories have changed in both the This difference is most evident in a comparison of hominoid and cercopithecoid lineages from a gibbons and monkeys, as the body mass range of primitive catarrhine condition that was broadly gibbons (approximately 5–10 kg) falls entirely intermediate, with life-history prolongation in within that of Old World monkeys, and average hominoids and acceleration in cercopithecoids. mass in the two groups is similar. For the timing of However, it is presently unclear that this presump- any given life-history trait in relation to body tion is warranted, an issue that will be further mass, gibbons lie above the primate regression line explored below. while Old World monkeys lie below (Fig. 1). It has The principal means for inferring the life histo- been hypothesized that the life-history divergence ries of fossil species has been through the chronol- between apes and Old World monkeys had its ogy of dental development. The timing of dental genesis soon after the cladogenesis of the two development in all mammals is highly correlated groups (Kelley, 1997), which probably took place with ontogeny as a whole; a functioning dentition in the late Oligocene to earliest Miocene (Kumar must be in place when an animal is weaned and and Hedges, 1998). This could plausibly be must develop in a way that will last for the J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 309 projected lifetime of the individual. The link 2002). However, the relatively late date for S. between dental development and ontogeny is evi- parvada limits its usefulness as a meaningful test denced by the correlations between aspects of of the hypothesis of an early life-history diver- dental development and individual life-history gence between apes and monkeys in the latest variables (Smith, 1989, 1991, 1992). Dental devel- Oligocene. opment is, in a sense, just another life-history trait The second fossil ape was an individual of (Smith and Tompkins, 1995), but one that is Afropithecus turkanensis from the early Miocene of preserved in the fossil record. While there is Kenya (Kelley, 1999, 2002). In the following analy- systematic variation in the relationship between sis we revise the earlier estimate of age at first dental development and various life-history molar emergence for this individual, which was attributes, primarily associated with variation in preliminary and lacked a full description of the diet (Godfrey et al., 2001), as well as occasional methods of analysis. The revised estimates re- idiosyncratic variation associated with specific ported here incorporate new data on molar crown ecological demands (Godfrey et al., 2002; formation in Afropithecus (see also Smith et al., Schwartz et al., 2002), within a broad framework 2003) and a more thorough and rigorous analysis the pace of dental development serves as a reliable of relevant comparative data. Knowing the age at proxy for the pace of life history as a whole. first molar emergence in Afropithecus is important Among living primates, it has been demonstrated because it nearly doubles the antiquity of such that age at first molar emergence is a particularly estimates for fossil apes, approaching the esti- good correlate of various life-history traits (Smith, mated date of divergence of apes and Old World 1989, 1991), emergence being defined as the initial monkeys. In addition, this analysis provides penetration of the oral gingiva by the molar cusps. further data for the documentation of dental Thus, if the average age at first molar emergence development in fossil apes, which complements can be established for a fossil species, then its information on developmental chronology and general life-history profile can be characterized as crown formation times derived from histological well. studies (Beynon et al., 1998; Zhao et al., 2000; The most straightforward approach to estimat- Kelley et al., 2001; Smith et al., 2001, 2003; ing age at first molar emergence in fossil species is Schwartz et al., in press). to determine the age at death for individuals that died while in the process of erupting their first molars, making necessary adjustments if the Materials and Methods stage of eruption differs from that associated with gingival emergence. Ages at death can be deter- The Afropithecus specimen used in this analysis mined with a high degree of precision using the is KNM-MO 26, a partial right mandibular corpus record of incremental growth lines that are pre- of an infant from the site of Moruorot in Kenya served in all teeth, including fossilized teeth (Fig. 2). Moruorot lies in the Lothidok Range west (Boyde, 1963; Bromage and Dean, 1985; Dean of Lake Turkana, approximately 10 km southeast et al., 1986, 1993b; Dean, 1987a, 1989; Beynon of Kalodirr, the site from which most remains of et al., 1991; Macho and Wood, 1995; Kelley, 1997, Afropithecus have been recovered (Leakey et al., 2002; Dirks, 1998; Antoine et al. 1999). 1988; Leakey and Walker, 1997). The Moruorot To date, age at first molar emergence has been localities lie within the lower part of the Kalodirr directly calculated for only two fossil apes. The Member of the Lothidok Formation and therefore first was an individual of Sivapithecus parvada date to approximately 17.5 Ma (Boschetto et al., from a 10 Ma locality in the Siwaliks of Pakistan. 1992), or late early Miocene. This individual was found to have an age at first KNM-MO 26 preserves the deciduous fourth molar emergence that was well within the range of premolar (dP4) and first molar (M1), as well as the extant chimpanzees, probably equal to or slightly permanent lateral incisor (I2), canine (C) and greater than the chimpanzee mean (Kelley, 1997, premolar (P) germs within their crypts (Figs. 2 310 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329

Fig. 2. Infant mandible of Afropithecus turkanensis (KNM-MO 26) from Moruorot, Kenya, showing the erupting M1 and the I2 germ within its crypt; (a) lingual, (b) occlusal.

and 3). The M1 was in the process of erupting the labial and lingual surfaces of the tooth and the when the individual died, with the cusp apices mesial aspect of the crown apex. lying just superior to the mandibular alveolar margin. The alveolar bone mesial to the I2 germ Calculating age at death was broken away, exposing the tooth within its crypt surrounded by a hardened matrix. The Age at death for the individual represented matrix within the crypt was carefully removed with by KNM-MO 26 can be determined using the a dental pick and needle probe, exposing much of I2 germ, which was still developing when the J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 311

Fig. 3. Radiograph of KNM-MO 26. individual died (the incisor crowns are still devel- among individuals within a species (Dean, oping when the first molar emerges in most higher 1987a,b, 1989; Dean and Beynon, 1991; Beynon primates). Age at death is calculated by adding the et al., 1991; Dean et al., 1993a; FitzGerald, 1998; time that elapsed between birth and the inception Schwartz et al., 2001). The surface manifestations of I2 crown mineralization (postnatal delay) to the of the Retzius lines are known as perikymata, duration of crown formation up until the time of which have, therefore, the same periodicity as the death. Retzius lines (Fig. 4). Crown formation time is determined using the Crown formation time in teeth that have not incremental growth lines preserved in the enamel. completed their development is calculated as the The use of incremental growth lines for determin- sum of the time required to form the cuspal enamel ing ages at death in infants is discussed in detail in plus the time to form the amount of lateral enamel Boyde (1963), Bromage and Dean (1985), Dean present at the time of death. Cuspal enamel is the et al. (1986, 1993a,b), Dean (1987b), Beynon et al. earliest formed enamel, in which successive Retzius (1991), Kelley (1997), Antoine et al. (1999), and lines are completely buried under subsequently Ramirez Rozzi (2002) and will be only briefly formed enamel. Lateral enamel is defined as the reviewed here. Enamel incremental lines include enamel formed subsequent to the first Retzius line both short-period lines known as cross-striations, that reaches the crown surface (see illustrations in which record daily increments of enamel depo- Bromage and Dean, 1985; Beynon and Wood, sition, and long-period lines known as striae of 1986; Macho and Wood, 1995; Ramirez Rozzi, Retzius or Retzius lines, which record brief, 2002; Smith et al., 2003). periodic disruptions in ameloblast secretion across It was not possible to section the teeth of the entire developing enamel front. Retzius line KNM-MO 26 to directly observe histological periodicity, which is the number of daily cross- structures. Therefore, values for certain growth striations between Retzius lines, is constant within parameters listed above had to be estimated from all teeth of an individual, but varies to some extent studies of dental development in extant apes 312 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329

Fig. 4. Naturally fractured surface of the lateral enamel of an Afropithecus turkanensis molar (KNM-WK 24300, RM2) showing striae of Retzius meeting the surface of the enamel and forming perikymata. Note that perikymata only form where striae of Retzius reach the tooth surface. The cervix of the tooth is below the bottom right edge of the image. The ﬁeld width of the image is approximately 750 µm.

and other fossil primates, and from information on I2 postnatal delay from all extant apes and derived from a histological study of two Afro- humans for which histological data were available pithecus molars (see Smith et al., 2003). For each [histological and radiological determinations of the of the growth parameters there is both intra and inception of mineralization can differ substan- interspecific variation. Thus, several estimates of tially, with histological determination being more age at death and age at M1 emergence were accurate (Beynon et al., 1998; Reid et al., calculated for KNM-MO 26, using combinations 1998a,b)]. Comparable data were also available for that incorporated conservative minimum values, as the I2 of Proconsul heseloni (Beynon et al., 1998), a well as average values. The estimates of age at M1 possibly closely related contemporary of Afro- emergence reported here therefore include a mini- pithecus (Begun et al., 1997; Leakey and Walker, mum estimate, as well as a range of more probable 1997; Harrison, 2002). estimates, since it is improbable that any one individual will express the minimum known values Duration of cuspal enamel formation for all growth parameters. We also calculated a This growth parameter also had to be estimated single maximum estimate for age at M1 emergence from published values for extant apes and humans, using the maximum known values for each esti- and for P. heseloni, since the I2 of KNM-MO 26 mated growth parameter. Specific issues pertaining could not be sectioned for direct observation. to the estimates or calculations of each of the Cuspal enamel formation time is related to enamel growth parameters are discussed below. thickness, although the relationship between the two will vary among species depending upon

Postnatal delay in the inception of I2 ameloblast secretion rates and the degree of mineralization sinuosity of the enamel prisms as they course from To estimate the postnatal delay in the the enamel-dentine junction (EDJ) to the tooth

KNM-MO 26 I2, we compiled comparative data surface (Dean, 1998). As stated above, the cuspal J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 313 enamel thickness of the KNM-MO 26 I2 is Appendix 1), combined with the spacing of unknown. While cuspal enamel thickness and the perikymata adjacent to the non-expressed region duration of cuspal enamel formation have been on the KNM-MO 26 I2. calculated for two A. turkanensis molars (Smith The periodicity of Retzius lines was determined et al., 2003), it cannot be assumed that incisor in two Afropithecus molars that were sectioned as cuspal enamel will have the same values (Dean and part of a separate study on enamel thickness and Reid, 2001). microstructure (Smith et al., 2003), using two methods described by Dean et al. (1993a,b) and Duration of lateral enamel formation Swindler and Beynon (1993). Where possible, This value is calculated by multiplying the direct counts of the number of cross-striations number of perikymata times the periodicity between adjacent Retzius lines were made by using (number of cross-striations between Retzius lines). both scanning electron and polarized light micro- To obtain a count of the perikymata on the scopic images. Additionally, the average spacing

KNM-MO 26 I2, the tooth was molded within its between Retzius lines was divided by the average crypt using Colte`ne President Plus Regular spacing of cross-striations measured from prisms Body, and a replica made using Ciba-Geigy in the same area, and the number rounded to the Araldite GY 506 epoxy resin cured with nearest whole integer (see Smith et al., 2003 for hardener HY 956. The replica was then sputter- details of specimen preparation and methodology). coated with a thin layer of gold-palladium and examined with a JEOL scanning electron micro- Estimating age at gingival ermergence of M1 scope. A photo montage was constructed of the based on age at death mesio-labial tooth surface from the developing cervical region to a point near the crown incisal As previously noted, the individual represented edge. by KNM-MO 26 died before the erupting M1 Perikymata were not expressed over the apical would have emerged from the gingiva. Since

2.5 mm of the I2 crown, due to a combination of gingival emergence is the standard for comparison post-depositional chemical weathering or abrasion of age at M1 eruption among extant species, the and a true fading out of perikymata expression age at which this would have occurred in over the apical-most portion of this interval. The KNM-MO 26 must be estimated. This was done latter is most likely due to the acute angle of using comparative data on M1 emergence in extant incidence of the Retzius lines to the tooth surface baboons, and compared to results from a previous apically, which sometimes results in the non- study on chimpanzees (Zuckerman, 1928). expression of perikymata for the apical-most From June, 1995 through June, 1999 the senior Retzius lines of lateral enamel. A similar trend in author carried out a longitudinal radiographic perikymata expression was seen in the two Afro- study of M1 emergence and root formation on a pithecus molars reported on by Smith et al. (2003). captive breeding colony of Papio anubis housed Even when there are no apical perikymata at the Biological Research Laboratories at the expressed, histological sections of hominoid University of Illinois at Chicago. Approximately permanent anterior teeth demonstrate that the every three months, all of the baboons in the ﬁrst Retzius line reaching the tooth surface colony were sedated for TB testing. While under (delineating cuspal from lateral enamel) is invari- sedation, all infant individuals between approxi- ably near the incisal edge (C. Dean, personal mately one and three years of age were given oral communication). Thus, the number of Retzius examinations, and periapical x-rays were taken of intervals in the apical-most 2.5 mm of the the M1 and lower deciduous premolars. KNM-MO 26 I2 crown had to be estimated. This On several occasions, examination coinci- was done using perikymata counts over the same dentally occurred either just as the M1 mesial cusps interval in two lower lateral incisor crowns of P. (the ﬁrst to present) were emerging from the heseloni and P. nyanzae (Beynon et al., 1998, gingiva, with emergence clearly having taken place 314 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329

Table 1 Table 2

Postnatal delay in the inception of I2 mineralization in Duration of I2 cuspal enamel formation in Proconsul heseloni Proconsul heseloni and extant hominoids and extant hominoids

Species Months Source Species Months Source Proconsul heseloni 1.5 Beynon et al., 1998 Proconsul heseloni 4.0 Beynon et al., 1998 Hylobates lar 3.7 Dirks, 1998 Hylobates lar 4.5 Dirks, 1998 Gorilla gorilla 11.0 Beynon et al., 1991 Gorilla gorilla 4.0 Beynon et al., 1991 Pan troglodytes 2.5 Reid et al., 1998a Pan troglodytes 5.3,6.4† Reid et al., 1998a 7.9 Reid et al., 1998a 5.8 Reid et al., 1998a 8.4 Reid et al., 1998a 6.0,6.8† Reid et al., 1998a Pongo pygmaeus 13.0 Beynon et al., 1991 6.4 Reid et al., 1998a Homo sapiens 0 Dean and Beynon, 1991 Homo sapiens 5.5 Reid et al., 1998b 4.8 Reid et al., 1998b 5.8 Reid et al., 1998b 8.3 Dean et al., 1993a 7.0,7.7† Reid et al., 1998b 8.0 Reid et al., 1998b †Antimeres from the same individuals. within the preceding few days, or when the M1 mesial cusps were visible beneath a thin layer of between the earliest and latest onsets among three gingival tissue, from which it was determined that human I s. These data make it diﬃcult to establish gingival emergence was imminent. For four of 2 any criteria by which to choose the most appropri- these individuals, there was a radiographic record ate estimate for the postnatal delay in the of previous examinations that could be used to KNM-MO 26 I . Therefore, in keeping with our estimate the time interval between gingival emer- 2 methodological protocol, we selected the minimum gence and the developmental stage equivalent to and average values of 0 and 6.6 months, respec- that of the KNM-MO 26 individual when it died; tively, as estimates for the Afropithecus infant. that is, with the M1 cusp apices just above the mandibular alveolar margin. Duration of cuspal enamel formation Similar, although less precise, data on M1 eruption and gingival emergence were reported for Data on the duration of I cuspal enamel two infant chimpanzees by Zuckerman (1928). 2 formation in extant apes and humans, and in P. heseloni, are shown in Table 2. Values for extant apes and humans range between 4.0 and 8.0 Results months, with a mean of 6.0 months. Both intra-

Postnatal delay in the inception of I2 speciﬁc and interspeciﬁc variation for the duration mineralization of cuspal enamel formation are substantially less

than for the postnatal delay in I2 mineralization. Data on the postnatal delay in I2 mineralization in extant apes and humans, and in P. heseloni, are Duration of lateral enamel formation shown in Table 1. Values among extant apes and humans range from 0 to 13.0 months, with a mean The labial face of the KNM-MO 26 I2 preserves of 6.6 months. Based on these limited data, there 82 perikymata (Fig. 5). Perikymata are expressed are no obvious associations between the duration from within about 2.5 mm of the crown apex down of the postnatal delay in mineralization and either to the last formed immature enamel close to the body size or phylogeny, although the two highest advancing cervical line. values are most likely the males of the two largest Our estimate of the number of Retzius lines extant apes. Where there are data on intraspeciﬁc unexpressed as surface perikymata over the apical- variation it appears to be substantial; for example, most 2.5 mm of lateral enamel is based in part on there is more than an eight month diﬀerence the number of perikymata over the same part of J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 315

the crown in P. nyanzae and P. heseloni. Starting from the incisal edge, perikymata counts in 1 mm increments over the apical 3.0 mm of the crown of

a P. nyanzae I2 (KNM-RU 1716) are 2, 8 and 8; the same values in an I2 of P. heseloni (KNM-RU 7290) are 8, 11 and 13 (Beynon et al., 1998). Eliminating half of the perikymata over the cervical-most 1 mm increment (to equal 2.5 mm) results in totals of 14 and 26 perikymata over this interval, with a mean of 20. We thus used 14 as our conservative estimate and 20 as an average estimate of the number of unexpressed perikymata in the apical 2.5 mm of lateral enamel in the Afro-

pithecus I2. The total number of perikymata plus lateral enamel Retzius lines not expressed as

perikymata in the KNM-MO 26 I2 was therefore estimated to be 96 or 102 (82 plus either 14 or 20).

Perikymata on the Afropithecus I2 itself suggest that the higher estimate is likely to be closer to the actual value. The apical-most 1 mm of the

KNM-MO 26 I2 crown over which perikymata are expressed (adjacent to the non-expressed region) contains 12 or 13 perikymata. Since perikymata spacing over the entire labial crown surface is highly uniform (see Fig. 5), it seems likely that perikymata numbers closer to those of the P. heseloni incisor would have been present in the

Afropithecus I2 as well. The Retzius line periodicities in the two sectioned Afropithecus molars were determined to be 7 and 8 days, respectively (Fig. 6; see also Smith et al., 2003). Multiplying these values by the lateral enamel perikymata estimates of 96 and 102 gives an estimated range for the duration of lateral

enamel formation in the KNM-MO 26 I2 of between 672 days (22.2 months) and 816 days (26.9 months) (Table 3).

Age at death of KNM-MO 26

A range of estimates of the age at death for the KNM-MO 26 individual using the calculated and estimated growth parameters detailed above is shown in Table 3. Two sets of estimates are given, one using the minimum values for the growth Fig. 5. SEM montage of the mesio-labial face of the KNM-MO parameters estimated from extant apes and Pro- 26 Afropithecus I2 showing the development of perikymata. The interval over which perikymata are expressed equals 5.8 mm. consul species, and the other using average values. Cervical is toward the bottom. The minimum estimates, using 7 and 8 day Retzius 316 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329

Fig. 6. Polarized light micrograph of the outer lateral enamel of Afropithecus turkanensis (KNM-WK 24300, RM2). The surface of the tooth is at the right and the cervix is toward the bottom. Enamel prisms are shown running from left to right, with striae of Retzius (large white arrows) and cross-striations (small white arrows) crossing the prisms. The periodicity of adjacent Retzius lines is eight cross-striations, each representing one daily increment of enamel deposition. Note that there are seven cross-striations between Retzius lines, the eighth being on the Retzius line. The ﬁeld of width of the image is approximately 155 µm.

line periodicities, are 26.2 and 29.4 months. Using of death and the eventual time of M1 emergence is the average values for the estimated growth based on the progression and timing of ﬁrst molar parameters results in estimated ages at death of eruption in extant baboons and chimpanzees.

34.8 to 39.5 months, also based on 7 and 8 day In all four of the baboons for which M1 gingival Retzius line periodicities. emergence was observed, the radiograph taken

three months prior reveals the M1 mesial cusps to Time from death to M1 emergence in be at or just below the level of the alveolar margin KNM-MO 26 (Figures 7 and 8). This is a slightly earlier eruption stage than had been reached by the KNM-MO 26

Our estimate of the time interval between the individual when it died, in which the M1 cusp developmental stage of KNM-MO 26 at the time apices are just above the alveolar margin (Figs. 2 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 317

Table 3 Estimated ages at death in KNM-MO 26 (months)

Retzius periodicity I2 postnatal delay I2 cuspal enamel I2 lateral enamel Age at death Minimum 7 0.0 4.0 22.2–23.6 26.2–27.6 8 0.0 4.0 25.4–26.9 29.4–30.9 Average 7 6.6 6.0 22.2–23.6 34.8–36.2 8 6.6 6.0 25.4–26.9 38.0–39.5

Retzius periodicity in days; all other values expressed in months. Age at death equals the sum of the I2 postnatal delay, the cuspal enamel formation time and the lateral enamel formation time. The ranges for the duration of I2 lateral enamel formation reﬂect the use of either 96 or 102 lateral enamel perikymata (see text). Minimum and average estimates explained in text.

and 3). Judging by the amount of tooth movement interval between M1 cresting the alveolar margin that took place in the four baboons during the and gingival emergence probably takes somewhere three month interval between the first x-ray and between four and five months in chimpanzees, in gingival emergence (Figures 7 and 8), we estimate contrast to the three months that it takes in that the small difference in the M1 eruption stage baboons. between KNM-MO 26 and the four baboons at Combining the baboon and chimpanzee data, the time of the first x-ray corresponds to at most the additional time that would have been needed to one month. According to the baboon eruption achieve M1 gingival emergence in the KNM-MO schedule, therefore, KNM-MO 26 would have 26 infant can be estimated to be between about died approximately two months prior to gingival two and four months. Adding two and four emergence. While the eruption schedules in these months to each of the estimates of the age at death captive baboons may be somewhat accelerated of KNM-MO 26 produces minimum estimates for compared to those of wild baboons (see data in the age at M1 emergence of 28.2 and 31.4 months Table 5; also Phillips-Conroy and Jolly, 1988), the using the baboon schedule of M1 eruption (based percentage difference is not likely to be significant on 7 and 8 day Retzius periodicity, respectively), for the time interval being discussed here of only a or 30.2 and 33.4 months using a chimpanzee few months (see further below). schedule (Table 4). Estimates of age at M1 emer- Zuckerman (1928) presented similar oral gence using average rather than minimum values examination and radiographic data on two for estimated I2 crown growth parameters range infant chimpanzees, which, as expected, suggest a between 36.8 and 43.5 months. Although not somewhat slower M1 eruption schedule than in included in the various calculated estimates, the baboons. The data for one animal in particular, maximum estimated age at M1 emergence in Clarence, are sufficiently precise for comparison KNM-MO 26 using the maximum values of all with the baboon results described above. An initial estimated growth parameters, combined with radiograph revealed the M1 “crown” to be level the highest estimate of missing lateral enamel with the alveolar margin, which we interpret to perikymata (26) and an 8 day periodicity, is 53.4 mean that the cusp apices were at the alveolar months. Since the principal concern of this study margin. In a second radiograph six months later was to determine whether or not age at M1 emer- the M1 was described as being “fully erupted,” gence in the KNM-MO 26 individual was earlier which presumably means that the tooth was level than in extant chimpanzees, the maximum esti- with the occlusal plane of the deciduous premo- mate will not be discussed further. We simply note lars. An oral examination was made sometime that it is equally unlikely that one individual would between five and six months after the first radio- uniformly express the maximum known values graph was taken and revealed that the M1s “were of all the estimated growth parameters as that cutting the gums” (Zuckerman, 1928, p. 25). it would uniformly express all the minimum Taken together, these observations suggest that the values. 318 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 319

Figures 7 and 8. Periapical radiographs of the mandibular deciduous premolars and erupting M1 in two infant Papio anubis, Nos. 6216 (Fig. 7) and 6219 (Fig. 8), housed at the Biological Research Laboratories, The University of Illinois at Chicago. For both individuals (a) was taken 6/24/96 and (b) was taken 9/25/96, the latter coincident with the initial gingival emergence of the ﬁrst cusp as revealed by oral examination. 320 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329

Table 4

Estimated ages at M1 emergence in KNM-MO 26 (months)

Retzius periodicity Age at death Age at M1 emergence (baboon model) Age at M1 emergence (chimpanzee model) Minimum7 26.2–27.6 28.2–29.6 30.2–31.6 8 29.4–30.9 31.4–32.9 33.4–34.9 Average 7 34.8–36.2 36.8–38.2 38.8–40.2 8 38.0–39.5 40.0–41.5 42.0–43.5 Retzius periodicity in days; all other values expressed in months. Age at death estimates from Table 3. The baboon model adds two months to the age at death while the chimpanzee model adds four months (see text for explanation). Minimum and average estimates explained in text.

Discussion Table 5 Age at M emergence (months) in Age at ﬁrst molar emergence and life history in 1 Afropithecus turkanensis, Pan troglodytes and extant cercopithecids Afropithecus Afropithecus turkanensis† 28.2–43.5 Extant species Mean Minimum Maximum While the minimum estimate for age at M1 emergence in KNM-MO 26 presented here is only Pan troglodytes 39.1 25.7 48.0 Macaca mulatta 16.2 12.5 22.6 slightly later than previously reported for this M. fascicularis 16.4 14 20 individual (Kelley, 1999, 2002), the other newly M. nemestrina 16.4 – 18.6+ calculated estimates greatly extend the range into M. fuscata 18.0 – <24 which the actual age probably falls. This extended Cercopithecus aethiops 10.0 7.9 12.0 range is partly due to the addition here of actual Papio anubis 1 20.0 >16 <25 P. anubis 2 16.7 15.7 <21 data on Retzius line periodicity in A. turkanensis, which is greater than the estimated value used in Sources: Pan troglodytes (Smith et al., 1994); Macaca the earlier reports. It also reﬂects increases in the mulatta, Cercopithecus aethiops (Hurme and van Wagenen, 1961); Macaca fascicularis (Bowen and Koch, 1970); Macaca estimates for the duration of incisor cuspal enamel nemestrina (Swindler, 1985; B.H. Smith, personal formation and the postnatal delay in incisor communication—based on Swindler data); Macaca fuscata mineralization, based on a more thorough analysis (Smith et al., 1994; B.H. Smith, personal communication— of the comparative data from extant species. based on data in Iwamoto et al., 1987); Papio anubis 1 (Smith The estimates for the age at M emergence in et al., 1994; B.H. Smith, personal communication—based on 1 data in Kahumbu and Eley, 1991); Papio anubis 2 (J. Kelley, KNM-MO 26 (28.2–43.5 months) fall within the unpublished data from a longitudinal study of M1 eruption range of Pan troglodytes (25.7–48.0 months), and in a captive colony at The University of Illinois at Chicago). encompass the chimpanzee mean of 38.9 months Reliability of range data for extant species varies depending (Table 5). The means of the range of estimates on methodology and sample size. †Range of estimates. using baboon and chimpanzee schedules of M1 eruption are, respectively, 34.9 and 36.9 months. What is most important from our perspective, however, is that even the absolute minimum esti- schedules. Phillips-Conroy and Jolly (1988) mate of 28.2 months is well outside the ranges of reported that the eruption schedules of captive

M1 emergence of even the largest extant cerco- baboons were accelerated compared to those of pithecids for which there are reliable data, the wild-living animals, but neither Kahumbu and maximum age being less than 25 months (Table 5). Eley (1991) nor Iwamoto et al. (1987) found any Most of the comparative data in Table 5, how- systematic differences between wild and captive ever, are from captive animals. As noted by Smith populations of, respectively, baboons and et al. (1994), data are equivocal regarding the macaques. The baboon data in Table 5 tend to degree to which wild and captive populations support accelerated dental development and erup- might be expected to differ in their dental eruption tion in captive animals. Of the two Papio anubis J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 321 populations reported upon in Table 5, P. anubis 1 was a wild population whereas P. anubis 2 is the breeding colony at The University of Illinois at Chicago. The mean percent acceleration in the age at M1 emergence in the captive population is 16.5%, but, as noted by Iwamoto et al. (1987), this may simply reflect genetic differences between these two particular populations, rather than a systematic effect to be expected from all wild-captive comparisons. Importantly, with respect to interpretation of KNM-MO 26, it is still the case that the upper limit of the cercopithecid range data is for the wild P. anubis population. Given the broad relationship within higher taxa between body size and life- history variables (including dental development), it Fig. 9. Least squares regression of age at weaning against age at is likely that the wild P. anubis maximum from M1 emergence, both in months and log-transformed, for 20 extant non-human primate species. Included species are those Table 5 is near the upper limit of the range of at from Table 6, with the following exclusions because of a lack of M1 emergence ages for all cercopithecoids. weaning age data: Cheirogaleus medius, Galago senegalensis, While it is possible that KNM-MO 26 repre- Macaca fuscata, and Homo sapiens. Age at weaning from Godfrey et al. (2001); age at M1 emergence from Smith et al. sents an individual that is near the maximum of (1994). the A. turkanensis range of M1 emergence ages, it is more probable as a simple consequence of central tendency that it is closer to the species mean. To life histories of fossil species, for example as Old produce a mean age of M1 emergence in A. turka- World monkey-like, ape-like or human-like. How- nensis that is within the cercopithecid range, and ever, the correlations between life-history variables therefore outside the chimpanzee range, would and M1 emergence in extant primates are insuffi- require that, (1) the minimum estimate for age at ciently robust, and the errors in estimates of age at

M1 emergence in KNM-MO 26 is the closest of the M1 emergence in fossil species are too large, to various estimates to the actual age, and (2) that reliably calculate the values of specific life-history even this age is near the maximum for the species variables in fossil species based solely on estimates as a whole. This is a possible, but much less of age at M1 emergence (see Smith et al., 1995; probable, set of circumstances, we conclude, Smith, 1996). The estimated age at M1 emergence therefore, that even though our analysis is for a in KNM-MO 26, and the implications of this single individual, the mean age of M1 emergence in estimate for characterizing age at M1 emergence in A. turkanensis was within the range of extant A. turkanensis as a whole, suggest that life history chimpanzees, and perhaps close to the chimpanzee in this early hominoid can be broadly character- mean. ized as having been like that of living great apes. As noted earlier, among primates, age at M1 emergence is correlated with a variety of life- First molar emergence and life-history evolution in history attributes (Smith, 1989, 1991; also Fig. 9). Hominoidea and other primates Because these analyses encompass species representing all major groups of extant primates, age at In the following discussion of life-history evolu- M1 emergence can be used to infer life history in tion, age at M1 emergence is used as a substitute or fossil primates that lie within the extant primate proxy variable for the overall pace of life history. radiation. The strength of the M1 emergence-life There are several reasons for doing this rather than history relationships show that M1 emergence data using the specific life-history traits with which age can legitimately be used to generally categorize the at M1 emergence is correlated, and that more 322 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 directly reflect reproductive and maturational of additional extant and fossil species will help to milestones. First, in many species, average age at elucidate the relationships between molar crown M1 emergence is known more reliably than are the formation times, ages at molar emergence, and life average values for many other life-history vari- history attributes. ables. Second, it is likely that there is more facul- It has been hypothesized that the slowed life tative intraspecific variation in reproductive and histories that characterize the extant apes, particu- maturational traits than there is in age at M1 larly the great apes, might have had their genesis emergence. Plasticity in life-history traits, even during the early evolutionary history of the over the course of individual life spans, appears to Hominoidea, and that they might in fact have been be an important aspect of life-history adaptation. the fundamental adaptive shift underlying the Such plasticity is not likely to be reflected in dental cladogenesis of hominoids and cercopithecoids development. Finally, at present, M1 emergence is (Kelley, 1997). Prior to the analysis described here, one of only two or perhaps three variables (the the oldest fossil ape for which age at M1 emergence others being brain size and possibly molar crown had been determined was a 10 Ma individual of formation time) by which extinct species can be Sivapithecus parvada from the Siwaliks of Pakistan included in discussions of life history. (Kelley, 1997, 2002). The estimate of age at M1 Regarding molar crown formation, a recent emergence for the 17.5 Ma individual of A. turka- study by Macho (2001) demonstrated that, within nensis nearly doubles the antiquity of such esti- primates as a whole, both M1 crown formation mates for fossil hominoids and hominids. The time and average molar crown formation time are finding of an age at M1 emergence that is essen- significantly correlated with a number of life- tially like that of extant chimpanzees, and the history traits. However, there are a number of inference therefore of an essentially modern reasons for withholding judgment on these results. great ape life history in A. turkanensis, could Among these are the use of the primate life-history be viewed as lending additional support to the data compiled by Harvey and Clutton-Brock above hypothesis of life-history evolution in the (1985), much of which is now known to be either in Hominoidea. error or at least unreliable (see, for example, Smith There are, however, a number of ways to inter- et al., 1995 and Smith and Jungers, 1997). More- pret the data on M1 emergence in primates as a over, many of the crown formation times in whole, with different implications for life-history Macho’s study are calculated estimates rather than evolution in the Hominoidea and other higher direct histological measurements (Shellis, 1998), primate taxa. Fig. 10 shows two different interpre- some of which are demonstrably in error when tations of the relationship between M1 emergence compared to known crown eruption ages (Smith and body mass in extant primates. Plotted in each et al., 1994). Finally, for some species (e.g., are the 24 extant primate species (including chimpanzees and humans) molar crown formation humans) for which there are reliable data on age at times simply do not reflect known differences in M1 emergence (Table 6). Fig. 10a shows a single life-history values, differences that are more or less linear regression for all the included extant species, concordant with average ages at M1 emergence. purposely left unidentified. Fig. 10b shows the The reason for this may have to do with variation various species identified by higher taxonomic in rates of root formation, especially initial root group, and also includes estimates of age at M1 formation. Disparities between crown formation emergence and body mass for A. turkanensis and times and inferred or calculated ages at M1 emer- S. parvada. Statistics for the various regressions gence are beginning to become apparent among depicted in Fig. 10 are shown in Table 7. fossil apes as well, exemplified by comparisons Based on Fig. 10a, the late age at M1 emergence between the similarly sized Proconsul nyanzae, in Afropithecus could be interpreted as a simple Dryopithecus laietanus, and Afropithecus turkanen- consequence of large body mass (estimated aver- sis (Beynon et al., 1998; Kelley et al., 2001; Smith age mass=30 kg), without any necessary phylo- et al., 2003). It can be expected that further study genetic implications. The correlation between age J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 323

at M1 emergence and body mass for all the included species is highly signiﬁcant (Table 7). By this interpretation, any primate with the body mass of a small chimpanzee would be expected

to have an age of M1 emergence within the chimpanzee range, and, by implication, an overall life history that was similar to that of a chimpanzee. In Fig. 10b, there is still a signiﬁcant relation-

ship between age at M1 emergence and body mass within the diﬀerent higher taxa (excepting hominoids, which is a simple consequence of small sample size). However, the four included higher taxa (plus Homo) are also characterized by a series

of apparent grade-shifts in age at M1 emergence. The grade-shifts from lemuriforms to each of the anthropoid groups, from cercopithecids to hominoids, and from non-human hominoids to Homo are clear. Interpretation of the cebid regression is less so, as it implies that a cebid the size of

a chimpanzee would have an age at M1 emergence that is greater than that of modern Homo. Understanding life-history evolution in platyrrhines is in fact critical for interpreting the significance of M1 emergence data among fossil hominoids, and for understanding life-history evolution in hominoids more generally. A plausible interpretation of the data in Fig. 10b is that both platyrrhines (here represented only by cebids) and hominoids broadly represent the primitive anthropoid condition and that cercopithecids are derived with respect to both, having accelerated life Fig. 10. Two possible interpretations of age at M1 emergence in histories (see also Fig. 1). In this case, the abso- relation to body mass in primates (all regressions are least squares—regression statistics in Table 7); (a): Phylogeny- lutely more prolonged life histories of apes relative neutral interpretation; (b): Phylogeny-based interpretation to platyrrhines would be most reasonably inter- revealing apparent grade-shifts in age at M1 emergence. preted as a simple consequence of increasing Symbols for (b): Lemuriformes (circles), Cebidae (triangles), Cercopithecoidea (squares), Afropithecus turkanensis (A), body size. However, the platyrrhine regression in Homo sapiens (H), Pan troglodytes (P), and Sivapithecus Fig. 10b has a very limited representation of taxa, parvada (S). For extant species, age at M1 emergence from consisting of several callitrichines and two species Smith et al. (1994) and Smith et al. (1995); body masses (average male and female mass) from Smith and Jungers (1997). of Cebus. The slope is substantially greater than in For A. turkanensis, average body mass estimated at 30 kg based either of the other anthropoid groups, but could be on postcranial size (Leakey et al., 1988) and an estimated male significantly altered with a more representative mass of 35 kg (Kappelman et al., in press); age at M1 emergence estimated at 36 months (see text). For S. parvada, average body sample of taxa. If life history has slowed in Cebus mass estimated at 61 kg based on postcranial size (see Kelley, relative to some or most other platyrrhine genera, 1988), and age at M1 emergence estimated at 43 months or if life history has accelerated in callitrichines, (Kelley, 1997; Kelley et al., in preparation). perhaps in association with dwarfing, then the slope is artificially high. If so, then the primitive condition for platyrrhines, and for catarrhines 324 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329

Table 6

Age at M1 emergence and body mass in extant primates

Species Age at M1 emergence (months) Average body mass (kg) Cheirogaleus medius 0.84 0.18 Varecia variegata 5.76 3.58 Lemur catta 4.08 2.21 Eulemur fulvus 5.04 2.13 E. macaco 5.16 1.82 Propithecus verreauxi 2.64 3.55 Galago senegalensis 1.20 0.21 Callithrix jacchus 3.72 0.37 Saguinus fuscicollis 4.10 0.35 S. nigricollis 3.35 0.48 Cebus albifrons 12.72 2.74 C. apella 13.80 3.09 Saimiri sciureus 4.44 0.72 Aotus trivirgatus 4.32 0.78 Cercopithecus aethiops 9.96 3.62 Macaca fascicularis 16.44 4.48 M. fuscata 18.00 9.51 M. mulatta 16.20 6.54 M. nemestrina 16.44 7.58 Papio anubis 20.04 18.10 P. cynocephalus 20.04 17.05 Trachypithecus cristata1 12.00 6.44 Pan troglodytes 39.12 53.00 Homo sapiens2 66.03 59.00

M1 emergence data from Smith et al. (1994) and body mass data from Smith and Jungers (1997), with the following exceptions: 1 M1 emergence data from Wolf (1984). 2 M1 emergence data from Smith et al. (1995).

Table 7 Statistics for Fig. 10 least squares regressions

Regression y-intercept Slope Correlation % Variance p Primate (ln)1 1.142 0.595 0.911 0.831 0.000 Primate (untransformed)2 5.144 0.876 0.941 0.886 0.000 Lemuriformes3 0.946 0.538 0.896 0.803 0.006 Cebidae3 1.833 0.636 0.960 0.922 0.001 Cercopithecidae3 2.053 0.341 0.799 0.638 0.017 Hominoidea (less Homo)3 2.438 0.305 0.972 0.944 0.152 1Fig. 10a. 2Not ﬁgured. 3Fig. 10b.

as well, might be intermediate between that of case, life-history evolution in catarrhines would hominoids and cercopithecids; note, for example, represent a true divergence, with life-history the position of the Alouattini with respect to acceleration in cercopithecoids and prolongation the other platyrrhine taxa in Fig. 1. In this in hominoids, and with both states being derived. J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 325

In the overall scheme of Fig. 10b, the age at M1 of equids, again even with substantial phyletic size emergence in Afropithecus may largely be a func- increase. tion of its being a hominoid. Certainly a lemuri- A similar pattern has recently been reported form or cercopithecoid with the body mass of in members of two lemuriform sister-taxa, one Afropithecus would be expected to have an age at extant (Indriidae) and one subfossil (Palaeo-

M1 emergence that is substantially earlier (see propithecidae) (Schwartz et al., 2002). Including further below). It is unclear at this point if this subfossil species, members of these two groups would also be the case for platyrrhines, or at least span the body mass range from average-sized for some platyrrhines. If platyrrhines and homi- monkeys to chimpanzees. Indriids are remarkable noids together broadly represent the primitive for having exceptionally precocious dental devel- anthropoid condition, then the M1 emergence age opment and an age at M1 emergence that is of Afropithecus would again largely be a function strongly temporally dissociated from many other of its size. life-history parameters (Godfrey et al., 2001, As important to this discussion is the homi- 2002). The recent confirmation of this phenom- noid regression. Since it includes only three enon in the chimpanzee-sized subfossil Palaeo- species, two of which are extinct species with propithecus (Schwartz et al., 2002), is compelling estimated ages at M1 emergence based on single additional evidence for the importance of individuals, its slope must also be regarded as phylogeny as well as body size in life-history highly uncertain. Moreover, the slope is largely evolution. determined by the body mass and age at M1 However, as noted earlier in the discussion of emergence estimates for A. turkanensis. Any the platyrrhine data, it cannot be assumed at changes in the estimates that we chose to repre- present that all members of the different primate sent A. turkanensis (30 kg average mass and 36 higher taxa plotted in Fig. 10b will fall within the months for age at M1 emergence, the overall life-history grade characteristic of the particular mean of the estimates) are likely to significantly taxon. Both the cebid and the lemuriform regres- alter the slope. The true nature of the relation- sion lines are based on two clusters of species of ship between body mass and age at M1 emer- substantially different body mass. Not only are the gence in hominoids will only become clear when regression lines uncertain as a consequence, but there are reliable data for the other great apes, the addition of species of intermediate or larger particularly Gorilla, and for gibbons. body mass might reveal departures among lower The possibility at least of phylogenetically taxonomic ranks (ie., genera/tribes) from whatever associated grade shifts in age at M1 emergence— dominant grade level that emerges. There might in and, by extension, in life history more generally— fact be a greater expectation of this in older, more that appear to have been established during the biologically diverse clades such as the platyrrhines early evolutionary history of the higher primate or lemuriforms than in more recent clades like the groups is interesting in a broader context. In a extant cercopithecids. series of empirical studies of life-history variation It can be anticipated that many plausible inter- in mammals as a whole, Harvey and colleagues pretations of life-history evolution in primates will

(Harvey et al., 1989a,b; Read and Harvey, 1989) be eliminated as the data on age at M1 emergence found that a disproportionate amount of the vari- improve for both extant and fossil primates. ation was at higher taxonomic levels, suggesting Extant taxa that are especially important for the early establishment of fundamental life-history improving the database include, in addition to the suites that are subject to comparatively little other great apes and gibbons noted above, atelids change during the subsequent evolutionary history and additional cebids, smaller cercopithecines, of the group, sometimes even in the event of additional colobines, and additional non-lemurid significant changes in body size. Likewise, Martin strepsirhines. Especially important fossil taxa and MacLarnon (1990) found evidence for con- include primitive catarrhines and species that servative life-history evolution in the fossil record extend the body mass ranges of their respective 326 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 groups, such as the fossil Theropithecus, very large It is proposed that life-history evolution in subfossil lemurs, and smaller fossil apes. primates more generally occurred as a series of grade shifts among higher level taxa. Some or all of these shifts may largely reflect phylogeny, Summary irrespective of body size. Others may reflect nothing more than changes in body size, still Estimates of the age at death were calculated for phylogenetically based, but along a common an infant of early Miocene A. turkanensis, trajectory for the pace of life history. Regardless of KNM-MO 26, that was in the process of erupting the predominant mode of change, it is becoming its first molar. The estimates were based on the increasingly clear that phylogeny as well as body perikymata preserved on the lateral incisor germ in size must be taken into account when attempting the mandible, combined with data on the duration to reconstruct the life histories of fossil primate of cuspal enamel formation and the length of the species. postnatal delay in the inception of I2 mineralization in extant apes and humans, as well as in species of Proconsul. A range of estimates was Acknowledgements calculated to accommodate both intra and interspecific variation in the latter growth parameters. We gratefully acknowledge the Government Perikymata periodicity was determined from histo- of Kenya and the National Museums of Kenya logical sections of two A. turkanensis molars. As for permission to study the Afropithecus fossils the eruption stage of the M1 in the A. turkanensis (Permit OP/13/001/10C 354 issued to JK). We infant reveals that it had not yet achieved gingival thank Meave Leakey and William Anyonge, past emergence when the animal died, estimates of the Heads of the Palaeontology Division, and Emma projected age at M1 emergence were calcu- Mbua, then Collections Manager of Palaeoanthro- lated from the age at death estimates combined pology, for facilitating our work at the KNM. with radiographic data on the progression of M1 JK expresses special gratitude to Jeff Fortman, eruption in baboons and chimpanzees. Associate Director, and Sam Rosado and the Estimates for the age at M1 emergence in other staff of the Biological Resources Laboratory, KNM-MO 26 ranged from approximately 28 to The University of Illinois at Chicago for their 43 months, well outside the ranges of large invaluable long-term assistance with the baboon extant cercopithecids for which there are radiographic project. We thank Chris Dean, comparable data, but comfortably within the Meave Leakey, Lawrence Martin, Don Reid, and range of chimpanzees. It is inferred from this result Holly Smith for innumerable valuable discussions that life history in A. turkanensis was essentially relating to the work reported here, and Terry like that of extant chimpanzees. This inference is Harrison plus three anonymous reviewers for compatible with the hypothesis that there was a comments on the manuscript. This research was shift to the prolonged life histories that charac- supported by National Science Foundation grant terize extant apes early in the evolution of the SBR-9408664 to JK. Hominoidea. However, this presumes a primitive condition for life history in catarrhines that was faster, when adjusted for body size, than in References chimpanzees. Limited data from extant platyrrhines may indicate that this is not the case. A Antoine, D., Dean, C., Hillson, S., 1999. The periodicity of plausible interpretation of these data is that extant incremental structures in dental enamel based on the devel- platyrrhines and hominoids together represent the oping dentition of post-medieval known-age children, in: Mayhall, J.T., Heikkinen, T. (Eds.), Dental Morphology primitive condition for life history, and that only 1998. Oulu University Press, Oulu, pp. 48–55. cercopithecoids among living anthropoids have Begun, D.R., Ward, C.V., Rose, M.D., 1997. Events in derived life histories, being highly accelerated. hominoid evolution, in: Begun, D.R., Ward, C.V., Rose, J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 327

M.D. (Eds.), Function, Phylogeny and Fossils: Miocene Dean, M.C., Reid, D.J., 2001. Perikymata spacing and distri- Hominoid Evolution and Adaptation. Plenum Press, New bution on hominid anterior teeth. Am. J. phys. Anthrop. York, pp. 389–415. 116, 209–215. Beynon, A.D., Wood, B.A., 1986. Variation in enamel thick- Dean, M.C., Stringer, C.B., Bromage, T.G., 1986. A new age at ness and structure in East African hominids. Am. J. phys. death for the Neanderthal child from the Devil’s Tower, Anthrop. 70, 177–193. Gibralter and the implications for studies of general growth Beynon, A.D., Dean, M.C., Reid, D.J., 1991. Histological and development in Neanderthals. Am. J. phys. Anthrop. study on the chronology of the developing dentition in 70, 301–309. Gorilla and Orangutan. Am. J. phys. Anthrop. 86, 189–203. Dirks, W., 1998. Histological reconstruction of dental develop- Beynon, A.D., Dean, M.C., Leakey, M.G., Reid, D.J., Walker, ment and age at death in a juvenile gibbon (Hylobates lar). A., 1998. Comparative dental development and micro- J. hum. Evol. 35, 411–425. structure of Proconsul teeth from Rusinga Island, Kenya. FitzGerald, C.M., 1998. Do enamel microstructures have J. hum. Evol. 35, 163–209. regular time dependency? Conclusions from the literature Bowen, W.H., Koch, G., 1970. Determination of age in and a large-scale study. J. hum. Evol. 35, 371–386. monkeys (Macaca irus) on the basis of dental development. Godfrey, L.R., Petto, A.J., Sutherland, M.R., 2002. Dental Lab. Anim. 4, 113–123. ontogeny and life-history strategies: The case of the giant Boschetto, H.B., Brown, F.H., McDougall, I., 1992. Stratigra- extinct indrioids of Madagascar, in: Plavcan, J.M., Kay, phy of the Lothidok Range, northern Kenya, and K/Ar ages R.F., van Schaik, C.P., Jungers, W.L. (Eds.), Reconstruct- of its Miocene primates. J. hum. Evol. 22, 47–71. ing Behavior in the Primate Fossil Record. Kluwer, New York, pp. 113–157. Boyde, A., 1963. Estimation of age at death of young human skeletal remains from incremental lines in the dental enamel. Godfrey, L.R., Samonds, K.E., Jungers, W.L., Sutherland, Proc. 3rd Int. Meet. Forensic Immunol. Med. Pathol. and M.R., 2001. Teeth, brains, and primate life histories. Am. J. Toxicol. Lond., Excerpta medica Int. Cong. Ser. 80, 36. phys. Anthrop. 114, 192–214. Harrison, T., 2002. Late Oligocene to middle Miocene Bromage, T.G., Dean, M.C., 1985. Re-evaluation of the age at catarrhines from Afro-Arabia, in: Hartwig, W.C. (Ed.), The death of immature fossil hominids. Nature 317, 525–527. Primate Fossil Record. Cambridge University Press, Caccone, A., Powell, J.R., 1989. DNA divergence among Cambridge, pp. 311–338. hominoids. Evolution 43, 925–942. Harvey, P.H., Clutton-Brock, T.H., 1985. Life history variation Charnov, E.L., Berrigan, D., 1993. Why do female primates in primates. Evolution 39, 559–581. have such long lifespans and so few babies? Or life in the Harvey, P.H., Read, A.F., Promislow, D.E.L., 1989a. Life slow lane. Evol. Anthrop. 1, 191–194. history variation in placental mammals: Unifying the data Dean, M.C., 1987a. The dental development status of six East with the theory. Oxford Surv. Evol. Biol. 6, 13–31. African juvenile fossil hominids. J. hum. Evol. 16, 197–213. Harvey, P.H., Promislow, D.E.L., Read, A.F., 1989b. Causes Dean, M.C., 1987b. Growth layers and incremental markings and correlates of life history differences among mammals, in hard tissues; a review of the literature and some prelimi- in: Foley, R., Standen, V. (Eds.), Comparative Socio- nary observations about enamel structure in Paranthropus ecology: The Behavioural Ecology of Humans and Other boisei. J. hum. Evol. 16, 157–172. Mammals. Blackwell, Oxford, pp. 305–318. Dean, M.C., 1989. The developing dentition and tooth Hurme, V.O., van Wagenen, G., 1961. Basic data on the structure in hominoids. Folia Primatol. 53, 160–177. emergence of permanent teeth in the rhesus monkey Dean, M.C., 1998. A comparative study of cross striation (Macaca mulatta). Proc. Am. Phil. Soc. 105, 105–140. spacings in cuspal enamel and of four methods of estimating Iwamoto, M., Watanabe, T., Hamada, Y., 1987. Eruption of the time taken to grow molar cuspal enamel in Pan, Pongo permanent teeth in Japanese monkeys (Macaca fuscata). and Homo. J. hum. Evol. 35, 449–462. Primate Res. 3, 18–28. Dean, M.C., Beynon, A.D., 1991. Histological reconstruction Kahumbu, P., Eley, R.M., 1991. Teeth emergence in wild live of crown formation times and initial root formation times baboons in Kenya and formulation of a dental schedule for in a modern human child. Am. J. phys. Anthrop. 86, aging wild baboon populations. Am. J. Primat. 23, 1–9. 215–228. Kappelman, J., Richmond, B., Seiffert, E.R., Maga, M. & Dean, M.C., Beynon, A.D., Reid, D.J., Whittaker, D.K., Ryan, T., In press. Fossil apes from the Sinap Formation, 1993a. A longitudinal study of tooth growth in a single in: Fortelius, M., Kappelman, J., Sen, S. Bernor, R.L. individual based on long- and short-period incremental (Eds.), Geology and Paleontology of the Miocene Sinap markings in dentine and enamel. Int. J. Osteoarchaeol. 3, Formation in Central Turkey. Columbia University Press, 249–264. New York. Dean, M.C., Beynon, A.D., Thackeray, J.F., Macho, G.A., Kelley, J., 1988. A new large species of Sivapithecus from the 1993b. Histological reconstruction of dental development Siwaliks of Pakistan. J. hum. Evol. 17, 305–324. and age at death of a juvenile Paranthropus robustus speci- Kelley, J., 1997. Paleobiological and phylogenetic significance men, SK 63, from Swartkrans, South Africa. Am. J. phys. of life history in Miocene hominoids, in: Begun, D.R., Anthrop. 91, 401–419. Ward, C.V., Rose, M.D. (Eds.), Function, Phylogeny, and 328 J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329

Fossils: Miocene Hominoid Evolution and Adaptations. Schwartz, G.T, Liu, W., Zheng, L., in press. Preliminary Plenum Press, New York, pp. 173–208. investigation of dental microstructure in the Yuanmou Kelley, J., 1999. Age at first molar emergence in Afropithecus hominoid, Yunnan Province, China. J. hum Evol. turkanensis. Am. J. phys. Anthrop. 28(Suppl.), 167. Schwartz, G.T., Reid, D.J., Dean, C., 2001. Developmental Kelley, J., 2002. Life-history evolution in Miocene and extant aspects of sexual dimorphism in hominoid canines. Int. J. apes, in: Minugh-Purvis, N., McNamara, K.J. (Eds.), Primatol. 22, 837–860. Human Evolution Through Developmental Change. Johns Schwartz, G.T., Samonds, K.E., Godfrey, L.R., Jungers, W.L., Hopkins University Press, Baltimore, pp. 223–248. Simons, E.L., 2002. Dental microstructure and life history Kelley, J., Dean, M.C., Reid, D.J., 2001. Molar growth in the in subfossil Malagasy lemurs. Proc. natl. Acad. Sci. 99, late Miocene hominoid, Dryopithecus laietanus, in: Brook, 6124–6129. A. (Ed.), Dental Morphology 2001. Sheffield Academic Shellis, R.P., 1998. Utilization of periodic markings in enamel Press, Sheffield, pp. 123–134. to obtain information on tooth growth. J. hum. Evol. 35, Kumar, S., Hedges, B., 1998. A molecular timescale for 387–400. vertebrate evolution. Nature 392, 917–920. Smith, B.H., 1989. Dental development as a measure of life Leakey, M., Walker, A., 1997. Afropithecus function and history in primates. Evolution 43, 683–688. phylogeny, in: Begun, D.R., Ward, C.V., Rose, M.D. Smith, B.H., 1991. Dental development and the evolution of life (Eds.), Function, Phylogeny, and Fossils: Miocene history in Hominidae. Am. J. phys. Anthrop. 86, 157–174. Hominoid Evolution and Adaptations. Plenum Press, New Smith, B.H., 1992. Life history and the evolution of human York, pp. 225–239. maturation. Evol. Anthrop. 1, 134–142. Leakey, R.E., Leakey, M.G., Walker, A.C., 1988. Morphology Smith, B.H., Tompkins, R.L., 1995. Toward a life history of the of Afropithecus turkanensis from Kenya. Am. J. phys. Hominidae. Ann. Rev. Anthrop. 24, 257–279. Anthrop. 76, 289–307. Smith, B.H., Crummet, T.L., Brandt, K.L., 1994. Age of Lee, P.C., Foley, R.A., 1993. Ecological energetics and eruption of primate teeth: A compendium for aging indi- extinction of giant gelada baboons, in: Jablonski, N.G. viduals and comparing life histories. Yrbk. phys. Anthrop. (Ed.), Theropithecus: The Rise and Fall of a Primate Genus. 37, 177–231. Cambridge University Press, Cambridge, pp. 487–498. Smith, R.J., 1996. Biology and body size in human evolution: Macho, G.A., 2001. Primate molar crown formation times and Statistical inference misapplied. Curr. Anthrop. 37, life history evolution revisited. Am. J. Primatol. 55, 451–460. 189–201. Smith, R.J., Jungers, W.L., 1997. Body mass in comparative Macho, G.A., Wood, B.A., 1995. The role of time and timing in primatology. J. hum. Evol. 32, 523–559. hominid dental evolution. Evol. Anthrop. 4, 17–31. Smith, R.J., Gannon, P.J., Smith, B.H., 1995. Ontogeny of Martin, R.D., MacLarnon, A.M., 1990. Reproductive patterns australopithecines and early Homo: evidence from cranial in primates and other mammals: The dichotomy between capacity and dental eruption. J. hum. Evol. 2, 155–168. altricial and precocial offspring, in: DeRousseau, C.J. (Ed.), Smith, T.M., Martin, L.B., de Bonis, L., Koufos, G.D., 2001. Primate Life History and Evolution. Wiley-Liss, New York, Dental development in Graecopithecus freybergi (Ourano- pp. 47–79. pithecus macedoniensis). Am. J. phys. Anthrop. 32(Suppl.), Phillips-Conroy, J.E., Jolly, C.J., 1988. Dental eruption 140. schedules of wild and captive baboons. Am. J. Primatol. 15, Smith, T.M., Martin, L.B., Leakey, M.G., 2003. Enamel thick- 17–29. ness, microstructure and development in Afropithecus Ramirez Rozzi, F., 2002. Enamel microstructure in hominids: turkanensis. J. hum. Evol. 44, 283–307. New characteristics for a new paradigm, in: Minugh-Purvis, Swindler, D.R., 1985. Nonhuman primate dental development N., McNamara, K.J. (Eds.), Human Evolution Through and its relationship to human dental development, in: Developmental Change. Johns Hopkins University Press, Watts, E.S. (Ed.), Nonhuman Primate Models for Human Baltimore, pp. 319–348. Growth and Development. Alan R. Liss, New York, pp. Read, A.F., Harvey, P.H., 1989. Life history differences among 67–94. the eutherian radiations. J. Zool. Lond. 219, 329–353. Swindler, D.A., Beynon, A.D., 1993. The development and Reid, D.J., Beynon, A.D., Ramirez Rozzi, F.V., 1998a. microstructure of the dentition of Theropithecus, in: Histological reconstruction of dental development in four Jablonski, N.G. (Ed.), Theropithecus: The Rise and Fall of individuals from a medieval site in Picardie, France. J. hum. a Primate Genus. Cambridge University Press, Cambridge, Evol. 35, 463–477. pp. 351–381. Reid, D.J., Schwartz, G.T., Dean, C., Chandrasekera, M.S., Watts, E.S., 1990. Evolutionary trends in primate growth and 1998b. A histological reconstruction of dental development development, in: DeRousseau, C.J. (Ed.), Primate Life in the common chimpanzee, Pan troglodytes. J. hum. Evol. History and Evolution. Wiley-Liss, New York, pp. 89–104. 35, 427–448. Wolf, K., 1984 Reproductive competition among co-resident Ross, C.R., 1998. Primate life histories. Evol. Anthrop. 6, male silvered leaf monkeys (Presbytis cristata). Ph.D. 54–63. Dissertation,Yale University. J. Kelley, T.M. Smith / Journal of Human Evolution 44 (2003) 307–329 329

Zhao, L., Lu, Q., Xu, Q., 2000. Enamel microstructure of Lufengpithecus lufengensis. Acta Anthrop. Sinica 19(Suppl.), 72–82. Zuckerman, S., 1928. Age-changes in the chimpanzee, with special reference to growth of brain, eruption of teeth, and estimation of age; with a note on the Taungs ape. Proc. Zool. Soc., Lond. 1928, 1–42.