A biomechanical perspective on molar emergence and primate life history.

The strong relationship between M1 emergence age and life history across primates provides a means of reconstructing fossil life history. The underlying process that leads to varying molar emergence schedules, however, remains elusive. Using three-dimensional data to quantify masticatory form in ontogenetic samples representing 21 primate species, we test the hypothesis that the location and timing of molar emergence are constrained to avoid potentially dangerous distractive forces at the temporomandibular joint (TMJ) throughout growth. We show that (i) molars emerge in a predictable position to safeguard the TMJ, (ii) the rate and duration of jaw growth determine the timing of molar emergence, and (iii) the rate and cessation age of jaw growth is related to life history. Thus, orofacial development is constrained by biomechanics throughout ontogeny. This integrative perspective on primate skull growth is consistent with a long sought-after causal explanation underlying the correlation between molar emergence and life history.

INTRODUCTION

The evolutionary timeline and adaptive basis of modern humans’ derived life history are among the most enduring mysteries in human origins research. As the most highly encephalized of primates, modern humans generally follow a “slow” life history profile, taking a long time to reach adult body and brain size and to initiate the reproductive portion of the lifespan. In small-scale, natural-fertility societies, human offspring are weaned at an early age, far younger than predicted for a primate of our body/brain size. This early weaning age allows subsequent reproductive events to occur in rapid succession, reducing birth spacing (interbirth interval, IBI) and increasing lifetime reproductive output; in general, these are attributes of species with a “fast” life history profile. Modern human life history is therefore neither “fast” nor “slow” per se but is, instead, a unique blend of both, which serves to increase fertility rates that are energetically subsidized through complex webs of intra- and intergenerational transfers of resources, alloparental care, and social learning involving kin and nonkin (–).

Although unique, aspects of human life history are tightly integrated with other components of organismal biology. For instance, across Mammalia, the pace of life history is intimately linked to the pattern and pace of dental development (–). The pace of dental development in primates, especially the age at which the permanent first molar (M1) emerges into the oral cavity, is strongly correlated with a suite of life history attributes such as weaning age, age at first reproduction, IBI, and mortality rates (–). Molar emergence is generally considered to be the best skeletal indicator of life history and is therefore used to infer life history attributes in fossil taxa (–). Shifts in molar emergence ages among extinct hominins provide key details about the evolution of human growth and thus of our life history. Fossil evidence currently suggests that our species’ unique life history profile did not appear until the Middle to Late Pleistocene at the earliest (, , –).

Although used for decades to track evolutionary shifts in primate (particularly hominin) life history, the specific developmental mechanism underlying the relationship between molar emergence age and life history remains unknown, apart from the broad suggestion that dental development is integrated into overall cranial and somatic growth as a whole (). Uncovering this long sought-after mechanism will shed light on why the age at M1 emergence is so deeply entangled with fundamental aspects of a species’ life history. Here, we provide a model of primate skull growth that explores the relationships among life history, dental development, and the biomechanics of the developing masticatory system. This model proposes a specific biomechanical constraint operating throughout craniofacial growth that dictates the spatial location and timing of molar emergence, defined here as when a molar’s occlusal surface approaches or reaches the occlusal plane. We then address whether the acquisition of delayed dental maturation in modern Homo sapiens is linked to an altered pattern of cranial-wide coordinated growth coupled with a delayed somatic growth profile.

As predicted by fundamental biomechanical principles and confirmed by comparative and experimental analyses, the adult masticatory system of mammals is configured to avoid distraction of the temporomandibular joint (TMJ) (i.e., when the mandibular condyle is pulled away from the articular eminence) during biting and chewing (–). These principles, referred to collectively as the constrained lever model (CLM), assign great importance to the position molars occupy along the maxillary and mandibular dental arches in relation to the point in biomechanical space coinciding with the resultant vector of the masticatory adductor (i.e., jaw closing) muscle forces (i.e., the adductor muscle resultant vector, AMRV). In adult primates, the position of the distalmost molar, the permanent third molar (M3), is constrained to be located anterior to the AMRV because biting posterior to this point produces distractive forces at the working (i.e., biting) side TMJ (see figs. S1 and S2 and the Supplementary Materials for a full description of the CLM) (, ). Because this constraint on molar position has been substantiated for adults across multiple primate taxa, we hypothesize that the availability of space created anterior to the AMRV throughout ontogeny will constrain where and when each successive molar emerges (, –).

Here, we provide the first systematic assessment of whether, as a result of coordinated growth of the entire masticatory apparatus, specific mechanical constraints regulate the scheduling of molar emergence across primates. The integrative model deployed here (Figs. 1 and 2 and fig. S1) provides a mechanical and developmental context for explaining spatial and temporal variation in molar emergence ages. It predicts that the location and age at which a molar emerges results directly from the rate and duration of growth in the dental arches in relation to the three-dimensional (3D) configuration of the masticatory system. In turn, the rate and duration of dental arch growth (for simplicity, we focus on mandibular arcade growth) should be tightly integrated into the overall package of somatic and neural growth including key aspects of organismal life history. We use a taxonomically, morphologically, and adaptively heterogeneous primate sample (n = 21 species; table S1) to test whether: (i) where molars emerge is constrained such that each successive molar always emerges anterior to the position of the AMRV throughout ontogeny, (ii) the rate at which the “mechanically viable” space (i.e., within the growing alveolar process anterior to the AMRV) is made available throughout ontogeny determines when molars emerge, and (iii) the rate at which mechanically viable space appears is highly correlated with life history parameters that track overall growth, maturation, and energetic trade-offs.

Fig. 1.

Macaque skull indicating study landmarks and masticatory muscle mechanics.

(A) Lateral view of skull (left), basal view of cranium (center), and occlusal view of mandible (right) illustrating the 38 landmarks (see table S8) used to capture the position of the three principal masticatory adductor muscles and overall masticatory configuration. (B) Lateral view of the position and orientation of muscle lines of action (MLAs) for the masseter (left), temporalis (center), and medial pterygoid (right) muscles, as well as the point (blue colored circle) where each MLA intersects the TOS plane (dashed line). Checkered circles indicate landmarks that are out of view. See text and the Supplementary Materials for details. Photo credit: H. Glowacka, University of Arizona.

Fig. 2.

Macaque skull indicating overall masticatory configuration and biomechanical variables.

(A) Intersection points of each MLA from Fig. 1 were projected onto the occlusal plane, and their average position was used to represent the point at which the AMRV (red arrow) crosses the TOS plane, projected onto the occlusal plane (red square). (B) Occlusal view of mandible illustrating (left half of image): the position of the AMRV at the level of the occlusal plane (red square), extended laterally by the red dashed line, and the main masticatory regions where the CLM predicts bite points could be (regions I and II) and should not be (region III) located; (right half of image) the measurement resultant molar taken from the position of the AMRV to the distalmost border of each successively emerging mandibular molar (dp4, M1-3) and MAL (gray dashed arrow), comprising the summed linear distance from the AMRV, through the last emerged molar and each interproximal space along the entire mandibular arch, terminating anteriorly at infraoral (point no. 29, the midline point at apex of the septum between the mandibular central incisors; see Fig. 1A and table S8). (C) Occlusal view of juvenile mandible mapping out the regions of the CLM and the location of the measurements, resultant molar, and MAL [colored regions and arrows indicate the same as in (B)]. Checkered circles in (B) and (C) indicate landmarks that are out of view. See text and the Supplementary Materials for details. Photo credit: H. Glowacka, University of Arizona.

Our model incorporates a suite of 3D landmark data to quantify the configuration of the masticatory system, including the lines of action of the three main mandibular adductor muscles: the temporalis, masseter (deep and superficial components combined), and medial pterygoid (Fig. 1B). The position of the AMRV is a function of the magnitudes, orientations, and positions of these three individual muscle force vectors. Muscle force magnitudes, often estimated from physiological cross-sectional area (PCSA), cannot be calculated with confidence, as few to no data are available for how PCSA changes throughout ontogeny for most primate taxa. The AMRV was therefore estimated here using only the vector orientations and positions and was calculated as the average position at which the three jaw adductor muscles’ lines of action (MLAs) cross what is termed the triangle of support (TOS) plane (defined by the most posteriorly available bite point and the two TMJs; dashed black line; Figs. 1B and 2A). The MLA points of intersection with the TOS plane were transposed to the occlusal plane, and their average location was used to represent the location of the AMRV (red box positioned on dashed line, Fig. 2), thereby defining key regions generally associated with the configuration of the masticatory system (Fig. 2B; see Materials and Methods and the Supplementary Materials). Regions I and II are mechanically viable spaces within which teeth are expected to be positioned; bites on teeth positioned in region III would cause distractive forces to accrue at the TMJ. The spatial location of the AMRV allowed us to quantify the distance between this point and the position of each emerging molar as it reaches the occlusal plane (termed “resultant molar”; heavy double arrows in Fig. 2, B and C) and thus to test whether or not there is a constraint on where molars emerge throughout ontogeny.

The rate at which space is made available for emerging molars is a function of the rate at which preceding molars advance anteriorly throughout orofacial growth. To quantify this, a subset of species (n = 5; table S2) comprising individuals of known age was used to calculate growth-related changes in mandibular arch length (MAL), where total MAL was measured as the cumulative set of segmented lengths along the alveolar process from the AMRV to the anteriormost tooth in the dental arch, projected onto the occlusal plane (Fig. 2B). The overall rate (slope) and duration (age at cessation) of growth are calculated from species-specific MAL growth curves that are truncated at the calculated end point of the growth phase. Tests for differences in slopes were matched against pairwise expectations of mandibular growth trajectories for taxa that differ in adult MALs and/or molar emergence ages (Table 1 and tables S3 and S4). The simplest scenario suggests that species with faster growing mandibular arches should exhibit younger ages at molar emergence, although multiple hypothetical scenarios relating the relative timing of molar emergence to variation in adult MAL are possible (see Materials and Methods and fig. S3). As most scenarios accounting for delayed molar emergence involve a later age of MAL growth cessation (fig. S3, A to D), we compare MAL growth curves for our included species to determine whether the rate at which mechanically viable space appears throughout growth modulates the timing of molar emergence.

Table 1.

Primate molar emergence and MAL growth cessation ages.

Species	M1 emergence age (years)	M2 emergence age (years)	M3 emergence age (years)	MAL growth cessation age (years)
M. mulatta*	1.32–1.37	3.15	5.4–5.81	6.74
P. cynocephalus*	1.58–1.75	3.75–3.83	6.17–6.25	7.50
G. beringei †	3.2–3.6		10.7	17.00
P. troglodytes ‡	2.5–3.3	5.9	10.1–11.0	12.46
P. troglodytes §	3.6–4.6
H. sapiens*	6.15–6.33	11.49–12.0	19.8–20.4	22.00
H. sapiens ||	5.9 (±0.9)	11.3 (±1.2)	17.8 (±2.6)	22.00

*Smith et al. (), sexes combined; mandibular molars.

†Emergence ages are derived from an assessment of radiographs provided in the supplemental atlas of Kralick et al. ().

‡Machanda et al. () and Smith et al. (), sexes combined and mandibular gingival emergence for free-living Kanyawara chimpanzees.

§Kelley et al. (), gingival emergence, unknown sex, and mandibular molars for free-living Liberian chimpanzees.

||Mean (±1 SD) from Liversidge () and partial eruption (midpoint between alveolar and occlusal planes).

The extent to which craniofacial growth, masticatory configuration, and the scheduling of molar emergence are interconnected with overall growth and life history was evaluated using a phylogenetic principal components analysis (pPCA) (), performed on the same reduced species dataset and incorporating the following life history and life history–related variables: brain size, age at first reproduction, IBI, and gestation length. All multidimensional PCs with eigenvalues >1 were used as predictor variables in phylogenetic generalized least squares (PGLS) analyses to determine how well aspects of masticatory system growth (i.e., MAL growth rate and age at MAL growth cessation) predict these key components of life history related to reproductive energetics.

RESULTS

For the full species dataset, recently emerged molars were consistently positioned significantly anterior to the adductor muscle resultant throughout the entire period of craniofacial growth (Fig. 3 and table S5). Of the 1258 individual specimens analyzed here, only three (<0.25%; one each of Sapajus apella, Gorilla beringei, and H. sapiens) violated the model’s expectations but only for the position of their recently emerged M3s. Overall, our results provide support for a strong mechanical constraint on where molars emerge into occlusion that protects the TMJ from compromised function. This constraint exists for both haplorhine and strepsirrhine taxa, despite the latter’s unfused mandibular symphyses and low ramal heights (see fig. S2).

Fig. 3.

Position of molar emergence throughout growth.

Boxplots of resultant molar for dp4 (top) to M3 (bottom) dental emergence categories for all included primate species. Red dashed line represents the position at which the AMRV crosses the TOS plane, projected onto the occlusal plane (see Fig. 2). Positive resultant molar values indicate that the distalmost molar (i.e., the most recently emerged molar) is positioned anterior to the AMRV.

Molars do not emerge at the AMRV but rather anterior to it. This is consistent with the presence of a mechanical buffer zone at the periphery of the TOS (see the Supplementary materials and fig. S1D). This zone provides a safety mechanism ensuring that the AMRV does not fall outside of the TOS’s boundaries as balancing-side/working-side muscle activities fluctuate throughout the masticatory cycle, especially at more posterior bite points (see fig. S1) (). The size of the buffer zone varies throughout ontogeny and across species and is a function of skull size, jaw gape, and food material properties, among other factors (). The rate at which “mechanically viable” space appears along the dental arcade matches expectations from molar emergence schedules. Analysis of covariance (ANCOVA) results indicated that G. beringei, Pan troglodytes, Macaca mulatta, and Papio cynocephalus all have significantly steeper slopes and thus faster rates of growth in MAL than does H. sapiens (Fig. 4 and table S6). Mandibles of G. beringei do not grow faster than those of P. troglodytes, but both African apes have significantly slower mandibular growth rates when compared to M. mulatta. These observations match expectations from molar emergence schedules, which, for this comparative sample, are the most accelerated in macaques and the most delayed in modern humans.

Fig. 4.

Growth of MAL.

MAL growth for ontogenetic series of known-age individuals representing a subset of five primate species. Breakpoints (indicated by vertical dashed lines) indicate ages at growth cessation and are determined by segmented regression (see text for details).

MAL growth cessation age estimates (i.e., “break points” in Fig. 4 and Table 1) were oldest in H. sapiens (22.00 years), followed by G. beringei (17.00 years), P. troglodytes (12.46 years), and P. cynocephalus (7.50 years) and the youngest in M. mulatta (6.74 years). These trends also matched expectations from molar emergence schedules, except for the G. beringei–P. troglodytes comparison, where our model predicts G. beringei MALs to grow at a faster rate and similar duration.

The shifting configuration of the masticatory system as a modulating factor on molar emergence ages is related to key aspects of primate life history. There was a significant positive relationship between PC1 and MAL growth rate and a significant negative relationship between PC1 and MAL growth cessation (Fig. 5 and table S7), with PC1 explaining 78% of the variation in both variables (table S7). These results are consistent with chewing biomechanics being the causal mechanism underlying the strong relationship between molar emergence and life history.

Fig. 5.

Relationship between life history and MAL growth rate/cessation.

PGLS results for the relationships between PC1 and MAL growth rate (left) and MAL growth cessation (right). PC1 scores derived from a PCA of five primate species and includes brain size and the following life history variables: age at first reproduction, IBI, and gestation length.

DISCUSSION

Our results support the idea that the biomechanical constraint on masticatory form in adult primates operates throughout the duration of craniofacial growth. This constraint regulates where molars can emerge safely into functional occlusion, which, when viewed within the mechanical context of overall orofacial growth, modulates the timing of when molars emerge. This integrative perspective on primate skull growth is consistent with a long sought-after causal explanation underlying the correlation between molar emergence and life history.

The distinctive modern human facial profile includes a retracted face that is tucked under an expanded and globular neurocranium [e.g., ()], which likely resulted from relaxed selection for robust faces or from selection for high-magnitude bite forces; see () for a discussion of competing hypotheses on facial reduction in the genus Homo. Modern human craniofacial growth is characterized by prolonged periods of dental development with delayed molar emergence schedules. The earliest known co-occurrence of the unique modern human cranial form and development is from the Middle Pleistocene of northern Africa, at the site of Jebel Irhoud (Morocco), dating to ca. 315 ka (). Despite their relatively primitive neuro- and endocranial anatomy, the Irhoud hominins are notable for their possession of a vertically short face that is retracted beneath the braincase. The tooth formation trajectories in a juvenile Irhoud hominin reveal prolonged dental development and delayed ages at molar emergence (). Our findings suggest that delayed molar emergence in H. sapiens is the result of extreme facial retraction coupled with a deceleration in, and extended duration of, orofacial growth. The facial anatomy and dental developmental schedule shared by the Irhoud hominins and extant humans suggests that the unique pattern of facial growth is deeply rooted in the evolutionary history of modern H. sapiens.

There was one notable exception to the predictions of our model: the difference between mountain gorillas and chimpanzees. On the basis of inferred molar emergence ages and adult MAL, we predicted that mountain gorilla mandibular arches would grow at a faster rate and for a similar duration compared to those of chimpanzees (table S4B). Mountain gorilla mandibular growth rates are instead statistically indistinguishable from those of chimpanzees, and mountain gorillas grow their mandibular arches for far longer. This result is more consistent with mountain gorillas having later ages at molar emergence than chimpanzees (fig. S3A). The mismatch is likely due to an incomplete understanding of molar emergence ages in mountain gorillas, differences in growth trajectories as a correlated outcome of different dietary ecologies, or a combination of the two. A recent radiographic study on molar development in the same Virunga mountain gorilla population found that the chronology of molar formation is not accelerated compared to chimpanzees (). The chronologies of molar development and molar emergence are two separate, although interrelated phenomena, and spatial availability does not necessarily constrain the former (), so only data on the timing of molar emergence can help unpack the complex relationships among extreme folivory, somatic development, socioecology, and orofacial growth rates in mountain gorillas (see Supplementary Text). An important caveat is that the gorilla and baboon samples are limited around the estimated breakpoint. As larger skeletal samples become available, we expect that the error around the breakpoint estimates will decrease.

Our findings from an analysis of primates reveals that it is the combination of orthognathic faces with protracted jaw growth that results in a delayed appearance of alveolar space in which molars can safely emerge. Primates, however, are not the only mammals for which these linked phenotypes exist. Recent work reveals a tight correlation between the same developmental parameter (later molar emergence ages) and the same orofacial anatomy (foreshortened jaw length) in ungulates (). Future work should focus on revealing the gene regulatory networks that coordinate cranial and dental morphogenesis [e.g., ()].

In her seminal study, Smith () extended earlier foundational work on primate growth biology (, –) by exploring the linkages between dental development and life history across primates. Smith opined that “tooth development must be completely integrated into the plan of growth and development, timed to growth of the skull, maturation of muscles of mastication, and somatic growth in general” [(), p. 683]. Despite these insights, comparative biologists have come no closer to elucidating the precise mechanism maintaining the proposed coordination between dental and craniofacial growth. Our study indicates that primate life history is related to ages at molar emergence through its influence on the growth rate and duration of a kinetically stable masticatory apparatus and so provides a powerful lens through which the long-known linkages among dental development, skull growth, and maturational profiles can be viewed.

MATERIALS AND METHODS

Sample

Data were collected from cross-sectional ontogenetic samples of primate skulls (n = 1258 specimens), representing 21 species across primates (see table S1 for species list and sample sizes). To determine mandibular arch growth rates and whether these are correlated with life history, data were collected from cross-sectional ontogenetic samples of primate skulls representing five species (H. sapiens, G. beringei, P. troglodytes, M. mulatta, and P. cynocephalus; see table S2 for specimen list and the Supplementary Materials for sample provenance). This second dataset consisted of individuals with known or estimated ages at death.

Data collection and statistical analyses

Following previously published methods (, ), data were collected on the spatial configuration of the primate masticatory system throughout ontogeny by digitizing 38 homologous landmarks (Fig. 1A and tables S8 and S9) characterizing overall skull size and masticatory configuration, including the position of teeth and the origins and insertions of the masticatory muscles. We used landmarks to estimate the centroids of attachment sites (origins and insertions) of the jaw adductor muscles. The 3D spatial position of each MLA was derived geometrically as vectors running between each muscle’s origin and insertion. The specific landmarks for each adductor muscle’s origin and insertion, respectively, are as follows: masseter (2 to 3; 24), anterior temporalis (4 to 5; 23), and medial pterygoid (6; 25) for the left side of the skull only. Full details are described in the Supplementary Materials (see table S8 for landmark list and definitions and the “Masticatory system measurements” section). Landmark data were obtained using a Microscribe G3X digitizer (Immersion Corp., San Jose, CA). Details of data collection and landmarks can be found in the Supplementary Materials. The coordinate data were used to calculate two key variables: (i) resultant molar, the distance from the adductor muscle resultant to the last molar (described in further detail in the Supplementary Materials); and (ii) MAL, the distance, along the dental arcade, from the point between the two mandibular central incisors (infradentale) to the adductor muscle resultant. All landmark data were analyzed using customized code written in R 3.6.1 () by H.G.

Spatial position of molars at emergence

A molar was scored as “emerged” if its occlusal surface was in the occlusal plane. For each species, specimens were divided into four dental emergence categories: dp4 emerged, M1 emerged, M2 emerged, and M3 emerged. Resultant molar was compared to zero, which represents the position of the muscle resultant, using one-sided two sample t tests to determine whether molars emerge significantly anterior to the point at which the muscle resultant intersects the TOS.

Growth rate and cessation

To determine whether the rate at which space is made available in the jaw coupled with the duration of jaw growth determine the timing of molar emergence, pairwise comparisons of MAL growth trajectories were performed between taxa that differ in adult MAL and/or molar emergence timing. The simplest scenario suggests that species with faster growing mandibular arches should exhibit younger ages at molar emergence. Species differences in adult MAL and arch growth duration, however, suggest that a species can grow mandibular arches faster than another species, but because its target MAL is longer, result in later ages at molar emergence. Multiple scenarios relating the relative timing of molar emergence to variation in adult MAL are possible (fig. S3). To summarize: Earlier molar emergence occurs either in a shorter mandibular arch that grows at the same, faster, or slower rate but for a shorter duration (fig. S3, A to C) or in a longer mandibular arch that grows at a faster rate and for a longer duration (fig. S3D). Later molar emergence occurs either in a longer mandibular arch that grows at the same, faster, or slower rate and for a longer duration (fig. S3, A to C) or in a shorter mandible that grows at a slower rate and for a longer duration (fig. S3D). Similar ages at molar emergence occur in a longer mandibular arch that grows at a faster rate and for the same duration (fig. S3E), in a shorter mandibular arch that grows at a slower rate and for the same duration (fig. S3E), or in mandibular arches of similar length that grow at similar rates and for similar durations (in which case, the trajectories in fig. S3 would be identical). Table 1 lists reported or estimated molar emergence ages in the literature for the five species examined in this study. On the basis of these data and adult MAL data (table S3), predictions were made for pairwise comparisons of MAL growth rate and growth duration (table S4).

MAL growth rate was determined using segmented (i.e., piecewise) regressions, which described species-specific MAL growth curves (), with age as the independent variable and MAL as the dependent variable. Segmented regressions were fit to the data using the segmented package in R (). Identified breakpoints were extracted as the variable MAL growth cessation and were used to represent the end point of growth and thus growth duration, for each species. After breakpoints were identified, datapoints from the growth phase of the curve were extracted. These data were then used in an ordinary least squares regression with age as the predictor variable and MAL as the response variable, and the resulting slopes (extracted as the variable MAL growth rate) were compared between sets of taxa using an ANCOVA to determine whether the interaction between species and age is significantly different from zero; in other words, if the slopes of the two species are significantly different from each other.

Life history

The final set of analyses was aimed at determining whether brain size and life history variables are significantly correlated to MAL growth rate and MAL growth cessation. Brain size and life history data (i.e., age at first reproduction, IBI, and gestation length) were collected from the literature (see table S10). Wherever possible, the data collected were from the same populations as the skeletal collections included in this study. Because this study considered MAL growth rate and MAL growth cessation for only five species, the number of predictor variables had to be reduced. Dimensionality reduction was performed through a pPCA (), implemented using the Phytools package for R (). Resulting PCs with eigenvalues >1 were then used as predictor variables in several PGLS analyses. Separate PGLS analyses were run with (i) MAL growth rate and (ii) MAL growth cessation as response variables and the PCs as predictor variables. A phylogenetic tree for the analysis was downloaded from 10kTrees (). Data were log10-transformed if they did not meet the assumption of normality.