Primate tooth crown nomenclature revisited

Materials

The study sample is shown in Table 1. The sample consists of mandibular first and second molars from 420 specimens, representing 71 primate species (a complete list of specimens is provided in the Supporting Information). The study sample was selected to include species from all major clades. Sex was not considered as there is no evidence that it impacts the presence of cusps. Sample sizes for some species are low due to difficulties in identifying and microCT scanning specimens with unworn mandibular molars (this is particularly challenging as most primate species are relatively thin-enamelled). Specimens with low to moderate attrition were included as it did not impact our ability to identify particular crown features with confidence. Due to these small sample sizes the frequency of traits is not assessed.

Methods

Specimens were scanned on a number of different microCT systems including beamline ID 19 at the European Synchrotron Radiation Facility (ESRF, Grenoble, France), or a BIR ACTIS 225/300 or SkyScan 1,172 microtomographic scanner at the Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology. Scanning was conducted under standard operating conditions following established protocols (Olejniczak et al., 2007). Scan resolution varied between 10–60 microns. This resolution is sufficient to identify small crown features, although it is acknowledged that very tiny dentine horns may be poorly resolved making their assessment difficult. Individual molars were initially cropped in Avizo 9.0 (http://www.thermofisher.com). To facilitate tissue segmentation, image stacks of each molar were then filtered using a 3D mean-of-least-variance filter with a kernel size of one. This process sharpens the boundaries between enamel and dentine (Schulze & Pearce, 1994), allowing for a clearer segmentation of tissue types. Filtering was implemented using MIA open-source software (Wollny et al., 2013). Enamel and dentine were segmented in Avizo 6.3 using a seed growing watershed algorithm employed via a custom Avizo plugin, before being manually checked. After segmentation, triangle-based surface models of the EDJ were produced using the generate surface module in Avizo, and then saved in polygon file format (.ply).

For the purpose of evaluating current nomenclatures and creating new nomenclatures, species are grouped at the highest taxonomic level possible where similarities in mandibular crown morphology allow. In the results section for each group, we first create a ‘current variation’ schematic based on a review of the published literature that acknowledges crown features that have previously been observed and discussed. In many cases, this has never been assessed at the taxonomic levels that we identify here as relevant and necessary. We then report on crown feature variation at the EDJ present in the study sample and propose a ‘new’ expected variation schematic for each group. Davies et al. (2021) recently proposed the adoption of conservative terms for accessory cusps on hominin lower molars due to difficulties in defining variations in dentine horn presence on the lingual and distal marginal ridge of the EDJ. Specifically, they adopted the terms distal accessory cusp(s) and lingual accessory cusp(s) instead of cusp 6 and cusp 7, respectively. Our observations of accessory cusp presence on lower molars in this study sample reinforce the utility of the use of such generic terms and here we expand this to both the mesial and buccal marginal ridges of the EDJ, as well as, the cingulum (Fig. 2). As a result, we propose the following terms to classify accessory cusps (AC) on the marginal ridge of the EDJ that runs between the four primary cusps (protoconid, metaconid, entoconid, hypoconid): DM—distal margin; LM—lingual margin; MM—mesial margin; BM—buccal margin. Additionally, we propose the following terms to classify accessory cusps on the cingulum: BC—buccal cingulum; LC—lingual cingulum.

Strepsirrhini

In the family Lemuridae, the two mesial primary cusps are compressed bucco-lingually, are set close together, and are connected by a transverse crest. A distinct crest also traverses down the mesial slope of the protoconid, where it eventually turns and proceeds disto-lingually as a broad ledge to the base of the metaconid. The hypoconid and entoconid are not connected by a crest, and the talonid basin is shallow and circumscribed by the marginal ridge. The entoconid is absent in Varecia variegate (Schwartz & Tattersall, 1985). While a paraconid was present in fossil notharctids of the early Eocene (Gregory, 1920), it is absent in extant Lemuridae. Cuozzo & Yamashita state that lemurids “have lost the paraconid and lack a hypoconulid” (2006, p. 76); however Swindler (2002, p. 69) describes the presence of a “distal heel-like process” in some members of this clade and considers it to be a true hypoconulid. Swindler (2002) also reports the presence of a tuberculum intermedium in all six of his Hapalemur griseus lower first molars, and in five of the six second molars. Similarly, Schwartz & Tattersall (1985) describe a thickening of the postmetacristid in Lemur catta molars that results in the expression of what they term a ‘metastylid’. Some form of buccal cingulid expression is noted in all Lemuridae molars. Thus, the current variation scheme can be summarized with four primary cusps, a hypoconulid on the distal ridge, and a tuberculum intermedium on the lingual ridge (Fig. 3A). Our observations support descriptions of a cusp on the distal marginal ridge of some specimens (Fig. 3D), which we identify and label as a DMAC in the new schematic for Lemuridae (Fig. 3B). Additionally, we also corroborate the reports of accessory cusp presence on the lingual marginal ridge, with LMACs observed in the Prolemur simus and Eulemur fulvus specimens in our sample (Figs. 3E and 3F respectively). The image of the Varecia variegate specimen (Fig. 3G) demonstrates the absence of an entoconid in this taxon, but the expression of several LMACs along the marginal ridge.

The family Lepilemuridae consists of only one extant genus, Lepilemur. The molars have four primary cusps, with a diagonal transverse crest connecting a mesially-placed protoconid to a comparatively more distally-placed metaconid. Swindler (2002) reports that the crest travelling down the mesial slope of the protoconid may connect with the mesio-buccal elevation of the cingulid, forming a mesiostylid where to two crests join. Descriptions also note a distinct and complete buccal cingulid, and a strong cristid obliqua that terminates on the lingual side of the protoconid (Schwartz & Tattersall, 1985). While there has been limited commentary regarding accessory cusps expression in this clade, the potential presence and identification of a fifth cusp has been extensively discussed (James, 1960; Seligsohn & Szalay, 1978; Schwartz & Tattersall, 1985). Unfortunately, these discussions remain focused on hypoconulid expression in lower third molars, and there is little mention of similar manifestations in first or second molars. A further topic of debate is the identification of the cusp positioned distal to the metaconid. Based on comparisons with other strepsirhines like indriids and Copelemur, Schwartz & Tattersall (1985) suggest that the entoconid is absent in Lepilemur and that the cusp distal to the metaconid is a metastylid. Swindler (2002, p. 72) argues that “an obvious developmental groove” that separates the two cusps at the outer enamel surface would suggest that Schwartz & Tattersall’s (1985) metastylid is the entoconid. The current variation schematic for this clade therefore includes a hypoconulid, and a potential entoconid or metastylid (Fig. 4A). Our two Lepilemur specimens do not demonstrate the presence of a mesiostylid or hypoconulid (Figs. 4C–4H). Nevertheless, due to sample size limitations, we cannot rule out their existence in other individuals and therefore include MMAC and DMAC features in the new schematic for Lepilemuridae (Fig. 4B). Figures 4D and 4G reveal lingual crown morphology at the EDJ surface, and in particular, the positioning and appearance of the cusp distal to the metaconid. Unlike LMAC expression in our Lemuridae sample, which are positioned on the distal slope of the metaconid, the disto-lingual cusp in Lepilemur sits further back on the crown and appears developmentally distinct from the metaconid. As such, we label it the entoconid in our new schematic (Fig. 4B). No other features of relevance were identified.

In the family Cheirogaleidae, the protoconid and metaconid are closely positioned, and are connected by a transverse crest that separates a small trigonid from a spacious talonid basin. In Cheirogaleus major, Schwartz & Tattersall (1985, p. 38) remark that the molar cusps “lack virtually all detail and delineation of individual features”, and that structures can therefore only be discussed in a vague sense. Descriptions of hypoconulid expression in cheirogalids are restricted to Microcebus lower third molars (James, 1960; Schwartz & Tattersall, 1985; Cuozzo et al., 2013), while Swindler (2002) reports the presence of a buccal cingulid in all lower molars. Thus, the current variation scheme can be summarized as a relatively simple tooth crown with only four primary cusps (Fig. 5A). Our assessment of Cheirogaleus molars is in agreement with the comments regarding a lack of discernible features by Schwartz & Tattersall (1985). Even from the detail provided by high resolution images of the EDJ, Cheirogaleus molars lack any clearly defined cusps (Figs. 5C and 5D). As such, commentary regarding potential accessory cusp expression in this taxon is avoided. Contrastingly, in our Phaner furcifer specimen, a prominent buccal cingulid accompanies four well-defined primary cusps. Figure 5E demonstrates the presence of a large BCAC on the buccal cingulid, which represents the only addition to the new expected variation schematic for Cheirogaleidae molars (Fig. 5B).

The molars of Indriidae have four well-formed primary cusps. In the lower first molars, the protoconid is mesial to the metaconid, and the mesial marginal ridge is incomplete mesio-lingually. At the terminus of this incomplete ridge, Swindler (2002) mentions the presence of a parastylid. In the lower second molars, the protoconid and metaconid are positioned at similar mesio-distal positions on the crown, the mesial marginal ridge is complete, and the cusps are also connected by a faint transverse crest. Hypoconulid presence is only referenced in relation to lower third molars (Swindler, 2002). Bennejeant (1936) mentions the frequent appearance of a tuberculum intermedium on the distal surface of the metaconid in Avahi, while Schwartz & Tattersall (1985) also report the presence of an equivalent feature at the terminus of a thick postmetacristid. The current variation scheme provided therefore incorporates both parastylids and tuberculum intermediums, in addition to four primary cusps (Fig. 6A). It should be noted that both in the relative positioning of the cusps, and the arrangement of the marginal ridge, these schematic diagrams remain relatively accurate for lower first molars but do not reflect the general shape and patterning of some Indriidae lower second molars. Our observations of Indriidae molar morphology at the EDJ found no evidence for hypoconulid (or DMAC) expression in our sample. Figure 6C demonstrates the incomplete mesial marginal ridge in an Indri indri first molar, while Fig. 6D shows the presence of a small MMAC on the complete marginal ridge of a second molar. No LMACs were observed in this sample, however Fig. 6E demonstrates the thick lingually deflected postmetacristid in Avahi described by Schwartz & Tattersall (1985). Despite not observing a lingual accessory cusp in this sample, the new schematic for Indriidae includes the presence of an LMAC (Fig. 6B), along with the MMAC observed in the Indri lower left second molar.

In the family Galagidae, the first and second molars have four primary cusps, and a well-developed cristid obliqua (Swindler, 2002). As is common within many strepsirrhini clades, a compressed crest descends down the mesial face of the protoconid before angling back toward the metaconid as a broad, horizontal ledge. Discussion regarding hypoconulid presence is limited to lower third molars, and mention of a centrally emplaced heel in Galago senegalensis, and more lingually displaced heel in Galago alleni (Schwartz & Tattersall, 1985). Schwartz & Tattersall (1985) also describe the presence of a protostylid in Euoticus elegantulus, as well as a low conulid on the cristid obliqua at the base of the protoconid in Galago crassicaudatus. The current variation schematic therefore depicts four primary cusps, a protostylid, and an unnamed conulid on the buccal margin (Fig. 7A). From our observations, we demonstrate the presence of multiple DMACs in a Galago senegalensis first molar (Fig. 7C). In the same specimen we also identify an MMAC positioned where the mesial marginal ridge sharply bends towards the metaconid (Fig. 7D). We find no evidence of a protostylid in any Galagidae specimen, however the lingual positioning of the cristid obliqua and buccal flaring of the protoconid would appear to be consistent with morphological conditions commonly associated with protostylid presence. As such, a BCAC is included in the new schematic (Fig. 7B). In relation to the conulid observed by Schwartz & Tattersall (1985) on the cristid obliqua of Galago crassicaudatus, we demonstrate the similar presence of a dentine horn distal to the protoconid in Euoticus elegantulus that we more appropriately label as a BMAC (Fig. 7E).

The family Lorisidae are stated to have four well-developed cusps, a hypoconulid that is restricted to third molars, a prominent cristid obliqua, and a transverse crest that separates the trigonid basin from a spacious talonid basin (Swindler, 2002). In Arctocebus calabarensis, Swindler (2002) reports a paracristid that extends down the protoconid and ends as a mesiostylid (Fig. 8A). While the presence of a cusp at the terminus or turning point of the paracristid is common in strepsirrhini, we find no evidence of a dentine horn in our current Lorisidae sample. Based on previous observations however, we include the presence of a mesio-bucally positioned MMAC in the new schematic (Fig. 8B). Of particular note and relevance in this clade are the observations of distal and disto-lingual accessory cusp expression. Loris tardigradus, Arctocebus calabarensis, Nycticebus coucong and Perodicticus potto specimens all demonstrate single and/or multiple DMAC expression on the distal border of the crown (Figs. 8C–8G). In one of the Perodicticus potto specimens (Fig. 8F), it may be unclear which of the two disto-lingual cusps is the entoconid, and therefore it is unclear whether this molar has a double DMAC configuration, or a single DMAC and LMAC pattern. However, as the entoconid is situated in a strongly lingual position in our other specimens, our new schematic includes a double DMAC pattern in Lorisidae (Fig. 8B).

Tarsiidae

In the family Tarsiidae, the lower molars are tribosphenic, with well-formed paraconid, protoconid, and metaconid cusps, and a broad lower talonid basin with entoconid and hypoconid cusps (Schwartz, 1984). Swindler (2002) describes the presence of a hypoconulid on lower third molars, but no mention of a distal cusp in M1-M2. A cristid obliqua is present, as well as a distinct buccal cingulid in all molars (Swindler, 2002). Thus, the current variation scheme can be summarized as a tooth crown with five primary cusps and no other cusp features (Fig. 9A). In addition to a prominent buccal cingulid that continues along the distal margin of the tooth in some of our sample, we report the presence of multiple accessory dentine horns in Tarsiidae molars. Figure 9A reveals the presence of an LMAC and BMAC in one Tarsius spectrum specimen. Figure 9D similarly demonstrates the presence of a BMAC in Tarsius, but also a BCAC close to the base of the protoconid. Figures 9E and 9F reveal patterns of multiple DMAC expression in a Tarsius spectrum lower first molar, and Tarsius syrichta lower second molar. Based on these observations, the new schematic for Tarsiidae has the addition of several accessory cusp features (Fig. 9B).

Ceboidea

In the subfamily Callitrichinae, the lower molars have four cusps, with the mesial primary cusps connected by a crest that separates a small trigonid from a much larger talonid basin. The only notable mention of additional features in this clade is the presence of buccal cingulids on the first and second molars in most taxa except Callimico (Kinzey, 1973; Swindler, 2002). Thus, the current variation schematic can be summarized as four cusped tooth (Fig. 10A). While our observations match previous comments regarding the common presence of buccal cingulids in Callitrichinae, we extend this description by demonstrating that in some cases, these buccal features may express themselves as small dentine horns along the buccal cingulum. Where present, we identify these structures as BCACs (Fig. 10C). In addition to the buccal features presented, we also reveal the presence of multiple LMACs on the lingual marginal ridge of a Leontopithecus rosalia lower second molar (Fig. 10D). Saguinas and Callithrix specimens in the sample had no discernible crown features of interest (Figs. 10E and 10F). Based on these observations, the new schematic has the addition of LMACs and BCACs (Fig. 10B).

Cebinae lower molars have four cusps, with a prominent crest that separates the taller trigonid from the lower positioned talonid basin. Swindler (2002) states that the hypoconulid is absent in this clade, as is any form of lingual cingulid. Buccal cingulids are however reported as being variably expressed in all molars (Kinzey, 1973; Orlosky, 1973). The current variation schematic is depicted as a simple tooth crown with the four primary cusps (Fig. 11A). Contrary to Swinder’s comment’s however, we identify the presence of what Swindler (2002) would consider a hypoconulid (or more accurately, a DMAC) in a Sapajus apella lower first molar (Fig. 11C). Although there was no discernible lingual cingulid in our sample, the same Sapajus apella specimen also exhibited a small dentine horn on the outer slope of the metaconid. While this could be identified as a LCAC, lingual cingulid cusps were not observed in other specimens and therefore we tentatively attribute this to developmental abnormality (Fig. 11D). Regarding buccal cingulid expression, we corroborate the comments regarding buccal cingulid expression in this group and again identify the presence of a dentine horn (or BCAC) along the buccal cingulum in a Saimiri sp. specimen (Fig. 11E). Figure 11B illustrates the new schematic that we consider more appropriate and applicable to the cusp configuration of Cebinae lower molars.

In the subfamily Pitheciinae, all molars are commonly stated as possessing four cusps, with a crest connecting the protoconid and metaconid that creates a comparatively narrow trigonid and spacious taloned basin (Swindler, 2002). Very little is mentioned of any additional cusp-like structures in this clade, which may reflect the difficulties presented in identifying dental crown feature at the outer enamel surface in taxa with short cusps and crenulated enamel. As such, the current variation schematic is depicted as a tooth with only the four primary cusps (Fig. 12A). Our observations at the EDJ however identify the frequent presence of accessory cusps on the marginal ridges of Pitheciinae first and second molars. Figures 12C and 12D demonstrates the presence of DMAC and LMAC features in one Cacajao calvus specimen, while Fig. 12E provides evidence of multiple BMAC cusps on the buccal marginal ridge of a Chiropotes satanas lower first molar. In all Pitheciinae specimens, we identify the presence of an accessory cusp directly mesial to the protoconid (Fig. 12F). Unlike the diminutive nature of most accessory features, this cusp is often comparable in size to the neighbouring protoconid and may be mistaken for the primary cusp in some cases. While we classify this feature as an MMAC in the schematic as it is positioned on the marginal ridge between the protoconid and metaconid, further mention and consideration of this cusp will be made in the discussion. Figure 12B represents a new schematic that we consider more appropriate and applicable to the cusp configuration of Pitheciinae lower molars.

The Callicebinae subfamily in this study is represented by a small number of Callicebus molock specimens, but nevertheless demonstrates the presence of numerous accessory cusp features at the EDJ surface. First and second molars are reported to have four cusps, a small trigonid basin, a larger talonid basin, and a well-defined cristid obliqua (Swindler, 2002). In a sample of 40 Callicebus torquatus specimens, a ‘distostylid’ was identified with a frequency of 56% in lower M1 and 83% in lower M2 (Kinzey, 1973), while Swindler (2002) also remarks on the presence of this cusp in a sample of four Callicebus moloch. Distostylids are therefore included in the current variation schematic for Callicebinae (Fig. 13A). Our observations at the EDJ point to the presence of several DMACs on the distal marginal ridge (Figs. 13E and 13G). Additionally, we identify the presence of several MMACs on the mesial ridge (Fig. 13C), and a LMAC on the lingual ridge of a lower second molar (Fig. 13F). Finally, we demonstrate the presence of a prominent BCAC on the same second molar (Fig. 13D). Figure 13B presents the new schematic of Callicebinae lower molars that we consider to represent crown patterning in this clade.

In the subfamily Atelinae, lower first and second molars have four cusps, with a prominent crest that separates the trigonid basin from a wide talonid basin. In Ateles, Orlosky (1973) identifies the presence of hypoconulids in all lower molars. In Alouatta third molars, Swindler (2002) reports the regular appearance of hypoconulid and tuberculum intermedium cusps, however there is no mention of these features in first and second molars. Clark (1971) also described the presence of paraconid cusps on the lower molars of Alouatta. Incorporating these observations, the current variation scheme can be summarized as a tooth crown with four primary cusps, a hypoconulid, and a potential paraconid (Fig. 14A). Our observations at the EDJ confirm the variable presence of DMACs on the distal ridge of Ateles and Alouatta (Figs. 14C and 14F, respectively). No MMAC or LMAC were observed in Atelinae first or second molars. Figure 14D demonstrates the lack of discernible features on the lingual ridge of an Alouatta specimen. In relation to Clark’s (1971) description of a paraconid in Alouatta, we find no examples of paraconid expression in our limited sample. Figure 14E does however demonstrate notable lipping and elevation of the mesial marginal ridge in an Alouatta individual. As this type of ridge morphology may resemble cusp-like structures at the outer enamel surface in some specimens, we tentatively attribute the descriptions of potential paraconid expression to this phenomenon and exclude them from the new schematic for Atelinae until they have been confidently detected (Fig. 14B).

Cercopithecidae

The two subfamilies of Cercopithecidae are presented separately, and the Cercopithecinae subfamily is further separated into their Cercopithecini and Papionini tribes, due to notable differences in molar morphology and accessory cusp expression between the groups.

In the Tribe Cercopithecini, the molars are bilophodont and have high, well-defined cusps. There are a limited number of studies on variation in crown morphology from this group but from a large sample of guenons, Swindler (2002) reported the lack of hypoconulid presence on all lower molars, but the common expression of a protostylid on the lower molars of Chlorocebus aethiops (85% of specimens) and Miopithecus talapoin (100% of specimens). Thus, the current variation scheme can be summarized as a relatively simple tooth crown with the four primary cusps and a protostylid located on the mesiobuccal corner of the crown (Fig. 15A). Our observations at the EDJ of guenon first and second molars correspond with Swindler’s (2002) observations regarding the lack of hypoconulid (or DMAC) presence (Figs. 15C, 15D, 15E and 15F). Whether this relates to developmental constraints associated with a notably small distal fovea in this clade, remains to be determined. Unlike Swindler’s (2002) observations in relation to protostylid expression, however, we find no evidence of a protostylid (or BCAC) in any specimen. This is consistently true, even when alternative definitions of protostylids are adopted (see Hlusko, 2004; Skinner et al., 2008 for discussion of protostylid definitions). Due to the limited sample size available for this study, we cannot rule out the presence of a BCAC in some individuals and thus it is included in our new Cercopithecini schematic (Fig. 15B). Unlike the other Cercopithecidae groups, no other cusp features beyond the four main cusps were observed.

In the Tribe Papionini, molars are bilophodont and display a pronounced buccal flare. Swindler (2002) reports that accessory cusps are variably expressed in several members of this tribe, although are more commonly observed in Macaca and Papio molars. In Macaca fuscata lower second molars, tuberculum intermedium presence on the lingual aspect of the crown was reported in 38% of specimens, and in 56.8% of Papio first lower molars (Swindler, 2002). These features are therefore incorporated into the current variation scheme for Papionini (Fig. 16A). While our observations match Swindler’s comments regarding the common presence of a tuberculum intermedium cusp (or LMAC) in Papionini molars, we extend this description by demonstrating the presence of multiple lingual accessory cusps on the marginal ridge between the metaconid and entoconid in some taxa (Fig. 16C). In these specimens, LMACs are often positioned either deep within the lingual fovea, or on the distal shoulder of the metaconid. Currently, no more than two LMACs have been observed on any Papionini lower M1 or M2, however the new nomenclature allows for the addition of extra cusps if observed. Regarding distal accessory cusp expression, Swindler (2002) comments that it is well known that a shelf or cusp extends from the distal surface of Papionini molars, and that in the lower M3, it is considered the hypoconulid. Szalay & Delson (2013) describe in Theropithecus molars the presence of a “large distal accessory cuspule, which projects backward towards the succeeding tooth” (p. 375). There is no indication of a name from Szalay & Delson (2013) regarding this structure and how they would define it; however, Swindler (1983) has suggested that the distal cusps on the lower M1-2 are serially homologous with the hypoconulid of the M3. It should be noted however, that this is based on topographical and functional associations, not phylogenetic. Our studies at the EDJ support the comments regarding the common observation of distal accessory cusps in Papionini molars. Developing on these descriptions, we report a variable and complex pattern of cusp expression in this clade, including the large single cuspules described by Szalay & Delson (2013), as well as the common occurrence of multiple dentine horn presence along the distal ridge (Figs. 16D, 16E, and 16F). Importantly, from the images of multiple distal accessory cusp expression in this clade, we would argue that even if one wishes to label these features within the current system of nomenclature, the identification and differentiation of a ‘hypoconulid’ from the other cusps present could not be made with confidence. As such, we consider this to support the adoption of the term ‘DMAC’ for all distal accessory cusps in this clade.

In relation to cusp features beyond the cusps commonly situated on the buccal and distal aspects of the crown, Hlusko (2002) studied variation in ‘interconulid’ expression among a large sample of Papio hamadryas, and identified some form of expression in almost the entire collection (95%). In this case, Hlusko (2002) used Swindler’s (1976) definition of an ‘interconulid’ as a stylid between the protoconid and hypoconid of mandibular molars. As previously mentioned however, photographs provided of interconulid expression types in this study appear to show features on the buccal cingulum and perhaps better reflect what has previously been termed an ‘ectostylid’ (Kinzey, 1973). Nevertheless, similar buccal features have also been observed at the EDJ in our sample. These include cusp-like structures on the buccal marginal ridge, and on the buccal cingulum (Fig. 16H). Based on the potential confusion regarding the term interconulid, and the acknowledgement of cusp-like features on the buccal marginal ridge (which do not appear to have a suitable or commonly referenced name), we incorporate BMAC (buccal marginal accessory cusp) and BCAC (buccal cingulum accessory cusp) terms into the new nomenclature to facilitate the identification and differentiation of these features (Fig. 16B). Additionally, we also demonstrate the presence of MMACs (mesial marginal accessory cusps) on the mesial marginal ridge of Papionini molars. Similar to DMAC presence in Papionini M1-M2, patterns of MMAC expression vary from absent to multiple dentine horn expression along the marginal ridge (Fig. 16G).

In the other subfamily of Cercopithecidae, the Colobinae, there is limited discussion in the literature of any particular morphological feature on the molar crowns that may be of interest to this study. The Colobinae are described as having four cusps on the first and second molars, with a variably expressed hypoconulid on the M3. Swindler (2002) notes the variable presence of a tuberculum intermedium in Rhinopithecus on the lower M1 (7%) and M2 (62%), and on the M1 (9%) of Pygathrix. The current nomenclature for this clade can therefore be summarized as a simple tooth crown with four primary cusps, and a potential C7 (aka LMAC) on the lingual marginal ridge (Fig. 17A). While Rhinopithecus is not included in our sample, we see no examples of what we would consider a LMAC in any of our Colobinae specimens. Nevertheless, due to low sample sizes among some groups, and incomplete representation of the whole subfamily of Colobinae, we cannot rule out the presence of lingual accessory cusps in some individuals, and would recommend the LMAC designation for these features in future studies. Figure 17C shows the presence of a DMAC in a Colobus guereza lower right second molar, and a DMAC in two Presbytis melaophos molars (Figs. 17D and 17E). Specimens F–H in Fig. 17 demonstrate the simple M1/M2 molar morphology in this clade, and lack of discernible accessory features within these specimens. Based on these findings, the new schematic has the inclusion of both LMAC and DMAC (Fig. 17B).

Hominoidea

In the family Hominidae, molars generally have five main cusps, arranged in a Y-5 pattern. Importantly, because the fifth cusp (or hypoconulid) is consistently expressed in M1-M2 in this clade, it is not considered to be an accessory feature and is included as a ‘hypoconulid’ in the following schematics. Further dialogue regarding the inclusion of the term hypoconulid in this clade follows in the discussion. In addition to these five cusps, the variable presence of a tuberculum sextum and tuberculum intermedium are commonly reported in certain members of this clade. Swindler (2002) reports that Gorilla have the highest frequency of C6, while Pongo have the least. Previous studies of dental trait expression at the EDJ have provided a more detailed analysis of C6 and C7 expression in hominoids, noting variation in the placement of the accessory dentine horns on their respective marginal ridges, as well as observations of double C6 in some Homininae specimens (Skinner et al., 2008). The current variation schematic can therefore be summarized as a tooth crown with four primary cusps, a variable C7 cusp, and multiple potential C6 cusps (Fig. 18A). Our observations corroborate the presence of multiple DMACs on the distal ridge between the hypoconulid and entoconid of some Homininae molars (Figs. 18C and 18D). Assuming that the larger of the two distal cusps is the hypoconulid, Fig. 18D demonstrates the rare presence of a DMAC between the hypoconulid and hypoconid. While variation in the patterning and placement of these cusps may reflect developmental differences in their formation and expression, we still consider the use of the term DMAC for all distal accessory cusps appropriate as it does not intend to imply homology. In addition to DMAC expression, we also note the presence of LMACs in several of our Hominidae sample. Figs. 18E and 18F demonstrate the presence of single LMACs in a Homo sapiens and Gorilla gorilla specimen respectively. No other accessory features were observed. As such, the new schematic of Hominidae lower molars does not include any new features but does replace C6 and C7 terms with the more appropriate DMAC and LMAC designations (Fig. 18B).

In the family Hylobatidae, the lower molars possess five cusps, a narrow trigonid, and a more spacious talonid basin. Reports of a Y-5 pattern follow frequencies of roughly 100% in LM1, and 97% in LM2 (Frisch, 1965; Swindler, 2002). Swindler (2002) reports tuberculum intermedium expression of 0.07% in LM1 and 18% in LM2 in a sample of Hylobates molars. Tuberculum sextum presence, or any other form of accessory trait expression, was not discussed. The current variation schematic can be summarized as a crown with four primary cusps, a prominent hypoconulid, and a potential tuberculum intermedium (Fig. 19A). From our observations, we identify the presence of a small DMAC between the hypoconulid and entoconid in a Hylobates muelleri M2 (Fig. 19C), and a single example of a very mesially-positioned LMAC in Hylobates muelleri M1 (Fig. 19D). No dentine horns beyond the five primary cusps were observed in our Symphalangus sample (Figs. 19E and 19F). Figure 19B represents the new schematic of Hylobatidae lower molars that we consider to represent a better reflection of crown patterning in this clade.