Introduction

The diversification of Cercopithecoidea occurred in the context of the climate-driven environmental changes between the Eocene and Holocene. The ancestral members of this superfamily lived in the wet, forested regions of the Late Eocene to Early Oligocene Fayum, perhaps a similar environment to modern-day Uganda, and postcranial evidence suggests most of them were arboreal quadrupeds inhabiting a swampy, densely vegetated paleolandscape with high rainfall (Simons 1995; Kay et al. 1997). By the Neogene, the paleoenvironment of the Fayum and Northern Africa in general had become drier and more open with forests giving way to sparse open woodlands inhabited by cercopithecoids like Victoriapithecus with semi-terrestrial locomotion (Benefit 1999; Jablonski 2002). The African paleoenvironment continued a trend towards more open woodland and grassy savanna (DuPont and Leroy 1995), where cercopithecoids like Dinopithecus were habitual terrestrial quadrupeds that likely lived in large baboon-like troops during the Plio-Pleistocene and Holocene (Elton 2012).

Adapting to more open landscapes and living in larger troops likely required behavioural adaptations associated with increased inter and intra-group competition and greater social complexity, which has been argued to be associated with larger absolute brain size (Benefit 1999). The relatively large brain in primates functions in reproductive and life history strategies associated with learning and cognitive and memory skills in species with prolonged growth and low reproductive turnover but complex social organisation (Shea 2006; Dunbar 2009). For stem-anthropoids inhabiting dense tropical rainforests such as those that existed during the Eocene-Oligocene transition, higher visual acuity and the ability to distinguish between closely spaced visual stimuli, coupled with colour vision that would have allowed the discrimination of red, orange ,and yellow fruits and between those which were ripe and unripe, would likely have been important brain functions (Dominy 2004; Kirk and Kay 2004). This shift away from forest habitats to larger, open-habitat group structures would have changed foraging and anti-predator strategies perhaps in ways that are found in among extant terrestrial primates that form troops the more open landscapes of Africa today (Cheney and Wrangham 1987; McKee and Keyser 1994; Gilbert et al. 2018). Extant cercopithecoids are known to rely extensively on visual communication such as determining colour changes related to sexual communication (Changizi et al. 2006; Moreira et al. 2022) and changes in facial expressions (Peterson et al. 2018), which were all at the expense of auditory and olfactory communication (Strier 2017). The Social Brain Hypothesis posits that the association between increased brain size and larger troop is a causal one, driven by more the complex social behaviours that some cercopithecoids would have started adopting in the Miocene (Barton 2006; Kudo and Dunbar 2001).

Many functions that can logically be linked to these behavioural transitions—memory, object recognition, perception of places and paths—are coordinated through the temporal lobe of the brain (e.g., Murray et al. 2007; Bryant and Pruess 2018; Schuurman and Bruner 2023). The temporal lobe is one of the least studied regions of the primate brain despite its importance in early anthropoid and human evolution, although there are now a few studies in non-human primate outgroups that are relevant to assessing evolution of the temporal lobe in hominin evolution (Rilling 2014; Braunsdorf 2021). Compared to other mammals, primates have greater visual acuity than most other mammals, and anthropoids have greater visual acuity relative to other primates (Kirk and Kay 2004; Veilleux and Kirk 2014). The temporal lobe in anthropoid primates also has an important role in processing visual, auditory and memory-related tasks (Murray et al. 2007). Visual acuity is related to specialised organisation of ganglion cells in the retina and information processed via two pathways, ventral and dorsal, in the middle temporal and inferior temporal cortexes (Pruess 2006). The dorsal pathway neurons are sensitive to object motion, while the ventral pathway neurons are sensitive to object form (Pruess 2006).

Despite the logic that both brain size and proportions of regions like the temporal lobe are likely to have been affected by the environmental transitions of the Cenozoic, examination of absolute or relative brain size in extinct and extant cercopithecoids remains has been pursued by only a few studies (Barton 2006; Grabwoksi 2016; Sansalone et al. 2020). The quantification of cerebral proportions relative to brain size is indicative of cerebral reorganisation, and extinct taxa have the potential to pinpoint when and where in evolution such reorganiation occurred and in what environmental context, allowing a understanding of when behavioural changes may have occurred alongside morphological adaptations. Relative changes to olfactory bulb size and frontal lobe size have been noted alongside increases in absolute brain size during cercopithecoid evolution (Simons 1995, 1997, 2001; Bush et al. 2004; Gonzales 2015; Harrington 2016). Notably, the identification of cerebral reorganisation in extinct catarrhines have functional, evolutionary, and phylogenetic significance for the important neuroanatomical changes in the Cercopithecoidea.

Whether the increase in brain size in crown cercopithecoids was associated with the Cenozoic environmental transition and whether it was accompanied by a relative increase or decrease in the temporal lobe remains largely unknown. However, the proposition that the opening of Neogene environmental and concomitant changes in cercopithecoid behaviours was accompanied by temporal lobe reorganization can be tested using both extinct and Recent taxa using the bony middle cranial fossa as a proxy. The Functional Craniology Hypothesis of Moss and Young (1960) suggests that a reciprocal association exists between brain and skull form, where the temporal lobe is in close spatial proximity to the middle cranial fossa (MCF) of the basicranium (Bastir et al. 2008). This association has allowed MCF metrics to predict temporal lobe volume (TLV) in extant and extinct anthropoids with reasonable accuracy (Rilling and Seligman 2002; Pearson et al. 2020, 2021).

The aim of this paper is to determine whether there was an increase in temporal lobe size relative to brain size between extinct and extant cercopithecoids and whether the potential temporal lobe size changes coincided with the paleoenvironmental transitions and whether there is any association with the emergence of anthropoid behavioural ecology. To examine these aims, we studied eight key extinct species from the Cercopithecoidea: Proteopithecus sylviae and Catopithecus browni from the Late Eocene of the Fayum in Egypt; Parapithecus grangeri, Apidium phiomense, and Aegyptopithecus zeuxis from the Early Oligocene of the Fayum; East African Middle Miocene Victoriapithecus macinnesi; South African Pliocene Dinopithecus ingens,; and Pleistocene Papio angusticeps. We also studied three extant cercopithecines: Cercocebus atys, Macaca mulatta, and Papio anubis. Evolutionary changes to relative temporal lobe size among stem anthropoids and extant cercopithecines may pinpoint the timing of cerebral reorganisation in the Cercopithecoidea and the role of phylogenetic and functional drivers in temporal lobe evolution.

Materials and methods

Two different imaging modalities were used in the initial stages of this study, micro-computed tomography (µCT) scans of ex vivo extinct and extant anthropoid crania (n = 142) and in vivo T1-structural magnetic resonance imaging (MRI) of the human and non-human primate brain (n = 78). The total sample (n = 222) represents 17 anthropoid species including eight extinct catarrhine species (Table 1). Nonhuman primate MRI scans were accessed from the National Chimpanzee Brain Resource with necessary animal ethics protocol approvals (Rilling and Insel 1999) and human MRI scans accessed from the Open Access Series of Imaging Studies (OASIS) with appropriate human ethics protocol approvals (Marcus et al. 2007). All processing of MRI scans was conducted in Brainsuite 17a (Shattuck and Leahy 2002).

Table 1 The total anthropoid sample (N = 222) for the 11 extant and eight extinct anthropoid species (in bold). Digital imaging modality (computed tomography/CT, micro-computed tomography/µCT, magnetic resonance imaging/MRI) sample sizes with sex distribution in parentheses. The ID column lists the species for extant taxa and the specimen number for extinct taxa (with full identification details in the manuscript). Superscripts: a Human in vivo MRI from the Open Access Series of Imaging Studies (OASIS); b Non-human primate in vivo MRI from the National Chimpanzee Brain Resource
Full size table

Fossil specimens

In this study, we analysed miro-computed tomography scans of the fossil specimens from collections housed at Duke Lemur Centre Division of Fossil Primates (DPC): DPC- 18651, DPC-9867, DPC-11388, DPC-14095, DPC-1208, and DPC-5401; The Cairo Geological Museum (CGM): CGM-40327 and CGM-85785; The Ditsong National Museum of Natural History (DNMNH): DNMNH:SK:554, CO100, CO135, and SK554; and the National Museums of Kenya (KNM): KNM-MB-29100.

Data collection

To predict right temporal lobe volume (TLV) in extant and extinct catarrhines, seven linear metrics were digitally measured (in mm) using Checkpoint (Stratovan 2021) on the right middle cranial fossa (MCF) surface in extant anthropoids (Fig. 1). TLV was predicted in the extinct taxa and compared to the predicted TLV in Three extant cercopithecoid species.

Fig. 1
figure 1

The seven metrics used to predict right temporal lobe volume, shown on the right middle cranial fossa (MCF). See Online Resource 1 for anatomical description of metrics used

Full size image

Endocranial volume in extant taxa was calculated (in cm3) through the generation of virtual endocasts from the µCT scans in Slicer 4.0 (Fedorov et al. 2012). For extinct taxa, endocranial volumes (in cm3) were taken from published literature (Online Resource 2, Table 1). For comparison with predicted TLV, the endocranial hemisphere volume (EHV) was calculated as the endocranial volume/2.

Statistical analysis

Preliminary analysis examined whether endocranial volume (EHV) was a suitable proxy for cerebral hemisphere volume. A phylogenetic generalized least-squares (PGLS) regression tested the correlation of extant anthropoid right cerebral hemisphere volume (in cm3) and right temporal lobe volume (in cm3) calculated by voxel from the MRI. The temporal lobe data were taken from Pearson et al. (2020), where a more detailed description of how we defined temporal lobe boundaries can be found. Intraobserver error was assessed by an analysis of variance, where measurement error was calculated as the proportion of the mean-squared differences between replicates relative to the total between-group variation (see Bailey and Byrnes 1990). For the seven MCF metrics, measurement error ranged from ≤ 1–4%. These were not considered error values of note and analysis proceeded.

A second PGLS regression examined the correlation between endocranial hemisphere volume (EHV) (in cm3), calculated from the endocasts, and predicted temporal lobe volume (in cm3), calculated from right MCF metrics. Both PGLS regressions indicated the predicted TLV and endocast EHV were suitable proxies for cerebral volumes (Online Resource 3).

All MCF variables were first converted from linear measurements (in mm) to natural logarithmic units. A multiple regression of MCF variables provided an equation to predict TLV in extant and extinct taxa. The cube-root was taken for TLV and EHV (in cm3), then converted to natural logarithmic units to linearise the scaling relationships for further analysis (Simpson et al. 2003).

PGLS regressions tested the correlation between TLV and EHV in extant and extinct taxa with PGLS regressions followed Martins and Hansen (1997). A consensus anthropoid tree was downloaded from 10kTrees version 3 (Arnold et al. 2010) containing Papio anubis, Cercocebus atys and Macaca mulatta with their molecular divergence dates. We then modified the consensus anthropoid tree in Mesquite version 3.70 (Maddison and Maddison 2021) to include extinct taxa with the generally accepted geological ages and taxonomic affinities used to estimate the phylogeny (Fig. 2).

Fig. 2
figure 2

Phylogenetic tree of the cercopithecines used in this study. Scale bar is in millions of years (MYA)

Full size image

Blomberg’s K (kappa) statistic was used to assess the phylogenetic correlations of the traits. K measures the proportion of the tip variance that matches the expectation of Brownian motion relative to the total variance and thus ranges from 0 to 1 (Blomberg, Garland et al. 2003). Values of K close to 0 indicate little congruence between the trait and the phylogenetic tree as expected under Brownian motion, whereas values closer to 1 indicate a strong congruence.

To examine if TLV was relatively larger or smaller than EHV in extinct taxa compared to extant taxa, further PGLS regressions tested whether the values for the slopes and intercepts were as high or higher than those observed by chance using random sampling (5,000 iterations).

All statistical analyses, PGLS regressions, K calculations, and trait- mapping were performed using Phylogenetics for Mathematica version 6.5 (Polly 2019) and performed in Mathematica version 13.1 (Wolfram 2021). For full Mathematica script, see Pearson (2023).

Results

We examined whether there was an increase or decrease in relative TLV to EHV in extinct taxa compared to extant cercopithecoids through trait-mapping, PGLS regressions with random sampling of regression slopes and intercepts between extant cercopithecines and stem-anthropoids.

Predicted temporal lobe volume

A multiple regression of naturally log-transformed TLV and seven MCF variables provided a statistically significant correlation (adjusted r2 = 0. 98, p < 0.01) in extant anthropoids (Table 2). Natural log-transformed variables for TLV and EHV were analysed with a multiple regression with the resultant prediction equation used to predict extant and extinct TLV (Table 3).

Table 2 Multiple regression statistics for temporal lobe prediction with statistically significant values in bold. Abbreviations: b = intercept; r2 = coefficient of determination; Adjusted r2 = coefficient of determination adjusted for multiple regression; SE = Standard error of regression; SEE = Standard error of the estimate
Full size table
Table 3 Natural log-transformed predicted TLV and EHV for Cercopithecoidea with predicted endocranial hemisphere volume (EHV) and temporal lobe volume (TLV).
Full size table

Evolution of brain size and temporal lobe size in the Cercopithecoidea

To examine the potential effects of phylogeny on EHV and TLV in the Cercopithecoidea, a PGLS regression of naturally log-transformed TLV and EHV produced a statistically significant correlation (r2 = 0.71, p ≤ 0.00107) in both extinct and extant taxa (Table 4; Fig. 3). Trait mapping indicated a stronger influence of phylogeny on EHV (k = 3.67) compared to a lesser influence on TLV (k = 0.60) in both extant and extinct taxa (Table 4; Fig. 4).

Table 4 Phylogenetic generalised least squares regression of temporal lobe volume (TLV) and hemisphere volume (EHV) in the Cercopithecoidea. Abbreviations: PGLS = phylogenetic generalised least-squares regression; a = slope; b = intercept; r2 = coefficient of determination; p = p-value; k = Blomberg’s Kappa
Full size table
Fig. 3
figure 3

Phylogenetic generalised least-squares regression of temporal lobe volume (TLV) on endocranial hemisphere volume (EHV) in the Cercopithecoidea

Full size image
Fig. 4
figure 4

Comparisons of trait-mapping in the Cercopithecoidea. a. EHV; b. TLV

Full size image

Relative temporal lobe size to brain size in the Cercopithecoidea

The difference in relative TLV to EHV in the Cercopithecoidea was conducted using two groups of taxa: (1) the clade of Miocene and later taxa, including the 3 extant cercopithecine species and 3 extinct cercopithecoids, and (2) the pre-Miocene stem-group of 5 extinct anthropoid species. PGLS regressions tested whether the slope and intercept values in the two groups were more different than expected by chance using a non-parametric bootstrapping test. The difference in slopes using random sampling with 5,000 iterations was statistically significant (p = 0.0014), while the difference in intercepts using random sampling with 5, 0000 iterations was not statistically significant (p = 0.27) between the two groups (Table 5; Fig. 5).

Table 5 PGLS Regression of temporal lobe volume (TLV) and hemisphere volume (EHV) in stem-anthropoids and extinct and extant cercopithecoids with test for difference in slopes and intercepts and bootstrap (5,000 iterations). Abbreviations: PGLS = phylogenetic generalised least-squares regression; a = slope; b = intercept; r2 = coefficient of determination; p = p-value
Full size table
Fig. 5
figure 5

Comparison of random sampling in slopes (a) and intercepts (b) of stem-anthropoids and extinct and extant cercopithecoids from phylogenetic generalised least squares regression. The red bar represents the slope and intercept values in Table 5

Full size image

Ultimately, these findings indicate temporal lobe size was larger relative to brain size in the extant cercopithecoid clade compared to the stem-anthropoids (Fig. 6). We note that the Late Eocene taxon Catopithecus and Early Oligocene taxa Apidium and Parapithecus lie well below the regression line as fit to the entire catarrhine clade. Differences in the slope for the stem-anthropoids was near zero and slightly negative which suggests that as brain size got larger there was either no relationship to temporal lobe size or that the temporal lobe size became proportionally smaller. In the crown-clade the slope was near 1.0, which suggests isometric scaling with body size: as brain got larger, the temporal lobe scaled 1-to-1 with it.

Fig. 6
figure 6

Phylogenetic generalised least squares regression depicting the regression line of TLV on EHV and comparison of slopes between stem-anthropoids (in blue) and crown-clade cercopithecoids (in orange) showing that relative TLV to EHV changed most noticeably between the stem-anthropoids and the crown-clade in the Late Eocene and Early Oligocene

Full size image

Discussion

Brain evolution in the Cercopithecoidea and paleoenvironment

In general, our findings agree with Radinsky (1974) and Falk (1978), who documented an absolute brain size increase in cercopithecoid evolution (Fig. 4b). Our results also suggest a striking shift in the relationship between an increase in temporal lobe size relative to brain size between the stem and crown groups. Because of the links of the temporal lobe to visual acuity, memory, object recognition, and spatial navigation, and the roles that these play in extant cercopithecoid behaviours, it seems plausible that the marked transition between stem and crown scaling of the lobe is related to changes in individual and social behaviours that were associated with the environmental transitions from the Paleogene to the Neogene.

The changes in the paleoenvironment were also likely associated with changes in absolute brain size and body mass. A period of cooling and drying during the Oligocene likely required stem-anthropoids to expand into newer environments, where changes from an arboreal to semi-terrestrial locomotion would likely have been associated with shifts in body mass and diet (Scott 1995; Fleagle 2013) Among extant cercopithecines, species with larger body masses are unable to forage in the thinner upper canopy branches, which is where smaller-bodied cercopithecines have used niche partitioning (Cheney and Wrangham 1987). During the Miocene the continuing drier climate resulted in an expansion of sub-Saharan grassland savannahs and a shift from dense rainforests to more open woodlands favouring semi-terrestrial locomotion rather than obligate arboreal with postcranial and dentition of the extinct cercopithecine Victoriapithecus resembling extant taxa which favour semi-terrestrial locomotion also a diet of hard fruit and seeds (Benefit and McCrossin 2002).

Today, terrestrial primate taxa often have larger group sizes and a larger body mass. Studies of Papio, with its large body mass and strong sexual dimorphism, show that it has high levels of intragroup aggression and diets that are reliant during the African dry season on poor vegetation consisting of grasses and subterranean foods, supplemented in other seasons with fruit, leaves and occasionally meat consumed by males (Melnick and Pearl 1987). In contrast, Macaca, which inhabits tropical rainforest and woodland environments that are dietarily appropriate for a much smaller and absence sexually dimorphic body size and intragroup aggression. The semi-arboreal Cercocebus inhabits wetland, tropical woodland forest to open woodlands with smaller body mass and is reliant on the consumption of fruit and leaves (Melnick and Pearl 1987). These comparisons suggest that extinct taxa associated with tropical and rainforest vegetation might have consumed nutrients that would be suitable for larger brain size compared to the lower quality foods of savannas and grasslands. However, we found that extant taxa Cercocebus and Macaca, which inhabit forests and consume fruit and leaves, had smaller relative brain sizes compared to the savannah taxon Papio. Clutton-Brock and Harvey (1980) argued that diet and intra-family brain size are associated more with diet, where folivores had comparatively smaller brain size than frugivores and a lesser association between range size. In a more recent study by Chambers et al. (2021), diet was a better indicator of brain size than range size or social structure, which is at odds with the expectation of the Social Brain Hypothesis (Chambers et al. 2021). However, both studies suggested that diet variation and consistent access to resources might play a more significant role in primate brain size than social system.

In the Paleocene, early primate and euprimate taxa are thought to have been small-bodied and nocturnal and to have had limited colour perception (Dominy 2004). Their reliance on prehensile tail or grasping hands and feet to navigate the upper canopies were coupled with keen olfaction to better exploit access to fruit, but in the Eocene, stem anthropoids inhabited forests, exploiting fleshy fruit from palms, figs, and more lipid-rich laurels (Dominy 2004). Extant anthropoid primates are mostly diurnal with a few exceptions such as the nocturnal platyrrhine Aotus, but the emergence of colour vision is a staple argument in the divergence of anthropoids from the strepsirrhine-like primates ( Fleagle 2013; Strier 2017). Climate cooling in the late Paleogene resulted in increased seasonality in Africa and the loss of the palm forests; this required stem anthropoids to consume of proteinaceous leaves instead of fruit, as observed in the Egyptian Fayum deposits where Early Eocene anthropoids show a shift from fruit and insect diet to Early Oligocene preference for fruit and leaves (Dominy 2004). The the development of trichromatic vision has likely implications for cercopithecoid temporal lobe evolution where the ability to discern young reddish leaves against a background of mature foliage likely selected for those taxa with higher visual acuity which were more able to discern young seasonal leaves with greater nutritional value (Dominy 2004).

Cercopithecine behaviour and implications for temporal lobe size

Our findings indicated an isometric scaling for brain size likely associated with an increase in body size throughout cercopithecoid evolution as previously noted by Radinsky (1974) and Falk (1978). Potential drivers for an increase in brain size in Middle Miocene and Plio-Pleistocene taxa include the emergence of sexual dimorphism in Aegyptopithecus (Simons et al. 2007), suggesting increased intragroup social aggression, which is commonly found in extant cercopithecines with larger troop sizes such as savanna baboons (Cheney and Wrangham 1987; Melnick and Pearl 1987; Meguerditchiana et al. 2002).

The temporal lobe in humans is a highly multimodal association cortex associated with language processing, memory tasks, and auditory and visual processing (Rilling 2014). In non-human primates, which are less studied than humans, distinguishing auditory, memory, visual processing has been more difficult (Braunsdorf et al. 2021). A comparison between Homo sapiens and Pan troglodytes temporal cortex noted homologous language regions hidden beneath a sulcal fold in chimpanzees but visible on the sulcal surface in modern humans (Rilling et al. 2012). Complex vocal communication is not restricted to modern humans, but our species does rely heavily on a language system to the detriment of visual communication. In human and nonhuman primates, visual signalling and the processing of visual communication occurs in the temporal cortex and is associated with facial recognition and memory (Pruess 2006). Thus, in the early emergence of stem anthropoids inhabiting dense tropical rainforest, the ability to distinguish and memorise the faces, facial expressions, and appearance of group members versus other groups would have become increasingly relied upon, especially with the reduction in olfactory reliance for visual senses (Fleagle 2013).

The evolutionary implications for catarrhine reliance on visual communication is related to visual acuity, or the ability to distinguish between closely spaced visual stimuli. Diurnal anthropoids have significantly higher visual acuity than other mammals (Kirk and Kay 2004). In dense tropical rainforests during the Eocene, higher visual acuity combined with colour vision would allow discrimination of red, orange, and yellow fruits (Dominy 2004).

In dense rainforests, low-frequency calls travel further than higher-frequency calls at the same amplitude (Seyfarth 1987; Strier 2017). The use of greater vocal communication to pinpoint other group members may have been traded for greater reliance on visual acuity. For example, Late Eocene Cateopithcus and Proteopithecus had temporal lobes which are proportionally much larger and smaller, respectively. These differences in relative temporal lobe size could be associated with behavioural attributes, where greater reliance on navigating upper canopies versus lower canopy subsistence could result in a shift towards higher visual acuity, greater memory reliance, and cognitive complexity rather than lower visual acuity but greater auditory reliance. The Late Oligocene stem anthropoid Parapithecus was likely diurnal with visual acuity as high as extant primates (Kirk and Kay 2004). Our findings suggest Parapithecus had a small temporal lobe relative to brain size, which raises questions about whether increased temporal lobe function and higher cognitive processes were directly related to an increase in cortex size or, more likely, whether Parapithecus represents a transitional shift in cerebral reorganisation in the stem group of the Cercopithecoidea.

We found extant Papio had a slightly larger relative temporal lobe for brain size. Since baboons are terrestrial quadrupeds that inhabit open grassland and savannah environments, we hypothesize that larger group size, more vocal and visual communication such as the visual pathways processing in the inferior temporal cortex associated with facial recognition (Pruess 2006), behavioural cues including lip-curling and eye-lid flips commonly observed aggression displays among baboons (Strier 2017) likely require greater cognitive demands on the temporal lobe. Our findings imply the presence of interconnections between paleoenvironment and eco-behavioural adaptations in stem anthropoids and throughout extinct cercopithecine evolution.

Relative temporal lobe size in the Cercopithecoidea

Our findings indicate differences in relative temporal lobe size to brain size between extinct and extant cercopithecines and stem anthropoids. The evolution of the stem anthropoid temporal cortex may be influenced by visual and auditory communication. The smaller-bodied Late Eocene taxa were habitually arboreal quadrupeds and leapers inhabiting dense canopies where the demands of a higher functioning temporal cortex was selected in favour of relatively large olfactory bulbs for brain size (Simons 1995).

We did not find a consistent observable trend in relative temporal lobe size among the stem anthropoids, but Late Eocene Proteopithecus had smaller relative temporal lobes to brain size compared to the contemporaneous Catopithecus, which had proportionally much larger temporal lobes for brain size. Early Oligocene Apidium and Parapithecus both had smaller relative temporal lobes compared to the sexually dimorphic Aegyptopithecus, which had temporal lobes proportional to brain size. In the extant and extinct cercopithecines, temporal lobe scaled proportionally with brain size, likely associated with increases in body mass as present in the Middle Miocene extinct cercopithecoid Victoriapithecus, Pliocene extinct cercopithecine Dinopithecus, Early Pleistocene Papio, and extant cercopithecines Macaca, Cercocebus and Papio.

Conclusion

Our findings about relative temporal lobe size show that the scaling relationship between temporal lobe and brain size changed between the stem cercopithecines of the Late Eocene and Oligocene compared to the crown group in the Late Miocene and Quaternary. The timing of the transition suggests that cerebral reorganisation occurred at several key stages during the evolution of the Cercopithecoidea but was most noticeable in the Late Eocene and Early Oligocene during the initial opening of paleoenvironments.

Through the application of virtual imaging methods and phylogenetic comparative methods, this study expands on previous work conducted on absolute cercopithecoid brain evolution. In general, we found in agreement with previous findings that absolute brain size increased throughout cercopithecoid evolution. However, we note caution as there is continued debate about the phylogenetic relationships among stem-anthropoid taxa, and because there is a range of phylogenetic, behavioural and neurological differences in the taxa we have studied.

This study provides the first quantification of evolutionary changes to relative brain size and temporal lobe size in the Cercopithecoidea. The inclusion of comparative neuroanatomy and cranial anatomy, phylogenetic comparative methods and robust statistical analyses allow further understanding of relative brain size and temporal lobe evolution in the Cercopithecoidea. Potential drivers for temporal lobe evolution include changes to the processing of visual and communication pathways in response to fluctuating in paleoenvironment and socioecology.