The overall goal of the present study is the identification of turbinals and associated structures (e.g., semicircular crest) among three primate lineages based on the observed developmental patterns of the mammalian nasal capsule. Because we are studying relatively rare and slowly reproducing species, our sample size is small; our results admittedly will suffer from the lack of a broad phylogenetic foundation, the same deficit that may be attributed to old literature on the chondrocranium (see further discussion of this point by Sánchez-Villagra and Forasiepi 2017). However, numerous studies have relied on samples with developmental stages to establish the identity of, or to infer the absence of, individual turbinals (e.g., Macrini 2012; Ito et al. 2021; Macrini et al. 2023). Identification of developmental patterns can, in turn, provide critical evidence for broader phylogenetic analyses (Macrini et al. 2023). Whereas establishing the precise homology of turbinals is beyond the scope of the present study, our developmental data provide critical context for the turbinal anatomy of adult primates, recently comprehensively studied by Lundeen and Kirk (2019) and Lundeen and Kay (2022), and will broaden our knowledge of comparative nasal development of primates, which is currently restricted mainly to a small number of strepsirrhines (Smith et al. 2007; Smith and Rossie 2008).

Ultimately, our study will also support future efforts to record traits as numeric codes across a large species sample (phenotyping approach by Sereno 2007; Stefen et al. 2022). The coding of descriptive (i.e., text-based) phenotypic traits makes them computer-parsable and accessible to analyzing software for genotype-phenotype alignments and phylogenetic analyses (Stößel et al. 2010; Stefen et al. 2022; Christmas et al. 2023). Previous studies correlated genomic data (e.g., number of olfactory receptor genes, OR genes) with continuous morphometric data like turbinal surface area measurements (e.g., Martinez et al. 2023, 2024a).

The present study is part of a project which aims to associate phenotypic traits of the nasal cavity in primates with their sequenced OR gene data. Therefore, the selection of primate species was based on two preconditions: the availability of a fully sequenced genome, followed by the access to image data for comprehensive age stages. Genome data are available for 45 primate species so far; OR gene data have been extracted for 43 species (Zoonomia Consortium: Genereux et al. 2020; Christmas et al. 2023). The phylogenetic relationships among the sampled species were adapted from the Zoonomia Consortium (https://zoonomiaproject.org/the-mammal-tree-view; Genereux et al. 2020) (Fig. 1).

Figure 1. 

Topology of the placental grandorder Euarchontoglires, that was adapted from the Zoonomia Project (https://zoonomiaproject.org/the-mammal-tree-view). The seven selected primate genera cover the higher lineages Haplorhini: Catarrhini (n = 2), Haplorhini: Platyrrhini (n = 3), and Strepsirrhini (n = 2). Glires includes Rodentia and Lagomorpha.

Obtaining a comprehensive intraspecific ontogenetic series of imaging data is challenging—especially when the sample selection was further limited by access to genomic data. Because we expect little variation between closely related species (e.g., Paulli 1900c), some of our age series were selected on genus level (i.e., sister species with unknown OR gene data) to increase the age range. For instance, we included an age series of tamarins that are especially closely related, the bare-faced tamarin group (Saguinus oedipus and S. geoffroyi, Rylands et al. 2016). Other species are selected for ontogenetic information on adult primates previously studied by other authors (e.g., Lundeen and Kirk 2019). In total, our current sample of imaging data included seven primate genera, for all of which OR gene data are available. Six of them cover four to five ontogenetic stages (fetal to adult; as inferred from exact age or laboratory notes like “late fetal”, see below and Table 1). They equally represent three higher taxa (1) Strepsirrhini (Lemur and Otolemur), (2) Haplorhini: Platyrrhini (Aotus and Saguinus), (3) Haplorhini: Catarrhini (Papio and Macaca) (Table 1, Fig. 1). One caveat is that similar life stages (e.g., “neonate”) are nevertheless at different points of development (heterochrony; Smith et al. 2016; Werneburg and Yaryhin 2018). The turbinal growth in each of the six ontogenetic series serves to identify possible lineage-specific patterns. A third platyrrhine, Pithecia, is represented by a neonate and an adult stage.

The selection of fetal and perinatal specimens was intended to provide novel information on the grundplan of the primate nasal capsule, in combination with prior observations (e.g., Maier 1980, 2000; Smith and Rossie 2008; Maier and Ruf 2014). Thirty cadaveric specimens were used to create anatomic reconstructions (Table 1). No specimens were sacrificed for this study. All fetal or newborn specimens and some juveniles and adults died of natural causes in captivity at zoos or primate research centers. Subadult and adolescent M. nemestrina were acquired from a laboratory conducting research unrelated to our study, after the animals were sacrificed (see Smith et al. 2001). For some specimens the exact age or at least the age stage (neonate, juvenile, etc.) was noted. For instance, the late fetal P. anubis Papio107 died at 150 days gestation (average total gestation length 180 days, Smuts and Nicolson 1989). Commonly, adult mammals are recognized by their fully erupted permanent dentition (Smith et al. 2020). However, in the M. nemestrina A3 (female, 4 years, 4 months, and 20 days) all M2/m2 are in occlusion, whereas no M3/m3 has yet erupted. Sexual maturity is reached at about 35 months in females (see table 1 in Harvey and Clutton-Brock 1985). Thus, this specimen was certainly sexually mature, although dentally it was not fully adult. Some skeletonized specimens were acquired for scanning directly from museums (or indirectly from MorphoSource, https://www.morphosource.org, for the for the Media ID see Table 1). Some specimens were prepared for histological study in prior studies (as cited in Table 1). Except for skeletonized museum specimens, all cadaveric specimens were stored in 10% buffered formalin prior to scanning and/or histology. Some specimens were frozen first prior to formalin immersion; these were excluded from histological methods.

Whole heads and bodies, respectively, have been scanned with the GE V|TOME|X M 240 nano-CT device housed at the Nanoscale Research Facility, University of Florida, Gainesville. Additional µCT data used in previous studies and stored in the laboratories of VBD/TDS were included, along with image volumes obtained from the database MorphoSource (Table 1). The resolution (cubic voxel size) of the scans ranged from 0.020500 to 0.093786 mm.

Some histological material has been investigated in previous studies (Table 1). The preparation of each specimen (here on the example of Aotus1) at the School of Physical Therapy, Slippery Rock University followed the protocol described in DeLeon and Smith (2014). After CT scanning, the specimen was restored to 10% neutral buffered formalin. The right hemiface (septum included) was trimmed with a scalpel from the nose tip up to the caudal end of the olfactory recess as inferred from the CT scans. Most of the soft tissues surrounding the nasal cavity (eye, skin, muscles, etc.) were removed and the zygomatic was cut. For the decalcification, 50 g sodium citrate (C₆H₅Na₃O₇) was dissolved in 125 ml formic acid (CH₂O₂) and diluted to 500 ml with distilled water (Evans and Krajian 1930). The solution was replaced weekly and simultaneously the progress of decalcification was tested by mixing 5 ml of the solution taken from the bottom of the jar with 1 ml of a sodium/aluminum oxalate solution (5 g sodium oxalate Na₂C₂O₄ dissolved in 100 ml distilled water, lightly boiled and cooled). After the decalcification was complete (sodium oxalate no longer precipitates), the specimen was stored in fresh decalcifier for one additional week and returned to 10% neutral buffered formalin. Prior to paraffin embedding, the specimen was dehydrated by a gradual series of ethanol concentrations (50% up to 100%), cleared in xylene for about 3 hours, and transferred to melted paraffin. The block was placed in a vacuum oven (at 60°C) to enable the paraffin to penetrate the tissue and replace the air. After the paraffin had completely hardened, Aotus1 was sectioned at 12 µm with a rotary microtome. The previously obtained histological specimens were sectioned at 10 or 12 µm (Table 1). Every 5th section was mounted on a glass slide and selected sections were stained either with Gill’s Hematoxylin-Eosin or with Gomori-Trichrome (Gomori 1950; Gill et al. 1974).

We used 3D Slicer software (version 5.4.0 to 5.6.0; https://www.slicer.org, Fedorov et al. 2012) to reconstruct the CT data volumes. First, for better comparison between histology and CT, each volume was transformed to match the coronal view. Afterwards, the resampled TIFF image stacks were reconstructed to virtual 3D models by extracting bone surfaces in the Segment Editor tool. The histological sections were examined with a Leica DMLB photomicroscope connected to an AxioCam NRc5 Firewire digital camera, and with a Zeiss stereo microscope connected to an AxioCam 105 color Firewire digital camera. Photographed sections were exported as JPEG image files. The availability of histological serial sections in addition to the CT scans provided detailed information on soft tissues and especially the cartilage in some preadult specimens. The CT scans in turn facilitated the comparative examination of the overall shape of the developing turbinal skeleton within the nasal cavity based on the reconstructed virtual 3D models.

The terminology of the turbinals follows the scheme used, e.g., by Voit (1909), Reinbach (1952a, 1952b), and Maier (1993a, 1993b). Turbinals emerge on the nasal side wall (paries nasi) of the cartilaginous nasal capsule and are identified according to ontogenetic patterns which refer to the common template observed in placentals and marsupials. (1) The tripartite nasal capsule is divided into the pars anterior (housing the maxilloturbinal and the nasoturbinal), the pars posterior (housing ethmoturbinals, ETs, and interturbinals, ITs), and the pars lateralis (pars intermedia; housing frontoturbinals, FTs, and ITs) (Voit 1909; Reinbach 1952a; Smith and Rossie 2008). (2) Both ETs and FTs exhibit a species-specific number and are enumerated from rostral to caudal, starting with ET I and FT 1, respectively (Paulli 1900a, 1900b, 1900c; Macrini 2012; Ruf 2014; Wagner and Ruf 2021) (3) The ITs develop at later stages and seldom achieve the size and medial expansion like the enclosing ETs and FTs (Reinbach 1952a, 1952b; Zeller 1983; Smith and Rossie 2008). The presence of ITs varies within a species, and individually between the left and the right nasal fossa (Ruf 2014; Wagner and Ruf 2019, 2021). The IT between ET I and II is most prominent, and, in association with three ETs, regarded as part of the placental template (Schrenk 1989; Ruf 2014). (4) We follow the scheme of Reinbach (1952a, 1952b) and other authors that ET I is rostrally separated into two laminae, lamina anterior (LA) and lamina posterior (LP), since both fuse to each other and continue caudally into the cribriform plate as a uniform turbinal. Each of them elongates rostrally and forms an anterior process (Reinbach 1952a, 1952b; Wagner and Ruf 2021). If no caudally positioned lamina merges with the LA, the LP was considered as being absent. (5) In several placental groups including primates, the semicircular crest (LS) forms an uncinate process (PU) (Maier and Ruf 2014; Paulli 1900a, 1900b, 1900c).

In-text abbreviations: ET, ethmoturbinal; FT, frontoturbinal; IT, interturbinal; LA, lamina anterior (of ET I); LP, lamina posterior (of ET I); LS, lamina semicircularis; PU, processus uncinatus.

The raw data for the intranasal development were collected descriptively. To compare the growth patterns between individual morphological structures and across the age series, we transformed the text-based ontogenetic stages to distinct character states according to Sereno’s (2007) approach of trait recording (“phenotyping”). The definition of seven ontogenetic character states was adapted from developmental mechanisms described by, e.g., Smith et al. (2021a) (Table 2). For each specimen, the developmental stage for each structure (ETs, FTs, IT, LS, PU) was assigned one of the seven character states. The results were illustrated in schemes.

In Strepsirrhines, the first structures of the turbinal skeleton, which start to ossify are the two laminae of ET I (LA and LP), and the LS with its associated PU (Fig. S15). The rostrally positioned PU is the first structure to complete ossification in both L. catta and O. crassicaudatus, whereas in the LS, ET I, and all more posterior turbinals endochondral ossification is still proceeding. All turbinals which are present in the ossified ethmoid emerge from the paries nasi within the cartilaginous nasal capsule (Fig. 9), and have completely replaced the cartilaginous precursors before adulthood (Fig. S15).

Figure 9. 

Phylogenetic tree (see Fig. 1 for reference) of the examined primates, that is expanded by three fossil species (dashed lines, topology according to Lundeen and Kirk 2019 for Rooneyia, Kirk and Lundeen 2020 for Shoshonius, and Lundeen and Kay 2022 for Homunculus). The presence of the turbinals within the cartilaginous nasal capsule during early development and within the ossified nasal cavity at late development are shown (for details on the individual age stages see Figs S15, S16). Strepsirrhines exhibit the plesiomorphic pattern of three ethmoturbinals (ET I to III; up to ET IV in some L. catta specimens) and one interturbinal (IT) between ET I and II. The plesiomorphic number of frontoturbinals (FT 1 to FT 3) remains yet unknown. In contrast, all haplorhines are recognized by the absence of all FTs, the IT, and the posterior lamina (LP) of ET I, which all have been present in stem haplorhines represented by Rooneyia (Lundeen and Kirk 2019) and Shoshonius (Kirk and Lundeen 2020). Whether the damaged pars lateralis of the examined Shoshonius specimen housed FTs cannot be assessed (Kirk and Lundeen 2020). The stem platyrrhine Homunculus housed three ETs in its nasal cavity (Lundeen and Kay 2022), whereas among extant platyrrhines ET III exhibits a markedly heterochronic pattern. The most distinctive observation is the presence of a third ET in the cartilaginous template of Saguinus, which is absent in the ossified ethmoid, and on the other hand the absence of ET III in the nasal capsule of Aotus and Pithecia, whereas a bony ET III-like lamina was identified after ossification of the nasal region has completed. The presence of an uncinate process (PU) in the fossils is not evident from the literature. Abbreviation: ET I (LA), lamina anterior of ET I.

Among platyrrhines (Fig. S16), endochondral ossification is nearly complete in the neonate S. geoffroyi, whereas in the similar-aged A. nancymaae and P. pithecia all turbinals, the LS, and the PU are still partly cartilaginous. Most conspicuously, in one neonate P. pithecia (Saki3), ET II is even at the earliest state of an epithelial bulge. Nevertheless, ossification proceeds comparatively fast at least in A. nancymaae, because ET I to III, the LS, and the PU are completely osseous in the 14-days-old Aotus104. The most striking variation is the heterogeneous turbinal pattern between the cartilaginous template and the ossified ethmoid, namely the presence of a cartilaginous ET III in the mid-fetal S. geoffroyi, whereas this turbinal is entirely absent in all postnatal stages after ossification has begun (Fig. 9). In A. nancymaae and P. pithecia, the opposite is observed, namely the absence of ET III in the neonates, whereas it is present as bone in older stages.

In the two investigated catarrhine species ossification is complete at birth (M. mulatta YN09-175) or already before birth (P. anubis Papio107) (Fig. S16). Whereas the PU was not observed at the youngest Macaca stage by Maier (2000), the PU is the first structure undergoing endochondral ossification in the fetal P. anubis (Maier 2000). In all other Macaca age stages in this study, a bony PU is present. In contrast to platyrrhines, the number of turbinals remains constant throughout development (two ETs in Macaca spp. and P. anubis) (Fig. 9); except the presence of a bony ET III in the older infant Papio108.

The tripartite nasal capsule of the therian grundplan consists of the pars anterior (maxilloturbinal, nasoturbinal), the pars posterior (ETs, ITs), and the pars lateralis (FTs, ITs); the latter is rostrally separated by the LS, a posterior projection of the lateral wall of the pars anterior (Voit 1909; Reinbach 1952a; Smith and Rossie 2008; Martinez et al. 2024b). Primates are visual specialists, and the stereoscopic orbits which enable a depth perception are closely linked to the shape of the nasal capsule which in turn functions as a temporary template (“Stemmkörper”) for the enclosing dermal bones (Maier 1993b, 2000; Smith et al. 2014b). If we assume that the basal placentals’ pars posterior housed three ETs and an IT between ET I and II (Schrenk 1989; Ruf 2014), strepsirrhines conserved the plesiomorphic number (Fig. 9). It is also possible that the placental grundplan exhibited four ETs, in which case ET IV became lost in crown primates. Our assumed number of three to four ETs and an IT between ET I and II for the earliest primates agrees with Lundeen and Kirk (2019). However, the plesiomorphic number of FTs remains so far unresolved, but for basal Euarchontoglires (Fig. 1) two FTs are supposed to have been present in the pars lateralis (Maier and Ruf 2014). Lundeen and Kirk (2019) also assert that two FTs might be the primitive condition for Euarchonta, but they point to a possible homoplastic pattern with regard to the variable FT number within strepsirrhines. Their assertion is confirmed by our data and other studies, reporting a range from one (e.g., O. garnettii, present study) to three FTs (L. catta, present study; Daubentonia, Maier and Ruf 2014).

Strepsirrhines (Otolemur and Lemur, present study; Daubentonia, Maier and Ruf 2014) exhibit a distinctively higher number and complexity of turbinals compared to haplorhines, though the current descriptive data need to be confirmed by morphometric analyses. On the contrary, the great degree of orbital convergence in platyrrhines and catarrhines (Ross 1995) compresses the olfactory recess and the ETs. Further, a rostro-caudal compression is associated with the reduced midfacial length (Smith et al. 2014b). In both haplorhine groups, the constrained space coincides with (1) the reduction of the pars lateralis with an associated loss of all FTs; (2) the narrowing of the ethmoturbinal recess which forms the caudal end of the nasal cavity dorsal to the transverse lamina; and (3) the simplification and size reduction of ETs to narrow ridges (Maier 1993b; Smith et al. 2014b).

Individual turbinals are still difficult to homologize among placental taxa because a comprehensive developmental pattern across a vast species sample is still lacking (Maier 1993b; Werneburg et al. 2013). However, ET I is identifiable in all mammals (excepting those that have lost turbinals) by virtue of its relationship with the rest of the nasal capsule. Its LA overlaps the pars intermedia and juts rostrally as the antero-most element of the pars posterior (e.g., Lagomorpha, Ruf 2014; Canis, Wagner and Ruf 2019, 2021; Chiroptera, Ito et al. 2021; Eulipotyphla, Ito et al. 2022). The spatial relationship in all therians makes its homology exceedingly likely. Although perhaps less certain, the LP of ET I (ET II, for some authors) likewise has consistent spatial relationships across many mammals suggesting homology. This element shares a basal lamella, at least for part of the nasal cavity, with the LA (e.g., Paulli 1900a, 1900b, 1900c; Maier 1993a; Smith and Rossie 2008; Macrini et al. 2023). Strikingly, no such turbinal bears this relationship to the LA of ET I in any haplorhine studied here. The second lamina within the pars posterior does not fuse to ET I and merges with the cribriform plate as an independent turbinal. This pattern corresponds to the topology observed in ET II in the therian template (Paulli 1900a, 1900b, 1900c). We confirm Maier’s (2000) identification of the second lamina within the pars posterior as ET II, though he highlighted that it might possibly be the LP of ET I. This leads us to infer that the LP of ET I is lost in haplorhines, as suggest by others (e.g., Maier and Ruf 2014), and as apparently occurs in parallel in many bats (Ito et al. 2021). We hypothesize that the LP of ET I, being a far smaller element than the LA, is more prone to phylogenetic loss. This will require testing using a broader sample of mammals across developmental age, while taking phylogeny into account (e.g., Homunculus, see below). Similarly, the catarrhine template is lacking ET III, though it is present in one Papio specimen (see below). Our current study of primates suggests that reduced and simplified structures may also be properly recognized if their ontogeny is thoroughly investigated (turbinals, Maier 1993b; teeth, Hautier et al. 2016), even though the identity of lost elements is not always certain. Indeed, inferences about the loss, gain, or possible fusion vs. duplication of specific turbinals still remain largely unresolved (Smith and Rossie 2008; Van Valkenburgh et al. 2014).

Whereas within the pars posterior ET I, ET II and the IT in-between are quite easily distinguished, the terminology within the pars lateralis remains obscure. The investigated age series in L. catta did not provide any differentiation between FTs and ITs. As in the youngest available stage (fetal) all three turbinals exhibit a similar size, single-scrolled shape, and topology, we identified them as three FTs. In the infant however, ossification is most advanced in FT 2 that projects the most rostrally. FT 1 is at the least advanced stage of ossification. Based on the present sample, it is so far impossible to elucidate which FT in L. catta might be homologous to the single FT in O. garnettii. For instance, the three FTs remain simple single scrolls in the adult L. catta, whereas in O. garnettii the FT is more complex (double scroll). We adapted the scheme emphasized by, e.g., Smith and Rossie (2008) and simply counted the turbinals from medial (turbinal next to LS labeled as FT 1) to lateral/ventral in L. catta. Similarly, O. garnettii houses a single FT within its pars lateralis, designated as FT “1” based solely on position and not necessarily homology.

The pars posterior of the two fossil primates Shoshonius cooperi (early Eocene) and Rooneyia viejaensis (late middle Eocene) housed three ETs (including the LP of ET I) and an IT between ET I and II (Fig. 9). Whereas Rooneyia exhibited one FT, the presence of turbinals within the largely damaged pars lateralis of Shoshonius remains yet unknown. Though the number and the complex morphology of all turbinals was more similar to strepsirrhines, both species were regarded as stem haplorhines (Lundeen and Kirk 2019; Kirk and Lundeen 2020). The Miocene Homunculus patagonicus represents a stem platyrrhine which conserved three ETs, whereas the LP of ET I, the IT between ET I and II, and the pars lateralis were absent (Lundeen and Kay 2022) (Fig. 9). The turbinal surface area was larger compared to extant platyrrhines (Lundeen and Kay 2022), and the morphology in cross-section most closely resembled that of adult Aotus1 in our sample. Though all adult catarrhines and one platyrrhine (S. imperator) are lacking ET III, we observed an osseous ET III in an infant P. anubis (Catarrhini) and a cartilaginous ET III in a mid-fetal S. geoffroyi (Platyrrhini). On the other hand, in the two platyrrhines in which ET III is present in adults (P. pithecia, A. nancymaae) the neonate nasal capsule shows no evidence of a cartilaginous or ossifying third ET (Fig. 9). The heterogeneous presence of ET III could result from two different evolutionary scenarios:

(1) Phylogenetic constraints. ET III has become lost convergently in catarrhines and platyrrhines due to its presence in Homunculus (Lundeen and Kay 2022) and the young S. geoffroyi. In S. geoffroyi, the cartilaginous ET III gets resorbed instead of ossifying, which might represent a transitional state of undergoing turbinal loss. On the other hand, in A. nancymaae and P. pithecia ET III does not emerge from a cartilaginous precursor within the nasal capsule but instead develops through appositional outgrowth after the cartilaginous nasal sidewall has largely been resorbed. This implies a different developmental origin of ET III—and in P. pithecia additionally of ET II—and reveals an apomorphic condition compared to turbinals formed from the cartilaginous nasal capsule at an earlier ontogenetic stage. Our observation too might explain the smaller size and varying topology of ET III in A. nancymaae, namely its position dorsal to the transverse lamina, which develops at an earlier ontogenetic stage as result of the ossified lamina transversalis posterior (partly connecting with the ossified vomer, Smith et al. 2021b). Similarly, ET II and III are conspicuously smaller compared to the endochondrally ossified ET I (LA) in P. pithecia, even though they partly continue into the nasopharyngeal duct ventral to the transverse lamina.

The phenomenon of ET III appearing in some platyrrhines is striking. Since ET III is not present as a cartilaginous structure in newborns of A. nancymaae, nor is it an osseous lamella at birth, we infer that it is absent in the cartilaginous template. There are other examples of portions of turbinals forming without cartilaginous precursors. Smith et al. (2016) surmised that tertiary lamellae of maxilloturbinals in some strepsirrhines form as bony outgrowths from nonchondrally ossified turbinals. Whereas the appositional nature of these accessory lamella was an inference based on comparisons of infants to adults (Smith et al. 2016), our Aotus sample appears to form the entire ET “III” through progressive stages of development. At birth, no trace of either a cartilaginous or osseous ET “III” is seen. In an early infant stage, ET “III” appears to be only a bony ridge (Fig. 6E). In one of the older infants, ET “III” is a small, upwardly projecting bony spur, appearing to be a smaller precursor to ET “III” of the adult, an upwardly projecting bony fold (Fig. 6F, G). Since such bone forms as appositional outgrowth of existing endochondral bone, Smith et al. (2016) noted a similarity to the occurrence of membranous bone that emanates away from the cartilaginous cranial base during fetal development, called Zuwachsknochen in the German literature (e.g., Maier 1993a). Recently, DeLeon et al. (2022) suggested the term “ectochondral” bone for bones that ossify from endochondral bone before or after it ossifies. It is contextually distinct from intramembranous bone which forms earlier in development, with no precursor. Ectochondral bone, in contrast, is a means of elaboration of osseous structure at fetal or later stages of development. As such, ectochondral bone (a.k.a., Zuwachsknochen) may be an especially important means for generating novel osseous structures, or in this case, “reacquisition” of a third ET.

Unlike the third ET described above for some haplorhines, in strepsirrhines with presumably derived additional turbinals, these structures were intrinsically part of the cartilage template. For example, the suggested apomorphic third FT in L. catta emerges from a cartilaginous precursor. This observation may imply either that varying turbinal numbers are based on heterochronic developmental timing (early as cartilage, or late as bone) or that stem primates (or earlier lineages up to therians) housed three FTs in their pars lateralis.

(2) Intraspecific variation. Previous studies asserted that there is a species-specific number of ETs and FTs (e.g., Ruf 2014; Wagner and Ruf 2021). However, the use of different terminologies impedes a generalization of this observation. For instance, Smith et al. (2016) observed a variation of turbinal number within the pars lateralis in three L. catta specimens (three vs. four); they were identified as FT1 to FT3 and FT4, respectively. The authors emphasized that two turbinals were conspicuously larger (apparently called FT1 and FT2), indicating that the smaller turbinals were more likely ITs whose number varies between one and two. Similarly, Macrini (2012) observed two and three ectoturbinals within one Notoryctes skull; suggesting that, beside two FTs, the third ectoturbinal may be an IT that is missing on one side of the nasal fossa. Our own observation of a variable ET number within the nasal cavity of a juvenile L. catta (four in the right and three in the left nasal fossa) further refutes the theory of uniform ET – and similarly FT – numbers among placental species; indicating that ETs, at least, might be subject to intraspecific variation.

Despite the still ongoing debate about a vision-olfaction tradeoff, morphological reduction may characterize all haplorhines and even most strepsirrhines (though to a lesser degree in the latter group) (Smith and Bhatnagar 2004; Smith et al. 2004, 2014a, 2019; Melin et al. 2009; Lundeen and Kirk 2019). A relaxed or neutral selective pressure promotes variation (polymorphism) of a structure (Darwin 1859), in the present example the turbinal skeleton. Evidence for this assertion is provided by the CT scans of adult P. pithecia, in which one specimen exhibited one ET, the second two ETs, and the third two ETs on one side of the nasal fossa and three on the other side (unpublished data). A field study confirmed that P. pithecia prefers unripe fruits, which emit low levels of aldehydes and are hence primarily selected based on visual stimuli (Urbani 2002). In this context, it may be critical to realize that anthropoids (at least platyrrhines) distribute much olfactory epithelium on non-turbinal surfaces, such as the septum and “roof” of the nasal cavity (Smith et al. 2004). Some authors suggest numbers of turbinals are unreliable proxies for olfactory capabilities, at least in some mammals (e.g., Wagner and Ruf 2019, 2021; Martinez et al. 2023, 2024a, 2024b). This may be the case for anthropoid primates as well.

Beside the turbinal number, the reduced nasal cavity in haplorhines affected the morphology of the ETs, which either simplify or remain less complex single scrolls, up to double scrolls at most, in cross-sectional view. Since the turbinals in tree shrews (Scandentia) have a larger surface area compared to primates and colugos (Dermoptera), a simplification linked with a decreased surface area of the turbinal skeleton in Primatomorpha seems evident (Lundeen and Kirk 2019) (Fig. 1). However, to our best knowledge, hypotheses about the plesiomorphic turbinal shape in placentals are scarce, although Lundeen and Kirk (2019) inferred that the ancestral morphology of ethmoturbinals in primates is bullar.

Due to lack of certainty about homology, ETs and FTs are still simply counted according to their growth series from rostral (ET I and FT 1) to caudal throughout the species. A particularly challenging terminology refers to structures in the pars lateralis, because FTs and ITs are difficult to distinguish, especially when pre-adult stages are lacking.

The current sample neither confirms nor refutes the assumed number of two FTs in basal Euarchontoglires (Maier and Ruf 2014). Assuming that number is correct, in L. catta a third FT emerged, O. garnettii lost one FT. We need to enlarge the sample of strepsirrhine species and to include more ontogenetic series to receive reliable evidence of evolutionary changes within this lineage’s pars lateralis (e.g., which FT is lost or gained).

The use of morphological traits of the nasal cavity in phenotypic studies (e.g., associations with OR genes, Christmas et al. 2023) requires their homology across the species. As early developmental stages closely resemble the placental template, they enable the reliable identification of individual structures of the intranasal skeleton (Reinbach 1952a, 1952b). Provided that comprehensive developmental series are available, the growth patterns from these basal and mostly simple morphologies and topologies can be easily reconstructed up to their fully-grown and sometimes highly diverged adult stage (Zeller 1983, 1989).

We revealed the terminology of intranasal morphologies to establish a primate-specific template for recording traits according to Sereno’s (2007) phenotyping approach. A trait (i.e., a morphological structure) is either absent or present, and its expression is described based on qualitative (shape) or quantitative (e.g., size) data (Sereno 2007; Stefen et al. 2022). Our data strikingly revealed two challenges for the use of the phenotyping approach. First, phenotypic data ignores polymorphism (e.g., three and four ETs in the Lemur series) because it requires the assignment of one character state (i.e., three or four) to a trait (number of ETs). Second, different developmental origins might indicate that although turbinals share a similar topology in adults, they may be not necessarily homologous (in our present example ET “III” in platyrrhines to ET III in strepsirrhines and non-primate placentals). Both findings may bias the analyses of phenotypic data like their alignment with genotypes or their use in phylogenetic analyses. The application of phenotyping still requires its user to take a close look at the primary (i.e., descriptive, text-based) data, too, in order to entirely unravel possible contrary results. For instance, we strongly suggest that ET III has become lost independently in catarrhines and platyrrhines (Lundeen and Kay 2022), and that a convergent turbinal has evolved in some platyrrhine lineages as an apomorphic trait (Fig. 9). However, this may apply to the osseous turbinals; we cannot extend the argument to turbinals as mucosal folds. The early, likely programmed outgrowth of the turbinal as a soft tissue structure may well still be homologous.