Introduction

Caviomorpha represents an extremely diversified endemic South American group of Hystricognathi. The oldest known South American caviomorphs have been described from the late Eocene deposits of Peru (Frailey and Campbell 2004). Older South American localities, albeit numerous and fossiliferous, never documented any fossil rodents. The early phase of the rodent immigration and radiation in South America is therefore not well-understood. This absence of fossil rodents in sites older than the Peruvian locality and the isolation of South America during most of the Cenozoic have raised critical issues regarding the timing of arrival of rodents and their routes of colonization.

The current preferred hypothesis favors an African origin for caviomorphs from phiomorph Hystricognathi and suggests that their ancestors rafted or migrated from Africa to South America during Eocene times. Molecular and fossil analyses clearly indicate that the caviomorph South American lineages are closely related to African forms (e.g., Lavocat 1969; Parent 1980; Wyss et al. 1993; Martin 1994; Huchon and Douzery 2001; Poux et al. 2006). They both seem to share a common Asian hystricognathous ancestor (Marivaux et al. 2002).

The phiomorphs represent an adaptive radiation of hystricognathous rodents that occurred during the Paleogene in Africa (Wood 1968; Jaeger et al. 1985; Hartenberger 1985; Tabuce et al. 2001), of which the extant Thryonomyidae (cane rat), Petromuridae (dassie rat), and Bathyergidae (mole rat) are the descendants.

The Jebel El Qatrani Formation in the Fayum depression, Egypt, has yielded the richest collection of late Paleogene phiomorphs (Osborn 1908; Schlosser 1911; Wood 1968; Holroyd 1994; Sallam et al. 2009). Late Eocene and early Oligocene hystricognathous fossils have also been reported from Libya (Fejfar 1987; Jaeger et al. 2010), from Oman (Thomas et al. 1989) and from Kenya (Ducrocq et al. 2010). However, the fossil record has not allowed clear identification of an African representative of Hystricognathi as the ancestor of the Caviomorpha.

Here we describe a new species of the hystricognathous rodent genus Gaudeamus from a newly discovered locality from the lower Oligocene of Zallah (Libya). This species displays a tetralophodont and taeniodont dental pattern which is nearly identical in its morphology to some of the earliest known South American caviomorphs recently described from the ?Eocene of Santa Rosa, Peru (Frailey and Campbell 2004). This discovery bears implications on molar crest homologies and allows discussing the nature of the primitive molar pattern of the caviomorph ancestors. It also brings new insights to the phylogenetic and paleobiogeographic debates on the early history of caviomorphs.

Systematic paleontology

Order Rodentia Bowdich 1821

Suborder Ctenohystrica Huchon, Catzelfis, and Douzery 2000

Infraorder Hystricognathi Tullberg 1899

Genus Gaudeamus Wood 1968

Type species Gaudeamus aegyptius Wood 1968

Gaudeamus lavocati sp. nov.

Material

Holotype: ZR5-69, right lower jaw fragment with M1–M3 (Fig. 1a)

Fig. 1
figure 1

Cheek teeth of Gaudeamus lavocati sp. nov. a ZR5-69, holotype, right lower jaw with M1–M3. b Z5R-133, right dP4. c Z5R-136, left P4. d Z5R-121, right M3. e Z5R-90, right M2. f Z5R-70, right M1. g Z5R-139, left dP4. h Z5R-144, left P4

Full size image

Referred material: 159 isolated teeth (see Online Resource 1 for a complete list of specimens referred to G. lavocati).

The material is temporarily deposited in the collections of iPHEP, University of Poitiers, and it will be deposited after publication in the Collections of El Fateh University in Tripoli (Libya).

Measurements: see Table S1 in the Online Resource 1.

Type locality

G. lavocati derives from Zallah 5 locality, a newly discovered micro-mammal locality in the Oligocene deltaic-continental deposits of Sirt basin, located near the oasis of Zallah (Central Libya). Several large and small mammal localities have been mentioned previously (Arnoult-Saget and Magnier 1961; Arambourg 1963; Fejfar 1987) from this area. Biostratigraphic evidences suggest an early Oligocene age for Zallah deposits.

Derivatio nominis

In honor of the late R. Lavocat who suggested and supported the idea of an African origin of South American caviomorph rodents.

Diagnosis

Species of Gaudeamus (M2 of the holotype = 2.19 × 2.14 mm) of larger size than G. aegyptius Wood 1968. It differs from the latter by its tetralophodont molar pattern (Fig. 1). Lower molars with a mesolophid which delimits an extra basin on the posterior side of the metaconid. Lack of connection between the hypolophid and the hypoconid. Mesoloph backwardly directed and connected to the posteroloph on upper molars. Development of a labial (lingual) wall in upper (lower) worn molars. Marked unilateral hypsodonty on upper cheek teeth. Three-rooted molars, except for a two-rooted M3. dP4 elongated, pentalophodont, molariform, differing from that of G. aegyptius by more oblique lophids.

G. lavocati is slightly larger in size than species of Eoincamys and differs from them in some respects: less oblique molar crests, absence of connection between the mesoloph and the anterior arm of the hypocone on upper molars, connection of the mesolophid with the metalophulid I rather than with the ectolophid and lack of hypoconid/hypolophid connection on lower molars, and dP4 with more oblique lophids.

(For detailed description of G. lavocati cheek teeth, see Online Resource 1)

Comparison

Occlusal dental pattern

The genus Gaudeamus was initially erected by A.E. Wood (1968) from the lower Oligocene of Fayum (Egypt). By its semi-hypsodonty, its great differentiation of transverse crests and its simplified occlusal molar pattern, Gaudeamus was considered as an evolutionarily advanced species compared to other Oligocene phiomyids (Wood 1968; Holroyd 1994).

G. aegyptius and G. lavocati exhibit a common pattern comprising three principal lophs in the molars: an anteroloph, a transverse loph connecting the paracone to the hypocone, and a posteroloph on upper molars (a metalophulid I, a continuous loph connecting the protoconid to the entoconid, and a posterolophid on lowers). In both species, the posteroloph runs from the hypocone to the metacone and the metaloph is fused with the posteroloph. However, G. lavocati displays, in comparison to G. aegyptius, a more lophodont pattern with a greater differentiation of transverse crests. The obliquity of the crests in its dP4 is more pronounced than in G. aegyptius (Fig. 2d, i). But the main difference between the two taxa lies in the presence of a supplementary fourth crest in G. lavocati molars. From the medial part of the labial margin of upper molars, a short crest extends disto-lingually toward the middle of the posteroloph. This crest has been previously described as a neoloph by Holroyd (1994) who mentioned a yet undescribed new tetralophodont species of Gaudeamus from the Fayum L-41 quarry (Holroyd 1994; Sallam et al. 2009). This crest is interpreted here as a reduced mesoloph because of its homology in crest position and on the basis of the presence in G. aegyptius of a minute mesostyle from which the mesoloph is likely to develop. It corresponds to a modification of the ancestral pentalophodont structure otherwise present in phiomyid rodents rather than to a new structure.

Fig. 2
figure 2

Sketches of lower deciduous premolars patterns of Santa Rosa rodents (white area) and Old World hystricognathous rodents (light gray area). Not to scale. a Eobranisamys romeropittmanae, LACM 143352. b Eobranisamys riverai, LACM 143371 (reversed drawing). c Eoespina woodi, LACM 143405 (reversed drawing). d Gaudeamus aegyptius, YPM 18024 (reversed drawing; Wood 1968). e Phiomys andrewsi, AMNH 13271 (Wood 1968). f Eobranisamys riverai, LACM 143369. g Eosallamys paulacoutoi, LACM 143420 (reversed drawing). h Eosallamys palacoutoi, LACM 143451. i Gaudeamus lavocati, Z5R-139 (reversed drawing). j Protophiomys durattalhahensis, DT-2-007 (Jaeger et al. 2010). k Eosallamys simpsoni, LACM 143452. l Eoespina woodi, LACM 143391 (reversed drawing). m Hystrix depereti, PR-147 (Sen 2001). n Phiomys hammudai, DT-1-011 (Jaeger et al. 2010). After Frailey and Campbell 2004, completed

Full size image

When considering the earliest known South American members of caviomorphs, it appears that the dental morphology of G. lavocati sp. nov. is very close to that of Eoincamys from the ?Eocene Santa Rosa fauna of Peru (Fig. 3c–h). Both display moderately high-crowned and unilaterally hypsodont cheek teeth and a taeniodont tetralophodont molar pattern. Frailey and Campbell (2004) nevertheless interpreted the third crest of Eoincamys molar as a metaloph/metalophid. Their identification implies a forward displacement of the metacone and raises questions concerning the nature of the cusp (if it is not a metacone) lying at the labial extremity of the posteroloph. We instead interpret this third loph as homologous to the mesoloph. From this point of view, Eoincamys dental pattern appears to be very close to that of G. lavocati but less derived. The mesoloph is less oblique, its lingual tip tends to join the anterior arm of a more distinct hypocone instead of being confluent with the posteroloph (Fig. 3e–h). Remnants of the former ectolophid are still distinguishable in Eoincamys. A raised portion in the posterior sinusid tends to connect the hypoconid to the hypolophid in advanced stage of wear. Deciduous lower premolars of Santa Rosa rodent taxa display slight variation of a pentalophodont theme very similar to that of Gaudeamus (Fig. 2). These taxa share in common an occlusal pattern composed of an anterolophid, a metalophulid II, a mesolophid, a hypolophid, and a posterolophid. Differences between some Santa Rosan genera and Gaudeamus in the lower deciduous premolar are small and consist mainly of different degrees in inclination of the crests.

Fig. 3
figure 3

Comparative upper (left) and lower (right) molar dental pattern of some Santa Rosan genera and some Eocene/Oligocene African Hystricognathi. a Gaudeamus aegyptius, YPM 18044, right M2 (reversed drawing). b CM 26920, left M2 (Wood 1968). c Gaudeamus lavocati, Z5R-98, right M2. d ZR5-69, right M2 (reversed drawings). e Eoincamys pascuali, 143319, left Mx. f 143306, left M x (Frailey and Campbell 2004). g Eoincamys ameghinoi, 143339, left Mx. h 149435, left Mx. (Frailey and Campbell 2004). i Eobranisamys romeropittmanae, 143362, right Mx (reversed drawing). j 143346, left M x (Frailey and Campbell 2004). k Eosallamys simpsoni, 143374, left Mx. l 143276, left M x (Frailey and Campbell 2004). m Phiomys hammudai, DT-1-003, right M1. n DT-1-006, right M2 (reversed drawing; Jaeger et al. 2010)

Full size image

The comparison of the newly discovered primitive forms of caviomorph teeth from South America which include pentalophodont and tetralophodont forms, to those of the African phiomorphs, provides valuable information on cheek teeth crest homology (Fig. 3). Phiomorphs and caviomorphs are widely recognized as related taxa (e.g., Luckett and Hartenberger 1993; Catzeflis et al. 1995; Nedbal et al. 1996; Huchon and Douzery 2001; Marivaux et al. 2002) but terminology used to describe their cheek teeth has always differed among authors in important aspects. These problems of dental structure recognition in earliest hystricognathous rodents have brought on confusion regarding their phylogenetic affinities (e.g., Wood and Patterson 1959; Hoffstetter and Lavocat 1970; Patterson and Wood 1982). The central point of the differing interpretations in caviomorph dental morphology has been the nature of the third and fourth lophs of the tetralophodont and pentalophodont forms (variously metaloph, mesoloph, or neoloph depending on the authors). Hoffstetter and Lavocat (1970) interpreted the fourth loph of pentalophodont caviomorphs (Branisamys) as a rudimentary metaloph on the anterior face of the posteroloph. By contrast, Patterson and Wood (1982) considered Branisamys molar pattern as a basic tetralophodont form with an additional small loph (a neoloph) that projects from the posteroloph. The loph described as a mesoloph by Hoffstetter and Lavocat (1970) is considered by them as a metaloph. The key to identify the homology of these dental structures seems to lie in the set of morphological stages displayed by earliest caviomorphs. The pentalophodont pattern, representing the plesiomorphic molar condition among Eocene Hystricognathi (Hoffstetter and Lavocat 1970; Bryant and McKenna 1995; Marivaux et al. 2002, 2004), shows an anteroloph, a protoloph, a mesoloph, a metaloph, and a posteroloph (Fig. 4). It is clearly expressed in Phiomys andrewsi or Phiomys hammudai (Fig. 3m). In caviomorph pentalophodont forms (Eobranisamys, Branisamys), the metaloph displays a derived state (Marivaux et al. 2004) being short and directed back to join the posteroloph (Fig. 3k, i). The metaloph, as in their close African relatives, show tendencies to fuse with the posteroloph. The simplification of the basic pentalophodont molar condition leads to a tetralophodont pattern by the loss of the metaloph by fusion with the posteroloph (Fig. 3e, g). The metaloph and the mesoloph are progressively confined to the postero-labial corner of the tooth and removed as the degree of unilateral hypsodonty increases. As demonstrated by Jaeger (1989), unilateral hypsodonty developed in caviomorph taxa results in decreasing the size of the postero-labial corner of the tooth as wear proceeds. The lingual features appear to shift toward the labial side of the tooth and are gradually wiped out. In G. aegyptius (Fig. 3a), and also in other African taxa like Paraulacodus or Neosciuromys, the loss of both metaloph and mesoloph finally leads to a trilophodont state.

Fig. 4
figure 4

Occlusal dental nomenclature (after Marivaux et al. 2004). a Upper molar: 1 protocone, 2 paracone, 3 metaconule, 4 mesostyle, 5 metacone, 6 hypocone, a anteroloph, b neo-endoloph, c protoloph, d mure, e posterior arm of paracone, f mesolophule, g mesoloph, h anterior arm of metacone, i metaloph, j anterior arm of hypocone, k posteroloph. b Lower molar: 7 protoconid, 8 metaconid, 9 mesostylid, 10 mesoconid, 11 entoconid, 12 hypoconid, 13 hypoconulid, l metalophulid I, m posterior arm of metaconid, n posterior arm of protoconid, o ectolophid, p mesolophid, q anterior arm of entoconid, r hypolophid, s anterior arm of hypoconid, t postelorophid

Full size image

These morphological stages, related to different degree of semi-hypsodonty, allows identification of the homology of dental crest in earliest caviomorphs, shedding new lights on crest homologies among hystricognath rodents.

Incisor enamel microstructure

Both lower and upper incisors of G. lavocati display double layered enamel with multiserial Hunter–Schreger bands (HSB) and a clearly distinct portio externa. This multiserial enamel is characterized by a slightly angular and anastomosing interprismatic matrix (IPM; see Online Resource 2).

The incisor enamel microstructure of G. lavocati is typical for hystricognath rodents (Martin 1993). In addition, G. lavocati displays a primitive subtype of multiserial HSB (defined by Martin 2004), with an IPM running at a slight angle to the prism long axes and slightly anastomosing between the prisms in the cross sections in both upper and lower incisors. Such a condition is congruent with the oldest enamel microstructures described for the Paleogene South American Caviomorpha. The caviomorph incisors discovered in the Late Eocene (or earliest Oligocene ?) locality of Santa Rosa indeed display an important diversity, with three subtypes of multiserial HSB (Martin 2004) including the primitive condition observed in G. lavocati. The locality of Santa Rosa delivered the two known species of the genus Eoincamys, but up to now the incisor material from there has not been attributed to taxa, therefore no direct comparison between the enamel microstructures of G. lavocati and Eoincamys is possible.

Phylogenetic analysis

The new information presented here enables a fuller test of the hypothesis of close relationship between caviomorphs and phiomorphs and an exploration of the affinities of Gaudeamus. A cladistic analysis including 65 dental characters and 37 taxa has been performed to assess these hypotheses (see Online Resource 1 for a complete character list and data matrix for all included taxa). Outgroups selected to study relationships within Hystricognathi are extinct members of the family Ctenodactylidae, represented by Birbalomys woodi, and of the Yuomyidae, represented by Yuomys cavioides. Ingroup taxa include 35 species which represent the majority of extinct Eocene–Oligocene hystricognathous rodents and the extant taxon Thryonomys. We mostly focused on the African Eocene–Oligocene phiomorphs (16 species, see Online Resource 1 for a complete list of the sampled species) and the Eocene–Oligocene South American caviomorphs (ten species). Earliest Asiatic members of Hystricognathi, the Baluchimyines, have also been included. In order to investigate the phylogenetic relationship recently assumed between Gaudeamus and Hystricidae (Sallam et al. 2009), three Old World porcupines (Hystrix primigenia, Hystrix depereti, and Hystrix cristata) were added to our analysis. The analysis was performed with PAUP (v.4.0b10, Swofford 1998) using heuristic search methods (1,000 replications with random stepwise addition and a randomized input order of taxa). Parsimony analyses result in one most parsimonious tree of 249 steps (Fig. 5; CI = 0.36; RI = 0.68).

Fig. 5
figure 5

Most parsimonious tree showing the temporal range and geographical extent of the 37 hystricognath taxa included in this study. Bremer support values (Bremer 1988), calculated using TNT (Goloboff et al. 2003), are labeled at each node

Full size image

The most significant aspect of the analysis is the nesting of Gaudeamus within the clade that groups all the oldest known South American caviomorphs. Caviomorpha clade including Gaudeamus is supported by the following unambiguous synapomorphies: crested pattern, cusps indistinct or absent (ch. 1), high-crowned cheek teeth (ch. 3), the absence of the mesolophule on upper molars (ch. 33), the absence or indistinctness of the anteroconid on lower deciduous premolars (ch. 37), the indistinctness of the hypoconulid on lower deciduous premolars (ch. 44), the absence of the hypoconulid on lower premolars (ch. 53), and the metaloph fused with the posteroloph on upper premolars (ch. 65; see Fig. S1 in the Online Resource 1).

The Caviomorpha present two basal divergent clades. The first clade is characterized by a lack or incomplete taeniodont pattern (ch. 2) and includes most of the Paleogene South American taxa described as primitive Octodontoidea. The second clade is characterized by fully taeniodont molars and includes the Paleogene South American taxa described as primitive Cavioidea plus the genera Sallamys and Platypittamys.

The Hystricidae constitute the sister taxon to the clade that clusters Gaudeamus and the caviomorphs. The immediately more inclusive clade shows the African species P. andrewsi and P. hammudai as the sister taxa to the clade Hystricidae + Caviomorpha The immediate sister taxon is a clade consisting of a Phiomorpha subset (Paraulacodus, Paraphiomys, Metaphiomys, Phiomys paraphiomyoides) including the extant Thryonomys. The clade Phiomorpha defined by Sallam et al. (2009) appears to be therefore paraphyletic in our analysis. The remaining taxa traditionally attributed to Phiomorpha are variously nested along with baluchimyines at the base of the Hystricognathi.

The Hystricidae are supported as the sister taxon of the Caviomorpha by our analysis. This result is not in agreement with the chronobiogeographically and backbone constrained phylogenetic analyses which support Gaudeamus as a sister taxon of Hystricidae, outside the Caviomorpha (Sallam et al. 2009; it is noteworthy however that these authors found both Gaudeamus and Hystricidae nested among Caviomorpha in various locations in their unconstrained morphological analysis). The Hystricidae are generally classified as a divergent and basal hystricognath lineage (Adkins et al. 2001, 2003; Huchon et al. 1999; Blanga-Kanfi et al. 2009); however, parasitology (Hugot 2002) and a subset of a molecular-based study (Huchon and Douzery 2001) support them as the sister taxa of Caviomorpha. Up to now, hystricid fossils are actually only recognized since middle Miocene (Flynn et al. 1998) and molecular studies project a ghost lineage originating in late Eocene (Huchon and Douzery 2001). Actually, the lack of Paleogene hystricids likely clouds their phylogenetic relationships among Hystricognathi. Concerning the geographic origin of Hystrix, their earliest occurrences are recorded from Central Europe and Spain (Van Weers and Montoya 1996). However, paleontologists hypothesized that hystricids have an Asian or Indian origin due to their extant diversity in these regions (Hussain et al. 1978; Winkler 1994).

Discussion

Tetralophodont and trilophodont molars in hystricognathous rodents correspond to derived structures evolved from an ancestral basic pentalophodont pattern, preserved in some Eocene and Oligocene African Phiomorpha (Hoffstetter and Lavocat 1970; Bryant and McKenna 1995; Marivaux et al. 2002), following two distinct pathways. On the first hand, the tetralophodont pattern is achieved in some Phiomorpha (e.g., Metaphiomys, Paraphiomys) by the reduction of the mesolophule (ch. 33 state 1) combined with the absence of the mesoloph (ch. 31 state 0). On the other hand, the tetralophodont pattern in South American caviomorphs is obtained with the merging of the metaloph and the posteroloph (ch. 34, state 3).

G. lavocati sp. nov. is the only African taxon displaying the latter tetralophodont condition currently known. Gaudeamus sp. nov. from Quarry L-41 (Fayum depression, Egypt) may also present such a pattern (Sallam et al. 2009) but is not described yet, while G. aegyptius displays a trilophodont pattern associated with the loss of mesoloph/id.

Gaudeamus was classically referred as a phiomorph and its type species, G. aegyptius, was grouped into the Phiomyidae family (Wood 1968) or into the Thryonomyidae family (Lavocat 1978). Phylogenetic analyses of dental and enamel microstructure characters found Gaudeamus to be nested among Caviomorpha (this study and unconstrained analysis in Sallam et al. 2009). This genus exhibits features present in the South American taxa and appears to be more closely related to these latter than to the African Phiomyidae. The nesting of this African taxon within the caviomorphs raises problems concerning its taxonomic assignment. At present, we prefer not to allocate it to a precise family, pending the revision of Phiomyidae which constitute a paraphyletic group.

According to molecular data, the arrival of caviomorphs in South America is believed to have taken place during the middle-late Eocene (Poux et al. 2006). The migration scheme to explain the current distributional pattern of hystricognathous rodents has been extensively discussed and various paleobiogeographic scenarii have been proposed for the caviomorph origin. A trans-Atlantic route is, until now, the prevailing hypothesis despite the fact that South America and Africa were already separated by the Atlantic Ocean barrier. Colonization by rafting would have been aided by marine paleocurrents, paleowinds, or “stepping stone” islands (e.g., Wyss et al. 1993; Houle 1999). Secondly, a migration route via Antarctica, which was connected to South America until late Eocene (Barker et al. 1991), has been pointed out (e.g., Houle 1999). This southern migration route seems however unlikely because of major water barriers separating Antarctica and Australia since the Early Eocene (Woodburne and Zinsmeister 1984) and because no relevant rodent remains have ever been reported from these continents. The alternative possible migration route, via the Bering Strait and North America (e.g., Hussain et al. 1978), is disregarded because of the obvious lack of hystricognath fossils in North America.

The fact that Paleogene South American and African hystricognathous rodents share a common Asian hystricognathous ancestor (Flynn et al. 1986; Bryant and McKenna 1995; Marivaux et al. 2002) has raised the question of whether caviomorphs are descended from an Asian group of hystricognaths and has opened the possibility of a dispersal of hystricognathous rodents to South America from Asia (Hussain et al. 1978; Marivaux et al. 2002). However, data of multiple sources support the hypothesis of a Paleogene African origin of caviomorphs (Lavocat 1969; Parent 1980; Wyss et al. 1993; Martin 1994; Houle 1999; Huchon and Douzery 2001; Poux et al. 2006).

Considering this hypothesis, it is noteworthy that in our analysis, some African Phiomorphs constitute the immediate outgroup to the Hystricidae + Caviomorpha clade. The dental morphology exhibited by some Phiomys species (P. hammudai and P. andrewsi) would identify them as good African candidates to be stem representatives of the Gaudeamus lineage and all other primitive caviomorphs. The earliest caviomorph might have arose on the African continent before Oligocene times and Gaudeamus ancestors may belong to the African group that migrated to South America during middle-late Eocene. Although an African origin for caviomorphs has been advocated in the past, ancestor group is lacking from the fossil record of that continent. Additional new fossil findings in middle and late Eocene deposits of African caviomorphs are required to test our hypothesis concerning the early paleobiogeographical history of that group. The high morphological similarities in dental characters and incisor enamel microstructure of Gaudeamus and some contemporaneous caviomorph taxa may indicate that the beginning of caviomorphs radiation occurred in Africa within the time period of Phiomyid radiation. This reinforces the hypothesis of a direct migration route between Africa and South America during Eocene.

However, there are still no evidences of migration corridors between Africa toward South America and the dental pattern of Gaudeamus could also illustrate the result of a parallel evolution of an African phiomorph lineage towards the specialized pattern of some caviomorphs. Evolution of Gaudeamus parallel to tetralophodont Caviomorpha would constitute an exceptional case of synchronized convergence, given the very close similarity between Eoincamys pascuali and G. lavocati. Such a convergent case of dental morphology has already been documented among Murid rodents, with the great similarity observed between extant Deomyinae and Murinae (e.g., Chevret et al. 1993; Jansa and Weskler 2004), but in this last case there is no evidence of synchronized convergence.

The latter hypothesis, also based on the implications of our cladistic analysis, of a South American caviomorph migration into Africa is hardly likely considering that the possibility of migration were rather limited.

At present, dental and enamel microstructure morphologies invariably indicate phylogenetic affinities between Gaudeamus and the earliest caviomorphs (this study and unconstrained analysis of Sallam et al. 2009) and Gaudeamus can be considered, tentatively, as an African caviomorph. New fossils from slightly older localities, both in Africa and South America, may shed more light on the exact affinities of Gaudeamus within Paleogene Hystricognathi.