Granastrapotherium Johnson & Madden, 1997

Goillot, Cyrielle, Antoine, Pierre-Olivier, Tejada, Julia, Pujos, François & Gismondi, Rodolfo Salas, 2011, Middle Miocene Uruguaytheriinae (Mammalia, Astrapotheria) from Peruvian Amazonia and a review of the astrapotheriid fossil record in northern South America, Geodiversitas 33 (2), pp. 331-345 : 338-342

Genus Granastrapotherium Johnson & Madden, 1997

TYPE AND ONLY SPECIES. — Granastrapotherium snorki Johnson & Madden, 1997 from the middle Miocene of Colombia, by original designation.

and its estimated dimensions once r econstructed ( Fig. 3C ) are 70 mm (ectoloph MDL), 60 mm (anterior LLL), and 38 mm (posterior LLL).

MUSM 1477 is the posterior part of the ectoloph of an M3, with a sharp and oblique occlusal edge. The enamel is finely striated vertically. The preserved height reaches 73 mm.

Granastrapotherium cf. snorki COMPARISON

( Fig. 3 ) MUSM 994 is “∏-shaped” (two transverse lophs and an ectoloph) and brachydont, and it displays

Granastrapotherium snorki (partim) – Antoine et al. a mesial valley (anterior constriction of the proto- 2007: 21.

cone), which allows us to interpret it as an upper

Granastrapotherium snorki – Negri et al. 2010: 247. molar of an astrapothere, rather than a to xodont tooth ( Fig. 3 ). The protocone is large and conical

MATERIAL EXAMINED. — Left br oken M2, without and the lingual valley is deep; despite the advanced ectoloph ( MUSM 994), IN-DTC 28 locality (float); fragmentary ectometaloph of a left M3 ( MUSM 1477), stage of wear, its sagittal elongation tends to indicate IN-DTC-32 locality. MUSM 994 is not an M1, while the pr esence of a conspicuous metaloph-hypocone cr est impedes

LOCALITY AND DISTRIBUTION. — Río Mapuya, 14 km identifying it as a M3 either. On the other hand, upstream the confluence with Río Inuya ( IN-DTC 28) and 7 km upstream the latter locality ( IN-DT C-32), the general shape of the tooth fits perfectly M2s of Atalaya Department, Ucayali, Peru, and Río Inuya. uruguaytheriine astrapotheres ( Johnson & Madden

1997). There is no lingual cingulum, contrarily to

FORMATION AND AGE. — Ipururo Formation, late mid- what is observed in all astrapotheriids, including dle Miocene, c. 13 Ma ( Antoine et al. 2007; Espurt et al. 2010). Xenastrapotherium ( Johnson & Madden 1997;

Kramarz & Bond 2009). Together with the large

DESCRIPTION size of the tooth (c. 70 × 60 mm, i.e. ex ceeding MUSM 994 is a worn and br oken tooth, lacking by 10% the largest M2 referred to X. kraglievichi , its labial part and its roots. The crown as a whole, but coinciding with the mean value for G. snorki ; except where worn or br oken, is surrounded by Johnson & Madden 1997: tables 22.1; 22.4), the finely wrinkled enamel ( Fig. 3A ). The complete absence of lingual cingulum strongly supports the crown is elongate sagittally. In occlusal view, the assignation of MUSM 994 to Granastrapotherium . labial third of the crown is broken; the protoloph is Accordingly, due to its very large dimensions (c. 30% nearly parallel to the preserved part of the ectoloph; larger than M3s of X. kraglievichi ; Johnson & Madthe anterior and posterior sides make an angle of den 1997; pers. obs. CG 2008), the fragmentar y c. 45° converging lingually; the protocone and the ectometaloph MUSM 1477 is likely to document labial part of the pr otoloph are very large at this the same large-sized taxon, tentatively referred to advanced wear stage; the protocone is constricted as Granastrapotherium cf. snorki . anteriorly by a large valley and it shows a lingual groove ( Fig. 3B ); a bulge is observed on the lingual side of the ectoloph; a large median v alley opens DISCUSSION lingually; a thick hypocone joins the ectoloph, thus forming a metaloph, separate from the ectoloph by a AGE OF THE FITZCARRALD LOCAL FAUNA, sharp groove; no lingual cingulum can be observed; EASTERN PERU thickness of the enamel varies from 0.5 mm on the The specimens described in the present work sublingual side of the ectoloph to 1.7 mm on the ante- stantiate the uruguaytheriine astrapotheriid record rior side of the protocone. The tooth is 60 mm long in Amazonian Peru, primarily based on float specimens from the Río Inuya/Mapuya area mentioned by Willard (1966) and subsequently referred to “cf. Granastrapotherium ” by Johnson & Madden (1997). The new remains document two uruguaytheriine taxa: Xenastrapotherium sp. and Granastrapotherium cf. snorki . In northern South America, bispecific astrapotheriid assemblages were so far restricted to the early Miocene of Venezuela (Castillo Formation; Sánchez-Villagra et al. 2004), and the late middle Miocene of Colombia (Honda Group; Johnson & Madden 1997).

Such co-occurrence recalls unequivocally the Xenastrapotherium kraglievichi-Granastrapotherium snorki assemblage, which characterizes the late middle Miocene “ Miocochilius Assemblage Zone ” of La Venta, Colombia ( Johnson & Madden 1997; Madden et al. 1997). By the way, the Fitzcarrald local fauna also includes other biostratigraphical markers of the Lav entan SALMA (13.5-11.8 Ma), such as the small interatheriid notoungulate Miocochilius anomopodus Stirton, 1953 , the large toxodontid notoungulate Pericotoxodon platignathus Madden, 1997 , and the dinomyid rodents “‘ Olenopsis ’ sp., large” sensu Walton 1997 and “ Scleromys ” schurmanni sensu Walton 1997 ( Antoine et al. 2007). This assemblage as a whole fur ther points to the 13.6-12.76 Ma interval in the Honda Group of Colombia (C5ABn-C5Ar.2r chrons; Madden et al. 1997: fi g. 29.6; Gradstein et al. 2005: 68, table 5.2). To a lesser extent, the Fitzcarrald local fauna resembles the Quebrada Honda assemblage of southern Bolivia, which notably yields another species of Miocochilius , M. federicoi Croft, 2007 , and a large uruguaytheriine astrapotheriid ( Antoine et al. 2007; Croft 2007).

THE ASTRAPOTHERIID FOSSIL RECORD

IN NORTHERN SOUTH AMERICA

The spatio-temporal distribution of uruguaytheriine astrapotheres is illustrated in the F igure 4 and further detailed in the section below. In high- and middle-latitudes (> 30°S; mostly fr om Argentina and Chile), the last astrapotheriids ar e reported during the “Friasian” stage (e.g., Pascual & Odreman Rivas 1971). This informal stage represents a short interval intercalated between the late early Miocene Santacrucian SALMA and the early middle Miocene Colloncuran SALMA ( Flynn et al. 2002, 2008; Croft 2007; Forasiepi et al. 2009; Kramarz & Bond 2009). Therefore, uruguaytheriines as conspicuous elements of the mammalian faunas ar e the last well-documented representatives of Astrapotheria , throughout the late middle M iocene (Laventan SALMA; Johnson & Madden 1997). They might persist until middle late Miocene times (Huayquerian SALMA) inVenezuela and Amazonian Brazil, with fragmentary remains only identified at family level and likely to be r eworked or misidentified ( Frailey 1986; Linares 2004; Carlos Jaramillo pers. com. 2008).

Xenastrapotherium is widely distributed, both stratigraphically (ranging at least fr om the early Miocene to the late middle M iocene) and geographically ( Venezuela, Colombia, Ecuador, Brazil, and Peru; Fig. 4 ). Xenastrapotherium chaparralensis , from the Chaparral local fauna of Colombia (Tuné Formation, Tolima Department), was formerly considered as the earliest uruguaytheriine, dating back from the Deseadan ( Stirton 1953). The associated mammalian assemblage suggests this fauna is rather Colhuehuapian or early Santacrucian in age (early Miocene; Johnson & Madden 1997: 365).

A single mandible described b y Stehlin (1928) documents X.christi . It originates from the “Cucharo Formation”, near Zaraza, northeastern Guárico State, Venezuela ( Stehlin 1928; Simpson 1940). This formation was later synonymiz ed with the Chaguaramas Formation, which spans the late Oligocene-early Miocene interval ( Isea 1987). As such, X. christi might be the earliest uruguaytheriine with Uruguaytherium , for which an Oligocene age cannot be discarded either (see hereafter; Simpson 1940).

Xenastrapotherium aequatorialis is also kno wn only by a partial right mandible with m1-m2, “reported to have been collected along the banks of the Río Burgay [...] 1 km south of Biblián, Province of Cañar, Ecuador ” ( Johnson & Madden 1997: 362). Several K-Ar datings on andesitic lavas at the top of the Biblián Formation, which crops out in that area, constrain an early Miocene age for this mandible (>19.7 ± 0.5 Ma; for review see Lavenu et al. 1995).

By contrast, the mandibular and dental morphology of X. kraglievichi is well documented, with about three dozens referred specimens from various localities of theVillavieja Formation, La Venta area,

Colombia ( Johnson & Madden 1997). Following the same authors and Madden et al. (1997: 508, fig. 29.5), a right M3 originating from the uppermost vertebrate locality of the underlying La Victoria Formation, in the same ar ea, is also r eferable to X. kraglievichi . The corresponding stratigraphical interval is magnetostratigraphically constrained, and ranging from 13.183 up to 12.73 Ma ( Madden et al. 1997; Gradstein et al. 2005).

Xenastrapotherium amazonense is known by a maxillary fragment with M3 fr om Pedra Pintada locality ( Paula Couto 1976) and several teeth from Cachoeira locality (= Torre Da Lua; Rancy 1981) found on banks of the Alto Río Juruá, in Amazonian Brazil ( Paula Couto 1976, 1982; Johnson & Madden 1997; Negri et al. 2010). Several fragmentary teeth originating from surrounding localities were either referred to as “? Synastrapotherium amazonense ” by Paula Couto (1976), “Astrapotheriinae indet.” (sensu Simpson 1945, i.e. corresponding to Astrapotheriidae sensu McKenna & Bell 1997 ) by Rancy (1981), or Astrapotheriidae incertae sedis by Paula Couto (1982). All these remains originate from middle or early late Miocene deposits of the Rio J uruá-Río Breu confluence area, at the Peruvian border ( Paula Couto 1976, 1982; Negri et al. 2010).

At last, the present work extends the geographical range to Amazonian Peru, with Xenastrapotherium sp. reported here from the Fitzcarrald local fauna (late middle Miocene, eastern Peru), in deposits roughly coeval to the Alto Juruá fossiliferous levels ( Fig. 4 ; Negri et al. 2010).

Granastrapotherium s.s. was so far restricted to the middle Miocene Honda Group of Colombia, with c. 30 cranio-dental and postcranial remains referred to G. snorki , ranging from the base of La Victoria Formation (near Coima, Tolima Department) up to the Polonia Red Beds of the upper Villavieja Formation in La Venta area, 50 km mor e to the South ( Johnson & Madden 1997; Madden et al. 1997). From a chronostratigraphical point of view, the corresponding interval ranges from 13.734 Ma up to 12.82 Ma ( Madden et al. 1997; Gradstein et al. 2005). Granastrapotherium is the only ur uguaytheriine for which appendicular skeleton is somewhat available ( Johnson & Madden 1997). In addition, two craniodental specimens collected by Willard (1966) in the Río Inuya/Mapuya area of eastern Peru – without stratigraphical contr ol neither opportunity for comparing the pr esent specimens – document a close ally, referred to as cf. Granastrapotherium by Johnson & Madden (1997). The remains described in the present work originate unambiguously from the same levels and might document the same tax on ( Granastrapotherium cf. snorki ).

Uruguaytherium beaulieui Kraglievich, 1928 is the type and only species referred to Uruguaytherium . It is based on a single specimen, a left mandible with m2-m3 originating from an unknown locality of western Uruguay (“ Río Negro Dept.”), assumed to be either Oligocene or Miocene in age ( Kraglievich 1928; McKenna & Bell 1997).

Other astrapotheriid remains with uncertain taxonomic affinities were reported from several Miocene localities of nor thern South America in the last decades ( Fig. 4 ): a bispecifi c assemblage based on five postcranials from the early Miocene Castillo Formation is described at Cerro La Cruz, western Venezuela (“ Astrapotheriidae indet. A” and “ Astrapotheriidae indet. B”; Sánchez-Villagra et al.

2004; Weston et al. 2004).

A large uruguaytheriine occurs in the middle Miocene of Quebrada Honda, South Bolivia: Hoffstetter (1977) first referred several isolated teeth to the “ Uruguaytherium-Xenastrapotherium group”.

Then, a badly preserved neurocranium, larger than those of Astrapotherium , was described as belonging to “? Xenastrapotherium ” by Frailey (1987). Given the large dimensions of the corresponding remains, notably its bizygomatic width, referral to the latter genus was discarded for the benefit of “ Uruguaytheriinae , gen. et sp.incertae sedis” ( Johnson & Madden 1997: 377). The age of the concerned deposits is extrapolated at c. 12.7-13.0 Ma (K/Ar and magnetostratigraphy; MacFadden et al. 1990), which coincides with the middle part of the late middle Miocene Laventan SALMA ( Croft 2007).

To our knowledge, only two late Miocene astrapotheriid mentions are reported in the literature.

“ Astrapotheriidae indet. A” and “ Astrapotheriidae indet. B” – distinct from the unidentified astrapotheriids found in association in the early Miocene of western Venezuela ( Fig. 4 ), as reported by Weston et al. (2004) and discussed abo ve – occur in the middle and the upper member of the U rumaco Formation of Northern Venezuela, respectively ( Linares 2004). The nature and number of the corresponding remains (grouped as “ Astrapotheriidae indet.” in Figure 4 ) are not detailed, ex cept that they are badly preserved and/or fragmentary, which might indicate they are reworked from underlying formations of the same area, such as the Laventan Socorro Formation, as it often occurs (Linar es 2004; Carlos Jaramillo pers. com. 2008). Owing to mammalian and foraminiferan biochronology, the concerned deposits are assumed to be middle late Miocene in age (Chasicoan-Huayquerian SAL- MAs; Linares 2004).

At last, an upper canine originating from the late Miocene Río Acre local fauna of Amazonian Brazil (Huayquerian SALMA; Cozzuol 2006), originally identified as “ Xenastrapotherium amazonense ” by Frailey (1986, fig. 26D), was subsequently “referred only questionably to Astrapotheria ” ( Johnson & Madden 1997: 366).

To sum up, the late Miocene astrapotheriid record is far from being well constrained, and a late middle Miocene – Laventan – age for the Last Appearance Datum of Astrapotheria cannot be discarded to date.

Figures

FIG. 3. — Granastrapotherium cf. snorki from the late middle Miocene Fitzcarrald local fauna,Peruvian Amazonia.Left broken M2 without ectoloph (MUSM 994),IN-DTC 28 locality (float):A, occlusal view; B, interpretative sketch of the occlusal view;C, tentative reconstruction of the whole tooth. Scale bar: 20 mm.

FIG. 4.— Location map and fossil record of northern South American Astrapotheriidae Ameghino,1887,based on data from Kraglievich (1928), Stehlin (1928), Scott (1937), Simpson (1940), Paula Couto (1976, 1982), Rancy (1981), Frailey (1986, 1987), Isea (1987), John- son & Madden (1997), Flynn et al. (2002, 2008), Sánchez-Villagra et al. (2004), Weston et al. (2004), Antoine et al. (2007), Croft (2007), and Negri et al. (2010). Abbreviations: LO, late Oligocene; EM, early Miocene; MM, middle Miocene; LM, late Miocene.