Paleogene Land Mammal Faunas of South America; a Response to Global Climatic Changes and Indigenous Floral Diversity

Abstract

An appraisal of Paleogene floral and land mammal faunal dynamics in South America suggests that both biotic elements responded at rate and extent generally comparable to that portrayed by the global climate pattern of the interval. A major difference in the South American record is the initial as well as subsequent much greater diversity of both Neotropical and Austral floras relative to North American counterparts. Conversely, the concurrent mammal faunas in South America did not match, much less exceed, the diversity seen to the north. It appears unlikely that this difference is solely due to the virtual absence of immigrants subsequent to the initial dispersal of mammals to South America, and cannot be explained solely by the different collecting histories of the two regions. Possible roles played by non-mammalian vertebrates in niche exploitation remain to be explored.

The Paleogene floras of Patagonia and Chile show a climatic pattern that approximates that of North America, with an increase in both Mean Annual Temperature (MAT) and Mean Annual Precipitation (MAP) from the Paleocene into the Early Eocene Climatic Optimum (EECO), although the Paleocene-Eocene Thermal Maximum (PETM) is not recognized in the available data set. Post-EECO temperatures declined in both regions, but more so in the north than the south, which also retained a higher rate of precipitation.

The South American Paleogene mammal faunas developed gradual, but distinct, changes in composition and diversity as the EECO was approached, but actually declined somewhat during its peak, contrary to the record in North America. At about 40 Ma, a post-EECO decline was recovered in both hemispheres, but the South American record achieved its greatest diversity then, rather than at the peak of the EECO as in the north. This post-EECO faunal turnover apparently was a response to the changing conditions when global climate was deteriorating toward the Oligocene. Under the progressively more temperate to seasonally arid conditions in South America, this turnover reflected a major change from the more archaic, and more tropical to subtropical-adapted mammals, to the beginning of the ultimately modern South American fauna, achieved completely by the Eocene-Oligocene transition. Interestingly, hypsodonty was achieved by South American cursorial mammals about 15–20 m.y. earlier than in North America. In addition to being composed of essentially different groups of mammals, those of the South American continent seem to have responded to the climatic changes associated with the ECCO and subsequent conditions in a pattern that was initially comparable to, but subsequently different from, their North American counterparts.

Appendix I

Paleogene South American stratigraphy, mammalian faunas, and biochronology

Paleogene

As indicated in Fig. 1, the Paleogene begins at about 65.6 Ma, and lasts until about 23 Ma. The age and magnetic polarity signatures for the Epochs and Stages follows Luterbacher et al. (2004: fig. 20.4). The Sparnacian has been added, following the recommendations of Thiry et al. (2006). The paleotemperature curve is after Zachos et al. (2001).

Oldest Mammal Fauna

The Grenier Farm site is the currently oldest occurrence of a South American therian mammal for which the age is well documented (Goin et al. 2006a). This site has produced only a single (metatherian) taxon: Cocatherium lefipanum, from the Lefipán Formation, Chubut Province, Patagonia. Based on the associated marine invertebrate fauna the Grenier Farm Fauna is considered to be of Danian age. The position shown on Fig. 1 reflects the 5 m separation of the Grenier Farm site above the likely position of the K/Pg boundary (Goin et al. 2006a: fig. 2).

Tiupampan SALMA

The Tiupampan SALMA is recognized in the Santa Lucia Formation of Bolivia (Gayet et al. 1991). Whereas this unit is not found in Patagonia, it is here taken as older than the Peligran SALMA of that region, following Gelfo et al. (2009).

The Tiupampa Fauna is found in the middle part of the Santa Lucía Formation of Bolivia (Fig. 1; 12, Fig. 2b). Marshall et al. (1997) and Sempere et al. (1997) reviewed the stratigraphy of the Santa Lucía Formation and associated strata. The Santa Lucía Formation is underlain by the El Molino Formation (?late Campanian; Maastrichtian—Danian) and unconformably overlain by the Cayara Formation (Thanetian equivalent). All of the Upper El Molino, the Santa Lucía and part of the Cayara formations are of reversed polarity, correlated to Chron C26r, or from 61.7–58.7 Ma in Luterbacher et al. 2004). The El Molino-Santa Lucía contact is correlated with a regression coeval with the Danian-Selandian boundary at 61.7 Ma, and with the base of chron C26r. Regarding the Tiupampan, this fauna is associated with the base of the Middle Santa Lucía Formation, for which an increase in grain size of the sediments is interpreted (Sempere et al. 1997) as a regression. The first major regression within chron C26r is near the top of the Selandian, or ca 59 Ma. This suggests that the Tiupampa SALMA is about 59 Ma old = very late Selandian; see also Marshall et al. (1997).

Mammal biochronology suggests an earlier age for the Tiupampan. The relative ages of the Peligran (Bonaparte et al. 1993) and Tiupampan (Pascual and Ortiz-Jaureguizar 1990) SALMAs have been revised by Gelfo et al. (2009). Qualitative analyses suggested that the Tiupampan is older than the Peligran, and may be phyletically closer to those of early Puercan age in North America. Molinodus is a member of the Panameriungulatan subfamily Kollpaniinae in the Tiupampa Fauna, and may represent an early dispersal event from North to South America (Muizon and Cifelli 2000: 145), perhaps of Puercan age. As summarized by Gayet et al. (1991) most metatherian and eutherian groups are permissive of an early Paleocene (essentially Puercan) age of many Tiupampan taxa, but Alcedidorbignya (Pantodonta) is first known in the Torrejonian of North America. A late Puercan to Torrejonian correlation for the Tiupampan is shown in Fig. 1, at about 64 Ma.

This scenario implies that the reversed magnetic polarity with which the Tiupampa Fauna is associated pertains to Chron 28r, rather than 26r (also Gelfo et al. 2009), and that the regression interpreted from the presence of coarse-grained material in the Tiupampa sediments is open to interpretations other than reflecting a global event.

On one hand, Sempere et al. (1997:719 L) indicated that the nonmarine sediments of the basal part of the Santa Lucía Formation reflect an eustatically-controlled regression, without tectonic influence. On the other hand, Sempere et al. (1997:715 L) portrayed the distribution of the red-brown lacustrine mudstones of the lower Santa Lucía Formation as having been controlled by remaining subsidence of the basin and by the structural framework in which they occur. This suggests a local, rather than regional, cause for these deposits and diminishes the interpretation for an eustatic origin. Thus, notwithstanding the proposed correlation of the Santa Lucía sediments with Chron C26r, there appears to be no compelling reason to equate the base of the Santa Lucía Formation with the Danian-Selandian boundary.

Sempere et al. (1997:715 L) indicated that the middle Santa Lucía Formation begins with a “somewhat coarser facies,” but discussed this in the context of “highly subsident” areas, and also indicated that the paleostructural corridor in which they were deposited was reactivated at the time the top of the Middle Santa Lucía sandstone beds were deposited. The sandstone interval, which contains the Tiupampa Fauna, is about 50 m thick, according to Sempere et al. (1997: fig. 5). Again, the evidence of tectonic activity associated with the Middle Santa Lucía Formation (also Sempere et al. 1997:719 L) diminishes the interpretation of its regional correlation at ca 59 Ma. In summary, the 64 Ma age for the Tiupampan SALMA (Fig. 3) based on its fossil mammals appears plausible, comparable to that proposed by Flynn and Swisher (1995).

Peligran Through Riochican

The Salamanca Formation and Río Chico Group in the San Jorge Basin of Patagonia (Figs. 3 and 5) contain a stratigraphic succession of Paleogene mammal faunas that is unequaled in South America. As reviewed by Bond et al. (1995), the succession begins with the largely marine Salamanca Formation (Fig. 5) of Danian age, with a basal glauconitic sandstone that rests unconformably on sediments of the Chubut Group. The basal sandstone is followed stratigraphically upward by about 150 m of siltstone and mudstone apparently deposited under transgressive conditions (Legarreta and Uliana 1994). K-Ar dates on basalts at the base of the Salamanca Formation are 64.0 ± 0.8 Ma and 62.8 ± 0.8 Ma (Marshall et al. 1981).

The Salamanca mudstones are overlain unconformably by Banco Verde green sandstones that cut down (locally deeply) into the underlying strata (Fig. 5), and are followed upward by about 8 m of mudstone that is unconformably overlain by the basal sandstone of the Banco Negro Inferior. Marshall et al. (1981) indicated that a tuff in the upper part of the Hansen Member yielded a K-Ar date of 61 ± 5 m.y. Marshall et al. (1981) also showed that strata of the Banco Negro Inferior and adjacent strata of the Hansen Member carry a reversed magnetic signature, and that this pattern continues 80 m into the overlying Río Chico Group. This interval of reversed polarity is followed by a pair of normally magnetized intervals about 19 and 13 m thick, respectively, separated by a reversed interval also about 13 m thick (Marshall et al. 1981: fig. 2). This suggests that the Peñas Coloradas Formation in Fig. 5 is of reversed polarity (unless one or more of the numerous unconformities in the sequence have cut out normally magnetized strata), and that the upper part of the Las Flores and Koluel Kaike formations each are of normal polarity. Under the interpretation of Marshall et al. (1981), the faunas from the Río Chico Formation of their terminology (and the Riochican SALMA) correlates to magnetic chrons 25 and 26, or from about 56.5–61.7 Ma. As shown in Fig. 1, the Riochican SALMA is here correlated to about chron C22r, or at about 50–51 Ma. This is discussed further below. Iglesias et al. (2007a) dated a tuff horizon 35 m above the Banco Negro Inferior in the western part of the basin. It yielded an Ar-Ar date of 57.8 ± 6 Ma, but the sanidines retrieved were clearly altered, and the resulting age has little interpretive value.

Peligran SALMA

The monotreme Monotrematum patagonicum and the sudamericid Sudamerica ameghinoi were recovered from the Banco Negro Inferior (Pascual and Ortiz-Jaureguizar 1991; Pascual et al. 1992; Bonaparte et al. 1993) and, along with other non-tribosphenic Gondwanan as well as therian mammals of Laurasian affinity (Table 1), constitute the basis for the Peligran SALMA (Gelfo et al. 2009). In that context, the Peligran occurs in the unconformity-bounded package that includes the Banco Verde and Banco Negro units (Fig. 5). Foraminiferal evidence yields a Danian age for the Salamanca Formation (Bertels 1975a, b), compatible with the 61 Ma age (± 5 m.y.) for the Hansen Member and its reversed magnetic polarity (Marshall et al. 1981). Soria (1989), Pascual and Ortiz-Jaureguizar (1990, 2007), and Pascual et al. (1996), tentatively correlated the land-mammal fauna of the Río Loro Formation to the Itaboraian SALMA. More recently Gelfo et al. (2009) summarized the evidence that favors a post-Tiupampan age for the Peligran SALMA comparable to its position shown in Fig. 1.

Carodnian

There is no SALMA for this zone. The next unconformity bounded unit, the Peñas Coloradas Formation, contains the “Carodnia” faunal zone (Fig. 5). Pascual and Ortiz-Jaureguizar (2007) indicated that the “Carodnia” biochron contains the first occurrence of the polydolopid polydolopimorphian, Amphidolops (also known from many younger units), the “didelphid” Derorhynchus (now considered as not being present; Table 1), the anisolambdid litoptern Wainka and the xenungulate Carodnia feruglioi. Wainka would be a FAD relative to its presence in the Riochican SALMA, and Carodnia would be a FAD relative to its occurrence in Riochican and younger biochrons. Bond et al. (1995) indicated that this unit correlates with that from the lower levels at Bajo de la Palangana and at Cerro Redondo, near Punto Peligro. Gelfo et al. (2009) summarized the presence of possibly correlative faunal sites in Colombia and northwestern Argentina (also Villarroel et al. 1987 - Bogotá Formation, Colombia; Soria 1989; Pascual and Ortiz-Jaureguizar 1990, 2007; Pascual et al. 1996; Bergqvist et al. 2004 - Río Loro Formation, Argentina).

Even though sparse, the taxa in the “Carodnian” fauna fit well as shown on Fig. 1, and this position is compatible with the general superposition of the Peñas Coloradas Formation indicated in Fig. 5 and the apparently reversed magnetic polarity of its sediments (Marshall et al. 1981). Figure 4 indicates that there are many unconformities of unknown duration in the Río Chico Group and the Hansen Member of the Salamanca Formation. Figure 5 also shows the stratigraphic location of the two normal magnetozones described by Marshall et al. (1981) as though no unconformities were present (none were recorded in Marshall et al. 1981). Based on additional field studies (Raigemborn et al. 2010), but also indicated by Simpson (1935: fig. 3), it is clear that numerous unconformities are present in Río Chico Group deposits at Cerro Redondo and elsewhere in the Golfo San Jorge Gulf Basin. In that context the integrity of the magnetic zonation proposed by Marshall et al. (1981) and utilized further in Marshall et al. (1997) appears to be doubtful. For the purposes of this report, location of the paleomagnetic signatures of Marshall et al. (1981) is accepted, but their use as a coherent profile for the purposes of correlation to patterns in the GPTS is strongly questioned.

Itaboraian SALMA

Patagonia. The Peñas Coloradas Formation is followed stratigraphically by the Las Flores Formation and its faunal unit, in part known as the “Kibenikhoria” faunal zone (Fig. 5), which is a correlative of the Itaboraian SALMA. Bond et al. (1995) indicated that this faunal zone is from Cañadón Hondo (Figs. 2 and 5). Correlative faunas in the Las Flores Formation near Gran Barranca (Fig. 6; Raigemborn et al. 2010) occur above an ash dated at 57 Ma. This appears compatible with a revised age of the Itaboraian SALMA as utilized here (Fig. 1). We note that the Las Flores fauna occurs above unconformity 10 in Fig. 5, in rocks apparently correlative with those having a normal magnetic polarity signature (Marshall et al. 1981). If this magnetozone is chron C24n, that would be compatible with evidence in favor of an age of about 53 Ma for the Itaboraian SALMA.

Itaboraí

The fresh water travertine deposits at São José de Itaboraí yield fossil mammals of the Itaboraian SALMA, conventionally considered to be about medial Paleocene in age (e.g., Marshall et al. 1997), who interpreted the succession as having been developed in a sequence stratigraphic context regardless of the nonmarine origin of the sediments. This correlation is maintained in Flynn and Swisher (1995: fig. 3). Medeiros and Bergqvist (1999) and Gomes Sant’Anna and Riccomini (2001) reiterated that the Itaboraí travertines formed in a tectonically active basin, which carries the implication that their genesis likely is locally controlled rather than in response to global eustatism. Medeiros and Bergqvist (1999) also demonstrated that the upper travertine unit is of Pleistocene age and inappropriate for Paleogene sequence stratigraphic analysis, and further indicate that an ankaramitic basalt flow rests unconformably on the travertines that elsewhere contain the post-depositional fissure fillings with Itaboraí mammals. Gomes Sant’Anna and Riccomini (2001) indicated that the basalt is dated at 52.6 ± 2.4 Ma. On faunal grounds (e.g., Marshall et al. 1997), the Itaboraí mammals are not of Cretaceous aspect. If the basalt flow also postdates the genesis of the fillings, then the Itaboraí mammals should be post-Cretaceous and pre-53 Ma in age, but the above tectonic and stratigraphic considerations dictate against invoking a sequence stratigraphic analysis for these deposits. If the scenario proposed by Gomes Sant’Anna and Riccomini (2001) is reasonable, and that the cementation of the Itaboraí fissure fillings (and the contained fossils) took place at about this time, then a 53 Ma age for the Itaboraí fossils is plausible and is used here (Fig. 1).

Riochican SALMA

Legarreta and Uliana (1994) indicated that the youngest unit in the faunal succession of the former Río Chico Formation or Group, the “Ernestokokenia” faunal zone of Simpson (1933), is contained in the Koluel Kaike Formation that unconformably overlies the Las Flores Formation and is unconformably overlain by the Casamayor Formation of the Sarmiento Group (Figs. 5 and 6), now designated as the Sarmiento Formation (Ré et al. 2010b). It also occurs at the upper levels at Bajo de la Palangania and at Cerro Redondo. This is the basis for the Riochican SALMA. See also Flynn and Swisher (1995: fig. 3, where the Riochican SALMA is considered as late Paleocene, rather than late early Eocene, as advocated here in Fig. 1).

In summary to this point, strata that contain type or correlated successions pertaining to the Peligran through Riochican SALMAs can be documented in superpositional relationships in the Golfo San Jorge Gulf Basin of Patagonia. The Tiupampan is reliably interpreted as preceding the Peligran. Whereas it cannot be directly shown as pre-Tiupampan, the documented early Paleocene age of the Grenier Farm site indicates that derived polydolopimorphian metatherians were present by that time in South America. This is compatible with the Late Cretaceous presence of that group in North America on the one hand and also with its plausibly Late Cretaceous dispersal to South America.

Sapoan

Now that the “Barrancan” episode of the Casamayoran SALMA is known to be at least 36 Ma old (Kay et al. 1999), that aspect of the Sarmiento Formation is separated by a substantial hiatus (unconformity 12 in Fig. 5) from the underlying Koluel Kaike Formation. As indicated on Fig. 1, the post Riochican - pre “Barrancan” interval has a duration of about 8 m.y., and that from the Riochican to the “Vacan” SALMA is about 5 m.y. long.

Tejedor et al. (2009) described fossil mammals derived from a sequence of tuffs and sediments that crop out in the Chubut River drainage west of Paso del Sapo in north central Chubut Province (33, Fig. 2c). These deposits appear to be extra-caldera counterparts of the Tufolitas Laguna del Hunco Formation, an important plant-bearing succession that formed within the Piedra Parada Caldera at about 52 Ma and persisted later. An ignimbrite that underlies the “Sapoan” mammal-bearing sediments at the type locality of Laguna Fría has been dated (⁴⁰Ar/³⁹Ar) at 49.51 ± 0.32 Ma. Correlative deposits with fossil mammals at a site known as La Barda occurs above the upper member of the Huancache Andesite, dated at 47.89 ± 1.21 Ma and is overlain by basalts dated at 43 Ma. Tejedor et al. (2009) interpreted the Paso del Sapo mammals to occur within an interval of from 47 to 49 Ma and correlate these with the mammal fauna from the La Meseta Formation of the Antarctic Peninsula (Marenssi 2006; Ivany et al. 2008).

Vacan

Cifelli (1985) distinguished the Vacan and Barrancan as sub-ages of the Casamayoran SALMA. “Vacan” fossils are preserved in the lower part of the Sarmiento Formation at Cañadon Vaca, located about 65 km northeast of the Gran Barranca (Fig. 1). The thickest and youngest part of the Sarmiento Formation crops out at Gran Barranca, with “Barrancan” fossils found near its local base (Fig. 7). Cifelli (1985) noted that the contact between the Sarmiento Formation and underlying Río Chico Formation is ‘sharp but planar contact’ (p. 16), and that fossils occur from about 5 m, 12 m, 18 m, and 86 m above the base of the contact in Section II. Most specimens are found from 5–19 m above the contact (p. 16), and all of those are of “Vacan” age (p. 23).

Cifelli (1985) considered that “Vacan” taxa are more similar to those then known from Río Chico Formation (Koluel Kaike Formation of current literature; e.g., Ré et al. 2010a) than to those of the Gran Barranca. Andreis (1977) noted an angular unconformity between his Cañadon Honda Formation (basal Sarmiento) and the Río Chico Formation in the Cañadon Hondo site, located about 25 km E. of Cañadon Vaca on the E. side of Río Chico (see Fig. 1 in Cifelli). Carlini et al. (2005, 2010) have proposed an age of 45 Ma for the “Vacan” SALMA. Based on the pre-”Barrancan” age of the “Vacan” SALMA in the Sarmiento Formation, and its earlier age relative to early “Barrancan” elements, the ‘conformable’ relationship portrayed in Ré et al. (2010b) at Gran Barranca must contain a hiatus (Fig. 7). Both faunally and stratigraphically the assemblages termed “Vacan” differ from those assigned to the “Barrancan.” If the 86 m level at Cañadon Vaca correlates with the basal part of the Sarmiento Formation at Gran Barranca, then nearly 86 m of section have been removed at the level of the unconformity between the Sarmiento and Koluel-Kaike formations shown on Fig. 7.

The Gran Barranca and the “Barrancan” through Pinturan SALMAs

The Gran Barranca of the Colhue Huapi Lake (Lago de Colhue Huapi) in southern Chubut Province, Argentina (37, Fig. 2c) exposes primary outcrops of the Sarmiento Formation. The Sarmiento Formation contains the type sections of the “Barrancan,” Mustersan, Deseadan, Colhuehuapian and Pinturan SALMAs, and fossils correlated with the Tinguirirican SALMA also occur in this unit (Ré et al. 2010a) as indicated in Fig. 7. The physical stratigraphy of the Sarmiento Formation at Gran Barranca is summarized by Bellosi and Madden (2005), Ré et al. (2005, 2010a, b), and Madden et al. (2005), and the following discussion is derived from these publications. They incorporated new litho-, bio-, magnetostratigraphic and geochronologic studies that build upon numerous prior investigations, including Ameghino (1906), Feruglio (1949), Simpson (1941), Spalletti and Mazzoni (1977, 1979), Cifelli (1985), Mazzoni (1985) and Kay et al. (1999).

The Sarmiento Formation (formerly Sarmiento Group) is a regionally extensive pyroclastic unit about 319 m thick that unconformably overlies the subjacent Koluel Kaike Formation (Río Chico Group), and is overlain unconformably by the marine Chenque Formation, followed by the nonmarine Santa Cruz Formation. The areally extensive subhorizontal Sarmiento succession contains numerous facies and hiatuses based upon lithologic changes and the development of paleosols (Ré et al. 2010a, b).

As indicated in Fig. 7 the Sarmiento Formation is considered in the context of six members wherein the SALMAs based on fossil mammals can be correlated to the global chronology via the interbedded paleomagnetic and radioisotopic data. Considered in this context, the “Barrancan” SALMA is contained within the Gran Barranca Member of the Sarmiento Formation. This unit extends from the base of the formation, a few meters below the VRS Tuff, and up to Disconformity 2, which is directly overlain by the Rosado Tuff and ranges in age from about 42 to 39 Ma (Fig. 7). Much of the type “Barrancan” fauna was recovered from the Y Tuff, but others occur both above and below that fossiliferous tuff bed (Ré et al. 2010a).

Mustersan SALMA

The Mustersan SALMA is contained within the Rosado Member and unit 1 of the Lower Puesto Almendra Member of the Sarmiento Formation, bracketed by discontinuities 2 and 5 (Fig. 7), with an age range of about 39 to 36.5 Ma. As discussed by Ré et al. (2010a), the Rosado Member of the Sarmiento Formation is about 7 m thick, and is bounded above by Disconformity 3 (U3, Fig. 7). The Lower Puesto Member occurs between Disconformity 3 and Disconformity 5 (U5, Fig. 7), and is about 30 m thick. The Rosado Tuff occurs a few meters above the base of the Rosado Member and yielded a mean ⁴⁰Ar/³⁹Ar age of 38.66 Ma. The Kay Tuff occurs in Unit 2 of the lower Puesto Alamendra Member of the Sarmiento Formation, about 2 m above Disconformity 4. The mean age for the Kay Tuff is 37.05 Ma. Ré et al. (2010a) assigned a tentative age of about 38.0 to 36.5 Ma for the Mustersan SALMA at Gran Barranca, with the tentative aspect being the uncertainty of when, within Chron 18n.1n the Mustersan actually occurs. The Rosado Member strata correlated with the lower (but not lowest) part of that chron contain a transitional “Barrancan”-Mustersan fauna (El Nuevo L.F., Figs. 1 and 7) here suggested as being about 38.5 Ma.

Tinguirirican SALMA

The Tinguirirican SALMA was nominated by Flynn et al. (2003) for fossil mammals recovered from the Abanico Formation in southeastern central Chile (Tinguiririca, 52, Fig. 2d), in a strongly volcaniclastic succession about 100 m thick. Radioisotopic dates (⁴⁰Ar/³⁹Ar) from the succession range from 35.6 ± 0.85 Ma to 31.34 ± 0.17 Ma, with the Tinguiririca Fauna being assigned an age of about 31.5 Ma, although suggested as likely ranging older. Based on its correlation to the Gran Barranca an age for the Tinguirirican SALMA is here taken as likely 34.5 to 31.5 Ma.

Fossil mammals correlated with the Tinguirirican SALMA occur in the Vera Member of the Sarmiento Formation which correlates to an age range of about 34.8 to 33.3 Ma. The La Cancha site that yields Tinguirirican mammals (Goin et al. 2010) corresponds to magnetic polarity Chron C13n, or from 33.7–33.3 Ma (Fig. 7). In Fig. 1 the age of the Tinguirirican SALMA is taken as from about 34–31 Ma.

Deseadan SALMA

The Upper Puesto Almendra Member is dated radioisotopically as 30.4 to 25.9 Ma, The member is composed of three units, 3, 4, and 5, with the base of Unit 3 bounded by Disconformity 6 (U6, Fig. 7). Disconformity 7, within Unit 3, is the base upon which a basalt rests dated at 29.18 ± 0.38 Ma. Two other basalts in this upper part of Unit 3 are dated at 27.78 ± 0.08 Ma and 26.34 ± 0.32 Ma, respectively. Although these basalts also rest locally on or near Disconformity 7, they represent different flow events. Disconformity 8 is cut into Unit 3 and forms the base of Unit 4. Disconformity 9 likewise forms the base of Unit 5. Neither units 4 or 5 are dated isotopically. Disconformity 10 separates the entire member from the overlying Colhue-Huapi Member (U9, Fig. 7).

Two fossil-bearing units occur in the Upper Puesto Almendra Member. The La Cantera L.F. occurs near the base of unit 3 below Disconformity 7, and just above Disconformity 6. The fauna is considered as pre-Deseadan in that it contains some Tinguirirican elements as well as Deseadan taxa. The La Cantera L.F. is considered to correlate with Chron C11n.1n to C11n.2n, and thus range in age from 31.1 to 29.5 Ma.

The Deseadan fossils are found in units 3 and 4 of the Upper Puesto Almendra Member of the Sarmiento Formation. The stratigraphic information supports a potential age range of 29.3 to about 26 Ma.

Neogene

Colhuehuapian SALMA

The Colhue-Huapi Member of the Sarmiento Formation crops out only in the west end of the Gran Barranca. Four tuff beds occur in the unit. The Basal tuff is about 10 M above the basal contact, Disconformity 10 (U10, Fig. 7), and has a mean age of 20.39 Ma. The Big Mammal Tuff is about 6 m above the Disconformity 10, with a mean age of 19.75 Ma (not shown on Fig. 7). The Monkey Tuff lies about 22 m above the basal contact and has a mean age of 19.81 Ma. The highest tuff, MHZ 24.5 lies about 48.5 m above the base of the unit and has a mean age of 19.30 Ma. In that this tuff lies between beds with Colhuehuapian mammals on the on hand and Pinturan mammals on the other, it provides a boundary age between these SALMAs.

The beds of the lower fossil zone are the type section for the Colhuehuapian SALMA. The fossil-bearing beds in the Gran Barranca are correlated as being 20.2–20.0 Ma.

Pinturan SALMA

Kramarz et al. (2010) strengthened a proposal for a new Pinturan SALMA, based on fossiliferous strata of the Pintura Formation, in the Pinturas River valley of Santa Cruz Province (Kramarz et al. 2010: fig. 18.1). This unit is considered to be post-Colhuehuapian and pre-Santacrucian in age, based on mammalian paleontology. The strata in the upper fossil zone of the Colhue-Huapi Member of the Sarmiento Formation contains taxa correlated with the Pinturan NALMA (Kramarz et al. 2010), who state that this unit thus ranges regionally in age from about 18.75 to 16.5 Ma.