Permian plants from the Chutani Formation ( Titicaca Group , northern Altiplano of Bolivia ) : II . The morphogenus Glossopteris

Fossil plants belonging to the morphogenera Glossopteris, Pecopteris and Asterotheca were collected from the upper part of the Chutani Formation (Titicaca Group), near the town of San Pablo de Tiquina, on the southeastern shore of Lake Titicaca (northern Altiplano, Bolivia). This paper presents the first description of specimens of the morphogenus Glossopteris from Bolivia. The Bolivian specimens of Glossopteris consist of poorly-preserved impressions, although they present the diagnostic features of this morphogenus. They are fragments of leaves with secondary veins of taeniopterid-type, typical of glossopterids from Late Permian deposits of Gondwana. The only species of Pecopteris confirmed in the first part of this study, i.e. P. dolianitii Rösler and Rohn (see Vieira et al. 2004), was previously reported from the Late Permian beds of the Rio do Rasto and Estrada Nova formations in the Paraná Basin (southern Brazil). Therefore, a Late Permian age is proposed for the fossil plant-bearing beds of the Chutani Formation based on the analyzed assemblage. The phytogeographic implications of this new find are briefly analyzed.


INTRODUCTION
The presence of the morphogenus Glossopteris in Bolivia had already been mentioned in the literature (Archangelsky 1984(Archangelsky , 1986)).However, none of the published syntheses on Bolivian Paleozoic floras mentions this floristic element (Suárez-Soruco 1974, Archangelsky 1993).Furthermore, the latest review affirms that, so far, there has not been any description of glossopterid plants from Bolivia (Archangelsky 1993).Consequently, this second Glossopteris, Pecopteris and Asterotheca.The fossils were found NE of San Pablo de Tiquina, a small town on the SE shore of Lake Titicaca (western Bolivia), where ferryboats cross the Strait of Tiquina to the Copacabana Peninsula.The fossil site is located approximately 80 m up-hill from the town's cemetery (for a location map see Vieira et al. 2004).The studied samples were collected from the upper part of the Chutani Formation (Titicaca Group -geological setting in Vieira et al. 2004).The fronds of Pecopteris and Asterotheca have been described by Vieira et al. (2004) and form the first part of this study.The leaves of Glossopteris were found as impressions on brownish yellow, hard, compact, silty dolostone.The specimens are fragmentary and stained by iron oxide.The impressions are usually unclear and we had difficulties to photograph them.All studied specimens are listed below and housed at the Museu de Paleontologia of the Departamento de Paleontologia e Estratigrafia of the Universidade Federal do Rio Grande do Sul (code prefix MP-Pb).

Description:
The specimen MP-Pb 3576 is the best preserved and represents a middle portion of a fragmentary leaf.The incomplete specimen is 8.1 cm long and approximately 5.32 cm wide.The shape of the leaf cannot be deduced, but it shows parallel lateral margins.The midrib is distinct, about 6.7 mm wide, and composed of a bundle of thick parallel veins (at least 7 or 8), with rare connecting veins.The venation pattern is taeniopteroid: secondary veins arise from the midrib at acute angles (20-22˚), but they immediately take a sharp bend and follow a straight and parallel course at angles of 80-85˚to the midrib, near the leaf margin they incline towards the apex, reaching the margin at angles of 65-70˚.There are about 28-30 secondary veins per cm half way across the lamina.The pattern of meshes is not well preserved, but apparently the secondary veins show rare and irregularly distributed oblique anastomoses, dichotomizing few times (5 at least) to form short and fusiform meshes near midrib and narrow-elongate ones for the remainder of the lamina.
Further morphological details such as shape and size of leaf, and pattern of meshes along the complete leaf, are still required to obtain a more precise determination.
Glossopteris sp. 1 Figs.1B, 2B, 2C, 4 , 3578, 3580, 3581, 3582, 3583, 3584, 3585 Description: There are a few well-preserved specimens in the present collection.Amongst them, specimen MP-Pb 3577A is the most complete.This specimen represents a long (13.4 cm preserved) and narrow (1.74 cm maximum width) fragmentary leaf of unknown base, tapering gently and distally to an acute apex.Midrib strong, 4 mm broad (max.),persistent up to the leaf apex, composed initially of six thick parallel veins at the base, immediately reducing to four veins and gradually contracting towards the apex where it becomes resolved in finer veins.Taeniopteroid veining pattern, secondary veins leave the midrib at narrow (22-35˚) to broad (55-65˚) acute angles and, after a gentle arch, pass straight, parallel inclined at 55-65˚to the midrib, reaching the margin at about 55˚.The secondary veins often show irreg-ularly distributed perpendicular and oblique anastomoses, dichotomizing several times (10 at least).
The concentration of veins is approximately 25-30 per cm in the middle part of the lamina.The meshes vary from elongate to short, polygonal to fusiform in shape, alternating irregularly between long and short ones across the lamina.The specimen MP-Pb 3580 represents a middle portion of the incomplete leaf and measures 3.4 cm in length and about 14.6 mm in width.The fragmentary leaf shows parallel lateral margins, a distinctive midrib with 2.6 mm in width, and is composed of a bundle of poorly preserved parallel veins.The veining pattern is taeniopteroid and equal to that of specimen MP-Pb 3577A.Secondary veins are not preserved near the midrib, in the remainder of the lamina they run at an angle of 60˚to the midrib until they reach the margin.Specimen MP-Pb 3577B corresponds to the base of a leaf.It is 6.7 cm long and 1.66 cm wide (max.), gradually contracting towards the base into a short and broad stalk (1 cm maximum width).The stalk is composed of a bundle of thick parallel veins, which resolve in veins of secondary veining and midrib (about 4 mm wide) just above the beginning of the leaf lamina.Secondary veining is badly preserved, showing a similar general orientation to that verified in other specimens (about 60˚).

Comparison:
The specimens under consideration resemble Glossopteris dorizonenesis Rohn, Oliveira-Babinski and Rösler, G. leptoneura Bunbury, and G. stricta Bunbury in possessing narrow leaf with acute apex, but all three species are distinct from them in venation pattern (Rohn et al. 1984, Chandra andSingh 1992).At first sight, the venation pattern reminds one of that of G. taeniopteroides Feismantel sensu Kovács-Endrödy (1976), who described these African leaves with less divergent secondary veins from the undivided midvein in comparison to the Indian representatives of the species.However, here the midrib is composed of longitudinal strands (her figures exhibit this feature too, fig.17), the shape of the leaf is entirely different, and the number of secondary veins per cm is higher.In terms of venation pattern, specimens here described can also be compared with G. damudica Feistmantel and G. nimishea Chandra and Surange (Maheshwari and Prakash 1965, Kovács-Endrödy 1976, Chandra and Singh 1992).Both species, however, are different in leaf shape, and the secondary veins are almost perpendicular to the midrib and margin.In all characters, the leaves under consideration agree closely with specimens informally described and illustrated by Broutin et al. (1995) as Glossopteris sp.1?, except for the greater density of secondary veins.However, on account of the incompleteness and poor preservation, the present specimens have not been classified under a new formal name.

Age of Flora
Despite the relatively low taxonomic diversity, the floristic elements are useful for age assignment.The glossopterid leaves under consideration show some diagnostic features mostly found in leaves of West-ern Gondwana assemblages from the end of Early Permian (Kungurian) and younger ages.These features include small size, narrow shape and taeniopteroid venation pattern.In some cases, these kinds of leaves dominate Late Permian assemblages in South America (Rohn and Rösler 1989).In the Bolivian flora, the total absence of typical glossopterid leaves from both Gondwana Lower Permian (Rubigea sp. and Gangamopteris sp.) and Triassic (Yabeiella sp.; Glossopteris spp.from Pant and Pant 1987, Holmes 1992) is also remarkable.A very similar glossopterid leaf assemblage was collected in the uppermost part of the Gharif Formation, in the Arabian Peninsula (Broutin et al. 1995).This part of the Gharif Formation was dated as early Late Permian by palynology and marine fossils from underlying and overlying units.All these considerations about the glossopterid leaves studied indicate an end of Early Permian to Early Late Permian (middle through Late Permian) age for the deposition of the strata here analyzed.
The pecopterid foliage corresponds to taxa considered endemic of the Paraná Basin, in southern Brazil (see the first part of this report; Vieira et al. 2004).Pecopteris cf.P. pedrasica Read, and Pecopteris cf.P. cadeadensis Rohn and Rösler, are taxa identified with uncertainty, and will not be considered in this discussion.Pecopteris dolianitii Rohn and Rösler, the unique clearly identified species, has also been described from deposits of the Estrada Nova and Rio do Rasto formations in the Paraná Basin (Rohn and Rösler 1986).A Late Permian age (Kazanian-Tatarian) was proposed for these units based on the occurrence of: (a) Kungurian-Kazanian palynofloras in the underlying Irati and Palermo formations (Daemon and Quadros 1970, Arai 1980, Marques-Toigo 1991), and (b) Tatarian vertebrate terrestrial fauna at the top of the Rio do Rasto Formation (Barberena et al. 1985, Langer 2000).Besides Sphenophyllum spp.occurs in the lower part of the Rio do Rasto Formation (Serrinha Member -Rohn and Rösler 2000), while leaid conchostraceans are reported from the upper part.Both fossil groups are restricted to Pa-An Acad Bras Cienc (2004) 76 (1) leozoic strata all over the world.In conclusion, a Late Permian age for the Chutani Formation is here proposed.

Phytogeography
Despite the relatively low taxonomic diversity, probably due to non-systematic and insufficient sampling, the composition of the floral assemblage from the Chutani Formation allows some phytogeographic discussion.The assemblage constitutes the northernmost record of ''Glossopteris Flora'' at the western edge of South America, considering the known distribution of this flora in Permian deposits of South America (Rohn andRösler 1987, Archangelsky 1990).According to the current evidence, the Bolivian assemblage was a typical Gond-wana flora, without any Euramerican or Cathaysian elements.It includes glossopterid leaves associated with pecopterid fronds of Gondwana affinity, mostly related to endemic species from the Paraná Basin (see the first part of this report; Vieira et al. 2004).Hence, our observations suggest a shift of the northwestern boundary of the Gondwana Realm towards the tropics.The lack of information from northern South America hinders the determination of the northwestern boundary of this realm, which could be even further northwards.
An important fact is the link between Late Permian floras from the Paraná and Titicaca basins.The same type, and maybe even the same species, of glossopterid and pecopterid forms are found in both geographic areas (see Paleobotanical Results, and the first part of this report; Vieira et al. 2004), suggesting (a) the possibility of a direct dispersion route between them, and (b) the presence of similar paleoenvironmental and paleoclimatic conditions.We therefore propose that these two areas may be regarded as belonging to the same phytogeographic unit.The paleogeographic interpretation for this distribution depends on the continental configuration under consideration.In accordance with most paleomagnetic data, there are two basic Late Permian configurations proposed for Pangea: Scotese's model, the classical Pangea ''A'' (Scotese andMcKerrow 1990, Golonka et al. 1994), and Pangea ''B'' model (Morel and Irving 1981).The model of Scotese places each basin at a different paleolatitude, and approximately 15˚distant from each other.The Pangea ''B'' model assumes a paleolatitudinal proximity between both areas, more in accordance with the paleobotanic results obtained herein.The Pangea ''B'' configuration incorporates an anticlockwise rotation of Pangea during the Late Permian, displacing western Gondwana back to higher latitudes, and placing Bolivia and southeastern Brazil (Paraná Basin area) at a similar southern latitude (Díaz-Martínez et al. 1993, Broutin et al. 1995, Fluteau et al. 2001).The Pangea ''A'' model considers that most of this rotation took place later, in the Early Triassic (Golonka et al. 1994).

CONCLUSIONS
We confirm the presence of Glossopteris in the Bolivian Central Andes (west-central South America), as well as the phytogeographic extension of the Gondwana Realm towards northwestern Gondwana (northern South America).We also suggest a Late Permian age for the fossil-bearing beds of the Chutani Formation, although we do not discard the possibility of a younger (Early Triassic) age for the uppermost part of this unit, as suggested by Sempere et al. (1992).Broutin et al. (1995) suggest a wide extension of the Glossopteris Flora in the Late Permian for other regions of Gondwana (northern Africa and Arabian Peninsula) and southern Europe (Spain, Turkey).This could be due to paleoclimatic changes (more uniform global climate conditions, expansion of seasonal climates) and paleogeographic changes (anti-clockwise rotation and northward migration of Gondwana) that took place during the Late Permian (Díaz-Martínez et al. 1993, 2000, Broutin et al. 1995).