Megaspores from coals of the Triunfo Member , Rio Bonito Formation ( Lower Permian ) , northeastern Paraná State , Brazil

This paper presents the results of a detailed study of megaspores occurring in coal seams of the Triunfo Member, Rio Bonito Formation at Figueira, Paraná State, Brazil. This coal-bearing sequence accumulated in a marine delta system during the Early Permian. Four species are described and illustrated: Lagenoisporites triunfensis, L. scutiformis, Sublagenicula cf. brasiliensis, andSetosisporites cf. furcatus. Of these, the two species of Lagenoisporites are predominant. Relationships to other megaspore species are discussed; and the temporal and spatial distributions of the four species in the Paraná Basin are documented.

The two other, less abundant species identified are Sublagenicula cf.brasiliensis (7 specimens) and Setosisporites cf.furcatus (3 specimens).Previous records of these two species show a wide temporal and geographic distribution.For example, they have been recorded from the Monte Mor exposure of the Itararé Subgroup in southeastern Brazil (Trindade 1970) as part of a Late Carboniferous flora.They are also components of Early Permian floras of the Rio Bonito Formation in southern Brazil (Dijkstra 1956, Trindade 1962); e.g., at Barro Branco (Santa Catarina State) and Leão and Candiota coal mines (Rio Grande do Sul State).

FIGUEIRA REGION: LOCALITY AND GEOLOGY
The Figueira region is located in southern Brazil, in the northeastern part of Paraná State (Fig. 1).The upper Paleozoic lithostratigraphic succession includes the Upper Carboniferous to Lower Permian Itararé Subgroup sediments overlain by the Lower Permian Rio Bonito Formation (Triunfo and Paraguaçu Members) (Fig. 2) and the Palermo, Irati, Serra Alta, and Teresina Formations (Soares and Cava 1982, Aborrage and Yamamoto 1982, Morrone and Daemon 1985).The megaspores described herein are from the lower part of Triunfo Member, specifically from siltstones immediately overlying the coal bed.
The depositional environment of the coal at Figueira (Soares and Cava 1982, Morrone and Daemon 1985, Della Fávera et al. 1993, Della Fávera and Chaves 1998) is thought to have been related to a delta-plain system wherein extensive swampy areas were covered by vegetation and possibly surrounded by sandy banks.Subsequent change in the drainage regime caused the collapse of the basin and effectively terminated peat deposition.

SOURCE OF THE PALYNIFEROUS SAMPLES
The megaspore-bearing samples were collected in the coalfields of Companhia Carbonífera do Cambuí (Fig. 1, Table I): Amando Simões mine (AS, well 01 and 06), Slope Plan mine 115 (SP-115), and from the general waste deposit (GWD) of the Amando Simões mine.The megaspores are from the grey siltstones located at the top of the coal seams.Also occurring with the megaspores are carbonized compressions of coniferous leaves, cones, seeds, branches, and microphylls, together with lycophyte shoots and glossopterid leaves.

MATERIALS AND METHODS
The samples were treated initially with 50% Schulze reagent for 5 to 6 days.The megaspores were retrieved from the residues using a fine paintbrush and then rinsed with distilled water.They were then immersed in a 20% solution of hydrofluoric acid for 24 hours in order to remove adherent mineral mat-ter.After further thorough washing to neutrality, the specimens were left to dry at room temperature (20-30 o C).
The 16 samples (6 from SP-115 mine, 5 from AS, and 5 from GWD) yielded a total of 352 megaspores.These were mounted on slides for observation by stereoscopic binocular microscopy.
Selected specimens were studied and photographed with a scanning electron microscope (SEM) in Centro de Pesquisas e Desenvolvimento Leopoldo A. Miguez de Mello (CENPES-PETROBRÁS), Rio de Janeiro, RJ.
The other 27 specimens are proximo-distal compressions (Fig. 3, picture 3).Among the specimens examined, 33 had evidently germinated, as indicated by their open gula.These were not measured, as measurements are con- ventionally based on specimens with closed gula.The latter account for approximately 67% of the specimens studied.
The smaller specimens were found among the megaspores of the samples GP/3T 2162 and 2165 (e.g., 956 µm long in the latter sample).These are considered aborted forms, following the criterion applied by Dijkstra (1956) to small specimens of Triletes brasiliensis.
All these megaspores are very similar to L. triunfensis, previously reported as occuring only at the type locality [i.e., the outcrop of S.J. do Triunfo (Rio Bonito Formation): Arai and Rösler 1984].The main difference is that the average size of the Figueira specimens is larger (1267-2689 µm × 1044-2133 µm) than typical L. triunfensis (1280-1800 µm × 1250-1800 µm).This difference can be explained by the number of specimens measured during this study: i.e., 123 compared with only four studied by Arai and Rösler (1984).Most probably, the differences reflect intraspecific size-variations.Other characters, such as shape, gula, arcuate ridges,  ornamentation, etc., justify inclusion of the present specimens in Lagenoisporites triunfensis.Thus the Figueira region is the second site from which this species has now been reported.L. triunfensis shows many similarities to L. scutiformis Trindade.However, salient differences lie in the ornamentation of the distal hemisphere.In L. triunfensis it is apiculate-capillate (Arai and Rösler 1984), but is smooth in L. scutiformis (Trindade 1970).Although size is considered a valid differential parameter, its validity decreases as the range of variation increases, as in L. triunfensis.In this case, size cannot be used as the sole criterion for distinguishing L. triunfensis from L. scutiformis.In other words, distal ornamentation remains the prime means of differentiating these two species.Dybová-Jachowicz et al. (1979) proposed a novel classification of gulate megaspores, based on the type of gula, According to their scheme, many species formerly included in the genus Lagenoisporites were reallocated to a new genus Sublagenicula, which includes many of the subgulate megaspores.According to these authors, L. triunfensis features an anguligula, typical of the genus Auritolagenicula, which ranges from Upper Devonian through Namurian A. In contrast to Lagenoisporites, Auritolagenicula typically exhibits a thick exine which is more or less densely ornamented.The Figueira specimens do not comply with this system, and accordingly we prefer to retain the specimens described above in Lagenoisporites.
L. triunfensis shows no resemblance to other species of megaspores that have been described from either Argentina or Brazil.However, it shows similarities, in shape and type of gula, to certain specimens that were described from the Namurian of the Illizi Basin (Candilier et al. 1982) and were included in Auritolagenicula.When comparing A. spinulata Candilier, Coquel and Loboziak and A. bacculata Candilier, Coquel and Loboziak with the specimens studied herein, we can observe several differences relating to size and sculpture of the contact faces and of the distal hemisphere.Concerning size, the specimens of A. spinulata (1360-1800 µm × 1240-1680 µm) and A. bacculata (800-1100 µm × 700-800 µm) are smaller than those of L. triunfensis.The ornamentation of A. spinulata consists of long spinules that cover both the contact faces and the distal hemisphere, while the baculate ornamentation of A. bacculata is limited to the distal hemisphere.L. triunfensis can also be set apart from A. bacculata by comparing their ornamentation, insofar as the latter species bears spinules that are borne irregularly on the distal hemisphere.
Beyond Gondwana, Auritolagenicula auricula, described by Winslow (1962) as Triletes auritulus from the Lower Mississipian of Ohio, U.S.A., shows some similarity to L. triunfensis in size and in the type of gula, but the ornamentation and shape of the auriculae are different.Trindade, 1970 Figure 4 (pictures 1-6) and Figure 5  Discussion: The 34 megaspores studied represent 18% of the total and constitute the second most numerous species.Twenty specimens appeared in lateral compression (Fig. 4, pictures 3, 5 and 6) and 7 in polar compression (Fig. 4, pictures 1 and 2).

Lagenoisporites scutiformis
The most common dimensions, for specimens in lateral compression, range from 1800 to 2200 µm in length (16 specimens) and 1400 to 1800 µm in width (14 specimens).The most common equatorial diameter ranges from 1800 to 2100 µm (4 specimens).
Based on such characters as shape, dimensions, and exine ornamentation, these specimens are attributable to Lagenoisporites scutiformis Trindade.This species has previously been recorded from Monte Mor (Itararé Subgroup), State of São Paulo, by Trindade (1970); and from S.J. do Triunfo (Rio Bonito Formation) Paraná State, by Arai and Rösler (1984).
As with L. triunfensis, L. scutiformis might also be included in the genus Auritolagenicula, according to the taxonomy proposed by Dybová-Jachowicz et al. (1979), because it shows an anguligula.However, its psilate to scabrate distal ornamentation precludes such reassignment.
Besides, if the synonymy between Lagenoisporites and Auritolagenicula was confirmed, the priority must be given to the former genus, because Lagenoisporites was erected in 1955 and Auritolagenicula in 1979.Trindade (1970) and Arai and Rösler (1984) alluded to resemblance between Auritolagenicula angulata (=Lagenicula) and L. triunfensis.Distinction between these two species is discussed above.The difference between L. scutiformis and A. angulata lies in the nature of the thick exinal ornamentation (psilate in the case of L. triunfensis).
A. angulata has been reported from the Lower Carboniferous of Egypt, North America, and Poland.It is possible that a relationship exists between it and L. scutiformis, but this can be estab-  But there are size differences (Table II): the studied specimens are larger than those reported previously from Brazil (Dijkstra 1956, Piérart and Dijkstra 1961, Trindade 1959, 1962, 1970, Cauduro and Zingano 1965, Marques-Toigo et al. 1975), Argentina (Spinner 1969, Archangelsky et al. 1989, Cúneo et al. 1991, García 1995), and South Africa (Pant and Srivastava 1962).Leinz (1940) illustrated megaspores from coal samples of the Rio Carvãozinho (Figueira region) that are very similar to our specimens.It seems unusual that this is the only previous record of these megaspores from the Figueira region.
It is important to note that differences in size may relate to climatic, ecological, or intraspecific variations.Taking into account the fact that S. brasiliensis has a broad distribution range, in both space and time (being known from the Lower Carboniferous of Chad, Egypt, and Nigeria and Carboniferous-Permian of Argentina, Australia, and South Africa), we prefer to identify the Figueira megaspores as Sublagenicula cf.brasiliensis.
An Acad Bras Cienc (2002) 74 (3) These Figueira megaspores are clearly attributable to Setosisporites, and are obviously closest to S. furcatus.However, in view of differences in size and exine ornamentation, it is prudent to designate them as Setosisporites cf. S. furcatus.

DISCUSSION AND CONCLUSIONS
This paper is a contribution to the knowledge of Early Permian megaspores of Brazil.Details are provided concerning their morphology and distribution, especially for the species Lagenoisporites triunfensis and L. scutiformis.
The megaspore flora discussed here originated from a coastal swamp vegetation characterized by lycophytes.Of the latter, Brasilodendrom cf.pedroanum is quite possibly related to at least one of the species of Lagenoisporites.However, the precise botanical affinities can only be established via in situ observation of the megaspores; so far, no suitable (fertile) megafossils have been found.
The Figueira megaspore flora is of low diversity, and is dominated by L. triunfensis.This concurs with the low diversity of the preserved lycophyte megaflora, and is a likely consequence of a somewhat stressed, tide-influenced coastal paleoenvironment.
The notable endemism of the taphoflora, alluded to in previous megaspore research, are reinforced by the present study.Neither L. triunfensis nor L. scutiformis is known outside of Brazil, nor even beyond the states of São Paulo and Paraná.Nevertheless, we do not exclude the possibility of finding these species in more meridional localities, because the swampy regions of the Early Permian (typified by the Triunfo Member) characterized the entire paleo-shoreline, extending into the state of Rio Grande do Sul.
Considering the interval between the early and middle Artinskian, related to S.J. do Triunfo (Arai and Rösler 1984), and that between the end of the Sakmarian and the beginning of the Artinskian in Figueira (Ricardi-Branco 1997), it seems reasonable to assume that both L. triunfensis and L. scutiformis are typical of the middle Early Permian.
Finally, it is interesting to note Piérart's (1962Piérart's ( , 1975Piérart's ( , 1981Piérart's ( , 1984) ) suggestion that Sublagenicula brasiliensis could belong to the group called "transgressive megaspores" (along with Setosisporites furcatus, among others) that migrated during the Carboniferous from regions close to the glaciation limits to regions subsequently colonized by the Glossopteris Flora.

ACKNOWLEDGMENTS
The authors wish to express their sincere gratitude to the following: Fundação de Amparo à Pesquisa no Estado de São Paulo (FAPESP) for research support; Compahia Carbonífera do Cambuí for providing samples; Dr. Samuel Branco for illustrations; Dr. Georges Lachkar (Université Paris VI, France) for contributing helpful comments; Professor Geoffrey Playford (The University of Queensland, Brisbane, Australia) for reviewing the manuscript and making linguistic and other amendments; and two anonymous referees of Anais da Academia Brasileira de Ciências for their suggestions.

Fig. 2 -
Fig. 2 -Lithostratigraphy of the Rio Bonito Formation in the Figueira region.Modified from: Morrone and Daemon 1985.