Information about fern spore banks is restricted to the soil systems. As the dispersion of spores occurs by means of air, it is possible to have viable spores on tree bark. Thus, it is important to know if on this kind of substrate, which is thinner and apparently more susceptible to desiccation than the soil, the spores can survive for any length of time, forming transient or persistent spore banks. Samples of bark were collected from ten angiosperm trees in March 1997 and from fifteen trees in February and September 1998. The samples collected in March 1997 contained from 0.05 to 7.19 gametophytes cm⁻² of cultured bark, those of February 1998 from 0.11 to 4.22 gametophytes cm⁻², and in September 1998 from 0.32 to 5.0 gametophytes per square cm. Although the cerrado region is characterized by climatic seasonality, this seasonality was not observed in relation to number of viable spores on barks. As a consequence of the casual spore dispersion pattern, the bark spore bank has a random distribution among the trunks. Ten species were identified on barks collected in February 1998 and fifteen in September 1998, one of them epiphytic (Phlebodium areolatum (Humb. & Bonpl. ex Willd.) J. Sm.) and the others terrestrial species. Thelypteris was the most frequent genus in the analyzed samples. The results obtained show the potential for these substrates to retain viable spores that can participate in the regeneration process and population dynamics of the pteridophyte flora. Moreover, the existence of viable spores of terrestrial species on tree bark does not answer an important question—why do terrestrial species not establish themselves on trees?

Field observations suggested that the epiphytic fern Trichomanes godmanii occurred more frequently and abundantly on the trunks of the palm Welfia georgii than on the trunks of dicotyledonous trees. We tested this observation statistically by randomly selecting 25 individuals of W. georgii and the nearest dicotyledonous tree of similar dbh, for a total of 50 trunks. For each trunk up to a height of three meters, we recorded the presence or absence of T. godmanii and, if present, we visually estimated percent cover using a ranked scale. We found that the fern occurred more frequently and abundantly on the palm than on dicotylendonous trees. No relationship was found between the diameter of the trunks and vegetative cover. This is one of the few host-specific preferences recorded among epiphytic ferns. We cannot fully explain why the fern occurs more frequently and abundantly on the trunks of Welfia georgii instead of dicot trees, but the fern's adhesive hairs on its rhizomes and petioles probably help attachment to the smooth trunk of the palm.

A new species, Adiantum thongthamii (Pteridaceae), known only from a small island in southeastern Thailand is described and illustrated.

Isozyme and plastid DNA analysis prove that true A. ceterach occurs on the Canary Islands, in addition to A. aureum and an octoploid taxon. Combining morphological and cytological observations leads to correct determination, but the exospore length alone also allows reliable identification of these Canarian species. Our allozyme data suggest that the Canarian A. ceterach population is not completely genetically isolated from the European ones. The holotype of Ceterach aureum var. parvifolium, formerly regarded as an octoploid taxon, proved to be A. ceterach, leaving the octoploid without a correct name. The recently described A. octoploideum shows monomorphic, presumably fixed heterozygosity for a combination of the patterns seen in A. ceterach and A. aureum at four loci (Aat, Skdh, Me, and Pgi-2) confirming its allo-octoploid nature. It most probably originated by chromosome doubling in a tetraploid hybrid between A. aureum and A. ceterach or via the union of their unreduced gametes. Pgi-2 indicates multiple origins of the allo-octoploid, implicating recurrent gene flow from tetraploids to octoploids.