maining but the exospore, which is smooth, and without lustre. None 
of those examined by me attain the maximum diameter, 2.25 mm, 
noted by him in the Arroio dos Ratos spscimens. 
On account of the large size of the megaspores they have been 
regarded by Zeiller as probably Sigillarian, instead of belonging to 
Lepidodendron, the resemblance to the megaspores of Siglarivstrobis 
Tieghemi Yeing very close. This expression on Zeiller’s part as to tha 
origin of the species was in effect prophetic. The presence of the genus 
Sigdlaria in Brazil is now fully demonstrated. Megaspores approaching 
in size the specimens figured may, I believe, also belong to the genus 
Lepidophlotos, another type present in the Brasilian coal flora; but the 
spores here illustrated may well have been derived from the cones of 
Sigillaria australis, which is found at the same point. 
To one of the species of Sigilaria described above is probably to be 
referred the leaf fragment shown in Pl. V, fig. 3. This represents the 
basal portion of a leaf nearly 3mm in width. From the base to the 
broken end is 8.5 cm, another fracture, with slight overlap, being 
secn in the lower part. The leaf tapers but little in the fragment 
the midrib is slender, and the stomatiferous furrows are wide 
set. There is little doubt that the specimen belongs to Sigil/aria or Lept- 
dophloios, and since it occurs in the same shale with Sigillaria australis 
and in most intimate association with some of the megaspores noted 
above I do not hesitate to refer it to the same genus. 
Locality : The spores are in a shale in the Bonito coal near Minas, 
Santa Catharina. 120 meters above the granite, and about 160 meters 
below the Iraty black shale horizon. Lot 3595. Also from Estrada Nova 
near Minas. About 135 meters above the granite. Lot. 3923. The leat 
is from the latter locality. 
FILICALES 
Ilereunder are included only the types which either do not 
themselves present Pteridospermic Cycadofilic characters or do not 
seem so closely related to other Pteridospermic types as to make it 
appear probable that they really belong to the same class. Accor- 
dingly Glossopteris and Gangamopteris, together with Newropteridium 
arein this report excluded from the true ferns. The reasons will be 
more fully discussed in connection with the discussion of each genus. 
5560 Bt 
