43 289 



miniim. The specimens fioiii Chili and Argentina, which I have seen, belong to 

 Ü. argentina. 



The essential characters of D. opposila are: Rhizome erect with many fasci- 

 culated leaves. The leaf narrows very gradually through a large number of reduced 

 pinnæ nearly to the base of the stem. Pinnæ opposite, horizontal, from a hastate 

 base, i. e. the basal pair of segments are always enlarged, gradually tapering to the 

 acuminate apex. Segments short, obtuse, with few (4 — 7) pair of simple veins. The 

 underside in general glandulose, and the vascular parts more or less pilose. Texture 

 firm, sometimes nearly coriaceous. To this species I refer a great many forms 

 showing all the characters named above, but varying in size, texture, pubescence 

 and the presence or want of an acrophore. Yet, the most different forms are con- 

 nected by numerous transitional forms, and it seems to me rather impossible to 

 distinguish good varieties, with exception of the var. geraensis described below. 

 Looking again and again over the numerous specimens examined, I believe I am 

 able to point out a somewhat constant ditTerence between the West-Indian and the 

 Brazilian plants, while the specimens from the continent outside Brazil are almost 

 intermediate between the two extreme forms. As this difference covers a ditïerent 

 geographical distribution, I find it right to separate out the Brazilian form as a 

 variety. The two forms may be distinguished as follows. 



1. lypica ( A. conter minum WiUd. = A. pohjphyllum Klf.) — [Fig. 15J. Smaller 

 than the next with the pinnæ longer and broader (6 — 10 cm. long by 1cm. broad); 

 texture thinner, the edges not revolute; no aërophore; segments oblong, obtuse, 

 rather oblique. Underside densely glandulose; pubescence sparely; 4-6 pair 

 of veins. 



2. var. riviilorum (Raddi) C. Chr. apud Rosenstock, Hedwigia 46: 120. 1906 — 

 [Fig. 16]. Leaf upto 1 m. long, 7—10 cm. broad, more rigid with revolute edges. 

 Pinnæ short, 3—4 cm., very acuminate, often with an aërophore at their base, 

 costæ and veins more or less hairy. Segments short, nearly triangular, a little 

 oblique, about 4 pair of veins. 



This form very common in southern Brazil is no doubt P. riviilorum Raddi, 

 Raddi's figure agreeing very well with it. Mettenius considered Raddi's plant a 

 distinct species, well characterized by the presence of an aërophore; but his Aspi- 

 dium riviilorum includes the numerous forms which I now refer to D. pachyrachis, 

 a species with only a tew pair of reduced pinnæ. One often finds, however, in 

 the larger specimens of this variety a distinct aërophore, and Prof. Rosenstock, 

 thinking that this character is accompanied by a nearly total want of hairs, has 

 therefore named such forms var. Mettenii Ros. Hedwigia 46: 120. 1906. Still it has 

 been impossible for me to find out constant characters by which the proposed 

 variety can be determined. 



A form distinguished by its size is var. longissima n. var. 



Columbia: Santa Marta, leg. H. H. Smith n. 1008 (HC). 



38* 



