80 
coste, from D. Blanchetiana by its soft, wool-like basal scales and sparse pubescence 
from D. subincisa by its basal scales and pinnules of basal pinne, shortened above 
not very produced below. With regard to pubescence it is, however, very variable; 
most specimens seen lack the long, whitish hairs beneath; and in some forms the 
cost above are shortly rusty-pubescent with Clenitis-hairs without long antrorse sete. 
I consider these differences little important; they are due, I think, to age and dif- 
ferent conditions of the growing-places. Also in cutting the species varies considerably, 
some specimens are nearly or fully tripinnate with the pinnules incised nearly or 
quite to the midrib. The largest specimen seen has the basal pinne 70 cm long, 
tripinnatifid, with the basal basiscopic pinnula 20 cm long. 
The species is widely distributed through the southern provinces of Brazil. I 
have thought it possible to distinguish local forms or varieties, but I have found it 
difficult to find good characters. In the mountains round Rio de Janeiro some forms 
occur that FEE has described as species (see synonyms); these »species« are, 
however, very slightly different from each other, and they all belong to the same 
form, which perhaps should be separated from the true connexa as a variety (var. 
eriopodia). This somewhat approaches D. grandis by the coste above being rusty-pu- 
bescent by short Cfenitis-hairs, while the leaf otherwise is quite glabrous, by the 
brighter, narrower, wool-like, basal scales and by the basal veins, which end in the 
leaf-tissue below the sinus. The specimens from Minas Geraes are the largest 
seen, more scaly than other forms and generally rather pubescent beneath on co- 
Stules and veins. P. Willsii Bak., a cultivated fern (Hort. Wills 1881), said to originate 
from Rio Janeiro, is to me a small form of typical D. connexa (Kew!). 
Below I enumerate some of the numerous species seen: 
Brazil: Para, Spruce nr. 22 (Kew) — Minas Geraes: LINDBERG nr. 547 (B), ie sige nr. 2184 (B, 
H, L, Rg, S, U, very large, perhaps iicce: 2186 (L, S, 2187 (Rg), 2188 (Rg, S, U) — Rio 
de Janeiro: Garpner nr. 134 (Kew), Guaziou nr. 1780, 2066 (H = Ph. at Fée), 2397 
(H, Rg = Ph. eriopodia oe ie (H, Rg= Ph. adnata Fée), 7246 (B, H), 7498 (H) — Sao 
Paulo: Reene xt nr. 50, fae hone! nr. 3091 (Rg), Heiner nr. 533 ae ULBRICHT nr. 
52, 94 (R), A. & C. cae; nr. ean . WackeET nr. 177 pt. (R), — Pa 4: Dusé&n nr. 
10339, 13531 (Rg), ANNIES nr. 36, se 5 (R) — Sta. Catharina: Ae nr. 193, 194 (B), 
SPANNAGEL nr. 105 (R), HAERCHEN nr. 79 a O. Moxrer nr. 31, 76, 123 (R), Gotrz nr. 39 (R) — 
Rio Grande do Sul: Jércens & Srien nr, 38, 48, 197,. 2381, "975 (R); Linpman or. A 869, 
Uruguay: Dep. Tacuarembo, Gruba de los helechos, C. Osten nr. 6619 (R), W. HERTER nr. 3534 a (RB) 
var. lateadnata (Christ). 
Syn. Phegeopteris late-adnata Christ, Ann. Cons. Jard. bot. Genéve 3: 36. 1899. 
Dryopteris latealata (err. typogr.!) C. Chr. Ind. 274. 1905. 
Differs from most Brazilian forms by its thinner texture, nearly all pinnules 
short-stalked, only the uppermost ones being adnate and decurrent, and by its 
Mey quite oT leaf; the costae above are finely puberulous, both surfaces 
