648 GRIGGS: SPECIES OF HELICONIA 
branch-bracts are broad like those of H. crassa, but no mention is 
made of the leaves, which would, therefore, be expected to be like 
those of H. psittacorum and very much narrower than those of the 
present species. 
\90 Heliconia Collinsiana sp. nov. 
The few plants of this species we saw were growing in a clump 
at the top of a steep bank by the roadside. They were lopping 
over upon the surrounding shrubbery as though unable to stand 
alone ; the leaves were spread out to catch the sun like those of 
H. Boringuena (f. 3). 
Whole plant about 5 m. long: leaf-bases 2.75 m., petiole I m., 
blade 1.25 m.; blade oblong, gradually narrowed to the tip, ob- 
liquely cordate at the base, decidedly inequilateral in the type, the 
sides being 17 and 22 cm. respectively, bright green above, below 
very glaucous, with a little fine fuzzy pubescence on the midrib, 
otherwise glabrous: inflorescence bright red, glabrous, 45 cm. 
long, pendent ona slender peduncle 30 cm. long ; branch-bracts 
lanceolate, the lowest linear-lanceolate, gradually narrowed to @ 
blunt tip, very distant on the very flexible but slightly flexuose 
rachis, lowest 30 cm. long, 4 cm. broad, those half way up 15 
cm. long and 3 cm. broad: flowers and their bracts not seen; few 
flowers (less than a dozen) to a branch-bract: pedicels 15-25 
mm. long: berries 15 mm. in diameter when fresh. 
v On the bank of the Cahabon river about five miles below the 
ford by which the road crosses to the town of Cahabon (70. 35 2). 
It gives me great pleasure to name this species after my friend and 
associate, Mr, G, N. Collins; but for his pointing it out to me I 
would have ridden by without seeing it. 
Donnell Smith's zo. 2702 from Masagua, Escuintla, is probably 
the same species. The leaf with his specimen is only 52 © 
long, 25 cm. wide, ovate and green below; its slender petiole, 
however, shows that it is not one of the large leaves of the plant, 
but from one of the small lateral suckers whose leaves are always 
shorter than those of the main stem. The same fact may also 
explain the absence of glaucescence, but the whole group is very 
variable in that respect. The inflorescence of this specimen f° 
sembles that of the type closely, but the branch-bracts are some 
what broader and each has more flowers. — 
According to the artificial key given by Schumann Heliconia 
Collinsiana could belong only with H. pulverulenta or 7. glauca, 
