HOWK : PUVCOLOGICAL STUDIES 95 



tion, without a corresponding change in the herbarium label. The 

 two specimens agree essentially with the cyathiform plants from 

 Bemini Harbor, Bahamas, shown in figure 2 ofour I'I.ate 3, which, 

 by the way, were found growing with more numerous plants of 

 U. conglutinatii, one of which is photographed on plate 2 (the 

 largest plant). The plants shown in the upper half of our plate 

 3 are rather strikingly different from the typical Udotea cyathifor- 

 inis ; they are, in fact, so Penicillus-Wkc in habit that on finding 

 them we were inclined for a time to look upon them as a probably 

 new cyathiform species of Peniciihis, but on a microscopic exami- 

 nation nothing could be found to distinguish them satisfactorily 

 from the typical U. cyathiformis. They were growing on sand- 

 covered rocks at the low-water line, a position that would pre- 

 sumably have been unfavorable to a full and normal development. 

 Specimens from Culebra Island, Porto Rico [no. 4Jj8 in herb. N. Y. 

 Bot. Garden), apparently bridge over the gap in habit between 

 these short Penidlliis-\\V^ plants and the normal U. cyathiformis. 

 The plants from Key West, Florida, figured by W. H. Harvey 

 (Ner. Bor.-Am. 3 : //. 40C) as Udotea conglutinata are, with little 

 doubt, U. cyathiformis, as would appear from the very distinct 

 longitudinal striations of the flabellum in the natural-size figures 

 and from the rather straight and rigid character of the filaments in 

 the detailed enlargement. 



Photographs of representatives of Udotea co7igliitiiiata and U. 

 cyathiformis are given herewith and below are descriptions and a 

 key which includes the more important diagnostic characters. 

 The characters of the stipe-cortex, which we have illustrated by 

 drawings (plate 8, figures 8-13) we have found especially 

 reliable in determining occasional forms whose relationships might 

 otherwise seem a little doubtful. 



Flabellum plane; transition from stipe to flabellum gradual; fila- 

 ments of flabellum 28-60 /z in diameter, interwoven and tortuous, 

 usually forming a superficial tomentulose nap ; branches of the 

 corticating filaments of the stipe somewhat loosely and irregularly 

 fastigiate, the ultimate divisions mostly finger-shaped, subacute, or 

 taper-pointed U. conglutinata. 



Flabellum more or less cyathiform or open on one side and almost 

 plane (then usually a little concavo convex, at least at base); tran- 

 sition from stipe to flabellum abrupt ; filaments of flabellum nearly 

 straight and parallel, mostly 60-100^ in diameter; branches of 



