MARINE DIATOMS OF THE PHILIPPINE ISLANDS 175 



of this, giving a suggestion of a stauros; in girdle view the two con- 

 vex halves are seen to be separateoVby the deep median sinus. 



Length 0.143; width 0.019; lines 10 in 0.01 mm. 



This diatom because of its median ridge and its coarse beading 

 resembles Navicula carinifera Grunow, except for its linear shape, 

 median constrictions, and tapering ends. It has some likeness to 

 T. lepidoptera (Gregory) Cleve, but its very coarse beading in pro- 

 portion to its size and its Achnanthes-iike constrictions at the middle 

 sufficiently separate them. In this last respect it reminds one of 

 " Achnanthes pennaeformis" Greville (Diat. So. Hemi., pi. 6, figs. 

 11-13). The specific name here given refers to the confusing qual- 

 ities, which seem to link it to Navicula, to Tropidoneis, and to Ach- 

 nanthes. It is rare. 



Type.— Cat. No. 43698, N. U.M.S. 



Genus WILLEMOESIA Castracane 



This genus, created by Castracane in his Report of the Challenger 

 Expedition, page 165, is not clearly described, although the first of 

 his three figures in plate 8 make sufficiently plain the nature of the 

 single species so far known. Rattray (Rev. Cose, p. 452) rejects this 

 genus and names the diatom Coscinodiscus humilis Rattray. It is a 

 serious stretch of the boundary of Coscinodiscus to include this form, 

 and the finding of three specimens in three separate dredgings from 

 the Philippine Islands shows we have to do with a persistent and 

 well-marked diatom. Van Heurck in his treatise, page 537, accepts 

 this genus. 



WILLEMOESIA ELONGATA (Grunow) Mann 



(Castracane, Chall. Exp., pi. 8, fig. 8; not 8a-86; Van Heurck, Synopsis, pL 

 125, figs. 14-17; Van Heurck, Treat., p. 537, fig. 284.) 



The two figures of Castracane's excluded above are suspiciously 

 like nondiatom plates frequently met with, one of which is well 

 illustrated in the same work, plate 10, figure 10. In no other 

 figures nor in the specimens I found are the borders hyaline, as they 

 are in these plates. They are also perfectly flat, not convex like 

 the present species. Rattray's specific name, humilis, should bo 

 rejected in favor of the earlier Coscinodiscus elongatus of Grunow 

 (1880), which is unquestionably the same diatom. As to the figures 

 in Van Heurck's Synopsis, which with some question he places in 

 Actinocyclus, it is a priori evident these unique specimens are sus- 

 piciously alike; and as for the "pseudonodule" because of which Van 

 Heurck with doubt assigns them to Actinocyclus, I am of the opinion 

 that the dot seen by Van Heurck is not a true pseudonodule, one of 

 my specimens having three such dots on one valve and none on the 

 other valve of the same frustule. I shall therefore assume, for the 

 present, that we are here dealing with a single species. 

 35035—25 12 



