164 BULLETIN 100, UNITED STATES NATIONAL MUSEUM 



considerable size, but because it arrests attention by its fine prismatic 

 coloring, the main portion of the valve being a bluish-green and the 

 more finely beaded angles a soft buff or straw color. 



It is much like the figure in Schmidt's Atlas (pi. 79, fig. 1) but not 

 quite identical. Schmidt incorrectly calls this Amphitetras graef- 

 feiana Witt. It may be well to state here that the true A. graeffeiana 

 Witt is a very coarse diatom, with well separated rectangular beads 

 (figured and described in Journ. Mus. Godeff., 1873, p. 69, pi. 8, figs- 

 2a-b). Witt says of it "Errinert in der Form an Tri. formosum var., 

 Bright (Micro. Journ., vol. 4, p. 274, pi. 17, fig. 8). Kann aber nicht 

 mit demselben vereint werden da diezellige Strukturvon Tri. formosum 

 (1) viel feiner, (2) nicht so deutlich radiirend ist." Perhaps the 

 Schmidt figure represents a quadrate form of T. formosum Brightwell. 

 All these belong to what we may call the T. arcticum Brightwell group, 

 including T. antarcticum Janisch, and some diatomists would group 

 them under that name. After a careful examination of T. arcticum 

 T. formosum Brightwell and this new species, numberless specimens 

 of all these being available in Philippine Islands material, I have come 

 to the conclusion that there are three well defined types which it 

 would be much better to hold as separate species. They are: (1) T. 

 diaphanum Mann, as above, the markings of which are very delicate, 

 composed of radiating rows of minute beads, and having at the center 

 an evident cluster of small spines. (2) T. formosum Brightwell, more 

 coarsely marked with a hexagonal radiating network, with no central 

 spines but generally having a slightly depressed central area in which 

 the markings are imperfectly formed beads or blotches; well illus- 

 trated under that name in Schmidt's Atlas, plate 79, figures 2-3. 

 The network is much finer than that of the next, and is never or 

 rarely filled with a secondary set of markings consisting of minute 

 beads appearing as a ring within each of the areolations, as is always 

 the case in the following: (3) T. arcticum Brightwell, much coarser, 

 with secondary internal beading within or, as a matter fact, beneath 

 the areolation, the center of the valve generally having a minute 

 hyaline space (Schmidt, Atlas, pi. 79, figs. 5-8, and pi. 81, figs. 3-4). 



As the old genus Triceratium is a hopeless complex of disassociated 

 forms, chiefly Biddulphiae, I see decided advantage in accepting 

 Cleve's suggestion of placing in the genus Trigonium that residue of 

 Triceratium which is BiddulpMoid, but entirely destitute of the horns 

 or other processes at the angles, a salient characteristic of the true 

 Biddulphia. Most members of Trigonium are uniformly triangular; 

 a few, like the above, show occasional specimens with four or more 

 angles. 



In this connection it is interesting to consider the statement of 

 Castracane in the Report of the Challenger Expedition, page 113, that 



