MARINE DIATOMS OF THE PHILIPPINE ISLANDS 165 



"we can not believe that the same species can assume sometimes one 

 form and sometimes another, or that from the same Stictodiscus 

 sometimes discord and sometimes triangular or polygonal forms arise." 

 The first half of this statement is not true; nor is the other half true, 

 unless he means by " the same Stictodiscus " the same individual. In 

 that case it evidently is true; for the usual method of diatom multi- 

 plication always results in the new individual having one old valve of 

 the parent frustule and one new valve. Consequently no biangular, 

 quadrate, or polygonal new valve could match up with a triangular 

 old valve, and any deviation in the number of sides or angles would 

 be impossible. But the same species with two, three, four, or m'ore 

 sides is so common that every diatomist is familiar with the fact. 

 Thus in some of the Pacific Ocean dredgings examined for my paper 

 on Diatoms of the Albatross Voyages the triangular and quadrate 

 forms of Biddulphiafavus (Ehrenberg) Van Heurck were both extremely 

 abundant. It is easy to understand the change from a triangular 

 to a quadrate or pentagonal form if we remember that in the auxo- 

 sporial method of reproduction we have all the necessary conditions for 

 this modification of contour without the loss of those other factors on 

 which depend the species' true characteristics. It was the failure to 

 take this into account that misled Ehrenberg into giving over one 

 hundred specific names to the same species Actinocylus ehrenbergii 

 Ralfs, and that lured De Toniinto the muddle of splitting up several 

 clearly defined species and grouping them in two impossible genera, 

 Amphitetras and AmpMpentas, according as individual specimens hap- 

 pened to have four or five angles. (See De Toni, Syl. Alg., p. 911.) 

 And as a triangular or other angled form of a species starting out 

 from its auxoporial original would retain that form in all its subse- 

 quent multiplications by fission, we see the reason why some one of 

 these is often abundant at a particular place in excess of, or even to 

 the exclusion of, the others. Thus in the case here considered, the 

 quadrangular phase shown in the illustrations is more frequent in the 

 Philippine Islands than the triangular or pentagonal forms. The 

 same is true in the case of T. hicoronatum where, as already stated, 

 the generally rare biangular form outnumbers the triangular. 

 Type— Cat. No. 43694, U.S.N.M. 



TRIGONIUM DISSIMILE (Grunow) Mann 



(Schmidt, Atlas, pi. 81, fig. 5.) 

 See under T. latum. 



TRIGONIUM DULCE (Greville) Mann 



(Micro. Journ., 1866, p. 9, pi. 2, fig. 20.) 



The type came from a fossil deposit at Barbados. 



