178 bulletin: museum of compakative zoology. 



Engleri, var. constricta, of which Fig. 14 is a gigantic form. The trans- 

 itions manifest themselves in all distinctive characters, viz., the shape of 

 the valve, its size, and the number of " costae " in a given area (e. g., 

 10 fi) ; this last character separates S. Engleri from the two others. 



If we admit that there are no distinct limits between the forms, I 

 think it is better to unite them as one species, retaining the most diverg- 

 ing forms as varieties, and I propose to name them S. nyassae 0. Mull, 

 with the var. Engleri (0. Mull.) and var. malombae (0. Mull.). 



Another question is, if we should subordinate them all under the 

 older species 8. bifrons (Ehbg.) Ktz. and S. constricta Ehbg. To 

 answer this question it would be necessary to take up for study a large 

 number of species of which I have no authentic material, but must rely 

 upon the drawings in A. Schmidt's Atlas, in W. Smith's British Diatom- 

 aceae, and in H. Van Heurck's Atlas, as well as earlier publications by 

 Ehrenberg, Kiltzing, etc. The chance for error is, I think, too great, and 

 I will not enter upon the matter, but will confine myself to the reduc- 

 tion already made. 



Besides, the large Surirellas of the plankton of the East African lakes 

 are so peculiar that they deserve to be distinguished, perhaps, also, by 

 maintaining the names given to them. 



Melosira. With regard to the Melosira forms of the Central African 

 lakes, much the same may be said, as in the case of Surirella. Our 

 knowledge of them is also due to Professor Otto Mliller, and in his 

 paper on Nyassa Diatoms (part II, 1904) he has given an exhaustive 

 study of the Melosira forms of this region. Further, he has examined 

 the Melosirae occurring in Victoria Nyanza slides sent him by me and 

 has given his notes on these forms in my paper (I. c, p. 338-339). 

 The most common form in these slides, which have been made from 

 Professor Borgert's material, is a slightly divergent form of M. nyassen- 

 sis 0. Mtill., named by Muller, var. Victoriae O. Muller. This form 

 occurs also in the samples taken by Dr. Agassiz, but is not the 

 dominant one ; it seems to attain its maximum later in the year. The 

 dominant Melosira of the February samples — and it is the most com- 

 mon species of the whole phytoplankton — is the same as Muller men- 

 tions as a new species in the Borgert samples, where it was not so 

 common. As it stands near M. granulata (Ehrbg.) Ralfs, I have re- 

 corded it in my paper as " M. aff. granulatam." Now when I have had 

 excellent and rich material at my disposal, I think myself justified in 

 describing it as a new species, which I name in honor of Dr. Alexander 

 Agassiz. 



