I9I3- No. 8. 



JAPANESE DIATOMS. 



rr 



The hexagonal foramen becomes more hke a narrow sht in a compact 

 chain, and the cross sections of the frustules become rounder as they 

 decrease in diameter. Neither character is by any means specific in the 

 case under discussion. 



There are, however, certain 

 local variations to be noticed, 

 though they are not strictly con- 

 stant. The thick interKing setæ 

 are much longer in the specimens 

 from the warmer seas, and have 

 sharph'-defined minute processes 

 on them, arranged spirally. 1 he 

 verrucose dots on the margin of 

 the primary valve of resting- 

 spores are nearly constant in the 

 Atlantic forms. In the material 

 from southern Japan, they are 

 ofcen quite imperceptible even in 

 the ignited preparations. On the 

 other hand, in the specimens from 

 Volcano Ba}' and Yeddo Bay, 

 the primary valve has not only 

 the small dots at the margin, but 

 also short spines distributed on 

 the upper valve. d 



Verv often the thickened f'ig- 4- Chaetocrras roniprcssinn Lauder. 



setæ at first diverge for a short 

 distance, and then run parallel 

 with the axis of the chain in 

 one direction. 



Loc. Volcano F3ay; Echigo 

 Prov., Mi.saki; Yeddo Bay; Mi- 

 sumi (Yendo). Tateyma (Oka- 

 mura). Enoshima; Akashi (Schrø- 

 der). Formosa Channel (Schrø- 

 der). 



(I. A portion of a chain from material 

 from Bay of Bengal. X 360. 



/). From material from Misumi. The 

 spore is perfectly smooth. X 360. 



r. From the same material. The 

 spores have minute processes on 

 the margin of the primary valves. 



(/. From a preparation by Cleve ; 

 material from Yeddo Bay. The 

 primary valves are finely spinul- 

 ated at the top, and the secondary 

 on the margin only. X 480. 



All from burnt preparations. 



