Affinities 119 



diatomaceous earth. The statement made by Castracane that he had found 

 the valves of diatoms in coal from the English Carboniferous strata, was 

 apparently made in error, and has never been verified (Cleve, '94 ; Seward, 

 '98). Edwards has stated that he has found the valves of diatoms in still 

 older rocks in New Jersey, but this observation also lacks verification. 

 Cleve has examined a large number of Silurian clays and limestones, 

 and also Rha?tic and Cretaceous rocks, without finding a trace of any diatom. 

 He also remarks upon the richness of the post-glacial strata of Sweden in 

 both freshwater and brackish diatoms. 



Cleve ('94) has emphasized the importance of the study of fossil diatoms 

 to geologists. He states that microscopical examination of the pre- and 

 inter- glacial deposits of northern Germany and Denmark has furnished 

 evidence that these strata were formed in inlets from the North Sea and not 

 from the Arctic Sea, and that accurate investigation of the geographical 

 distribution of the living freshwater forms 1 will enable the geologist to 

 ascertain the climate of the periods in which the numerous freshwater 

 deposits were accumulated. 



AFFINITIES. The Bacillariea3 Avere at one time included within the 

 Phseophyceae, but modern knowledge of their structure and life-history shows 

 them to be widely removed from the brown seaweeds. Certain authors have 

 placed them in close proximity to the Conjugate, and in 1904 Oltmanns 

 so far extended the scope of the ' Akonta? ' as to embrace the Bacillariese as 

 well as the Conjugate. 



Diatoms agree with Desmids in their unicellular and colonial habit, and 

 in the fact that each cell consists of a newer and an older half. They multiply 

 by cell-division in much the same way, but the comparison instituted by 

 Oltmanns between the intercalary bands of the diatom-girdle and the inter- 

 calary pieces of cell-wall developed in certain species of Closterium and Penium 

 scarcely holds good. The intercalary bands of the diatom are formed in 

 definite order during cell-division, and they are developed before the actual 

 connecting band of each new half is laid down. On the other hand, the 

 so-called ' girdle-band ' possessed by a few desmids is not a paired structure, 

 even though two segments of it may be present in an old cell ; but it is 

 a cylindrical piece of cell-wall, arising subsequent to cell-division, and inter- 

 calated between the new and the old half-cells. Moreover, less than one-half 

 per cent, of the known species of desmids ever develop such a structure, and 

 in those in which it occurs only one ' girdle-band ' may be formed during 

 a period covering upwards of twenty divisions. 



The cell-wall of most desmids is furnished with pores, but in contrast 



1 It is a surprising fact that the existing freshwater diatoms are not so well known as the 

 fossil forms. Our knowledge of them is very incomplete and far from accurate. 



