132 T. H. BtJFFHAM ON NEWLY-DISCOVERED PHENOMENA IN 



this composite frustule the granular contents are somewhat re- 

 stricted when moving in the chambers. 



The mode of self-division is similar to that of diatoms with 

 simple frustules excepting that the " connecting membrane " (or 

 " hoop ") may be formed at other parts of the widened frustule 

 than the middle. Nor must we overlook the gelatinous cushions 

 or " isthmi " which connect the four corners of adjoining frustules. 



To come now to the process of conjugation. Prof. W. Smith, 

 in his invaluable " Synopsis of British Diatomacege ' (Intro. Vol. 

 ii., p. 13.) says : " From a single frustule two sporangia are 

 produced in the process of conjugation : this takes place in 

 Achnanthes and Rhabdonema." How far I am compelled to con- 

 sider this description inadequate will presently appear. 



In March, 1883, I collected at Folkestone, Bhab, arcuatum 

 growing on Cladophora, Polysiphonia, and Ceramium in rich pro- 

 fusion. After carefully washing in sea-water I placed them in 

 glycerine to await examination. This was done after mounting in 

 the same medium, but chiefly in Deane's gelatine. As objection 

 might be taken to results obtained from preserved and no longer 

 living specimens I may here remark that the only effects of 

 glycerine, as far as I can see, are to make the endochrome more 

 transparent and slightly greener, and but very little to alter the 

 disposition of the granules. And Deane's gelatine, after glycerine, 

 very well preserves the same appearance. At all events it will 

 not be argued that these media will differentiate substances that 

 are identical in structure and condition. Relying to a large extent 

 on this principle I am not wholly dependent on it. 



It would be of no interest to relate here the various attempts 

 made to reconcile one's earlier observations with the statements of 

 Prof. Smith, and I therefore proceed at once to give the results of 

 the examination of many thousands of filaments, continued at in- 

 tervals since collecting the specimens, with here and there an 

 instance of conjugation, accumulating, however, in the aggregate, 

 to a considerable number, and exhibiting almost every possible 

 stage. In this examination I have been assisted by my friend Mr. 

 W. H. Gilburt, to whom I am indebted for several important 

 suggestions. 



I. There are filaments which differ from the ordinary kind in 

 consisting chiefly of frustules whose breadth (reckoned along the 

 axis of the filament) is usually about equal to the length, and the 

 terminal half-dozen or so are connected only by one angle to each 



