134 T. H. BUFFHAM ON NEWLY-DISCOVERED PHENOMENA IN 



one case there were five, besides a zigzag chain of ten — one only 

 of this chain being attached to the female. It would be impos- 

 sible to understand the various positions where many are crowded 

 on excepting by the use of the binocular. The appearance with 

 a -i-inch objective and paraboloid x 120 is extremely curious as 

 the whole can be seen nearly in focus at once. In Fig 5 ( X 200) 

 is seen the earliest stage of conjugation, only that for clearness the 

 male frustules on the under side are not drawn. 



An interesting point now suggests itself. One cannot yet foresee 

 whether one or two sporangia will be developed. Prof. Smith only 

 knew of the latter result, but the instances are about equal. 

 What, then, is the determining cause ? As far as I have seen 

 there is no case where conjugation occurs excepting when the 

 female frustule is greatly widened and presents the broad band 

 mentioned above. This may be a stage in self-division, yet there 

 is no trace of new valves, and the band is noticeably wider than in 

 those vegetating frustules which goon self- dividing. Prof. Smith 

 says (loc. cit., p. xiv) : " Self-division occurring during the pro- 

 gress of conjugation the endochrome becomes segregated in the 

 very act of intermingling, and a single frustule whose contents 

 have been already differentiated gives rise to two sporangia as in 

 Achnanthes and Rhabdonema" It will already have seemed 

 probable that this description does not include the true process of 

 conjugation, but I gladly note that it suggests the explanation we 

 want. It is a fact that in about half the cases only that half of 

 the frustule which is nearer the base of the filament produces a 

 sporangium, as shown in Figs. 6 and 7. The terminal half — 

 although having males attached — falls away. We can only infer 

 that its vitality was not enough, probably from the nucleus not 

 having divided, to carry on to its completion the production of a 

 sporangium. If, however, the nucleus has divided — or in whatever 

 way the necessary vitality of the upper half has been obtained — 

 there will result two sporangia, as in Plate VIII., Figs. 1 and 2. 



.Returning from this digression to the course of development of 

 a single sporangium following the earliest stage shown in Fig. 5 

 we have already noted that the terminal half of the female frustule 

 falls away. The isthmi attaching the male frustules enlarge, the 

 endochrome contracts, and the central pale nucleus is lost. The 

 endochrome of the female moves towards the band, and the now 

 open end of the latter secretes a gelatinous portion which closes 



