THE VICTORIAN NATURALIST. 103 



biologist — and, in fact, every worker with the microscope — has 

 been correspondingly benefited. 



But in regard to diatoms, as I said before, much remains to be 

 done. Finality, even in respect to a knowledge of their structure 

 merely, lias not yet been attained. To quote Dr. Dallinger 

 again: — "It would be extremely unwise to give absolute adhesion 

 to any present interpretation of what is now held by some 

 students of diatom structure, of no mean repute and of unrivalled 

 manipulative skill, to be the absolute structure of some of the 

 larger forms." 



We have also much to learn as to their mode of reproduction- 

 Reproduction — or, more correctly, multiplication — usually takes 

 place by autofission or division, and until recently was supposed 

 to result in the production of a series of diminishing frustules, 

 which, on reaching their minimum size, gave rise to what was 

 called an anxospore, which, having been preceded by conjugation, 

 was capable of reproducing two sporangial frustules of consider- 

 ably larger size, these again giving rise by fission to a new series 

 of diminishing frustules. This unconfirmed auxospore theory (for 

 it does not appear ever to have been founded upon actual 

 observation) rests — ist, upon the supposed inability of the 

 silicious walls of diatoms to expand ; secondly, on the idea that 

 all diatoms are capable of binary subdivision ; and, thirdly, on 

 the supposition that there is no mode of reproduction except by 

 auxospores. That the silicious valve is capable of distention 

 may be proved by an examination of some of the many varieties 

 of Melosira, particularly the varieties subflexilis and varians ; and 

 that binary subdivision cannot take place in genera with unequal 

 valves is universally acknowledged, as the two valves which are 

 formed in the process of binary subdivision must stereotype them- 

 selves on the old valves, and for this reason this process cannot 

 take place in those genera in which the axes cross one another, 

 like Campy lodiscus, and is confirmed by the fact that while there 

 are recorded not less than 75 observations of the process of 

 division, not one of them affords an exception to the rule. 



I may explain that binary subdivision consists in the elongation 

 of the cell and the formation of a hoop adherent to each end 

 valve, so that the two valves are separated by a band which pro- 

 gressively increases in breadth by addition to the free edges of the 

 hoops, until the original form of the cell is completely changed, 

 leading to a separation of the endochrome and nucleus into two 

 halves ; the primordial utricle folding in, and finally forming a 

 complete double partition, upon the adjacent surfaces of which 

 a new silicious valve is formed, so that the original frustule is 

 replaced by two frustules, each of which has one old and one 

 new valve. Generally speaking the new valves are a little 



