T. F. SMITH ON DIATOM STRUCTURE. 129 



you cannot see clearly, you see what you want to see, whether 

 beads or perforations. 



Now, the structure of Aulacodiscus KiHonii is very fine, and 

 stands to-day in relation to the defining power of the widest 

 angled oil immersions as Coscinodiscus oculus iridis stood in 

 relation to the objectives of 1873, just outside the limits of 

 their power. And as we cannot in this case trust to appear- 

 ances, analogy should be our guide, and we must argue from 

 the known to the unknown. Nobody now looking at a Coscino- 

 discus with an oil immersion would conclude that the hexagons 

 were beaded, and I think we may reasonably reject the black 

 dots when we see them in the finer diatoms. Of course, I do 

 not reject the paler markings, as they are in accordance with 

 well-known structure, and belong to the membrane on the top 

 of the hexagons. The inner layer of eye-spots is the only thing 

 clearly seen ; and, given your layer of eye-spots, to construct 

 your diatom you place your hexagons on this layer, and then 

 stretch a fine perforated membrane on the top, and that is all. 

 But there is other evidence on which I rely, and for this proof 

 I refer you to the January number of our Journal for this 

 year, where, in Plate II, figs. 8 and 9, you will find an Aula- 

 codiscus, which the authors of the paper have named cellulosus. 

 Fig. 8 is the whole disc magnified 400 diameters, and Fig. 9 a 

 part magnified 1,000. Both figures show clearly the straight- 

 sided cells, and Fig. 9 shows fine dots inside the hexagons ; 

 and although I have not examined the diatom, I have no doubt 

 but what it will show all the details belonging to the hexagonal 

 structure, from the eye-spots in the inner layer to the fine 

 perforated membrane in the outer. Another thing which con- 

 firms me in this opinion is the centre rosette, which this 

 diatom has in common with Aulacodiscus Kittonii ; and as far 

 as my experience goes, this always means the same sort 

 of structure. 



I have not this last diatom under a microscope at present, 

 but I have a very interesting slide through which I shall be 

 most happy to pilot you directly. When I bought it, it formed 

 a symmetrical group, but owing to the slipping of the cover it is 

 now in fragments, out of which I hope to be able to find enough 

 to confirm my views. You will see that some of the hexagons 

 are torn clean away, and some of the bits have turned over, and 



