ZOOLOGY AND BOTANY, MICROSCOPY, ETC. 
105 
of the valve, firstly, in producing an optical image which is more com- 
plete and hence more real, and further by reducing more nearly to a 
mathematical plane the portion of the valve that can be seen with one 
and the same focal adjustment. 
The new results obtained and confirmed by photography — which all 
serious observers now regard as the best criterion — still further simplify 
our opinions, and enable me to summarize them as follows : — 
(1) Diatom valves consist of two membranes or thin films, and of an 
intermediate layer, the latter being pierced with openings. 
The outer membrane, which is often very delicate, may readily be 
destroyed by the action of acids in cleaning, or by friction, &c. It may 
be also that this membrane exists only in a very rudimentary state. 
Specialists on this subject are generally agreed in supposing that these 
membranes may be sufficiently permeable to allow circulation by end- 
osmose from the interior to the exterior of the valve, though they have 
no real openings during the life of the diatom and whilst it remains intact. 
(2) When the openings of the interior portion are arranged in'alternate 
rows, they assume the hexagonal form ; when in straight rows, then the 
openings are square or oblong. 
The hexagonal form, which occurs so frequently in nature, seems to 
be the typical form of the openings in the interior portion, and this form 
obtains mostly in large valves, which are not otherwise provided with 
strengthening ribs. Even in the forms having square openings we 
frequently perceive deviations, and the tendency to recur to the hexagonal 
type on certain portions of the valve. It may be that the interior consists 
of several layers superposed, formed successively and very closely joined, 
but so far I have not met with any form exhibiting sui^erposed layers 
differing from each other in tyj^e. 
This description seems to me to represent in broad outline the struc- 
ture of diatom valves. But this structure may appear complicated, 
either by the presence of secondary internal valves (“ Regenerations- 
hiille or by deposits of silica on various parts of the valve. These 
deposits originate the “ thorns ” met with in divers forms (such as 
Triceratium), the rings found on the under membrane of certain forms of 
Coscinodiscus* and the anastomosed ribs of Navicula aspera Ehr. {Stau~ 
roneis pulcliella W. Sm.|) &c. 
All these deposits are merely secondary silicious products which have 
nothing to do with modifying the general structure of the valve in its 
primordial elements. 
Description of the Plates. 
Plate II. — (1) Amphipleura pellucida Kiitz. resolved into beads 
X 2000. The insufficient magnification shows a good general view, but 
the beads are not so sharp in the print as in the negative. 
I have observed the rings in Pleurosigma formosum referred to by Mr. T. F. 
Smith, but the new objective (1-G3 N.A., medium 2 -4) when employed on valves 
that were purposely broken, shows them lying Hat on the under membrane, precisely 
as in Coscinodiscus. Possibly these rings are portions of secondary valves. I have 
not been able yet to determine the point. 
t Tlie valves of Navicula aspera Ehr. appear at first siglit very complicated, and 
they have hitherto been erroneously ligured by all writers. ]\[y latest examinations 
would show tliat the appearances observed are due to the mixing up in vision of 
more or less distinct views of ribs or thickenings regularly anastomosed so as to form 
rings more or less alternate. 
