Some Methods of Preparing Diatoms , dec. Bij G. II. Gill. 427 
allow the whole to stand for some hours. Wash and levigate as in 
the other methods. The mercurous sulphide thus formed is a black 
amorphous precipitate, which fills the “ lacunae ” of the diatoms with 
an almost completely opaque stopping. Mercuric sulphide is apt to 
become red and crystalline ; hence the necessity of the precautions to 
avoid the conversion of one into the other, which are detailed above. 
The only fault of this method is that the sulphide is somewhat apt to 
clot and become difficult to remove from the outside of the valves by 
washing. Perhaps this would be avoided by using weaker solutions 
than those I have worked with. 
(4) Silver Nitrate Method. — A strong solution of silver nitrate 
(about 100 grains to the oz.) may be substituted for the mercurous 
nitrate, but on the whole does not serve so well except for those 
diatoms having the finest markings, e. g. Pleurosigma angulatum. 
The silver sulphide formed is brown and less opaque than the mer- 
curous sulphide, but is not so apt to clot over the surface of the object. 
By none of these methods will every diatom in a batch be equally 
well charged. 
Diatoms treated by one or other of these methods exhibit very 
clearly that all “ striae,” “ dots,” &c., are, as stated in the first para- 
graph, cavities of some kind, which, in default of a better name, 
might be called “ lacunae ”or “ pores.” 
Whether these lacunae are complete perforations through the 
silicious test or mere pits or depressions on the inner or outer surface 
of the valve, or more or less flask-formed cavities communicating by 
one or more canals with the inner or outer surface, or with both, 
cannot at present be resolved with any degree of certainty in the case 
of those diatoms which have the finer markings. But in the case of 
some large Coscinodiscs it can be shown that the valve has a structure 
which may be described as cellular. Where the areolae are widely 
separated from one another, a fragment of a charged valve viewed 
edgeways presents the appearance of a number of mammaeform 
cells springing from the inner side of the outer face of the valve by 
their wider extremity, and terminating in a more or less conical 
perforated apex at the end facing inwards. Fig. 1 shows a valve of 
this description on the flat. All my edge specimens have spoilt 
themselves by rolling over. 
Where the areolae are very close together, so as to cause one 
another to assume the hexagonal form, the cells which constitute 
their prolongation partake of the same form, and their inner faces join 
together to form a perforated plate of considerable substance. The 
whole structure presents a close resemblance to a single layer of 
honeycomb cells with their cappings and bases complete but per- 
forated. Figo 2 exhibits an edge view of a fragment showing 
this structure. 
The outer face or surface of these cells, very commonly if not 
universally, consists of a thin silicious membrane pierced with a 
2 h 2 
