428 
Transactions of the Society. 
VIII . — On the Unreliability of certain Characters, generally accepted 
for Specific Diagnosis in the Diatomacede. 
By Thomas Comber, F.R.M.S., F.L.S. 
(Read 20/A June , 1894.) 
When the systematic study of diatoms was first taken up, those 
engaged in it, being mostly concerned with the discrimination of the 
various forms which they met with, naturally adopted, for their 
diagnostic characters, such features as were the most readily per- 
ceptible. Amongst these, there were none more easy of recognition, 
and of description, than those regarding number and size ; and 
accordingly we find, in the works of Ehrenberg and Kiitzing, many 
specific characters based upon these. 
Any differences in the number, whether of the elevations on the 
dorsal margins of the valves of Eunotia and Himantidium ; of the 
rays or radial segments of those of Actinocyclus, Actinoptychus , 
Asterolampra, and other discoid genera; of the processes in Eupo- 
discns and Aulacodiscus ; or of the constrictions m Biddulphia and 
Eunotogramma ; were alike regarded as justifying the formation of so 
many distinct species. 
As an instance of the extent to which this was carried, may be 
cited Ehrenberg’s species quoted by Balfs as synonyms of his Actino- 
cyclus Ehrenbergii. They are 119 in number, and all based upon 
the varying number of the rays, from A. ternarius with three rays to 
A. Panhelion with 120. 
Nor was a difference in number regarded as constituting a 
character of only specific value, for, according to the number of angles 
which a valve might have, it was referred to the different genera 
Triceratium, Amphitetras, or Amphipentas; although such valves 
may now be regarded as not even specifically distinct. Ehrenberg also 
proposed the genera Tripodiscus, Tetrapodiscus, and Pentapodiscus, 
for what is now seen to be a single species, Aulacodiscus Argus 
A. Schmidt, according to the number of processes. Yet this may be 
different in the two valves of the same frustule. 
Prof. W. Smith, in his ‘Synopsis of the British Diatomaceae.’ 
followed the older observers in using such characters for the separation 
of species, but with some doubt, for he remarks regarding Actino- 
ptychus duodenarius, A. sedenarius, and A. octodenarius Ehrenberg, 
that the “ three forms are probably the same species in different 
stages.” He even goes so far as to refuse to recognize as distinct 
species Biddulphia tri-locularis, B. quinque-locularis, and B. septem- 
locularis Kiitzing, and refers them all to B. pulchella Gray. 
A few years later, llalfs, in Pritchard’s 4 Infusoria,’ rejected such 
characters, explaining that he considered species founded on them as 
“altogether unscientific and erroneous.” He therefore constituted 
