Limits of Species in the JDiatomaceee. By Thomas Cornier. 461 
In previous papers which have appeared in the Society’s ‘ Trans- 
actions,’ I have endeavoured to show how misleading the indiscrimi- 
nate system of species-making has proved. In the first, ‘On the 
Unreliability of Characters generally accepted for Diagnosis,’ * your 
attention was drawn to the frequent relinquishment of species esta- 
blished by the earlier observers in such genera as Actinocyclus, 
Eunotia, &c., and to the great variation, both in relief and striation, 
which sometimes exists between the two valves of the same frustule, 
or between different frustules of the same filament. In the second, 
‘ % On the Development of the Young Yalve of Tracliyneis cispera,' f 
instances were given of the changes which take place in the course of 
growth of the valve, sufficient to cause what are merely stages of 
growth to be regarded as distinct species. In the third, ‘ On the 
Occurrence of Endocysts in Thalassiosira ,’ J I recalled the observa- 
tions long ago made by Mr. Lauder, and since confirmed by other 
observers, that there exists in the genera Bacteriastrum and Chseto- 
ceros what Mr. George Murray has termed a sort of dimorphism ; 
so that forms which had previously been referred, not only to 
different species, but even to distinct genera, were in reality but 
different phases in the life-history of a single species. 
This has led me to the conviction that to establish new species for 
each trifling variation, whether in outline, in striation, or in the re- 
lative size or form of blank spaces, is entirely misleading and detri- 
mental to the interests of science. While some diatomists have con- 
tinued to follow [this course, others have taken, more or less, the 
broader view which I have endeavoured to justify. The very different 
estimates which have been attached to the term “ species ” by the best 
known authors on diatoms, the unequal criteria which they have 
adopted, and the consequent confusion which has prevailed in the 
nomenclature of the order, will be seen from the following particulars. 
Prof. W. Smith, in his ‘ Synopsis of the British Diatomaceae,’ seldom 
explicitly mentions any series of intermediate forms ; but any charac- 
ter with respect to which such intermediates occur he usually refers 
to as “variable” or “ inconstant,” and does not rely on it for his 
diagnosis of species. Thus he writes of Tabellaria flocculosa and 
T. ventricosa, that “ with regard to the character on which Kiitzing 
relies as separating ” them, “ viz. the relative sizes of the central and 
terminal inflations, I have only to say that, in the numerous speci- 
mens in my possession, I find these characters exceedingly variable. 
... I therefore feel myself obliged to unite all these specimens 
under one designation.” § Again, concerning Pinnularia biceps Greg. : 
“ The absence or presence of the costae at the centre of the valve 
appears to be an accidental circumstance, as numerous frustules may 
be found in which the interruption is more or less complete. I can- 
not, therefore, at present admit Dr. Gregory’s new species.” || 
* Journ. R. Micros. Soc., 1894, pp. 428-32. t Op. cit., 1895, pp. 400-3. 
X Op. cit., 1896, pp. 489-91. § Svnops. Brit. Diat., ii. p. 45. 
[| Tom. cit., p. 96. 
