14 
Gregory, on a remarkable Group 
else to form a subgenus, characterised by its markings, 
that is, by its structure, including, as species, the chief types 
of outline to be found in the group. Some system of sub- 
division must be employed, in order to avoid confusion. I 
dare not venture, in the present imperfectly investigated con- 
dition of these forms, to prepare a permanent nomenclature 
for them. I content myself with directing attention to the 
subject, merely using the name N. varians as a convenient 
symbol for the group. T entertain no doubt that other 
analogous groups will be detected by careful examination. 
It is quite plain that in such groups form, shape, or outline 
cannot be regarded as a trustworthy specific character, 
although it is probable that many species exist in which the 
form, being constant, may be safely used in this way. 
It is an important question how far other characters, such 
as the number of the striae, or their arrangement, or the 
general aspect, may be depended on as specific characters. 
In this case, as in that of form, there seem to be many 
examples in which the characters are constant. But yet other 
cases occur where the tendency to vary seems to extend to 
these characters also. Thus, I have more than once pointed 
out that Pinnularia divergens^ W. Smith, which, as it occurs 
in the Premnay peat, has, according to W. Smith, 11 striae in 
•001'', occurs abundantly in the Mull deposit, and in many 
recent gatherings, with every detail, and especially the very 
peculiar arrangement of the striae, which have three centres of 
divergence, precisely as figured in the Synopsis, while the 
number of striae is from 24 to 26 in •001". I have now 
repeatedly met with both varieties, and although the number 
of striae seems never to fall so low as stated by Mr. Smith, yet 
there is a very marked difference. 
It would be out of place here to enter minutely on this 
question, which, however, is well worthy of attention. It will 
probably be found that in certain cases none of the characters 
above alluded to are constant ; while in many they appear to 
have a great degree of uniformity. 
But it is strictly within the scope of this paper to notice a 
group allied to that of N. varians, and differing from it chiefly 
in the number of striae. 
I have already stated that some of the forms now figured 
had been referred to such species as N. semen, N. rhyncocephala^ 
' P. gracilis, and P. radiosa, in all of which the normal number 
of striae is from 24 to 26 in -001." 
Now, I find, occurring generally with those forms which I 
refer to the group of N. varians^ others, having, like them, all 
or most of the varied shapes I have alluded to, and yet having 
