19 
a weathered, longitudinal surface, which certainly is due to the cause sug- 
gested for C. borealis (Cf. Plate III, figure 5a). It is not clear from the 
description whether corallites both distant and in contact occur in the 
same corallum, but certainly both conditions are considered typical of 
the species. From the figures it seems that Whitfield has two varieties. 
Figures 3, 5, and 6 are indistinguishable either by description or appearance 
from C. canadensis. There seems to be a well-developed inter-corallite 
“ wall ” in figure 3 and this recalls clearly one of Reusch's specimens from 
0vre Vikenes (P.M.O. 3), sections of which I have seen and which are 
figured by Kiaer (1929, Plate II, figure 5). Figures 7 and 4 cannot be 
distinguished from C. canadensis var. anticostiensis ; figure 7 might possibly 
be the forma arctica, but this depends on where Whitfield measured his 
corallite diameters. His maximum figure of 4 mm. makes it seem likely 
that forma arctica was represented in the collection. 
The only way in which specimens of C. borealis Whitfield without 
coenenchyme can be distinguished from C. canadensis Bill, from the 
description is in the diameter of the corallites. It has been pointed out 
several times in this paper that in the material studied no change in 
structure accompanies a variation in corallite diameter and that this 
variation is regarded as individual. The writer is unable, therefore, to 
accept Whitfield’s species as distinct from C. canadensis until it is shown 
from the original specimens that a peculiar internal structure is present. 
The coralla with small corallites have a coenenchyme, according to Whit- 
field, and these, again, have not been distinguished from the previously 
described C, anticostiensis, or C. canadensis var. anticostiensis, as under- 
stood here. 
Troedsson (1928, 124) followed Whitfield’s identification on corallite 
diameter. 
“ The collection from cape Calhoun [he writes! contains more than a dozen speci- 
mens. These are easily divided into three groups on account of the general appearance, 
which varies with the size of the corallites. Therefore, I have referred the group to 
as many species. This classification is also supported by the microscopic structure. 
In each group, however, the distance between the corallites differs strongly, and this 
is also the case with other features. The first species consi.sts of small corallites, 
diameter T5 to 2*5 millimetres, and has certainly to be identified with the North 
American Richmond form Calapoecia huronensis. The second, with the corallites hav- 
ing a diameter of 3 or 4 millimetres, is Calapoecia borealis Whitfield; and the third, 
with very large corallites, their diameter usually being 5 or 6 millimetres, rarely 4, 
is a new one, Calapoecia arctica n. sp.” 
It is with great diffidence that the writer ventures to disagree with 
Troedsson’s conclusions, which, in his opinion, were based on an insuffici- 
ency of material. The present study has shown how variable the genus 
Calapoecia can be, and that, although specimens in many cases seem very 
distinct when considered individually, they lose this individuality when 
viewed in a series of many examples. In the light of these results the 
writer cannot accept Troedsson’s statement that his classification is sup- 
ported by microscopic structure.” None of his excellent plates shows a 
single instance of internal structure that may be taken as implying^ a 
specific difference between his groups. Each thin section figured by him 
can be related directly to the scheme proposed above. Troedsson (123) 
apparently examined no types. In two specimens of “ C. borealis Whit- 
