26 
between the Sardnulae. He mentions two forms, the Swedish organwrij 
which corresponds to the more advanced of the examples described above, 
and the Norwegian form, considered by him not to be 8 . organum, cor- 
responding with the more primitive (Ca/apoecia-like). Evidently Troedsson 
had studied very few thin sections. 
It must be emphasized that these remarks on Sarcinula organum are 
made in order to compare that species with Calapoecia and not as a study in 
itself. The suggested conclusions would have to be checked with a study 
of a great many more specimens than the writer has been able yet to see, 
and also stratigraphical questions would have to be closely considered. But 
the facts gathered so far may be summarized. 
The more developed forms of Calapoecia have in common with Sarci- 
nula organum: a corallite stereozone and no epitheca; “ costse ” which ex- 
tend over a polygonal area around each corallite, this area being conti- 
guous in juxtaposed corallites; an inter-corallite region composed of 
horizontal diaphragms; and well-developed mural pores connecting the 
corallites with the ‘ coenenchyme.’ They differ on the following points: 
the septa in Calapoecia are twenty in number, those in Sarcinula are 
twenty-four; the wall in Calapoecia is much more openly cribriform than 
in Sarcinula (it is not inconceivable that a more C alapoecia~lik.e Sarcinula 
might be found in which the wall is less compact) ; this is because the septa 
in Sarcinula are in lateral contact for practically all their radial extent, 
whereas those of Calapoecia end free; the tabulse are constantly more 
remote and rarely incomplete in Sarcinula. 
The difference in number of septa seems the most important point here; 
for just as the astraeoid trend may be expressed by different lineages of 
Rugose corals, so in these earlier forms (Tabulates) the advanced Cala- 
poecia condition might have been reached by two different stocks at roughly 
the same time. The only structures that are present and, therefore, liable 
to vary are the same in each case — septa and tabulfe, the diaphragms being 
considered as only extra-stereozone tabulae. One variation, then, working 
on these two forms would be likely to produce a very similar result, and 
perhaps structures originally disposed in different ways would be brought to 
conform. Throughout the variations from typical C. canadensis to forma 
arctica the number of septa remains constantly twenty. S. organum has 
twenty-four. In view of these facts and possibilities Calapoecia and Sar- 
cinula cannot be regarded as congeneric. They are probably quite closely 
related, however. Here the matter must rest until a Sarcinula more primi- 
tive than the examples yet described is found. 
The Devonian Thecostegites bouchardi (Michelin) is closer to Sarcinula 
than to Calapoecia in the structure of its corallites; the intercorallite tissue, 
however, is not very dissimilar from some of the less regular examples of 
the latter genus, but the horizontal elements are more distorted (B.M. 
R6291). 
The nearest parallel to the “ wall ” structure of the typical C. cana- 
densis is seen in Cleistopora geometrica (Edwards and Haime), which was 
the subject of a recent paper by Smyth (1933, 169) . 
"A horizontal section of the common wall between two corallites (figure 10) 
composed of two sets of septa baok to back. Sometimes there appears to be a common 
zone between them in which the fibres are perpendicular to the plane of the wall, and 
