286 MR. W. D, LANG ON GROWTH-STAGES [ Mar. 2, 
given :—“Save for the appearance of the dissepimental ring (a 
feature confined to the mature part of the corallum) none of the 
phases above described is truly ontogenetic, for, as will now be 
explained, they are intimately connected with the habits of growth 
of the corallum.” This is only an expression of the common difti- 
culty in the determination of species of any group, especially 
among fossils, the environment of which when alive is so 
uncertainly understood—that of knowing whether a given cha- 
racter 1s directly the result of heredity or caused by environment, 
These “habits of growth” are shown to be the external shapes 
of the corallum at a given size, two shapes being mentioned, 
Trochoid and Cylindrical. And it is obvious from the examination 
of simple corals that a Trochoid or a Cylindrical shape-stage 1s 
always present, and if both, the former always precedes the 
latter. For simple corals, then, two definite shape-stages may be 
postulated ; nor can the environment directly reverse the order 
in which they oceur. Other growth-stages in the shape of the 
corallum may be found in simple corals; for instance, an Inverse- 
trochoid shape-stage occurs in some Montlivaltia following the 
Cylindrical shape-stage, and a Discoid shape-stage in other M/ontli- 
valtia as an episode in the Trochoid ; but these need not here be 
considered. 
So far as Caninia cornucopie is concerned, it appears that the 
environment determines the size reached by the corallum before 
it passes from the Trochoid to the Cylindrical shape-stage. If this 
were all, Carruthers’ “ phases” in C. cornucopie might be adopted 
unhesitatingly as “truly ontogenetic” growth-stages, and the 
determination of the species only complicated by the wide limits 
of variation of the diameter, so that a comparatively large spe- 
cimen might be found in an early growth-stage, and vice versa. 
But it is stated that “it often happens that examples showing 
the long septa of the dwmonti-phase are found to have the am- 
plexoid septa,” characteristic of the next—the nystiana-stage— 
“‘in their lower portions.” This appearence of an older stage, 
earlier than, that is proximal to, a younger one, might seem to deny 
that the “ phases” are growth-stages. The sequence in which the 
“phases ” occur, however, is nowhere stated to be irregular, that is, 
haphazard and different in different individuals; on the contrary, 
except for the above modification, it is implied that the order of 
phases is in regular sequence. An explanation of the apparent 
irregularity is suggested by the behaviour of the coste in the 
English species of Chalk Parasmilia, which alter in their orna- 
mentation and general shape in a definite way when traced from 
the proximal to the distal end of the corallum. 
Before describing these it may be said that Carruthers mentions 
other difficulties in the way of accepting his “ phases” as true onto- 
genetic growth-forms. It appears that some internal characters 
change in correspondence with the change in general shape, others 
with change in the actual size of the corallum. Consequently, in 
extreme cases of non-correspondence between the size and shape 
