44 Papers from the Department of Marine Biology. 
The main criticism of the above statement which I should like to 
make is that Semper is mistaken in supposing that the crab moves 
about so little in the cavity that one or two small fissures or apertures 
alone suffice for the respiratory current. In my experience, in Pocil- 
lopora the cavity is large enough for the crab to move about a good deal 
in one plane and by observing the inhabitant of a nearly closed gall I 
could actually follow such movements. I have never seen such scars 
as Semper describes as due to the ‘‘continual scratching in one spot,” 
nor did Calman. In saying that “as in all the crabs of this group”’ 
(whichever group is meant) the respiratory current enters the branchial 
cavity anteriorly (close to the mouth) and passes out posteriorly, the 
case is exactly the reverse. Why it should be stated that both the 
exhalant and inhalant streams should always flow in one and the same 
direction I can not understand. As stated above, experiments with 
carmine particles show that the exhalant current flows upwards and 
outwards and the inhalant current laterally. Finally, in no galls that I 
have seen are the apertures reduced to two. In Pocillopora there is 
nearly always a considerable number (up to 10), of equal size and pre- 
sumably importance, situated along a line which encircles the gall but 
which is interrupted by the stalk. Only in a few old galls where inter- 
nal growth has restricted the space do some of the apertures close up. 
GALL FORMATION IN STYLOPHORA AND SIDEROPORA. 
A species of Stylophora (S. raristella Dep. var wilson J. S. Gard.) in 
the Cambridge Museum shows very well the formation of galls in 
branching corals where the branches are more massive than in Pocillo- 
pora and Seriatopora. The gall is produced apically, the branches 
broadening out as in Pocillopora; but when fully formed the gall is 
hardly distinct from the branch which bears it, its breadth and width 
being little greater. The completed gall differs, moreover, from that of 
Pocillopora in the fact that the approximated lips never show local 
fusion, a narrow fissure of uniform width remaining between them 
throughout the life of the gall crab. Here, then, the gall is truly 
formed by the broadening of existing branches and not by the produc- 
tion and fusion of new ones, and this is due to the larger scale of branch- 
ing characteristic of the genus. This description applies also to the 
genus Sideropora, as far as can be seen from the account and figures 
given by Semper of galls in Szderopora.palmata (p. 218, fig. 67); they 
also occur in S. digitata. 
ACTION OF RESPIRATORY CURRENT ON INDIVIDUAL POLYPS. 
Speaking of the thece which occur in the interior of the walls of the 
gall, Semper says: 
“Not one of the cups is normal in structure; the depression, which in the 
external polyps is very deep, is here no more than a shallow pit, and the 
septa (or party walls) of the cup are very slightly developed. Hence it follows, 
