128 DR. L. A. BOERADAILE ON THE 
along the axis, between the leaflets. The arthrobranch of the second 
maxilliped has leaflets on the posterior side only, the anterior side, which is 
applied to the wall o£ the chamber, being flat; but along its edge an 
epibranchial channel runs as a gutter. The hypobranchial channels are part 
of the hypobranchial space and are in communication with one another 
under the gills, especially at the bases of the latter, which are there a little 
arched, leaving a longitudinal corridor in which lies normally the epipodite 
of the third maxilliped. The epibranchial channels are part of the epi- 
branchial space, and will communicate over the gills unless the roof be 
lowered on to the latter. At the hinder end of the chamber there is a space 
which contains no gills. This sj^ace is very shallow and in it lies the 
enigmatical fold of the body-wall known as the " pericardial lobe." It is 
shielded from the current entering over the last leg by a ridge which, as I 
shall presently show, directs the water forwards and downwards, and it 
probably plays no important part in the circulation of the water about the 
gills. The branchiostegite fits closely against the bases of the gills, which 
are flattened back to receive it. Above most of the rest of the surface of 
the gills the epibranchial space is probably as a rule deep enough to allow 
the epipodite of the first maxilliped to play freelj', but by the action of the 
dorso-ventral muscles varies in depth from time to time owing to circum- 
stances of which nothing is known, and which may be related to other 
functions than respiration. At the anterior end of the chamber, where the 
arthrobranchs of the third maxilliped and cheliped face partly forwards 
towards the roof as it falls to the opening of the exhal^nt channel, there is 
a deeper part of the cavity. Since its roof is flexible like that of the rest of 
the epibranchial space, it is possible that this does not always exist, Lut I 
have always found it, and I believe that it is kept in being, in its lower part 
at least, by the attachment of the roof to the arched sclerite that I have 
mentioned. It slants downwards, forwards, and outwards to the opening of 
the exhalent channel. From it water must be drawn by the action of the 
scaphognathite into the exhalent channel, which, as I have shown, caimot 
receive water from the region directly behind itself. Thus the water from 
all parts of the chamber must pass through this space, and, though it is not 
sharply defined from the rest of the epibranchial space, it may be dis- 
tinguished as the " collecting space." 
The gill stream *. The water which bathes the gills normally makes entry 
under the edge of the branchiostegite in the thoracic region and leaves in 
the same way in the preoral region. According to the classical account of 
this process, given by Milne-Edwards (12), the entry of the water takes place 
only by an opening which lies in front of the coxa of the cheliped and 
* It was not till I had written the paragraphs on this subject that I saw the paper of 
Mr. R. K. S. Lim, published in 1918 (10). As my work extends as well as coutirms 
that of Mr. Lim, I have left unaltered what I had written. 
