MOUTH-PAETS OF THE SIIOHE CRAB. 131 
leg and the gill which adjoins it in front ; and through this it passes into the 
hypobranchial space. The water which enters the opening iDehind the 
cheliped flows similarly forward over the arthrodial membrane, which is 
shaped to provide a kind of duct for it, to the entry of the hypobranchial 
channel between the anterior arthrobranch of the cheliped and the posterior 
arthrobranch of the third maxilliped*. At this very large portal of the 
hypobranchial space it meets the water from Milne-Edwards's opening. 
From, that aperture the current passes, when the third maxillipeds are 
opposed, obliquely backwards through the portal just mentioned, though a 
little may stray forwards to pass between the arthrobranchs of the third 
maxilliped. When, "liowever, the maxilliped is divaricated and the opening 
thus fully uncovered, water passes also to the channels between the arthro- 
branchs of the third and second maxillipeds, and to the podobranehs of those 
limbs. The water which enters the opening behind the last leg is prevented 
from flowing directly forwards by a ridge on the epimeral region, passes over 
the articulation of the last leg, which is not prominent like the others, joins the 
stream which enters between the last two legs, and reaches the hypobranchial 
space by passing forwards tlirough a definite entrance formed by the shaping 
of the hinder leaflets of the last gill at a spot near ils base and at the end of 
the longitudinal corridor mentioned above. The arrangements may be 
summed up by the statement that the water which enters over each limb 
passes into the hypobranchial space by an opening between its gill and the 
gill of the limb in front of it, except that most of the water entering above 
the third maxilliped passes behind its gills. 
In each hypobranchial channel, the water flows, I believe, upwards, mingling 
to some extent under the gills with that in the adjacent channels, and 
diminishing as it goes by loss between the leaflets of the gills to the 
epibranchial space. Since this loss is hindered by friction against the 
leaflets, the water does not all escape till the top of the gills is reached, and 
is thus distributed over as wide a gill-surface as possible. 
The course of the water over the outer surface of the gills presents a very 
difficult problem. The shape of the chamber gives no convincing indication 
of the direction in which tlie stream gets through it, and its size^ as a whole 
and in each part, depends upon the changing form of its roof. I can only 
offer some suggestions on this subject. 
Normally, the water will pass direct from the point at which it issues fioni 
between the leaflets of the gills to, the collecting-space. If, however, the 
roof be lowered on to the surface of any part of the gills, the water which 
is passing through them is probably got away by the epibranchial channels. 
The appearance of these structures strongly suggests that a current flows 
* The entry to the hypobranchial channel between the arthrobranchs of the cheliped is 
very small, and that channel probably receives its -water from the longitudinal corridor. 
9* 
y 
( 
