134 DK. L. A. BOKRAUAILE ON THR 
which it is almost or quite in contact. As the maxilliped returns, the 
epipodite is drawn back to its former position. The epipodite of the third 
maxilliped reaches over the epimeron of the cheliped to lie against the 
posterior region of the inner wall. When the maxilliped is in the normal 
position the epipodite lies in the longitudinal corridor at the base of the gills. 
If the movement of the limb were simply outward,, and if the epipodite were 
not hinged to it, the effect of its divarication from its fellow would be merely 
to press the epipodite more firmly inward against the wall above the bases 
of the gills, with which it would therefore tend not to be in close contact. 
Actually, however, the maxilliped, as I have stated above, rotates downwards 
as well as outwai'ds, and thus moves its epipodite upwards over the flank of 
the body, while the movement presently brings the stout base of the epipodite 
against the articulation of the cheliped, and this opjDOsition flexes it at its 
hinge and directs it outward as well as upward. Thds when the maxilliped 
is divaricated, the blade of its epipodite moves u[>wards and presses outwards 
against the under surface of the gills, which it sweeps and lifts a little from 
the thoracic wall, thereby flushing these parts with water. As the maxilliped 
returns to the normal position, the base of the epipodite is pressed against the 
arthrobranchs of the cheliped, and thus bent back into the longitudinal 
direction. The position of the ejiipodite of the second maxilliped upon the 
forwardly-facing epimera makes unnecessary any such special mechanism to 
bring it against the inner surface of the gills. 
No doubt the movements of the epipodites have the effect of mingling and 
distributing the water in the gill-chamber while they clean the gills, but it is 
not clear that this has any such physiological importance as has been 
attributed to it. 
The variations of t]ie gill-stream are a very striking feature of the 
physiology of the crab. They may be studied either by the carmine method 
or by watching the movements of the flagella of the exopodites of the 
maxillipeds, and, after making a window in the branchiostegite, those of the 
scaphognathite. Since the latter method studies directly the working of the 
principal agent, it is the most instructive, though it is open to the obvious 
objection that the action of the scaphognathite is affected by the operation. 
But when the shock of the latter has passed off, its effects are less serious 
than might be expected. The most remarkable feature of the action of the 
sca[)hognathite is its extreme sensitiveness. Any rough or sudden handling 
of the crab is liable to cause it to stoj) — in the face of danger the creature 
holds its breath. Yet this does not always happen when it is expected. 
Other variations are brought about by less obvious causes. The scapho- 
gnathites work independently, and either of them may cease working while 
the other continues. Their beating changes its rate and force from time to 
time in the same individual, and dili'ers in different individuals in the same 
vessel at the same time. Clearly the causes are sometimes internal : probably 
