286 MEMOIRS OF THE CARNEGIE MUSEUM 
extends upward as far as the base of the primary limb. Its distal margin in the 
outer gill is fused with the mantle, and thus a small space is cut off from the bran- 
chial chamber, which appears as a longitudinal canal along the base of the outer 
gill—the swprabranchial or cloacal canal of the outer gill, into which the water- 
tubes of this gill open. In the case of the inner gill, the secondary (inner) lamella 
is always connected anteriorly with the wall of the visceral mass (abdominal sac) 
for a shorter or longer distance; behind this connection it is sometimes free as 
far as the posterior end of the foot; and behind this point, the two inner laminz 
of the right and left inner gill are connected with each other up to the posterior 
end. Thus the inner gill also has a swprabranchial or cloacal canal, which is double 
(right and left) anteriorly, lying on either side of the foot and abdominal sac, 
but becomes united back of the foot into one cavity. In many cases this canal is 
not entirely closed, but communicates with the branchial chamber through a slit 
along the sides of the abdominal sac. 
There are a number of variations in the development of the characters men- 
tioned in the last paragraph. 
The outer lamina of the outer gill is united with the mantle up to the posterior 
end of the gill in practically all species known to me, the only exception being 
in the case of Margaritana margaritifera. Here (Compare fig. 6) the outer lamella 
is free posteriorly for a considerable distance, and the posterior end of the gills 
assumes in consequence a very peculiar shape (See below, under “‘diaphragm’’). 
As has been stated, the connection of the inner limb of the inner gill with 
the abdominal sac is variable. In the most primitive condition, such a connection 
is present only at the anterior end of the gill for a short distance, and in this case 
the inner lamina is free and leaves a slit-like communication open between the 
suprabranchial canal and the branchial chamber. In other forms, this slit is more 
or less, often entirely, closed; and then the whole inner lamina of the inner gill 
is connected, first to the abdominal sac, and farther behind, back of the hind end 
of the foot with that of the gill of the other side, and thus the suprabranchial 
canal of the inner gill is entirely separated from the branchial chamber. 
Long ago Louis Agassiz called attention to these differences of structure. Later 
on their value was doubted by Lea as well as by Simpson, and Sterki also does not 
consider them as of much systematic significance. I have investigated my material 
with reference to this question, and find that the connection of the inner lamina 
of the inner gill with the abdominal sac is indeed variable, and that sometimes it 
is differently developed in the same individual on the right and left side. But it 
is only in certain forms that this variability occurs. In others this character is quite 
