636 JAMES L. KELLOGG 
arrows point forward, as well as downward. The reason for 
such a roundabout course seems to be that particles could not 
well be carried directly downward across the current; so they 
proceed with it for a distance before turning posteriorly. 
Siphon membrane and waste canal. A waste canal occurs in 
several species. Wherever it is present there is also a well 
developed curtain-like structure (sm), at the base of the incur- 
rent or branchial siphon. This siphon membrane may be raised 
to admit the incurrent stream freely, or may be drawn down- 
ward so as to throw the stream toward the mantle edges. It 
seems probable that its function is to throw the current down- 
ward on to the mantle edges, and away from the gills, when 
much sediment is present. A relatively large amount of it 
would then be deposited on the mantle, and would quickly be 
taken backward to the small bay 6b. But a downwardly directed 
current from the incurrent siphon would tend to wash the mantle 
collections forward in the mantle chamber. In order to prevent 
this, the covering walls of the waste canal have been developed. 
The siphon membrane probably should be regarded as an organ 
developed to aid in the removal of waste matter. 
An examination of figure 38 will show an exception to the rule 
that siphon membrane and waste canal occur together; but this 
cannot be used as an objection to the explanation offered of the 
function of these organs. Cardium buries only a small portion 
of the shell in the bottom. The posterior end of the body lies 
in the water, and mantle collections, instead of being held to 
accumulate in a mass, to be discharged through the incurrent 
siphon, as is usual in lamellibranchs, are steadily carried over 
the mantle edges to the exterior, on a ciliated tract lying below 
the siphon membrane, where the waste canal is usually found. 
All the mantle collections finally arrive at the base of the in- 
current siphon, where they are out of the way of the incoming 
stream. In burrowed forms, the mass cannot be discharged 
between the ventral Shell edges as in Cardium, Unio, and others, 
because of the close investment of sand, or fused mantle edges. 
The walls of the siphon itself are not ciliated, for it would be 
impossible for cilia to move material against the stream in the 
