114 Papers from the Department of Marine Biology. 
at the bottom. Fishes, often of large size, always swam towards them, 
but when within a few inches turned and swam away. Either by 
sight, or some other sense, perhaps stimulated by some exhalation from 
the comatulids, these animals were recognized as inedible. 
FOOD AND FEEDING. 
Examination of the contents of the stomachs of living crinoids were 
made on two different occasions. One day two crinoids, which had 
been in the live-car for several days, were brought to the laboratory 
and the contents of their stomachs were carefully examined under the 
microscope. No essential difference between the two specimens was 
noted. In each case, the greater part of the food material was green 
alge, chiefly unicellular, but some linear forms (thread-algz) were also 
noted; a few diatoms were detected and some foraminifera; in one of 
the stomachs, several radiolarians were also seen. There was no indi- 
cation that other than living food material had been taken into the 
stomach. All of the material was virtually undigested. On another 
day a study was made of the contents of the stomachs of two crinoids 
just brought into the laboratory from the reef. In both individuals 
the food material was identical with that of the crinoids which had 
been living in the live-car. In one specimen a piece of a red alga was 
noted and in the other some fragments of minute crustacea. These last 
were the only particles of animal food noted and were the only things 
observed that could possibly have been taken in as dead plankton. 
It was made perfectly clear that the comatulids at Maér, at least in the 
dry season, are all vegetable feeders, and that the percentage of animal 
food is negligible. But it is possible, of course, that during the rainy 
season there may be a change in the composition of their diet; on that 
point there is no evidence to offer. 
It is not easy to determine beyond question the manner in which 
the food reaches the stomach. It is easy to suppose that the food is 
simply swept into the mouth by the current of water flowing along 
the ciliated furrows of the arms and disk, but I did not succeed in 
demonstrating such a movement, although there is no reason for 
doubting its reality. The multibrachiate forms keep the arms more 
or less widely spread out when they are at rest, but in the ten-armed 
species they are often more or less erect and sometimes quite rigid. 
In the aquaria and live-car the arms are frequently if not constantly 
curving in towards the mouth and the tips sweep the surface of the 
disk lightly. Individuals living on the lower surface of rock-fragments, 
since they are attached by the aboral side, rest with the ciliated grooves 
and mouth downward, away from the surface of the rock. It is 
therefore evident that, unlike starfishes and echini found in similar 
situations, they get their food from the water directly and not from 
the surface on which they rest. Taking all the facts observed with 
reference to the nature of the food, the position and movements of 
