110 Papers from the Department of Marine Biology. 
small Comatella maculata, which, after wearing the cork float in the 
live-car for 24 hours, was dropped into deep water. It immediately 
began swimming, but spasmodically and not in the beautifully codérdi- 
nated manner of a Stephanometra. Apparently these comasterids do 
not, under normal conditions, swim at all, but could with sufficient 
effort be taught to do so. 
On the other hand, all of the species of the other families of comat- 
ulids observed at Maér are good swimmers and do not creep about as 
the comasterids do. They are usually found at rest on the lower side 
of rock-fragments or on a branch of coral with the arms more or less 
erect, but sometimes the arms are opened out flat against the rock. 
They maintain their position by means of the cirri, with which they 
often cling so tightly that it is difficult to dislodge them without 
damage. Under no conditions did they seek to escape by swimming, 
but when once dislodged, swimming seemed to be their only means 
of locomotion. Placed in a basin or a deep bucket of sea-water, a 
very slight mechanical stimulus served to set them in motion, and 
swimming continued for a more or less brief interval, varying greatly 
with different individuals. It seemed as though the frequent contact 
with the side of the basin or pail served as a deterrent, and I think 
that they would undoubtedly have gone a greater distance in open 
water. Nevertheless it was clear that the individuals observed were 
“sprinters, ’’? and not long-distance swimmers, the movements being 
more like the flitting of small birds in shrubbery than like long-sustained 
flight. The gracefulness and beauty of the movements were their most 
notable feature, but their rapidity and force were also remarkable. 
In the ten-armed specimens, swimming was accomplished by using 
the arms in sets of five alternately, so that when arms 1, 3, 5, 7, and 
9 were brought up almost vertically over the disk, arms 2, 4, 6, 8, and 
10 struck backward forcibly, with pinnules fully extended, until they 
nearly met behind the cirri. Only a single stroke was made, but as 
arms 2, 4, 6, 8, and 10 were relaxed and drawn in and upward over the 
disk, another similar stroke was made by the alternating five arms. 
Of course the movement was much more rapid than a description indi- 
cates, but it decreased in rapidity as the comatulid became fatigued. 
At the start the strokes were at the rate of perhaps 100 per minute, 
but they rapidly dropped to much less than that and usually ceased 
altogether in less than a minute. Each stroke appeared to carry the 
individual about the length of its own arms, so that an individual with 
arms 50 mm. long started out at the rate of about 5 meters per minute. 
But the longest distance any specimen was seen to travel continuously 
was less than 3 meters. 
In the multibrachiate Stephanometride and Mariametride, the 
movements were very similar, but exceedingly difficult to analyze 
satisfactorily. The arms seemed to be used in sets of five in rapid 
