838 
MR. G. J. ROMANES AND PROFESSOR J. C. EWART 
horizontal position—the ambulacral surface of the rays which are thrown back being 
then of course turned up, so as to face the surface of the water. If the rays succeed 
in finding a solid body, they will perhaps—though not necessarily—fasten upon it, 
and when their hold is secure, the rays which hitherto held the animal to the side of 
the tank relax their suckers, so that the Star-fish swings from its old to its new 
surface of attachment. The activity and coordination which the rays manifest in 
executing these various acrobatic movements is surprising, and give to the animal an 
almost intelligent appearance. 
If a Star-fish is turned over on its dorsal surface upon the flat floor of a tank, it 
almost immediately begins to right itself. Its method of doing so (see Plate 81, 
fig. 20) is to twist round the tip of one or more of its rays (a), until the ambulacral 
feet there situated are able to get a firm hold of the floor of the tank ; then by a 
successive and similar action of the ambulacral feet further back in the series, the 
whole end of the ray is twisted round ( b ), so that the ambulacral surface of the end is 
applied flat against the flat surface of the tank (c). The manoeuvre continuing, the 
semi-turn or spiral travels progressively all the way down the ray. Usually two or 
three adjacent rays perform this manoeuvre simultaneously; but if—as is sometimes 
the case—-two opposite rays begin to do so at first, one of them soon ceases to continue 
the manoeuvre, and one or both of the rays adjacent to the other takes it up instead. 
The spirals of all these rays being turned in the same direction (see a, b, c ), the result 
is, when they have proceeded sufficiently far down the rays, .to drag over the disc and 
the remaining rays (d, e), which abandon their hold of thefibottom of the tank, so as 
not to offer any resistance to the lifting action of the other rays; the animal, therefore, 
turns a complete somersault—the disc and inactive rays being thrown over the active 
ones with considerable rapidity. The whole movement—from the first twisting round 
of the tips of the active rays to the final turning over of the whole animal—does not 
usually occupy more than about half a minute. It will be seen that this whole move¬ 
ment implies no small amount of co-ordination, and it is therefore of interest to con¬ 
sider it in this connexion. Asa general rule, the rays are from the first co-ordinated 
to effect the righting movement in the direction in which it is finally to take place— 
the rays which are to be the active ones alone twisting over, and so twisting that all 
their spirals turn in the same direction. This, however, although usually, is by -no 
means invariably the case; for at the commencement of the righting movement 
different rays may act in antagonistic ways—twisting their spirals in opposite direc¬ 
tions, and doubling their ends under, without reference to the direction in which the 
somersault is eventually to be turned. But in all cases a definite plan, so to speak, is 
very soon made—the opposition rays, as previously stated, leaving go their hold, the 
antagonistic spirals of adjacent rays being unwound or reversed, while any antagonistic 
doublings are straightened out; so that the whole righting movement in fresh speci¬ 
mens never, at the most, occupies more than a minute. 
(B.) Sun-stars ( Solaster ).—All the remarks which have been made on the natural 
