Form Regulation in Harenactis attenuata 387 
part from the aboral end, would undoubtedly be impossible if the 
original constitution and correlations existed. 
The occasional appearance of forked or branched oral-aboral 
tentacles inwhich the oral and aboral parts of the tentacle are sep- 
arate for more or less of its length are of some interest as “abnor- 
malities.”’ But forked and branched tentacles are by no means 
confined to the line of union between oral and aboral ends in rings. 
They occur not infrequently in other regions in rings, in Fig. 18, 
for example, in an oral tentacle, in Fig. 21 a and in Fig. 29 a and} 
in groups which are wholly on the oral side of the line of union. 
Moreover, they are often seen in other cases of restitution. 
Though their method of origin differs in different cases certain 
features are, | think, common to all, whether they are oral-aboral 
or in the usual position. [he growth of tentacles in restitution 
occurs chiefly in the more proximal regions, the distal portions of 
the tentacle being the first to appear: in fact, there is considerable 
evidence, which I hope to present at another time, that in at least 
some actinians the tentacles are throughout life growing at the 
base and undergoing atrophy at the tip. It follows from the 
method of growth of the tentacles that the branched portion was 
formed before the single basal part, 1. e., such tentacles began to 
form as two distinct tentacles but later united. In a number of 
cases this sequence of events has actually been observed, so that no 
doubt of its occurrence can exist. 
The factors which determine the occurrence of fusion or union 
between the two parts which are at first distinct are, so far as I can 
determine, the following: first, the diameter of the first portions 
of the tentacle to appear, 1. e., the distal portions, is usually less 
than that of later, more proximal portions, especially if the develop- 
ing tentacle is growing rapidly; if two regions of tentacle-formation 
are localized near each other they may form separate tentacles 
at first, but as each increases in size they may meet and unite to 
form a single region, so that further growth is single, not double. 
A second factor producing essentially the same result is the divi- 
sion into two parts of a tentacle-forming region in its earlier stages. 
Such division often occurs in restitution in consequence of folds 
and wrinkles in the body-wall which result from the contraction 
