RESTITUTION-BODIES AND FREE TISSUE-CULTURE IN SYCON 301 
to form segments of spheres. The external epithelium of the 
solid remainder might be formed: (a) by collar-cells only, 
(b) by dermal-cells only, (¢) by patches of both. 
Other types have now been observed. The most interesting 
are the dermal blow-outs. These appear to be formed 
whenever the mass contains a preponderance of dermal cells. 
A mass of collar-cells generally fills most of the interior ; it is 
covered closely with a single layer of dermal epithelium, which 
at one point is swollen out to form a segment of a sphere 
which thus consists entirely or almost entirely of dermal 
cells. Often cells are to be seen adhering to its inner surface ; 
these were sometimes rounded and of a fair size, presumably 
typical amoebocytes, oftener of the minute elongated type 
which I have called finger-cells (see p. 304). A few detached 
cells might sometimes occur in the cavity. These were ocea- 
sionally seen to be forming spicules. A typical mass of this 
kind is shown im fig. 12. 
Shaking caused contraction and disappearance of the blown- 
out regions, as with the collar-cell blow-outs. 
One very peculiar mass was seen (at Wood’s Hole). This 
was isolated together with a number of others shortly after 
concrescence, when they were solid and irregular in shape. 
Four days later this was found to have a large hemispherical 
collar-cell blow-out, which in its turn showed a small dermal 
blow-out on one side. Under the surface finger-cells were 
visible. It would thus appear that local as well as general 
excess of dermal cells can occur, leading to the formation of 
mixed blow-outs. 
Previous experience (Huxley, 8) has led me to conclude that 
when a small proportion of dermal cells is present in a culture, 
they exercise an attraction for each other. This leads to the 
production of a few normal regenerates in a culture consisting 
mainly of collar-cell blow-outs. In a similar way this con- 
gregation of dermal cells can lead to the formation of dermal 
blow-outs. This was so in the mass shown in fig. 12. 
This and other facts would indicate that the formation 
of dermal blow-outs is mainly a matter of the number of dermal 
