DEVELOPMENT OF CUCUMARIA ECHINATA 203 
directed transversely. The second broad part then divides into 
three lobes almost equal in breadth, while another lobe arises 
on the left edge, behind the first lobe (Pl. 9, fig. 19). Thus the 
hydrocoele is now fan-shaped, with a narrow conical process 
directed postero-dorsally, which is the future pore-canal, and 
an expanded anterior margin, wavy with six lobes as just 
deseribed. From that transverse lobe, at first numbered third, 
is formed the mid-ventral radial canal (mv), while the other 
five lobes are rudiments of the five primary tentacles (tf). Except 
the mid-ventral one no other radial canals appear at this stage. 
The pore-canal opens to the exterior through the dorsal pore 
about the end of these changes (dp). 
Kowalewsky (1%, p. 4) and Selenka (45, p. 171) are 
of the same opinion that,inC. kirchsbergiiandC. planei 
respectively, there are first formed only three tentacles situated 
dorsally, and the remaining two appear after the hydrococle 
ring has closed. Ludwig (22, p. 608) found, on the contrary, 
that in C. planci the five primary tentacles appear simul- 
taneously as outgrowths of the radial canals. This divergence 
in view from other observers results probably from the fact that 
he originally dealt with eighth-day larvae without examining 
any earlier stages. I agree with Kowalewsky and Selenka 
in assuming the appearance of the tentacles to be earlier than 
that of the radial canals, but those three lobes which first 
appeared are, In my opinion, not the dorsal three of the primary 
tentacles. Ludwig’s account of the early features of the hydro- 
coele is very incomplete owing to the lack of any intermediate 
stages. According to him the hydrocoele is of an irregular 
horseshoe shape, whose arched part les dorsally, the right limb 
is short, stretching obliquely antero-ventrad (‘nach vorn und 
unten ’), and the left limb is longer, stretching postero-ventrad 
(‘nach unten und hinten’). I never observed such a condition, 
and am convinced that he was in error in these statements, 
from which he drew an incorrect conclusion that the hydrocoele 
ring probably closed on the right side. 
Enterocoele.—psoon after being separated from the hydro- 
coele the enterocoele divides into two vesicles, one larger and 
