YOUNG STAGES OF LIMNOCODIUM AND GERYONIA, 195 
but states that its walls form as an upgrowth around the 
oral area, after the mouth has been formed by a rupture of 
the ectoderm and endoderm at the centre of that area. 
The observations which I have made upon young stages 
of the freshwater Trachomedusa, Limnocodium Sowerbii, 
lead me to think it probable that, after all, Haeckel’s 
observations are correct, and that the sub-umbrellar cavity 
is formed as a closed space between two layers of the ecto- 
derm. It would not by any means be necessary to accept 
the interpretation of appearances given by Haeckel to the 
effect, viz. that the endodermal sac is the sub-umbrellar 
cavity, but we have to suppose a step in development 
intermediate between his fig. 28 and 29 (of Taf. iv, ‘Je- 
naische Zeitschrift,’ vol. 11). In his fig. 28 the internal 
sac which is drawn 7s the archenteron or endodermal sac. 
In figs. 29 and 80 an ectodermal sac (formed, I would 
suggest, by a hollowing out of the ectoderm, probably 
without any opening to the exterior) has taken the place of 
the primitive endodermal sac, which has become flattened 
and otherwise modified to form the stomach and gastro- 
vascular canal system (see Fol’s fig. 17, plate xxv, ‘Jen. 
Zeits.,, vol. vii). The probability of the correctness of 
Haeckel’s observations and inference as to the first ap- 
pearance of the sub-umbrellar cavity as a closed sac, is not 
only supported by my observations on Limnocodium, in 
which this certainly is its condition at one period of its 
development, but when we examine carefully the accounts 
given by both Metschnikoff and Fol, we find that it is 
precisely at the critical period which would enable them to 
deal decisively with Haeckel’s observations that their series 
of embryonic Geryoniz in both cases is deficient. Neither 
Metschnikoff nor Fol have seen stages corresponding to 
Haeckel’s figs. 29 and 30, with only four tentacles. They 
both give series in which there is a sudden break at what 
is the critical period for this matter; they pass at once from 
the condition in which no tentacles are present to that in 
which six are already manifest. (Fol’s figure 18 of an 
embryo 82 hours old, with no tentacles, is succeeded by 
figure 19 of an embryo, 156 hours old, with six tentacles ; 
and Metschnikoff’s figure 10, which represents an embryo 
about 50 hours old, having no tentacles, is followed by 
figure 11, representing, as he states, an embryo 190 hours 
old, and possessing six tentacles.) 
It seems to me obvious from these facts that one of the 
most important stages in the development of Geryonia is 
still almost entirely unknown, and, accordingly, the stage 
