274 Proceedings of Royal Society of Edinburgh. [sess. 
animal scale. Examples taken from the Adinozoa, Echinoder- 
mata, Polyzoa, Lamellibranchiata , and T unicat a, of species which 
have again acquired a locomotor habit tend to show that the 
loss of organs undergone during the sedentary stages is irremedi- 
able and limits the further advance in differentiation. In other 
words, as has been recently pointed out,* we have no proof that 
organs once lost have ever been re-evolved. 
It therefore follows that although there are numerous instances 
of sedentary types again assuming a pelagic or littoral existence, 
yet the ancestral history of such forms as the Arthropoda and 
Chordata probably has been passed in the pelagic or littoral 
regions. In the case of this earliest coelomate ancestor, its radial 
symmetry leaves us no choice between these two ; a pelagic 
environment is the only possible alternative. 
We are therefore led to regard the remote ancestor of the 
Coelomata as having been a radially symmetrical tetramerous 
coelenterate with a ventral axial mouth (with functions of mouth 
and anus) and four gut-pouches separating off from the alimentary 
canal to form four ccelomic pouches, f All four of these 1 coelomic 5 
pouches will form reproductive and muscular elements, and will 
contain a nutritive circulatory fluid. Belonging to the pelagic 
plankton, it must have had an existence somewhat similar to that 
of many pelagic medusae of the present day.J (Stage I.) 
The next differentiation is that of the mouth (blastopore) into 
mouth and anus. The first step in this is the approximation of the 
lips of the stomodaeum in two opposite radii in correlation to the 
hypertrophy of the ingestive and egestive functions, respectively, 
* E. W. Macbride, Natural Science, January 1897. 
t Cf. the suggestive remarks of Korschelt and Heider (pp. 344-345, 
English translation). 
I By this it is not implied that this organism was morphologically compar- 
able to the medusae of the present day. To this view (Kleinenberg, Balfour) 
it has been objected by Korschelt and Heider (pp. 342-343, English transla- 
tion) that the medusa presents a higher type of locomotion, and we may add 
of ingestion. Another indicated difficulty is the absence, in medusae, of an 
apical nervous system. It has been assumed below that the apical ganglion 
was present in the pelagic ccelomate ancestor, and that in such a form as the 
trochosphore it has become secondarily shifted, to the new apical pole at the 
apex of the pre-oral lobe. For figure of this Stage I,, see Quart. Journ. 
Micros. Science, vol. xxxviii. p. 325. 
