298 
Proceedings of Poycd Society of Edinburgh. [sess. 
The Larval Forms. 
One outstanding feature about the Ar chi-coelomata is the con- 
stant occurrence of free pelagic larval forms, and the comparison of 
the demersal larva of Balanoglossus Kowalevski with a Tornaria , 
and a Bipinnaria or Brachiolaria with the demersal type of larva, 
such as that of Asterina gibbosa, enables one to conjecture to a 
large extent what are the secondary or coenogenetic features of 
these pelagic larvse. We have assumed that the archi-coelomate 
ancestor was pelagic in its habits, and with this fact in view the 
pelagic habits of so great a number of larvse in the Ar chi-coelomata 
acquire a phyletic significance. In other words, the vast majority of 
the Ar chi-coelomata pass through a pelagic stage in their ontogeny, 
because their common ancestor was pelagic. Amongst these larvse 
there are endless instances of the archimeric segmentation into 
three parts, protomere, mesomere, and metamere, whether exem- 
plified by actual division of the body into three parts by con- 
strictions, as in Actinotrocha (fig. 7), larval Balanoglossus (fig. 8), 
the larval of the Brachiopoda (figs. 9, 10, 11, 12), or whether 
indicated by the structure of the three ciliated bands, prototroch, 
mesotroch, and metatroch, as in Tornaria, Echinoderm larvse, Actino- 
trocha, and Polyzoan larvse. 
This constant occurrence of archimeric segmentation, in however 
modified a manner it appears, must indicate a common descent 
from a form with this archimeric segmentation fully developed. 
( C .) Evidence may be derived from morphology and from ontogeny 
of the metamerically segmented groups that a secondary segmenta- 
tion ( metameric ) has been phyletically superposed upon the primary 
( archimeric ). Thus, these animals may be shoivn to pass through 
ontogenetic stages ivhich closely resemble the archi-coelomate type, and 
to possess in their morphology more or less vestigial traces of the 
archimeric segments. 
More than two years ago I remarked* : — “I believe that these 
resemblances will all eventually find expression in the constitution 
of one primitive group .... having the mesoderm divided up 
into one pre-oral and two pos-toral coelomic pouches, all, primitively, 
* Froc. Royal Soc. Edin ., 1895-96, p. 63. 
