No. 3.] THE EMBRYOLOGY OF THE EARTHWORM. 437 
Lumbricus, and in Polygordius, without perceiving that it is 
essentially the same in all, though in Polygordius the ectoblast 
extends over the ventral surface at an even earlier period than 
in Lumbricus. In all, the mesoblastic bands lie at an early 
period on opposite sides of the body throughout the trunk- 
region, and their subsequent growth and union along the median 
ventral line in Polygordius and the other larval types is closely 
similar to the corresponding process in Lwmbricus, and scarcely 
less striking than in the epibolic types. There is, however, no 
food-yolk (Eupomatus, Hydroides), the gastrulation is of the 
embolic type, the blastopore never actually extends between the 
mesoblastic bands (since they are still undeveloped at the time 
of its closure), and their wide separation must be due to some 
other cause. The essential agreement in the history of the 
mesoblastic bands between forms so different both in structure 
and in the conditions of embryonic development as Clepsine, 
Lumobricus, and Polygordius, is very strong evidence that meso- 
blastic concrescence has some ancestral meaning, and was not 
originally caused, though afterwards it was undoubtedly in many 
cases modified and rendered more conspicuous, by accumulation 
of food-yolk in the ovum.? 
All of the evidence seems, therefore, to indicate that the 
mesoblast (with the contained coelomic cavities) originally lay 
in two masses that extend along the sides of the alimentary 
canal and joined in front of the mouth and also at the posterior 
extremity. Concrescence is the ontogenetic repetition of the 
-process by which these originally separate masses extended 
dorso-ventrally around the archenteron and fused in the middle 
line, and the explanation of this process must, I believe, be 
sought in the relation of the germ-bands to the blastopore. 
The posterior limit of the ventral surface of the embryo ZLam- 
bricus may be placed at the point of union of the two primary 
mesoblasts, and the anterior limit at the mouth. The blastopore, 
therefore, occupies at first nearly the whole ventral surface, its 
anterior lip corresponding with the anterior lip of the mouth, its 
posterior lip lying just anterior to the primary mosoblasts. In 
1] am aware that I have used the word “concrescence”’ in a somewhat broader 
sense than that employed by Professor Whitman. The word, however, cannot logi- 
cally be restricted to such cases as those of C/efsime, and to make such an arbitrary 
limitation would be simply to ignore the real problem to be solved. 
