No. 3.] ZHE EMBRYOLOGY OF THE EARTHWORM. 441 
pleted. In Lumbricus the germ-bands are present in the gas- 
trula, but are relatively much less developed than in Clepsine, 
and they do not unite anteriorly until after the establishment of 
the mouth. In Lupomatus, finally, the mesoblastic bands are 
wholly rudimentary at the time the blastopore closes, but they 
are nevertheless represented by the primary mesoblasts. 
I am not prepared to discuss the case of Lofadorhynchus and 
allied forms in which no teloblasts have been observed, for it is 
very far from certain that they are not really present (see p. 443), 
and their very general occurrence in the Trochosphere indicates 
at present that they are to be regarded as characteristic of the 
larva. : 
To sum up: as:far as our knowledge goes, the development of 
Lumbricus can be most simply and clearly interpreted in accord- 
ance with Sedgwick’s hypothesis, as follows: (1) The ancestral 
form possessed an elongated ventral blastopore that gave rise to 
both mouth and anus by closure in the middle region ; (2) the 
mesoblast and the nervous system originally formed a ring 
around this blastopore, subsequently undergoing concrescence 
throughout its middle portion as the blastopore closed; (3) the 
coelomic cavities were arranged in a continuous series in the 
mesoblastic ring, each lateral cavity lying opposite a correspond- 
ing cavity on the other side of the body, and a single anterior 
cavity lying in front of the mouth and giving rise to the head- 
cavity ; (4) the larval Trochosphere is secondarily derived from 
such a form by retardation or temporary suppression of the 
trunk-region and early and extensive differentiation of the head- 
region.! 
Further than this I shall venture no conjecture as to the 
character of the adult ancestral form, except to state that the 
views suggested are reconcilable with the derivation of annelids 
either directly from Ccelenterata, or from Platyhelminths, in ac- 
cordance with the views of Balfour and Sedgwick, or Lang. 
The essential feature of all these views is the identification of 
the principal or longitudinal axis of the body with one of the 
1 The apparent non-extension of the blastopore into the trunk-region in the larval 
Trochosphere is under this view owing simply to the rudimentary condition of the 
trunk at the time the blastopore is formed. The early closure of the blastopore is 
probably due to the advantage of the earliest possible establishment of the mouth in 
the free-swimming larva, which must procure its own food. 
