398 WILSON. [Vot. III. 
sine (No. 50), and Vejdovsky on Rhynchelmis (Nos. 45, 48), 
shows that it is impossible to establish any certain homology 
between these forms and Lumbricus. At first sight the four- 
celled stage shown in Fig. 3 might seem to correspond with that 
of Rhynchelmis, but the subsequent development fails to estab- 
lish the comparison. The eight-celled stage (consisting of four 
macromeres and four micromeres), which may be taken as char- 
acteristic of the typical unequal cleavage, seems to have been 
entirely masked by secondary changes in the cleavage periods, 
and the facts are still too incompletely known to justify any 
speculation as to the history of these changes. 
The one positive result that seems to be brought out by the 
comparison is that the phenomena of cleavage in Luwméricus are 
of a highly modified character, and this conclusion is also indi- 
cated by the individual variability of the segmenting ova. This 
conclusion, unsatisfactory as it is, may have an important bear- 
ing on the general interpretation of the process of gastrulation. 
2. Origin of the Layers.— My observations on the origin of 
the primary mesoblasts agree closely with those of Kleinenberg, 
and show beyond all question that the mesoblast is differentiated 
long before the invagination of the gastrula. 
The changes resulting in the formation of the blastula usually 
occupy from 18 to 24 hours. During the following 24 hours 
(more or less) the cells increase rapidly in number by radial 
divisions, the blastula remaining one-layered and nearly spher- 
ical. Two of the larger cells lying side by side near the equator 
of the blastula (Fig. 23) undergo no radial division, and their 
inner ends soon project into the blastoccel. The inner ends 
are then separated from the peripheral parts by tangential 
divisions, which take place nearly at the same time in the two 
cells. The two smaller cells thus formed lie side by side in the 
blastoccel, which they nearly or quite fill (Fig. 25). The two 
larger cells, which still lie at the surface of the blastula, are the 
“primary mesoblasts,” which by continued divisions give rise 
to the entire mesoblast. I have failed to trace the origin of 
these cells in the process of cleavage, the original character of 
which has been so altered that it is impossible to determine 
the relation of the primary mesoblasts to the micromeres and 
macromeres of the typical unequal cleavage. Fortunately, how- 
ever, Vejdovsky has succeeded in following their mode of origin 
