572 E. W. MACBItlDE. 



proportion. The characteristic longitudinal ciliated band by- 

 means ol' which locomotion is effected has already appeared 

 as a transverse ridge, which rapidly becomes a horizontal 

 loop (PI. 31, figs. 4 and 5) ; as the postero-lateral arms grow, 

 they receive extensions of this loop, and so the swimming 

 powers of the larva are increased. It may be partly owing 

 to this circumstance that the vacuolated crest disappears, 

 but it is also partly replaced by a conversion of the ecto- 

 derm of the posterior end of the larva into an appendage of 

 vacuolated cells, which persists until metamorphosis is com- 

 plete (figs. 29 and 44, vac.) . 



Leaving the consideration of eggs which have received 

 natural fertilisation, and turning to those which have been 

 artificially fertilised, we find that they are distinguished by- 

 such an early and abundant formation of mesenchyme that 

 sections through earlier stages (fig. 18) show a practically 

 solid mass of blastomeres, and when segmentation is complete 

 there is an outer covering of cylindrical ectodermal cells, and 

 an inner mass of rounded cells, so that an appearance is 

 presented which strongly recalls that of a Ccelenterate 

 planula (PI. 32, fig. 19). A flattening and thickening of 

 one pole ntiw follows, together with a renewed production of 

 mesenchyme (fig. 20, mes.), and then invagination takes 

 place in such a way as to leave a wedge of cells projecting 

 into the cavity of thearchenteron (fig. 21). This results from 

 the fact that the endodermic plate of cells becomes in one 

 meridian several cells deep before it is invaginated. In the 

 next stage the postero-lateral arms make their appearance 

 (fig. 22) and receive most of the primary mesenchyme, whilst 

 soon after the coelom appears as a vesicle at the apex of the 

 archenteron. Part of the projecting tongue of cells is 

 included within it and part within the lower portion of the 

 archenteron, both parts being eventually absorbed. At no 

 time is there the slightest trace of the vacuolated crest of cells 

 so characteristic of normal development. Caswell Grave, in 

 his paper on the development of Ophiura brevis (12), 

 describes a tongue of cells projecting into the archenteron as 



