108 John H. Geroüld, 



sections of the elong'ated trochophore of Ph. gouldii (bl hours) 

 (Fig. 87), but, since this transitory phenomenon is seen at a slightly 

 later period (51—57 hours), I shall discuss it in the next section. 



The two pairs of permanent retractor muscles are formed in 

 the very young trochophore, chieflj^ from groups of ectoderm cells 

 upon each side of the apical plate (Fig. 44, 48). They form a part 

 of the zone of cells which bear the preoral circlet of cilia. These 

 cells of the apical plate become greatly elongated, sink beneath 

 the surface, extend directly backward beneath the prototroch but 

 outside of the still solid bands of coelomesoblast, terminating at four 

 equidistant points in the region of the postoral circlet, two being 

 dorsal and two ventral. The terminal growth of the trunk carries 

 these points further and further back, till they occupy the position 

 shown in Fig. 88. Each muscle consists at first of only three or 

 four elongated cells which, as thej' sink beneath the surface, become 

 drawn out into spindle-shaped fibres, at the middle of which the 

 nucleus lies in the midst of a little granular protoplasm. 



Thus the ectomesoblast cells which give rise to the retractors 

 are derived from the first quartet of ectomeres, as is the case with 

 certain of the myoblasts in Thalassema (Toreey, 1902). although they 

 are differentiated in Phascolosoma at a later period. It is probable 

 that other ectomesoblast cells in Phascolosoma also enter into the 

 formation of these muscles, notably certain ones which lie in the 

 zone immediately behind the prototroch (Fig. 44, 48). 



Completely differentiated muscle cells are seen at the earliest 

 in longitudinal sections of trochophores of about thirty-six hours, 

 when the hitherto spherical embryo begins to become elongated. 

 The myoblasts are at that time spindle-shaped, their nuclei lying 

 half way between the apical and somatic plates, and immediately 

 outside of the coelomesoblast. Their position at that time, near the 

 coelomesoblast, invites the conclusion that thej' may have been derived 

 from cells of the somatic portion of the latter, and I consider that 

 the possibility of the derivation of the middle portion of the retractors 

 from the coelomesoblast is not entirely excluded, though it does not 

 seem probable. 



Hatschek (1883) assumes that the retractor muscles in Siptmculus 

 arise from the somatic layer of mesoderm. When they make their 

 appearance in this animal the coelom has already been formed, and 

 within it the retractor muscles were first observed. Hatschek states, 

 however, that he did not see them during the earliest stages of 



