710 PROCEEDINGS OF THE ACADEMY OF [Oct., 



erroneous, and believed that the cereliral jjan^lia arise from the cross. 

 I.ater investigators have been unaV)le to either pro\'e or disprove tliis 

 statement. It is at least certain that the brain is formed from the 

 first cjuartette of micromeres, and its first rudiment is found in the 

 neighborhood of the rosette. These facts, however, can be gathered 

 from the works of the older investigatf)rs (Hatschek, Salensky, ct ol). 

 In Dinophilus the rudiment of the cerebral ganglia involves more 

 than the cells of the cross alone. Its development, however, is typi- 

 cally annelidan, since the ectodermal thickening which represents the 

 earliest rudiment of the brain appears beneath the rosette, which in 

 the trochophore bears the apical tuft of cilia. This 7'udiment then is 

 to be regarded as the ontogenetic representative of the '' Scheitelplatte" of 

 the annelid trochophore. 



(2) The PrototroeJi. 



The head of both the larval and adult individuals of Dinophilus 

 gyrociliatus (apatris) is encircled by two narrow transverse l^ands of 

 long cilia, similar to those which are found on the metameres of the 

 trunk, as illustrated in text fig. ^T. Both of these bands are preoral; 

 the first being situated just anterior to the eyes, the second surrounds 

 the head near its juncture with the trvmk, and passes ventrally just 

 anterior to the mouth. This second preoral band corresponds in posi- 

 tion and function to the prototroch of the trochophore. It first appears 

 at a stage nearly corresponding to that figured in fig. 59 as two delicate 

 tufts of cilia on each side of the head. 



The history of the cells composing this second preoral ciliated band, 

 which I have already taken the liberty of calling the prototroch, will 

 be considered under the head of the first quartette. I have done this 

 chiefly because of the important ]ilace given in the literature of cell- 

 lineage to those components of the annelid prototroch and the mollus- 

 can velum derived from the first quartette, the "primary trocho- 

 blasts" (Mead, 1897). 



At the 29-cell stage the primary tr()chol)lasts are already beginning 

 to divide (figs. 14 and 15). The division is equal in all. The posterior 

 pair divide meridionally, while in the anterior pair the spindle is in- 

 clined from the horizontal plane in a dexiotrojiic direction. While 

 the posterior pair begin to divide as early as the stage shown in fig. 14 

 (26 cells), lb" does not complete its division until a stage numbering 

 54 or more cells is reached. Meanwhile the cells of the second quar- 

 tette in the A and C quadrants have each given off a small cell above, by 

 a nearly equatorial cleavage (figs. 14, 15, 16, IS and 19). 2b later also 

 divides, but this cleavage, instead of4)einge(|uatorial. is strongly dexio- 



