16 WILSON. 



and transparent, are easily procurable in large numbers, and de- 

 velop so slowly that the successive stages may be very accu- 

 rately followed in life, while every point may be repeatedly 

 verified in a large number of specimens. The results of a study 

 of these eggs not only help still further to set aside the ap- 

 parent contradiction between the polyclade and the annelid- 

 mollusk type, but, w^hen taken in connection with the foregoing 

 observations on annelids and gasteropods, also raise some highly 

 interesting questions regarding the relation of cell-lineage to an- 

 cestral reminiscence. 



I shall not here describe the cleavage of Lcptoplana in detail,, 

 but will only indicate its leading features. Up to the thirty- 

 two-cell stage, and for some distance beyond, the cleavage is a 

 most beautiful example of the symmetrical spiral type, agree- 

 ing very exactly wuth Discoccclis as described by Lang, except- 

 ing in the fact that in the four-cell stage the cross -furrow is 

 inconstant and often wanting. The first three quartets of mi- 

 cromeres are formed exactly as in an annelid, and have the 

 same position and relative size as in Discoccelis (Fig. 5, A), w^hile 

 the four large cells remaining give rise to the archenteron. 

 Regarding the morphological value of these three quartets,, 

 however, my results differ very considerably from Lang's and 

 are such as to bring the polyclade cell-lineage into direct rela- 

 tion with that of the annelid, gasteropod and lamellibranch. 

 As in these groups all three of tJie quartets give rise to ectoblast^ 

 the first and third apparently to ectoblast alone, though I am 

 not certain that the third quartet may not give rise also to a 

 small modicum of mesoblast-cells. The principal interest 

 centers in the second quartet, from w^hich, as Hallez, Gotte and 

 Lang have shown, the principal mass of the mesoblast is formed. 

 What these observers have failed to observe is the fact that each 

 cell of this quartet gives rise to several ectoblast-cells — at least 

 three, and probably four — before sinking into the interior to 

 form mesoblast. These divisions are of constant form, as fol- 

 lows : During the fifth cleavage each cell divides unequally 

 towards the left as viewed from the side (/. r., clockwise, as seen 

 from above) to form an ectoblast-cell (** 2^ ") that abuts against a 



