Experimental Studies on Germinal Localization. 21 



to the changed condition of surface-tension in the dwarf embryos 

 and larvae, as is shown by the readiness with which they dis- 

 integrate. (I have several times seen an actively SAvimming dwarf 

 larva suddenly fly to pieces on coming in contact with an obstacle 

 or even with the surface of the water.) For these reasons, de- 

 spite the great ease with which the eggs may be operated, it is 

 difficult to base trustworthy conclusions regarding the more spe- 

 cial features of the egg-localization on the defects observ^ed in 

 the individual partial larvae. I have therefore in the following 

 work restricted my account in the main to the results that appear 

 with unmistakable clearness, and appear in so large a proportion 

 of the larvae as to remove all reasonable doubt. Beyond this, 

 owing to the importance of following the development of the 

 living larvae as far as possible, the number preserved for section- 

 ing was not very large, and the technical difficulties indicated 

 above, in case of the normal larvae, here appear in aggravated 

 form. This explanation is necessary to account for certain ob- 

 vious gaps in the work, which I hope to fill out by further in- 

 vestigation, especially those relating to the mesoblast, regarding 

 which I can at present offer only somewhat provisional conclu- 

 sions. 



III. 



EFFECT OF REMOVING THE POLAR LOBE. 



(a) General History of the lobeless Larvae. — During the 

 trefoil stage of the first cleavage the polar lobe may easily be 

 removed, wholly in part, by means of a fine scalpel. Complete 

 removal of the lobe produces a highly characteristic and constant, 

 though in one respect very unexpected, result. Exactly as Cramp- 

 ton earlier found in Ilyanassa, the egg continues to segment after 

 this operation quite symmetrically, in a manner similar to the 

 normal cleavage of such forms as Patella or Lymnaea, giving rise 

 by typically alternating spiral cleavages to successive symmetrical 

 quartets of micromeres (Figs. 20-26). These cleavages differ 

 constantly in two respects from the normal, namely, that ( i ) 

 no trace of a polar lobe is formed at either the second or the 

 third cleavage, and (2) the members of the D-quadrant are no 



