346 H. H. NEWMAN 



in other mammals, and remains single while it descends the 

 oviduct and even after it reaches the uterus. When the egg 

 reaches the uterus it has undergone 'embryonic germ-layer 

 inversion/ so that the ectodermic vesicle is at least partially 

 surrounded by endoderm. Patterson ('16) has made the very 

 significant observation that the embryo stops developing at an 

 early gastrula stage and remains quiescent for several weeks. 

 Toward the close of October, after having been at a develop- 

 mental standstill for a long time, floating free in the uterus 

 without any nutritive or respiratory connection, a significant 

 event takes place: the embryo accomplishes placentation. De- 

 velopment is resumed when the egg undergoes the process of 

 primary placentation in October and new growth vigor is de- 

 rived from the nutritive supply thus obtained. The period of 

 retardation has, however, been so prolonged that the embry- 

 onic axis has been largely obliterated and, when rejuvena- 

 tion or recovery takes place, one location in the ectodermic 

 vesicle is as likely to become the apical point of the new axis 

 as another. As a matter of fact, two points, favorably situ- 

 ated with reference to the conformation of the uterus, assume 

 the initiative and become two new heads ends or apical points. 

 A little later two other points, not so favorable situated, become 

 similarly physiologically isolated and form two secondary head 

 ends. The result is the formation of one primary and one sec- 

 ondary individual out of each lateral half of the ectodermic 

 vesicle. The four apical points grow as though each were a 

 whole independent embryo and each forms its own amnion and 

 its entirely separate placenta, though they remain enwrapped 

 in a common birth robe or chorion. The phenomenon is one 

 involving physiological isolation, through retardation, of parts 

 of the ectodermic vesicle, that part of the embryo that takes the 

 initiative in development. The succeeding processes of complete 

 separation of the four quadruplets are the normal sequelae of 

 the above-described initial steps and need not concern us here. 

 Do the results described above for the starfish Patiria tend to 

 strengthen this theory as to the causes of twinning? I believe 

 they go far in that direction. Each of the three methods of 



