TWINNING IN OTHER SPECIES OF ARMADILLO 71 



canals are forked just as are those of a pair of embryos 

 of D. novemcinctus that come from a primary outgrowth, 

 a fact that lends probability to the view that the develop- 

 ment of polyembryony in the two si)ecies of Dasypus 

 is practically identical in character. It should also 

 be said that, even in sets with eight or more fetuses, 

 there is no exception to the rule that all from a single 

 egg are of the same sex. Unfortunately nothing is 

 known about the heredity of armor characters, nor 

 about the symmetry relations existing between the 

 different members of a polyembryonic set. Such a 

 study of the species, if correlated with what has been 

 published in these connections about D. novemcinctus, 

 would be important. 



We are indebted to Fernandez for a clear under- 

 standing of the interrelations of germ-layers and of the 

 peculiarities of amnia and allantois in D. hyhridus. 

 The diagram (Fig. 26) is adapted from Fernandez' 

 first paper. It would serve, however, almost equally 

 well for D. novemcinctus. Only two embryos that lie 

 to right and left of the egg are shown. Note the external 

 endoderm {en), the rudimentary allantois (j/), and the 

 short yolk stalk opening into the yolk sac, which is in 

 this case inverted so as to form the external layer of 

 the vesicle. The belly-stalk {bs) or primitive umbilicus 

 is estabhshing a relation with the Trager, but as yet 

 no umbilical vessels have invaded it. A posterior 

 prolongation of the amnion {p am) goes back toward 

 the Trager, but appears to have no part in the formation 

 of an umbiHcus. The amniotic connecting canals and 

 common amnion are showm wath both ectodermal and 

 mesodermal layers present. 



