112 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 



sented bj (1) two dorsal tubules (the secretory portion), whose fused 

 ends are continued backward as the pronephric duct, and (2), ventral 

 to these, two tubules (the sexual portion) arising from Miillerian 

 evaginations, whose caudal ends join the pronephric duct some distance 

 behind its coiled, anterior end. It would seem that in the ancestors of 

 the Urodela and Anura there was originally one pronephric tubule for 

 each of several segments and that the number was increased by the ad- 

 dition of a tubule in each of the segments already containing one (Figs. 

 i?, S, and T). These additional tubules lost their secretory function and 

 took on a sexual one. The exact position of the sexual tubules in the 

 somite is not significant. It seems probable that they arose ventral to 

 the secreting set, for that is their position in Amblystoma and Rana 

 sylvatica. In Hyla versicolor, it is true, they arise dorsally, but this 

 position may have come about secondarily, as is suggested by the foct 

 that in one case a Miillerian evagination was seen which was ventral to 

 the nephrostome. 



Such a duplication of tubules as postulated above seems to take place 

 in Amphiuma, where, according to Field ('94), three somites on each 

 side possess two tubules each, one dorsal and one ventral, making six in 

 all. Whether one of these sets, the dorsal or the ventral, gives rise to 

 the Miillerian duct or not, we do not know, as the development of the 

 duct has not been studied. Such a result would not, however, be essen- 

 tial to a confirmation of the theory here set forth, as it is quite possible 

 that in groups other than the Urodela and Anura the Miillerian evagi- 

 nations have been derived from a third set of pronephric tubules, or from 

 tubules of the posterior or anterior end of the single original series. In 

 the latter case there would be no necessary correspondence in number 

 between Miillerian evaginations and secretory tubules. 



In Amblystoma (and probably all other forms above the elasmo- 

 branchs) the appearance of the sexual tubules is delayed, so that they 

 arise much later than the secretory set. Their distal ends fuse with 

 each other (to form the anterior end of the Miillerian duct) before they 

 reach the Wolffian duct in all the cases that I have examined ; but as 

 the point of fusion is quite variable, they may sometimes reach the duct 

 separately.^ The original formation of the Miillerian duct in the re- 

 mote ancestors of the Amphibia by a splitting of the Wolffian duct is 

 indicated by the fact that the latter contributes cells to the growing 

 Miillerian duct in the region where the two come in contact. Behind 

 this point the ^liillerian duct grows back independently. This inde- 



1 Kip ('94, '94») notes such cases in mammals (see page 106). 



