58 R- R- HUMPHREY. 



rounded by sheaths of interstitial cells. Testes of the type shown 

 in Figs. 15 and 16 are frequently encountered in June or July. In 

 another year the testis of Fig. 15 would reach the stage shown in 

 Fig. 17. Then, by the disappearance of the degenerating lobules 

 and interstitial cells during the summer and early fall, all trace of 

 the lobe finally vanishes, as illustrated in Figs. 18 and 19, and only 

 a slender germ-cell cord remains. 



The number of lobes that may possibly be formed, then, is 

 limited by the extent of the primary germ-cell cord. This may 

 and doubtless does increase by growth in length of its anterior 

 portion as the animal develops. In any event, it varies consider- 

 ably in different males, since the anterior lobe may appear as in 

 Fig. 1 6 or 17 even when only two lobes are present, or it may show 

 no indication of running out in a testis of from three to four lobes. 

 An unusually long germ-cell cord permits the development of 

 numerous lobes; a short cord reduces the number, though in no 

 case has it been apparent that a very large old Dcsmognathus male 

 possessed but a single lobe. Animals have been found, however, 

 with multiple testes (as in Fig. 20) consisting of one large func- 

 tional lobe, a very small lobe developing caudally, and a trace of 

 an anterior lobe. Similar to these, but one step nearer to a simple 

 testis, is that represented in Fig. 21. In this case no caudal lobe 

 has yet begun its growth, and the anterior lobe is reduced almost 

 to total disappearance. This type of testis in Desmognathus is 

 exceptional, only one having been encountered in an examination 

 of over a hundred males. 



From the preceding description it is readily seen that a combi- 

 nation of two factors is essential in the production of the multiple 

 testis of Desmognathus. These two factors, let me repeat, are the 

 slow movement of the spermatogenetic wave and the delayed re- 

 generation of the emptied lobules. Neither of them alone could 

 give rise to a multiple testis. If, for example, the slow spermato- 

 genetic wave is combined with rapid regeneration of the lobules, 

 the length of the testis is merely increased each year by additions 

 from the anterior germ-cell cord, but remains a unit, as in the 

 Plcthodons. If, on the other hand, a rapid spermatogenetic wave 

 passed the complete length of the testis in a single season, as in 

 Cryptobranchus, but lobule regeneration were then delayed for the 



