VOL. i.] Esterly. Poixnn (Unmix t>f I'lrfhtiilon. 239 



is seen that the mass of dilators is often concave in outline 

 toward the center of the duct (PL XXIII, Fig. 28: I'l. XXI, Fig. 

 !.">, ilil.m.), so that in contraction the fibres first mentioned 

 pull in the general direction of the major axis and toward the 

 center of the ellipse. Other dilator fibres attach at the edges of 

 the duct near the end (PI. XXI, Fig. 14), and in shortening pull 

 in the direction of the minor axis of the ellipse, thus widening 

 the lumen by spreading its walls at the tips of the oval. (PI. XXI, 

 Fig. 14. ) The entire effect of the dilator fibres is to make the 

 aperture of the duct nearly circular, thus offering freer exit to 

 the secretion. Their action would be to open the duct from the 

 form shown in PI. XXI, Fig. 16, to that, for example, in PI. 

 XXIII, Fig. 30. 



The fact that the constrictor and dilator fibres lie entirely 

 within the epidermis need not militate against their having the 

 function of muscles, for in the case of the intrinsic gland muscu- 

 lature it has been well established that it has an ectodermal 

 origin. It is certain that the arrangement and appearance of 

 the fibres described as constrictor and dilator muscles are such as 

 to suggest very strongly both that nature and function. The 

 coloration in Mallory is exactly that of the smooth muscles of 

 the glands ; and the convergence of the constrictor fibres to their 

 insertion in a position where contraction would close the duct; 

 the endings of the dilators in places to be of greatest advantage 

 in widening it when the muscles contract all these facts lead 

 one to conclude that he has to deal with an apparatus for closing 

 and opening the ducts of the glands. 



The muscles of the poison glands, as has been said, immedi- 

 ately envelop the secretory cells. The entire gland is filled with 

 enormous cells, the generally recognized "Giftzellen" of many 

 authors or the "giant cells" of Leydig. In such glands a lumen 

 does not exist; this is especially well shown in sections of the 

 tail of a tadpole 38 mm. in length, in which the cell boundaries 

 are distinct, the secretion not yet being present in sufficient 

 quantities to obliterate them. There it will be seen that the ends 

 of the cells are in contact with the middle of the gland, thus 

 doing away with any trace of a lumen (Nicoglu '93; Seeck '91; 

 Calmels '83). A glance at the figures will serve to distinguish 



