OE THE MULLEBJAN DUCT OF AMPHIBIANS. 521 



Another Axolotl (S), 45 mm. in length, has likewise a pronephros that is much 

 reduced. The glomerulus is enclosed in a very small dorsal diverticulum of the coelom, 

 and the pronephros is encompassed by a thick layer of secondary tissue, so that it is 

 considerably below the epithelium. A few sections posterior to the beginning of the 

 glomerulus a very slight middle diverticulum begins ; and on the lateral wall of this 

 a thickened plate of epithelium represents the anterior of the Miillerian duct. After 

 seven sections the Anlage comes to be dorsal ; and in the 1 Oth it begins to rise up as 

 a ridge, which seems to rest on the outgrowth of the radix mesenterii.* Soon after the 

 twentieth section the glomerulus disappears, and in the twenty-seventh the outgrowth 

 of the radix mesenterii ends. The ridge is furrowed with a groove that becomes deeper 

 as it gets further back, till about the fortieth section the edges of the depression close 

 and form a tube,t and this is continued backwards by a solid Anlage which follows the 

 segmental duct. The segmental duct is at the base of the ridge from the termination 

 of the outgrowth of the radix mesenterii ; and posterior to the closing in of the groove 

 it lies within the ridge, and in close proximity to the Miillerian duct. 



An Axolotl (e) 51 mm. in length shows yet a further stage in the development of 

 the anterior end. It illustrates, namely, how the lateral Anlage is undermined by a 

 diverticulum of the ccelom, and by unequal growth of the walls of the body-cavity is 

 first drawn into a more antero-posterior line and then ruptured along the external 

 margin, so that the band of thickened epithelium projects free into the coelom and is 

 attached only by its inner edge ; the band becomes grooved, and closes in to form the 

 Miillerian duct in due course. I have been able to follow this process better in 

 Salamandra atra than in Axolotl, but the facts that follow bear out the view that 

 the development is similar in both cases. 



The Miillerian duct begins as usual in the middle diverticulum of the body- 

 cavity ; but in the fourth section the lung becomes free and this diverticulum terminates. 

 At the very beginning of the Anlage the lateral plate is undermined by another diverti- 

 culum of the coelom (fig. 17). After a few sections, however, the epithelial plate loses 

 its lateral connection, and retaining its inner basis, passes back as an oblique process 

 (fig. 18). For eleven sections after the break in the plate appears there is a lateral 

 remnant to show the former connection of the Anlage (r, fig. 18). In the twenty-sixth 

 section, in this case, the groove closes : it begins almost at the anterior, but does not 

 become deep, as in " 8," during its short course. The transition from groove to rod 

 here is so gradual that no tube is visible. It is impossible to say with certainty if a 

 considerable portion of the anterior Anlage has gone through the stages of grooving and 

 closing over, because the segmental duct has disappeared and all other local marks ; but 

 from a comparison with " 8," it seems probable that this is so. 



The development of the anterior end in Salamandra atra has already been 

 described up to the stage where, after the suppression of the inner parts of the lateral 



* See figs. 5 and 6 in my preliminary account already cited. 

 t See figs. 7-10 in the above-cited work. 



