262 INEZ WHIPPLE WILDER. 



doors to the laboratory, the mother goes back to them again, 

 even though the nest and all of its surroundings may have been 

 reconstructed. I have never had the opportunity to further test 

 the sense of ownership of eggs in a mother by exchanging the eggs 

 of two individuals, but the experiment would certainly be an 

 interesting one. 



The function of the mother in incubating the eggs is probably 

 mainly that of insuring the proper degree of moisture, since eggs 

 when removed from the mother and kept moist undergo normal 

 development. It is possible, also, that she may guard them 

 from inundation; for the first eight or ten days eggs will develop 

 normally, even though quite immersed in water, but during the 

 later development a short immersion in water kills them, ap- 

 parently through a rapid change in osmotic pressure between the 

 cells of the embryo and the fluid which by this time surrounds the 

 embryo within the protective envelopes. 



I have no evidence that the habit of eating the eggs, such as 

 Smith ('07) found in the case of Cryptobranchus , is ever indulged 

 in by Desmognathus, although, as will be shown below, this species 

 is by no means to be exempted from the charge of cannibalism. 



THE EGGS. 



The egg membranes, already very carefully described by 

 previous writers, are three in number, the outer one being 

 continued into a stalk which connects the eggs with the others, 

 and is thus a common membrane for the whole mass. Smith, 

 ('12) in comparing the egg strings of Cryptobranchus allegheni- 

 ensis with the arrangement of the eggs of other Urodeles, suggests 

 the probability that the most closely related egg arrangement is 

 such as is shown by the stalked egg enclosures of Desmognathus 

 rather than in the jelly-enclosed masses of certain other forms. 

 The accompanying drawings (Fig. 4) of the two egg masses of 

 a batch of eggs studied by me suggest the probability of the 

 derivation of their arrangement from a more primitive rosary 

 arrangement, such as Cryptobranchus possesses. A few eggs are 

 still seen in the main axis of the bunch, but the majority have 

 pushed out to one side, as if through an excess of lateral pressure 

 brought to bear upon the outer envelope from within. Each 



