36 /. H. Powers 



larvae with truncated muzzles this mode of respiration is probably 

 increased at the same time that branchiate respiration is decreased. 

 This latter is inevitably brought about by low nutrition which, 



considerable movement accelerates it. The regularity, however, surprised 

 me, not only m the same specimen from time to time, but in diverse animals 

 under somewhat different conditions. I tested a number of wild adults 

 just taken in water, during the breeding season; also a very large wild 

 adult which had shared an ordinary fish globe throughout the winter with 

 a mammoth larva. Besides these 1 tested a number of my specimens kept 

 permanently in jars since the time of their metamorphosis. The move- 

 ments were strikingly similar in all, even in the case of specimens kept in 

 water but just deep enough to cover the bodies. A mere lift of the head 

 was here sufficient to take in air in the ordinary fashion, and the nares 

 could be easily raised, and often were raised, above the water for hours 

 of air breathing. Yet, when the heads were under the respiratory move- 

 ments did not differ from those of the specimens in deeper water where 

 the indolent animals seemed so loath to rise to the top for air. Instead of 

 ten seconds, Hay gives five or six seconds as the period for A. tigrinttm. 

 For A. microstoinum, however, he gives the period as eight to twelve sec- 

 onds. Another point of difference is that Hay describes the expulsion of 

 the water as by the mouth, whereas, by the use of colored fluids, I ob- 

 served that it was, in all cases, at least partially by the nostrils, and often 

 entirely so, the mouth not being opened at all. 



It was while observing these movements that I chanced to observe that 

 identical movements took place periodically in the throat of the large larva 

 (the one whose head is shown in figure 3, plate VI), and the same move- 

 ments were verified immediatelji in all the larvae I had at hand, a dozen 

 or more. The large specimen was the only one sufficiently quiet to enable 

 me to time them ; they were regular but much slower than in the adult, 

 averaging very close to one in twenty seconds. I was surprised to find 

 that, in the larva as well as the adult the nostril was employed for the 

 expiration as well as the inspiration, although the former is by the gill 

 clefts as well. By pressing a large globule of majenta solution from a 

 pipette in such a manner that it remained cleanly suspended just in front 

 of the nares, the whole process could be perfectly observed. The globule 

 becomes slowly drawn out into a fine point, which, approaching the nos- 

 tril, is drawn in continuously, as a fine stream, during the whole interval 

 until the moment of sudden contraction by the buccal floor. This move- 

 ment sends strong jets of the colored fluid out again through the nostril, 

 and at the same time small clouds of fluid are expelled from the gill open- 

 ings. I have since tested many larvae, down to ten centimeters in length, 

 and I find the process common to all, though differing noticeably in the force 

 and size of the streams. For in some the nostrils are very fine. I w?s 

 interested to note further the independence of tliis process from any con- 

 nection with branchial respiration. For when watched from below, or 

 from the side, it becomes quite plain that the clouds of colored fluid ex- 

 pelled do not touch, by a long way. the gills, not even when these chance 

 to be dependent rather than arching upward, etc. The entire mass of col- 

 ored fluid is always shot strongly downward as well as backward, passing 

 usually exactly under the fore leg. I may add that this bucc il respiration 

 is very probably common to all branchiate Urodela. I fTnd the same phe- 

 nomena in Necturas. 



232 



