12 ANNAL8 NEW YORK ACADEMY OF SCIENCES 



terior dorsal fin. Hyde ('08) has shown that all the centers governing 

 respiration in the skate, though of a segmental nature, are located in the 

 medulla. Since in the above operation only the posterior two-thirds of 

 the cord was destroyed, the nervous structures that govern respiration 

 were not affected. 



After the cord was destroyed, the tail was removed, the caudal artery 

 and vein being thus exposed. Blood was then taken for the determina- 

 tion of its freezing point. After this, the caudal artery was closed with a 

 small wooden plug covered with absorbent cotton. The animal with the 

 exception of the posterior part of the body was then placed in the tank 

 containing the experimental solution. After the desired time, a second 

 sample of blood was taken for a second determination of its freezing 

 point. The difference between the first and the second was a measure of 

 the change in the osmotic pressure of the blood of the particular animal 

 for the given time and the given solution. In a number of cases as many 

 as six samples of blood, usually about 5 c. c. each, varying with the size 

 of the fish, were taken from one specimen. The blood was drawn into a 

 small beaker and placed in an ice bath until the caudal artery of the fish 

 could be closed and the fish could be transferred back to the water. The 

 common freezing tube with the side neck for the insertion of an ice crys- 

 tal was not used on account of the large amount of blood that would thus 

 be necessary for each determination. A test tube with a smaller diameter 

 was used instead. Duplicate determinations of the freezing point of the 

 blood and distilled water demonstrated that the error due to undercooling 

 must have been small. The experiment was repeated in a number of 

 cases with uniform results, as will be shown later. SeYeral clean dry test 

 tubes were kept at hand in order to facilitate the determination of the 

 freezing point of a number of samples in the shortest space of time. I 

 found that about fifteen minutes were required for all the steps in the 

 making of a single determination. On account of necessary interrup- 

 tions, it was not possible to make the time intervals equal in all cases. 



The whole blood, including corpuscles and plasma, was used in the ex- 

 periments that follow. Hamburger ('95), Eoth (^99) and others have 

 asserted that the corpuscles are inert in determinations of the freezing 

 point. Moore ('08) found that the corpuscles of pig's blood had a A of 

 from 0.02° to 0.03° lower than that of the serum. Since in all the fol- 

 lowing experiments A was obtained in the way already indicated, the 

 error due to the presence of corpuscles would be approximately constant 

 in cases where the corpuscles were not laked. It would have been prac- 

 tically impossible to make the frequent determinations of A in these ex- 

 periments, had I stopped in each case to defibrinate and centrifuge each 



