__237 
blood, the result might be more conclusive. I therefore ligatured the 
gland arteries and veins of three fish which had been weighted as 
usual for six hours, and of one unweighted (control) fish and left the 
preparations for an hour. I also ligatured the caudal arteries and 
veins of two out of the three weighted fish so as to be able to make 
smears of blood subjected to the same conditions as that from the 
active gland save that it was in contact with another tissue, though 
this precaution of ligaturing the tail vessels was hardly necessary since 
in all cases the heart had been removed. On making smears of the 
blood contained in these three activated glands, the inactive gland 
and the tails of two of the fish which possessed the active glands I 
could detect no evidence of haemolysis in any case. 
Finally I made an extract of the activated gas gland in order to 
determine if it had any effect on fresh blood. I ground up with clean 
silver sand and 7 cc. of normal saline the activated glands of seven 
large Pollack, and added 2,5 cc. of the filtrate to 1 cc. of freshly 
aerated blood of the Pollack. I left this mixture, and also a control 
preparation in which the 2,5 cc. of the filtrate was replaced by 2,5 ce. 
of normal saline, for two and a half hours, but on making films of the 
two preparations no sign of haemalysis was discernible. 
I may also mention here that I employed the greater part of the 
extract just mentioned to see if it had any action on haemolysed ar- 
terial blood similar to that of ferricyanide. It is common knowledge 
(see HALDANE, 3) that a concentrated solution of ferricyanide, if allowed 
to come into contact with an equal volume of a mixture of equal parts 
of fresh arterial blood and water (this latter haemolyses the blood 
and so enables the ferricyanide to come into contact with the oxy- 
haemoglobin, as before stated), the oxygen is at once evolved with 
great energy. According to the theory of gas production I have out- 
lined at the beginning of this paper, the gland cells should secrete a 
substance similar in its properties to ferricyanide. I employed the 
simple form of gas volumetric apparatus designed by BARCRoFT and 
RoBERTS (Journ. Physiol., Vol. 39, 1909) for blood gas analysis but 
failed to obtain conclusive results. All I can say is that I obtained 
precisely similar results on adding a saturated ferricyanide solution 
(instead of extract) to a similar quantity of blood. 
In addition to these experiments upon the activated glands of 
Pollack, others were conducted in connection with the glands of some 
other fish, all unfortunately with the same result — no signs of haemo- 
lysis were observed. Two Mullets (Mugil chelo) were weighted for 
5 hours, one small Wrasse (Ctenolabrus rupestris) was weighted for 
