102 THE SIMPLER NATURAL BASES 



col ori metric processes which have been suggested more recently (see 

 p. 92). The quantitative estimation of adrenaline is of importance 

 in many physiological and pathological investigations. 



The most obvious, accurate and reliable method is based on a 

 comparison of the pressor effects of intravenous injections ; the 

 peculiarly evanescent nature of this adrenaline action greatly favours 

 accurate comparison, and in a suitably prepared animal equal sub- 

 maximal doses will produce time after time practically identical 

 effects ; this method is, however, inapplicable to very dilute adrenaline 

 solutions. The blood pressure of a cat, with brain and spinal cord 

 destroyed and without anaesthetic, reacts according to Elliott [1912] 

 " with mechanical accuracy," and by comparison with a standard 

 solution, Elliott assays the adrenaline content of the cat's supra-renal 

 gland with an error of croi mg., which is 3-4 per cent, of the total 

 amount present. 



The accurate pharmacological assay of preparations of the supra- 

 renal gland by means of the blood pressure was first carried out by 

 Houghton [1901] ; the blood pressure has further been used especially 

 by Elliott [1912], Hunt [1906], Sollmann and Brown [1906], Cushny 

 [1908, 1909], Schultz [1909, I], and Dale [Barger and Dale, 1910, I]. 

 Schultz employed dogs (with morphine, ether and curari) and cats 

 (with ether), Elliott and Dale almost exclusively decerebrate cats. The 

 doses are T^rsV m g- for dogs and ^V?V mg. for cats (which are more 

 resistant than dogs). Other blood-pressure methods, such as the de- 

 termination of the dose required to compensate for the vaso-dilator 

 action of a given quantity of nitroglycerine (Cameron [1906]) and the 

 determination of the minimal dose necessary to give a perceptible 

 pressor effect, are much less accurate. 



A second method employing the circulatory system but depending 

 on vaso-constriction instead of on blood pressure is due to Lawen 

 [1903-4] and has been improved by Trendelenburg [1910]. The rate 

 is measured at which, under a constant hydrostatic pressure, blood flows 

 through the vessels of a frog, of which the brain and spinal cord have 

 been destroyed ; the adrenaline to be estimated is added to the blood. 

 This method appears to yield moderately accurate results, but is la- 

 borious when many estimations have to be performed. The significance 

 of determinations by this method of adrenaline in serum has recently 

 been questioned by O'Connor [191 1, 1912, i] who finds that serum itself 

 causes vaso-constriction, quite apart from the addition of adrenaline 

 (see also Handovsky and Pick [1913, Ch. I]). Stewart [1912], and 

 Dale and Laidlaw [1912, 2] agree with O'Connor's objections to the 



