CONDITIONS PAVOUEING rEKMENTATION. 27 



obviated in the future, first, by a more strict attention than I at 

 first bestowed upon obtaining or providing for the precise degree 

 of neutralization desired — secondly, by attention to a fact of the 

 full significance of which I was not at first aware, viz. that the 

 acidity of the urine in certain cases becomes notably diminished 

 by boiling. This diminution in acidity seems invariably to occur 

 where acid phosphates are precipitated by boiling. The result is, 

 that, wherever this occurs, the liquor potassse is liable to be added 

 to the boiled urine in a harmful excess, since the quantity sup- 

 posed to have been required, and enclosed within the liquor- 

 potassse tube, was calculated upon the degree of acidity of the 

 fluid before it had been lessened by boiling. In cases of pro- 

 longed heating this diminution of acidity may have been very ap- 

 preciable. In a specimen of phosphatic urine whose acidity was 

 equal to 5| minims of liquor potassse to the fluid-ounce, I found 

 the acidity only equivalent to 4 minims to the ounce after the urine 

 had been boiled for only 2 minutes. In all cases, therefore, in 

 which, after a preliminarytrial,phosphates are found to be deposited 

 by boiling the urine, it would be proper to estimate the acidity 

 with a specimen of the fluid which had already been boiled for 

 the period determined upon as that to be employed in the expe- 

 riments about to be made. This is a precaution which should 

 in future never be omitted"*. ... "It must not be forgotten 

 also that in boiling the measured quantity of urine in the retort 

 before its neck is sealed, any excessive spurting away of fluid may 

 cause an alteration in the proper ratio between urine and liquor 

 potassse." 



Another cause of variation in the time of fermentation in spe- 

 cimens of the same set is the diff"erent rate of boiling to which 

 the fluids have been subjected during the first period — that is, 

 whilst the fluid is being boiled over the flame. Eapid boiling in 

 a retort or flask with a capillary extremity will very frequently 

 cause the temperature to rise as much as 3^° 0. (nearly 7°F.) 

 above the boiling-point, as I first ascertained in 1873t. Acci- 

 dental variations in the rate of boiling, causing the temperature 

 to be raised to difierent points, may thus afiect successive fluids of 

 the same stock diff'erently. This, indeed, is another reason why it 

 is desirable to curtail the period of boiling over the flame and 



* I would rather say now that it is better not to use such a fluid at all in 

 these particular experiments. 



t See ' Nature,' vol. ii. (1870j p. 227. 



