EXPERIMENTAL DIABETES 371 



If mineral acids be given, by mouth, to rabbits, the animals lose 

 the power of moving about, become comatose, and ultimately die 

 of collapse ; a symptom complex, at least, suggestive of diabetic 

 coma. If alkali be administered to acid-poisoned animals these 

 symptoms disappear. 



If the urine of such poisoned animals be examined it will be 

 found to contain an excess of ammonia. This must mean that, for 

 its neutralisation, the acid has not only used up all the fixed 

 alkali (Na, K, &c.) in the organism, but has appropriated some of 

 the ammonia which normally would have been transformed into 

 urea. The excretion of an excess of ammonia can, therefore, be 

 taken to indicate the neutralisation of all the fixed alkali by the 

 acid administered. 



A direct estimation of the alkalinity of the blood in acid 

 intoxication will show that most of the alkali has been neutralised. 

 One of the most important functions of the fixed alkali, in the 

 blood at least, is the carrying of carbonic acid, and since the 

 mineral acids possess so very much stronger affinities for alkali 

 than carbonic acid does, we would expect to find very little 

 carbonic acid in the blood of acid-poisoned animals. As a matter 

 of fact, Walter ( 31 ) found only 23 vols. per cent., the normal 

 amount being 30-40 vols. per cent. The immediate cause of .the 

 coma may therefore be the accumulation of carbonic acid in the 

 tissues; it remains where it is produced, for the blood, deprived 

 of its alkali, can no longer transport it quickly enough to the 

 lungs. 



Is there, then, really an acid intoxication in diabetic coma ? 

 The indications of this in the experimentally produced condi- 

 tion we have seen to be: (1) Excess of ammonia in the urine. 



(2) Marked diminution in the percentage of C0 2 in the blood. 



(3) Diminution of alkalinity of blood. Do these conditions exist 

 in diabetic coma ? 



It has been known for long that an excess of ammonia is 

 excreted in severe diabetes. In normal human urine, the amount 

 of nitrogen excreted as ammonia is only 3-6 per cent, of the total 

 nitrogen ; in diabetes, it may form 20-30 per cent, of the total 

 nitrogen (Magnus Levy 39 ). Since this represents ammonia diverted 

 from its normal metabolism into urea for the purpose of neutralis- 

 ing the excess of acid, it would be expected that alkali administra- 

 tion to diabetics would cause it to diminish. Such has frequently 



