Excretion of Creatine in Carbohydrate Starvation. 



211 



These figures have been plotted on the fig. 1 previously referred to and they 

 lie fairly close to the curve. The figures for the second day of Experiment I 

 and for both days of Experiment II lie somewhat below the curve, while the 

 figure for the third day of Experiment I lies a little above the curve, but the 

 difference is in no case great. 



'005 0-1 0-13 0-2 0-25 0-3 0-35 0-4°;, 



Fig. 1. — The curve shows the error in grammes per 100 c.c. in the estimation of creatinine 

 caused by increasing amounts of aceto-acetic acid. Abscissae : percentage concen- 

 trations of aceto-acetic acid. Ordinates : the error in the creatinine determinations 

 expressed in grammes per 100 c.c. © = the error in the creatinine determinations 

 on second and third days of Expt. I (Table VII). □ = the error on first, second, 

 and third days of Expt. II (Table VIII). 



It must be remembered that the figure for the concentration of the aceto- 

 acetic acid in the diet experiments was obtained by the Messinger-Huppert 

 method, which makes no distinction between acetone and aceto-acetic acid. It 

 is true that acetone is only present in urine in small amounts, but in consider- 

 ing the effect of the aceto-acetic acid some allowance should be made for the 

 amount of acetone present. A diminution in the concentration of the aceto- 

 acetic acid would make the points below the curve more nearly approximate 

 to the curve but would displace the point above the curve away from the 

 curve. 



The general agreement between the errors produced by aceto-acetic acid in 

 these experiments, and the errors produced when sodium aceto-acetate con- 

 taining ethyl alcohol is added to urine, point to the conclusion that it is the 

 aceto-acetic acid itself and not the alcohol that causes the errors, in the latter 

 case. 



Our experiments agree with those of Greenwald and show that the aceto- 



