Excretion of Creatine in Carbohydrate Starvation. 209 



it because other observers have added this substance to urine instead of 

 aceto-acetic acid. 



The sodium salt of aceto-acetic acid produces a much more marked effect 

 than the other acetone bodies. Even when added to urine in small amounts 

 the colour obtained with picric acid and soda is not darker, as stated by 

 Krause(16) and others, but is actually lighter, and when present in large 

 amounts the colour is very much lighter (Table II and fig. 1). The error 

 is not eliminated on standing but increases. 



Table II shows the Effect of Increasing Amounts of Aceto-acetic Acid on the 

 Estimation of Creatinine in Urine. 



Aceto-acetic acid added 

 to urine. 



Scale 

 reading. 



Creatinine. 



Error in the creatinine 

 determination. 



giro, per 



grm. per 24 hrs. 





grm. per 



grm. per 





100 c.c. 



in 1500 c.c. 





100 c.c. 



100 c.c. 



per cent. 











7 



0-116 







-0234 



0-35 



7 -33 



0-111 



0-005 



4-3 



-0468 



-702 



" 8 



-101 



0-015 



12 -9 



0-093 



1 -4 



9 



0-09 



026 



22 -2 



0-187 



2 -8 



11-3 



0-072 



0-044 



38 



0-374 



5-6 



14 66 



0-055 



0-061 



52 -6 



Thus if the concentration of the sodium aceto-acetate is only 0'02o4 or 

 0468 per cent, the creatinine estimation is too low, the actual errors being 

 - 005 and O'Olo grm. respectively. The error produced by larger amounts 

 is very striking, for if the concentration is increased to - 374 per cent, the 

 error is as great as 0'061 grm., and the percentage error in this case is 

 52*6 per cent. As amounts of aceto-acetic acid up to a concentration of 

 - 4 per cent, may be excreted in diabetes, the error caused in such cases 

 must be very great. If the error in the creatinine determination be plotted 

 against the concentration of the aceto-acetic acid the resulting curve is 

 almost a straight line (fig. 1). 



The chemistry of this action is at present engaging our attention. 



It was not possible to isolate the pure acid and add it to urine, but this 

 does not matter, as the aceto-acetic acid is excreted in the urine partly as 

 the free acid and partly as a salt. Moreover, in the process of estimating 

 the creatinine an excess of caustic soda is added, and this must convert all 

 the free acid into the sodium salt. 



The solution of sodium aceto-acetate used in our experiments also con- 

 tained 0'37 grm. ethyl alcohol to 1 grm. of aceto-acetic acid. 



The presence of ethyl alcohol in the sodium aceto-acetate solution (due to 

 its mode of preparation from aceto-acetic ethyl ester) is a possible disturbing 



