MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 271 



Sources of uncertainty in this kind of eooperimenting. — One point wnich has hardly received 

 tln> attention il deserves in discussions of this kind is the uncertainty of the nitrogen balance in 

 any given cum' as a measure of the actual influence of a given condition upon nitrogen metab- 

 olism. This has been emphasized elsewhere in the present memoir (see pp. 393 and 394). Differ- 

 ences which look large in a table of figures are often far inside the unavoidable variations in 

 actual experimenting. 



Even when the differences are significant the interpretation may be erroneous. A striking 

 illustration of the danger of such error is found in the current discussion of the question we are 

 now considering. For a number of years past writers upon this subject have insisted most 

 positively that alcohol, instead of being a protector of protein, is a protein poison. This theory 

 is based almost wholly upon the experiments of Miura, Schmidt, and Schoencseiffen. The experi- 

 ments of Neumann, Rosenfeld-Chotzen, Clopatt, and Rosemann, not to speak of others, including 

 our own, have shown that this theory was wrong and have given us a very plausible hypothesis 

 to explain why it was wrong. 



We can not insist too strongly upon the danger of drawing positive conclusions from figures 

 for nitrogen balance as a measure of protein protection by either alcohol or sugar or starch or 

 fat. Certainty comes only with careful planning and execution and manifold repetition of 

 experiments. 



Incidentally, it is to be noted that the excretion of nitrogen in the urine is not necessarily an 

 exact measure of the amount of proteid broken down in short periods, since the time between the 

 disintegration of the protein and the appearance of the nitrogen in the urine, the so-called 

 nitrogen lag, varies widely. The longer the experimental period the less the error from this 

 source. 



Finally, there is the unsettled question as to how much of the protein metabolized is that of 

 food and how much comes from organized tissue. 



F'uml conclusions regarding th<_ influence of alcohol upon prof, in notoholmn. — The 

 experiments and considerations above cited seem to us to warrant the following conclusions: 



1. The power of alcohol to protect the protein of food or body tissue, or both, from 

 consumption is clearly demonstrated. Its action in this respect appears to be similar to that of 

 the carbohydrates and fats; that is to say T , in its oxidation it yields energy needed by the body, 

 and thus saves other substances from oxidation. In this way alcohol serves the body as food. 

 Just how moderate quantities of alcohol compare with isodynamic amounts of sugar, starch, and 

 fat in the power to protect protein from catabolism is not yet settled. Apparently it is in some 

 cases equal, in others inferior, to these substances. It is by no means certain that the fats and 

 carbohydrates are always equal to each other in this power. 



2. Alcohol appears also to exert at times a special action as a drug. In large quantities it is 

 positively toxic, and may retard or even prevent metabolism in general and proteid metabolism 

 in particular. In small doses it seems at times to have an opposite influence, tending to increase 

 the disintegration of protein. This action, though not conclusively demonstrated, is very 

 probable. It offers a satisfactory explanation for the occasional failure of alcohol to protect 

 protein, the assumption being that the two tendencies counteract each other. The only 

 justification for calling alcohol a proteid poison is found in this disintegrating tendency. This 

 pharmacodynamic action of alcohol appears to be temporary and most apt to occur with people 

 little accustomed to its use. The circumstances under which such action occurs can not now be 

 fully defined. 



Influence of coffei upon protein metabolism in flux, experiments. — In some of these experi- 

 ments alcohol was administered with coffee, in others with water. It might be thought that the 

 presence of the coffee would interfere with the action of the alcohol. a The figures give no 

 support for this view, as is shown in the following tabular statement. 



- e W Ibury and Egbert, A Physiologic Consideration of the Food Value of Alcohol, Jour. Am. Med. Assc., 



Mar. 31, 1900. 



