10 INFLUENCE OF FOOD PRESERVATIVES ON HEALTH. 



experimental treatment except as a punishment for crime. In the 

 second place, the intelligence of the human animal may also be utilized 

 in the study of 'the effects produced. Symptoms which the lower 

 animals might have of distress or malaise when in the incipient stage 

 might escape notice altogether, whereas similar symptoms in a man 

 would be described. Further, it must be admitted that animals under 

 confinement, as is necessarily the case when experiments are made 

 with them, are not wholly in a normal state, whereas the man who 

 volunteers for an experiment of this kind would not chafe or become 

 restive under confinement. Again, it must be considered that as the 

 investigations above outlined are particular^ applicable to the diges- 

 tion and health of man, it is evident that the experiments made upon 

 man himself would be the most decisive in all cases. 



The one great disadvantage of experiments of this kind is the 

 inability to absolutely control the " experimentee." Where a large 

 number of persons is to be considered and the experiment is to extend 

 over a long period it is evidently impracticable to secure a direct per- 

 sonal control of every action of each one during the whole time. In 

 the present case the young men selected (who volunteered for the 

 experiment) continued their usual vocations. They were simply placed 

 upon their honor and neither watched nor confined. The data which 

 are obtained in this way are therefore open to the objection, in some 

 cases, that the rules and regulations set for the conduct of the experi- 

 ment may have been transgressed without the knowledge or consent 

 of the observer. While this is a valid objection and should have full 

 consideration, it must not be forgotten that among the twelve young 

 men upon whom the experiments were conducted it is not likely that 

 the violations of their pledge of honor would be sufficiently numerous 

 to affect in any marked degree the results as a whole. Further, it 

 must be remembered that the greater number of those upon whom 

 experiments were made were young men of approved character, many 

 of whom had college training, and a large majority of whom were 

 engaged in scientific pursuits. All these facts are of more or less 

 importance in considering the character of the data secured. It would 

 be unwise to claim that among so many persons and amid so man} 7 

 temptations no violation of the pledge took place, yet it must be 

 admitted that upon the whole we can be reasonably certain that the 

 obligations voluntarily assumed were discharged faithfully and con- 

 scientiously. Any departures from the set rules of conduct which 

 might occur would not be made with any design of affecting the data, 

 and therefore, as a whole, the errors which might arise from this 

 source would, according to the doctrine of probabilities, be largely 

 compensatory. Thus, while in any individual case the data might be 

 rendered unreliable by reason of such departures from the set rules, 

 the results as a whole would not be seriousty affected. The plan of 



