34 K. G. HOSKINS 



which this factor was ignored or inadequately controlled. The number of 

 animals in a cage, inequality in the size of animals in a given group, varia- 

 tions in lighting, exposure to drafts, differences in the distance from the 

 source of artificial heat, differences in cleanliness of cages, different de- 

 grees of infestation with parasites these are some of the variables which 

 can easily creep into growth experiments. Sex differences, which may be 

 marked, should always be taken into account. The desirability of selecting 

 experimental and control animals from the same litters is obvious, but 

 sometimes is overlooked. 



In the last analysis experiments of the type discussed amount to 

 statistical studies. Accordingly, all the criteria of statistical science 

 should be observed in such researches. The observance of such criteria 

 has, as a matter of fact, been the exception rather than the rule, and many 

 of the reported results are of correspondingly slight value. A high degree 

 of "spontaneous variability" is characteristic of the laboratory animals 

 that are most commonly employed. To demonstrate, therefore, that the 

 administration of a given gland substance brings about any characteristic 

 change in growth or organ weight, a large number of cases is usually es- 

 sential. In the face of such high variability and of the possibility of 

 mere coincidence, the attempt to draw valid conclusions from observations 

 on, say 10 animals, betrays an ignorance of the elementary principles of 

 statistical science. "Averages," under such circumstances, are meaning- 

 less, though frequently cited as the gist of results secured. 



As a variant upon oral administration, siiblingual is sometimes em- 

 ployed. It is said that the oxytocic effect of pituitrin can be obtained 

 in this way. The method has been little used either clinically or in 

 experimentation. 



Administration by Rectum'. The effects of the digestive enzymes 

 upon the endocrin gland substances are largely unknown. In case of the 

 thyroid the fact that ingestion of gland substance is capable of warding off 

 the ill effects of deficient secretion shows that we are dealing with a 

 stable body. In the case of epinephrin, on the other hand, oral adminis- 

 tration is quite without effect, as judged by pharmacologic criteria. It is 

 true that many claims have been made of the efficacy of such administra- 

 tion clinically, but therapeutic evidence of this sort is proverbially falla- 

 cious. It has been stated that in the case of most of the endocrin glands 

 better results are seen rod by rectal than by oral administration. This 

 possibility should ho thoroughly investigated, both on account of its clin- 

 ical bearing and its importance in experimental work. The writer has 

 recently ohserved that in case of epinephrin rectal administration has 

 approximately the same degree of efficacy as has hypodermatic. In some 

 aperimental animals no effect could he detected from doses as large as 5 

 milligrams, even when so sensitive- a test object as the denervated gut was 

 used. In other cases slight chano-es in blood pressure and marked in- 



