VITAMIN A 265 



On the whole, we find no certain gain in uniformity to balance the 

 undoubted loss in delicacy when the higher rate of growth is employed 

 as a standard by means of which to make quantitative comparisons of 

 vitamin A values. 



In determining the relative vitamin potency of two or more mate- 

 rials, we recommend the determination by a suitable number of con- 

 trolled experiments, of the quantity of each material which contains 

 respectively, the amount of vitamin A needed to induce (say) 3 grams 

 gain per week. The relative vitamin contents of two given test materials 

 are inversely proportional to the weights of the material required to 

 support this arbitrary rate of gain. 



For an accurate quantitative determination, such as should be ex- 

 pected when the effect of some treatment or variation of condition 

 upon the vitamin value of a product is being studied, it would seem 

 best to feed in parallel 10 or more animals on each of the amounts 

 which are being used for direct comparison, but the number logically 

 should depend on the uniformity of the results. The average deviation 

 of gains in body weights in an 8-weeks experiment which aims at a 

 total gain of 24 grams should never be over 10 grams. If all conditions 

 are of the best this can be reduced to about 5 grams. 



The results of experiments carried out as here described may be 

 expressed in terms of the conventional standard units, a unit of vita- 

 min A being defined as that amount which when fed daily to a standard 

 test animal prepared as described, will suffice to support an average 

 gain in body weight of 3 grams per week during the test period of 4 

 to 8 weeks. (Obviously one may always express vitamin A values in 

 terms of this standard unit, and yet may state the number of such 

 units contained in one gram, or one kilogram, one ounce, one pound, 

 or in a 100-Calorie portion of the food, or in whatever other amount of 

 food may be regarded as convenient — as for example, the number of 

 units of vitamin A in a dozen eggs. In this monograph vitamin A 

 values are expressed in terms of units per gram.) 



However, it is to be kept in mind that one who desires may use 

 a higher level of feeding and make comparisons at rates of growth 

 higher than 3 grams per week, and yet may express the vitamin A 

 values thus found in terms of the now-recognized standard units by 

 establishing (or accepting as established) the relative amounts of 

 vitamin A which must be fed to induce the two rates of gain in ques- 

 tion, i.e., a gain of 3 grams per week and the higher rate at which the 

 experimenter has worked or proposes to work. 



Occasionally also, there are practical difficulties which make it im- 



