220 VITAMIN D GROUP 



more vitamin D than the reference oil. But how much more? By the curve, 

 it is seen that 44% ash corresponds to 11.6 I.U. However, the reference 

 group in this assay showed only 42.3 % ash, which corresponds to 9 I.U. 

 Apparently, then, the chicks, for some unknown reason, are responding a 

 little low, and a correction is called for: X:11.6: :10:9, or X = 12.9. Thus 

 the unknown oil is found to be supplying 12.9 I.U. of vitamin D per 100 

 g. of the diet. 



The example given above is a practical, everyday illustration of the use 

 of response curves, and it introduces the important subject of error. For a 

 discussion of error, the reader is referred to the paper by Massengale and 

 Bills^^ from which Fig. 17 was taken, and especially l)o the more professional 

 discussions of statistical methods given by Coward-"* and Bliss.^^ Sugges- 

 tions on means of minimizing error are given in the technical articles 

 previously mentioned,-^ ■ ^^ and the chapter by Guerrant in the second 

 volume of Gyorgy's book.^ 



In 1951 a method for the determination of vitamin D with radiophos- 

 phorus was described by Snyder et alr^ Applied as yet only to the rat, it 

 would appear to be adaptable also to the chick. This method is the fifth 

 satisfactory means of measuring the response of animals to vitamin D, the 

 others being the line test, x-ray, bone ash, and growth methods. The new 

 method is claimed to give results comparable in accuracy to the other 

 methods, and it has the advantages of requiring distinctly less time and 

 labor, and covering an extraordinarily wide span of units (0.5 to 50 units) 

 in one test. 



The procedure is as follows. Young (21-day) rats are placed on ricketo- 

 genic diet, and, after 16 days, are given by mouth a single dose of the test 

 solution of vitamin D. Forty-eight hours later a solution containing ap- 

 proximately 20 nc. of P^- is injected intraperitoneally. After another 10 days 

 without further treatment, the forepaws of the still-living rats are located 

 beneath the window of a Geiger counter. The observed radiation count is 

 interpreted by means of a response curve established from points at the 

 upper and lower limits of linearity at the time of the assay. 



In the course of twelve series of experiments over a year it was noted 

 that the slope of the response curves varied considerably. This is another 

 way of saying that the response of animals, or even of groups of animals, 

 is variable, as is well known. It is because of this variability that it is 



26 O. N. Massengale and C. E. Bills, J. Nutrition 12, 429 (1936). 



2^ C. I. Bliss, in Vitamin Methods, Vol. 2, p. 445. Academic Press, New York, 



1951. 

 28 R. H. Snyder, H. J. Eisner, and H. Steenbock, /. Nutrition 45, 305 (1951). 



