Cclxxxiii. the Influence of the Length of the Test Period on the Accuracy Obtainable in a Vitamin B1 Test

IT was shown by Coward [1933] that the accuracy obtainable in a vitamin A test by the "increase in weight" method was increased very little by prolonging the test period beyond three weeks. Since then, enough results of vitamin B1 determinations by the method of the "increase in weight" of rats have been accumulated to make it possible to carry out a similarexamination ofthe duration of a vitamin B1 test. The technique of the test has been described by Coward et al. [1933]. Briefly, it consists of measuring the increase in weight of rats which, having become steady in weight on a diet deficient in vitamin B1, are then given doses of (a) the International Standard (I.s.) for vitamin B1 or (b) the substance under examination. With the description of the test was published an estimate of the accuracy obtainable. It was much greater than that obtainable in a vitamin A test of the same length of time. Thus a smaller number of animals would be required for a vitamin B1 test for a given degree of accuracy than would be required for a vitamin A test of similar accuracy. In the determination of the vitamin B1 content of a substance we have generally divided the rats into 4 or 5 groups, two of the groups being given doses of 0x01 and 0-02 g. respectively of the I.s. and two of the groups doses in the ratio of 1: 2 of the substance under examination or, if no information on the possible potency of the substance was available, the doses chosen for a preliminary test were in the ratio of 1: 3 or 1: 5. A group of rats given no dose was included when it was desired to show whether even a large dose of the test substance was or was not better than no dose at all. The different groups have generally contained equal numbers of bucks and does but it was shown in the paper by Coward et al. [1933] that the curve of response of bucks was very little steeper than that of does; hence it was not really necessary to have equal numbers of each in each group. By this means we have obtained a large amount of information not only on the variation in response to doses of vitamin B1 but also on the variation in the slopes of the curve of response. It has confirmed our previous estimate of the high degree of accuracy attainable and it has shown that the accuracy obtainable by carrying on the test for three weeks is only very slightly greater than that obtainable by carrying it on for two weeks.