The absence of a satisfactory animal test for vitamin P has had several disadvantages. Clinical trials have been made not only with hesperidin but also with citrus concentrates that must have contained at least one other active substance and may even have been devoid ofhesperidin. This compound, although its chemical structure is established, cannot itself be estimated by any sufficiently specific chemical or physical tests. Thus the vitamin P activity of the concentrates used, and possibly also of the hesperidin, can only be expressed in terms of the clinical-results attained; such expression is of very limited precision, and is, moreover, dangerously liable to lead to argument in a circle. Furthermore, attempts to concentrate the active principle from natural sources, to say nothing of isolation and synthesis, have been handicapped or frustrated by the inability to relate chemical properties with physiological activity. The same difficulties have naturally also impeded clinical investigation, and the pioneer work of Armentanp, Bentsath, Beres, Rusznyik & Szent-Gyorgyi [1936] and of Scarborough [1939] has been followed by only a few other trials [cf. Rapaport & Klein, 1941;Goldfarb, 1941]. Even their limited results, however, have given a more than academic interest to the problem of producing experimental avitaminosis-P in laboratory animals.The early attempts of Szent-Gyorgyi and his co-workers [Bentsath, Rusznyak & Szent-Gyorgyi, 1936; 1937] to demonstrate vitamin P deficiency in guinea-pigs were not found convincing: Zilva and Moll [1937] were both unable to confirm SzentGydrgyi's results, and Szent-Gyorgyi [1938] was himself unable to repeat them later. We were therefore interested in a paper by Zacho [1939] who used as a criterion of vitamin P deficiency the 'specific' symptom of capillary fragility, sometimes confused, as Scarborough [1941 a] has pointed out, with increased capillary permeability. The criteria used by the earlier investigators-increased survival time or a reduced number of haemorrhages in scorbutic guinea-pigs-were clearly unsatisfactory even on theoretical grounds: death of guinea-pigs from scurvy is complicated by intercurrent infectioris, and estimation of the animals' haemorrhagic 'score' post mortem must be largely a matter of subjective judgement. Both would be accompanied by a high variance, so that clearcut results would be difficult to get even with large numbers of animals.Interest in Zacho's paper has also been aroused in the Hungarian workers, andRusznyaik & Benk6 [1941] have stated that they were able to confirm Zacho's findings, not only on guinea-pigs but also on rats. When their note appeared, we had already overcome certain difficulties encountered in trying to apply Zacho's technique, but we had considered it inadvisable at that stage to publish our results because we wished to try to put the method on to a quantitative basis, in our opinion the acid test of any claim for an animal response to a physiologically active substance. Without intending to detract at a...