IN considering the agglutination-test for bacillary dysentery, the War Office Committee on Dysentery in 1918 decided, on the available evidence (Kruse, C. J. Martin, Andrewes and Inman, Gettings, Murray), that five strains of Flexner-Y bacilli, representing five separable, though closely related, serological types, might be taken as adequately covering the whole serological range of the species. The Vaccine Department of the R.A.M. College, Millbank, and the Standards Laboratory at Oxford accepted these findings as basis for practical work, and the latter institution proceeded to issue the five varieties of "Standard agglutinable cultures" and "Standard agglutinating sera."At that time two types only were being produced, viz. "Flexner" and Y. These strains later on were relabelled V and W respectively in the new classification. Before any standard cultures of dysentery bacilli at all were issued, an adequate number (120) of sera of healthy human adults had been tested upon standard cultures of B. Shiga and B. Flexner (later V), and a calculation of the results made it possible to fix a limit above which the agglutination by a patient's serum of any standardized suspension of these bacilli could be taken as indicating a dysentery infection. The figures thus obtained cannot be applied to work with unstandardized suspensions, for not only do different strains of one serological type differ profoundly in agglutinability, but different suspensions of the same strain show considerable variations in this respect. The "normal limit," if it is to apply to all standardized suspensions of a type, must be expressed in standard units. Each serological type must be treated as an entirely separate entity.The absolute magnitude of the reduction-factor for agglutinability of the original arbitrarily chosen standard suspension of any bacillary type may be fixed at anything one pleases. In practice it is decided by convenience. For B. Shiga and B. Flexner the size of the reduction factor of the original standard suspension was arranged so that the normal human sera should not be found