It is usually accepted that in precipitin reactions a definite quantity of precipitin interacts with a definite quantity of homologous proteid, and further that in ordinary circumstances the interacting quantities combine to give rise to a precipitate, the precipitum. From this it follows that, if a known weight of serum containing precipitin be allowed to react with various weights of homologous proteid and if the quantities of proteid be sufficiently varied, certain reactions will be completed and others will be incomplete. The addition of precipitin and of homologous proteid to the superfluid of each reaction will determine the presence of excess of precipitin or of homologous proteid in such fluid. Conversely, if a known weight of the homologous proteid be allowed to interact with various weights of serum containing precipitin, it will be possible to determine in a similar manner those reactions which have been completed and those in which the homologous proteid or the precipitin is in excess. If the first stated proposition be correct, the results of these two series should coincide. A definite weight of serum containing precipitin should interact with a definite weight of proteid to from the precipitum.