It has been shown that passive transfer by these sera is completely unrelated to the egg albumin-anti-egg albumin system, as demonstrated by a failure of passive transfer by some antisera containing ample anti-egg albumin and persistence of passive transfer in other antisera from which all anti-egg albumin had been removed by precipitation with homologous antigen. Three preparations of non-precipitating anti-egg albumin have been shown to have sensitizing capacities which bear no relation to their non-precipitating anti-egg albumin contents. From a portion of one of these the non-precipitating anti-egg albumin was removed without impairing its sensitizing ability, while in another portion obliteration of the sensitizing capacity was accomplished without reducing the anti-egg albumin.

Evidence is presented to show that there are at least two possible antibodies in anti-egg albumin sera which are capable of inducing skin sensitivity and that they are antibodies against egg white impurities in crystalline egg albumin other than anti-conalbumin, anti-ovomucoid, and anti-lysozyme.

The usefulness of a suitable quantitative precipitin technic for the analysis for antibodies against antigen impurities and for their selective absorption from sera is illustrated. The principle governing the procedure is described. The technic allows for the determination of a given trace antibody by working with such small concentrations of its purified specific antigen that whatever other antigen-antibody compounds are formed simultaneously with that to be determined will be below their solubility levels and consequently will not contribute appreciably to the precipitate.