Infectivity titrations of milk-borne murine mammary tumor virus (MuMTV) from different sources or prepared in different ways or stored for periods of time have been compared. Titration curves were in general reproducible for MuMTVs of different sources or handled in different ways and for different methods of measurement, such as hyperplastic alveolar nodule (HAN) development, tumor development, or MuMTV antigen secretion in third-lactation milk, The curves had characteristic shapes with a low incidence of infection at low dilutions of milk, high incidences at intermediate dilutions, and low incidences at high dilutions. Infectivity incidences were unaffected by dilution over the range 10(-2) to 10(-5). The curves did not change appreciably with time of storage of milk at liquid N2 temperature for periods up to 3 years. Rate zonal fractionation of RIII milk gave zones with bioactivities which were not proportional to B-particle content. Upon dilution, the bioactivity of Zone 3, rich in B particles, and Zone 5, poor in B particles, increased, while the bioactivity of all the other zones usually decreased with dilution. The low incidence of infection at low dilutions may have been due, in part, to an immune response of the inoculated mouse. Administration of inactivated virions 4 h prior to or with MuMTV inoculations gave some evidence in support of this hypothesis but the complexity of the bioassay system for MuMTV lends uncertainty to interpretation of results.