The Ity locus affects the net increase in numbers of Salmonella typhimurium in the liver and spleen of infected mice. There has been controversy, however, about whether the effects of this locus are due to differential killing of S. typhimurium or differential growth rates of S. typhimurium in mice. Our studies using S. typhimurium aroA mutants, which do not grow in vivo, demonstrate that growth of the infecting salmonella is necessary for the observation of the Ity phenotype. To examine the effects of the Ity locus on the growth and killing of fully virulent salmonella, we infected Ity-congenic mice i.v. with stationary phase S. typhimurium containing a single copy of the plasmid pHSG422. This plasmid exhibits defective replication at body temperature and is diluted out during salmonella growth in vivo. Thus, the frequency of plasmid-containing salmonella recovered from mice provides a measure of salmonella cell divisions in vivo. Inasmuch as the numbers of plasmid-containing salmonella are only slightly affected by bacterial division, any decline in the numbers of plasmid-containing salmonella is an unbiased measure of killing. By infecting mice with these plasmid-containing salmonella we observed that: 1) during the first four h post infection (during blood clearance of injected salmonella) there is about 3-fold more killing of salmonella in Ity(r) mice than in Ity(s) mice; 2) from 4 to 44 h postinfection (after blood clearance is completed) there is little if any additional killing in either Ity(s) or Ity(r) mice; and 3) during the first 48 h postinfection there is about 18-fold more growth of salmonella in Ity(s) mice than in Ity(r) mice. Thus, the major effect of the Ity locus on resistance to salmonella, is the regulation of growth within a 'safe' (relatively nonbactericidal) site in the liver and spleen.