Production of Shiga-like toxin (SLT) by enterohemorrhagic Escherichia coli (EHEC) is controlled by phage conversion, and specific phages carry either the SLT-I or SLT-II operon. EHEC strain 933 produces both SLT-I and SLT-II. Previous studies demonstrated that the vast majority of phages recovered from strain 993 have hexagonal heads with short tails and encode SLT-II. However, conflicting results were obtained concerning the properties of SLT-I converting phages from strain 933. The present study reexamined the recovery of phages from 933 by various methods and characterized the restriction fragments from strain 933 DNA that hybridized with radiolabeled DNA from the SLT-I converting phage 933J, which has an elongated head with a long tail, and the SLT-II converting phage 933W. In the present study, only SLT-II converting phages like 933W were recovered from strain 933. A set of restriction fragments that hybridized with DNA from phage 933J but not 933W was present both in wild type strain 933 and in the variant 933D, which produces only SLT-I and was shown here to be cured of phage 933W. The sizes of the restriction fragments in strain 933 that were homologous with phage 933J differed, however, from those of phage 933J. These data indicate that the phage we isolated and named 933J probably did not originate from strain 933 as we originally reported. The present evidence demonstrates that strain 933 contains both the SLT-II converting phage 933W and other sequences of DNA homologous with phage 933J that probably represent a defective SLT-I converting phage.