Phage Mu, a temperate phage, possesses an invertible region of DNA 3kb in length, the G region. The orientation of the G region determines the host specificities of phage Mu (Van de Putte et al . , 1980). The G region codes for tail polypeptides, the products of the S and U genes. In one orientation ( + ) the phage adsorbs to and infects E. coli K12 (Mu.K). In the opposite orientation (-) the phage adsorbs to and infects a strain of C. freundii (Mu.F). However Mu also plates on E. coli C, Serratia marcescens, E. cloacae and several other hosts. Rice (1980) was able to raise specific antisera to the Mu.K form and the W.C (Mu.C equivalent) form of Mu. However no antisera was raised to the Mu.F form. We attempted to isolate gin̄ mutants, in which the G region is not inverted; such a mutant would provide a specific antigen for the G(-) Mu.F phage. Presumptive gin̄ mutants were identified as phage that could plate only on one host, E. coli K12 or C. freundii. A specific antisera was raised to the MH₄₄₀₀gin̄ G(-) phage. Attempts were made to isolate the Mu.KC phage as seen by Jamieson (1971) and Rice (1980). For this purpose, heat induced lysogens of E. coli K12 were used, as opposed to lytic Mu.K lysates; the former give high titre lysates of G(-) and G(+) phage. The resulting Mu.KC lysogens were examined for their plating behaviour and neutralisation by the anti W.C serum. Two classes of phage were isolated, Mu.KC'' and Mu.KC'''. These phage differed from the Mu.KC phage seen by Jamieson and Rice as judged from their limited neutralisation by the anti W.C serum, and they also differed from each other with respect to their plating on C. freundii and the extent to which they are neutralised by the anti W.C serum. The Mu.KC" forms do not plate on C. freundii , e.o.p. <8 x 10̄⁹, and are not neutralised by the anti W.C serum. Whereas the Mu.KC''' forms do plate on C. freundii, e.o.p. 10̄¹-10̄², and exhibit a low level of neutralisation with the anti W.C serum. E. coli C lysogens of phage Mu.KFC were also isolated. These also fell into two distinct classes as judged by their plating ability on C. freundii ; Mu.KFC'' forms do not plate on this host, e.o.p. <5 x 10̄⁸, while the Mu.KFC''' forms do plate on C. freundii, with an e.o.p. 1 x 10̄¹ (comparable to the Mu.KC'' and Mu.KC''' mentioned above). During the study, differences in the plating ability of lytic and induced lysates propagated on the same host were observed. For example, compare the E. coli K12 system; lytic Mu.K phage plate on C. freundii with an e.o.p. of <10̄⁹, while induced Mu.K plate on this host with an e.o.p. of 1. Again, lytic Mu.KF lysates plate on E. coli K12, e.o.p. 10̄¹ , and an E. coli C, e.o.p. 2 x 10̄⁵. However, induced Mu.KF lysates plate on E. coli K12, e.o.p. 10¹, but these lysates were not observed to plate on E. coli c, e.o.p. <2.5 x 10̄⁹ . And yet again, lytic Mu.KC'' and Mu.KC''' plate on E. coli K12 with an e.o.p. of 10̄² -10̄³, however induced Mu.KF'' and Mu.KFC''' lysates plate on this host with a e.o.p. of 2.0 - 8.5. The MH₄₄₀₀ strain, gin̄ (albeit leaky), enabled the detection of the role of G inversion during plating Mu onto different hosts. This phage plates with a reduced frequency when G inversion is involved, e.g. induced MH₄₄⁰⁰.K plated on C. freundii with an e.o.p. of 10̄³. The e.o.p. of induced MH₄₄⁰⁰ .KFC on E. coli K12 (4.5 x 10̄⁴), was observed to be significantly lower than the wild type Mu (Mucts61) on this host (e.o.p. 2.0 - 8.5).