Papilio memnon is a Swallowtail butterfly widely distributed in South-East Asia. The females are highly polymorphic and many of them are Batesian mimics. A previous paper gives an account of the mode of inheritance of seventeen of these female forms and here we describe the inheritance of nine from Java and Sumatra, an area we had not investigated before. We have also obtained further information on some of the forms whose modes of inheritance are not clear in the earlier paper. The nine new forms and one of those previously investigated (f. thunbergi) have been shown to be determined by the same complex locus or super-gene which controls the polymorphism already studied. The results from the present genetic investigation confirm the previous findings in P. memnon and P. dardanus that the resemblance of a form to its model is greater in the gene complex of the race in which the form is found than it is in intra-specific hybrid gene complexes. This suggests that the detailed resemblance of the mimic to its model has been built up gradually by the accumulation of modifiers. Of particular interest in this connexion is the mimic f. achates in Java and South Sumatra in which the heterozygote is a better mimic than the homozygote in the gene complex of its own race. Evidence has also been accumulated on the nature of dominance. Previously, complete dominance was present between most sympatric forms and absent between allopatric ones, both in P. dardanus and P. memnon. In contrast, the forms found in Java and Sumatra show only partial dominance between sympatric forms in most instances. However, in a hybrid gene complex the heterozygotes are even more intermediate. Thus, on these two islands the evolution of dominance has proceeded less far than elsewhere. There is ample evidence for the presence of dominance modifiers, but we are uncertain whether the absence of full dominance results from the modifiers having other, more powerful, selective forces acting upon them, or whether insufficient time has elapsed for full dominance to have evolved. In a complex Batesian mimetic polymorphism the forms must either be controlled by multiple allelomorphs at a single locus, by a group of closely linked loci (a super-gene), or by independent genes having complex epistatic interactions. In both P. dardanus and P. memnon it has been established that super-genes have been evolved. The present study throws further light on the super-gene in P. memnon. We have reason to believe that a number of very rare forms (some of which we have bred) result from crossing over rather than from point mutations, since if mutations were responsible we should have expected some of the patterns to be controlled by loci unassociated with the super-gene. This is not the case. Furthermore, in one instance, a double mutation would be required to explain the pattern. On the assumption that crossing over is the explanation of these rare forms we have postulated five loci. These control the presence or absence of tails (T), hindwing pattern (W), forewing pattern (F), colour of the basal triangle or epaulette on the forewing (E), and abdomen colour (B). Previously we had deduced that W must lie between T and B. The present study has given independent support to this order and also suggested that neither W nor B lies between E and F, and that T does not lie between W and E, the probable order being TWFEB. If the hypothesis is correct the initial mimic of P. sycorax could have been produced by a single crossover, and f. anceus, the modern mimic, evolved by subsequent minor modifications. In our investigations into the genetics of Batesian mimicry we have found that the polymorphism is controlled by a multiple allelic series at one locus or by a super-gene in P. dardanus, P. memnon and P. polytes (unpublished). This contrasts with the situation in Mullerian mimicry. In order to see whether the evolution of such super-genes is a general phenomenon or whether in other genera of Lepidoptera, and perhaps in other orders, complex epistatic interactions have been developed rather than super-genes, it will be necessary to extend this investigation outside the genus Papilio. We have already started to study Pseudacraea eurytus from Africa, which seems a particularly suitable species to use since the Batesian relationships of its forms are well documented.