sackie or related virus infections other seasonal factors or combinations of factors must be responsible, and these may be viral, non-viral, or both. In this case the role of Coxsackie virus would appear to be that of an initiating factor, producing subclinical islet cell damage in patients in whom diabetes is subsequently precipitated by other seasonal factors. The results of our antibody studies throw little light on this problem as neither the present study nor our previous results indicate when Coxsackie virus infection occurred in relation to the onset of diabetes.If viruses do induce diabetes in man, the question arises as to what proportion of the total incidence of diabetes is associated with Coxsackie virus infection. There are at present few indications that maturity-onset diabetes is associated with viral infection, but in our previous investigations (Gamble et al., 1969) we found higher Coxsackie B4 virus antibody titres in non-insulin-dependent diabetics of acute onset than in controls. The numbers were too small for a significant difference to be found but the difference was of the same order as that found in the insulin-dependent diabetics and this type of patient merits further study. Of insulin-dependent diabetes, our results suggest that a substantial proportion may be associated with Coxsackie B virus infection and, if this is so, a search for a viral aetiology in the remainder would clearly be indicated. The numbers studied to date have been too small to show whether other members of the Coxsackie B groups are implicated but it is unlikely that they could account for many cases without it being apparent from our results. Other picornaviruses are obvious candidates, particularly in cases occurring in autumn, in whom we found no excess of Coxsackie virus antibodies.Mumps virus is not, however, a member of the picorna group and the possibility that other groups of virus may be involved should not be overlooked, particularly in younger children.We gratefully acknowledge the help we have received from the many physicians who have provided information and specimens from their patients for this study. Our 'ournal, 1973, 4, 262-264 Summary Feeding dietary fibre in the form of bran induced changes in bile salt metabolism in five people with intact gall bladders. There was evidence of reduced dehydroxylation of bile salts; the proportion of deoxycholate conjugates in bile was reduced and the transfer of radioactivity from labelled taurocholate to deoxycholate was decreased. These findings, which were independent of changes in intestinal transit rate, imply that bran reduced the degradation of bile salts by colonic bacteria. This property of bran accords with recent theories that fibre-depleted diets favour the degradation of bile salts in the colon. These findings may be relevant to the aetiology of large bowel cancer.