SIUMMARYSurface waters in 14 selected sites were examined for the presence of salmonella using modified Moore's swabs. The sites included an upland impounding reservoir, 3 rivers and 10 streams within Lancashire and Cheshire, selected because of their accessibility to farm livestock. Salmonellas were isolated from 22 out of the 57 swabs examined representing 10 sites. The probable source of pollution was shown to be sewage or farm effluent and an examination of sites over a wider area may be expected to produce similar results.The significance of these findings is discussed in relation to the epidemiology of salmonella infections and the possible disinfection of effluent discharges.

Improvements in performance of a lysine conductance medium for the detection of salmonellas were achieved from a study of the effects of its various components. When sodium biselenite was included as an inhibitor for non‐salmonella organisms conductance signals were depressed. The inclusion of sodium chloride reduced this toxicity and improved conductance responses. Increasing the pH to pH 7.0 prevented the medium becoming too acidic and inhibitory to salmonellas. The new medium detected 70‐0% of salmonella‐positive animal protein samples.

No significant differences in electrical conductance parameters were observed when multiple portions of animal proteins were examined for salmonella on three Malthus 2000 analysers. No differences were noted in the isolation rates from conductance tubes on the three analysers. Results also indicates the importance of examining multiple samples due to uneven distribution of organisms in a dry matrix.

No abstract

SUMMARYTwo trials were carried out by 9 and 47 laboratories to evaluate the reproducibility and sensitivity of a standard technique for the isolation of salmonellas from animal feeding stuffs.The trials involved the examinations of test samples containing known numbers of salmonellas by a buffered peptone water pre-enrichment technique. Statistical analysis of the results of the trials showed that the method was capable of detecting salmonellas in most probable numbers as low as 2·3 per 100 g and that no significant differences in performance were observed between laboratories.The standard technique described is recommended for the examination of animal feed and conforms to international recommendations.