134 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 
THE PROBABLE ERROR OF A BACTERIOLOGICAL 
ANALYSIS. 
By Jas. JoHNSTONE, D.Sc. 
There has been a tendency of late years to depreciate 
the value of bacteriological analyses in relation to questions 
of shell-fish pollution, and to attach greater significance to 
conclusions based upon topographical and epidemiological 
investigation. This tendency is due partly to the lack of any 
really convincing and generally accepted method of analysis ; 
partly to the reaction against an exaggerated estimate of the 
utility of bacteriological tests, and partly to the fact that 
analysts have not usually attempted to.make estimates of the 
magnitude of the error to which their conclusions are always 
subject. This error may be a rather large one, and neglect of 
it has led to apparently anomalous results which have done 
much to create the tendency to which I allude. 
But one must recognise that topographical data are very 
difficult to apply, for it is most difficult to foresee exceptional 
conditions and to reckon on the probability with which these 
may be expected to occur. Epidemiological evidence is still 
more difficult to interpret even when it has been collected with 
scrupulous care—and the local public health services cannot 
be said to be organised yet with such thoroughness as to 
justify us in expecting this scrupulous care. ‘There are several 
well-known examples—Bulstrode’s investigation of the Win- 
chester and Emsworth outbreaks, and Hamer’s work on the 
Bethnal Green epidemic—where the enquiry was a really fine 
piece of research. But the majority of cases of illness attribut- — 
able to shell-fish pollutions are sporadic, and it must be 
exceedingly difficult to trace the cause to shell-fish alone, 
eliminating all other possible explanations. Naturally, a local 
authority will tend to make such an hypothesis as will remove 
