In 1925 a debate erupted in the correspondence columns of the British Medical Journal concerning the effectiveness of eating raw pancreas as a treatment for diabetes. Enthusiasts were predominantly general practitioners , who claimed success for the therapy on the basis of their clinical impressions. Their detractors were laboratory‐oriented ‘biochemist‐physicians,’ who considered that their own experiments demonstrated that raw pancreas therapy was ineffective. The biochemist‐physicians consistently dismissed the GPs' observations as inadequately ‘controlled’. They did not define the meaning of ‘control’ (...) in this context, although it clearly did not have the term's present‐day meaning of a trial employing an untreated comparison group of patients. Rather, the physician‐biochemists' ‘properly controlled’ experiments involved careful regulation of their patients' diet and other environmental factors, and evaluation of the therapy's success through biochemical, rather than just clinical, criteria. However, my analysis suggests that these factors alone are inadequate to account for the biochemist‐physicians' dismissal of the GPs' work as ‘uncontrolled’. I suggest that the biochemist‐physicians were deliberately exploiting the powerful rhetorical connotations of the term ‘control’. Ultimately, they implied that only a trial which they themselves had conducted could be deemed ‘adequately controlled’. (shrink)