ANY system of diagnosis depends upon the detection, measurement and assimilation of properties which are characteristic of the pathological process involved. These may be either physical, chemical or biological. Attention has been drawn, recently, to the finding that malignant tumours in the breast are hotter than benign tumours, and it has been suggested that this difference may be sufficiently characteristic to permit a pre-operative diagnosis. The accurate detection and mensuration of this temperature elevation is therefore fundamental if such a criterion is to be used in diagnosis.As it is impractical, and virtually impossible, to obtain an accurate assessment of thermal conditions at tumour level, the thermal conditions at the skin surface overlying the tumour have been used with the contralateral uninvolved breast as a comparative control.To measure the thermal gradients at the skin surface one may either use heat conducted to the instrument as with a thermocouple, or heat radiated to the instrument. The latter method depends upon the Stephan-Boltzman law which states that the quantity of infra-red emitted by a surface varies directly with the fourth power of its temperature in degrees Absolute and the emissivity of the surface. A full account of the physics involved in infra-red thermometry is given by Barnes (1964). Such infra-red thermometers may be either hand held whereby a small area of temperature is measured, or incorporated into an optical scanning system whereby the thermal contours of a large surface area are measured. Such measurements are usually recorded in the form of a grey scale, and if this is calibrated by including a series of black bodies of known temperature with the scan, the temperature of any spot on the recording or thermograph can be interpreted by measuring the greyness with a densitometer.Lawson (1957) was the first to detect an elevated skin temperature overlying a malignant tumour in the breast when compared to a symmetrically identical area in the opposite breast. He did not find such an elevation if the tumour were benign.Lloyd Williams et al. (1961) repeated Lawson's observations on a series of 100 women with a lump in one breast. They found that 54 out of 57 carcinomas were associated with a temperature elevation of over 1°C. and 23 out of 24 cysts and adenomas were associated with a thermal difference of less than 1°C. They used a hand held infra-red detector, the Schwarz thermopile and a comparative assessment of the thermal contours of the skin between right and left breasts required the charting of a series of individual measurements on a proforma. By using such * Present address: Mulago Hospital, Kampala, Uganda.