On Methods of measuring Skin Temperature

For the measurement of the temperature of exposed skin surfaces, we have reached the conclusion that the most accurate instrument is a radiation thermopile suitably screened from the effects of wind. As the thermopile does not actually touch the skin it does not interfere with the cutaneous circulation; neither does it hinder the heat-loss from the skin surface. Readings are simply and rapidly made; with the Moll thermopile and the Cambridge unipivot galvanometer used by us a reading can be taken in 4–6 sec. The readings are not influenced by wind if the thermopile is screened, and as the temperatures are measured as differences from that of a blackened surface of known temperature, change of air temperature does not upset the results. The apparatus is easily portable. By rolling a mercury thermometer over a small area of skin the temperature can be estimated with but very slight error. This method is a reliable one for use where other apparatus is not available; but, it has the disadvantage that each observation takes from 1 to 1½ min. Accurate estimates cannot be obtained by this method in the special cases when the skin temperature is subject to rapid variation. Under steady conditions accurate estimates can also be made by means of a clinical thermometer protected by a cork wedge, but each observation takes from 4 to 5 min. Of the types of thermo-junction tested by us the simple Lewis type of junction attached and protected by a strip of surgical tape appears to give the most accurate results. The average error, taking the thermopile values as standard, was only 0·5–0·8° C, and the thermo-junction temperatures were consistently higher than the thermopile values by about this amount. The Benedict type of junction also gives results which compare reasonably well with those obtained from radiation measurements. In our observations the average error with this type of junction was 0·7–1·3° C, and the error usually had the effect of giving too high a value for the thermo-j unction temperature. Other observers, however, have measured temperatures with this type of junction which appear to be somewhat low, and it appears possible that the temperature of the observer's hand may have influenced the readings obtained. Forms of apparatus in which the thermo-junction and its leads are exposed to the effects of the air are liable to give very erroneous estimates of skin temperature. In our own comparisons the skin temperature was underestimated by 3–3·5° C. when the Strauss thermo-junction was used; and examples have been drawn from the literature which suggest that, with other apparatus in which the junction was exposed, the temperature was underestimated. If such forms of apparatus are used for measuring skin temperatures beneath the clothing, they may yield fairly reliable results, but in exposed positions considerable errors may arise. This liability to error is particularly serious where there is a wide range of air temperature and of air movement. Skin temperatures can be measured with considerable accuracy by means of an electrical resistance thermometer.