In the course of the medical work of this institution, there have been found, not infrequently, patients whose temperatures are too low to be registered by the ordinary clinical thermometer. Observations with crude substitutes in one or two cases showed that surprisingly low temperatures were reached. In order that these cases could be properly recorded, special thermometers were procured, reading to a very low point, and an order issued that in all cases in which the temperature was too low for the ordinary instrument, one of the special thermometers should be obtained and a physician on the service should control the readings by personal observation.
Various errors may enter into such work, as follows :
1. Inaccurate instruments. The instruments used are standardized only between 95 and 110 and had the ordinary certificate of accuracy. The lower readings were not tested. One of the instruments used, however, has been sent