NO. II 



BODY RADIATION ALDRICH 



ACCURACY OF SKIX TEMPERATURE AIEASUREMENTS 



In the study of body radiation above referred to, skin and clothing 

 temperatures were measured by a special thermoelement device sug- 

 gested by Dr. C. G. Abbot. For convenience I quote the following 

 illustrated description of the instrument from my previous publi- 

 cation : 



For the direct measurement of skin and clothing temperatures, a special device 

 was prepared with the help of Mr. Kramer, the Observatory mechanician, and 

 embodying Dr. Abbot's suggestions. The device is shown in Figure i.' It con- 

 sists of a specially mounted copper-nickel thermoelement of fine drawn wire. 

 A frame of German silver is bent as shown in the figure and fastened in a 

 wooden handle, JV. Two silk threads are stretched to form a cross between the 



Fig. I.— Thermoelement device for measuring surface temperatures. 



P — Spring steel projection. 

 6^— Silk thread. 

 T — Thermoelement. 



F — Fibre rings. 

 S — German silver frame. 

 IV — Wooden handle. 



four spring-wire posts, p. The thermoelement wires are fastened symmetrically 

 to these silk threads with the junction straddling the lengthwise thread. The 

 wires lead out through fibre rings, F, and through the wooden handle. The 

 copper wire (see fig. 2) leads through a switch to a sensitive type Leeds and 

 Northrup D'Arsonval galvanometer and thence to the constant temperature 

 junction in a stirred kerosene bath as shown in Figure 3. The Cu-Ni wires are 

 sufficiently long so that all desired positions can be reached without moving 

 the constant temperature bath. Holding the device by the wooden handle, one 

 presses lightly the four prongs of spring wire p upon the surface whose tem- 

 perature is desired. This places the junction in excellent contact with the sur- 

 face. There is no backing to the junction save a single silk thread, and thus no 

 possibility of heat piling up and causing too high temperatures. For about 4 cm. 

 on each side of the junction, the wire also touches the surface and assumes 

 the surface temperature, thus eliminating error due to cooling of the junction 

 by conduction along the wires. 



' The figure numbers of the original publication have here been changed to 

 accord with the arrangement in this paper. 



