3 2 4 
Journal of Agricultural Research 
Vol. V, No. 8 
tube sufficient to ruin the bulb for accurate measurement of temperature. 
After the bulbs had been in use a short time they would produce an 
electromotive force as if they were primary or secondary cells. 
In another type of thermometer bulb the resistance wire was inclosed 
in a thin-walled small-bore copper tube, which was filled with Wood’s 
metal to exclude moisture from the tube and to render the thermometer 
more sensitive by increasing the conduction of heat to the wire. This 
bulb did not prove satisfactory because, though the thermal conductance 
from the water to the wire may have been improved, the sensitiveness of 
the thermometer was not, the mass of metal apparently causing a lag in 
response to temperature change. Furthermore, the Wood’s metal 
apparently did not completely exclude moisture, for ultimately the wire 
in this thermometer also became short-circuited with the metal. Another 
serious objection was the possibility that the resistance wire might be 
stretched by the unequal expansion of the metal in which it was embedded. 
The bulb which was finally used with utmost satisfaction was con¬ 
structed in accordance with the specifications of the one developed by 
Dickinson and Mueller (n, 12, 13) in connection with investigations on 
calorimetry at the United States Bureau of Standards, which was designed 
especially for use in determining the temperature at a definite point of 
liquid flowing in a tube in a continuous-flow calorimeter. The bulbs 
were designed especially to combine constancy, freedom from lag, and 
intimate contact with the entire water flow. The platinum resistance 
wire was wound on a thin strip of mica, and this coil, laid between two 
similar mica strips, was inclosed in a flat sheath of thin silver which 
pressed the mica insulating strips firmly against the resistance wire, thus 
affording opportunity for rapid conduction of heat between the case and 
the wire. The silver case terminated at the top in a tube which was sealed 
to a glass tube, on the end of which was a bulb containing phosphorus 
pentoxid, the purpose of which was to exclude moisture from the space 
in which t£& resistance wire was inclosed. The flat part of this bulb, 
which was about 10 cm. long, 10 mm. wide, and 1 mm. thick, and con¬ 
tained the sensitive part of the thermometer, was inserted in a brass tube 
with a constricted channel, like that for the final heater described on page 
319, so that the sensitive portion of the thermometer was surrounded by a 
space about 1 mm. across; and water flowing through this space was thus 
brought into intimate contact with the thermometer, which very rapidly 
acquired the temperature of the water and responded instantly to changes 
in temperature and integrated stream lines of temperature, if any existed. 
The two thermometers, one in the ingoing and the other in the outgoing 
water, had exactly the same resistance, about 25.5 ohms at 20° C., and 
the same coefficient of change of resistance with change in temperature, 
about 0.0039 P er degree for the range of temperature in which they would 
be used, the resistance change of each thermometer being 0.1 ohm per 
