66 



Therinoineters can be made, of course, with bulbs longer or shorter 

 than the one described. We adopted the length of 6 inches placed 3 

 inches below the surface, as in our experience that represents a layer 

 of soil in Avhich most of the roots of the cotton plants are contained. 

 We expect to distribute a number of these instruments through the 

 State [South Carolina] and have records kept for us near signal- 

 service stations in our typical soils — a method wdiich could hardly 

 have been arranged with the old form. The instrument is mounted 

 on a neat metal backing, and is made by H. J. Green, of New York. 

 It cost $10 without packing or express charges. The great trouble 

 about the instrument is the danger in transportation of having the 

 index get doAvn in the mercury column. For this reason it has to be 

 transported in a box on gimbals to swing freely w^ithin a larger box, 

 so that it will always remain upright. We had such a box made, 

 capable of carrying eight or ten instruments, for $5. 



From experiments at Houghton Farm (Agr. Sci., Vol. II, p. 50) 

 F. E. Emory finds that the thermoelectric couple and galvanometer, 

 as used by Becquerel, consumed much time and was frequently use- 

 less owing to atmospheric electricity and ground currents. Short- 

 stem graduated thermometers, with bulbs immersed in oil and fas- 

 tened at the lower end of a light w^ooden rod, gave good results when 

 the temperature at the thermometer was not warmer than that of the 

 overlying soil or the atmosphere ; otherwise a circulation of air takes 

 place. He finds that the telethermometer, giving a continuous rec- 

 ord, answers his needs, but we know nothing of its accuracy. 



T. C. Mendenhall (1885) describes a modified form of thermometer 

 for observing the temperature of the soil at any depth, which he calls 

 the " differential resistance thermometer." Experiments w^ith this 

 instrument at Washington, D. C, have shown him that it is miich 

 less troublesome than Becquerel's electric method, but still too trou- 

 blesome to be recommended to any but persons accustomed to electric 

 measurements. Mendenhall's arrangement consists essentially in util- 

 izing the varying resistance of a platinum wire which extends from 

 the upper end of an ordinary mercurial thermometer down into its 

 bulb. The total resistance diminishes as the temperature rises and 

 allows the current to flow through less platinum but more mercury. 

 The changes in the resistance are measured by the galvanometer, but 

 he hopes to substitute for this the telephone, which wdll make the 

 apparatus more convenient for general use. 



[It is desirable that this or Becquerel's method or the thermo- 

 phone be provided in connection with the ordinary buried long- 

 stem thermometers in order that by an annual or more frequent set 

 of comparative observations the changes in the zero point of ordi- 

 nary thermometers may be detected. — C. A.] 



