668 STATE BOARD OF AGRICULTURE. 



taking 25" pieces of the i*e.spective wires and joining them in series of 

 three. That is, three coup)les were joined together in series. The joints 

 were first dipped in ZnCL, then soldered and finally coated with shel- 

 lac. Above the joints, which were to be buried in the soil, were placed 

 glass tubes G inches long and % inch in diameter. The opposite ends 

 of these glass tubes were made waterproof with shellac. These glass 

 tubes were employed in order to facilitate the placing of the ends of the 

 couples into the soil and also to keep the wires of the latter well insu 

 latcd. To each series of couples were joined coy>per wires 12 feet long. 

 All Ihe couples were then standardized. This was accomplished by 

 dipping one end of the series of couples in ice and Ihe other in a con- 

 stant temperature bath. It was found that they were all sensitive to 

 about 0.2° C. 



Those ends of the couples having the glass tubes were placed into the 

 soil five inches deep from the lowest end of the point or four inches 

 from the upper part of the joint. The three couples were placed at 

 different places in the soil in order to obtain a more nearly average 

 record. The temperature readings were made by dipping the long bended 

 end — as shown in the diagram — into an ice bath, connecting the two 

 12 foot wires to the galvanometer, and reading the deflection on the 

 latter. The galvanometer employed was of D'Arsonval type and was 

 of very high grade. Its needle came to rest in a few seconds. It 

 took only about 40 minuies to complete the reading of the 40 thermo- 

 couples. At the beginning of the experiment there were two readings 

 made, one in the morning at 8:00 and one in the afternoon at 3:30. 

 Later, however, the readings were made only in the afternoon. At the 

 beginning also, the readings were taken every day but after the max- 

 imum temperature was attained they were taken every other day for 

 some time and occasionally afterwards. The experiment lasted alto- 

 gether from December 2Sth to April 10th, or 103 days. The data ob 

 tained are given herewith : 



