[CALLBNDAR & m'leod] OBSERVATIONS OF SOIL TEMPERATURES 33 



A few samples of the records obtained have been reproduced. The 

 scale of the record may readil}' be made much more open than in the 

 case of ordinary recording instruments wilhout loss of accuracy or sensi- 

 tiveness. We have observed curious differences in the minute detail of 

 several of the curves, which could not have been detected with less sensi- 

 tive and accurate instruments, which appear to be connected in some way 

 with the type of weather prevalent at the time in each case. We hope to 

 be able to find some meaning in these minute characteristics, as it appears 

 that they may be of use in forecasting. 



With regax'dto these records, it will be observed that the co-ordinates 

 on the chart are rectangular, a point of great convenience in many ways. 

 The scale of the record may also be very readily changed at any moment. 

 For instance, we prefer to use a scale of 5 mm. to the degree Fahr. for 

 the atmospheric temperature. This can be changed in a minute to one of 

 two or three centimetres to a degree if it is required to obtain a record of 

 one of the buried thermometers, which are otherwise exactly similar to 

 the air thermometer except that their temperature changes are much 

 slower and of smaller rang-e. 



