18S2 ] H. F. Blanford — Some fiirther results of f lie sun -thermometer. 73 



the figures discussed were the actual readings of the sun-thermometers, 

 readings which notoriously depend, not only on the intensity of the sun, 

 but also on the temperature of the air ; and it has been shewn by Koppen 

 and others, that there is a cyclical variation of air temperature, of the 

 opposite character to that disclosed in the curve, resulting from the regis- 

 ters of insolation temperatures. Hence it is at least probable that, the 

 deduction of the air temperatures, and tlie discussion of the residual excess 

 of temperature due to the solar action would have resulted in a curve of 

 the same ij^Q, and of still greater amplitude. 



Since this paper was published, I have attempted to carrj^ on the 

 comparison of the insolation temperatures, from year to year, by a rough 

 and ready method ; but as I am now convinced, one of very precarious 

 validity. In the first place, all sun -thermometers are compared before 

 being issued, with a common standard, by actual exposure to the sun, side 

 by side, for 30 or 40 days, and their registers are corrected for the differ- 

 ences thus determined. All readings are recorded as excess temperatures 

 (above those of the maximum thermometer in the shade), and in order to 

 avoid the tedious process of picking out days of comparative clearness, I 

 have taken simplj;^ tlie highest difference recorded at each station in each 

 month, and the average of all these monthly maxima, as representing the 

 solar intensity for the year. This method is, however, open to many objec- 

 tions, which I need not here specify ; and I have therefore now reverted to 

 my former method, (with one essential improvement), as the only one 

 which is calculated to yield any trustworthy information, on the question 

 of the supposed variation of the solar heat. 



In the present paper, which is to be regarded only as a first instalment, 

 I have taken the registers of eight stations, representing a great variety of 

 climates, and which fulfil the three essential conditions, that the register of 

 each station is that of the same instrument throughout ; that it is exposed 

 in the same manner, and also at the same ^Iucq. Those of one and the 

 same station are therefore as rigorously comparable in consecutive years, 

 as can be ensured by the ordinary arrangements of our observations. The 

 selected readings are those of days, on which the average estimated cloud 

 at 10 A. M. and 4 p. m. did not exceed one-fifth of the sky expanse ; and 

 the figures compared, are those of the excess temperature, sliewn by de- 

 ducting the self-registered maximum shade temperature, for each day, from 

 the reading of the maximum black-bulb thermometer in vacuo,* on the 

 same day. It is unnecessary to give these first results in eoctenso. As an 

 example of the data thus obtained for one year at a single station, I 

 reproduce the following, which is a fair specimen of the whole : 



* Except in the case of Vizagapatam, where the thermometer is not enelosed in an 

 exhauiitcd Lube. 



