82 On the lest Means for conducting meteorological 



our observations relate to the wliole of that space; which he 

 would therefore divide into J'dur equrd parts, to each of which a 

 &et of observations should distinctly relate. I suppose, in order 

 to obtain far better comparison, the heat of the day, the cold of 

 the night, and the two intervening spaces of middle temperature, 

 as also the duration of other jjhcenomena, with reference to these 

 four periods. He deprecates the objertion of making the ob- 

 servations by this means too laborious, though it is clear, that only 

 persons of fixed or sedentary habits could comply with such a 

 rule; whichy'or that reason it would not be advisable to make 

 a general one. 



With regard to the variations of the barometer and thermo- 

 meter, he thinks that, for want of knowing the elevation of the 

 place above the sea in the former case, and its annual mean 

 ternpcratme in tiie latter, we are often at a loss duly to appre- 

 ciate the absolute measures, usually given of the height of each 

 in the scale : he would therefore have the variations ail put down 

 relatively ; that is, the barometer at such a time so many tenths 

 plus or minus, with reference to a fixed mean height ; and the 

 temperature in like manner so many degrees plus or minus, 

 with reference to a given vican temperature. This method 

 would accord perfectly well with the mode of expressing the 

 variation by curves, of which I gave a specimen, with a view to a 

 particular object many years since, in the Philosophical Maga- 

 zine*; and which I have since applied more extensively iu the 

 second volume of the Climate of London. But, before it can 

 be brought into successful practice for anv given station, we 

 must haveybr that station a sufficient series of observations with 

 good instruments to enable us to fix pretty accurately its local 

 mean, as to both instruments; which being obtained, such plus 

 and minus results would speak a very intelligible language, and 

 in the case of curves being used with a given scale, an universal 

 one. 



The writer of the letter proceeds to observe, that the mean 

 temperature of a year, taken as a whole, has a certain influence 

 on its productiv'.ness in an agricultural sense; but that the ir- 

 regular distribution of the heat, even in a warm year, may ren- 

 der it a barren one. To pursue the inquiry into this interesting 

 jjart of the subject, he views the day whether in winter or sum- 

 mer as v.nrm or cold, vcith reference to the mean heat of the 

 decade in which it stands ; a method which may sometimes give 

 more luarni days in the winter than in the summer half year. 



In order to the establis'.nnent of a common mode of gradua- 

 tion in meteorological instrunicms, he proposes that decimal 

 divisions of the scale be adopted in all cases. As he emplovb 



* Vol. vii. p. 365, &c. 



Six's 



