Appendix A. METEOROLOGICAL INSTRUMENTS. 359 



the white cotton : the means of a number of observations taken by- 

 each approximated closely, but the difference between individual 

 observations often amounted to 2°. 



Observations again indicative of the radiation from grass, whether 

 dewed or dry, are not strictly comparable ; not only does the power 

 of radiation vary with the species, but much more with the luxu- 

 riance and length of the blades, with the situation, whether on a 

 plane surface or raised, and with the subjacent soil. Of the great 

 effect of the soil I had frequent instances ; similar tufts of the same 

 species of grass radiating more powerfully on the dry sandy bed of 

 the Soane, than on the alluvium on its banks ; the exposure being 

 equal in both instances. Experiments for the surface-temperature of 

 the soil itself, are least satisfactory of any : — adjoining localities being 

 no less affected by the nature, than by the state of disintegration 

 of the surface, and by the amount of vegetation in proximity to the 

 instrument. 



The power of the sun's rays in India is so considerable, and 

 protracted through so long a period of the day, that I did not find 

 the temperature of springs, or of running water, even of large deep 

 rivers, so constant as was to be expected. 



The temperature of the earth was taken by sinking a brass tube a 

 yard long in the soil. 



A thermometer with the bulb blackened affords the only means 

 the traveller can generally compass, of measuring the power of the 

 siui's rays. It should be screened or put in a blackened box, or laid 

 on black wool. 



A good Photometer being still a desideratum, I had recourse to 

 the old wedge of coloured glass, of an uniform neutral tint, the 

 distance between whose extremes, or between transparency and 

 total opacity, was one foot. A moveable arm carrying a brass plate 

 with a slit and a vernier, enables the observer to read off at the 

 vanishing point of the sun's limb, to one five-hundredth of an inch. 

 I generally took the mean of five readings as one, and the mean of 

 five of these again I regarded as one observation ; but I place little 

 dependence upon the results. The causes of error are quite obvious. 

 As far as the effects of the sun's light on vegetation are concerned, 

 I am inclined to think that it is of more importance to register 



