262 



.Mr. \V. Hrcnnand. 



were obtained by taking the inverse of these times for the clx 

 action. 



15. The table is only a first approximation; yet I have much 

 greater confidence in the values, than in those given by anyone ol> 

 vation, the table itself being deduced from a very large number 

 experiments. 



Sir H. Roscoe believes (Bakerian Lecture, 1865) that he brought 

 the errors due to matching shades to within 2 per cent, correct ; and 

 in graduating strips, the mean error was found by him not to exceed 

 1 per cent, of the measured intensity. I am not satisfied that my 

 separate observations were always so closely accurate in the matching 

 of shades. I employed my daughters independently, to match shades, 

 and compared them with my own reading, and found that the 

 readings sometimes differed more than 2 per cent. 



The photographic paper employed, varied somewhat in tint ; ths 

 exposed to the candle being a little redder than that exposed to tt 

 Sun and sky ; the same intensity in the darkening was sought 

 every case. 1 suppose the difference in tint to have been due to 

 heat of the candle. 



16. The effect of the sky observed, was that due to the effect 

 each elemental area of it multiplied by the sine of the angle between 

 that elemental area and the normal to the plane of exposure, these 

 infinitesimal effects being summed throughout the visible sky within 

 90 of the Sun. 



The chemical action of the sky (i.e., of the portion of it thus in- 

 cluded) is seen to be half that of the Sun at 45 altitude ; and at 

 altitudes of the Sun below 13, where little more than half the sky is 

 included in each observation, to be greater than that of the Sun. 



17. I found the chemical action of the Sun, exactly the same fo 

 the same altitude, at all seasons of the year and at all hours of 

 day, as far as the experiments went at Dacca, and I find in Somer 

 shire the same chemical action of the Sun at the same altitude as 

 Dacca. I have not been able to get exactly the same candle 



I used at Dacca ; and a difference in the composition of the stear 

 might possibly cause a small difference in the results, but I belie 

 not one of much importance. 



[The observations in Table J below in the postscript show that 

 difference is absolutely nil. "27th October. 1890.] 



In the ' Phil. Trans.,' 1867, pp. 558562, Roscoe says that for eqi 

 altitudes of the Sun the chemical intensities are equal ; and he 

 sumes " that the same " relation between the Sun's altitude and chei 

 cal intensity holds good at Kew, Heidelberg, and Para." These result 

 of Roscoe are confirmed by my observations ; he obtained them onlj 

 by "averaging" numerous observations taken at Kew, and assumii 

 that the effects of cloud, <fec., in the long run were self-destractivt 



