Photometric Observations of the Sun and Sky. 263 



Boscoe supposes that a marked difference which he found in intensity 

 between spring and autumn might be due to a difference in trans- 

 parency. I can only explain some of Roscoe's results by supposing 

 that the sky was not perfectly clear at the time of the observations. 

 Indeed, from the description, many of Roscoe's observations would 

 appear to have measured the effects of cloudiness rather than of Sun 

 and sky. I have no anomalies in the results of my observation 

 except such as I think I may fairly attribute to cloud or haze. My 

 experience in England is that it requires months of watching to 

 catch a sky that will give results similar to those I obtained regularly 

 in Dacca during the cold season. 



In the ' Phil. Trans.,' 1867, p. 559, Roscoe finds (by the same 

 method of ''averaging") that "the relation between the Sun's 

 altitude and the chemical intensity of total daylight is graphically 

 represented by a right line." And in the ' Phil. Trans.,' 1870, p. 315, 

 Roscoe and Thorpe say that the relation between altitude and total 

 chemical intensity, for altitudes above 10, is seen to be accurately 



ipresented by a straight line. 



Table B indicates, and Table G below proves, that the straight line is 



ly a first approximation to the truth. The calculation from my 

 Table B of the chemical action of the whole visible sky (and Sun) on 

 the horizontal plane can be effected, as shown farther on in the 



esent paper. 



18. Various observations had led me to expect that the chemical 

 action of the sky at the same moment was different in different parts 

 of it. To investigate this suspicion, I designed an instrument which 

 I call the Mitrailleuse Actinometer (fig. 2) ; 1 place in the President's 



thand photographs of two of these instruments. 

 I mount a number of similar cylindric tubes in one plane in a semi- 

 circle, to the centre of which each tube is directed. One extremity 

 of each tube lies on the circumference of the circle ; the other 

 extremities lie on a concentric circle of about half the radius. In the 

 circumference of this smaller circle, is a semicircular series of holes, 

 against which a semicircular block, carrying the sensitised slip of 

 paper, is pressed by a screw. Each cylinder in the first Dacca mitrail- 

 leuse cut out of the sky a circle of 8 28' angular diameter. One of the 

 tubes near its top, carries a small plate of wood, on which stands a 

 style parallel to the tube, by means of which the particular tube can 

 be brought into a line with the Sun. By another motion the plane of 

 all the tubes can be adjusted to the plane of symmetry (or else- 

 where). 



[A vertical plane through the Sun at any time divides the visible 

 sky into two exactly similar portions. I will call it the plane of 

 symmetry]. 



19. The observations (Table C) were taken 23rd December, 1864, 



VOL. XLTX. T 



? 



