50.2 Mr. J. J. Waterston on Solm* Radiation. 



With r = 20° the thickness would be double this amount, and so 

 on. Thus the presumed extra atmospheric value of r being 67° 

 gives 001 24 inch thickness melted per minute. 



From June to December the amount may be expected to vary 

 T Vth. corresponding to alteration of sun's distance. In Her- 

 schel s ( Meteorology ' the probable thickness is stated to be 

 •0109 inch. 



§ 8. If the law indicated by straight lines on the chart is true, 

 it would require extremely accurate observations to give the extra 

 atmospheric constant of solar radiation with precision*. Prom a 

 singfe observation made in Bombay some years ago, I am disposed 

 to believe it may exceed 67° considerably. 



§ 9. The mode of approaching the law of absorption is as fol- 

 lows : — Project the values of r as ordinates to the secants of 

 zenith distances as abscissae : the resulting curve is evidently 

 hyperbolic in character. If it is the conic hyperbola, the reci- 

 procals of the ordinates laid off to the same abscissse should 

 range in a straight line. The obvious plan is therefore to lay 

 off the reciprocals of r in this way, and see how far their range 

 agrees with the straight ; and if it differs, the character of the 

 divergence might lead us to the true function that expresses 

 the natural law, if it was not very complicated, and if the 

 condition of the atmosphere did not vary so rapidly as to 

 obscure it. 



The observations, though taken under unfavourable conditions, 

 favour the simple hyperbolaf. 



§ 10. It will be remarked, on inspecting the Chart, that the 

 value of r at the same altitude of the sun diminishes with the 

 declination as the season advances. If continuous observations 

 were possible for a few hours each day, when the altitude of the 

 sun ranged between 15° and 45°, we might expect to see the 

 projection of the equalized observations range each day in a 

 different line ; but these lines ought all to converge on nearly 

 the same point in the ordinate at the zero of the secant scale, 

 if the law holds good. 



* That is in this climate, where r is comparatively small, and the trend 

 of consecutive observations slopes from a part of the chart, where the scale 

 of r is large, to where it is small. In a tropical station, such as Bombay, 

 where even during the winter months the value of r for the same altitude 

 is double what it is at our summer solstice, the trend must incline much 

 less to the axis, and consequently errors of observation will be but slightly 

 magnified in the value given to R, the extra atmospheric value of r. 



t The equation of which is r cosec altitude =k a constant, hence cosec 



altitude oc - ; and thus the reciprocals of r laid off as ordinates to the secants 

 r 



of zenith distances may be expected to range in a straight line. See Chart, 



fi 2 3. 



