82 MR ROBERT COCKBURN MOSSMAN ON 



Stokes sunshine recorder, which occupies a good exposure at my meteorological station 

 in the south side of Edinburgh. The hourly values have been tabulated for the six 

 years ending with July 1896, the means given in the Table being for this period. 

 Looking at the seasonal values, it will be seen that about four per cent, more sunshine 

 is recorded after noon than before it, except in winter, when the afternoon hours are 

 sunnier than the forenoon by nearly ten per cent. There is little doubt that the 

 relatively greater clearness of the afternoons in winter is due to the prevalence of fog 

 and haze during the morning hours. It will be observed that there is a well-marked 

 seasonal swing in the hour characterised by the greatest amount of sunshine, which 

 approximates closely to the time of highest mean temperature. Attention may also be 

 drawn to the slow rate at which the sky clears in summer, compared with other seasons 

 of the year. Thus in April, the mean amount of sunshine for the hours ending 7 a.m. 

 and 11 a.m. is 37 hours and 13'1 hours, respectively, while in June the corresponding 

 values are 9 '2 hours and 10*4 hours. This is probably due to the condensation 

 accompanying the strong ascending currents so prevalent during summer. 



In Table XL VI. the number of days with different percentages of sunshine is 

 shown for the six years ending with July 1896. It will be seen from the maximum 

 values that on practically cloudless days in summer at least ten per cent, of the possible 

 sunshine is lost, owing to haze at the horizon ; while in winter the amount so lost is 

 about 25 per cent. Days with from 1 to 10 per cent, of the possible sunshine are the 

 most frequent at all seasons of the year, sunless days excepted. The latter are at a 

 maximum in winter when no sunshine is recorded in 42 per cent, of the cases. 



Rainband Observations. 



Observations of the thickness of the rainband in the spectrum of sunlight have been 

 made three or four times a day since August 1887. The hours of observation were 

 9 a.m., noon, 3 p.m., and 6 p.m., the latter observation being dispensed with in the 

 winter owing to lack of sunlight. The instrument employed was a direct vision 

 spectroscope, which was pointed to the N. W. at an angle of from 40° to 50°. The scale 

 was an arbitrary one, ranging from to 6. The rainband was compared with the lines 

 B, b, and F, to which values corresponding to 1, 2, and 3 were given. The following 

 are the means for the ten years ending July 1896 : — 



Jan. 



Feb. 



Mar. 



Apr. 



May. 



June. 



July. 



Ang. 



Sept. 



Oct. 



Nov. 



Dec. 



Year. 



1-14 



0-92 



114 



0-98 



1-15 



118 



1-28 



1-27 



110 



114 



1-12 



1-12 



1-13 



There is little doubt that the rainband spectroscope is a valuable auxiliary to the 

 ordinary instruments for forecasting weather. The following Table gives certain 

 particulars for the days on which rainband observations were made during the three 

 years 1888-90:— 



