Winifred E. Brenchley 



'2V3 



Tlie close associatiou of liouis ol .sun.shine with iiiaxiiiiuin temperature 

 is to be expected ; nevertheless, taking into consideration that the mere 

 total of weekly sunshine ignores its distribution between the several 

 days of the week and the hours of its incidence on individual davs, the 



Oct. 6 Nov. 

 1915 



Hours 



3 Dec. 1 Jan. 5 Feb. 2 Mar. 1 Apr. 5 May 3 June 7 July 5 Aug. 2 Sept. 60ct.4 Nov.1 Dec. 6 Jan. 3 

 1916 1917 



Maxiiniuii 



Minimum 



Fig. 1. Average weekly temperatures in roof water-culture house, 

 Sept. 29th, 1915— Jan. 10th, 1917. 



Oct. 6 Nov. 2 Dec. 7 Jan.A Feb.1 Mar. 7 Apr. 4 May 2 June6 July 4 Aug.l Sept. 5 Oct. 10 Nov. 7 Dec. 5 Jan. 2 

 1916 1917 



Fig. 2. Total hours of bright sunshine per week, Sept. 28th, 1915— Jan. 9th, 1917. 



value obtained, -7560, is remarkably high. The correlation between 

 mean maximum and mean minimum temperatures is also high, as is 

 naturally the case ; but the correlation between bright sunshine and mean 

 minimum temperature would be higher, owing to the close association 



15—2 



