584 



C. MICHIE SMITH ON 



remarkably well with each other, so that the diurnal curve for the forenoon 

 may be considered to represent the true curve very fairly; but this cannot be 

 said for the part for the afternoon hours, for on four days out of the nine the 

 readings have almost no value for determining the normal variation. I have 

 thought it well, however, to give a curve (No. 2, PI. CXLVI.) showing the mean 

 of all the observations, indicating the number of disturbed readings in each case by 

 a corresponding number of marks on the small circle surrounding the point mark- 

 ing the mean. The curve is drawn from the following means of the readings: — 



Hours, .... 



7 



8 



9 



10 



11 



12 



13 



14 



15 



16 



17 



18 



19 



20 



21 



22 



Readings, .... 



44 



49 



64 



71 



84 



91 



95 



99 



102 



98 



114 



98 



82 



95 



75 



59 



The other curve (No. 1) shows the means of the two hourly observations on the 

 4th, 5th, 6th, 7th, and 9th, which are nearly free from the effects of mists. 

 The following table gives the actual observations on these days : — 



Hours, 







7 



8 



9 



10 



11 



12 



[13 



14 



15 j 16 



17 



18 

 64 



19 



20 



59 



21 



22 



23 



24 



January 4, 

 „ 5, 

 „ 6, 

 „ 7. 

 >> 8, 

 „ 9, 



Mean 

 omitting the 









52 





52 





75 





68 





62 





51 





72 





125 





173 



170 



132 



115 



91 



49 



63 



73 



67 











60 



65 



60 



74 



78 



84 



78 



68 



74 



70 



75 



40 



62 



66 



74 



[91] 



76 



75 



92 



83 



79 



64 



54 



52 



[85] '[95] 



58 







44 



42 



72 



106 



97 



115 



[142] 



[110] 



[87] 



[84] 



[124] 



[106] 



74 





90 



59 



69 



62 



47 



49 



65 



77 



96 



111 



109 



129 



125 



120 



98 



60 



62 



67 



56 









'8th, 



1 





55 





67 





93 





108 



93 





69 















N. B. — Those readings within brackets were taken in mist. 



If we take this curve as most nearly the true one, we see that there is a 

 well-marked maximum about 14 h M.M.T., and that the curve is fairly sym- 

 metrical on the two sides of this maximum for at least four hours ; and there is 

 some evidence, though it is by no means conclusive, that the maximum normal 

 readings are obtained at the time of maximum temperature. The observations 

 made do not fix the time of minimum satisfactorily ; but probably there are 

 two minima — one early in the morning and another about 19 h , with a small 

 secondary maximum two or three hours afterwards. This diurnal curve differs 

 entirely from that for Madras, which shows a minimum between 9 h and 10 h , and 

 a maximum about 18 h . The average readings at Dodabetta are also much 

 higher than those at Madras ; and, as is to be expected, fine-weather readings 

 in Madras are much more variable than those on Dodabetta. 



The readings made during the prevalence of mists were begun simply with 

 the object of testing the assertions frequently made, that while all clouds are 

 positively electrified, they are surrounded by a zone negatively electrified. So 



