68 MR ROBERT COC'KBURN MOSSMAN ON 



The following are the plus corrections for each series : — 



Years. 



-Tan. 



Feb. 



Mar. 



Apr. 



May. 



June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 









• 







o 



= 























o 



O 



1785-1793 



1-6 



33 



4-6 



6-6 



6-6 



6-8 



6-3 



6-1 



5-1 



3-9 



2-0 



1-6 



February 

 1793-1798 



2-0 



33 



4-6 



3-9 



38 



4-1 



49 



5-1 



4-3 



2-9 



1-6 



1-8 



May 

 1798-1816 



1*8 



2-0 



3-2 



4-5 



5-5 



6-0 



6-8 



6-8 



5-7 



3-7 



1-9 



1-3 



A register was kept in Edinburgh by Mr George Waterston from 1799 to 1850 

 (see Table XL). The hours of observation were 8 a.m., 2 p.m., and 10 p.m. They have 

 been utilised from 1806 to 1820, and were corrected by comparing them with Adie's 

 mean temperatures for the ten years 1821-30. The following are the monthly correc- 

 tions obtained after smoothing the curve : — 



Jan. 



Feb. 



Mar. 



Apr. 



May. 



June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 



-2°-2 



-l°-8 



-r-5 



-l°-6 



-2°0 



-2°-2 



-2° 5 



-2°-3 



- 2 0, 2 



-2°0 



-2°-2 



-2°-4 



From June 1781 to December 1784 no observations are known to have been made 

 in Edinburgh. It was therefore necessary to interpolate from the records of contiguous 

 stations. A register was kept at Branxholm from 1775 to 1783, the results being given 

 in vol. I. of the Trans. Roy. Soc. Edin. A comparison of the means there given with 

 Hoy's corrected values from 1775 to June 1781 gave the following smoothed corrections 

 which were applied to the observations made from June 1781 to December 1783, 

 as follows : — 



Jan. 



Feb. 



Mar. 



Apr. 



May. 



June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 



+ 3°-9 



+ 3°-0 



+ 2°'7 



+ 2°-l 



+ l°-8 



+ l°-5 



+ 2 0, 2 



+ 2 0> 4 



+ l°-6 



+ 2°-2 



+ 3° -4 



+ 3°-6 



A register was kept at Glendoich from May 1783 to 1817, from which means (see 

 Table XII.) have been calculated from May 1783 to 1794, the corrections to Edinburgh 

 mean temperature being obtained by a comparison with the Edinburgh Advertiser 

 means from 1788 to 1793, thus: — 



Jan. Feb. Mar. Apr. May. June. July. Aug. Sept. Oct. Nov. Dec. 



-0°-7 -0°6 -0°-6 -0°-6 -0°-6 -0"-9 - 1°0 - 1 0> 5 - 1°-4 - 1°-3 - 0"-9 -0°-6 



The corrected means from all these sources having been obtained, final means were 

 calculated from them. For example, from 1799 to 1804 the means given in Table VII. are 

 the average of the corrected means deduced from Adie's and the Edinburgh Advertiser 

 8 a.m. and 8 p.m. observations, along with the Barnton observations, all being brought 

 to the mean of the maximum and minimum by the corrections already given. The 

 results are remarkably accordant in the majority of cases. Forbes' adopted temperatures 

 were utilised as a check from 1805 to 1820. The Kinfauns Castle record was further 

 brought to the mean of the Edinburgh record for the period 1813 to 1821 by a comparison 



