Annual Variations for the Temperature of the Air. 



Ixxxv 



178. We may employ the monthly means in the 11th column of Table 72, for the purpose of predicting 

 the mean temperature for a coming month, the probable error of the predicted temperature for each month as 

 deduced approximately from the Table, being as follows : * — 



Jan. 



Feb. 



Mai'ch. 



April. 



May. 



June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 



2°-0 



2°-9 



1°1 



l"-9 



l°-9 



l°-7 



r-4 



r-5 



r-5 



r-1 



r-2 



3°-2 



Thus, at Makerstoun, there are equal chances that the mean temperature of any month of March will not 

 be more than 1°-1 from 39°-5. The months of March, October, and November, shew the least variation of 

 monthly mean temperature ; the months of December, January, and February shew the greatest variation. 



Table 73. — Monthly Means of the Diurnal Ranges of Temperature, as deduced from the 

 Observations of the Register Thermometers, for the Years 1843-6. 



Year. 



Jan. 



Feb. 



March. 



April. 



May. 



June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 



Mean. 



1843 



9-0 



7-8 



14-1 



15-2 



13-3 



13-6 



16-7 



20-2 



20-8 



13-6 



12-4 



8-1 



13-7 



1844 



10-9 



11-7 



13-1 



18-4 



19-4 



18-2 



18-7 



19-5 



15-9 



13-4 



8-4 



6-8 



14-5 



1845 



11-9 



121 



14-0 



20-2 



14-2 



18-7 



171 



171 



17-9 



120 



12-6 



11-0 



14-9 



1846 



8-6 



10-6 



14-7 



14-3 



18-1 



24-9 



16-6 



18-8 



19-3 



13-9 



9-0 



11-0 



150 



Mean 



10-1 



10-5 



140 



17.0 



16-2 



18-8 



17-3 



18-9 



18-2 



13-2 



10-6 



9-2 



14-5 



179. Annual Variation of the Diurnal Ra^ige of Temperature. — From the last line of Table 73, the mean of 

 the diurnal ranges of temperature was least in December, and it was greatest in June and August. It appears 

 probable, however, that when a sufficient number of years' observations is considered, the mean of the diurnal 

 ranges will be found to vary little from April till September. This result is analogous to that obtained for 

 the ranges of the mean undisturbed diurnal variations of the magnetic elements. 



The ranges of the monthly mean diurnal variations, from the hourly observations in the two years 1844-5, 

 are as follow : — 



Jan. 



5°-60 



Feb. 



7°-05 



March. 



9°-30 



April. 



14°-70 



May. 



11°80 



June. 



12°-15 



July. 

 12°-00 



Aug. 

 12°- 35 



Sept. 



12°-60 



Oct. 

 8°-20 



Nov. 

 5°-05 



Dec. 



3°-70 



These quantities indicate a result quite similar to that obtained from Table 73, though, as might be expected, 

 the ranges are considerably smaller. December has the least range, and May, June, and July have rather 

 less ranges than April, August, and September. 



Table 74. — Mean Differences of the Daily Mean Temperature from the Monthly Mean for 



each Month in the Years 1843-6. 



Year. 



Jan. 



Feb. 



March. 



April. 



May. 



June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 



Mean. 



1843 



5-45 



4-40 



4-63 



4-21 



2-08 



3-44 



1-57 



2-48 



4-42 



6-98 



4.44 



2-69 



3-90 



1844 



4-30 



4-20 



4-43 



2-62 



2-69 



3-00 



3-33 



1-62 



2-07 



4-50 



3-90 



3-20 



3-32 



1845 



6-02 



3.21 



6-08 



3-97 



2-46 



3-34 



2-16 



1-86 



3-32 



3-04 



4-69 



3-12 



3-61 



1846 



4-21 



4-20 



4-67 



2-85 



2-15 



4-47 



2-57 



2-80 



3-61 



3-87 



4-82 



4-60 



3-73 



Mean 



500 



4-00 



4-95 



3-41 



2-34 



3-56 



2-40 



2-19 



3-35 



4-62 



4-46 



3-40 



3-64 



* These numbers divided by 3 will give approximately the probable errors of the monthly means in the 11th column of Table 72 

 as the true monthly mean temperatures at Makerstoun. 



