ABNORMAL AT TORONTO. Ill 



monthly and annual mean changes on the average of the six years, 

 would not, without greater labour than the occasion warranted, give 

 the annual mean changes from G a.m. to 6 a.m. for the separate 

 years. On this account, for the years 1854 to 1859, the annual means 

 of the diurnal change in the temperature and barometer have been 

 given for the interval commencing at 2 p.m.* 



The normals to which reference is made in the temperature-tables, 

 are deduced from the table of twenty-four-hour daily means given by 

 General Sabine An his paperf on the Periodic and Non-peinodie 

 Variations of Temperature at Toronto, by applying the diurnal 

 variations given (though with a contrary sign,) in the same paper. 



The approximate normals of reference for the other elements are 

 simply the monthly means at each of the six observation hours, derived 

 from an average of several years. 



The normals thus computed are tabulated and kept as standards 

 to which the observed values of the elements are referred ; the 

 abnormal variations with their proper signs being entered in the daily 

 register side by side with the observed values. 



TABLES I. TO VII., ON TEMPERATURE 



From table I we see that the average extent of an abnormal digres- 

 sion of temperature, without regard to sign, and irrespective of the 

 hour and season, was 6°.5 on the average of nine years, and that the 

 digression in different years never differed more than 0*^.6 from this 

 average. 



In table II., which gives the abnormal variations without regard to 

 sign for the different months, double weight is given to the earlier 

 series in computing the means from 1854 to 1862. Double weight is 

 also given to the earlier series wherever in subsequent tables the results 

 of the earlier series are combined with those of the years 1860 to 1862. 



The progression from month to month, though it shews that the 

 digressions are decidedly larger in the winter than in the summer 

 months, is not perfectly continuous. If the monthly means be 

 collected in quarterly groups we have 6°. 1 as the average digression in 

 spring, 4°. 9 in summer, 5°.8 in autumn, and 9°.l in winter. 



In table III. we have for each series separately, as well as for the 

 two combined, the yearly and half-yearly mean abnormal variations at 



* TlirouRhout both series observations have hoen made on Sundays at 6 a.m. and at 2 P M., 

 so that no break on account of Sundays or holidays has occurred, 

 t Philosophical Transactions for 1853, pp. loi to 159, and pp. 145, 14G. 



