Variation of Temperature at the Cape. 73 



temperature at the Cape deduced from observations extending from 

 1841 to 1870 inclusive. I have carefully examined the zero-points 

 of all the thermometers which have been employed in this series of 

 observations. I have then deduced the rate of change of these ther- 

 mometers, from a comparison of the index-errors thus found and 

 those given originally or obtained in 1852 by Sir Thomas Maclear, 

 when he compared the principal thermometers at the Observatory 

 with the readings of a standard "Regnault" which had been sent 

 out to the Observatory for that purpose by you. These indications 

 of change have been carefully checked by all the comparisons made, 

 at different times and for different purposes, of these thermometers 

 inter se and with others which still remain at the Observatory. 

 From the agreement of the different results thus checked, I have no 

 doubt upon my own mind of the systematic character and sensible 

 amount of the increase of readings of thermometers with age thus 

 indicated. In some cases the change appears to amount to as 

 much as o, 05 F. per annum. From these results I have deduced the 

 index-errors of the different thermometers for the different periods, 

 and applied these corrections throughout. I have also corrected the 

 mean results of the five observations made daily since 1847 in order 

 to deduce the true daily mean. 



The results thus reduced on a general system, and extending over 

 thirty years, appeared likely to afford information respecting any 

 connexion which might exist between the mean temperature and the 

 frequency of solar spots. I have therefore constructed the curves of 

 variation of mean annual temperature, and the inverse curve of solar- 

 spot frequency for comparison. The latter curve has been founded 

 upon Wolf's observations. 



The observations of temperature from 1841 to 1851 inclusive were 

 made in the original Meteorological Observatory, which was burnt 

 down in 1852, March 11. 



The observations from 1852, April 24, to 1858, August 31, were 

 made in a wooden shed erected for the purpose on the site of the old 

 Observatory. 



The observations from 1858, August 31st, to the present time 

 have been made in the crib before the south-west window of the 

 Transit-Circle Room. 



These changes are so far unfortunate, that there is clearly a change 

 of mean temperature arising from the different circumstances of expo- 

 sure. I have therefore referred each set of observations to the mean 

 temperature deduced from all the observations made under the same 

 circumstances of exposure. The deviations of the mean temperature 

 for each year from the mean of the whole period of similar exposure 

 are then laid down as ordinates on the scale of one division of the 

 ruled paper to 0°*05 F. To smooth down the irregularities, I have 

 joined the points thus laid down, and bisected the lines thus joining 

 these points whenever the corresponding mean temperatures were 

 deduced from a full year's observations. In other cases the tempe- 

 ratures corresponding to the deficient months have been supplied 

 from the adjoining years, and the resulting mean temperature allowed 



