372 Dr BREWSTER on the Register of the Thermometer kept 



upper curve of the Plate. Each point of it is the mean of 

 365 observations. Its resemblance and general parallelism to 

 that of 1824, cannot fail to strike the reader, and proves how 

 nearly these observations have conducted us to the form of the 

 daily curve. 



The intermediate curve, which is laid down from the last co- 

 lumn of Table VI., and is the mean of the two curves, is nearly 

 free from the very slight inequalities in the afternoon branch of 

 both curves, and may be considered as representing with great 

 accuracy the mean annual daily curve for the Latitude of Leith, 

 and at the level of the sea. 



In order to observe the variation in the form of the daily 

 curves for different seasons, I have given in Plate XV. and XVI. 

 their projections for every month in 1824 and 1825; and in 

 Plate XVII. the mean of the monthly curves in 1824 and 1825. 

 These curves obviously divide themselves into three groupes or 

 classes, viz. 1. The Curves of High Temperature, such as those of 

 June, July, August and September ; 2. The Curves of Low Tem- 

 perature, such as those of November, December, January, February, 

 March ; and, 3. The Curves of Moderate Temperature, such as 

 those of April, May, and October. 



But though the curves thus group themselves into three va- 

 rieties of temperature, yet it is obvious that those of April and 

 May have the same form as those of the summer months, while 

 the curve of October resembles the flat curves of the winter 

 months. In order, therefore, to obtain a type of the daily curves 

 of summer and winter, I have joined October to the winter 

 months, and April and May to the summer ones, as in the fol- 

 lowing Table for 1824 and 1825*. 



* It is probable that the Royal Society of Edinburgh will publish a series of 

 Plates representing the daily curve for each day of 1824 and 1825. 



