Variation of Temperature in the British Isles, &c. 



83 



when the planetar}' effect reaches its maximum. Drawing then the line 

 AC corresponding to a maximum on June 21, we know that the sea 

 effect cannot exceed BC in magnitude, and that the effect which is 

 independent of the sea, cannot be less than AC. 



Fig. i is a scale drawing representing this analysis of the tem- 

 perature oscillation for Kew into two parts. It will be noticed in 

 Table II that the sea temperature reaches its maximum at the Shet- 

 lands, at Scilly, and at Yarmouth on from August 10 to 13. Taking 

 August 12 as a mean date, the sea effect reaches its maximum 20 

 days after the maximum of the first-order curve at Kew ; the line BC 

 is therefore ruled at 20 to AB. AC, being unknown, is dotted at an 

 angle of 32, corresponding to 32 days, the lag of the time of maximum 

 at Kew behind the solstice (June 21). It follows from the measure- 

 ments of the diagram that the sea effect at Kew cannot exceed 8 0< 3 F., 

 and the original effect, apart from sea, cannot be less than 5*3 F. 



The mean amplitude of the sea variation is nearly 8 D F., so that if 

 the lag of the seasons at Kew is wholly due to the sea, the whole 

 variation of temperature of the sea is superposed upon the initial 

 variation. This would no doubt be an exaggeration, and the initial 

 variation at Kesv must be greater than the 5'3 F. shown in the 

 diagram. 



A corresponding diagram drawn for Scilly from the air and sea 

 temperatures there (fig. ii)- represents nearly the whole temperature 



FIG. ii. 



