"Range of Atmospheric Temperature. 



587 



perhaps conclude that, taking the whole year, the lunar variation presents 

 the appearance of a curve with a double period superposed upon one of a 

 single period. The range of this variation is not, however, very great. 

 Series (A) is represented in Pig. 1. 



Fig. 1 denotes the lunar variation for the whole year. 

 Fig. 2 denotes the lunar variation for the summer months. 

 Fig. 3 denotes the lunar variation for the winter months. 



D. Semiannual Lunar Variation. 

 10. If we now make use of the lunations corresponding to the six 

 winter months (October to March), employing for this purpose lunations 

 1-2, 10-15, 22-27, 34-40, 47-52, 59-64, 71-77, 84-89, 96-101, 109- 

 114, 121-126, 133-139, 146-151, 158-163, 170-176, 183-188, 195- 

 200, 208-213, 220-225, 232-237, 245-250, 257-259, we obtain the fol- 

 lowing result : — 



Phase of lunation... (0) (1) (2) (3) (4) (5) (6) (7) 



Value of range 11-28 ii"4i ii'32 10-82 10-44 10-43 io'79 ii"09 (1*) 



But before making use of these numbers, we must apply to them the fol- 

 lowing small correction : — 



The winter lunations represent observations of which new moon cor- 

 responds, we may say, on an average to the beginning of the various 

 winter months, while full moon corresponds to the middle and (7) to a 

 few days before the end. It is possible, however, that the sum of the 

 various new-moon observations for any six winter months (inasmuch as 

 they occur at dates preceding those of the corresponding full -moon ob- 

 servations, or observations for other phases) may be alfected differently 

 from the latter by the annual variation which is indicated in Table I. 



The correction applicable on this account has been obtained in the fol- 

 lowing manner : — The mean monthly values of Table I. have been em- 

 bodied in a curve, and from this curve has been tabulated the values of 

 temperature-range corresponding to the eight divisions of each of the six 

 winter months. 



By this means it has been found that the following correction is ap- 

 plicable to the numbers of (D) : — 



Phase of lunation... (0) (1) (2) (3) (4) (5) (6) (7) 

 Correction appH- 1 



cable to winter I — "lo —-04 00 -|--o6 -l-'oS -|--o6 00 —•04 (E) 

 months. J 



