70 DISCUSSION OF T II H I N K L T K N (' !: OF THE MOON 



If AVC give Aveight to the annual means according to the number of observations, 

 they Avould be; one for the first and second year, three-fourths for the third year, 

 one and three-fourths for the next year, and tAvo for the last year: a general exam- 

 ination, hoAvever, shoAVS that, owing to the disturbing effect of the progressive 

 change, the monthly means are very nearly of equal A'alue, derived either from the 

 In-hourly or the hourly series. It Avill also be shown in the sequel that the lunar 

 diurnal variation is nearly the same in. the summer and Avinter seasons; the means 

 of Table V. and the final means of Table VIII. have therefore been adopted with- 

 out reference to combinations or AA'cights. 



A comparison of the values of Table VIII. among themselves shows them to be 

 very irregular, although derived from many thousand obserA'ations ; afiA'e year scries 

 of observations seems barely sufficient to exhibit a tolerably regular progression. 

 In the folloAving table tAvo groups have been formed, one of results from three 

 years, 1840 to 1843, comprising 8,797 observations, the other from the remaining 

 tAvo years comprising 13,248 observations. From these it appears that the lunar 

 diurnal variation during these tAvo periods exhibits the same general character. 



Before proceeding to the analysis of the final result of Table VIII. the separate 

 results have been combined into summer and winter groups ; the first group com- 

 prising the months from April to September, the second group the months from 

 October to March. 



Table IX. exhibits the lunar diurnal variation of the horizontal force during the 

 summer and winter seasons. 



