Probable error of any single representation of the Philadel- 

 phia values =±0-25'Jiv= ±0-000009 parts of the horizontal force 

 = ±0-000038 in absolute measure. 



Investigation of the horizontal force in reference to the lunar places. 



The following process of reduction has been adopted. After 

 marking the days of the full and new moon and also the days 

 preceding and following, the daily means of the horizontal force 

 readings were taken (already corrected for difference of tempera- 

 ture and progressive change) for the place of any disturbed obser- 

 vation, the monthly normal, belonging to the respective hour, 

 was substituted before taking the daily mean. All accidental 

 omissions in the record of the hourly or bihourly series were 

 supplied by the hourly i 

 obtained, " ^ 



thus 



I independent of the solar diurnal variation. The 

 monthly normal was next compared with each daily mean and 

 the differences (normal minus mean) were tabulated. 



A positive sign signifies a greater, a negative sign a less force 

 than the normal value. As the results deduced from a smgle 

 year are yet too much affected by the incidental irregularities of 

 tlie observations, the collective results from the five year serie3 

 (1840-45) are herewith presented. 



Table No. U.—Infuence of the lunar phases on the horizontal force. 



The J 





+2-4 

 39 



-Paruofthehf 



In ab.ol. mo.,.. 



Onedaya'fter " " 



OnedlySter « " 

 Difference for new-fnll moon. 



oooor46 



-000023 



umber of observations from which any > 



i average i 



lOve six means were ueuuucu lo wy^^ ■^^■^ ^. 



error, in scale divisions, of any one of the results is ±0-7 nearly 



From the Makerstoun observations Broun found for theyeai 



1S43-46 a minimum at the time of the full moon :md a max 



■ iiu at the time of the new moon; K-'' ^■■'- --' ' '-^'^'i'^ <^» 



' vatlons, between 1843-46 found * - ' ^ 



e. It must be remarked, ho we V - ± 



- il found that the signs were rever- ^^l 



. .trs that the lunar influence on tnc .. _ - ■ i^-^ 



-^ a cycle of short period. This last remar.v <.oe= i .j. ..,yo t 



t^e effect of the moon's declination and variation in di:,tunee. 



