494 



Dr. G. Neumayer 07i the Lunar 



[May 16, 



The curves derived from the results of this Table are shown on Plate XI., 

 with the exception of the semiannual and annual curves which may be 

 studied on Plate X. 



Glancing at the various curves thus resulting, we are first struck by the 

 great conformity of some of them, whilst others present irregularities 

 apparently quite irreconcileable with what we feel inclined to adopt as the 

 law. There is, however, in all cases manifested a progressive change, 

 evidently depending on the moon's hour-angle in the first instance, calhng 

 for a rigorous examination. The semiannual curves of the lunar-diurnal 

 variation of atmospheric pressure may be taken as representing the prin- 

 cipal types of the various monthly curves. During the sun's absence from 

 the hemisphere (in our case, when the sun's declination is north), from 

 April to September, the lunar variation reaches its maximum at about 23^ lo''^ 

 or 45™ prior to the moon's upper transit, its minimum value occurring .at 

 and a secondary one at 2^ with a range of 0*00653 inch. The curves 

 for the single months appertaining to this semiannual period exhibit, 

 generally speaking, the same characteristics, though somewhat irregular, and 

 showing, in some instances, deviations of considerable extent ; so, for in- 

 stance, the curves for August and September. The summer semiannual 

 curve (while the sun's declination is south) exhibits an essentially different 

 character, there being no strongly expressed maximum noticeable, whilst a 

 decided minimum occurs at 0'' 30"^ or 30'" past the moon's upper passage, 

 the maximal pressure taking place at and a secondary one between IS'^ 

 and 19^. The amplitude of oscillation amounts to 0*00432 inch. But in 

 this period of the year we notice a great difference in the lunar-diurnal 

 variation of the barometer, when we examine the single months somewhat 

 more closely ; thus, for instance, the curve for the month of December shows 

 such characteristics as to cause it to be more like the curves for the winter 

 period, and, on the other hand, we perceive that the curve for the 

 month of November is exactly of the opposite character as that for De- 

 cember. The remaining four months show more or less irregularity, and 

 make a greater or smaller approach towards the general type for the class 

 under consideration. 



Although there is undoubtedly in all of these cases strong evidence of an 

 influence of the moon on our atmosphere, I could not rest satisfied, con- 

 sidering that this evidence is seemingly of a somewhat conflicting nature. 

 As already explained, the monthly values have, for the purpose of further 

 inquiry, been combined into quarters, and the results for these quarters 

 were again united in mean values, arranging the two quarters in which 

 epochs of solstice occur, and the remaining two, comprising the equinoxes, 

 respectively, in two groups. Thus we obtain six monthly mean A'alues of 

 the lunar-diurnal variation for the '^solstitial and the equinoctial quarters." 

 I was prompted to adopt this course because of the great similarity of the 

 curves for December and June in the one case, and of March and Septem- 

 ber in the other, though in by far less a degree. This similarity may best 



