1 94 Miscellanies. 



Another question is, whether a change of weather is more liable to 

 happen on the four principal days of the lunar phases than on com- 

 mon days. But it must first be decided what is meant by the term 

 change of weather. This term should, the author thinks, be limited 

 to a change from clear weather to rain, or from rain to clear weather, 

 and not be understood to include, as some meteorologists make it, 

 all changes, such as that from calm to windy, or from clear to cloudy, 

 hci As the author accepts it, the weather must have been steady 

 during two days at least; that is, that the weather has been clear, or 

 that it has rained more or less during two consecutive days. For 

 example, a week has passed without rain ; it rains on the eighth day, 

 and on the ninth the weather is again fine. In this case, according 

 to the author's definition, there is no change of weather. So, also, 

 if it has rained during five successive days, the sixth and the seventh 

 must be clear, in order to constitute a change of weather. This may 

 be arbitrary, but at least it is not vague, and if practised it will pre- 

 vent, in the balancing of calculations, any leaning to a favorite hy- 

 pothesis. To avoid another error, into which some have fallen, the 

 author marks no change as occurring on lunar phases but those which 

 take place on the very day, and never those which may happen on 

 the evening before or on the next day. 



With these precautions he finds that during the 34 years or 12,419 

 days, there have been 1458 changes of weather. Of this number, 

 105 have taken place at the epoch of the two principal lunar phases, 

 viz. 54 at the new moon, and 51 at the full moon. Now the whole 

 number of principal phases during the 34 years is 840, therefore, 



As 12419 : 840: 11458 : 98.6 the number of changes which 

 would have taken place at new and full moon, had these lunar phases 

 had no more than the share of common days, but instead of which, 

 the number was 105. 



Of the 54 changes at new moon, 32 were from rain to fine weath- 

 er, and 22 from fine weather to rain. Of the 51 at full moon 31 

 were from rain to clear, and 20 from clear to rain. Thus at the new 

 and full moon the changes to fine weather are to those to rain as 63 

 to 42. 



Having thus proved that the epoch of new and full moon are not 

 absolutely without some effect on the weather, the author examined 

 whether this effect was confined to those very days, or extended to 

 the day following. On " the days following the new and full moon, 

 there were 129 changes, instead of 98.6 which would have been the 

 number, had these shared the proportion only of common days. 



