112 SCIENCE OF COMMON THINGS. 



Effect of moonlight on animals and vegetables. Equinoctial storm. 



more frequently about four days before full moon, and 

 less frequently about four or five days before new 

 moon, than at other parts of the month ; but this can- 

 not be considered as an established fact : in other 

 respects the changes of the moon cannot be shown to 

 have influenced in any way the production of rain. 



713 Does the bright moonlight in any ivay hasten the putrefaction of 

 animal or vegetable substances ? 



It is generally supposed to do so ; but the fact is, 

 that- on bright, clear nights, when the moon shines 

 brilliantly, dew is more freely deposited on these sub- 

 stances than 'at other times, and in this way putrefac- 

 tion may be accelerated. With this the moon has no 

 connexion. 



713 Is there any foundation for the belief that the appearance of the 

 aurora borealis is followed by a change in tlie weather? 



Meteorological registers conclusively show T that there 

 is no such connexion, and that the appearance of the 

 aurora is as often followed by fair weather as by foul. 



714 Is there any truth in the traditional notion that a long and violent 

 storm usually accompanies the period of the equinoxes ? 



The examination of weather-records for sixty-four 

 years shows that no particular day can be pointed out 

 in the month of September, (when the " equinoctial 

 storm " is said to occur) upon which there ever was, or 

 ever will be, a so-called equinoctial storm. The fact, 

 however, should not be concealed, that taking the ave- 

 rage of the five days embracing the equinox for the 

 period above stated, the amount of rain is greater than 

 for any other five days, by three per cent., throughout 

 the month. 



715 Is there any reason for believing that cold and warm seasons alter- 

 nate? 



Meteorological records, kept for eighty years at the 

 observatory of Greenwich, England, seem to show that 

 groups of warm years alternate with cold ones in such 

 a way as to render it most probable that the mean an- 

 nual temperatures rise and fall in a series of curves, 

 corresponding to periods of about fourteen years. 



