224 Mr. J. P. Harrison Lunar Influence on Temperature. [May 4, 



In 1856 the frequent recurrence of higher temperatures about the 

 eighth or ninth day of the moon's age, led to an examination and compa- 

 rison of the mean temperatures of the third day before, and the second 

 day after first quarter of the moon, for a series of seven years at Chiswick, 

 and sixteen years at Dublin. The results showed conclusively that the 

 temperature of the second day after first quarter was higher than the 

 temperature of the third day before that phase during the years in 

 question. 



On extending the investigation to the remaining days of the lunation, 

 the maximum was found to occur, at both stations, at the period when 

 heat was first observed, and the minimum after full moon and last 

 quarter. 



The long series of mean temperatures which had been determined by 

 Mr. Glaisher for the British Meteorological Society from observations 

 taken at Greenwich between 1814 and 1856, were next arranged in tables 

 constructed for the purpose. These observations, though corrected by an 

 arbitrary rule totally irrespective of the moon, and in a measure therefore 

 eliminating influences that may have been exerted on the observed tempe- 

 ratures, appeared on the whole the best, as they were also the most exten- 

 sive printed series existing. 



The method pursued. The Tables were constructed in the following 

 manner : The mean temperatures of the days on which the moon entered 

 her four principal phases having been first inserted in columns arranged 

 at equal distances, the mean temperatures of the first, second, and third 

 days before and after each of the quarters were entered in the columns 

 adjoining on either side ; and any remaining observations in octant columns 

 midway between the quarters*. The deficiency occasionally occurring in 

 an equal number of six observations between the quarters, was supplied by 

 repeating the observation of mean temperature of the third day after, or 

 third day before the quarters, the same observation in such cases being 

 used for both those days. Thus an equal number of observations was 

 secured for twenty-eight days out of 29 '5, at all the seasons of the year, 

 a point of no little importance as regards the next process, viz., obtaining 

 true means of the temperatures of the several days. This was done in 

 the usual way, by adding together the observations of mean temperature 

 in each column, and dividing the sums by the number of lunations the 

 temperatures of which had been tabulated. 



The last operation consisted in laying down the mean line on scale-paper, 

 and marking above or below it the mean temperatures belonging to the 

 several columns on vertical lines, representing the several days of the luna- 

 tion preceding or following the four quarters. The points thus marked 



* On an average, the number of observations in each of the octant columns equals 

 half the number of observations in the other columns. Their means were not made use 

 of iu forming the curves of temperature. 



