396 ANNUAL OF SCIENTIFIC DISCOVERT. 



claimed that the following points must be regarded as established meteoi'olog- 

 ical facts : 1. That the temperature before the first quarter is lower than that 

 of the second day after it. 2. That this fall and rise prevails most in the win- 

 ter months, and in the month of May. 3. That a reciprocity of action takes 

 place between corresponding days of the moon's age. Thus, whilst it was 

 found, both at Dublin and Greenwich, that for twenty-one consecutive years 

 the mean temperature rose at the first quarter in more instances than it fell, 

 it fell at the last quarter in more instances than it rose; and in the only two 

 years in which a fall occurred instead of a rise at the first quarter, there was 

 a rise instead of a fall at the last quarter. Between new and full moon this 

 reciprocity of action was still more apparent. Here, for the same series of 

 years, there was a fall in thirteen years after new moon, and a rise in thir- 

 teen years after full moon ; and in five out of the eight instances in which a 

 rise occurred instead of a fall at new moon, a fall instead of a rise took place 

 at full moon. Also a like principle appeared to hold good in individual 

 months. For example, in twenty-one consecutive Januarys a fall occurred 

 in seventeen at new moon, while a rise took place in sixteen at full moon. 

 The action thus apparent at different periods of the lunation was shown 

 clearly in curves of temperature of each day of the moon's age. A curve 

 of ten years' mean temperature at Greenwich, for 1837 1846, was exhibited 

 in juxtaposition with one sent to the Dublin meeting, which was also formed 

 of ten years' mean temperature, at the latter station, for 18471856. At first 

 and last quarters the curves corresponded in a most remarkable manner at 

 both stations. At new and full moon they alternated; the fall in the Dublin 

 curve being at the new moon, and the rise at full moon ; in the Greenwich 

 curve the rise at new moon, and the fall at full moon. Leaving the consid- 

 eration of daily mean, temperatures, on extracting the maxima and minima 

 mean temperatures for the month, it was found that more maxima occurred 

 after first quarters than before; the proportion of maxima to minima, on the 

 second day after that phase, being more than 2'1 both at Dublin and Green- 

 wich, for the respective periods of twenty -two and forty -three years. The 

 twenty-four highest and lowest maxima and minima in the month at Green- 

 wich were then taken for the same forty-three years, forty-eight per cent, 

 found to occur at first quarter, and minima only before the day of the change. 

 Similar results were obtained from the highest and lowest mean tempera- 

 tures at Dublin, and at Toronto from 1843 to 18 18. Another point elicited 

 during the progress of the inquiry, was the recurrence of high and low tem- 

 peratures on the same days of the lunation. Taking first the maxima and 

 minima mean temperatures for the month during twenty years at Green- 

 wich from 1837 to 1856 the whole number found recurring on corre- 

 sponding days (many of them three and four times in each period of twelve 

 lunations), amounted to 236, averaging about twelve for each year, or half 

 the maxima and minima for the month. To illustrate this recurrence of 

 high and low temperatures, several years were selected, which presented the 

 strongest evidence of system. Thus, in the two consecutive years com- 

 mencing November 1847 and ending October 1848, maxima and minima 

 occurred: In 1847, twice on the third day before new moon; twice on the 

 second day before new moon; three times on the day after new moon; twice 

 on the third day after new moon; three times on the second day before full 

 moon; twice on the third day after full moon. In 1818, three times on the 

 day of new moon; twice on the day after new moon; three times on the scc- 

 onl day before full moon; twice on the day before full moon; twice on the 



