in the Magnetic Forces of the Earth. 225 



the suppositions made the average quantity of dew, for the year 

 that must be deposited in the interval between 5 p. m. and 5 a. m. 

 to satisfy the requirements of the theory, is about r |^ of an inch, 

 and therefore about three times greater than the actual average 

 for July, as ascertained from observation ; — or rather, strictly 

 speaking, about three times greater than the average of the re- 

 sults of Professor Brocklesby's experiments, which, it is probable, 

 is somewhat higher than the actual average for this month. On 

 glancing at the curves of diurnal variation of temperature for 

 1844, (see p. 42,) it will be seen that the average fall of tempera- 

 ture about the time of sundown, varies, from one quarter of the 

 year to another, from 1£° to 3°. This difference is probably th^ 

 result of the joint action of three different causes, viz., the une- 

 qual evaporation that obtains, at this hour, on some days, the un- 

 equal deposition of dew, at the same hour, on other days, and as- 

 censional currents varying in velocity. The first and last, so far 

 as they act, will tend to make the fall of temperature at sundown 

 greater in summer than it would be from radiation alone ; and 

 the second tends to make it less than this, perhaps, during the 

 greater part of the colder half of the year. It is probable that 

 the first tendency prevails over the second, and therefore that the 

 average loss of heat, from radiation alone, is not greater than 2° 

 per hour, the average loss of heat at sundown for the entire year. 

 It is certain that it must be less than 3°. On the supposition that 

 it is 3°, the entire loss, from radiation, would be 108°, instead of 

 ~2°. and the amount of dew required in# 06. On the other hand, 

 there can be no question that the estimate I have made of the 

 actual amount of heat lost is too high, for it is the sum of the 

 entire variations of temperature of the different layers, to the 

 depth of 18 in , in the course of a day, whereas the actual loss of 

 heat during the night is equal to the sum of the actual losses of 

 the same layers in the course of the night, and the loss of each 

 layer in this interval of time will be less than its entire loss, with 

 the single exception of the layer at the surface. From these 

 considerations it appears that the average quantity of dew neces- 

 sary to reduce the average loss of temperature due to nocturnal 

 radiation down to the actual average loss, for the year, is certainly 

 materially less than in *06, and is probably less than J "03. As 

 to the quantities required for the other seasons, taking the radia- 

 tion at 72°, 15° for the average nocturnal loss of temperature in 

 the interval between the vernal and autumnal equinox, and 9° 

 f °r the same in the interval between the autumnal and vernal 

 equinox, f find that the heating effects of the dew during the 

 n *ght ; in these two periods must be, respectively, 27° and 45°, 

 and the quantities of dew in *04. and in 024. The average for 

 Jl *ty, to correspond to an average of in 024 for the six warm 

 months, must be less than in 02, and probably, judging from the 



Second Series, Vol. VIII, No. 23,— Sept., 1849. 29 



