6 E. Loomis—Resulis from an examination of the 
area of high barometer there is a downward movement, not gen- 
erally violent, but steady and of long continuance. The result 
of such a downward movement must generally be a consid- 
erable fall of the thermometer at the suriace of the earth. The 
atmosphere generally loses heat most rapidly at the top, in con- 
sequence of radiation into space, and it receives its heat chiefly 
at the bottom from the effect of the sun’s rays upon the sur- 
face of the earth, so that at an elevation of a few miles a pound 
of air usually contains a less amount of heat than a pound of 
air near the earth’s surface. Hence we conclude that within 
an area of high barometer the thermometer must generally 
stand below its mean height. 
Monthly minima of temperature. 
In order to test the preceding conclusion, I have taken the 
monthly minima of temperature as observed at New Haven for the 
years 1872, 3, and 4, and have compared them with the state ofthe 
barometer and such other circumstances as could be supposed 
to influence the result. These observations of temperature 
were all made with a self-registering thermometer. The precise 
hour of minimum is therefore unknown, but it occurred some- 
time during the night preceding the date given in column first. 
In the following table, column first shows the date of mini- 
mum temperature for each month; column second shows the 
lowest temperature recorded ; column third shows the direction 
of the wise as recorded at the morning observation, which in 
summer was generally 6 A. M., and in winter about 7 A. M.; 
column fourth shows the degree of cloudiness at the same hour; 
column fifth shows the height of the barometer at New Haven; 
column sixth shows the position of the nearest center of high 
barometer as indicated by the U.S. weather maps; and col- 
umn seventh shows the highest isobar represented on the map 
or the corresponding date. 
on comparing these observations, we perceive that in all 
but five cases the sky at the date mentioned was cloudless, and 
in only two cases did the clouds cover as much as one-half of 
the sky. It is probable that in each of these cases, during a 
eae of the preceding night the sky had been entirely cloud- 
ess. We must conclude, then, that these monthly minima of 
temperature were in part the result of radiation ; but this cause 
alone will not account for the very low temperature observed, 
perature of the same months on cloudless nights. We fin 
maining cases an,area of high barometer existed at the west 
and was rapidly approaching New Haven, although the barom- 
