30 REPORT OF THE SECRETARY. 



Correct general information of this kind would furnish material for 

 the preparation of an interesting map, showing at a glance the 

 regions where tornadoes annually prevail, and those where they oc- 

 cur only occasionally, or not at all. In meteorological investiga- 

 tion we always make a step in advance when the region to which 

 phenorhena are limited can be accurately defined. 



Since the latter part of October, a series of daily observations has 

 been made for the Institution, by Mr. Force, on the temperature of 

 the hydrant water in Washington. The water flows some six or 

 seven miles through iron pipes, under ground, before reaching the 

 place of delivery, and acquires, approximately, the temperature of 

 the soil at the depth at which the pipes are laid. The observations 

 were made at a hydrant .out of doors, about 7 o'clock every morning, 

 the water being allowed to flow two or three minutes before noting 

 the temperature. Vfhen the curves for the water and the air are 

 plotted on the same diagram and the same scale, they exhibit to the 

 eye, in a very striking manner, the extreme fluctuations of the tem- 

 perature of the air, compared with that of the ground, at the depth 

 of only two or three feet. They also show the slowness with which 

 variations of temperature penetrate the earth. The change pro- 

 duced by a decided increase or decrease of temperature being gene- 

 rally indicated by the water two or three days after it occurred in 

 the open air. 



Among the contributions relative to meteorology which have been 

 received at the Institution, is a series of tri-daily observations made 

 at Brunswick, Maine, by the late Professor Parker Cleaveland, from 

 "1808 to 1859. The records include observations on the ordi- 

 nary thermometer, the maximum and minimum thermometer, the 

 barometer, rain-gauge, wind-vane, &c. It is proposed to publish these 

 in full, as a part of the second volume of the "Meteorological Re- 

 sults," but for this purpose it is desirable that the means for months, 

 years, and the whole period, should be calculated, and this work is 

 now in progress, and will probably be finished before the printer 

 will call for the copy. The publication of these observations is in 

 accordance with the plan adopted to print in full a number of long 

 series of observations such as those we have already published, viz: 

 for Providence, Rhode Island, and Washington, Arkansas. They 

 will be of much interest for solving various meteorological problems, 

 such as the recurrence of particular phenomena, changes of seasons, 



