The question of wind pressures and wind velocities is a most important one in these days of great engineering problems, particularly in connection with the stability of bridges and other large structures.
Experimental determination of the constants of anemometric formulæ have recently been made both in England and this country. From results obtained in the English experiments it was concluded that the very widely used Robinson anemometer is not as satisfactory and reliable an instrument as a different form of anemometer devised by Mr. Dines. These conclusions, however, are not sustained by the American experiments, which were made by Professor C. F. Marvin, Signal Office, by means of a whirling apparatus, and under the most favorable circumstances, which yielded highly satisfactory results. Professor Marvin has lately made very careful open air comparisons of anemometers previously tested on the whirling machine, which have shown that, owing in part to the irregular and gusty character of the wind movement in the open air, taken in connection with the effects arising from the moment of inertia of the cups, and the length of the arms of the anemometer, the constants determined by whirling machine methods need slight corrections and alterations to conform to the altered conditions of exposure of the instruments in the open air. This latter problem is now being experimentally studied at the Signal Office, and final results will soon be worked out.
Professor Langley has also made very elaborate observations of pressures on plane and other surfaces inclined to the normal, which it is believed will prove important contributions to this question, but the results have not yet been published. It is important in this connection to note experiments made by Cooper on the Frith of Forth Bridge, where a surface of 24 square metres, during a high wind, experienced a maximum pressure of 132 kilogrammes per square metre, while a surface of 14 square decimeters showed, under similar conditions, 200 kilogrammes per square metre, by one instrument, and 170 by another. The opinion expressed by Cooper that in general the more surface exposed to the wind, the less the pressure per unit of surface, seems reasonable, and if verified by more elaborate experiments must have an important bearing.
There are questions in connection with which even negative results are of an important character, particularly when such results are quite definite, and tend to remove one of many unknown elements from physical problems of an intricate character. In this class may be placed atmospheric electricity, with particular reference to its value in connection with the forecast of coming weather. The Signal Office, through Professor T. C. Mendenhall, a distinguished scientist peculiarly fitted for work of this character, has been able to carry out a series of observations, which have received from him careful attention, both as to the conditions under which the observations were made and in the elaboration of methods to be followed.
Professor Mendenhall also supervised the reduction of these observations, and after careful study presented a full report of the work to the National Academy of Sciences, in whose proceedings this detailed report will appear. Professor Mendenhall says, "Taking all the facts into consideration, it seems to be proved that the electrical phenomena of the atmosphere are generally local in their character. They do not promise, therefore, to be useful in weather forecasts, although a close distribution of a large number of observers over a comparatively small area would be useful in removing any doubt which may still exist as to this question." It may be added that Professor Mendenhall's conclusions bear out the opinions expressed to the speaker, in a discussion of this question, by Professor Mascart, the distinguished physicist.
It has been generally admitted that the aqueous vapor in the atmosphere plays a most important part in bringing about the formation of storms and maintaining their energy. It has been frequently commented on by the forecast officials of the Signal Service, that storms passing over the United States were in general preceded by an increase in moisture, but unfortunately little effort had been made on the part of previous investigators to determine any quantitative relation between the actual humidity and the amount of precipitation or its relation to the storm movement. It has long been regretted that the direct relations of this to other meteorological phenomena were not more fully defined. During the past year Captain James Allen, of the Signal Office, has endeavored to apply the results of his investigations and theories to the practical forecasts of storm conditions. Captain Allen has carefully studied the relations of the potential energy of the surface air, as represented by the total quantity of heat it contained, to the movement of storm centres and the extent of accompanying rain areas. In his first investigations the potential energy per cubic foot was estimated as follows: Supposing the air to have been originally 32° and the moisture in it as water at 32°, the total quantity of heat applied to reduce to the state of observation will be A = (t-32)/6 + Q in which A is total heat per unit volume; t is the temperature of the air, Q the total heat of vapor, and the specific heat of air at constant volume being taken as one-sixth (.168). From Regnault's formula we have Q = 1091.7 + .305(t-32).
For the mechanical equivalent we have J = 772A. If we divide J by the pressure estimated in pounds per square foot, it will give the height through which the pressure can be lifted if all the heat is spent in work by expanding the air.
An approximate expression for the upward velocity V may be obtained from Torrecelli's theorem from which we have V2 = 2gh, h in this case being the height through which the pressure would be lifted if all the heat is spent in work. The theory has been that the storm centre will move over that section of the country where V is the greatest, and that the time of occurrence and amount of rain have a relation of conformity to the changes in Q and its actual amount.
Auxiliary charts were also made showing for each station the following values of Q:
1st. Highest Q not followed by rain in 24 hours.