Rotation-anemometers are those in which the wind sets in motion plane or curved metallic blades. The earliest form resembled that of Dinglinger. mentioned by Leupold in 1724, in that it used the Polish water-wheel with vertical axis, but differed essentially in that Dinglinger pre vented the rotation of the arms and measured the pressure required to keep them quiet, whereas d'Ons-en-Bray, in 1734, allowed them to rotate continuously. Since that time two essentially different varieties of the rotation-anemometer have been developed, namely (a) those of Scho ber and Woltmann, Combes, Casella, Whewell, or Biram, in all which sets of plane plates in clined to an axis are forced to revolve about it by the wind blowing in the direction of the axis. This form is much used in studies on ventilation of mines and building(. The most important meteorological application of this style is that manufactured by Richard for use at the French observing stations.
(b) The Robin s o n anemometer, brought out by Dr. Robinson in 1846, b u t suggested t o him by Edgeworth many years before. This has come into very general use by English and American meteoro logical observers, as the Robinson hemi spherieal cup ane mometer. In this instrument a ver tical spindle car ries at its upper end four horizontal arms at right angles to each other; each arm carries at its ex tremity a hollow hemispherical cup of thin sheet metal whose circular rim is in a vertical plane passing through the common vertical axis of ro tation of the spindle. The wind rotates these cups so that the convex side of each cup goes for ward. Numerous experiments have been made to determine the relation between the velocity of the wind and that of the cups. The instru ment makers have generally followed Dr. Rob inson's conclusion, that the linear motion of the centre of the cup is one-third of that of the wind ; but observation and experiment, as well as the ory, show that this cannot be true. The most intelligent and satisfactory investigation of this important subject has been carried out by Pro fessor C. F. Marvin, of the United States Weath er Bureau. Combining his results with those of European students, we must conclude that in perfectly uniform winds the general average ra tio between the velocity of the wind and that of the cups varies with the length of the arm and the size of the cups between 2.5 and 3.5. so that it is necessary to determine the ratio by aetual experiment upon each respective type of ane mometer.
Professor Marvin shows, besides, that the ratio varies according as the anemometer is exposed to a uniform wind or to one that is variable and gusty. He finds that in the latter case the ratio depends not merely upon the di mensions of the arms and cups. but especially upon the moment of inertia of the revolving sys tem; that is to say. on the mass of the cups. For gusty winds, the recorded wind velocity is always too great. This is explained by the faet that the gusts give to the revolving cups a great velocity. which they, by reason of their momentum, retain after the gust has ceased. It would seem, there fore, that rotating anemometers should be stand ardized not merely in quiet air, but also out of doors in ordinary gusty winds. By such compari
sons Professor Marvin has compiled a table. of which the following is an abstract, showing the correet wind velocity for records of anemome ters in the ordinary or average gustiness of the wind at Washington. If the observed wind ve locities are indicated on dials constructed on the assumption that the eentres of the cups move with one-third the velocity of the wind. then the corrected wind velocities are given by the follow ing table: Observations on strong winds on the summit of Mount Washington indicate that the velocities given in this table apply also to that high ele vation, so that there is no evidence that the Rob inson anemometer is appreciably influenced by changes in the density of the air ; but, of course, the wind pressures for a given velocity are smaller in proportion to the density. In order to determine the for computing wind pressure at high velocities, Marvin conducted special measurements at the summit of Mount Washington, using both large and small-pressure plates, and obtaining automatic simultaneous rev. ords on the same sheet of paper for both the pressure and the velocity. Ile finds that when the air has the standard density for 32° F. and 30 inches of pressure, the wind pressure on a plane flat surface is equal to 0.0040 pound to the square foot multiplied by the square of the velocity of the wind in miles per hour and by the area of the plate; this formula gives the pressures printed in the preceding table. For further details, see Professor Marvin's paper on wind-pressures and wind-veloeities, printed in the annual report of the chief sig nal officer of the army for 1890.) A gen eral review of the subject of anemometry is given in Abbe's; Treatise on Meteorological Ap paratus and Methods (Washington. 1887). The Robinson anemometer, as originally manufac tured by James Green, of New York, and reduced by Professor Marvin's table of wind veloeities, is that adopted at all Weather Bureau stations. In order to obtain the general velocity of the wind, free from all local effects. these anemometers are, if in a city. placed as high as practicable above the roofs of tall buildings, or if located in the country. on the tallest available support. The velocities thus obtained are considerably higher than the average at the surface of the earth, but the winds at the surface arc much weakened by resistances. and these higher locations are needed in order to give us a clear idea of the gen eral motion of the air under the action of the barometric pressures indicated by the isobars.
When no anemometer is available, the pressure velocity or force of the wind is estimated and recorded on sonic arbitrary scale, such as that which was introduced into the British navy by Admiral Beaufort about 1800, and is almost uni versally used at sea. (See BEAUFORT SCALE.) The Weather Bureau has used various scales of numbers and terms, hut the tendency is not to de part from the Beaufort scale. The anemometer is easily made to register its own indications on a sheet of paper, and thus becomes an anemo graph. This is done mechanically in the Kew pattern used at. British stations of the first order. but is done electrically at the Weather Bureau stations.