Dr. Robinson has proved, both by theory and experiment, that the centre of any one of the cups so mounted and set in motion by the wind, revolves with one-third of the wind's velocity. If, therefore, the diametrical distance between the centres of the cups be one foot, the circle described by the centres in one revolution is 3.1416 feet, and the velocity of the wind will be three times this, or 9.42 feet, which must be referred to time for the absolute rate. The instrument is sometimes made with the centres of the cups 1.12 feet apart, so that the circle described is 1/1500 of a mile in circumference. Hence, to produce one revolution of the cups, the wind must travel three times as fast, or 1/500 of a mile. Therefore, 500 revolutions will be produced by one mile of wind; so that the dials may be graduated to register the velocity in miles and tenths of miles. The simplest arrangement is with five dials, recording respectively 10, 100, 1,000, 10,000 and 100,000 revolutions. _Directions for using Robinson's Anemometer._--The dials read off in the same manner as the register of a gas meter, commencing with the dial farthest from the endless screw. "The figures on the first dial indicate so many hundreds of thousands of revolutions; those on the second dial so many tens of thousands; those on the third, thousands; those on the fourth, hundreds; and those on the fifth so many tens. "The instrument should be read every morning at 9 o'clock; and, usually, it will only be necessary to read the first three dials. The figures can be entered as they are read off. Should the index point _between_ two figures, the less of the two is to be taken. "For example, if the first dial points to 7, or between 7 and 8; while the second dial indicates 4; and the third, 5; the entry to be made is 745 (indicative of 745 _thousand_ revolutions). "Every time the index of the first dial is found to have passed zero (0), a cross or star is to be prefixed to the next (a lower) reading. "To ascertain how many _thousands_ of revolutions have been made during the month, it will simply be necessary to subtract the first reading from the last, and prefix to the three figures thus obtained a figure corresponding to the number of stars in the column. For every _thousand_ revolutions there are two miles of wind: we have therefore only to multiply by 2 to find how many miles of wind have passed during the month. "Two entries must be made for the last day of each month (the one being written under the other), so as to bring the readings down to 9 A.M. on the 1st of the following month. The same entry which ends one month, will therefore begin the next. This repetition of one entry is necessary, in order to prevent losing a day's wind.