"The accompanying example of the 687 readings of an Anemometer for 13 days 773 will illustrate the method of making 822 the entries, &c. 855 900 "In this instance, the first reading 953 (687) is less than the last (793). 990 When the first reading is greater than *066 the last, it will be necessary to borrow 197 1,000 in making the subtractions, 323 and then deduct one from the number 414 of stars. Thus, if the first reading 597 of the series on the margin had 712 been 887, the result would have been 793 906 instead of 1106. ---- 1106 thousands of revolutions. 2 +----- 13 | 2212 miles of wind in period. +----- 170 miles of wind per day, on an average. "The foregoing directions are all which require to be regularly attended to. But it may be interesting at times to find the velocity of the wind during a period of a few minutes. This may be ascertained by observing the difference of two readings of all the dials, with an interval of some minutes between them, when a very brief calculation will suffice; but perhaps the simplest method is the following:-- "Take two readings, with an interval of 12 minutes between them. The difference of these readings, divided by 10, is the velocity of the wind in miles per hour. Thus--if the reading of the five dials (from left to right) at noon is 15206, and at 12 minutes past 12 is 15348, the velocity of the wind is 14.2 miles per hour."--_Admiral FitzRoy, F.R.S._ A lever and clutch are sometimes fitted to this anemometer, as in fig. 88, for throwing the train out of gear when not required to register. It may also be connected with clock-work so as to be self-recording, by causing the mechanism to impress a mark upon prepared paper moved by the apparatus, at certain intervals of time. [Illustration: Fig. 88.] This anemometer should be fixed in an exposed situation, as high above ground as may be convenient for reading. It may be made very portable, by the arms which carry the cups being fitted to unscrew or to fold down. When fitted in gimbals, it can be used at sea with much advantage. The pressure of the wind has been experimentally proved to vary as the square of the velocity; the relation being _V squared_ = 200 x _P_. From this formula, therefore, the pressure can be calculated corresponding to the observed velocity.