S. P. Langley — Internal Work of the Wind. 51 



sive instants of its passing the fixed anemometer, which velo- 

 cities, indeed, were probably nearly the same for a few seconds 

 before and after registry, but which incessantly passed into — and 

 were replaced by — others, in a continuous flow of change. But 

 although the observations do not show the actual changes of 

 velocity which any given particle experiences in any assigned 

 interval, these fluctuations cannot be materially different in 

 character from those which are observed at a fixed point, and 

 are shown in the diagram. It may perhaps still further aid us 

 in fixing our ideas, to consider two material particles as start- 

 ing at the same time over this two mile course ; the one mov- 

 ing with the uniform velocity of 22 6 miles an hour (33 feet 

 per second), which is the average velocity of the wind as 

 observed for the interval between 12 hrs 10 m ' iDS 18 secs — and 12 hrs 

 15 mins 45 seos on February 4 ; the other, during the same inter- 

 val, having the continuously changing velocities actually 

 indicated by the light anemometer as shown on Plate III. 

 Their positions at any time may, if desired, be conveniently 

 represented in a diagram where the abscissa of any point 

 represents the elapsed time in seconds, and the ordinates show 

 the distance in feet, of the material particle from the starting 

 point. The path of the first particle will thus be represented 

 by a straight line, while the path of the second particle will be 

 an irregularly curved line, at one time above, and at another 

 time below the mean straight line just described, but terminat- 

 ing .in coincidence with it at the end of the interval. If, now, 

 all the particles in two miles of wind were simultaneously 

 accelerated and retarded in the same way as this second parti- 

 cle, that is, if the wind were an inelastic fluid, and moved like 

 a solid cylinder, the velocities recorded by the anemometer 

 would be identical with those that obtained along the whole 

 region specified. But the actual circumstance must evidently 

 be far different from this, since the air is an elastic, and nearly 

 perfect fluid, subject to condensation and rarefaction. Hence 

 the successive velocities of any given particle (which are in 

 reality the resultant of incessant changes in all directions), 

 must be conceived as evanescent, taking on something like the 

 sequence recorded by these curves, a very brief time before 

 this air reached the anemometer, and losing it as soon after. 



It has not been my purpose in this paper to enter upon any 

 inquiry as to the cause of this non-homogeneity of the wind. 

 The irregularities of the surface topography (including build- 

 ings, and every other surface obstruction) are commonly 

 adduced as a sufficient explanation of the chief irregularities 

 of the surface wind ; yet I believe that at a considerable dis- 

 tance above the earth's surface (e. g. one mile), the wind may 

 not even be approximately homogeneous, nor have an even 



