236 THE STRUCTURE AND LIFE OF BIRDS cnap. 
inflated with hydrogen gas in different degrees, and 
allowing them to race: those which were more dis- 
tended, rising to a height of 500 feet or more, quite 
outpaced the smaller ones floating below them. But, 
as far as I could judge, smal! differences of elevation 
at the higher levels did not much affect the velocity. 
"Since making these experiments, I have found that 
far more elaborate ones, with kite-wire suspended 
anemometers, had already been recorded.. They 
showed that the velocity of the wind increased up to 
a height of nearly 1,100 feet above the ground, but 
that the rate of increase there was very slight, the 
difference being only forty-four feet per minute ° 
between the velocities at altitudes of 795 feet and 
1,095, and eighty-five between 549 and 795.} 
It seems probable, then, that the wind increases in 
velocity up to a certain altitude, but that above that 
there is no appreciable increase. This being so, we 
must look for some other irregularity of which a bird 
soaring over a level plain can make use when he has 
passed this limit. It is within every one’s experience 
that wind often comes in gusts.. But we not unfre- 
quently speak of a steady breeze, meaning one which 
is almost or entirely free from such fitfulness. Our 
experiments with an anemometer at New Romney 
went to show that a steady wind, in the strict sense 
of the term, does not exist. We had to expose the 
instrument for not less than a minute at a time, make 
a series of experiments, and strike averages, in order 
to obtain dependable figures. Professor Langley has 
by means of a delicate anemometer made more thorough 
1 Nature, April 22, 1886, 
