176 THE STRUCTURE AND LIFE OF BIRDS cuab. 
numbers 1, 2, 3, 4, but the resistance of the air was, 
as we have seen, equal to the weights of the globes 
which were as I, 4, 9, 16, the squares of the numbers 
which represent the velocities. Experiments at once 
more elaborate and more accurate have been made 
since. Professor Marey concludes from many made 
by himself, that the resistance increases in a less 
proportion for velocities between o and Io metres per 
second ; when Io metres per second is exceeded, then 
the rule of the square of the velocity under-represents 
the rate of increase of the air’s resistance. When 
the speed attained is very great, in-the case of a 
bullet for instance, Newton’s law does not hold at all: 
the -rate of increase of resistance altogether outpaces 
the square of the velocity. The rate of movement of 
a wing is comparatively moderate, so that here it 
might seem that we should be safe in applying 
Newton’s law. There is liability to error, however 
from another cause. A wing is very different from 
the glass globes with which he experimented—it 
presents a concave and irregular surface with rough 
edges. Such an object passing through the air, which 
is not a perfect fluid, but viscous, must, like an oar 
forced through the water; produce eddies, and this 
complicates the problem so much that our greatest 
authorities confess that we know very little of the 
resistance to a surface like that of a wing. It is 
necessary to say this, since the rule “resistance of air 
increases as the square of the velocity” is often 
quoted as if it held true of all surfaces and all 
velocities. Nevertheless it comes near enough to the 
facts to be of great value, and probably when we 
