VII 



FLIGHT 263 



giving what I consider to be the explanation, I shall 

 first try to show that what is accepted as the solution 

 of the problem by some noted ornithologists in reality 

 leaves it just where it was. 



If you take two cubes, a side of one of which is 

 twice the length of a side of the other, the larger 

 one is in bulk eight times the smaller, but its surface 

 area is only four times as great (see figure 29, p. 113). 

 This will hold of other figures of three dimensions. 

 Magnify a bird till it is eight times its former size : 

 yet you will only have multiplied the surface area by 

 four. This is no doubt a true principle in geometry, 

 and it might be applied to the present case if 

 symmetry were the only thing under consideration. 

 We should then be doing right if we took the pro- 

 portion of bulk (not area) of wing to bulk of body, and 

 it would turn out that the build of big birds and small 

 is not very different. 1 But the present question is 

 really one of dynamics. The problem which nature 

 has to solve when she increases a bird's size is : " If 

 the weight of the bird be multiplied by so many, how 

 much will the area of the supporting surfaces have to 

 be increased ? " In other words, how much must be 

 added to the wing area, in order that the bird may be 

 able to fly ? And the answer is (except for modifying 

 circumstances which I shall pass on to soon), that if 

 the weight be doubled the supporting surfaces must 

 also be doubled. 



There is an undeniable interest in the fact that 

 when we compare birds' bodies and wings (the legs are 



1 Those who wish to apply this principle should take the cube 

 root of the bird's weight and the square root of the wing area. 



