FLYING TO BED 149 



these last flights of the starlings, which appears to 

 me to ofl^er additional evidence of this being the 

 case. Supposing there to be a hedge, or any other 

 shelter, in the bird's course, one can, by stooping 

 behind it, remain concealed or unthought of, whilst 

 they pass directly overhead. One then notices that 

 there is a constant and, to some extent, regular 

 rising and sinking of the rushing noise made by 

 their wings. It is like rush after rush, a maximum 

 roar of sound, quickly diminishing, then another 

 roar, and so on, in unvarying or but little varying 

 succession. Why should this be .? That, at more 

 or less regular intervals^ those birds which happened 

 to be passing just above one, should fly faster, thereby 

 increasing the sound made by their wings, and that 

 this should continue during the whole flight, does 

 not seem likely. It would be method without 

 meaning. But supposing that, at certain points, 

 the living stream were composed of greater multi- 

 tudes of birds than in the intermediate spaces, then, 

 at intervals, as these greater multitudes passed above 

 one, there would be an accentuation of the uniform 

 rushing sound. Now in a moderate-sized band of 

 starlings, flying rapidly, there is often a thin for- 

 ward, or apex, end, which increases gradually, or, 

 sometimes, rather suddenly, to the maximum bulk 

 in the centre, and a hinder or tail end, decreasing 

 in the same manner. If hundreds of these bands 

 were to fly up so quickly, one after another, that 

 their vanguards and rearguards became intermingled, 

 yet, still, the numbers of each main body ought 

 largely to preponderate over those of the combined 

 portions, so that here we should have a cause 



