THE SENSE OF DIRECTION. 51 



that migration was arrested when the sky was completely over- 

 cast, but it was resumed after less or more of starlit sky became 

 visible. Possibly, we are here in touch with a visual factor of 

 great importance, though of an unknown nature. It appears as 

 if the sea, the sky, and the horizon, in whole or in part, furnish 

 surfaces and lines of definite value, and as effective for the 

 second component as the fixed surfaces and lines of the land- 

 scape. Most often in Nature the second component is disabled 

 by the presence of haze or fog. Probably the absence of rela- 

 tively fixed points in an atmosphere laden with condensing 

 vapour makes the keeping of a true course impossible. If the 

 bird continue its journey, drift may set in and pass unnoticed. 



We have seen that the environment is able to furnish sur- 

 faces and lines for the maintenance of a true course. To a 

 rapidly moving bird or insect these surfaces and lines must 

 present the appearance of running contrary to the animal's 

 direction of motion. Any divergence from the proper trajectory 

 of flight alters the relative distances and the apparent direction 

 of motion in the environment which must have the widest pos- 

 sible extension in order to prevent local peculiarities throwing 

 the animal off its course. Now, if we suppose a bird in flight 

 to be pointing towards its objective, and to be drifting before a 

 wind blowing across its course, the definitive surfaces and lines 

 coming within the visual fields will appear to run not directly 

 but to curve obliquely backwards, the obliquity being inclined 

 towards the wind. The oblique distortion of the environment 

 and the peculiar alteration of the lateral distances relative to 

 the bird will compel it by force of habit to correct the displace- 

 ment. This the bird can only do by altering the axis of its 

 body to an alignment lying between the direction of the wind 

 and the course towards the objective. As it happens, birds 

 actually do so whenever the cross-wind is of sufficient strength 

 to set up drift. In general terms the deviation of the bird's 

 axis, as estimated at the anterior end, is always towards the 

 wind, whether it blows before or abaft the beam. Hence the 

 bird has a constant tendency to eat into the wind a little, and 

 the appearance of birds flying under these conditions suggests 

 that the deviation to windward is proportional to the strength 

 and direction of the wind, and automatically sufficient to 



