WATCHING BIRDS AT A STRAW-STACK 213 



unlikely that there should be a leader, and both this 

 and the other explanation, which at first sight seems 

 satisfactory, are destroyed by the salient fact that in 

 hardly any case do all the birds rise and fly away 

 together, but only the great majority. Almost invari- 

 ably a certain number of them, though sometimes only 

 half-a-dozen or so, or even less, remain, nor has this 

 anything to do with the particular species of bird. 

 Moreover, the flying up of any bird from the crowd 

 does not, of itself, communicate alarm to the others, 

 for first one and then another and often several at a 

 time may constantly be doing so, whilst the rest feed 

 quietly and take no notice. It may be said that it is 

 only when a bird flies off in alarm that its flight com- 

 municates alarm to its companions. That it does so 

 necessarily, even in such a case, I, from general obser- 

 vation, very much doubt, and also, if the facts as I 

 have given them be a little considered, it will be seen 

 that the diiificulties are not met by this view of the 

 case. 



The theory of a leader seems more applicable to 

 birds like rooks, which are gregarious, and may be con- 

 stantly watched in large numbers together, without the 

 intermixture of any other species. The same diffi- 

 culties, however, apply here, and even to a greater ex- 

 tent, for the movements of rooks are more complicated, 

 whilst alarm or any such primary impulse as the origin 

 (I do not say the explanation) of them, is in most cases 

 quite out of the question. An instance or two of these 

 sudden and quite simultaneous nTovements of bodies 

 of rooks I have noted down directly after observing 

 them. They would be much in place here, but as I 

 have two chapters devoted to these birds, and, more- 



