108 THE MIGRATION OF BIRDS 



weather, and therefore are not trustworthy indica- 

 tions of the density at any time or place of migration. 

 Out of 115 song thrushes killed at the lights and 

 sent to him, 80 per cent struck during the fourth and 

 first quarters of the moon, and the same rule holds 

 good for other species. The intimate relation 

 between the lunar phases and the number of examples 

 killed was shown by statistics from 1888 to 1894. 

 Out of 673 specimens received only 116 were killed 

 when the moon was more than half full. 



Apart from fog or cloud, birds may fail to hit the 

 land aimed at, either through accidental diverg- 

 ence from correct direction or wind drift. In 

 November 1884 Mr Barrington received information 

 of large numbers of rooks passing simultaneously 

 at the Tearaght and Skelligs Lights- -island 

 stations 20 miles apart and each 9 miles off the 

 Kerry coast. The birds arrived in continuous 

 flocks from the westward the open Atlantic- 

 and passed in an easterly and landward direction ; 

 they came in small parties and in flocks numbering 

 two or three hundred, on many days between the 

 2nd and 25th of the month. A few birds were 

 noticed at the same time at stations on the south 

 and east Irish coasts, and all alike making for the 

 land. From similar observations made in other 

 years he concludes that these were portions of hosts 

 which had overshot the mark, and failing to find 



