us BRITISH BIRDS. [vol. xv. 



by aii}^ species before it can be expected to be remunerati^'e : 

 it must be procurable for marking in large numbers, and it 

 must ^deld a good percentage of reappearance records : both 

 these are points on which sufficient information is now avail- 

 able. A third requirement is that the species should 

 present some interesting problem ; some species which 

 are good in the respects previously mentioned tend to 

 fail here. The point is, of course, open to dispute in that 

 there may be no agreement as to the nature of the problems 

 to be studied, and it may be admitted that to some extent the 

 problems are unknown and that their scope cannot therefore 

 be predetermined. There are, however, some obvious gaps 

 in the material available for a theory of migration, and the 

 facts required to fill these may well be taken as an immediate 

 object of inquiry. One example will suffice : much in the 

 theory of the origin and nature of migration depends on whether 

 it be true or not that the members of a species breeding 

 farthest north " winter " farthest south, other members 

 being stationary midway. The British Isles present a 

 particular aspect of this question and one which confronts 

 their students of migration at every turn. Very many 

 species are found all the year round in the British Isles but 

 are known in autumn both as immigrants from the north 

 and as emigrants to the south, and vice versa in spring : do 

 the visitors from the north replace our native birds while 

 these journey southwards, or do our native birds remain 

 while the others pass through to still more distant winter- 

 quarters ? The question is here stated in its crudest form, 

 but is yet enough to indicate a fruitful field of inquiry to 

 which the marking method would seem to be peculiarly well 

 adapted. 



There is, however, still another requirement which actual 

 experience of working the n^etliod has revealed, namely 

 that the circumstances must be such as will not invalidate 

 the results bv the introduction of too many uncertain factors. 

 This point requires elucidation, but a few examples from the 

 Aberdeen University data may make it plain. 



In working up his results the writer found that the only 

 satisfactory means of correlating his records for any species 

 was by classifying them in homogeneous groups within each 

 of which one record would be strictly comparable with another, 

 and between which some comparison on broader lines could 

 also, perhaps, be made. In some cases this was found to be 

 impossible. Starlings marked in winter were, for instance, 

 quite likely to be a mixture of native birds and of birds bred 



