BY R. J. TILLYARD. 43 



are only severed by it, (i.e., the group is divided into two distinct 

 portions) the barrier is only a Secondary Barrier. It is clear 

 that a barrier can only be a Primary Barrier to those groups 

 whose arrival in its neighbourhood is of later date than the 

 uprising of tlie barrier; while, even to such groups, if they possess 

 special facilities for passing the barrier, it may only play the 

 part of a Secondary Barrier, or even be no barrier at all. On 

 the other hand, if the date of the uprising of the barrier be 

 later than that of the arrival of the group, it cannot rank higher 

 than as a Secondary Barrier. A recognised barrier, such as 

 Wallace's line, might be carefully treated in this manner for a 

 large number of groups with very valuable results. 



A further suggestion as to a valuable use of this contour 

 method is offered by the author for the case of migrating groups 

 of birds. With sufficient records, two separate contour maps 

 might be drawn up for, say, one of the genera of the Fjnngillidce, 

 showing (a) the contour of the group during the nesting season; 

 {b) its contour during the winter. These two contours, drawn 

 on large maps and exhibited side by side, would bring home to 

 us, more clearly than pages of records, the movements of the 

 group during the changing seasons of the year. Probably the 

 records available in Europe and America for such a contour will 

 be found to be quite sufficient. 



In conclusion, the author contends that the study of zoogeo- 

 graphical distribution will be advanced by the method outlined 

 in this paper, and that contours of groups are a more natural 

 unit for study than theoretical subdivisions of regions into 

 separate portions. 



