525 
1921 .] Bird-Migration by the Marking Method. 
already been pointed out, the number of birds in a, given 
area at a particular time is only imperfectly represented by 
the mortality rate, because the latter varies with time of 
year and other circumstances. Moreover, the true mortality 
rate is still less adequately represented by the “ recorded 
mortality ” rate which forms the datum of the method. These 
difficulties cannot be altogether overcome, but they will be 
minimised where the problem takes the form of comparisons 
between sets of results collected under approximately similar 
circumstances. There is virtue, too, in the mere numbers of 
records, and concentration of effort should lead to an increase 
in these for the particular species selected. Where the 
total numbers are small the quantitative errors due to mere 
chance will obviously be great, and the extent of the possible 
discrepancies from this source is strikingly shown by the 
figures for the Lapwing (cf. Table II. p. 485), one of 
the best subjects for study. 
The selection of suitable species does not exhaust the 
question, for it is also important that the reappearance 
records should fall into groups which are homogeneous as 
regards the circumstances of marking. A concentration of 
effort on particular areas would therefore have its advantages. 
The question of season is probably still more important, and 
a serious objection must be considered in the case of most 
marking of birds at other times than the breeding season, 
namely, that the birds marked may consist of a mixture 
of sedentary natives and visiting immigrants from other 
summer-quarters. 
There are various other points that need not be laboured. 
The unreliability of isolated records which may well be 
abnormal will, for instance, be sufficiently obvious. The 
danger of laying stress on negative evidence except in very 
clear cases may also be mentioned : in considering the 
Hedge-Sparrow and the Blue Titmouse, for instance, it has 
been seen that almost negligible numbers of marked birds of 
these species are recovered where no special efforts are made, 
whereas an exceedingly high figure may result if continuous 
trapping be carried out. The absence of records from a 
