One is left with the impression that the whole subject is still unsettled. While earlier workers sought to reduce the entire matter to law, the latest disclaim the possibility of generalization. After summarizing Brewster's and Gätke's opinions, Thomson (1926:79) wrote: "It would seem, in any event, that no general rule can be laid down." Meinertzhagen's summation (1930:56-57) still represents fairly well the status of our knowledge: "Order of sex and age on autumn passage is very difficult to arrive at, as evidence is conflicting. But, on the whole, it seems that birds flock together, old and young, preparatory to moving south, and do in many cases initiate migration in company.... But once movement is initiated, among birds which do not habitually fraternise in flocks, adults, and especially males, will naturally outstrip the less virile females and still less virile offspring.... The consequence is that any observer at an intermediate station such as Heligoland is, in noting birds of the year as first arrivals, has not had an opportunity of noting the flocks of adults which have passed without alighting. On the other hand, there is very definite evidence to show that among certain species, adults follow their offspring on migration. The reason for different behaviour among different types of birds remains obscure." We regard much of this as still theory.

Fig. 1. Composition by age and sex as found in one or more series of each of eight species of birds included in the Topeka sample. Each separate series is represented by a single histogram, the histograms for a species being grouped with the earliest series on the left. Each histogram expresses the numbers of adults (left-hand column) and immatures (right-hand column) in terms of percentage of the whole series. Thus the two bars of each couplet add up to 100 per cent. The hatched portion of each bar represents males, the clear portion females.

It would be difficult to imagine a better way of resolving the problems and uncertainties just reviewed than by the detailed analysis of large samples of migratory birds killed at random at various points and times. An analysis of the sample of birds accidentally killed at Topeka is presented here as an initial step in this direction. Although the conclusions based on this sample are tentative and may in time be altered, the data themselves are definite. If this general type of analysis is repeated again and again—applied to samples taken on many dates and in many localities—a mass of hitherto unparalleled evidence for the study of migration will emerge.

Differential Migration OF Sex- and Age-classes as shown by the Topeka Sample.—Smaller samples have not been treated. Species affording samples seemingly large enough to justify at least preliminary analysis were: Catbird, Red-eyed Vireo, Mourning Warbler, Dickcissel, Nashville Warbler, Orange-crowned Warbler, Yellow-throat, and Lincoln Sparrow (Fig. 1). For all of these except the Catbird and Dickcissel, at least two samples from a week or more apart were available for comparison in an effort to detect trends in migration. Fig. 1 shows the actual ratios of sex- and age-classes observed in samples of the species listed above. Each of the last four species provided two separate samples, of sufficient size to warrant an attempt at measuring the statistical significance of the observed changes in adult-immature ratios (Table 2).

Table 2.—Statistics of the Ratios of Adults to Immatures in Four Species

SpeciesDates of samplesTotal numberNumber and percentage of adults[1]Difference (in %)[2]P[3]
Nashville WarblerOct. 1 (93)
Oct. 5-7 (63)		15645 (.484)
26 (.413)		.071.36
Orange-crowned WarblerSept. 25-Oct. 1 (19)
Oct. 5-9 (25)		443 (.158)
11 (.440)		.282.05
Yellow-throatOct. 1 (115)
Oct. 5-8 (44)		15962 (.540)
20 (.455)		.085.34
Lincoln SparrowOct. 1-3 (44)
Oct. 6-10 (27)		7127 (.614)
8 (.296)		.318.01

[1] Percentage of immatures equals 1.000 minus percentage of adults.

[2] Standard error of the difference between ratios was computed by the formula

σe =√Pe Qe (1+1),
N1N2