(2) An early preponderance of immatures, followed by a preponderance of adults. The several species of birds at Topeka that display this pattern conform with the conclusions of Gätke and other early Old World ornithologists that in most species immatures precede adults in migration. In the present sample two variations of this pattern occur.
(a) In the Dickcissel (Fig. 1, F) and the Mourning Warbler (Fig. 1, G), immatures decrease more markedly than adults (visible in samples of Mourning Warbler; inferred in Dickcissel), leaving the adults in the majority in the closing phase of migration. The distinctive and interesting feature in each of these two species is the ascendancy in numbers of adults despite the almost complete disappearance of adult males. The relative increase of adults is here caused by a retarded migration of adult females, which linger conspicuously behind all other classes. Something of this nature was suggested, in theory, by Dixon (1892:70) who thought that adult females are delayed by "maternal duties." It was hinted at also by Dwight (1900:127) who thought that in some species females molted later than males as a result of prolongation of parental responsibilities. As mentioned already, there is need for caution in interpreting the present samples because the Dickcissel is represented only by one sample and two of the three samples of Mourning Warblers are small. In the case of the Mourning Warbler, the samples may be regarded as one, nearly lacking in adult males. The progressive increase of adult females, however, may be significant; at least there are enough of these to make division of the birds into three samples enlightening. There is, of course, some chance that the majority of adult males have not yet migrated, or are migrating by a different route. This seems unlikely in both cases. October 1 is late in the migration of the Dickcissel and it seems that large-scale migration would not occur much later, and in the case of the Mourning Warbler adult males are rare in all three samples, extending over a considerable period and reaching late into the probable migration period of the species. It is interesting to conjecture just when and where adult male Mourning Warblers do migrate in autumn. Brewster (1886:16) wrote: "This species arrives at Cambridge [Massachusetts] about September 12, and during the remainder of the month is ... abundant.... The adults, however, are so very uncommon that I have never known them [to] represent more than five per cent of the total number of individuals. They do not seem to be more numerous in the earlier flights than towards the close of the month, and I am very sure that they cannot be found in this locality before the young begin to appear." While the present samples show an abundance of adult females of this species (could Brewster have failed to recognize these as adults?) the whereabouts of the adult males remains a mystery.
(b) Another variation is displayed by the Orange-crowned Warbler (Fig. 1, H). Here also there is an increase of adults towards the end of migration, but this increase is marked by a growing percentage not of females but of males. Locally this species is a late migrant compared with most others of the Parulidae. Thus the first sample, composed of birds taken September 25-October 1, may be regarded as fairly early in the fall migration. Immature birds compose 84.2 per cent of this sample, there being no adult males at all. By October 5-9 the picture has changed markedly, the sample being composed of 44 per cent adults (82 per cent of which, in turn, are males) and 56 per cent immatures. In view of this trend one can not help suspecting that a still later sample would show a majority of adults, perhaps nearly all males. This of course does not necessarily follow; the migration of immatures could simply be more protracted, and could have commenced earlier, than that of adults.
Little imagination is required to see how enlightening it might be could we analyze thoroughly the patterns of all migrating species. When the detailed facts are available, it seems likely that general trends will emerge which may be of great significance to the study of migration in general. A final point which must eventually be clarified is determination of the extent of variability in the pattern of each species from year to year and locality to locality.
Once patterns of precedence in migration of different classes are established, search into the life-histories of the species concerned may help to explain the peculiarities discovered. In the present case, for instance, we find a possible clue to the reason for the high proportion of adult females of the Dickcissel late in migration, as shown by our sample. Gross (1921:14-15) presented evidence that adult female Dickcissels molt considerably later than their mates, and we have independent evidence that individuals of this species are at times almost flightless when molting the remiges!
Molt in Relation to Migration
General comment.—The exact relationship between molt and migration seems not to have been definitely established. The subject has received cursory attention in the literature and conflicting opinions have been expressed. Dwight (1900:126-128) believed that molt is completed or nearly completed before migration in nearly all passerine species that occur in New York save for certain swallows and flycatchers. Molt has since been found to precede migration of at least one of the flycatchers (Empidonax virescens) considered by Dwight to be an exception to this rule (Mengel, 1952). In Great Britain the subject of molt in migration was considered in some detail by Rintoul and Baxter (1914) and Ticehurst (1916), who arrived at conclusions similar to Dwight's. These workers also found certain swallows to be exceptions to the rule.
The above authors and others have shown that, at least among passerines, some body molt is frequently found in migrating individuals but that molt of tail feathers is much less often found and molt of remiges almost nonexistent. Baxter and Rintoul noted only four cases of molting wing feathers among hundreds of migrants. Among the diverse non-passerine orders the picture seems to be more complicated, as might be expected. We do not, however, comprehend the reasoning which led Meinertzhagen (1930:56) to summarize: "... on the whole it can be said that though birds seldom migrate when flight feathers are in quill, moult in general does not influence migration." This seems to us an obvious non sequitur. Meinertzhagen (loc. cit.) went on to say: "Males and females of one species are believed to moult simultaneously [see, however, Dwight, 1900:127], and there is no doubt that in some cases the two sexes migrate at slightly different times, and occasionally prefer different winter quarters. Birds of the year never moult their quills previous to their first autumn migration [Consultation of Dwight, 1900, who gives many examples of this, would have spared Meinertzhagen this error.], and yet they frequently follow adults on passage and sometimes precede them. There are no grounds for believing that adults have moulted their quills before birds of the year are prepared to migrate [but there are, in many cases; cf. Dwight, 1900:127], in the case where adults precede the latter. Neither is there any evidence to show that adults have not moulted their quills till after their offspring are ready for passage, in the case where they follow their offspring. It does not, therefore, appear that moult is an important factor."
Comments interpolated above show our feeling that this summary is inadequate and misleading. To us it seems that the extreme rarity in migration of birds with remiges in molt is strong evidence that molt does influence at least the time of migration. It is immaterial whether this molt occurs before or after migration, although in the majority of cases it seems to take place before. Much more needs to be known of the migration pattern and molt of each species before generalizations can safely be made.