close allies of Florida birds found in Cuba, belong to the latter. The fact that many of the older genera have representatives on each of the islands would seem to indicate a past direct or indirect connection between the islands of the group. The generally close relationships which exist between the species of these genera points to the conclusion that they are derived from a common ancestor differing but slightly from the present type.

If we assume that the West Indies were separated sometime during the Middle Tertiary, we may then regard these older forms as survivors of the fauna of that period, which have been preserved to us through the isolation afforded by an insular life. This supposition is supported by the fact that they are quite as distinct from existing genera as are the genera of birds which have been described from the Miocene.

The isolation which has protected these old types has also resulted in the differentiation of the species derived through migration. Thus while the West Indies have preserved to us species which on the mainland have succumbed to the continental struggle for existence, they have given us many new forms which have been differentiated from their mainland ancestors under the influences of a new environment.

Article XVII.-ON A COLLECTION OF BIRDS FROM CHAPADA, MATTO GROSSO, BRAZIL, MADE BY MR. H. H. SMITH.

By JOEL ASAPH ALLEN.

PART II.-TYRANNIDÆ.

(Continued from Vol. III, p. 380.)

The present Part treats only of the family Tyrannidæ, which is represented by 45 species. In treating of Myiarchus tyrannula, reference is also made to its West Indian, Mexican and United States allies, at one time considered by several ornithologists of repute as merely subspecies of the M. tyrannula group.

Since the publication of Part I the Museum has purchased the remainder of the Smith Collection of Chapada birds (see this Bulletin, Vol. III, p. 337), thus increasing the number of duplicates available for exchange.'

88. Tænioptera nengeta (Linn.).-A series of 59 specimens, divided about equally between males and females, represents every month in the year except October, as follows: January, 4; February, 4; March, 7; April, 9; May, 5; June, 7; July, 4; August, 4; September, 2; October, o; November, 6; December, 6.

The plumage in the November and December specimens is much faded and worn. The molt begins in December, and the new clothing plumage is pretty well acquired before the end of January, but the molt of the quills continues irregularly till into April. As a rule, however, March and April birds are in good condition, but the highest condition of plumage is seen in the May and June specimens. The females average a little smaller than the males, but there is no very appreciable difference in coloration. By August the plumage begins to show signs of deterioration. In old. males the tips of the first and second primaries are more or less incised on the inner vane, but much less so than in some other species of the genus.

I have learned since the publication of Part I that the credit for gathering and preparing the collection is about equally due to Mrs. H. H. Smith and Mr. W. C. Smith, the latter preparing most of the birds collected during 1883-84.

A bird of the year, in first plumage, taken Dec. 19, has the white of the lower plumage tinged with pale buff, and the gray of the upper plumage with brownish, the rump and upper tailcoverts being decidedly brown; the tips of the wing-coverts are buffy white instead of grayish white.

The measurements of 10 adults of each sex are as follows: Males wing, 5.02-5.68, averaging 5.39; tail, 3.20-3.76, averaging 3.58; culmen, .70-.82, averaging.76. Females: wing, 4.90-5.34, averaging 5.19; tail, 3.42-3.74, averaging 3.56; culmen, .70-.80, averaging .75.

89. Tænioptera velata (Licht.).-Three specimens-Cachoeria, & ad., Feb. 4, 1886; Chapada, & ad., Sept. 23, 1883, ad., Dec. 6, 1883.

90. Fluvicola albiventris (Spix).-One specimen, Corumba, March 24, 1886.

91. Arundinicola leucocephala (Linn.).—Four specimens, 2 males and 2 females, Corumba, March, 1886.

92. Alectrurus tricolor Vieill.-One specimen, &, June, 1883.

93. Cnipolegus comatus (Licht.).-Six specimens, 5 of which are males, taken as follows: January, 1; July, 1; August, 3; October, 1.

94. Copurus colonus (Vieill.).-Three specimens, November, 1882 and 1883. Two are adults and the other young, in the black plumage=C. funebris Cab. & Heine (Mus. Hein., II, 1859, p. 41).

95. Platyrhynchus bifasciatus Allen.-As already stated (this Bulletin, II, No. 3, 1889, pp. 141, 142) this species is represented by 18 specimens, 11 males and 7 females, all from Chapada. The series having been already fully described (1. c.) further remarks are not necessary.

96. Todirostrum cinereum (Linn.).-The series of 13 specimens was collected as follows: Chapada, April (2 specimens) and August (5 specimens); Corumba, 6 specimens February,

March, and April. Of these specimens 7 are sexed as males, 4 as females, and 2 are not marked for sex, but are doubtless females. In all there is a very narrow frontal band of yellow at the base of the bill, in some, however, scarcely discernable. The females are distinguishable from the males by a very small spot of white (sometimes yellowish white) on the middle of the crown, but apparently are not otherwise different from the males. There is, however, much variation in color and size, and especially in the size of the bill, but it is apparently individual and seasonal rather than sexual. The 7 males measure as follows: wing, 1.58– 1.88, averaging 1.77; tail, 1.36–1.46, averaging 1.43; culmen, .52– .58, averaging .55. The 6 females measure; wing, 1.66-1.80, averaging 1.70; tail, 1.28-1.40, averaging 1.34; culmen, .52-.57, averaging .54.

mens.

97. Euscarthmus ochropterus Allen.—Chapada, 15 speciThere is nothing to add to the description of this species already given (see this Bulletin, II, No. 3, June, 1889, pp. 143, 144).

98. Euscarthmus pelzelni Scl.- One specimen, Chapada, May 23, 1885.

99. Euscarthmus striaticollis (Lafr.).—Two specimens, Chapada, & ad., Feb. 18; 9 ad. (no date).

100. Hapalocercus meloryphus (Wied).-One specimen, Chapada, Aug. 25.

101. Habrura pectoralis (Vieill.).-Chapada, 10 specimens -May, 2; July, 2; August, 4; September, 2. These specimens have already been commented upon in a comparison between H. pectoralis and H. superciliaris (Wied). (See this Bulletin, II., No. 3, June, 1889, p. 146.)

102. Culicivora stenura (Temm.).—One specimen, Chapada, ad., April 13.

103. Serpophaga albogrisea Scl. & Salv.-Three specimens, taken as follows: Abrilongo, & ad., Feb. 28; Chapada, ad., July 20; juv. in first plumage, Nov. 6. The young bird.

is brown above, with each feather narrowly edged with whitish; below, throat and breast dark ashy brown, faintly cross-barred with darker brown; rest of lower parts whitish with a faint tinge of greenish buff; under wing-coverts greenish yellow; wings and tail dusky brown.

104. Leptopogon amaurocephalus Cab. - Chapada, 10 specimens, collected as follows: February, 2; May, 1; June, 2; August, 1; November, 4. There is apparently very little seasonal or sexual variation, the series being remarkably uniform as regards both size and coloration. In November specimens the cap is paler, and the wing-bars are pale yellowish instead of ochraceous, as in the birds in fresher plumage. One of the specimens has the cap but little darker than the back and the wingbars yellowish, thus resembling closely the description of L. tristis of Sclater and Salvin from Bolivia.

Six males measure as follows: wing, 2.47-2.67, averaging 2.56; tail, 2.06-2.33, averaging 2.20; culmen, .47-52, averaging .50. Three females measure as follows: wing, 2.40-2.45, averaging 2.42; tail, 2.07-2.09, averaging 2.08; culmen, .50-.52, averaging

.50.

On comparison with Panama specimens of L. pileatus Cab., the differences prove to be almost inappreciable. The brown-headed group of this genus, embracing L. amaurocephalus, L. pileatus and L. tristis, apparently constitute a widely dispersed species, perhaps barely separable into two or three geographical subspecies, the Central American form standing as L. a. pileatus, and the Bolivian form, if really separable, as L. a. tristis.

105. Myiopatis semifusca Sel.-Chapada, 6 specimensApril, 2; May, 1; August, 1; September, 2. The abdomen varies in different specimens from silky grayish white, faintly tinged with greenish, to a deep tinge of pale sulphur yellow.

106. Ornithion cinerascens (Wied)=0. obsoletum auct. (See this Bulletin, II, No. 3, June, 1889, pp. 148, 149.)

This species is represented by 12 specimens, taken at Chapada, as follows: February, 1; March, 1; April, 1; May, 2; June, 3; July, 1; August, 2; September, 1.