IdbT.] PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 87 



But, before proceeding any farther, I may first introduce a synopsis 

 of these five forms in order to make the explanation following more 

 intelligible: 



PROVISIONAL KEY TO THE HAWAIIAN SPECITCS OF THE GENUS 



CRASIEMPI8. 



fl'. Wiug-markiugs pure white. 

 bK A whitish mark above, or in front of, the eye. 



c' . " Upper parts of the body brown." Ch. sandwichensis (Gmel. ). 



c^. Upper parts of the body deep smoky gray Ch. doJei Stejn. [sp. u.]. 



fc2. No whitish mark above, or in front of, the eye ; " forehead and feathers over the 



eye chestnut" Ch. ridgiva yi Steju. [Ibis, 1885, pi. i, fig. 1]. 



a^. Wing-markings tawny, or " ferruginous white." 

 6'. Sides of head, including ear-coverts, uniform bright tawny, without distinct 



superciliary stripe Ch. sclateri Ridgw. 



¥. Ear-coverts dusky, lores and a distinct superciliary stripe pale tawny 



Ch. iiidis Stejn. [Ibis, 1885, pi. i, fig. 2]. 



In the first place, I do not think there are any observations on record, 

 which at the present time justify us in regarding the white-rumped 

 specimens as males and the tawny-rumj)ed ones as females. On the 

 contrary, the only published observation that I am aware of is strongly 

 against such a supposition, for the two specimens collected on Hawaii 

 by the naturalists of the "Challenger" — the form which I call Ch. ridg- 

 icayi — are said to have white rumps and white wing markings, but both 

 are determined as 29 by the collector. It would also seem as if the 

 9 of the pair in tbe Vienna Museum has a white rump (Pelzeln, Ibis, 

 1874, p. 462). Mr. Knudsen's observations in regard to the four speci- 

 mens (two white-rumped Ch. dolei and two tawny-rumped Ch. sclateri), 

 as related in letter to Mr. R. Ridgway, are to the following effect: 



"2 Amakahi [C/i. sclateri \ — all the birds that follow are male and female — . . . 

 "2 Apekepeke [C'/i. dolei'\, also flycatcher, as the above. They live together and by 

 many are considered as the female of Amakahi. 

 These are male and female, as I have seen by the ovary, &c." 



I will suggest the possibility of the tawny-rumped specimens being 

 the younger birds, but until the question be settled one way or another 

 by competent observers on the spot, I feel not at liberty to substitute 

 one uncertain theory for another, and shall therefore keep the two styles 

 of birds apart provisionally. 



This point being decided, there can hardly be any doubt as to the pro- 

 priety of recognizing three different species with white- wing markings 

 (and probably white rumps). We have first the brown and chestnut 

 colored bird from Hawaii, Ch. ridgicaiji, figured on plate i. Ibis, 1885. 

 This bird has the sides of the head entirely dark, " the forehead and 

 feathers over the eye chestnut, and feathers below the eye blackish 

 washed with chestnut," and the color of its breast and flanks is "chest- 

 nut," consequently it cannot be identified with Latham's -^ Sandwich 

 Flycatcher," which he describes as having "the forehead buff'-colored ; 

 over the eye a white line," and "breast rust-color." Then we have the 



