84 BULLETIN 159, UNITED STATES NATIONAL MUSEUM 



Geographic distribution. — Islands of Banka and Billiton, east of 

 Sumatra. 



12. Kittacincla malabarica tricolor (Vieillot). 



Turdus tricolor Vieillot, Nouveau dictionnaire d'histoire naturellc, vol. 20, 



p. 291, 1818 (based on "pi. 114 des Oiseaux d'Afrique de Levaillant"). 



("iles de la mer du Sud.") 

 [Grillivora] longicauda Swainson, Natural history and classification of birds, 



vol. 2, p. 238, July 1, 1837 (based on "O. d'Af. pi. 114"). (Type locality 



now first designated, Bantam, western Java.) 



Geographic distribution. — Western Java. 



The use of the subspecific name tricolor for this race from western 

 Java requires a little explanation. Vieillot's Turdus tricolor ^^ was 

 based on Levaillant's Oiseaux d'Afrique, vol. 3, p. 67, pi. 114, 1802, 

 with a good description, and its locality was cited as follows: 



"On soupgonne que cette esp^ce se trouve dans les iles de la mer 

 du Sud." 



In the text to Levaillant's plate 114 there occurs this statement 

 concerning the origin of the specimen that formed the basis of Levail- 

 lant's description: "cette espece appartient encore aux climats du 

 sud faisoit partie de I'envoi fait au cit. Gigot-Dorcy par M. Wood- 

 fort." From the text to the preceding plate we learn that the bird 

 thereon figured came from "des iles de la mer du Sud." The state- 

 ment in the text to plate 114, page 67 — "aux climats du sud" — was 

 therefore merely another way of saying "iles de la mer du Sud." 

 This being the case, it seems logical to consider that the type locality 

 of "le merle tricolor a longue queue" of Levaillant, plate 114, which is 

 in turn the basis of Turdus tricolor Vieillot, is some one of the islands 

 in the southern seas from the Indian Ocean to the Pacific Ocean, all 

 of which area was formerly" included in the expression "south seas." 

 From this viewpoint Hartert ^* was hardly within his rights of subse- 

 quent designation of the type locality as "India," which can not be 

 considered as one of "des iles de la mer de la Sud." 



It seems, therefore, that Kloss '^ has properly restricted the t3'^pe 

 locality to Bantam, West Java, which island is the one most likely to 

 have been the origin of the specimen used by Levaillant as the basis 

 of his description and plate. The name Turdus tricolor thus becomes 

 available for the subspecies from western Java, as above indicated. 



13. Kittacincla malabarica javana Kloss. 



Kittacincla malabarica javana Kloss, Journ. Federated Malay States Mus., 

 vol. 10, pt. 3, p. 210, June, 1921. ("Karangbolang, South Coast of Mid- 

 Java.") 



Geographic distribution. — Central southern Java. 



" Nouveau dictionnaire d'histoire naturelle, vol. 20, p. 291, 1818. 



2' Nov. Zool., vol. 9, no. 3, p. 572, Dec. 16, 1902. 



« Journ. Federated Malay States Mus., vol. 10, pt. 3, p. 210, June, 1921. 



