Remarks.—No specimens have been examined by me. The island of Agiguan is a very small one lying offshore from Tinian and not far from Saipan, where A. l. luscinia occurs. A. l. nijoi is given tentative recognition, on the basis of the measurements of the five adult specimens given by Yamashina. These indicate that the population has a distinctly shorter bill.
Evolutionary history of Acrocephalus luscinia.—The species of Acrocephalus in Micronesia and Polynesia have received several taxonomic treatments. In regard to the Micronesian forms, Quoy and Gaimard called the population at Guam Thryothorus while Kittlitz called the population in the Carolines, Sylvia. Evidently to emphasize the distinctness of these two birds, Reichenbach in 1850 renamed the bird in the Marianas as Hybristes and the bird in the Carolines as Eparnetes. The birds were later placed in the genus, Tatare, by Hartlaub, Gray, Sharpe and other workers. Gray also used the name, Calamodyta, for the bird in the Carolines. The generic term, Calamoherpe, was employed also by a number of workers for the Caroline population. Sharpe (1883:525) placed the reed-warblers in the family Timelidae and retained the name, Tatare, for the Micronesian and Polynesian forms. In distinguishing Acrocephalus from Tatare he has the following to say of Acrocephalus: "besides having a much shorter bill, possesses a very much more pointed wing, the distance between the primaries and the secondaries being much more than the length of the hind toe and claw; whereas in Tatare the wing is much more obtuse, and the distance between the primaries and the secondaries is less than the length of the hind toe and claw." More recent authors have followed Sharpe using the generic name, Conopoderas (= Tatare, old name preoccupied). However, Tristram (1883:38-46) regarded the separation of these oceanic forms from Acrocephalus as a taxonomic error. He said that this is "one of the very few links (the others being the solitary Hirundo tahitica and the Merulae) between the avifauna of Oceania and our own; and it has a much wider range east and west than either of the other links, extending from the Carolines in the east to the Marquesas in the west." Mayr has pointed out (orally to the writer) that the separation of the Oceanic reed-warblers from Acrocephalus is an unnatural one, although it is perfectly true that the extreme members (A. caffra and A. l. luscinia) have a very long bill, but forms with shorter bills like A. l. syrinx point to the close affinity between the continental species and these insular birds. This has also been noted by Hartert (1898:58). Mayr (in litt.) comments that "There is no difference between Acrocephalus and Conopoderas in regard to the wing formula, provided that we compare the Polynesian species with the tropical forms of Acrocephalus (such as toxopei and cervinus). The character mentioned by Sharpe is very artificial and merely indicates the difference in the wing between a migrant of the temperate zone and a resident of the tropics. There is no denying that some of the warblers of eastern Polynesia are no longer reed-warblers but have become dwellers of trees and bushes. However, this same tendency prevails among some of the unquestionable species of Acrocephalus (scirpaceus and palustris) and at any rate a slight change in habits is not sufficient for generic separation." Earlier, Mayr (1942b:169) used Conopoderas as one of the several genera that is based on "morphologically distinct geographic forms." The degree of modification that has occurred in these oceanic reed-warblers, would, if the birds were in a continental area, undoubtedly be considered worthy of specific or even generic rank by some authors; however, as Mayr (1942b:162) points out, "the majority of well-isolated subspecies have all the characters of good species and are indeed considered to be such by the more conservative systematists." Owing to their differentiation, the Micronesian and Polynesian reed-warblers might not be considered by some ornithologists as belonging to a single superspecies; however, all evidence seems to point to the origin of this group by a single invasion from Asia."
Tristram (1883:41) was the first worker to recognize the relationship of the Micronesian and Polynesian reed-warblers to the continental forms, when he placed them within the genus Acrocephalus. Rothschild (1893:2) further stated, "Tatare cannot be separated generically from Acrocephalus." In discussing the status of the Hawaiian species, A. familiaris, Hartert (1898:58) also follows this treatment. Bryan (1941:187) also comments on the relationship of the "miller" birds of Laysan and Nihoa to species at Guam, Christmas and other islands of the Pacific.
The reed-warblers of Polynesia and Micronesia represent an ancient invasion from Asia. The continental form, Acrocephalus arundinaceous, is apparently closest to the ancestral stock of these oceanic birds. This species resembles the oceanic populations in size, general coloring, shape of bill, and wing and tail structure. Some of the continental races of this species have a shorter first primary which is similar to that in the oceanic forms. How rapid the spread was of the reed warbler through the large insular area that it now occupies is unknown. A. syrinx of Micronesia has a shortened wing and some populations have a long bill. Species in Polynesia have stronger wings than the one in Micronesia, but have become differentiated in other ways, as, for example, by the long bill of A. caffra and the small size of A. aequinoctialis. In addition, call notes have become varied, as noted by Chapin (in Mayr, 1942b:54). Also certain of the reed-warblers have become bush and tree-living birds. The Hawaiian birds are reduced in size and have become tree-living in a manner similar to that of other Polynesian species. These modifications of the reed-warblers of the Oceanic area appear, according to Murphy and Mathews (1929), to indicate their long-time residence in the islands, as compared with subspecies of A. arundinaceous that are found in Melanesia. The latter birds, which are not ancestral to the Polynesian birds, resemble closely their Asiatic ancestors and have also retained their swamp-living habits. This would seem to indicate that the birds in Melanesia may be of more recent occurrence. Stresemann (1939b:324) presents a map of the distribution of A. arundinaceous in southeastern Asia and adjacent islands. The original stock came from a point in China, north of Indochina, spreading to the Philippines and to Celebes, from where it reached the Solomons and New Guinea via the Lesser Sundas and Australia.
Fig. 15. Geographic distribution of Acrocephalus in the Pacific area and routes of its dispersal. (1) A. arundinaceus; (2) A. luscinia; (3) ranges of A. atypha, A. caffra, and A. vaughani; (4) A. aequinoctialis; (5) A. familiaris.
The path of invasion of Oceania by the reed-warbler is pictured in [figure 15]. Probably the birds became established in Micronesia by an invasion from the Bonins, where A. arundinaceus orientalis is known to occur today. From the Marianas and Carolines, the birds spread to Polynesia; A. l. rehsei of the Gilbert Islands (Nauru) might well be a connecting link. Possibly, the Hawaiian birds came as a separate invasion via the Volcano and Bonin islands or through the Micronesia Chain, or through the Line and Christmas islands from the south. It seems evident, however, that owing to their geographic proximity and comparative structural similarity, the species in Hawaii is closest to A. luscinia of Micronesia. The absence of reed-warblers from the western Carolines and Palaus seems to reduce the possibility of an invasion from the Philippine region. However, reed-warblers are absent from the Marshall and the northern Gilbert islands, where there is undoubtedly suitable habitat for their occurrence. Possibly these islands were once occupied by the birds but they were eliminated by natural causes or by man and his land uses.