I now simply give the name of the bird as determined under the above rules, and note the occurrence of the bird as verified by myself with the help of the other members of the party—Messrs. W. R. B. Oliver, S. R. Oliver, W. L. Wallace, and C. E. Warden—and especially of the island settlers, Messrs. Roy and King Bell.

The latest complete list of the Kermadec avifauna previous to our visit was that of Mr. T. F. Cheeseman, published in these Transactions, vol. 23, p. 216 et seq., 1891. In that paper no fewer than forty species were totalled, but I was not so fortunate as to collate such a number.

I had instituted comparisons of this avifauna with those of Lord Howe and Norfolk Islands, but deferred making use of my data until such time as further study reinforced my conclusions.

I find that my friend Mr. W. R. B. Oliver had undertaken a résumé of these avifaunas, and from a study of literature arrived at conclusions quite compatible with those of my own, but somewhat different from those he had propounded from his botanical studies alone. I propose here to touch upon these results, as with access to specimens and advice not available to my friend I am able to indicate some improvements and alterations, though in the main I confirm his conclusions.

Firstly, I would reject all the doubtful records which he not only includes but amplifies. Thus, in his paper on “The Geographic Relationships of the Birds of Lord Howe, Norfolk, and the Kermadec Islands” (Trans. N.Z. Inst., vol. 44, p. 215, 1912) Oliver adds six names of birds not hitherto recorded from the group—three from skins received by Mr. Cheeseman from Mr. Roy Bell, the fourth a record by himself (which, however, had been already noted six years previously by Ogilvie Grant, Ibis, 1905, p. 554), and two others on Mr. Roy Bell's authority.

Oliver has fairly well gauged the affinities of the Lord Howe avifauna, and I would add the following information.

Nesolimnas sylvestris Sclater has been shown by Mathews (“Birds of Australia,” vol. 1, p. 191, 1911) to be no relation of the Neozelanic Nesolimnas nor Gallirallus (= Ocydromus), but congeneric with the New Caledonian Tricholimnas lafresnayanus Verr. & Des Murs., a semi-flightless descendant of the north Australian Eulabeornis.

In the same place Mathews (p. 247 et seq.) has gone into the history and examined the supposed specimens of Notornis alba, and has conclusively proven that the reference of the unique specimen of Fulica alba White to the genus Notornis was incorrect, and that the bird was apparently a fixed albinistic species of Porphyrio closely allied to P. melanotus Temminck, and that there can be no good reason for considering it to have any more relationship with New Zealand than with Australia or New Caledonia. I prefer the latter source.

With regard to Cyanoramphus subflavescens Salv., it has more relationship with C. cooki Gray of Norfolk Island and C. saisseti Verreaux, the New Caledonian form, than with the Neozelanic forms. I would consider the whole of the red-fronted parrakeets as representing one species, but that the above three are closer to each other than to the Neozelanic forms, which would include the Kermadec race. I might note here that I have examined all the birds mentioned in this paper in connection with the forms compared, and that herein my conclusions are given to the best of my ability.

Where other workers are mentioned these must be regarded as confirmatory evidence of my own views.

As regards the “two” species of Gerygone, there is only one (Mathews, Novit. Zool., vol. 18, p. 448, 1912) which is more closely allied to New Caledonian forms than to Neozelanic species, whilst the Zosterops has very little affinity to the New Zealand Z. lateralis, which, moreover, can only be considered a recent immigrant to New Zealand, and consequently should not enter into comparisons regarding ancient land connections. Ninox albaria I have not yet seen, and its affinities are not yet well known.

The Rhipidura belongs to the Austro - New Zealand species R. flabellifera. Examination of long series of Australian Rhipidura point to the fact that the New Zealand form is a comparatively recent settler from Australia, as north Australian races differ more from south Australian than the latter do from the Neozelanic race. It would seem that this would also be a dubious factor to use in basal comparisons. The Pachycephala, which Oliver writes is “probably related to Australian forms,” is only subspecifically distinct from the common Australian P. gutturalis, and was only separated as recently as 1898.

Oliver concluded: “Numerically the Australian, New Zealand, and New Caledonian elements in the endemic birds of Lord Howe Island are about equal, or overwhelmingly in favour of a New Caledonia - New Zealand migration as against an Australian immigration. The two flightless rails turn the balance in favour of New Zealand…. The existence of two brevipinnate rails belonging to genera found elsewhere only in New Zealand is sufficient proof of a former land connection with that country.” As the rails are little related to Neozelanic forms, as shown above, I hope that the natural conclusion will be accepted, and that the sentence “Lord Howe Island would therefore properly belong to the New Zealand biological region” will be altered to “Lord Howe Island cannot be accepted as part of the New Zealand biological region.” I have written critically, so that it will be clearly understood that there is, practically speaking, no endemic Neozelanic element in the avifauna of Lord Howe Island. The supposed Neozelanic basal element is purely New Caledonian, and the true relationships of the Lord Howe bird-life will be fully developed in a paper, now in manuscript, by Mr. G. M. Mathews and myself.

In the Norfolk Island avifauna there is undoubtedly present a Neozelanic basis, and this constitutes a most intricate factor in the disposition of this faunula. However, the New Caledonian influence is so predominant that there can be little hesitation in preferring that fauna as being the closest and the most natural to which it might be attached. The only endemic Neozelanic genus worth consideration is Nestor, and against this must be placed Turdus, Diaphoropterus, and Aplonis, genera which occur in New Caledonia but not in New Zealand or Australia. Hemiphaga is very closely related to Carpophaga, and these pigeons are birds of quite a considerable power of flight. The Zosterops are quite a peculiar little group, notable for their very large size. It might be noted that the two species of Turdus on Lord Howe and Norfolk Islands are not what are generally understood as representative forms — i.e., subspecies — but have had a different origin, as have also the species of Gerygone and Zosterops on the two islands. Further evidence will be put forward in the paper noted above, but the conclusions arrived at by Oliver must be revised, and the attachment of Lord Howe and Norfolk Islands to the New Zealand biological

Puffinus pacificus pacificus (Gmelin).

Procellaria pacifica Gmelin, Syst. Nat., p. 560, 1789: Kermadec Islands breeding. Puffinus carneipes Cheeseman, loc. cit., p. 226. P. chlororhynchus, ibid.; Buller, loc. cit., p. 105. P. c. iredali Mathews, Bull. Brit. Orn. Club, vol. 27, p. 40, 1910.

Probably the most abundant bird breeding on the island. Though Hutton, Buller, Salvin, and Goodman all admitted that the Kermadec-breeding bird was quite distinct, it did not receive a name until Mathews, comparing it with the type of P. chlororhynchus Lesson from Shark Bay, Western Australia, named it as above. However, when he came to monograph the petrels in his “Birds of Australia” he conclusively proved that Gmelin's P. pacifica, which had been previously ignored, should be accepted as the species name, and, moreover, that it was best applied to this large Kermadec race (loc. cit., p. 80). This course I fully approve of.

In the paper quoted in the introduction to this account Oliver includes as breeding on Lord Howe, Norfolk, and the Kermadec Islands Puffinus sphenurus. Puffinus sphenurus Gould is an absolute synonym of P. chlororhynchus Lesson, which name must be restricted to the west-Australian-breeding bird.

Mathews has named the east-Australian-breeding form Puffinus pacificus royanus (Birds Austr., vol. 2, p. 85, 1912), and Lord Howe Island birds agree fairly—one Norfolk Island specimen not quite as close; but with these typical Kermadec birds cannot be confused. So that as far as comparisons go the Lord Howe and Norfolk Island breeding forms must not be confused with the Kermadec form, whether they be called P. pacificus or P. chlororhynchus subspp. The name P. sphenurus must not be used.

Pterodroma neglecta (Schlegel).

Procellaria neglecta Schlegel, Mus. de Pays Bas, vol. 6, Procell., p. 10, 1863: Kermadec Islands. Oestrelata mollis Cheeseman, loc. cit., p. 225. Oestrelata sp., ibid. O. neglecta, ibid. O. mollis Buller, loc. cit., p. 112. O. neglecta, ibid., p. 115. O. phillipi, ibid., p. 119.

The synonymy shows the confusion that has existed regarding the forms of surface-breeding petrels living at the Kermadec Islands. In the Emu, vol. 10, p. 13, 1910, I briefly sketched the problem, and my conclusions that only one species without any well-marked varieties could be recognized by me. Further study has not more enlightened me, but has decided me to withhold the exhaustive account of my researches I had drawn up until I feel better able to give some explanation of the anomalies presented.

The disuse of the familiar Aestrelata is due to the investigations of Mathews (Birds Austr., vol. 2, p. 129, 1912). Mathews has also shown that there is (or was) a bird breeding on Norfolk Island very closely allied to the present species, but that O. montana Hull (= O. phillipi (Gray) = O. solandri Gould = P. melanopus Gmelin) may not at present breed on Norfolk Island, and has confirmed my conclusion that the bird is quite different from P. neglecta Schlegel, and has given a full history of the species under the name P. melanopus Gmelin.

Pterodroma cookii nigripennis (Rothschild).

Oestrelata nigripennis Rothschild, Bull. Brit. Orn. Club, No. x, p. lvii, 1893: Kermadec Islands. O. cookii Cheeseman, loc. cit., p. 224. O. nigripennis Buller, loc. cit., p. 113.

This beautiful little bird is more numerous on the outlying islands and on Macauley Island and Curtis Island than on Sunday Island. It breeds

during the summer months, whilst Puffinus a. assimilis only breeds in the winter on Meyer Island, so that Cheeseman's remark, “Breeds… more sparingly on Sunday Island in company with Puffinus assimilis,” needs correction.

Pterodroma macroptera gouldi (Hutton).

Aestrelata gouldi Hutton, Ibis, 1869, p. 351: New Zealand. Oestrelata gouldi Buller, loc. cit., p. 111. O. fuliginosa, id., ib., p. 118. O. macroptera Ogilvie Grant, Ibis, 1905, p. 554; Oliver, Trans. N.Z. Inst., vol. 44, p. 215, 1912.

A specimen washed up on the beach on the 7th August, 1908, proved sufficient for identification. Another bird, too much damaged for preservation, had been noted on the 25th July. However, it had already been added to the Kermadec avifaunal list by Ogilvie Grant, whose specimen, which I have examined, also proves to be a washed-up bird. It belongs to the New Zealand race, so that I do not doubt Oliver's bird is also referable to that form. The typical subspecies inhabits the Cape seas, whilst Mathews (Birds Austr., vol. 2, p. 139, 1912) has described a west-Australian-breeding race as P. m. albani.

Pterodroma externa cervicalis (Salvin).

Oestrelata cervicalis Salvin, Ibis, 1891, p. 192: Kermadec Island. Oestrelata sp. Cheeseman, loc. cit., p. 224. O. cervicalis Buller, loc. cit., p. 114.

A beautiful bird, which appeared to be decreasing in numbers through the ravages of cats, only a few small scattered colonies now, being known. Its closest relative breeds on Juan Fernandez Island, where also a very close ally of P. neglecta is recorded as breeding.

Diomedea exulans rothschildi Mathews.

Diomedea exulans rothschildi Mathews, Birds Austr., vol. 2, p. 246, 1912: Australian seas. D. exulans Cheeseman, loc. cit., p. 224; Buller, loc. cit., p. 128.

A specimen which had been washed up on the beach at Denham Bay previous to our arrival is the basis of this record. At the place quoted Mathews has given a good history of this bird and of the allied species commonly known under the name of Diomedea regia Buller, but which must bear the older name of D. epomophora Lesson.

Onychoprion fuscatus serratus (Wagler).

Sterna serrata Wagler, Naturl. Syst. Amphib., p. 89, note, 1830: New Caledonia. S. fuliginosa Cheeseman, loc. cit., p. 221; Buller, loc. cit., p. 159.

Bred abundantly on Denham Bay beach, and sparingly on the rocks off the north-west corner and on Meyer Island. The rejection of the well-known specific name of fuliginosa Gmelin is unavoidable, as Linné had previously named a young bird from the Island of Domingo, West Indies, Sterna fuscata. This was founded on the Sterna fusca of Brisson (a post-Linnean (1758) non-binomial writer), and Brisson's description and figure are admirable and unmistakable to one who has seen the young, as I have. The South Pacific birds are easily separable from typical birds, and the name to be used is the one I have given.

Procelsterna cerulea cinerea Gou'd.

Anous cinereus Gould, Proc. Zool. Soc. (Lond.), 1845, p. 104: northeast coasts, Australia; Cheeseman, loc. cit., p. 222. Procelsterna cinerea Buller, loc. cit., p. 161.

Bred sparingly on the cliffs at each end of Denham Bay; more commonly on Meyer Island; also at Macauley and Curtis Islands.

Megalopterus minutus minutus (Boie).

Anous minutus Boie, Isis, 1844: north-east Australia. A. melanogenys Cheeseman, loc. cit., p. 221. Micranous leucocapillus Buller, loc. cit., p. 163.

An increasing colony bred on Meyer Island and one of the other outlying islets, though not on Sunday Island. A flight was seen at Macauley Island, but as there are no trees I do not think it was breeding there. This species has been commonly known as Micranous leucocapillus Gould, but Mathews (Novit. Zool., vol. 18, p. 4, 1911) showed that the correct generic name was Megalopterus Boie, which had priority; whilst more recently (Birds Austr., vol. 2, 1912) the same writer has accepted Boie's specific name, which has also priority over Gould's specific name.

Gygis alba royana Mathews.

Gygis alba royana Mathews, Birds Austr., vol. 2, p. 433, 1912: Kermadec Islands. G. candida Cheeseman, loc. cit., p. 222. G. alba Buller, loc. cit., p. 163.

Only bred sparingly round the sea-coast of Sunday Island. A most delightful account of the habits of this bird has recently appeared from the pen of Mr. Roy Bell (Emu, vol. 12, pp. 26–30, 1912), whose photos, considering the difficulties under which he worked, have scarcely been excelled by any bird-observer. A splendid review of the species of Gygis will be found in the place above quoted, where Mathews has separated the Kermadec-breeding bird under the name given. I have carefully considered all the points, and fully endorse Mathews's conclusions.

Pluvialis dominicus fulvus (Gmelin).

Charadrius fulvus Gmelin, Syst. Nat., p. 687, 1789: Tahiti; Cheeseman, loc. cit., p. 220. C. dominicus Buller, loc. cit., p. 174.

This was the wader met with most frequently on Sunday Island, thirteen being noted during September and October. A flock of thirteen waders seen off Macauley Island on the 12th November seemed to consist mostly of this bird.

Eupoda vereda (Gould).

Charadrius veredus Gould, Proc. Zool. Soc. (Lond.), 1848, p. 38: north Australia.

On the 22nd April Mr. W. R. B. Oliver shot a specimen of this bird on Denham Bay beach. Mathews (Novit. Zool., vol. 18, p. 5, 1911) has shown that Eupoda Brandt, 1845, has seven years' priority over the more familiar Ochthodromus Reichenback, 1852.

Numenius phaeopus variegatus (Scopoli).

Tantalus variegatus Scopoli, Del Flor. Faun. Insule, fasc. ii, p. 92, 1786: Luzon, Philippine Islands. Numenius variegatus Buller, loc. cit., p. 181.

One specimen (out of a pair) was shot on the north coast by Oliver on the 24th September; a few days later three similar birds were seen.

Pisobia maculata acuminata (Horsfield).

Totanus acuminatus Horsfield, Trans. Linn. Soc. (Lond.), vol. 13, p. 192, 1821: Java. Heteropygia acuminata Buller, loc. cit., p. 187.

On the 25th October Oliver shot a specimen on Denham Bay beach. Mathews (Novit. Zool., vol. 18, p. 7, 1911) has shown that Sharpe's identification of the Watling drawing upon which was based Latham's T. aurita was purely an error, and hence the usage of Latham's name incorrect. This is the bird Oliver (loc. cit., p. 221) catalogues under the name Erolia aurita. The generic name Erolia was introduced for the curlew sandpiper, and the present species cannot be considered congeneric with that bird.

Anas superciliosa Gmelin subsp.?

Anas superciliosa Gmelin, Syst. Nat., p. 537, 1789: Dusky Sound, south New Zealand; Cheeseman, loc. cit., p. 221; Buller, Suppl., vol. 2, p. 5, 1906.

This bird was constantly seen on the crater-lakes, and, although noted all the year round, was not observed to breed. I do not know whether the birds were referable to the New Zealand or Australian form.

Sula dactylatra personata Gould.

Sula personata Gould, Proc. Zool. Soc. (Lond.), 1846, p. 21: north Australia. S. cyanops Cheeseman, loc. cit., p. 223; Buller, loc. cit., p. 49.

Did not breed on Sunday Island; a couple bred on Meyer Island; plentifully on one of the less islets, hence known to the settlers as Gannet Island. A fair number breed on Macauley Island and Curtis Island. Mathews (Novit. Zool., vol. 18, p. 9, 1911) has shown that the species name to be used is dactylatra Lesson, that name having six years' priority over the more familiar cyanops, both being given to Ascension Island, Atlantic Ocean, breeding birds.

Phaethon rubricauda novaehollandiae Brandt.

Phaethon novaehollandiae Brandt, Mem. Acad. Sci. St. Petersb., ser. 6, vol. 5, pt. 2, p. 272, 1840: Lord Howe Island. P. rubricauda Cheeseman, loc. cit., p. 223; Buller, loc. cit., p. 53.

Breeding sparsely all round Sunday Island, and more plentifully on Meyer Island. Rothschild separated (Avifauna Laysan, pl. 3, p. 296, 1900) the Kermadec, Lord Howe, and Norfolk Island breeding birds under the name P. rubricauda erubescens. Mathews (Novit. Zool., vol. 18, p. 243, 1912) has shown that Brandt had previously introduced the name here accepted for a young bird described by Latham from a drawing made at Lord Howe Island. The subspecies is well differentiated by its larger size and brighter coloration.

Circus sp.

Circus gouldi Cheeseman, loc. cit., p. 218; Buller, loc. cit., p. 54.

There cannot be much doubt that the harrier which frequented Sunday Island during the winter months was referable to the species Circus approximans Peale, but to which race cannot be decided without study of specimens. The species has a wide range over Australia, New Caledonia, Fiji, and New Zealand.

Mathews (Novit. Zool., vol. 18, p. 10, 1911) has noted that Circus approximans Peale, 1848, given to the Fijian race, has priority and must be used as a species name, whilst C. gouldi Bonaparte, 1850, can be utilized, but restricted for the east Australian form. It will be most interesting to learn from which source come the birds which travel to and from the Kermadecs.

Cyanoramphus novaezelandiae cyanurus Salvadori.

Cyanorhamphus cyanurus Salvadori, Ann. Mag. Nat. Hist., ser. 6, vol. 7, p. 68, 1891: Raoul Island, Kermadec Group. Platycercus novaezealandiae Cheeseman, loc. cit., p. 218. Cyanorhamphus cyanurus Buller, loc. cit., p. 87.

A parrakeet bred on Meyer Island; it very rarely occurred in the autumn on Sunday Island. On Macauley Island also not uncommonly was seen a similar parrakeet. Salvadori named a bird procured by Macgillivray at Raoul Island C. cyanurus. The wing-measurement is there given as 6.6 in. This is copied into the Cat. Birds Brit. Mus., vol. 20, p. 587, but reference to the type specimen shows this to be a misprint for 5.6 in. In the Cat. Birds Brit. Mus. the habitat of cyanurus is given: “Raoul Island, of the Kermadec Group, and perhaps also Sunday Island (where it has been exterminated by wild cats), Meyer, and Macauley Islands (Cheeseman). Whether the birds from Sunday, Meyer, and Macauley Islands, also from the Kermadec Group, belong to the same species as those from Raoul Island remains to be ascertained.”

Of course, Raoul Island is Sunday Island, and the bird so labelled by Macgillivray was probably procured on Meyer Island.

Buller's accounts in the Supplement are too confused to be intelligible. On p. 84, under C. novaezealandiae, he wrote: “Specimens brought from Macauley Island, in the Kermadec Group, do not differ from the New Zealand bird…. I have examined a caged parrakeet, brought by Mr. Ernest Bell from Curtis Island, situated a few miles from Sunday Island, in the Kermadec, Group, where also this parrakeet was abundant till the introduction of the domestic cat, which soon killed it off. I can detect no difference from the New Zealand bird…. Macauley Island, where a distinct species closely allied to C. novaezealandiae is said to exist, lies about a degree distant from Sunday Island.” Then on the following page is written: “Mr. Bethune [sic], of the ‘Hinemoa,’ declares that the Macauley Island parrakeet is quite distinct from the one inhabiting Sunday Island. He says he could readily pick a specimen out of a hundred of the others…. Mr. Cheeseman records that he found this parrakeet very plentiful on Macauley Island…. He states that all his specimens were larger than New Zealand ones, but that he could not detect any other difference. It is highly probable, therefore, that this was C. cyanurus, which differs only in having the tail of a bluish hue.” Then on p. 86 occurs: “That from Macauley Island (Kermadec Group), of which several were

Daption capensis Linné.

Diomedea melanophrys Boie.

Phoebetria fuliginosa Gmelin.

Under the above names Cheeseman (loc. cit., p. 224) records having observed petrels at sea in Kermadec waters. They were not noted by any member of our party.

Mathews (Birds Austr., vol. 2, 1912) has shown that the correct name for the bird noted above as Diomedea melanophrys Boie is Thalassarche melanophris impavida Mathews; whilst until specimens are procured we can only guess at the identity of the last-named. The same author has carefully reviewed the literature, and by means of ample material has shown that two distinct species have been confused under the name P. fuliginosa Gmelin: that the correct name for one is P. palpebrata Forster, of which P. fuliginosa Gmelin is an absolute synonym; whilst the other must bear the name P. fusca Hilsenberg. The New-Zealand-breeding races have been named by Mathews Phoebetria palpebrata huttoni and Phoebetria fusca campbelli. Which of these two occurs in Kermadec waters is problematical, the probability being the latter.

Sterna caspia Pallas.

Two years ago I made the following comment in manuscript: “Cheeseman (p. 221) included this bird on the authority of Mr. Bell as having been noted. No use can be made of such records, as we are unable to judge whether the bird seen (!) was referable to the New Zealand or Australian subspecies, whilst it is only guesswork to have called it S. caspia at all!”

Oliver has now recorded Sterna bergii (p. 215), and has omitted S. caspia altogether, though including doubtful records. It would therefore appear that he has concluded that the record of S. caspia applies to his new record. The nomenclature of the two species has been elaborated by Mathews (Birds Austr., vol. 2, 1912), and the bird commonly known as S. caspia must be called Hydroprogne tschegrava Lepechin, the Australian subspecies being known as H. tschegrava strenuus Gould. S. bergii is referred to the genus Thalasseus, and the races discussed, the north Australian race being called Thalasseus bergii pelecanoides King, the south-east Australian Thalasseus bergii poliocercus Gould, and the Fijian form Thalasseus bergii rectirostris Peale. As the species does not occur in New Zealand, the straggler procured must belong to one of these.

Whilst on the subject of terns, I might note that Oliver (p. 220) has included as a visitor to Norfolk Island the New Zealand Sterna albistriata. If this is based on the record in the Cat. Birds Brit. Mus. it would be better expunged, as the specimen which I have carefully examined is in very immature plumage, and it has no history! I myself would not admit it, and Mathews, in his “List of the Birds of the Phillipian Region” (Novit. Zool., vol. 18, pp. 447–52, 1912), has rejected it without comment, as Saunders himself did not believe in it.

Anous stolidus Linné.

At the end of his account of Anous cinereus Gould (p. 222) Cheeseman noted it was probable that A. stolidus might also breed on the group. By a misreading of what was written, Buller included the species in the Supple

ment (p. 162) as breeding on the Kermadecs on Cheeseman's authority, and it has been included in the most recent list of New Zealand birds. But this bird does not yet breed at the Kermadecs, and has been omitted by Oliver without explanation.

Limosa novaezealandiae Gray.

Cheeseman (p. 220) included this bird on Mr. Bell's authority. Such a common New Zealand migrant should occur, but we did not meet with it during the time we were on the island. This points to it being but a scarce straggler to the group, and that the group does not lie in the line of its migration. The correct name to be used for this bird is Limosa lapponica baueri Naumann.

Tringa canutus Linné.

Oliver (p. 215) has added this visitor. For a fortnight in September, 1908, I endeavoured, unsuccessfully, to shoot a bird which I thought might prove to be this species. It associated with a small flock of Charadrius dominicus fulvus Gmelin which frequented Denham Bay beach. The correct name of the species is Canutus canutus (Linné) (Mathews, Novit. Zool., vol. 18, p. 5, 1911).

Ardea sacra Gmelin.

Sula serrator Gray.

Admitted by Cheeseman (p. 220) on Mr. Bell's authority. These must be omitted until skins are actually received from the group.

Sula leucogaster.

Phalacrocorax sulcirostris.

These have been added by Oliver (p. 215) on Mr. Roy Bell's authority. As both of my friends know, I do not admit hearsay records, however much faith I have in the observer; consequently these two records are inadmissible in my list of the Kermadec avifauna.

Tachypetes aquilus Linné.

Mr. Cheeseman introduced this (p. 223) on Mr. Bell's authority. No one can possibly separate the two species of frigate-birds on the wing, and they are not easily differentiated when in the hand. Both species have occurred in New Zealand.

Chrysococcyx lucidus Gmelin.

Cheeseman's record (p. 219), upon the same authority as the preceding, needs verification by means of skins. No one can separate the Australian form from the New Zealand forms without careful examination of specimens. The correct name would be Lamprococcyx lucidus Gmelin for the New Zealand bird.

Zosterops caerulescens Latham.

Anthus novaezealandiae Gmelin.

These two species were met with by Cheeseman himself (p. 218). They were apparently only stragglers from New Zealand which have failed to