New Zealand Flycatchers of the Genus Petroica Swainson Part I

Transactions and Proceedings of the Royal Society of New Zealand, Volume 78, 1950, Page 14

New Zealand Flycatchers of the Genus Petroica Swainson Part I By C. A. Fleming, Wellington [Read before Wellington Branch, October 25, 1949; received by the Editor, March 24, 1948; issued separately, February, 1950.] Abstract Field observations and study of skins have led to the proposal of a revised classification of the New Zealand Tomtits and Robins, previously listed as species and subspecies of the genera Petroica or Myiomoira, Miro and Nesomiro. The tits are considered subspecies of Petroica macrocephala (Gm.), of which the Snares Island Black “Robin” is shown to be a melanistic race. Miro is recognised as a subgenus of Petroica to include australis Sparrm. (three races, main islands of New Zealand) and the Chatham Island traversi Buller, a degnerate melanistic form with no close affinity to the Snares Island bird, with which it had been associated. The affinities of P. macrocephala are with the Australian and Pacific P. multicolor, and Miro also arose from Australian Petroica at an earlier date: both stocks have undergone specialisation and loss of dispersal ability after invading the New Zealand area. Much of the subspeciation has occurred in post-glacial time, and no great geological antiquity is attributed to the earliest invasion of New Zealand by ancestors of the Robins, which nevertheless constitute one of the “older elements” in the New Zealand avifauna. Contents Part I Introduction History Taxonomic Characters Material Method Acknowledgments Systematics. Fam. Muscicapidae. Genus Petroica Swainson Petroica macrocephala (Gm.) P. macrocephala macrocephala (Gm.), Yellow-breasted Tit. P. macrocephala toitoi (Less.), Pied Tit P. macrocephala chathamensis n.subsp., Chatham Island Tit. P. macrocephala marrineri (M. and I.), Auckland Island Tit P. macrocephala dannefaerdi (Roths.), Snares Black Tit. References Part II Sub-genus Miro Lesson Petroica (Miro) australis (Sparrm.) P. (Miro) australis australis (Sparrm.), South Island Robin P. (Miro) australis steadi n.subsp., Stewart Island Robin P. (Miro) australis longipes (Less.), North Island Robin P. (Miro) traversi (Buller), Chatham Island Black Robin The Phylogeny of the New Zealand Tits and Robins 1. Correlated Characters 2. Evolutionary History 3. Dates in the Phylogenetic History References

Introduction One of the results of the war-time occupation of the Auckland Islands by coast-watching parties constituting the secret “Cape Expedition” was the acquisition by scientific members of the units of observations and specimens, permitting for the first time an adequate comparison of the resident Petroica with other races of the New Zealand Tomtit. In the course of field work at the Auckland Islands in 1942–43, I found that the local tomtit lacks the marked sexual dimorphism which characterises other New Zealand races, the female resembling the male so closely that discrimination is difficult. This discovery, not hinted at by previous visitors to the Aucklands, nor by Mathews and Iredale (1913) when they named the bird, stimulated investigation of the systematics of the New Zealand tomtits. In 1944, Dr R. A. Falla spent a few hours on Snares Island, and found that The Snares Black “Robin” (“Miro” dannefaerdi Roths.) has the appearance, behaviour and song of a tomtit (Petroica). Previously, systematists had considered the Snares bird closely related to the Chatham Island Black Robin (Miro traversi Buller). “With field experience of the latter bird on its island refuge, Little Mangare, I had no difficulty in reconciling its behaviour and song with closer affinity to the New Zealand Robins (Miro) than to the tits (Fleming, 1939, pp. 507–8). Investigation of the affinities of the two melanistic forms led to a reassessment of the characters of Miro and Petroica, and, eventually, to a review of the other species of Miro. The work was originally planned with Dr R. A. Falla, and, although this paper appears under a single authorship, I have had the benefit of frequent consultations with Dr Falla, who has offered valuable suggestions, data, and criticism. My participation in the 1947 Snares Islands Expedition, after this paper had been written, allowed me to gain personal experience of the Black Tomtit in the field, and a number of obserations and a photograph have been incorporated. Finally in October, 1948, I examined certain skins at the British Museum (Natural History), and this necessitated a revision of the nomenclature of the Chatham Island Tit. History The pioneer explorers and naturalists found in New Zealand two types of essentially similar insectivorous birds which the settlers later distinguished as “Robins” and “Tits.” The Robins are larger birds with similar sexes, dark grey with cream breasts and long legs; the Tits smaller, glossy black with yellow or white breasts, and with distinct brown-backed females. The application of these Old World bird names in New Zealand was perhaps influenced by Gmelin's description of the Robin as a Turdus, and of the Tit as a Parus, and in the case of the former, the confiding habits and perky stance of the New Zealand bird amply justified the implied resemblance to Erithacus. The names have become established, although the affinities of both are now known to be with the flycatchers. The scientific nomenclature of the group has suffered more vicissitudes than the vernacular. Table I shows some of the stages in the development of the taxonomy of the New Zealand Tits and Robins.

Table i.—summary of the history of the nomenclature of new zealand tomtits and robins Authority Pied Tit Yellow-breasted Tit Chatham Island Tit Auckland Island Tit Snares Island Black Tit Chatham Island Robin South Island Robin North Island Robin G. R. Gray, 1843, in Dieffenbach's Travels in New Zealand Miro toitoi (Garnot) Miro forsterorum n.n. for Turdus minutus Forst. Miro dieffenbachii n.sp. Miro albifrons (Gmel.) Miro longipes (Lesson) G. R. Gray, 1845, in Voy. Erchus and Terror, Birds Petroica toitoi (Garn.) Petroica macrocephala (Gmel.) Petroica dieffenbachii (Gray) Petroica macrocephala (Gmel.) Petroica albifrons (Gmel.) Petroica australis (Sparrm.) W. L. Buller, 1873, History of the Birds of New Zealand (1st ed.) Myiomoira toitoi (Garn.) Myiomoira macrocephala (Gmel.) Myiomoira macrocephala (Gmel.) Myiomoira macrocephala (Gmel.) Miro traversi n.sp. Miro albifrons (Gmel.) Miro longipes (Garnot) O. Finsch, 1872–1874, Papers in Journ. J. Orn. and Trans. N.Z. Inst. Myiomoira toitoi (Garn.) Myiomoira macrocephala (Gmel.) Myiomoira macrocephala (Gmel.) Myiomoira macrocephala (Gmel.) Myiomoira traversi Hutton Myioscopus albifrous (Gmel.) Myioscopus longipes (Garnot) W. L. Buller, 1888, History of the Birds of New Zealand (2nd ed.) Myiomoira toitoi (Garn.) Myiomoira macrocephala (Gmel.) Myiomoira macrocephala (Gmel.) Myiomoira macrocephala (Gmel.) Miro traversi Buller Miro albifrons (Gmel.) Miro australis (Sparrm.) W. L. Buller, 1905, Supplement to the Birds of New Zealand Petroeca toitoi (Garn.) Petroeca macrocephala (Gmel.) Petroeca macrocephala (Gmel.) Petroeca macrocephala (Gmel.) Miro traversi Buller Miro dannefordi Roths. Miro traversi Buller Miro albifrons (Gmel.) Miro bulleri “Sharpe” Miro australis (Sparrm.) G. M. Mathews and T. Iredale, 1913, Reference List of the Birds of New Zealand Myiomoira toitoi (Lesson) Myiomoira macrocephala macrocephala (Gmel.) Myiomoira macrocephala dieffenbachii (Gray) Myiomoira macrocephala marrineri subsp. n. Nesomiro traversi dannefaerdi (Roths.) Nesomiro traversi traversi (Buller) Miro australis australis (Sparrm.) Miro australis bulleri (Buller) Miro australis longipes (Lesson) W. R. B. Oliver. 1930, New Zealand Birds Petroica toitoi (Lesson) Petroica macrocephala macrocephala (Gmel.) Petroica macrocephala dieffenbachii (Gray) Petroica macrocephala marrineri (M. and I.) Miro dannefaerdi (Roths.) Miro traversi (Buller) Miro australis (Sparrm.) Miro longipes (Lesson) G. M. Mathews, 1931, List of Birds of Australasia Myiomoira toitoi (Lesson) Myiomoira macrocephala macrocephala (Gmel.) Myiomoira macrocephala dieffenbachii (Gray) Myiomoira macrocephala marrineri (M. and I.) Myiomoira macrocephala enderbyi Mathews Nesomiro dannefaerdi (Roths.) Nesomiro traversi (Buller) Miro australis australis (Sparrm.) Miro australis bulleri (Buller) Miro australis longipes (Lesson) C. A. Fleming, 1950 (this paper) Petroica macrocephala toitoi (Lesson) Petroica macrocephala macrocephala (Gmel.) Petroica macrocephala chathamensis n.subsp. Petroica macrocephala marrineri (M. and I.) Petroica macrocephala dannefaerdi (Roths.) P. (Miro) traversi (Buller) P. (Miro) australis australis (Sparrm.) P. (Miro) australis longipes (Lesson)

It is not proposed to discuss the changes in nomenclature in detail, but a few comments on the systematic issues involved are necessary to introduce what follows. After the confusion caused by the independent naming of material collected by the early expeditions from that of Cook (1769–77) to the Erebus and Terror (1840–41), the nomenclature became somewhat more stable in the middle of the century. Distinct North and South Island species of Robin and Tit were recognised, and were grouped together, either under Miro (Lesson, proposed in 1830 for the North Island Robin) or under Petroica (Swainson, 1829, genotype, Muscicapa multicolor Gm., the Norfolk Id. Robin). Insular forms from the Auckland and Chatham Islands were lumped with the South Island Tit, the latter having been named (but not adequately characterised) by Gray in 1843 (but see p. 39). In 1850 Reichenbach proposed the genus Myiomoira for the North Island (Pied) Tit, and Myioscopus for the North Island Robin; later writers have maintained the generic separation of the Robins (Miro-Myioscopus Reich.) from the Tits, which were listed as Myiomoira or Petroica (“Pefroeca”) in confusing alternation in accordance with successive authors' opinions as to the importance of the differences between the New Zealand birds and their admitted relatives in Australia and Norfolk Island. Emphasis on the differences between the Robins and the Tits culminated in Finsch's allocation of Myioscopus to the Lusciniidae and Petroica to the Muscicapidae. Even in recent years, when the affinity of the two genera has been admitted, Miro has been separated from Petroica and its allies by the Gerygone group of flycatchers, and by Smicrornis, in the lists of Mathews and Iredale (1913) and Mathews (1931). In 1871 H. H. Travers discovered on Mangare, Chatham Islands. a flycatcher intermediate in size between the mainland species of Miro and Petroica and differing from both in its uniform sooty black colour. Buller (1872) described the novelty as Miro traversi, and Hutton (1872) as Petroica traversi. Finsch (1874) also considered traversi a Myiomoira (i.e. Petroica.) and was followed by Sharpe (1875). In 1888 Reischek captured what he himself called a “black tomtit” at the Snares Islands. This bird, and others subsequently obtained from the same islands, were recorded as Miro traversi by both Buller (1890, 1893)* One specimen, received by Buller “in spirits,” was listed by him (1892) as Rhipidura fuliginosa (Sparrman), but was almost certainly the Snares Island Black “Robin,” as no fantails have subsequently been seen at the Snares or any other Subantarctic Island. and Finsch (1888), but in 1894 the Snares race was distinguished as Miro dannefaerdi by Rothschild, who pointed out constant differences from the Chatham Island Robin, differences which Buller (1906) was reluctant to admit and which Rothschild himself later (1907) thought to merit only subspecific rank. With the wholesale application of trinomial nomenclature to New Zealand birds by Mathews and Iredale in 1913, Miro was restricted to the large main island robins, which were made subspecies of M. australis (Sparrman). Under Myiomoira was listed macrocephala Gmelin, with three subspecies: the nominate race (South Island). dieffenbachii (Gray) [Chatham Islands] and marrineri subsp. nov.

[Auckland Islands]; the North Island toitoi Lesson was retained somewhat inconsistently as a second full species. Nesomiro gen. nov. was proposed for the Chatham and Snares “Black Robins,” which were given trinomial status as subspecies of the former, the genotype, Iraversi Buller. Mathews and Iredale stated that the two forms separated under Nesomiro differ more from Miro than Miro does from Petroica. Their apparently detailed diagnosis is perhaps based on average characters of the two forms which they believed closely allied. Later writers have not questioned the affinity of the two black forms, though they were listed as full species (under Miro by Oliver, 1930, and Fleming, 1939), and under Nesomiro by Mathews, 1931). Perhaps the first suspicion that the two black forms might not be as closely allied as they appear is contained in Mathews and Iredale's (1921) statement that they “appear to have evolved independently on similar lines in achieving an entire black plumage on far distant islands,” but the next sentence shows that these authorities nevertheless believed both the black forms to be related more to Petroica than to the genotype of Miro. To round off the somewhat confused history outlined above, a brief summary of the results of the present investigation may be given in advance of detailed discussion. The New Zealand tits are grouped as a polytypic species in the Australo-Polynesian genus Petroica; they are considered quite closely related to the type species of that genus, with which they still retain an allopatric relationship, and with which they share important characters and tendencies such as heterogynism, progressive and retarded plumages, variation in tail pattern, etc. The Snares Island Black “Robin” (“Miro” dannefaerdi Rothschild) falls easily into place as a melanistic subspecies of Petroica macrocephala, with little, other than its blackness, to separate it from the other subantarctic form, P. m. marrineri. The robins have had a much longer history of deviation from other Petroica. Under Miro as a subgenus of Petroica I include not only the robins of the main islands of New Zealand, as races of australis, but, also the allopatric Chatham Island Black Robin (M. traversi), which is a melanistic form, in part primitive, in part degenerate, sharing the distinctive characters of the mainland races but separated from them by a much greater morphological and phyletic gap than they are from each other. The taxonomy adopted, adequately expresses the relationships of the New Zealand robins and tits without the use of the names Myiomoira and Nesomiro either as genera or subgenera. The subgeneric separation of the two superficially similar melanistic forms is the chief innovation in the present arrangement, which, it may be again noted, springs from, and confirms, the conclusions derived from study of the live birds in the field. Taxonomic characters Characters of individual variation, secondary sexual characters, and characters of stages in ontogeny are essentially the same as those distinguishing subspecies, species, and genera. In Petroica few morphological characters may be designated solely “specific,” “subspecific,” “generic,” “sexual,” or “juvenile,” Among New Zealand

species the following characters distinguish taxonomic categories of varying rank: size, proportions of appendages, primary formula, number of cmarginate primaries, relative lengths of primaries, extent of tail graduation, dorsal and ventral plumage, colour, size of frontal spot, tail and wing pattern. Normally immature characters of Petroica (streakiness, ill-defined breast pattern) are subspecific characters of the adults of Miro. On the other hand, normally adult characters are accelerated and occur in nestlings of some subspecies (marrineri). Sexual dimorphism, well marked in “normal” forms, is minimised in the melanistic races and in marrineri, in which race the female has advanced, male-like plumage characters, some of which occur as rare mutants in female chathamensis. The opposite tendency, loss of pronounced sexual dimorphism by suppression of “male” characters, has been recorded in Pacific forms of P. multicolor (Mayr, 1934), and occurs in the subgenus Miro. It is not difficult to apply the polytypic species concept to Petroica macrocephala, since the five forms are allopatric and close morphologically, forming a more or less continuous series in many of their characters (see, for instance, Fig. 10). So, too, the mainland robins are obviously geographic representative forms of a polytypic species. Miro traversi of the Chatham Islands is, however, a further allopatric Robin and falls into the class of allopatric forms which are morphologically so distinct that it must be inferred that their divergence has gone beyond the specific threshhold. The New Zealand and Chatham Robins form a “superspecies” (Mayr, 1931, p. 2), and this conclusion is crystallised in the nomenclature by the recognition of Miro as a subgenus of Petroica. Material Over 310 study skins in New Zealand collections have been examined in the course of this revision. This total represents an average of 34 specimens for each of the 9 forms recognised. Though this amount of material is small compared with that used in the comprehensive revisions of Ticehurst (Phylloscopus, 1938) and Miller (Junco, 1941) it compares favourably with the number available to other recent reviewers of passerine genera (e.g. Mayr and Serventy, 1938, The Genus Acanthiza, 17·5 specimens per subspecies). On the other hand, the large number of skins in the New Zealand collections lacking essential data is a constant irritation and the large public collections are worse offenders in this respect than are the much smaller private collections. For example, 43% of skins of Miro and Petroica in one of the largest museum collections are deficient in data for either sex, date, or locality, whereas the same figure for the largest private collection examined is only 9%. Of primary types, I have seen the types of Miro dieffenbachii Gray. Myiomoira macrocephala marrineri Mathews and Iredale, and M. m. enderbyi Mathews (British Museum) the syntypes of Miro traversi Buller, collected by Travers in 1871, which are in the Dominion Museum, and the syntypes of Miro bulleri Buller, which are in the Canterbury Museum. New Zealand collections contain more or less exact topotypes of some of the named forms. The exceptions are Turdus australis Sparrman from Dusky Sound and Muscicapa

longipes Lesson and M. toitoi Lesson, both from the Bay of Islands, where they have long been extinet. The last record of toitoi in the Bay of Islands is of a pair collected by Pycroft (Buller, 1905, p. 114), but these have not been traced. The Australian Museum, Sydney, and Dr D. L. Serventy arranged loans of representative skins of some of the Australian species of Petroica and allied forms, and other Australian specimens of Petroica were available in Auckland and Canterbury Museums. In Auckland Museum, too, there are representatives of Petroica multicolor multicolor (Norfolk Id.) and of the Fiji and Samoa races (kleinschmidti Finsch and pusilla Peale) which proved useful. The synonymies and references in A List of the Birds of Australasia (London, 1931), by G. M. Mathews, have been my guide to the intricacies of the early nineteenth century nomenclature, although most of the important publications after 1843 have been seen. Some of the systematic work of G. M. Mathews may call for revision in the light of field and cabinet studies, but the summaries of bibliographical research recorded in his lists are of immense help to one attempting systematics with inadequate library facilities. Method The problem was to assess the affinities between the various New Zealand tits and robins and their extralimital relations, by finding the characters which distinguish, and the range of variation within, each natural unit; and then to use the information to arrange them into the simplest taxonomic system capable of expressing their inferred phyletic relationship. Small series of skins representing each population were taken as an arbitrary standard sample, examined and compared in detail, and the similarities and differences between them tested on larger series as opportunity offered. As a result, some characters were found to vary indiscriminately, while others were found constant for each sex, age, race, or species. The conception of what races, species and genera were to be recognised changed as the work proceeded. Each specimen has been measured for wing, tail, and tarsus, and some also for total length, exposed culmen, bill from skull, and midtoe and claw: these last were of little use in discriminating subspecies, but useful in analysing differences in proportions. Tail/wing and tarsus/wing percentages were worked out for most specimens and also recorded in “scatter-diagram” form. Characters such as primary formula, relative length of first and second primaries, and extent of emargination, which have been found to distinguish the species and subgenera rather than subspecies, have not been recorded for every specimen, since examination of a representative series showed a limited amount of individual variation. Plumage characters were examined in relation to the arbitrary standard sample (which was not always a typical sample), and notes made of such items as colour of dorsal and ventral surface, size of frontal spot, extent of light shaft-streaking, presence or absence of white-tipped secondary coverts, and detail of tail pattern. When the limits of each race had been found from adequately labelled material, many of the unlabelled and poorly labelled skins

could be located with certainty in their correct subspecies. It is always possible to distinguish dannefaerdi from traversi even when skins of the former have been labelled in error, “Chatham Island Robin”; bleached males of macrocephala can be distinguished from toitoi by dimensions; females of chathamensis from those of macrocephala by dimensions, colour and tail pattern. Discrimination of unlabelled Miro australis is made less confidently. Presentation of full tables of measurements and other information for all skins examined is neither expedient nor feasible; on the other hand, it is desirable to have some data on individual specimens to enable future workers to make objective comparisons with my standards. To compromise, I have tabulated a selection of measurements of individuals of each form, but have also expressed the information for each series in terms of means, observed range, and standard deviation (σ). Dimensions are taken as indicated in the Handbook of British Birds (Vol. 1, 1938, pp. xxxiv-xxxv). Where colours mentioned are followed by (R) they have been named by comparison with the plates of the copy of Ridgeway's Colour Standards and Colour Nomenclature (1912) in the Dominion Museum. Primaries are numbered in the usual way, the most distal, short, remex being the first. But in measuring primaries for comparing their lengths, I have inserted a steel tape between the two to be measured, to their bases, instead of measuring from the tips of the longest coverts as is sometimes recommended. During the course of the work, an attempt was made to record fully the present distribution of the New Zealand robins, and the distributional data included in this paper are based on an inquiry conducted throng) the Ornithological Society of New Zealand, the results of which have already been published in the Society's journal (N.Z. Bird Notes, Vol. 2, pp. 24–25, 1946, and pp. 150–151, 1948). Acknowledgments I wish to record my indebtedness to my fellow members of the 1942–43 Auckland Island coast-watching parties for assistance and facilities freely given in the field, and to members of later relief parties who have passed on information. Dr R. A. Falla not only made available the information and material collected during his visit to the Subantarctic Islands in 1943–44, but has helped by discussion and criticism during the preparation of this paper. He has generously allowed me to incorporate the results of his examination of critical skins at the British Museum in 1947. Drs W. R. Oliver, G. Archey, and R. A. Falla provided facilities for study of the material in the Dominion, Canterbury, and Auckland Museums; Major G. A. Buddle and Mr E. G. Turbott located and packed for loan material in the latter institution. The late Mr E. F. Stead, Christchurch, allowed me to incorporate information from his collection. I am grateful to members of the staff in the Bird Room, British Museum (Natural History) for facilities provided in October, 1948.

The Director, Australian Museum, Sydney, kindly arranged a loan of Australian skins, and Dr D. L. Serventy also loaned specimens from the Serventy-Whittell Collection. Mr Michael Sharland measured skins of Petroica multicolor leggei in the Tasmanian Museum, Hobart, and Mr K. A. Hindwood provided important information on the distribution of Australian Petroica. Many members of the Ornithological Society of New Zealand contributed distributional data in response to an inquiry made in New Zealand Bird Notes. I am grateful to Mr A. S. Wilkinson and to Major G. A. Buddle for permission to use their copyright photographs of North Island Robin and Pied Tit. Systematics Family Muscicapidae Subfamily Muscicapinae Mayr (1941, p. 106) has briefly discussed the difficulty of applying the conventional family names to Australasian genera and has advocated the use of “subfamily names to group seemingly related genera.” Petroica and allied Australian genera seem sufficiently close to Muscicapa in structure, plumage and ethology to make the subfamily designation quite a happy one. The point is mentioned in passing because the two groups here discussed have, at one time or another, been placed in other families, have been associated with other Austral flycatchers in a special subfamily, Petroicinae (Mathews, 1920), and, what is of more importance, have been separated, by some reviewers, in different families. Finsch (1872, p. 112, 1874, p. 171) was apparently the first to emphasize the differences between the New Zealand robins and tomtits by allocating the former to the Lusciniidae (“nearest to Erithacus”) and the latter to the Muscicapidae. The resemblance of Miro to Palaearctic robins (Erithacus) extends to posture, disposition, and habits, but is chiefly due to the long tarsus (over 36% of wing length), a feature apparently unique among the flycatchers. Actually Miro and Petroica are so close that the melanistic races of each have been grouped together for years: they are here accorded subgeneric separation. The family separation of Miro from Petroica was accepted by Bowdler Sharpe (1875), and in faunal lists from Buller to Mathews the two are separated by the “Sylviine” genus Gerygone and its allies. Petroica and its subgenus Miro depart from current definitions of the Muscicapidae (e.g. Handbook of British Birds, i, p. 299, 1938) in their relatively narrow bill, moderate to long tarsi, and streaked, rather than spotted, young. Genus Petroica Swainson Petroica Swainson, 1829, Zool. Illustr., (2), 1, p. 39. Type (by monotypy): Muscicapa multicolor Gmelin. Norfolk Island. Myiomoira Reichenbach, 1850, Syst. Avium, plate lxvii. Type (by monotypy): Muscicapa toitoi Lesson. North Island, New Zealand. Subgenus Petroica s.str. Characters: Small and medium-sized flycatchers with a relatively small bill, about as wide as high at nostrils; tarsi weak to moderately strong (23–37% of wing length); wing and tail short, wing usually

with white alar bar; first primary less than 53% second; third to fifth longest, third to sixth emarginate; tail often with light pattern on outer feathers; plumage soft and fluffy; adult males usually with coloured breasts and white, rarely coloured, frontal spot; immature plumage streaked. Range: Australia, Tasmania, mountains of New Guinea, New Zealand and outlying islands. Western and Central Polynesia. (Fig. 3.) Fig. 3—The distribution of the genus Petroica Swains. All described species are indicated on the map, which shows: (a) the concentration of species in south-eastern Australia, and (b) the inferred derivation of forms inhabiting the peripheral islands (from New Guinea east to Samoa and south to the Auckland Islands) from the headquarters of the genus on the continent of Australia. Remarks: The above diagnosis is a modification of that of Mayr, designed to include the genotype, P. macrocephala, and all of the

Australian species currently listed under Belchera, Whiteornis, Littlera, and Erythrodryas. I have, admittedly, examined only a few skins of the Australian forms, and have no experience of them in the field, but am inclined to consider them all congeneric with multicolor. Plumage characters in the above diagnosis do not, of course, apply to the melanistic race. Affinities: Full discussion of the affinities of Petroica is beyond the scope of this paper. None of the other Australian flycatchers seems as close as Miro. Petroica macrocephala (Gm.). New Zealand Tomtits Characters: The five races which make up this species are moderate to large-sized Petroica typical in their general proportions, bill and (except the melanic race) in plumage pattern, with small frontal spots, black upper parts, and pale orange-yellow or white breasts; tail with oblique white bar on 4th to 6th rectrices; primary formula 2 = 7/8 or 2 = 8/9 (usually 4>5>6>3>7>8>2), indicating rounder, shorter wing than genotype (2 = 6/7 or 7/8; 4>5>3>6>7>8>2); first primary 48–55% second; tarsus proportionately longer than in most other Petroica (except nominate race of genotype), see Fig. 9; tail more graduated than in genotype. Soft parts seem to be identical in all races: tarsi dark purplish brown with distal anterior streak of orange; feet bright orange, with purplish brown on phalangeal scutes; iris dark brown; rictal membrane, throat and tongue deep orange tending to crimson; bill brownish black, purplish horn at base. Distribution: North Island, South Island, Stewart Island, Chatham Islands, Snares Islands, and Auckland Islands. Moult and Plumages: There is inadequate data to describe the moult sequence in each race; this composite account, based on information for all races, must serve until comparative studies are possible. An adult female (December 17, Chatham Islands) is renewing central rectrices, 7th, 8th, and 9th primaries, and beginning to replace interscapulars, but December males, in worn feather, show no sign of moult. March males (Stewart Island) have a complete new plumage with full-length tails and remiges, except that 4th and 5th primaries are short; much of body feather is still in sheath. Both sexes were completing moult at Auckland Islands from March to April, and, by May, neither sex showed signs of the recent moult. North and South Island birds (both sexes) are also in fresh completely moulted plumage in May. Between May and December there is increased wear, but no sign of moult. The distinctive fledgling (juvenile) plumages are not represented in collections except between November and February. A female from Enderby Island on January 24 (R.A.F.51) is moulting primaries and body plumage from worn streaked brownish plumage and cream breast, washed with brownish (paler in all respects than November female fledglings), into what appears to be a normal adult female plumage. I take this to be the normal post-juvenile moult, from faded fledgling into first winter plumage, occurring within a few months of leaving the nest. In that case a live bird (Auckland Islands), handled and described by Dr Falla on April 10 as having “forehead, crown, foreneck, and chin flecked with greyish white,” but otherwise like an adult

female, was probably in its first winter plumage. Old World flycatchers (Witherby et al., 1938, pp. 302–9) and other passerines usually retain some of the juvenile feathers in the first winter. Apart from this instance no suggestion of distinctive first winter and first summer plumages has been seen.* But see below, under P. m. dannefaerdi. This is a further possible difference between P. macrocephala and some races of multicolor, in which Mayr has described such plumages, and in one of which breeding in immature (or retarded) plumage has been reported. Habitat: New Zealand Tomtits inhabit a fairly wide range of biotopes in coastal scrub (as on outlying islands from North Auckland to the Subantarctic), in mixed rain forest and Nothofagus forest, and in some areas of cultural vegetation (exotic Pinus forest margins). They do not require large areas of arboreal vegetation, and have persisted in a number of small relict bush and scrub patches far from the larger indigenous forests, as in Southern Hawke's Bay, Wairarapa, Marlborough, and, until recently, Canterbury Plains. This tolerance makes it hard to explain the almost complete disappearance of the bird from the North Auckland Peninsula, an area where several other formerly widespread birds find refuge. General Habits: No detailed investigation of the habits of any of the races of P. macrocephala has yet been made. General and partial accounts have been furnished by Buller, Potts (1870, 1887), McLean (1911, pp. 72–76), Guthrie Smith (1914, pp. 187–199), and Wilkinson (1927, pp. 239–40). The following notes are included because they are relevant to the classification of dannefaerdi as a Petroica,† See also Stead, 1948. and to the affinities of P. macrocephala with P. multicolor. The species is sedentary, undertaking few small-scale and no large-scale seasonal movements; in this it agrees with P. multicolor among Australian species (e.g. Fletcher, 1939, Serventy, 1937) and may be contrasted with phoenecia, rosea, and goodenovii (Alexander, 1932, Campbell, 1937, Chisholm, 1938, Howe, 1932, Lord, 1943, etc.). There is no gregariousness at any time of the year; pairs seem to remain together, in the general area of their nesting territory, all winter (Wilkinson, op. cit.), although this may not be the case if there is any form of altitudinal migration. Petroica macrocephala customarily feeds by darting from a watching perch to pick up insects on the bark or limbs of trees, or, particularly, on the forest floor. Rarely is food taken in the air, this habit of the Muscicapidae having largely been lost in the New Zealand forests, virtually free from predators. Aggressive behaviour towards other males, tolerance of the female and incipient sexual activity have been noted in May (Rotorua and Gisborne) and in every autumn and winter month (Auckland Is., C.A.F.). On the other hand, the pairs were somewhat loosely bound in winter (Auckland Is.). McLean (op. cit.) mentions inter-male tolerance during winter in North Island hill country, and Potts (1887, p. 130) records antipathy between the sexes in autumn and winter. Male song, territorial defence, and sexual display become intense in July-August; nest building began as early as August 22 (1942)

at Auckland Island, and laying a month later (September 20, C.A.F.), but most egg dates are later (October–November), and apparently further broods may carry breeding into the summer (Hen Island, and Chatham Islands, new-fledged young in January and December). Song (male) includes a “somewhat startling ‘see see see”’ call (McLean) uttered during most of the year, but largely replaced by a characteristic high-pitched warbled trill in the breeding season. The latter, variously rendered by McLean (op. cit., p. 77) and Andersen (1926, pp. 61–62), seems to be a true territory-proclaiming song, uttered by males from a succession of perches, often in high trees, in a definite beat, in spring and during building and incubation. It is much reduced during feeding of young and moult, but may be heard, occasionally, throughout most of the year. Though the song is a high pitched whistle, the vocalisations “ti oly oly ho” (Reischek, 1887, p. 185) and “yodi yodi yodi” are not inappropriate (pace Andersen). The female has a call of “see,” soft, and high pitched, and occasionally, when excited, utters a whispered warble similar to the male's song. This female subsong is not confined to the South Island race (as Guthrie Smith averred, 1914, p. 187). Aggressive behaviour towards other tomtits and towards other animals (including man) takes the form of raising and displaying the frontal feathers, wing fluttering, and darting flights with arched wings, accompanied by bill snapping. Display, such has been noted, includes wing quivering by female, chasing “butterfly flights” by male, with wings spread awkwardly, high and forward, and persistent feeding of female by male (Fig. 4). The last continues through incubation and even through feeding of young in nest, when the female may be called off the nest or fed while incubating or brooding. The male shares in feeding young, but usually not in building and incubation. In one case (marrineri) the male was noted to chase away the one male fledgling of a brood of three, while still feeding the females. Nest Site: Usually at least partially sheltered from above, in a hollowed “knee,” trunk, or limb, in a tree-fern head, on a cliff ledge, or, less commonly, open to the sky. Eggs: the eggs of P. macrocephala are ovoid to ovoido-conical, cream, spotted with yellowish-brown to purplish-brown, usually in a denser zone at the larger end. In size they range from about 18 × 15 mm. in toitoi to 23 × 16·5 mm. in dannefaerdi; the eggs of marrineri closely approach, if they do not equal, the latter measurements. Clutch: three to five, often four. According to Howe (1932) the Australian species all lay three eggs. Affinities: Of the other species of Petroica, the genotype, P. multicolor, is closest to the New Zealand species. P. macrocephala shares with multicolor a sedentary habit in fairly humid forested regions, and both species have shown a marked ability for long-range transoceanic colonisation of small islands, which is not a characteristic of any of the other Australian species even though they may be normally nomadic or migratory to a greater extent than is multicolor. The two species share the fundamental plumage pattern of the genus with other species and each has certain distinctive and specialised features, but the other Australian species all diverge in some way or other from

the common denominator in a manner which makes it unlikely that any of them is closer to macrocephala, than is multicolor. The two species are similar in general size and proportions, the larger size of macrocephala being probably correlated with the cooler climate of New Zealand and the relatively long tarsus with “insular” conditions. Within each species a similar type of geographic variation is shown. Thus, smaller races tend to inhabit warmer localities (Bergmann's Law) and a longer tarsus is found in some outlying insular races. Variation in female, rather than in male, plumage (heterogynism), differences in size of frontal spot and intensity of breast colour, and a tendency for the usual sexual dimorphism in plumage to be lessened by the development of “advanced” plumage in females: these are important phases of geographic variation shared in both species. There are however a few discrepancies. The very large frontal spot of multicolor is shared by goodenovii, and the small spot of macrocephala is present in the remaining Australian species. In multicolor however the frontal spot varies geographically, and at its smallest (campbelli) is not much larger than its best development in macrocephala. On the whole it seems reasonable to consider a small frontal spot primitive and to regard the large spot as a secondary development in multicolor and in goodenovii. Orange-yellow soles characterise the feet of macrocephala. Among Australian species rhodinogaster has “soles of feet ochraceous” (North, 1903) and rosea has “feet dark brown, soles yellow” (specimen so labelled in Serventy-Whittell Collection), but goodenovii has dark soles and P. multicolor campbelli has “feet black, soles grey” (Serventy-Whittell Coll.) and so has the New South Wales race, judging from skins. Two skins of the Samoan P. multicolor pusilla Peale and two of the Fijian P. m. kleinschmidti Finsch have been seen; the soles of their feet had definitely been yellow or orange in life and Dr E. Mayr (pers. comm.) has informed me that the San Cristabel birds (P. m. polymorpha Mayr) collected by him had orange soles. A watercolour sketch of a specimen in the flesh of the nominate race of multicolor, made by Mr A. A. Thompson at Norfolk Island in 1944 and lodged in the Dominion Museum, Wellington, shows distinctly ochraceous but not bright orange soles to the feet. Apparently, therefore, pale yellowish or orange soles are normal in the genus and the darker pigmentation of the continental Australian races of multicolor and of goodenovii is secondary.* As the Australian races of multicolor and the species goodenovii share both large frontal patches and dark (grey or black) soles to the feet, they may be more closely related than their difference in size and habitat preference (and habits?) would suggest. P. goodenovii is tolerant of much greater aridity than multicolor and may have arisen comparatively lately by geographic isolation of part of the multicolor population in an arid area, and attained a sympatric relationship by subsequent expansion and overlap. In size goodenovii is as close to the Pacific races of multicolor as they are to continental multicolor. Petroica macrocephala macrocephala (Gmelin). Yellow-breasted Tit 1789. Parus macrocephalus Gm., Syst. Nat., vol. 1, pt. ii, p. 1013 (Queen Charlotte Sound). 1826. Pachycephalus? auslralis Stephens, in Shaw's Gen. Zool., vol. xiii (2), p. 207 (n.n. for macrocephalus). 1843. Miro forsterorum Gray, in Dieffenbach's Travels in N.Z., vol. ii, p. 191 (Queen Charlotte Sound).

1843. Miro dieffenbachii Gray, ib. (“Chatham Islands,” error = New Zealand). 1844. Turdus minutus Forster, Descrip. Animal., ed. Licht., p. 83 (Queen Charlotte Sound). Adult Male: Crown, back, scapulars, rump, upper tail coverts, sides of head, lores, ear-coverts, throat, uppermost breast and rictal bristles, black with slight gloss, bases of feathers slate grey (R), shafts whitish proximally; a small white frontal spot as broad as base of bill and about as high, separated from culmen by a broken line of small black feathers; remiges black, shafts black, margins of webs brownish black, crossed by a sickle-shaped white bar, convex in front, made up of white wedges of increasing size on outer webs of eighth and following flight feathers, about a third the distance from their bases at first, widening to form lenses half the length of the seventh and eighth secondaries; inner webs with similar white areas of increasing size between eighth primary and seventh secondary; lesser, median and greater coverts black, white tips on a few secondary major coverts; under surface of wing deep mouse grey (R) with triangular white bar; axillaries and lining coverts white, lesser under coverts black, some with whitish tips; rectrices black, inner three pairs wholly so, outer three with oblique white bar (Plate 6, Fig. 6); breast orange immediately below line of demarkation with black of neck, grading, ventrally, within a space of a few millimeters, into Naples yellow (R) on the midbreast and ivory yellow (R) on flanks, venter and under tail coverts. (Peel Forest, May.) Adult Female: Crown, back, scapulars, rump, fuscous (R), bases of feathers grey, basal shafts whitish; some rump feathers with dark olive buff tips (R); lores, cheeks, ear coverts brownish olive mottled with olive buff; frontal spot light olive buff, smaller than in male; remiges blackish brown, paler on inner webs, alar bar as in male, but made up of pale ochraceous buff (R) areas on outer web and whitish areas on inner web; upper wing coverts blackish brown, under coverts and axillaries whitish, lesser under coverts mottled white and hair brown; rectrices blackish brown, with oblique bar as in male; throat, breast and belly maize yellow (R), grading into dorsal colouration on sides of neck, into hair brown on flanks, and washed faintly with buff, especially on the breast; under tail coverts cartridge buff (R). (Stewart Island, July.) Fledgling Male: Crown, nape, scapulars, back, rump, dull brownish black, with light buff shaft lines, most conspicuous on head; frontal spot smaller than in adult; upper tail coverts with pale buffy tips; wing and tail as in adult; throat dull brownish black with pale feather shafts; demarkation between neck and breast ill-defined; breast buff yellow grading ventrally into maize yellow (R) and to whitish at venter, mottled in front with subterminal bands of fuscous. (Nelson, February.) Fledgling Female: General pattern and colour similar to adult female, but with conspicuous light shaft areas on fuscous feathers of crown and nape, less conspicuous towards rump; frontal spot virtually absent; cheeks fuscous, mottled with light central areas to feathers; throat mottled fuscous and buff; breast light buff, widely washed and

mottled with drab and buffy brown; belly light buff with ochraceous tinge towards flanks and under tail coverts. (Nelson, undated.) Variation: Colour of breast varies in intensity in both sexes. In males, the palest appear whitebreasted in the field. Some of the variation is due to age of the individual (Stead, in Oliver, 1930) and to seasonal fading. A rare condition is represented by C.M. 0.1080.12, ♂ ad., Otago, in which the breast is salmon (R) above, passing into light ochraceous salmon below. The same phenotype has been recorded in marrineri and chathamensis: apparently the genes controlling it are present in all three races. Occasionally a faint second alar bar is represented by small white spots on the outer webs of secondaries, distal to the main alar bar, and reminiscent of the second bar characteristic of the Australian Scarlet Robin, P. multicolor subsp. Variation in tail pattern chiefly concerns the character of the tip of the fourth rectrix (Table 2). In over 80% of males and in all females examined the inner web of this feather is white to the tip, the conditions tabulated as “clear” and “nearly clear” in Table 2. Table 2—Variation in Tail Pattern In Races of Petroica macrocephala Gm. Subspecies: toitoi macrocephala chathamensis marrineri Sex: ♂ ♀ ♂ ♀ ♂ ♀ ♂ ♀ Number examined: 31 23 29 8 13 9 7 4 4th Rectrices: inner web clear to tip. 11 20 17 8 0 0 0 1 4th Rectrices: nearly clear to tip. 9 3 7 0 0 1 3 1 4th Rectrices: inner web smoky or with incomplete black tip. 11 0 5 0 1 3 3 2 4th Rectrices: inner web with black tip. 0 0 0 0 12 5 1 0 3rd Rectrices: with white spot. 1 1 0 1 0 0 3 2 5th Rectrices: inner web clear to tip. 1 1 1 1 0 0 0 0 Percentage with dominantly clear webs on 4th rectrices. 78 86 5 45 Percentage with dark tipped inner webs on 4th rectrices. 22 14 95 55 Percentage with white spot on 3rd rectrices. 4 3 0 45 Percentage with clear webs on 5th rectrices (as in P. multicolor subsp.). 4 5 0 0 In the remaining 20% of males the black of the outer web encroaches a little on to the inner web (incomplete black tip). One female has a white spot on the third rectrix, a character of marrineri. One of each sex has the fifth rectrix “clear”; this apparently represents an ancestral condition, since the Australian species rosea, phoenecia, and races of multicolor have the inner webs of their patterned rectrices white to the tip. (Plate 6.) Female dorsal plumage varies a little from fuscous to olive brown and mummy brown (R). The frontal spot shows some slight variation.

Albinism and partial albinism have been noted in a number of birds; in one case the flight feathers were unaffected. I have been unable to detect any variation correlated with geographic distribution. The Stewart Island skins are few, and the males in moult, but are adequate to show that any differentiation on either side of Foveaux Strait is slight and probably below the taxonomic threshhold. Nor are such variable features as intensity of breast colour, size of frontal spot, detail of tail pattern, or dimensions related to geography. Table 3—Dimensions of Petroica macrocephala macrocephala Gm. Collection and number Locality Date Sex Wing Tail Tarsus Mid-toe Culmen H.G.D.9 Picton (virtual topotype) ♂ 73 57 23.5 Dom. Mus. Nelson April, 1905 ♂ 72 58 24.5 20 11 C.M.1080.23 Nelson 20/5/31 ♂ 71 54 22 C.M.1080.11 Karamea, Nelson 1895 ♂ 75 57 22 18 10 C.M.1080.24 Blenheim 31/8/35 ♂ 71 56 25 A.M.85.17 Blenheim 14/9/35 ♂ 70 54 22.5 15 12 Aust. Mus. Lewis Pass ♂ 73 55 21 P.392 Peel Forest 21/5/38 ♂ 74.5 59 22 17.5 12 C.M.1080.10 Otago 1891 ♂ 73 52 20 18 10 C.M.1080.12 Otago 1891 ♂ 72 55 23 19 12 A.M.85.14 Invercargill ♂ 73 55 23 17.5 12 P.720 Stewart Island ♂ 72 52 23 18.5 11.5 P.721 Stewart Island ♂ 70 55 24.5 19.5 11 C.M.1080.26 Nelson 21/7/26 ♂ 67 52 23 Dom. Mus. Nelson Mar., 1904 ♀ 71 54 24.2 19 A.M.85.18 Blenheim 14/9/35 ♀ 67 52 22.5 16.5 11.5 Aust. Mus. Mt. Fyffe ♀ 67 51 20 11 C.M.1080.13 Otago 1891 ♀ 70 52 20 17 9 A.M.85.3 Otago ♀ 67 52 22.5 16.5 11.5 Dom. Mus. Stewart Island July, 1899 ♀ 71 56 24 18.5 11 Aust. Mus. Codfish Island June, 1938 ♀ 73 53 21 10 Note: In this and other tables, the following abbreviations are used: A.M. = Auckland War Memorial Museum Collection. C.M. = Canterbury Museum, Christchurch. Dom. Mus.: = Dominion Museum, Wellington. Aust. Mus. = Australian Museum, Sydney (Wardlaw Collection). H.G.D. = Drew Collection, Dominion Museum, Wellington. E.F.S. = Stead Collection (Christchurch). Auck. Univ. = Auckland University College, Dept. of Zoology Museum. P. = Various private collections. Dimensions: Selected individual dimensions are presented in Table 3. The following is an analysis of the dimensions of the whole series examined. Males (26 measured): wing, 70–75 (mean 72·5, σ = 1·54); tail, 52–59 (mean 55·7, σ = 2·12); tarsus, 20–25 (mean 23·1, σ = 1·4); mid-toe, 15–20; culmen, 10–13; bill from skull, 12–14. Females (12 measured): wing, 67–72 (mean 69·9, σ = 1·98); tail, 51–57 (mean 53·4, σ = 1·85); tarsus, 20–24·2 (mean, 22·3, σ = 1·53); mid-toe, 16–19; culmen, 9–13; bill from skull, 11–14. The sexual differences in length of wing and tail are statistically significant (t = 3·95 for wing and 3·4 for tail); see Fig. 5.

Proportions: Tail/wing ratio (100 X tail\wing] in 26 skins, 71·4–81, mean 76·1, σ = 2·72; tarsus/wing ratio (100 X tarsus\wing], 27·2–35, mean 31·7, σ = 2·19. Distribution: South Island of New Zealand, Stewart Island, and outlying islands. Detailed records: D'Urville Island, Pickersgill Island, Picton, The Grove, Mount Stokes, White Bluffs, Kaikoura, Pelorus Sound, Blenheim, Hopai, Dun Mountain, Tophouse, Karamea, Scatter diagram to illustrate sexual dimorphism in mean length of wing & tail in adults of petroica macrocephala macrocephala Fig. 5.

Fig. 1—Subspecific plumage differences in Petroica macrocephala (Gmelin).

Fig. 2—Ventral views of seven New Zealand Tomtits (Petroica macrocephala) to show plumage phases of some of the subspecies. (a) P. m. dannefaerdi, Snares Island Black Tit, adult male. (h) The same, fledgling female (?). (c) P. m. marrineri, Auckland Island Tit, adult male (May). (d) The same, adult female. (e) P. m. macrocephala, Yellow-breasted Tit, adult male. (f) P. m. toitoi, Pied Tit, adult male. (g) The same, adult female.

Fig. 4—Adult male and female of the Pied Tit, Petroica macrocephala toitoi (Lesson), type of Reichenbach's Myiomoira, showing the marked sexual dimorphism in plumage characteristic of most races of New Zealand Tit, and the formal feeding of the female by the male. Photographed at Kapiti Island by A. S. Wilkinson.

Fig. 6—Tail Pattern in Petroica. Upper four figures: races of P. macrocephala. Right halves of upper figures represent variations mentioned in text. Lower three figures: Australian species with related tail patterns.

Fig. 8—The Snares Island Black Tit, Petroica macrocephala dannefaerdi (Rothschild), adult male. Snares Island, Dec. 4, 1947. Photo by C. A. Fleming.

Whakapuaka, Rotokahu (Reefton), Stockton, Akaroa, Christchurch, Mt. Grey, Ashley Gorge, Peel Forest. Mt. Fyffe. Hanmer, Lewis Pass, Brunner, Teremakau, Kumara, Paparoa Range, Kanieri, Ross, Waiho, Weheka, Okarito, Glenary, Waikaia, Dunedin, Catlins, Cromwell, Golden Downs, Wye Creek, Wakatipu, Rees and Dart Valleys, Hilderthorpe, Kononi, Invercargill, Green Island, Puysegur Point, many Fiordland localities. Stewart Island: Ulva, Halfmoon Bay, Pegasus, Bench Island, Lord's River, Codfish Island, South Cape Islands, Moki Island. Solander Island, Foveaux Strait (no specimens seen). In 1840 the yellow-breasted Tit was almost ubiquitous wherever there was arboreal vegetation within its range; though now much reduced, it is still widespread. Most of the above records apply to the period 1930–45. Reports of disappearance in recent years come from the east side of the Alps in Canterbury, where the bird had long persisted in isolated bush patches.* A Christchurch suburban record in March, 1946, is given in New Zealand Bird Notes, vol. 2, no. 3, p. 51 (1947), and Mr R. J. Scarlett supplies a further record (Port Hills Road). Characters and Affinities: The nominate race of macrocephala is also the most generalised, other races being distinguished by various specialised characters, readily derivable from the macrocephala type. So, too, of all the New Zealand forms, this shows most resemblances to the Australian races of multicolor. Of special note are the occasional variants with a second alar bar, with two patterned tail feathers clear to the tip on inner webs, and with uniform salmon breast. Of New Zealand forms, toitoi, the nearest geographically, is closest, agreeing in proportions, female dorsal colour, and general plumage pattern, and differing solely in dimensions, with considerable overlap, and in its lack of lipochrome in both sexes. P. m. chathamensis, with a superficially similar male, differs more fundamentally in its longer tarsus, tail pattern, and female plumage. Petroica macrocephala toitoi (Lesson). Pied Tit 1828. Muscicapa toitoi Less., Man. d'Ornith., vol. i. p. 188 (Bay of Islands). Adult Male: General pattern and colour of upper parts, head, neck, wing, tail as in P. m. macrocephala; white frontal spot perhaps a little smaller; lower breast, flanks, belly, undertail coverts and thighs white, with grey-black bases to the feathers. (Rotorua, May.) Adult Female: Similar to female of macrocephala, but lacking yellow pigment; dorsal colouration tending to olive brown (R) or chaetura drab (R), rather than mummy brown; throat whitish, grading into hair brown laterally and washed faintly with buff, especially on chest. (Rotorua, May, and Wellington, May.) Fledgling Male: Crown, back of head, brownish black with faint fuscous shaft streaks; frontal spot small; back with obscure shaft-streaks; sides of head, lores, ear coverts, brownish black with fuscous centres to feathers; throat black with light shafts and tips to feathers, giving mottled appearance; upper breast cream with hair-brown tips and pepperings forming an indistinct band across the chest; lower breast white with faint pepperings; wing and tail as in adult. (Wairoa, Clevedon, January, 1890.)

Fledgling Female: Frontal spot of negligible size; dorsal plumage warmer in tone than in adult female, with light ochraceous buff shaft streaks from crown to rump; throat and breast widely washed with hair brown and light ochraceous buff on the tips of the feathers, forming an indefinite breast band. (Akatarawa, undated.) Variation: Several males have a faint peppering of white on the tips of black feathers above the demarkation line between black and white on the breast; otherwise they seem good adults. Female dorsal colour varies a little, and separation from females of macrocephala on this feature alone cannot be guaranteed, but the yellow breast of the latter is usually an immediate criterion. Tail pattern varies a little; 60% of males and all females examined have clear or almost clear inner webs on antipenultimate (4th) rectrices (Fig. 6, upper left), other males have that feather “half-wedged,” a condition illustrated by the right half of the upper right figure in Fig. 6. Single examples of each sex have the penultimate (5th) rectrix “clear,” and the 3rd rectrix with a white spot. Albinism and partial albinism are not very rare. There is some suggestion of a weak cline in wing measurements from north (shorter) to south. Six males from Auckland Province average 67·5, five from Wellington 70·1, but the differences are not statistically significant. Table 4—Dimensions of Petroica macrocephala toitoi (Less.) Collection and number Locality Date Sex Wing Tail Tarsus Mid-toe Culmen E.F.S. Taranga (Hen) Island 27/9/03 ♂ 68 47 22 A.M.20.7 Waitakere, Auckland April, 1880 ♂ 66 52 22.5 17 11.5 A.M.20.3 Wairoa (Clevedon) ♂ 69 53 22 18 11.5 A.M.20.5. Wairoa (Clevedon) ♂ 65 52 22 17.5 12 P.482 Mamaku May ♂ 68 52 22.5 17 10 P.29 Rotorua ♂ 69 52 22 E.F.S. Kereru, Hawke's Bay May, 1909 ♂ 69 52 22 H.G.D.58 Wanganui ♂ 67 52 22 H.G.D.199 Longacre, Wanganui ♂ 66 54 21.7 C.M.1081.1 Eketahuna 1891 ♂ 71 49 22.5 Dom. Mus. Wellington May, 1906 ♂ 65 50 21.5 19 10 Dom. Mus. Wellington May, 1906 ♂ 71 53 22.5 18.5 10 C.M.1081.12 Akatarawa 13/10/15 ♂ 73 54 21.5 E.F.S. Silverstream 17/9/09 ♂ 70.5 53 20 A.M.20.4 Wairoa ♀ 65 52 21.5 17.5 12 P.54 Mukatiti, Rotorua May ♀ 65 49 22 18 10 P.483 Mamaku May ♀ 68 47 21.5 18 10.5 H.G.D.59 Wanganui ♀ 65 55 21 Dom. Mus. Wellington May, 1906 ♀ 67 52 22 19 9.5 Dom. Mus. Wellington May, 1906 ♀ 66 51 22 17 10.5 Dimensions: See Table 4 for selected individual measurements. Males (38 measured): wing, 65–73 (mean 69·2, σ = 2·12); tail, 47–55 (mean 52·1, σ = 1·86); tarsus 20–23 (mean 21·8, σ = 1·75); mid toe, 16–19; culmen, 10–12; bill from skull, 12–13.

Females (10 measured): wing, 65–69 (mean 66·8, σ = 1·4); tail, 47–53 (mean 51·1, σ = 1·74); tarsus, 21–22 (mean 21·6, σ = 1·25); mid toe, 17–18; culmen, 9·5–12; bill from skull, 12·5–13. The sexual difference in wing length is significant (t = 4·4). Differences between males of toitoi and macrocephala are significant for wing, tail and tarsal measurements (t = 7, 7 and 3) and so are those between wing and tail measurements of females (t = 4 and 3 respectively). Proportions: Tail/wing percentage in 33 skins, 70–81·7, mean 75·8; tarsus/wing percentage 30–34·1, mean 31·9. In these and other structural details there are no significant differences from the nominate race. Distribution: North Island of New Zealand and its outlying islands. Detailed records: Bay of Islands (before 1900), Waipoua, Huia Valley (Waitakere Range). Hen and Chickens Islands, Little and Great Barrier Islands, Motuihi Island (before 1935)), Wairoa South, Ness Valley, Mount London, Moumoukai Valley, Coromandel Range, Waingaro, Te Aroha Mountain, Pirongia, Whareorino, Aria (and elsewhere in King Country), Rangitoto and Hauhangaroa Ranges, Arataki, Rotorua-Taupo District, Raukumara Range, Urewera Country, Kaingaroa, Egmont National Park (to 4,000 ft.), Waitotara Valley, Pipiriki, Kereru, Waikaremoana, Ruahine Range, Wanstead, Ohakune, National Park, Raetahi, Wanganui District, Rangataua, Manawatu Gorge, Puketoi Range, Wairarapa District, Tararua Range, Rimutaka Range, Orongorongo, Silverstream, Kapiti Island. Virtually ubiquitous in 1840, the Pied Tit is still very generally distributed, though destruction of its habitat has divided its range into many discrete “islands” separated by many miles of country in which no tits are to be found, and which they seldom if ever cross. Such relict populations have a precarious existence if small—the single pair seen on Motuihi, for instance, in 1935, twenty or so miles, mainly across water, from the nearest of their species, had succumbed by 1941 (P. C. Bull, pers. comm.). Similar small populations many miles from the nearest “reservoir” of tits have been reported from a number of places nearer the centre of the island, and in southern Hawke's Bay and East Wellington. Possibly there is occasional reinforcement of such small populations from the nearest bush areas. Mr R. H. D. Stidolph (pers. comm.) found a dead tit in Masterton, at least five miles from their nearest known habitat, and I have a credible report of a male tit being seen in an Auckland suburb about ten miles from Huia, where the species occurs. If some of the small populations are able to survive only by occasional reinforcement from larger bush areas (whether by chance wandering or by overflow of surplus population), it may help to explain why the Pied Tit has so largely disappeared from North Auckland, where there are many areas of patchy bush such as are occupied by tits in the central volcanic region (for instance). In the north there are no large “reservoir areas” from which relict populations of tits in conditions bordering on the limits of viability for the species could be restocked.

On the Hen Island (Taranga), where ecological conditions may approach those of pre-European New Zealand, at least in respect to predators and accompanying bird species, E. G. Turbott (1940) found the population density of the Pied Tit 14 and 16 pairs per hundred acres in forest and second-growth scrub respectively. Characters and Affinities: The characters of the race toitoi are absence of lipochrome pigment and small average size. The North Island Tit, in spite of the superficial difference from the South Island form which has resulted in its being maintained as a distinct species by all previous reviewers, is actually much closer to that form than are the other “yellow-breasted” races. In female plumage tint, dimensions, and proportions, the two main island races are as closely allied as their geographic proximity would lead one to anticipate. Petroica macrocephala chathamensis n.subsp. The type of Miro dieffenbachii, which Gray noted was “found on the Chatham Islands,” turns out to be a specimen of P. m. macrocephala from the South Island. Gray's characterisation of dieffenbachii as a small form (1843, p. 191) is at variance with the known dimensions of the Chatham Island population, as pointed out in 1939 (Fleming, 1939, p. 506), and the selection of a South Island skin as type of dieffenbachii leaves the Chatham Island Tit without a name. Holotype: Adult male skin, Rangatira (= South-East) Island, Chatham Islands, March 15, 1926, Dominion Museum, New Zealand, Register No. D.M.1832. Collected by Whitney South Sea Expedition. Dimensions of Holotype: Wing, 75; tail, 57; tarsus, 26; mid-toe and claw, 19·5; culmen, 12. Adult Male: General pattern and colour as in P. m. macrocephala; line of demarkation between black of neck and orange of upper breast perhaps higher than in that race and rectrices usually with a slightly different pattern (Fig. 6); otherwise to be distinguished by slightly larger size, and especially longer tarsus. (South-East Island, September and December.) Adult Female: Crown, nape, scapulars, back, rump, upper tail coverts sepia (R), much warmer in tone than female toitoi and macrocephala; frontal spot larger than in females of other races, outlined above by a narrow band of brownish black; wing and tail as in other races, but fourth rectrix usually with inner web tipped with wedge of pigment as in male, or at least smoky in that area; throat, breast and belly cream colour, paler than in macrocephala and clearer, lacking the wash of buffy or hair brown of that race. (Normal phase, South-East Island, March, September, December.) Fledgling Male: Crown, back, scapulars, rump, upper tail coverts, sides of head, lores, dull brownish black with light fuscous shafts, conspicuous on crown and scapulars, and with faint ochraceous buff tips, especially on rump and tail coverts; frontal spot white, small; wings and tail as in adult; throat fuscous, mottled whitish; breast, flanks, belly, venter, under tail coverts, white. (South-East Island, December; ventral surface pale yellow in a January skin from Mangare.)

Fledgling Female: Similar to normal phase of adult female, but dorsal colouring Prout's brown to cinnamon brown (R) with conspicuous light ochraceous buff shafts; wing and tail as in adult, alar bar somewhat more buffy; throat white, peppered with brown; breast, belly and venter white, washed with light buff and flecked with drab on flanks and across breast; thighs and under tail coverts pinkish buff. (South-East Island, December; Whaingamarino, December.) Variation: As in other yellow-breasted races, intensity of colour on upper breast varies. In one male, upper breast is salmon orange, rapidly grading into maize yellow (R); in another the tint is between bittersweet orange (R) and flame scarlet (R), grading into pale yellow orange (R), and there are sporadic buff tips on lower throat feathers. A salmon-breasted bird was noted in the field (South-East Island, December, 1937). A number of females depart from the normal condition by exhibiting tendencies towards “maleness” in the plumage of the sides of the head and throat. The feathers of the cheeks and upper neck have subterminal bands of blackish brown of varying width, giving a mottled effect. The tendency culminates in a Dominion Museum skin labelled “Chatham Islands, female, 2/9/1900,” in which the plumage is that of a normal female except that the lores, ear coverts and upper neck are brownish black, with light feather bases in the centre of the throat, and light basal shafts elsewhere (Fig. 1). This extreme form is unique among some score of New Zealand skins and was not noted in the field, but there is another in the British Museum (1903.12.10.13). Table 5—Dimensions of Petroica macrocephala chathamensis Fleming Collection and number Locality Date Sex Wing Tail Tarsus Mid-toe Culmen Dom. Mus. Chatham Islds. Sept., 1900 ♂ 78 58 26 19 11.5 Dom. Mus. Chatham Islds. ♂ 75 57 25 19 11 Dom. Mus. A.242 Chatham Islds. ♂ 76 56 26 20.5 12 A.M.85.5 Chatham Islds. ♂ 76 56 26 20 12 A.M.85.6 Chatham Islds. ♂ 75 60 23 18 13 Dom. Mus. A.240 Pitt Island ♂ 77 56 26.5 22 11.5 Dom. Mus. A.241 Pitt Island 14/9/71 ♂ 76 60 27 19 12 C.M.1080.14 Mangare Island 6/1/24 ♂ 1 mm. 75 55 23 19 10 Dom. Mus. 1832 Rangatira Is. (S.E. Island) Mar., 1926 ♂ 75 57 26 19.5 12 P.314 South-East Is. Dec. ♂ 77 58 26.5 20 12 P.315 South-East Is. Dec. ♂ 74 54 26.5 19 11 C.M.1080.18 Whaingamarino 29/12/24 ♂ 73 56 24 19 10 Dom. Mus. A.244 Chatham Islds. ♂ 75 60 25 19.5 11 Dom. Mus. Chatham Islds. 2/9/09 ♀ 74 58 24 19.5 11.2 Dom. Mus. 1833 Rangatira Is. (S.E. Island) Mar., 1926 71 54 24.5 20 11 Dom. Mus. A.243 South-East Is. ♀ 76 58 26 19.5 11.5 Auck. Univ. South-East Is. ♀ 74 55 24.5 17 12 Note: C.M.1080.14, an immature bird, was not included in the series from which means and extremes were calculated. Dom. Mus. 1833, collected by the Whitney South Seas Expedition, is labelled ♂, but has typical female plumage. Dom. Mus. 1832 is the holotype.

A Dominion Museum skin in normal adult female plumage, collected by the Whitney South Sea Expedition at South-East Island in March, 1926, is labelled “♂”. It would be unwise to consider the male plumage as variable as this skin suggests. The skin is doubtless wrongly sexed. Dimensions: See Table 5. Males (12 measured): wing 74–78 (mean, 76·1, σ = 1·08); tail 54–60 (mean, 57·3, σ = 1·72); tarsus 23–27 (mean 25·8, σ = 1·07); mid-toe 18–22; culmen 11–13; bill from skull, 13–15. Females (15 measured): wing 71–76 (mean 73·3, σ = 1·59); tail 54–60 (mean 56·5, σ = 1·57; tarsus 24–26 (mean 25·3, σ = 0·76); mid-toe 17–20; culmen 10–12; bill from skull 12–13·5. Sexual difference in wing length is significant (t = 5). Differences between chathamensis and macrocephala (both sexes) are fully significant for wing, tarsus, and tail. Proportions: Tail/wing percentage in 28 skins, 70·5–80, mean 76, σ = 2·3. These figures are practically identical with those for macrocephala. Tarsus/wing ratio in same series, 32·5–35·7, mean 34·3, σ = 1·14. The difference between this ratio and that for macrocephala (mean 31·7) is fully significant. Distribution: Chatham Islands: Main Chatham Island, once abundant, now confined to parts of Southern Block; Pitt Island, Mangare, Little Mangare, and South-East Island. Characters and Affinities: The Chatham Island Tit is distinguished from the South Island Yellow-breasted Tit by its female plumage, distinctive tail pattern, slightly larger size, and, especially, by its disproportionately long tarsus (and therefore greater tarsus/wing ratio). The last feature seems to be unique in Petroica, s.str., and comparison of other dimensions shows that it is the tarsus which is larger, and not the other organs that are more reduced, than in other races. Of other races, the South Island macrocephala is closest. Remarks: Gray's original description of Miro dieffenbachii (Dieffenbach's Travels in New Zealand, 1843, vol. 2, p. 191) reads: “This species is very like the preceding, but is altogether of a smaller size and the colour of the chest is darker with the base of the lower mandible pale. Found on the Chatham Islands.” “The preceding” is Miro forsterorum n.sp., with an adequate description of the bird we know as macrocephala, from Queen Charlotte Sound. The fallacy that the Chatham Island race was smaller than the South Island one was perpetuated until 1939 (Fleming, 1939, p. 506). Now, when Gray described and distinguished his Miro forsterorum and M. dieffenbachii he had before him skins from the Chatham Islands and from the South Island (Queen Charlotte Sound) collected by Dieffenbach and Sinclair. He distinguished two species on size and colour, but we may never know why the smaller was attributed to the Chatham Islands, and the larger to the South Island: it may have been due to inadequate or faulty labelling (some of Dieffenbach's Chatham skins were labelled simply “New Zealand”), moulting specimens, or to a lapsus. The dimensions given for forsterorum are: total length 5 ¼ inches; bill 7 lines; wings 1 ½ inch; tarsi 1 inch. Even a moulting or

fledgling bird is unlikely to have a wing as short as 1 ½ inches (38 mm.) and dieffenbachii cannot have been smaller in that dimension. When Gray wrote the appendix to Travels in New Zealand he expressly stated that it was a list of the birds of New Zealand, the Chatham Islands, and Auckland Island, and no tomtit was listed from Auckland Island, so it is unlikely that he was basing his comparison on Auckland Island skins (which are larger than Chatham Island birds), unless he had already seen, but not recorded, the material then recently brought back by the “Erebus and Terror,” or unless some of that material was already in the British Museum, perhaps inadequately labelled: both possibilities are unlikely since the Auckland Island material was registered at the museum less than a month before the appearance of Dieffenbach's Travels. When, however, Gray wrote the report on the “Birds” of the “Erebus and Terror” (1845), he certainly confirmed his ideas on the size differences between dieffenbachii and macrocephala (as Dr Falla pointed out to me) by including the larger Auckland Island birds within his conception of the latter form. Comparison of the two works suggests that the description of macrocephala was amended on the basis of the original Auckland Island skin (wing, 3 ¼ inches; tarsi, 13 lines). By this time, it seems, the small size, rather than the Chatham Island locality, had become the essential character of dieffenbachii, and a type was selected from the original New Zealand Company material. It is a small bird, from “New Zealand,” collected by Dr Sinclair. The writing on the label, including the word “type,” is in Gray's hand, according to Captain C. H. B. Grant, and the ticket is of the style used for the other skins registered in the eighteen-forties. Whatever the date of its selection, it must be considered the valid type of dieffenbachii, because it was listed as such in the Catalogue of the Birds of the British Museum (Bowdler Sharpe, 1879, p. 177, specimen “d”). The specimen is a male skin, with the diagnostic tail pattern of macrocephala (inner web of 4th rectrices white) and with the dimensions: wing, 65 (difficult to measure); tail, 53·5; tarsus, 22·5 mm., which are those of macrocephala rather than of chathamensis. The British Museum registration number is 42.10.30.10. The mounted pair of P. macrocephala listed as a and b by Bowdler Sharpe were not found, and the only other early skin seen is 42.9.29.10, a female skin labelled “New Zealand, New Zealand Co.,” with dorsal plumage tone and tail pattern suggesting that it is a specimen of chathamensis. The history of Miro dieffenbachii emphasizes the change of outlook brought to taxonomy by the recognition of the generally allopatric relationship between closely related forms, and by the adoption of a polytypic species concept. When Gray wrote, the slight morphological differences claimed seemed more important than the differences in locality; the significance of “geographic isolation” was not appreciated. For about 30 years some ornithologists sought and claimed to recognise two sympatric tomtits in the South Island; Finsch, for example, maintained the separation of macrocephala and dieffenbachii in 1870, describing Banks Peninsula skins as the latter species, but synonymised them in 1872. No nineteenth century worker seems to have critically compared the available series of tits from the South

Island and the Chathams. In contrast, when, in 1913, Mathews and Iredale presented a Reference List of the Birds of New Zealand using trinomial nomenclature (applied with an appreciation of what we now call a “polytypic species concept”), they seized on Gray's locality for dieffenbachii and used the name subspecifically for the Chatham Island Tit, but the actual morphological characters of the race were still not investigated. Petroica macrocephala marrineri (Mathews and Iredale). Auckland Island Tit. 1913. Myiomoira macrocephala marrineri Mathews and Iredale, Ibis, p. 436. (Auckland Islands.) 1930. Myiomoira macrocephala enderbyi Mathews, Bull. Brit. Orn. Club, vol. li, p. 21. (Enderby Island.) Adult Male: Plumage generally similar to that of males of the nominate race; frontal spot small; breast paler and lacking the band of intense colour in front; demarkation with throat perhaps lower than in macrocephala and appreciably lower than in chathamensis; breast about maize yellow (R), with subterminal bands of neutral grey peppering feathers of lower flanks; under tail coverts black with faint whitish tips; wing as in other races, but white tips on secondary coverts seldom present; tail with oblique white bar often prolonged on to third rectrices as spot of variable size. (Carney Harbour, May.) Adult Female: Superficially similar to male and thus strikingly different from females of other races; upper parts dull black, blackish brown, or chaetura black (R) instead of glossy black of male; tail coverts with buff tips; throat black, darker than upper parts, giving way to light breast colour at a sharp line on upper breast, as in male; breast and belly white to ivory yellow; wing as in male, but with occasional white tips to secondary coverts, and alar bar with ill-defined edges of light ochraceous buff which dominates over white on outer webs of primaries; tail similar to male. (Auckland Islands, composite description.) Fledgling Male: Differs from adult male in having faint light shaft lines on head and minute fuscous tips to feathers of upper surface from crown to scapulars; frontal spot smaller; breast maize yellow, as in adult, but peppered with blackish just below line between neck and breast, and lacking more ventral pepperings of adult; wing and tail as in adult. (Port Ross, November.) Fledgling Female: Crown and rest of upper surface dull blackish brown with light shafts, particularly on crown; frontal spot small and ill-defined; lores and neck black with lighter centres to feathers; breast ivory yellow with deep greyish olive tips to feathers giving mottled effect; line of demarkation between breast and neck ill-defined; wing and tail as in adult female. (Port Ross, November.) Variation: One male (Near Crozier Point, 4/3/43, No. 124) is similar to the aberrant macrocephala (C.M.1080.12), having a light ochraceous salmon breast, and agrees with the nominate race in having an odd white tipped secondary covert. Guthrie Smith (1936, p. 210) and H. W. Wenham (oral communication) saw similar salmon-breasted birds in the field. The tail is somewhat variable, the spot on the third rectrices being present in about 45% of skins, and a fine dark margiing to the white area on the fourth feather being usually present.

(Fig. 6.) The terminal peppering of ventral feathers on fresh moulted specimens is apparently soon lost by wear. Females, in fresh plumage, vary from blackish brown to an uncompromising black. Both adult and fledgling females become paler with fading and wear, but are at most times difficult to distinguish from males in the field. November males are paler on the breast and back than are some fresh moulted females. No systematic differences have been found in skins from the various islands of the Auckland Group: specimens have been examined from Adams Island and Enderby Island, as well as from the main island. Mathews distinguished Myiomoira macrocephala enderbyi (Enderby Island, New Zealand) from the nominate race on the basis of its longer wing, making no mention of differences from marrineri, either in ignorance of, or indifference to, the geographic issues involved. Enderby Island, Rose Island, and Auckland Island are separated by channels a few chains in width. As shown below, the type of enderbyi is almost certainly a topotype of marrineri, of which the former name becomes a synonym, irrespective of any differences between Enderby Island and Auckland Island birds. Table 6—Dimensions of Petroicu macrocephala marrineri (M. and I.) Collection and number Locality Date Sex Wing Tail Tarsus Mid-toe Culmen C.M. Probably from Enderby In. Jan., 1902 ♂ 81 64 26 22 11.5 C.M. (F. W. Hutton) Jan., 1902 ♂ 78 63 23 12 R.A.F.239 Enderby Island 19/8/43 ♂ 76 60 24.5 19.5 11.5 R.A.F.124 Near Crozier Pt., Auck. Island 4/4/43 ♂ 79 62 26.5 21 12 F.729 Crozier Pt. 19/11/42 ♂ 75 (58) 25 19 11.5 F.725 Carnley Harbour, Auckland Is. 10/5/42 ♂ 78 60 26 18 12 F.727 Musgrave Penin. 16/7/42 ♂ 78 62 26 21 11.5 P.7 Carnley Harbour 22/5/42 ♂ 77 61 25 19.5 12 E.G.T.43 Western Harbour, Auckland Is. 9/9/44 ♀ 78 62 26 R.A.F.51 Enderby Island 23/1/43 ♀ 76 61 25 E.G.T.37 Kekeno Point, Auckland Is. 27/7/44 ♀ 77 60 24 F.726 Carnley Harbour 4/4/42 ♀ 75 60 26 20 11.5 F.730 Carnley Harbour 10/5/42 ♀ 74 58 24 20 11 F.728 Carnley Harbour 10/5/42 ♀ 77 59 25 20 11 P.2 Carnley Harbour 17/4/42 ♀ 76 59 25 20 12 E.G.T.50 Adams Island 25/9/44 ♀ 77 51 26 Note: Skins collected in 1942, 1943, and 1944 have not as yet been allocated permanent collection register numbers, and are designated by the field numbers used by the collectors. Dimensions: See Table 6. Males (9 measured): wing 75–81 (mean 77·7, σ = 1·72); tail 60–64 (mean 61·8, σ = 1·32); tarsus 24·5–26·5 (mean 25·6, σ = 0·69); mid-toe 18–23; culmen 11·5–12; bill from skull 14–15. Females (7 measured): wing 74–77 (mean 76, σ = 1·15); tail 58–61 (mean 59·7, σ = 1·11); tarsus 24–26 (mean 25, σ = 0·81); mid-toe 20; culmen 11–12; bill from skull 14–15. For the whole series, both sexes, the means and standard deviations are: wing 77 (1·7); tail 60·8 (1·6); tarsus 25·3 (0·79). Except that

chathamensis has as long a tarsus, marrineri is larger than the other races so far discussed. Proportions: Tail/wing ratio in 16 skins, 76·5–81·5, mean 78·8, σ = 1·3, i.e. practically identical with that of dannefaerdi, but higher than in all other races. Tarsus/wing ratio, 31–34·5, mean 32·8 (σ = 0·83) is also the same as in dannefaerdi. Distribution: Auckland Islands; littoral forest from sea-level to about 150 m., rarely at greater altitudes, on main Auckland Island, Adams Island, Ocean, Rose, Ewing, and Enderby Islands. Characters and Affinities: The Auckland Island Tit is a distinctive race, only equalled in size by dannefaerdi and differing from other yellow-breasted races in its male-plumaged female, black under tail coverts (present in some chathamensis), and secondary coverts, tail pattern, and tail/wing ratio. Fledgling plumages are equally distinctive, differing from those of other races in being more “advanced,” i.e. more nearly resembling the normal adult male condition. In its advanced plumages, suppression of sexual dimorphism, and increased black pigmentation (coverts, throat, reduced frontal spot), in tail/wing ratio and general size, marrineri shows relationship with the completely melanic dannefaerdi. Remarks: The notes on Auckland Island specimens of Petroica in the British Museum made by Dr R. A. Falla in January, 1947, are here reproduced. No. 42.12.16.48 “Auckland Is.” ♂ Wing 79 mm. 1901.10.21.32 Enderby Is., 14 Jan., 1901 (Ranfurly) ad. ♂ 1903.7.17.3 " " 8 " 1902 ” ad. ♂ Wing 76, Tarsus 26. 1903.7.17.4 " " " "" " " " " 1903.7.17.5 " " " "" " " " " 1903.7.17.6 " " " "" " " " is not sexed, but is a “female-plumaged” female. “The first one is an ‘Erebus and Terror' skin and almost certainly came from Enderby Island also, as McCormick (1884), who mentions all the birds he shot, describes how he got one ‘blackcap’ on Enderby. So marrineri is based on a probable Enderby Island bird and enderbyi at a later date on others in a series that is really topotypical. The only difference left between Auckland Island and Enderby Island specimens is that at least two ‘female-plumaged’ females have now turned up from Enderby. The Enderby males in the above series have pale lemon breasts, and one (1903.7.17.3) is quite bright, although none equal to the brightest from the main island.” One of the “female-plumaged” birds has been referred to above (p. 25); it is moulting into a blacker apparently normal female plumage, and is interpreted as a faded and foxed fledgling. The variation in colour in adult females includes some which are blackish brown when fresh, so that foxing of an immature January female skin, over 40 years old, might readily result in a brownish plumage. I examined this skin in November, 1948, and noted as a character of immaturity the light shafts of the dorsal plumage. The throat is mottled somewhat as in abnormal females of chathamensis. On September 26, 1942, I watched a female in normal marrineri plumage

Fig. 7—Left wings of (a) P. m. macrocephala, (b) P. m. dannefaerdi. (c) P. (Miro) traversi, (d) P. (Miro) australis longipes. P. m. dannefaerdi and P. (Miro) traversi have been considered closely related because of their uniform black colouration, but wing shape, relative length of the first primary, and lack of emargination on the seventh primary of dannefaerdi indicate closer relationship with macrocephala.

industriously building a nest in an Enderby Island sinkhole, and since none of the other adult tits seen on the island from 1942 to 1945 have been reported as differing from those of Auckland Island, I take the view that there is no evidence to justify taxonomic separation of the two populations. The type of marrineri Mathews and Iredale is the only skin labelled “Auckland Is.” in the British Museum, i.e. 42.12.16.48. The type of enderbyi Mathews is so labelled, a male skin presented by the Earl of Ranfurly and collected at Enderby Island on January 3, 1901. Petroica macrocephala dannefaerdi (Rothschild). Snares Island Black Tomtit. 1890. Miro traversi Buller (not of 1872), Trans. N.Z. Inst., vol. 23, p. 36. 1894. Miro dannefaerdi Rothschild, Novit. Zool., vol. i, p. 688. (Snares Islands.) Adult Male: Whole plumage glossy black, bases of feathers grey with light shafts; wing and tail brownish black with no pattern. (Snares, June.) Adult Female: Similar to male, but with slightly less gloss, especially on ventral surface. (Snares, June.) Fledgling Male (?): Dull black, with slight gloss on back; feathers of crown, throat and breast with pale grey shaft-streaks; ventral flank feathers dull black with fuscous tips; less extensive than in female; nestling down, adhering, greyish brown. (Snares Island, examined in field, December, 1947.) Fledgling Female (?): Similar to male, but ventral plumage more extensively tipped with fuscous and buff, giving a distinctive pale-breasted aspect. This bird is figured in Plate 4, Fig. b. (Snares Island, December, 1947.) The Canterbury Museum collection contains a number of skins, largely deficient in data, which are clearly immature judging by their enlarged orange rictal membranes and weakly pigmented bills, but which have lost (apparently by moult) the characteristic pale-tipped ventral plumage of the fledglings described above. Recognition of such post-fledgling plumages in P. m. macrocephala and P. m. toitoi must await collection of better series than are at present available, and it is possible that the birds described (and figured in Plate 3) as fledglings of these races may include “first winter” Table 7—Dimensions of Petroica macrocephala dannefaerdi (Roths.) Collection and number Locality Date Sex Wing Tail Tarsus Mid-toe Culmen Dom. Mus. A.269 Snares Island June, 1896 ♂ 82 66 25.5 22.5 12.5 Dom. Mus. A.271 Snares Island ♂ 76 62 26 21 C.M.1085.5 The Snares 1894 ♂ 81.5 62 27 20.5 12.5 P.632 (Unlocalised) ♂ 82 66 26 c. 22 13 Dom. Mus. A.272 Snares Islands June, 1896 ♀ 79 64 26 20 12 Dom. Mus. A.270 Labelled “Miro dannefaerdi” ♀ 79 62 26 22 12 C.M.1085.6 The Snares 1898 ♀ imm. 79 61.5 26 20.5 11.5 Note: The rest of the skins in New Zealand collections are either indefinitely localised, although clearly referable to this form, or are lacking data for sex and date.

plumages. One of the Canterbury skins is described below as a “first winter” bird, even though it is labelled female with no recognition of its immaturity, and no detailed date. Fig. 10—Scatter diagrams (“variation polygons”) for the dimensions “length of wing” and “length of tarsus” in New Zealand and Australian forms of Petroica, showing the close affinity of the five races of P. macrocephala and their relationship with Australian species. P. australis and P. traversi, with their discrepant large size and high tarsus/wing ratio are grouped in the subgenus Miro. First Winter (?): Similar to adult, but with somewhat less gloss, and with faint light feather shafts on the crown, becoming inconspicuous

on the back. (Canterbury Museum; Snares, 1895, labelled female.) Variation: No plumage variation has been noted. Dimensions: See Table 7. Insufficient sexed skins are available to present male and female measurements separately. For 18 skins of both sexes the dimensions are: wing 76–82 (mean 78·9, σ = 2·15); tail 58–66 (mean 61·6, σ = 2·59); tarsus 25–26 (mean 25·5, σ = 1·87); mid-toe 20–22·5; culmen 12–14; bill from skull 13·5–16. Differences between mean lengths of wing, tail and tarsus of dannefaerdi and marrineri are not statistically significant, even when both sexes of the latter are grouped to make the comparison a fair one. The difference in wing length may be significant when larger series are compared. The culmen of dannefaerdi is notably longer than that of marrineri. Differences from all other races are significant for every dimension tested. Proportions: Tail/wing percentage 75·4–81·7, mean 78·7, σ = 2·28 (17 skins). This ratio is the same as that of marrineri, and significantly higher than in other races. Tarsus/wing percentage 30·8–34·2, mean 32·5, σ = 1·0. This ratio is also practically identical with that of marrineri (mean 32·8) and a little higher than in macrocephala, though not significantly so. Distribution: The Snares Islands, abundant throughout the main island, chiefly in Olearia lyallii scrub, but also nesting in rocks within tussock areas. Observed nests were some 70 metres apart, which suggests a population in the order of some hundreds of breeding pairs. Characters and Affinities: The outstanding feature of the Snares Island Tomtit is, of course, its melanism; apart from that there is little to separate it from the Auckland Island marrineri, which is certainly its closest ally. The discoverers of P. m. dannefaerdi correctly assessed its affinities by referring to the bird as a “black tomtit” (Reischek, 1889, Chapman, 1891). On his visit to The Snares in 1944, Dr R. A. Falla found that the song and habit of this bird agree with those of the New Zealand Tomtits, and on the Snares Islands Expedition of November–December, 1947, we had the opportunity of confirming this in every aspect of behaviour (see also Stead, 1948). The song is a typical tomtit warble, which could be matched within the range of variation in other races of macrocephala. The Black Tomtit of the Snares resembles the Black Robin of the Chatham Islands chiefly in colour, and differences between them in dimensions, proportions (Fig. 10), foot colour and behaviour prohibit close relationship. The presence of a tomtit at the Auckland Islands resembling the Snares bird in size, but retaining the plumage pattern of mainland tomtits confirms the taxonomic arrangement here adopted. References Alexander, W. B., 1932. Movements of Species of Petroica in Queensland. Emu, vol. 32, pp. 113–4. Andersen, Johannes C., 1926. Birdsong and New Zealand Song Birds. 215 pp. Whitcombe and Tombs Ltd. Buller, W. L., 1872. A History of the Birds of New Zealand, 1st ed. — 1890. An Exhibition of New and Interesting Forms of New Zealand Birds, with Remarks Thereon. Trans. N.Z. Inst., vol. 23, pp. 36–43. — 1892. Further Notes and Observations on Certain Species of New Zealand Birds. Trans. N.Z. Inst., vol. 24, pp. 75–91.

Buller, 1893. Notes on New Zealand Birds. Trans. N.Z. Inst., vol. 25, pp. 53–88. — 1906. Supplement to the Birds of New Zealand, London. Campbell, A. G., 1937. Further Notes on the Birds of Toolern Vale. Emu, vol. 37, pp. 131–132. Chapman, F. R., 1890. On the Outlying Islands South of New Zealand. Trans. N.Z. Inst., vol. 23, pp. 491–522. Chisholm, E. G., 1938. The Birds of Barellan, New South Wales. Emu, vol. 37, pp. 301–313. Finsch, Otto, 1872. Revision der Vogel Neuseelands. Journ. fur Orn., no. 116, pp. 81–112; no. 117, pp. 161–272. — 1874. Zusatze und Berichtigungen zu Revision der Vogel Neuseelands. Journ. fur Orn., 1874, pp. 167–224. — 1888. Ibis for 1888, p. 308. Fleming, C. A., 1939. Birds of the Chatham Islands, Part 2. Emu, vol. 38, pp. 492–509. Fletcher, J. A., 1939. Notes from Eaglehawk Neck. Emu, vol. 38, pp. 421–422. Gray, G. R., 1843. List of the Birds hitherto Recorded as Found in New Zealand, Chatham and Auckland Islands, with their Synonyms. Dieffenbach's Travels in New Zealand, vol. 2, pp. 186–201. — 1845. The Zoology of the Voyage of H.M.S. Erebus and Terror: Birds. Guthrie Smith, H., 1914. Muttonbirds and Other Birds. Whitcombe & Tombs Ltd. Hindwood, K. A., and Mayr, E., 1946. A Revision of the Striped-crowned Pardalotes. Emu, vol. 46, pp. 49–67. Howe, F. E., 1932. Observations on the Genus Petroica, with Notes on Nidification. Emu, vol. 32, p. 39. Hutton, F. W., 1872. Notes on some Birds from the Chatham Islands, collected by H. H. Travers, Esq., with Descriptions of two new Species. Ibis., vol. 2 (third series), no. 7, pp. 243–250. Lord, E. A. R., 1943. Migratory Notes (1941–2) from the Murphy's Creek District, Queensland. Emu, vol. 43, pp. 18–23. Mathews, Gregory M., 1920. Birds of Australia, vol. 8, 1931. A List of the Birds of Australasia, 562 pp. London. Mathews, G. M., and Iredale, T., 1913. A Reference List of the Birds of New Zealand. Ibis, April, 1913, pp. 201–263, and July, pp. 402–452. — 1924. The Nature of the New Zealand Avifauna. Emu, vol. 20, pp. 210–221. Mayr, E., 1931. Notes on Halcyon chloris and some of its Subspecies. Am. Mus. Novitates, no. 469. — 1934. Notes on the Genus Petroica. Am. Mus. Novitates, no. 714. — 1941. List of New Guinea Birds. Am. Mus. Nat. Hist., New York. Mayr, E., and Serventy, D. L., 1938. A Review of the Genus Acanthiza Vigors and Horsfield. Emu, vol. 38, pp. 245–292. Miller, A. H., 1941. Speciation in the Avian Genus Junco. Univ. Calif. Publ. Zool., 44, pp. 173–434. McLean, J. C., 1911. Bush Birds of New Zealand, Part 2. Emu, vol. 11, pp. 65–78. North, A. J., 1903. Nests and Eggs of Birds found Breeding in Australia and Tasmania, Part 3. Spec. Cat. No. 1, Aust. Mus., pp. 121–201. Oliver, W. R. B., 1930. New Zealand Birds, 541 pp. Wellington. Potts, T. H., 1870. On the Birds of New Zealand. Trans. N.Z. Inst., vol. 3, p. 59. Reischek, A., 1887. Notes on Ornithology. Trans. N.Z. Inst., vol. 19, pp. 184–188. — 1889. Notes on the Islands to the South of New Zealand. Trans. N.Z. Inst., vol. 21, pp. 378–389. Ridgeway, R., 1912. Color Standards and Color Nomenclature. Washington. Rothschild, W., 1907. Extinct Birds. Serventy, D. L., 1937. Local Migration in the Perth District, W.A. Emu, vol. 37, pp. 90–93. Sharpe, R. B., 1875. Appendix to Gray's Birds (Zool. Ereb and Ter.). — 1879. Cat. Birds. Brit. Mus., vol. 4. Stead, Edgar F., 1948. Bird Life on The Snares. N.Z. Bird Notes, vol. 3, no. 3. Ticehurst, C. B., 1938. A Systematic Review of the Genus Phylloscopus. Brit. Mus. Nat. Hist. Turbott, E. G., 1940. A Bird Census on Taranga (The Hen). Emu, vol. 40, pp. 158–161. Wilkinson, A. S., 1927. Birds of Kapiti Island. Emu, vol. 26, pp. 237–258. Witherby, H. F., et al., 1938. The Handbook of British Birds, vol. 1. London.