THE IBIS 109: 141-167

LONG-DISTANCE VAGRANCY IN THE PETRELS

W. R. P. BOURNE

Received on 18 October 1965

INTRODUCTION

This paper is intended primarily as a review of the status of the rarer petrels reported in Great Britain, and secondarily as a general survey of long-distance vagrancy and the validity of other extralimital records of petrels elsewhere. The "British List" traditionally includes not only a number of species known to occur in the area regularly, but also some ten species of vagrants of various degrees of plausibility, stretched to the limit by two otherwise collected only in the central Pacific; and the situation is much the same with several other national lists elsewhere. Faith in the British records recently received a rude shock when they were found to figure prominently among the rare birds which occurred in a more than statistically probable profusion around Hastings in southern England at the beginning of the century (Nicholson & Ferguson Lees 1962), and when the "Hastings records" had been eliminated it seemed wise to assess the remaining records anew. An examination of the original data led to a comparison with similar records elsewhere, and the discovery of other doubtful reports there as well, until I have now been through all the more celebrated cases of long-distance dispersal of petrels, by whatever means, still quoted in the more important national lists of the world; though it is difficult to trace and disentangle all of them, especially the older ones. It is hoped that the results of such a comparative survey may be useful even beyond the British Isles, since it appears that errors and impostures have been no respecters of nationality in this field.

I have to acknowledge correspondence with Mr. E. Eisenmann, M. Christian Jouanin, Dr. R. S. Palmer and Prof. F. Salomonsen before this survey started, the stimulus from Mr. P. A. D. Hollom, Secretary of the B.O.U. British Records Committee, to undertake it, and various assistance and advice from Mr. Roger Bailey, Mr. E. K. Barth, Mr. E. Brun, Dr. J. F. Monk, Mr. P. Gould, Mr. P. W. Post, Major R. F. Ruttledge, Prof. K. H. Voous, Mr. J. Warham and Dr. J. Wattel. I am also grateful to Dr. P. Humphrey for permission to cite unpublished records collected during the course of the Smithsonian Institution Pacific Project under his direction communicated by Mr. Gould. The unpopular verdicts are all my own. I shall follow the English and scientific nomenclature of the second (1954) edition of 'The Birds of the Ocean' by W. B. Alexander with minor modifications, and the classification and order of higher groups recommended by Alexander et al. (1965).

GENERAL CONSIDERATIONS

On first examination it might be supposed that sea-birds would wander freely throughout the oceans so that long-distance vagrancy would be common among them. However, most of them are normally restricted to very limited sea-areas by strict preferences for certain types and temperatures of surface water, and except under special conditions they do not wander far from them (Murphy 1936). The conditions which may initiate long-distance vagrancy appear to include abnormal disturbances of the oceanic and atmospheric circulations, especially cyclonic storms, and perhaps also inefficient orientation and navigation in the more migratory species. Once they have gone astray, the larger and sturdier species, with high powers of endurance adapted for vast oceans swept by gales for weeks on end, may of course continue to wander further and further afield searching for congenial areas until they come to grief a very long way from home a long time after they first went astray; but otherwise most sea-birds wander very little.

In addition to their natural wanderings, birds are of course inevitably bound to be transported outside their normal range by human agency, especially on ships bound for distant ports. Such events may lead to both innocent and fraudulent erroneous records of extralimital vagrancy. In order to assess the reliability of records it is essential to try and form a clear idea of the way in which mistaken records as well as vagrancy can occur; most mistaken records in the past have clearly arisen mainly as a result of a failure to allow adequately for all possible sources of error, and it is also important to remember that many obviously unacceptable records were originally put forward in perfectly good faith by people who unfortunately did not know as much as we do now. Some preposterous records on national lists have clearly survived only because informed people have never paused to look at them carefully in the light of modern knowledge. For now that we have many past records available to study, it becomes clear that certain rules can be laid down covering both acceptable and unacceptable records.

Thus "normal" vagrancy appears to occur mainly by dispersal along coastlines and along the main current and wind systems of the oceans in the direction of water and air flow. As a result of the latter effect, birds also tend to move west in high latitudes and the subtropical trade-wind zones, and east in the belt of temperate westerlies. In closed oceans they also tend to move clockwise around the anticyclone stationary in middle lattitudes in the northern hemisphere, but anticlockwise in the south. They are particularly liable to be blown west and into higher latitudes in the subtropical zones, and then sometimes back east again, along the paths of hurricanes and other tropical storms. Pelagic sea-birds are probably not very likely to be forced across land-masses from one ocean to another by major tropical storms alone, because these do not tend to cross land-masses, but to swing back out to sea in the higher latitudes of the same ocean. It seems remotely possible that birds might rise to great altitudes in the rising air in the centre of a tropical storm, and then cross the major land masses in a jet stream afterwards; but, as far as I know, nobody has yet produced evidence for this happening. Hence, some other mechanism than a direct crossing must be sought to explain the occurrence of vagrant sea-birds by natural means beyond a major land barrier.

There are very many ways by which sea-birds may be transported by unnatural means. They are still regularly recorded on voyages in some parts of the world at the present time, notably in regions where there is much fog at sea, and many more must have been seen on the longer voyages of the past. The bodies of some of them which die on deck may be carried some distance before they are thrown overboard. In addition, sailors of many nations used to make a recreation of catching them with various different types of hooks and snares towed behind ships, and found a source of profit in the construction of various articles from their remains; for example, the wing-bones of albatrosses were used as pipe stems, and the dried extremities were often brought home as souvenirs, while the skins of smaller species made tobacco pouches; and the plumage cf all of them must have been greatly in demand for millinery until the trade was stopped at the end of the last century. Bligh (1792) describes how the sailors of the 'Bounty' caught petrels off Staten Land and fed them on grain to remove the fishy taste before eating them; Green (1887) describes the situation in the middle of the last century, mentioning that he had seen albatrosses liberated north of the equator, and a Black-browed Albatross Diomedea melanophris and a Giant Petrel Macronectes giganteus taken to the London Zoo; and in a famous passage, Hutton (1867) describes how several Cape Pigeons Daption capensis were liberated in the English Channel, with the result that northern records have never been accepted in Europe since.

In the early days of collecting, naturalists had little understanding of the preparation of preserved skins, and even the accomplished crew of Cook's first expedition brought few or none home, though they compiled copious written notes and drawings to show the scale on which the sailors of that day could catch birds (Lysaght 1959). There is a list of birds brought back from Cook's later voyages among Banks' papers in the British Museum (Natural History) which still requires more study, as they include many new species described by Latham and Gmelin; but here a new hazard appears, as they were clearly often inadequately labelled, apparently got mixed up together in a "bird tub", and later were often said to originate in a variety of places selected at random, such as "Kerguelen", "Queen Charlotte's" or "King George's" Sounds (which are at opposite extremities of the Pacific), Tahiti or Christmas Island. In some cases this situation was made even worse by the transposition of localities for two species in lists, as with the types of Procellaria furcata and P. fuliginosa from the Bering Strait and Tahiti (Bourne 1957). In this manner a bird like the Gentoo Penguin eventually acquired a name like "Pygoscelis papua" (Lysaght 1952).

The situation was made worse by intrigue and rivalry between early explorers, as in the case of Titian Ramsay Peale, ornithologist to the United States Exploring Expedition of 1838-1842, one of the major figures in the discovery of ornithology of the South Seas (Poesch 1961). He was no scientist but a good enough naturalist by the standards of his day, and far ahead of it in his desire to describe and illustrate habitats. Before he set out he was party to an agreement that participants in the expedition should hand over their results to be published in a series of official reports; a few of these were to be distributed to public bodies and foreign nations, while the authors sold the remainder for their private profit. During the expedition Peale fell out with all his colleagues, and when he returned he found that a favourable government had fallen. His colleagues failed to supply their notes, specimens sent ahead of him had been lost, given away, or parted from their labels, his employers lost patience with the time and expense required to prepare his lavishly illustrated report, and ultimately he was supplanted in the job he expected. Disgusted and impoverished, Peale failed or was unable to take up the spare copies of his report for sale. They were destroyed in a fire, and the report was handed over to be revised by a contemptuous competitor, who blackened Peale's reputation and consigned many of his new species unwarrantably to synonymies. Eventually, he "sought refuge in other pursuits, not so congenial, but more harmonious", leaving only a private record of his fate, "to show how future generations may be benefited by the expensive blunders of the past". He did not live to see millionaires competing for the few surviving copies of his work, now among the most prized treasures of great libraries, and correspondingly hard to consult.

If collectors originally failed to appreciate the need for accurate data with their specimens, it is hardly surprising that sailors who brought them back for milliners and curiosity shops failed to supply much information with them. In the last century the distribution of most sea-birds was little understood, so that it mas correspondingly hard to assess reports of their origin, and there are not just a few but repeated instances of even the best ornithologists accepting such birds uncritically as likely to have been obtained in the vicinity of the ports where they appeared. The most conspicuous case is perhaps J. J. Audubon's records (1839) of a number of southern species said to have been collected by Townsend "a few days' sail" from the mouth of the Columbia River, which figured in the North American list for many years until challenged by Stone (1930); and I have recently also questioned other North American records by Lawrence (Bourne 1964). Among European examples are the Light-mantled Sooty Albatross Phoebetria palpebrata still quoted in many lists though it was discredited long ago by Mayaud (1936), and similar records still survive unquestioned (Fisher & Lockley 1954). Often they are obviously unacceptable because several exotic species of similar origin were reported together; the Sooty Albatross was accompanied by a White-chinned Petrel Procellaria aequinoctialis which has received less attention, for example (van Kempen 1889). It is harder to detect single importations, but by and large all sea-port rarities seem to demand scrutiny, because ports form the first chance for disposal for the live or dead birds acquired deliberately or accidentally during long voyages from foreign parts.

In addition to sea-ports as a place for the disposal of sailors' accumulated pets and wares, attention had also to be paid to food markets as a source of live or dead birds. Until the recent imposition of restrictive legislation the first reaction of the peasantry of most parts of the world to any strange bird seems to have been to send it to market; and in some areas they came to be involved in trade with far-distant parts. Thus it is possible to cite at least one, and then a group of three albatrosses, a petrel, and a group of stormpetrels which appeared in Leadenhall Market, London, in the past (Stubbs 1913, Hartert 1926, Bourne 1964), at least some of which hardly seem likely to have occurred in the northern hemisphere naturally, though in view of the demonstrable trend for at least some stray petrels to appear in the most unexpected places it seems quite unpredictable where others came from. It is also notable that some at least of these market specimens were found by taxidermists, who in these cases admitted where they came from.

In assessing the reliability of past records, the difficulties of taxidermists must also be remembered. Past bird-stuffers were simple but highly-skilled craftsman plying a respectable trade much in demand, not learned scientists, and too much must not be expected of them. They were as vulnerable as anyone to the stray lounger in search of a dishonest penny offering a bird with questionable antecedents for sale, and possibly less qualified than some others to detect impostures. If they acted as middlemen themselves and offered the bird for resale after stuffing, it is not clear that they always implied that the bird was locally taken, even if the purchaser insisted on assuming as much. Furthermore, skins might easily become confused in the course of making-up in an untidy workshop, while if he failed to preserve some prized but mangled specimen, the taxideimist would be under great temptation to replace it with something that he thought was similar, rather than confessing his failure; and this would be likely to pass undetected in view of the normal difference in appearance of a skin before and after preparation. A taxidermist might identify a stray skin wrongly and sell it as something taken locally before he realized what he was doing. If he was caught out in some mistake or petty deception he might feel it unwise to admit this later, if he was ever informed of what happened at all; and, of course, members of a trade never inform on their less honest brethren, although they may know only too well what is going on.

Even if an important specimen chanced by good fortune to fall straight into the hands of educated people capable of appreciating its importance from the start, it would not necessarily be recorded safely, speedily and accurately, though it would now at least be more likely to be sent to a museum than a market. A private collector might skin it and add it to his own collection without either reporting it at the time or attaching proper data to the specimen. A minor collector or museum might send it to a bird-stuffer for preparation, with the increased hazard of swapping specimens, and again fail either to record the specimen or attach proper data to it. It is also an unfortunate fact that some great and famous institutions are so busy, or the staff are so preoccupied, that specimens may be put aside for years before they receive sufficient attention by experts to reveal their importance; for this reason the first examination of a major collection is always an exciting experience, since the most surprising things may lie unnoticed there. If there is a lapse of time before a specimen is recorded, and especially if the information concerning it is passed on by intermediaries, or the custodian of the collection has changed, there is a much increased chance that labels or specimens may become swapped, or the information concerning them become garbled. Therefore, in addition to specimens originating in ports or commercial centres, delayed or second-hand reports are also peculiarly suspect.

In the circumstances it seems desirable to apply stricter criteria to especially old records than has sometimes been customary in the past. Sight records more than about 50 years old require a great deal of substantiation, as few people possessed binoculars or knew much about the appearance of live birds then. Even specimens require careful investigation, to be sure that they are the birds that were collected, and that they were identified correctly. It is also often overlooked that a minute examination of a specimen may provide much useful information. Thus the first, and indeed the only, direct evidence for the fraudulence of the "Hastings records" was the observation by the custodian of the Dyke Road Museum, Brighton, that one was stuffed "in an unmistakeable oriental manner" (Griffith 1927). Birds are commonly stuffed in distinctive ways and with distinctive local products; dissection of some of our most respected specimens might reveal some surprising things, even if it is seldom the ideal dated local newspaper. I have not attempted to examine the interior of any historic specimens, but the idea deserves more consideration in the future.

Ideally, therefore, all the following information should be available before a record is accepted; and there seems no reason why ancient records which fail to meet acceptable criteria should receive privileged treatment compared with those of the present day:-

Who first saw the bird, where, when, and how;

its subsequent history and fate, as far as they can be traced;

how it was identified, if a specimen, as the bird collected, and as a species;

how it was recorded, with details of confirmatory evidence from an independent source, if available;

whether the whole story appears circumstantially probable, in view of past and present information on the subject.

It seems impossible to test all old records fully according to modern criteria now; but at least we should require a reasonably full and detailed story, logically consistent in itself, to be attached to them. A good many highly respected records which have been vouched for by good authorities seem to lack even this. In such cases the only treatment seems to be to suspend judgment and to await the best and most reliable form of confirmation, further records. If the story was true it may happen again.

ALBATROSSES : FAMILY DIOMEDEIDAE

DIOMEDEA EXULANS Wandering Albatross

This is the largest and best-known of the common albatrosses, dispersed throughout the Southern Ocean and migrating into moderately low latitudes so that it might reasonably be expected to cross the equator at intervals. There are many doubtful old records and two good recent ones.

Fossil remains have been reported by Bell (1915) from Pleistocene brick-earths of Ilford, Essex, England; no details are given, it is not clear who identified them, and I am unable to trace the specunens. It is not clear that they were distinguished from remains of the fossil species Diomedea anglica, also originally identified as D . exulans.

It seems likely the locality is erroneous. This may also be the basis of vague subsequent reports of one in Denmark, formerly overlord of Norway.

Degland & Gerbe (1867) report three near Chaumont in November 1858.

Brunnich (1764) reported one, without details, from Norway. It seems likely the locality is erroneous. This may also be the basis of vague subsequent reports of one in Denmark, formerly overlord of Norway.

Degland & Gerbe (1867) quote one killed by a customs official at Dieppe in November about 1830. This is accepted by Mayaud (1938) for the French list, although the record is at fourth hand and a customs official seems particularly likely to obtain specimens from sailors.

Boie (1835) says that Herr Drapiez reported one killed with oars by fishermen off Antwerp the previous September, and comments that it had probably escaped from a ship. This is also quoted by Degland & Gerbe (1867), and apparently really occurred in 1833, at much the same time as the Dieppe record nearby. Schlegel (1844) dismissed these birds as imported.

Ridgway (in Baird, Brewster & Ridgway, 1884) quotes a second-hand report of a bird recently killed at Tampa Bay, Florida, and reports that the skull was preserved but was not yet available. Apparently it was never seen.

Coues (1885, 1895) repeated in the 'Auk' two reports, one of a bird killed on the Columbia River on 13 February 1813, and one of a bird seen below Jacksonville, Florida, in May 1885. The first could have been a Short-tailed Albatross Diomedea albatrus; the second may or may not have had some connection with Ridgway's record, and has also never been confirmed. The species appeared in the North American List,for some years on the basis of these records, but has now been relegated to the "Hypothetical List".

Bannerman (1920) cites an undocumented report by Serra (1879-1882) of a bird of this species and a Giant Petrel Mucronectes giganteus from Tenerife in the Canaries, presumably the one quoted by Molinerux (1930).

The Vauchers (1915) report that an Arab brought in a skull collected on the west coast of Morocco about 1885. Apparently his skull was later identified as that of the very similar Royal Albatross Diomedea epomophora= D. regia by Hartert, who doubted its origin, though it is still quoted in some local lists (Heim de Balsac & Mayaud 1962).

Dubois (1890) reports a second bird from Belgium killed at Blankenberghe on 27 April 1887. The two Belgian records are accepted by most national textbooks but are questioned by Lippens (1954) as possible importations.

Stubbs (1913) reported one hanging, dripping blood, among Turkeys in Leadenhall Market, London, in December 1909.

Orlando (1958) reports in detail with photographs an immature male of the small subantarctic type breeding in the Tristan area killed off Palermo, Sicily, on 4 October 1957.

E. P. Agate (in Bourne 1966b) reports that an immature followed H.M.S. Protector for four hours at 37° 40' N., 9° 45' W. about 50 miles (80 km.) off the southwest coast of Portugal on l8 October 1963. It was described as larger than a Gannet Sula bassana, brown all over with white cheeks and underwings, a pale cream bill, and light brown legs; the observer was very familiar with the species.

In general, until one saw Orlando's photographs, it was possible to doubt whether this species had ever reached the northern hemisphere unaided, since all the earlier reports were second-hand, inadequately documented, or near important ports. The two Belgian records seem the most convincing of them, since they were reported with some detail at the time, though this is a rather surprising area to find not just one but two birds. Now the Palermo record while occurring in an equally surprising place seems incontestable, and the recent Portuguese sight-record was much where it might have been expected. It appears that the species must be a genuine transequatorial vagrant at times.

DIOMEDEA NIGRIPES Black-footed Albatross

This is the North Pacific albatross with the most southerly distribution. Mathews (1936) reports two skins supposedly from Australasia at Vienna, one labelled "New Holland" obtained from a Mr. Smith of London in 1835, and a second collected by Reischeck at Dusky Sound, New Zealand, in July 1884. The latter still figures in the New Zealand list although despite the efforts of an increasing number of observers it has not been reported there again. It seems a likely transequatorial vagrant.

DIOMEDEA MELANOPHRIS Black-browed Albatross

The second largest, the commonest, and the most highly migratory of the southern mollymauks, breeding in the far south, and migrating north towards the tropics. It is now clearly the commonest southern vagrant in the North Atlantic, so that it is surprising that it does not figure more prominently among the records for the last century, though it was among the southern sea-birds known to have been brought alive to London (Green 1887). Those records which did occur then differ rather markedly from most others of transequatorial vagrants in coming from much remoter places, often in the far north.

There are now some seven specimen records:-

One killed by Captain David Grey of the whaler 'Eclipse' at 80° 11' N., 00° 04' (or 04°?) E. on 15 June 1878 and presented to the Peterhead Museum, cited by many authors.

An adult female which visited the Gannet Sula bassana colony on Myggenaes in the Faroes regularly from 1860 until it was killed there on 11 May 1894 (see among others the long English account by Anderson 1895).

A sub-adult caught exhausted inland in a field at Linton near Cambridge on 9 July 1897 (Butler 1897).

An adult female shot on the Faroe Bank 40 miles S.W. of the Faroes on 10 May 1900 (Salomonsen A sub-adult female shot at Lille Hellefiskebanke at 66° N. on the west coast of Greenland in late An adult male taken about two miles south of Ferder Lighthouse in the outer Oslo Fiord, Bond (1959) reports that he has examined one killed out of a small flock off Vauchin, Martinique. The presence of a flock would be strange.

There are about three times as many possible sight records, though these are less certain because many refer to immatures which are hard to tell from those of the Grey-headed Albatross Diomedea chrysostoma among other unlikely species. These include the following:-

A second bird seen by Grey at 74° N. on 2 May 1885; an immature seen by Harvie Brown (1895) 20 miles N.W. of, and apparently within sight of, the Orkneys on 18 July 1894; one seen by George Bolam (1912) off Holy Island, Northumberland, on 21 February 1895; one seen off Oesel in the Baltic in 1911 (Molineaux 1930); an immature at Fair Isle on 14 September 1949 (Williamson 1950); an immature found alive inland at Staveley, Derbyshire, in mid-August 1952 (Macdonald 1953; see also under Yellow-nosed Albatross below); a bird seen at 40° N., 35" and 50° W. on 14 and 23 May 1953 and another at 46.5° N., 39° W. on 13 January 1955 ('Sea Swallow' 7: 2, 8: 12); an adult seen off Cape Clear and an immature seen off Malin Head, Ireland, on 24 and 26 September, and one seen off Spurn Head, Yorkshire, on 10 November 1963; one seen off Brandon Bay, Ireland, on 15 August 1964; one seen off Morte Point, Devon, on 25 April, and one seen off Spurn Point, Yorkshire during 4-7 November 1965 ('Irish Bird Report' and reports on rare birds in 'British Birds' for these years); an immature seen off Robin Hood's Bay, Yorkshire on 15 September 1966 (C. Feare in litt.); one filmed by Harold Pollock feeding with gulls around fishing boats off the Lofotens, Norway, in July 1960 (J. Warham in litt.); one photographed by P. Steingrimmsson in a gannetry in the Westman Islands off Iceland on 20 July 1966 ('Animals', 9: 596); and one off Malin Head, Ireland, 13 September 1966 (R. O'Connor, in litt.).

Until recently there have been no comparable records for the North Pacific, presumably because there are fewer observers and the birds are liable to be overlooked among the numerous similar Laysan Albatrosses Diomedea immutabilis there. Mr. P. J. Gould has now supplied some possible records from among the extensive data now being collected from the centre of this area by the Pacific Project of the Smithsonian Institution, as follows:-

5 and 6 November 1963, at ca. 14° 47' N., 168° 57' W., and 16° 26' N., 169° 34' W., a bird or birds which were perhaps larger than Laysan and Black-footed Albatrosses Diomedea immutabilis and D. nigripes with the head all dark, the neck white, the back, upperwings and tail dark, the rump and most of the underparts white, with an indication of an incomplete dark breast-band, and the underwing with heavy dark borders and a white centre, the wings showing signs of moult. (Other alternatives are especially a hybrid between the Laysan and Black-footed Albatrosses, or possibly young Grey-headed or Short-tailed Albatrosses Diomedea chrysostoma or D. albatrus).

23 January 1965, at ca. 24° 58' N., 156° 55' W., an albatross which in general looked like a Laysan, but with a different, large, deep yellow bill which was dark at the tip, seen by Mr. Gould himself, but so briefly that other features were not noticed (the account of the bill seems distinctive).

7 August 1966 at 1° 15'S., 161° 01' W., two birds noticeably smaller than Laysan Albatrosses which flew close to the fantail of the ship and continued N.W., banking frequently and sharply. Upperwing and back dark brown or black, white on head, rump and underparts, underwing all dark with a narrow line of white running most of the length of the wing. Bill not seen. (A Black-browed Albatross is larger than the others, but otherwise seems likely.)

DIOMEDEA CAUTA Shy Albatross

This is the largest of the southern mollymauks, which breeds around New Zealand and Tasmania, and disperses east and west at other seasons to feed off the coasts of the southern continents. An adult female was collected at 47° 55' N., 125° 37 'W. off the coast of Washington on 1 September 1951 (Slipp 1952).

DIOMEDEA CHLORORHYNCHOS Yellow-nosed Albatross

This is the smallest southern mollymauk, which disperses at sea in the subtropical South Atlantic and Indian Ocean, wandering north towards the tropics. After the Black-browed it seems the species which most often crosses the equator in the Atlantic. It stays, however, in lower latitudes than the Black-browed, so that it is the species which has occurred in the south and west of the North Atlantic where the other frequents the north and east. It was the only mollymauk described in the books before about 1850, so early records of all species tended to be put down as this one, whereas later the Grey-headed Albatross Diomedea chrysostoma became better known under the name D. culminata Gould, and specimens tended to be re-identified as that species (see below).

The first mollymauk for the northern hemisphere was thus inevitably reported under the name D. chlororhynchos when it appeared along the River Trent at Stockwith near Gainsborough on the Nottingham-Lincoln border in England, and was (equally inevitably) shot opposite the Chesterfield Canal basin on 25 November 1836 ('Analyst' 6: 160). Unfortunately no further information can be traced concerning the specimen, unless it was the albatross later alleged to have been shot at Chesterfield on 2 November 1870 which was later reported to be a discarded museum specimen ('Field' 1870: 26 Nov., 4 and 17 Dec.; 'Zoologist' 1871: 2527-2563).

Professor Salomonsen originally consulted me about the second record before ever I planned this paper. The bird was collected in Vestmannaeyrar, near Iceland, about 1844, and the skeleton sent the next year to Professor Eschrischt at Copenhagen where it still is. He inevitably identified it as D. chlororhynchos at that date, and subsequent authors have not only shifted it between that species and D. chrysostoma, but also confused it with the famous example of D. melanophris which spent so long in the Faroes. It is difficult to identify the skeleton in the absence of a large series of other mollymauks to determine their distinctive characters and individual variation. The dimensions for bones, however, reported by Professor Salomonsen, humerus 253 mm., ulna 260 mm., carpometacarpus 101 mm., femur 74 mm., tibio-tarsus 154 mm., tarso-metatarsus 79 mm., length of foramen prelachrymale 13 mm., suggest that the bird was proportionately big in the wing but small in the leg like members of the D. chlororhynchos-chrysostoma group of mollymauks, while excluding D. melanophris or D. cauta. While the size comes in between the average for D. chlororhynchos and D. chrysostoma, the tarsus is short at 79 mm., compared with a range of 72-82 (av. 77.3) for 17 D. chlororhynchos and 80-91 (av. 85.2) for 17 D. chrysostoma (though there was also one young D. chrysostoma of 76 mm.).

The next specimen was taken in the estuary of the St. Lawrence River, Canada, in August 1884 or 1885 (Palmer 1962), and apparently the only problem about it is the year.

A bird obtained at the Barre de l'Adour, Basses Pyrenees, France, apparently in August 1889, was originally reported by Granger (1893) as D. chlororhynchos. When, however, he came to examine its supposed skin in the Bayonne Museum, Mayaud found first (1927) that this was of D. chrysostoma and second (1928) that the skin did not agree with eye-witness reports of the original specimen. It seems possible that the latter was destroyed in a fire in the museum in December 1889, and was later replaced by another skin of foreign origin. Thus the bird was first identified as D. chlororhynchos while there is now considerable doubt whether the existing skin of D. chrysostoma is the same specimen. It seems rather likely to have been D. chlororhynchos since the other is much less likely to have occurred in France; but it is impossible to prove this now.

The remaining records are again North American, including one shot at the mouth of the Bay of Fundy on 1 August 1913, a female stranded 40 miles inland in Maine on 23 July 1934, a feather probably from this species washed up on Jones Beach, Long Island, on 7 November 1948, one off New Smyrna, Florida, on 13 July 1958, different birds off Moneghan Island, Maine, on 21 March and 21 May 1960, and one, perhaps the same as the preceding, photographed off Freeport, Long Island on 29 May 1960 (Murphy 1922, Norton 1934, Nichols 1950, Johnson 1958, Bull 1961, Palmer 1962, and unpublished records by M. Libby, supplied by P. W. Post who will be reviewing all records in 'Kingbird').

A mollymauk found entangled in telephone lines at Staveley, Derbyshire, in mid-August 1952, and photographed before release, caused some controversy, since it appeared to have a yellow nose ('Br. Birds' 46: 110, 307); the parties concerned have now examined the photograph together and decided that it was a young Black-browed Albatross. It is also now agreed that another young bird with grey on the head seen off Malin Head, Ireland, also cited under the Black-browed Albatross cannot be a Yellow-nosed either. Thus although there are now at least six certain North American records of this species, a probable one for Iceland and a possible one for France, this rather likely species cannot yet be confirmed for the British Isles.

DIOMEDEA CHRYSOSTOMA Grey-headed Albatross

This is a close relative of Diomedea chlororhynchos, but its range lies further south. It breeds at much the same stations as D. melanophris, but appears to disperse east and west at sea in the same latitudes instead of migrating north, so that it is rare north of about 40° S., and in consequence seems extremely unlikely to cross the equator spontaneously. It is, however, collected rather commonly at sea in the Southern Ocean, so that it is rather likely to be brought home by sailors. The northern records are in consequence inevitably suspect.

Supposed records for Iceland and France, and those birds which were actually D. chlororhynchos for America, have been discussed under that species. The bird reported by Audubon (1839) from off the mouth of the Columbia River in the North Pacific seems unlikely to have come from there (Stone 1930), and the identity of the skull reported by Cooper (1868) from California remains unconfirmed. Mollymauks with grey heads reported in British waters in recent years at Fair Isle, Malin Head, and Robin Hood's Bay might be this species, but seem much more likely to be juvenile D. melanophris, since it may apparently have grey on the crown in the first plumage (Warham, Bourne & Elliott 1966), and they are therefore listed under that species. There remains one record only for all the northern hemisphere.

Messrs. Einar Brun and E. K. Barth inform me that the only northern specimen is supposed to have been killed at Fiskumvann in Eiker near Kongsberg southwest of Oslo in Norway in either 1834 or 1837. It was first reported independently by Esmark and Rasch in the same journal in 1838. The former reports briefly that two specimens were knocked to death in April 1837, one of which was sent to him by Dr. Boeck, while the other was so damaged that it was thrown away; Rasch confirms the year. The specimen was (as was usual at that date) identified as D. chlorohynchos, and Mr. Barth reports that the first place where he can find the identification corrected is by Schaaning (1913), who identifies it correctly as an adult D. chrysostoma, cites the date actually given on the stand, 1834, and later (1916) was also the first to add the information that it was killed while settled on the ice.

It is difficult to place complete faith in this story as sufficient to establish the right of a bird of the far south to a place in the fauna of the Northern Hemisphere. It seems unlikely that the only two specimens to cross the equator would be found sitting together on ice in south Norway. It seems possible that the bird was actually collected, perhaps with another, sitting on ice, and by Dr. Boeck, in 1834 somewhere in the Southern Hemisphere, and only forwarded with rather vague data to Oslo from Fiskumvann in 1837, where it was incorrectly assumed to have been killed locally at the time. It would be interesting to know where Schaaning got the date 1834 and the information about ice, and whether he put the date 1834 on the stand, or found it there already, and in either case whether the bird was actually the same as the one reported in 1838. In any case, it seems intrinsically improbable that a species whose breeding range is roughly the same as that of the King Penguin Aptenodytes patagonica, and which does not seem to migrate north to any greater extent, can any longer be accepted as likely to cross the equator naturally.

PHOEBETRIPIA PALPEBRATA Light-mantled Sooty Albatross

This species has a range in the south virtually identical with that of D. chrysostoma, and seems about equally likely to cross the equator, though the closely allied Sooty Albatross P. fusca, which has a distribution to the north of it rather like that of D. chlororhynchos, seems much more likely to do so. So far, however, it is the southern and not the more northern species which has been reported in the north. The specimens in collections of exotic species said to come from the region of Oregon (Audubon 1839) and Dunkerque (van Kempen 1889) continued to be cited as collected locally, despite the attempts by Stone (1930) and Mayaud (1936) to discredit them. It seems extraordinarily improbable that P. palpebrata would reach the northern hemisphere before P. fusca, which seems rather likely to occur sooner or later.

THE TYPICAL PETRELS: FAMILY PROCELLARIIDAE

MACRONECTES GIGANTEUS Giant Petrel

This species breeds widely and abundantly in the far south, and many ringing returns now show that young birds migrate north to about 10° S., so that it is therefore a rather likely species to cross the equator into the northern hemisphere. The only records, however, for the North Atlantic area so far are one reported in the company of a Wandering Albatross Diomedea exulans without data from the Canaries by Serra (1879-1882, in Bannerman 1920), and one identified in the London Zoo by Green (1887). Early North Pacific records off the mouth of the Columbia River by Audubon (1839) and of birds waiting on whale-flensing off California by Cooper (1871) do not bear serious examination (Stone 1930), but one, presumably the same bird, reported from Midway Atoll in December 1959, 1961 and 1962 by Fisher (1965) can hardly be mistaken, and is presumably a stray from the south which settled there.

DAPTION CAPENSIS Cape Pigeon or Pintado Petrel

This is the commonest of the southern fulmars, breeding abundantly in the far south and migrating north strongly to the margin of the tropics. It is an obvious species to stray across the equator, and since it is one of the most familiar southern sea-birds also an obvious one to be brought home by sailors. There are about a dozen records from the northern hemisphere, as follows:-

In 1809 Nozeman & Sepp illustrated a report of the occurrence of the Manx Shearwater Puffinus puffinus in Holland with a figure of this species (van Oordt 1930).

About 1825 Verreaux, a dealer notoriously careless with his data, reported two killed at Bercy on the Seine, records very properly rejected by Mayaud (1936).

In October 1844 M. Besson shot one near Hykres in southern France (Jaubert 1853), a record which may deserve more consideration.

In 1853 Lawrence reported the receipt of a specimen lacking all data except that it came from Monterey in a collection received from California, which also contained another southern species otherwise unknown in the north, Procellaria cinereu (Bourne 1964).

In 1870 Allan Hume reported that Theo Carter had sent him one said to originate between Ceylon and India; this is still in the British Museum (Natural History).

A bird said to have been shot at Harpswell, Maine, in the U.S.A. in June 1873 (Norton 1922) appears to have been reported first as a Manx Shearwater. The reasoning used to infer its place and date of capture and subsequent history many years later seem so complicated that they leave much room for error (Bourne 1964).

J. H. Gurney (1901) includes in a catalogue of the French records two from the Department of Sarthe before 1878 that otherwise I cannot place.

A bird in the collection of Sir Pryse Pryse at Gogerddau, Wales, was said to have been shot on the Dovey Estuary in 1879 (Salter 1895).

One was shot at Crumlin, west of Dublin, Ireland, by Mr. William Kelly on 20 October 1881, and reported by the irreproachable observer A. G. More (1882) with the opinion that it came off a ship. It is still in the Dublin Museum, and has still not been accepted as a valid occurrence.

In 1887 Lucas quoted reports that birds had followed a vessel north to Acapulco at 16° N. on the west coast of Mexico.

In 1894 Mather reported the receipt from a birdstuffer of a photograph of a bird shot at the Old Harry Rocks, Bournemouth, England; otherwise unknown.

In 1930 E. D. van Oordt reported the discovery between mid-August and mid-September that year of an imperfect skull on the beach near the Hook of Holland, resurrecting Nozeman & Sepp's record that starts this series at the same time. Professor K. H. Voous has sent the skull for examination, and it does appear to belong to this species, but it was found in a region where a large volume of shipping would be clearing decks before entering one of the major ports of Europe.

The record from Acapulco, if the birds really followed a ship to the vicinity of that port, is interesting as it shows how these birds might follow a ship across the equator. Of the ten from north of the tropics, perhaps that from Hykres in 1884, that from Maine in 1873, those from Wales and Ireland in 1879 and 1881, and the skull from the Hook of Holland in 1930 deserve serious consideration. It is notable that only the last occurred in this century, and that they are all either poorly documented or come from the vicinity of major ports; it is notable that the person who reported by far the best record, that from Dublin, expressed the view that it was an escape; while at the time all the others were also discarded because Hutton (1876) reported that he had actually seen six or seven liberated in the English Channel.

In spite of a steadily growing volume of observations nobody has yet reported seeing one of these unmistakeable birds at sea north of the tropics, though they are famous ship-followers in the south. It seems likely that sooner or later one will be found at sea in the north, but so far we lack completely satisfactory records. The Dublin record, however, is an undoubted example of one found flying free inland; in the first and third versions of the B.O.U. 'British Check-List' the species was placed in brackets, while in the second it was included with a note that it may have been imported. I agree with Bannerman (1959) that it may have been over-strict to eliminate it entirely from the last one. It still seems likely that one day there will be a better record.

FULMARUS GLACIALOIDES The Southern Fulmar or Silver-grey Petrel

This is a species whose claim to have occurred in the northern hemisphere seems totally unacceptable. It is a much more southerly species than the Cape Pigeon, though it follows cool currents north to Peru. It was reported by Audubon (1839) among the southern species supposed to have been taken off the mouth of the Columbia River, while according to Godman (1910) Cooper also picked up a skeleton on the beach of Catalina Island, bringing his score of unsubstantiated southern petrels to three. Godman also cites a specimen from Mazatlan, Mexico, passed to the British Museum (Natural History) in the Salvin-Godman collection, but although it is listed (without any other data) in the catalogue, I cannot trace it now. None of these records seem remotely acceptable.

HALOBAENA COERULEA Blue Petrel

This species of the Southern Ocean was reported in the central Pacific by Layard (1876 a, b) and various subsequent authors; apparently a misidentification of Gould's Petrel Pterodroma leucopera, discussed below.

PACHYPTILIA SPP. Prions

These equally southerly species are occasionally reported from low latitudes, as for example at 10° S. off East Africa on 14 July 1965 by Voous (1966). Those reported by G. S. Willis in this area which he cites were seen at 2° 40' S., 46° 50' E. on 5 September 1952. The type of Pachyptila brevirostris (Gould 1855) said to come from Madeira agrees with typical Fairy Prions Puchyptila turtur from Australasia, and presumably the locality is wrong.

PROCELLARIA AEQUINOCTIALIS White-chinned Petrel or Shoemaker

This species of the Southern Ocean was reported with other exotic birds from Dunkerque by van Kempen (1889) and Molineaux (1930), without further data. It migrates into low latitudes, and has even been reported by Novaes (1959) from the lower Amazon on the equator, so it might reach the northern hemisphere at times. One of the latter birds was 250 miles from the sea up the Rio Tacantin; possibly it had followed the east coast of South America north and then entered the river when it ried to turn south again.

PROCELLARIA CINEREA Brown Petrel or Pediunker

Like the Shoemaker this species may wander north from the Southern Ocean along cool currents, and has even been taken near Ascension. The record by Lawrence (1853) of one accompanied by a Cape Pigeon but with no other data from Monterey, California, still quoted in the last A.O.U. North American check-list seems inadequately substantiated (Bourne 1964).

CALONECTRIS DIOMEDEA=PUFFINUS KUHLII Cory's or the Mediterranean Shearwater

This species not onIy migrates regularly from the North Atlantic and Mediterranean to South Africa, in January 1934 it reached New Zealand. The locality of the reported records from Kerguelen seems unreliable (Bourne 1955). It is perhaps worth pointing out that while this species is usually accounted rare in Britain, recent examples of large influxes in autumn from southwest Ireland have possible precedents as early as 1854 (Andrews 1860).

PUFFINUS PACIFICUS Wedge-tailed Shearwater

This species has a very wide distribution in the tropical Pacific and Indian Oceans. There is one report of it wandering into higher latitudes, of a bird labelled "Vancouver Island, British Columbia" bought in a taxidermist's shop in Eugene, Ontario, which figured in the North American list for some years (Jewett 1929). Falla (1962) has also suggested that a pale-breasted bird found in New Zealand might be of North Pacific origin.

PUFFINUS CARNEIPES Flesh-footed Shearwater

A bird of this species, which breeds in western Australia and the Tasman Sea, and migrates to the Arabian Sea and North Pacific, frequented Port Elisabeth, South Africa at night for some time before it was killed in 1909 (McLachlan & Liversidge 1957).

PUFFINUS TENUIROSTRIS Short-tailed Shearwater

This species, which normally performs a vast figure-of-eight migration between the Bass Straits and the Aleutians in the Pacific (Serventy, in Palmer 1962), has twice been found in the Indian Ocean, once in Ceylon and once in West Pakistan, wrecked on the beach at the time of the northward migration in May (Jouanin 1957). A clear example of movement north from the Southern Ocean into the wrong ocean.

PUFFINUSU PUFFINUS Manx Shearwater

The typical race, which breeds off northwest Europe and in the North Atlantic archipelagos, and winters off the Argentine, has wandered in the west wind zone once to South Africa (McLachlan & Liversidge 1957); and there has even been a ringing recovery in South Australia (Spencer 1962). It is well known that the western Mediterranean race P. p. mauretanicus moves west and north up the coast of Europe after breeding, individuals reaching Scotland, Denmark and Norway (Ash & Rooke 1954, Curry Lindahl 1963).

PUFFINUS ASSIMILIS and P. LHERMINIERI Little or Dusky and Audubon's Shearwaters

These are closely related to the Manx Shearwater, with an allopatric distribution in warmer seas. They are usually treated as two species, though I agree with Vaurie (1965) that it is more convenient to treat them as conspecific. Snow (1965) has recently reported that on the equator the Galapagos race P. (a?) subalaris breeds continuously, individuals nesting at nine month intervals; elsewhere, although there is much variation, they often appear to breed in the winter or early spring. In the North Atlantic they usually lay about February, though they may breed continuously in the northeast Canaries (Bannerman 1914; Palmer 1962), and here at least, they may disperse into higher latitudes afterwards.

After surveying the western records P. W. Post (in press) has found that typical P. (a?) lherminieri from the West Indies move north off the south-east U.S.A. in late summer, while the race P. a. baroli of the central Atlantic archipelagoes has been collected in South Carolina in August 1883 (Peters 1924) and off Nova Scotia on 1 September 1896 (Dwight 1897). Leslie Tuck saw a small shearwater off Onion Point in the Strait of Belle Isle on 21 August 1953, and E. I. S. Rees one in Hermitage Bay, eastern Newfoundland on 4 August 1960, these last two occurring after storms of tropical origin. There are now at least five sight records from the central North Atlantic steamer track around 45° N., 35° W., three in August, one in November, and one in early December (Rankin & Duffey 1948, Grayce 1950, R. Denham & E. D. Macdonald in litt.). In Europe P. a. baroli has been taken in Italy in October 1892 and 1895 (Arrigoni degli Oddi 1929) and in the Skagerak off Denmark on 18 September 1912 (Curry Lindahl 1963).

Until recently there were six specimens and one sight record of P. a. baroli for the British Isles; (Hollom 1960, Thorne & Nedderman 1961) one in April, five in May, and one in August; one off southeast Ireland, one in Wales, one in Cheshire, one in Kent, and three in Norfolk P. (a.) lherminieri was also reported to have been found alive on the beach near Hastings, Sussex, by a friend of Mr. G. Bristow on 7 January 1936 (Harrison 1936). There have also recently been an increasing number of sight records around the British Isles in late summer (when the birds rarely seem to come ashore), starting with one off St. Mary's Isle, Northumberland, on 15 August 1961 (Parrack 1964), followed by others off the east coast and especially off western Ireland. Perrins et al. (1965) have suggested that these were really small Manx Shearwaters, but since there are past specimen records of Little but not of small Manx Shearwaters, it seems more likely that they are the former, and that like the western Mediterranean Manx Shearwaters they move into higher latitudes after breeding, so that they are not infrequent in British waters in late summer (Bourne 1966 c).

Elsewhere a specimen of the southern race P. (a?) baillonii of the Mascarene Islands has also been found in higher southern latitudes at East London, South Africa, in March 1952, so that it seems possible that the movement into higher latitudes after breeding may occur south of the equator as well (Courteney-Latimer 1952; specimen examined).

PTERODROMA INCERTA Schlegel's Petrel

This species breeds on Gough Island and Tristan da Cunha in the southern winter and disperses at sea in that region. Thus the specimen found in a local collection from Zelinki or Dolinnki, Zips, then Hungary and now Czechoslovakia, in 1870 presents one of the stranger examples of possible vagrancy. There appears to have been little data with it, but it was accepted by good authorities at the time (Clarke in 1884 as Pterodroma hasitata; Godman in 1910 corrected to P. incerta). Dr. Andrew Keve informs me that he believes the record is now considered unreliable, though he is unable to say why.

PTERODROMA SOLANDRI Solander's Petrel or Bird of Providence

This close relative of Schlegel's Petrel breeds in the central Tasman Sea and disperses locally as far as the New South Wales coast; reports from the region of the Austral group to the east appear to be due to confusion with Pterodroma ultima in the past. It is said to breed in the southern winter (McKean & Hindwood 1965) so that it would not be expected to migrate far at this season; it is therefore difficult to explain the capture by Nagahisa Kuroda of two females in moult at 41° N., 150° E. off northern Japan on 9 July 1954, although one I have examined agrees exactly with the type from Lord Howe Island. Dr. Kuroda informs me that a petrel reported as P. phillipii from Minami-Daitojima further south on 3 August 1931 (Kuroda 1932) was also the species, which suggests it occurs regularly in the area. Probably these are immature birds.

PTERODROMA ULTIMA Murphy's Petrel

This little known species, discovered during the Whitney Expedition off the Bass Rocks and at Rapa and Oeno Islands in the Austral group and at Timoe and Maria Atolls in the Tuamotus in the south Pacific, has now been recorded from the north Pacific as well, where a female was flying low over the vegetation and calling at Green Island, Kure Atoll in the Hawaiian leeward islands on 7 October 1963 (Gould, in press). Few similar birds were recorded at sea in the area during the extensive observations by the Smithsonian Institution Pacific Project personnel, so it seems likely to have been a straggler.

PTERODROMA INEXPECTATA Mottled or Peale's Petrel

This species breeds in the New Zealand area, and disperses south to the Antarctic ice (van Oordt & Kruijt 1953). In the days when it was commoner it was taken east to 59° S. off Tierra del Fuego on 1 February and 36° 49' S., 111° 30' W. on 5 March in 1769 during Cook's first expedition (Lysaght 1959), so the locality reported by Peale (1848), 68° S., 95° W. near Peter Island in the Bellingshausen Sea on 21 March 1839 (when he noted the capture of petrels in his diary; Poesch 1961) may well be correct. There are now a growing number of records of it wintering in force at the opposite extremity of the Pacific around the Aleutians, so it must clearly be one of the most mobile of petrels; the apparently well-attested if third-hand record of a bird walking up a furrow in a ploughed field in New York State in early April 1880 (Brewster 1881) may involve an attempt to reach the winter quarters overland from the Atlantic sector of the Southern Ocean.

PTERODROMA NEGLECTA Kermadec Petrel

This species breeds abundantly across a narrow belt at about 30° S. in the South Pacific, and has been collected in the past north to 21° N. in the east Pacific; it is pointed out above that a supposed record from the northwest Pacific was really P. solandri. Gould (in press) now adds a record of a male taken at Kure Atoll in the Hawaiian Leeward Islands by A. Wetmore in April 1923, and reports that the personnel of the Smithsonian Institution Pacific Project saw a limited number at sea in the area throughout much of the year, with none in March, building up gradually to a maximum in January, followed by a sudden decline in February. He then suggests that since the birds were seen to the northern extremity of the area of survey they may winter further north, and occur there mainly on passage; the late date of the peak of occurences for a bird which appears to have a prolonged breeding season but to nest mainly in the southern summer suggests that these were in fact probably immature non-breeding birds.

This new evidence for a fairly strong migration by this little-known species may throw a new light on two extraordinary extra-limital reports, the bird said to have been found dead under a tree at Tarporley, Cheshire, England on or about 1 April 1908 (Newstead & Coward 1908); and the second filmed soaring over Hawk Mountain Sanctuary, Pennsylvania, U.S.A. on 3 October 1959 (Heintzelman 1961). The first record was attacked by Iredale as early as 1914; no named person saw the bird until it was bought in Chester Market on the fourth day after discovery and at the time it had never occurred away from the south Pacific. According to Coward's diary in the Edward Grey Institute, Oxford, for 14 April, the eyes had already begun to shrink when it was picked up, and it was only seen in the flesh by Arthur Newstead who bought it, not by Coward himself, though he saw the skin before it was dry. Many witnesses agree that the bird, which is still in the Chester museum, is correctly identified, but the correspondence attached to the skin (Bourne 1963) is more concerned with the cost of stuffing it than with attempts to verify its origin. Tarporley is well inland in Cheshire, away from the Mersey Ports and Manchester, which are visited by vessels from all parts of the world.

The equally bizarre record for North America occurred three days after a hurricane and is documented with a photograph which resembles only this and P. arminjoniana of known petrels, but shows perhaps more white on the wing that would be expected of the latter. Both of these are highly variable birds difficult to tell apart even in the hand, however, and P. arminjoniana often shows nearly as much white on the flight-feathers as P. neglecta, and this might be expected to show up unusually well when it was seen by transmitted light or in an over-exposed photograph, as in this case. The published discussion of its identity ignores doubts which had already been expressed whether it was P. neglecta (Amadon 1961), and in general it seems rather more likely that the bird was P. arminjoniana, already known to stray at times into the North Atlantic, than that it was P. neglecta from the Pacific.

PTERODROMA ARMINJONIANA Trinidade Petrel

This close ally of P. neglecta breeds in a narrow belt around the world between 20-30° S., overlapping with it and apparently sometimes hybridizing with it in the South Pacific, where its breeding range also extends north to the Marquesas at 10° S. It does not normally seem to wander much, and the personnel of the Smithsonian Institution Pacific Project have few possible records in the central North Pacific (Gould, in press), but in the North Atlantic area one was taken far off the Antilles at 20° 51 'N., 43° 35' W. on 31 December 1905 (Lowe 1911), another was found at Ithaca, New York State, U.S.A. after a hurricane on 24 August 1933 (Allen 1934), and a bird now in the Rothschild Collection at the American Museum of Natural History was brought by Brazenor Brothers, taxidermists of Brighton, in Leadenhall Market, London, on 26 December 1889 (Bourne 1963). While the latter specimen may have come from almost anywhere, and is mainly evidence for where taxidermists might obtain skins, it seems not impossible it came from Britain. Like P. neglecta, this species also appears to have a prolonged breeding season, and it seems possible that it may also have an unusually extensive post-juvenile dispersal.

PTERODROMA HASITATA Capped Petrel

This species, long supposed rare or extinct although in fact it still breeds in force in the inland cliffs of Hispaniola (Wingate 1964) has fairly frequently been swept inland all along the east coast of the U.S.A. in late summer storms at the time when it is moulting (Palmer 1962, and various specimens examined). The detailed account by the young Alfred Newton in his first major scientific paper (1852) of a bird found entangled in bushes on a heath in Norfolk, England, in March or April 1852 seems convincing, though Mayaud (1946) does not accept the origin of the bird in Boulogne Museum, France, reported by Dresser (1881) to have been shot locally by one Lebeau Longuety. The supposed Hungarian record (Clarke 1884) was a misidentification of P. incerta, discussed above.

PTERODROMA EXTERNA White-necked Petrel

This form, a close ally of P. hasitata and P. phaeopygia, breeds at much the same stations as P. neglecta in the South Pacific, with a much more clearly defined summer breeding season, and has recently been shown by the personnel of the Smithsonian Institution Pacific Project to migrate far north across the equator in the southern winter (P. J, Gould, pers. comm.). A moulting specimen that agrees with the type of the Kermadec race cervicalis was collected inland in Honshu, Japan, on 29 July 1962 (Kuroda 1963), and Mathews (1931, 1932) has also described a bird brought back by whalers said to have been taken at Tristan da Cunha in 1919 as a local race tristani. The specimen, kindly lent by Dr. A. S. Clarke of the Royal Scottish Museum, appears to differ from the immature type of externa from Juan Fernandes mainly in being adult (wing 315 mm., tail 141 mm., exposed culmen 37 mm., tarsus 41.4 mm.), and its origin seems inadequately proven, though it seems possible that a stray that reached the Atlantic might try to settle on Tristan.

PTERODROMA PHAEOPYGIA Hawaiian or Dark-rumped Petrel

A moulting male in the Leiden Museum, labelled "Ternate, 17 April 1862" and reported in various works as P. leucoptera, was lent by the late Dr. G. C. A. Junge, and proves to belong to the Hawaiian race sandwichensis, recently rediscovered breeding in the hills by Richardson & Woodside (1954). It is not clear why it should occur in the Moluccas (though there is also a specimen of Bulweria bulwerii labelled "Ternate" at Copenhagen), but since this is a summer breeder it seems possible that there is an undetected winter quarters somewhere offshore, and possibly also an undetected breeding station somewhere to the north in the region of the Marianas.

PTERODROMA LEUCOPTERA Gould's, White-winged, or Collared Petrel

This is a member of a complex, little known group of highly migratory forms from the Pacific, last reviewed by Falla in 1942. One close ally, Cook's Petrel P. cookii, has a population only known to breed in New Zealand, collected commonly off western South America, at sea off Mexico, and once in what may be the wintering area at Adak Island in the Aleutians. A second, Stejneger's Petrel P. longirostris, appears to perform similar movements in lower latitudes; and the typical race of P. leucoptera breeds in the summer off New South Wales, while very similar birds have been collected in moult at 4° 20' S., 93° 30' W. near the Galapagos on 11 June 1906 (Loomis 1918; skins seen); it is not clear whether they had crossed the Pacific or come from an undiscovered breeding station in the area. Intermediate birds have been collected in the region of the New Hebrides, and a small, more or less melanietic race P. l. brevipes, breeding in the Fiji group, was first reported from the far south (Peale 1848), and is said to have occurred in Wales (Salvin 1891).

The locality alleged for the type and a paratype P. l. brevipes, 68° S., 95° W. in the Bellingshausen Sea on 21 March 1839, is preposterous, According to Peale's diary (Poesch 1961) he did indeed catch a number of petrels there then, including the "cyaneus of Lesson", perhaps the "Prion coerulea Lesson" from the coasts of Patagonia and Tierra del Fuego mentioned in the list of specimens at the back of his report; but he also mentions in his report that the labels were missing from his "Procellaria brevipes", so that it seems likely that he confused the "cyaneus of Lesson" with other birds from the central Pacific. It is not clear where they did come from, though it might well be Fiji; however, while I can find no mention of petrels from Fiji in Peale's journal for that place, in 1876 Layard reported "Procellaria coerulea or what passes for it in these seas" from Samoa as well, and Peale's journal for Samoa mentions that on 20 October 1839 "the natives had a great variety of birds" at Upolu, and ten days later that he saw tropic-birds and petrels visit a great precipice with a 900 ft. waterfall near the town of Uimanga inland; and he may well have obtained his birds here.

In view of the subsequent record from Wales, it is notable that Layard began a second paper in 1876 "there are several collectors of birds now in Fiji, who are transmitting specimens of birds to Europe and Australia...", and also that there were a number of missionaries in the area who are known to have sent home birds as well as various other curios. A considerable number of specimens of P. l. brevipes reached several different museums in the period 1850-1888 by devious routes, including those at London, Cambridge, Paris and Leiden, and it must have become quite a familiar species, and almost a drug on the market. The Welsh bird was apparently destroyed in London by the bombing during the second world war (Bannerman 1959), but Salvin's original account and excellent figure and the subsequent information collected by Witherby, Jourdain, Ticehurst & Tucker (1940) suggest that its identity is not in doubt. Its origin seems more doubtful; it is said to have been shot by an unnamed longshore shooter, who showed it to the vicar, who advised him to take it to the Aberystwyth bird-stuffer, who eventually sold it to Mr. Willis Bund, Q.C., something over two months later as a Sooty Shearwater Puffinus griseus, and nobody enquired any further than this.

PTERODROMA ROSTRATA Tahiti Petrel

This is a large petrel breeding with P. leucoptera on the mountain tops of some of the higher central Pacific archipelagos, and judging by some recent records of the Smithsonian Institution Pacific Project ranging widely through equatorial seas. A female was caught on a fishing boat between Takao (Kaohsiung) and the Little Ryukyu Islands northeast of Formosa on 21 May 1937 (Kazano 1938, Hachisuka & Udagawa 1951), presumably brought by a storm from tropical seas.

BULWERIA BULWERII Bulwer's Petrel

This is a small petrel superficially resembling the gadfly petrels which breeds in the central North Atlantic and Pacific archipelagoes and disperses at sea in the tropics in the winter, avoiding the continental coast. It is rarely recorded anywhere; in Europe one occurred at a lightship between Corsica and Genoa in 3 June 1898 (Jourdain 1912), a body was found on a rock off Trinidad in the West Indies on 23 January 1961 (ffrench 1963), there is one in the Copenhagen Museum labelled "Ternate, Herr Rodeimer" which may or may not come from the Moluccas (which has also produced a specimen of P. phaeopygia, breeding in the same region to the north), and one was caught exhausted in the Maldives on the surprising date of 23 August 1958 (Phillips 1959), perhaps an immature non-breeder summering in the winter range. The locality of a supposed North American specimen from Greenland has long remained very doubtful; unlike some species breeding with it, such as the Madeiran Storm-petrel Oceanodroma castro, it is not cast ashore by tropical storms there.

The first British specimen shot inland at Tanfield, Yorkshire, on 8 May 1837 was long lost, and in view of the fact that it was recorded by Gould the locality might be doubted, but it was rediscovered and fully authenticated by Newton (1887) 50 years later.

A second Yorkshire bird was reported without details from Scarborough by a bird-stuffer in 1849 (Higgins 1849), and a third was washed up at Scalby Mills near Scarborough on 28 February 1908, though not recorded for 14 years (Collinge 1922); the first and last of these are apparently in the York Museum and are usually accepted. There are also five records from the vicinity of Hastings in Sussex between 1903 and 1914, four of which were rejected among the Hastings Records (Nicholson & Ferguson Lees 1963), though the first, reported to have been picked up under Beachey Head by some unnamed person on 3 February 1903, "shown in the flesh" to N. F. Ticehurst (1903) and sold to the uncritical collector Vauncey Crewe (Walpole Bond 1938) only escaped this fate because it occurred outside the area selected as characteristic for "Hastings Records". One was also seen at sea off Cape Clear, Ireland, on 26 August 1965 (Clements 1966). The number of Yorkshire reports is curious; possible as with some other rarities from that region they were seeking to escape from the North Sea.

STORM-PETRELS: FAMILY HYDROBATIDAE

OCEANITES OCEANICUS Wilson's Storm-petrel

Although so widespread and numerous, breeding in the far south and migrating north abundantly to about 45° N. in the Atlantic and eastern Mediterranean, and rather less far north in other oceans, as remarked by Eliot (1939) and Boyd (1954) this species seems to wander ashore extraordinarily seldom compared to the northern species of storm-petrel; although its flight appears weak, it must be very well adapted for bad weather. Although it is abundant off the east coast of the U.S.A. in summer, staying into the hurricane season, there are only about five records inland, no more than for the Madeiran Storm-petrel Oceanodroma castro from the far side of the Atlantic (Palmer 1962); and there appear to be equally few for Europe, even where it is common offshore in summer in the south. Mayaud (1938) only accepted two for France, from Biarritz and Arcachon on the Biscay coast on 3 December 1872 and 6 September 1883, and among the substantial number for the British Isles cited by Witherby et al. (1940) among others extraordinarily few stand up to critical scrutiny, none in this century.

The species was first reported off Land's End in May 1838 by Gould (1839), who distinguished the British Storm petrel Hydrobates pelagicus as well, so the record may be valid. One subsequently came ashore in a wreck of British Storm-petrels at Polperro, Cornwall, in mid-August that year, permitting a careful comparative description by Couch (1838), and Yarrell (1885) also reported two skins brought in by a sea-captain from the English Channel about that time, so that it seems possible that there was an invasion then. Thereafter there are half a dozen or more records from various parts of England during the remainder of the century which almost invariably lack adequate data, until three birds were recorded at Mossvale and Lough Erne in Northern Ireland and one in Jura in the Hebrides during a "wreck" of Leach's Storm-petrel Oceanodroma leucorhoa on 1-2 October 1891. The first two are still preserved in the Dublin and Belfast museums, and Major R. F. Ruttledge and Mr. J. W. Greaves inform me that their wing-length is 148 and 159 mm; this suggests that they belong to the large Antarctic populations which I think should be called 0. o. exasperatus Mathews (Bourne 1964). Apparently the only British record remaining unquestioned since this time is one said to have been picked up on the shore at Allonby, Cumberland, by Moore Kitchen in the winter of 1932 (Blezard 1943). I understand from Mr. Blezard that it was apparently never submitted for examination by an expert, and has been lost with the disposal of the owner's collection at his death. This is a curiously small total for a bird abundant in the Bay of Biscay for half the year.

PELAGODROMA MARINA White-faced Storm-petrel or Frigate Petrel

This species perfmms considerable migrations from temperate breeding-places into lower latitudes in the winter, while the North Atlantic populations may disperse north in late summer (Bourne 1966 a), yet in spite of this, like Wilson's Storm-petrel, it wanders only rarely. Two records from New England on 3 September 1885 and 18 August 1953 (Ridgway 1885, Gordon 1955) may represent extremes of the regular late summer dispersal in the North Atlantic. Two have been recorded later in the season in Britain; one dead with a Wilson's Storm-petrel in a wreck of birds after a gale at Walney Island, Cumberland, in November 1890 (Macpherson 1891), and one alive by a stream on Colonsay in the Inner Hebrides after a southwest gale on 1 January 1897 (Clarke 1897). The first was never shown to an ornithologist until the following July, when it figured among a collection of birds bought for half a soverign in Barrow-in-Furness, then a considerable port, and it is compromised by the company of a Wilson's Petrel at a time when nobody else noticed a wreck, so that it seems totally unacceptable. The second was sent at once in the flesh to Eagle Clarke at the Royal Scottish Museum, and has been lent to me so that I can confirm that it is an immature of the race P. m. hypoleuca from the Salvages. The late date is surprising, but other storm-petrels also tend to occur at the Irish and Danish coastal lights late in the year (Barrington 1900; Hansen 1954), perhaps sickly young birds which have failed to make their way against the prevailing southwest winds on autumn migration.

FREGETTA GRALLARIA White-bellied Storm-petrel

FREGETTA TROPICA Black-bellied Storm-petrel

These sibling species breed in the Southern Ocean and migrate north to winter in the tropics; owing to their close similarity in size and marked chromatic polymorphism and geographical variation they are commonly confused with each other (Murphy & Snyder 1952; Bourne, in Palmer 1962) and I shall discuss most records elsewhere. I have already expressed the view that it is highly improbable that any members of this highly pelagic group not otherwise known to wander much would be found together in numbers in Florida in the way reported by Lawrence (1851), while supporting his identification of the birds which probably came from the tropical Atlantic, as F. g. leucogaster which breeds at Tristan da Cunha (Bourne 1964).

OCEANODROMA TETHYS Galapagos Storm-petrel

This species, which breeds off the coast of Peru and in the Galapagos, has wandered or migrated at least as far north as the Revilla Gigedos and Guadelupe Islands off Mexico (Huey 1952); it seems not impossible it might even breed there.

OCEANODROMA CASTRO Madeiran or Harcourt's Storm-petrel

This is a highly pelagic species breeding in the central Atlantic and Pacific archipelagoes, and rarely seen near the major land-masses. Like other northern storm-petrels O. castro seems to be rather more prone to wander or drift with storms than the southern species are. It has already been pointed out that as many have occurred inland in North America, some at least after hurricanes, as with Wilson's storm-petrel Oceunites oceanicus, which is infinitely commoner offshore (Palmer 1962). It has been recorded twice in the British Isles; one is said to have been picked up on Milford Beach, Hampshire, on 19 November 1911 by Roland Follett (Munn 1912), though few details of the occurrence are recorded; and one in the Dublin Museum struck the Blackrock Light, Co. Mayo, western Ireland, on 18 October 1931 (Humphreys 1932). Other reports for Europe are apparently now discredited.

OCEANODROMA LEUCORHOA Leach's Storm-petrel

This is the commonest and most widespread of the northern storm-petrels, breeding around the northern periphery of the Atlantic and Pacific, and wintering widely in the tropics to the south. In consequence it is the species which occurs most commonly in "wrecks" in bad weather, though other northern species and occasionally southern ones may occur with it. The most famous of these in Britain occurred in the late autumns of 1891 and 1952 (Evans 1892, Boyd 1954), the first in the north and the second in the south, when birds were scattered over much of western Europe. Local wrecks involving fewer birds are reported at much shorter intervals. But while hundreds or thousands of birds may be involved in the larger incidents, they are rarely spread very far, most coming down within 500 miles of the sea, while none have travelled the distances in excess of 1000 miles involved in some other cases of supposed vagrancy, such as those found in the genus Pterodroma or the albatrosses.

The extreme examples of vagrancy in this species come from the south, including a sight record at 57° 40' S., 5° E. near Bouvet Island in the Southern Ocean on 1 January 1948 (Bierman & Voous 1950), one at Tower Hill, Port Fairey, Victoria, Australia on 5 July 1965 (Serventy & Whittell 1965) and one on Muriwai Beach, New Zealand, in August 1922 (Oliver 1955). Dr. Serventy thinks that the Australian bird may have come from the South Atlantic in the west wind zone; the New Zealand one at least seems as likely to have come from the little-known tropical Pacific wintering area.

OCEANODROMA (LEAUCORHOA?) MONORHIS Swinhoe's Storm-petrel

This close ally or race of Leach's Petrel breeds to the south of it in Japan and the East China Sea and winters to the south around the East Indies, while there is a growing volume of evidence that part at least of the population regularly penetrates the tropical Indian Ocean and even the Arabian Sea (Bailey 1966). There is one extraordinary record of an exhausted bird caught at the head of the Gulf of Aqaba in January 1958 (Merom 1960), possibly one that attempted to take the wrong route home out of the Indian Ocean.

DIVING-PETRELS: FAMILY PELECANOIDIDAE

Members of this family usually appear to be comparatively sedentary, and records outside the recognized range often seem to indicate the presence of undetected breeding colonies, though they clearly wander hundreds of miles at times. Thus Dr. D. L. Serventy tells me that the Common Diving-petrel Pelecanoides urinatrix which in Australia only hreeds in the Bass Strait region has been taken north along the east coast to Queensland, while Gibson & Sefton (1959) report a specimen of the even more southerly Georgian Diving-petrel P. georgicus, which breeds no nearer than the Auckland Islands, on a beach south of Sydney, eastern Australia, on 28 December 1958.

CONCLUSION

These records present some difficult problems. Some are obviously likely to be good, and these include some of the most remarkable examples of vagrancy found among birds; many are obviously unacceptable, and these include some of the most remarkable examples of improbable events ever accepted by naturalists, while a large intermediate class can only be a matter of opinion. Some of the more dubious records were unchallenged for many years by good authorities, and some still figure in a variety of national lists, while other records at least as good are refused admission even to hypothetical lists or in small print and square brackets. Attempts to change the status quo are rarely greeted with favour by its custodians, though often with unseemly glee by their critics. In most cases I have not so much tried to state a conclusion as to present the essential facts relating to some remarkable events (whatever their nature) for re-examination in a context of comparable records. Whether all the records are valid or otherwise, certain trends can be distinguished among them which permit the formulation of some conclusion about the nature of such records in general, in the way already suggested by Fitter (1960) for assessing large series of sight records; though I do not agree with him that large series of dubious records suggest that some are necessarily good; they may also indicate the common occurrence of a particular type of error.

There are some criteria separating probable good records from likely errors. Sometimes some obvious natural phenomenon such as a hurricane causes a number of exotic birds to appear together in its wake in a "wreck"; in this case their appearance together reinforces belief in their authenticity and the cause for the "wreck" should be obvious. Otherwise rare birds usually appear singly, not in groups, either of the same or different species, distributed at random over the sea and land, and certainly not concentrated in the vicinity of such sources of importations as major ports; the status of birds found in markets is always hard to assess, for they may or may not originate locally. It is rare for one man to record many vagrants unless he has quite exceptional opportunities for meeting them, eyesight or powers of credulity. The frequency of records of each species has usually tended to remain the same or even increase over the course of time with increasing observations for the genuine wanderers, as with the Black-browed Albatross Diomedea melanophris, though of course this is not necessarily always going to be the case, as with that notorious irruptor of the last century, Pailas' Sandgrouse Symhaptes paradoxus in another group, which has ceased to favour us with its intermittent appearance in western Europe for a long time now.

Obvious errors tend to occur in groups of the same or related species, often in defiance of geographical probability and other experience for the new area either before or since, and the same person may be found involved, either innocently or deliberately with several of them. The birds are often reported along the coast, especially around ports, or are produced in the course of commercial transactions, or in the course of handling numbers of specimens where data may be mislaid or especially swapped: swapped data are an important and sometimes detectable source of mistakes. The evidence for the bird's real origin and the early stages of their history may be obscure, though this may also occur with good records originating in remote places, and a long and complicated sequence of reasoning to establish their validity is suspicious, because it leaves room for error at several stages in the sequence. Dubious records have commonly achieved recognition purely because some distinguished authority chooses to take it upon himself to sponsor them and people are reluctant to doubt him. Once they have obtained a place in respectable literature, such as a regional, national or international list, errors are extraordinarily hard to discredit because they are subsequently quoted back and forth between uncritical authors faster than attempts to disprove them can be published. This last fact alone seems to provide one justification for examining all outstanding petrel records together again afresh in this paper.

Examining records of vagrant petrels, it is notable that although there are more sea-birds and their movements are better developed in the Southern Ocean than the North Pacific, and in the Pacific than the Atlantic, it is the last ocean that has produced most records of vagrants, perhaps because more observations are made there and vagrants are less likely to be overlooked among a host of little-known allied forms. It is also noticeable that it is usually the larger species that have been detected wandering really long distances, either because they have greater powers of endurance, or because they are more noticeable, although the most prodigious feats of endurance are reported among medium-sized gadfly-petrels and shearwaters. Although smaller species are sometimes wrecked in vast numbers at short distances with storms, they have less often been reported travelling really far in the past, either for lack of stamina, or because they have been overlooked. There is slowly accumulating evidence, however, that some of the notorious drifters such as Leach's Storm-petrel Oceanodroma leucorhoa may travel really long distances at times.

Most acceptable records of vagrant sea-birds can be explained either as the result of movement with recognized wind or current systems, or the malfunctioning of recognized migrations, either because the birds set off in the wrong direction, or in the right direction from the wrong place. It also seems possible that some records provide indications of hitherto undetected ranges and migrations. Some of the birds are eventually detected dispersed at random at sea or where they eventually come to grief after wandering inland, but with improved understanding of the factors which may lead to local concentrations of all migrants and especially sea-birds in recent years increased numbers of records are beginning to be reported from places where the birds are known to congregate.

Examples of possible undetected ranges or migrations are the movements of Capped Petrels Pterodroma hasitata, Little Shearwaters Puffinis assimilis and White-faced Storm-petrels Pelagodroma marina north in the late summer after breeding in the North Atlantic, or of White-necked, Kermadec and Solander's Petrels Pterodroma externa, P. neglecta and P. solandri from southern breeding places into the North Pacific, the last two perhaps mainly in the pre-breeding period, in the way that British Kittiwakes Rissa tridactyla and Great Skuas Catharacta skua may find their way north and then west to Greenland (Coulson 1966, Thomson 1966). It will require some such movements as these to bring the more distant wanderers within the influence of the northern hemisphere hurricanes and cyclones which may then cast them ashore in such regions as the eastern United States and Japan, where a good many records of rarities are known to follow the passage of storms of tropical origin (see for example the North American record of the Trinidade Petrel Pterodroma arminjoniana, Allen 1934). It seems likely more will be found to do so as the result of further study.

Those birds which are not cast ashore immediately may also be carried into unfamiliar weather-systems with the result that they travel even further afield; birds brought to the east coast of North America by hurricanes are then likely to drift northeast in the west wind zone to reach the coast of Europe some time later: although by the time they arrive it may be too late to link their arrival with the passage of the storm that originally caused them to go astray. This is the possible case with the storm-petrels which appear at the European lights in winter (Barrington 1900, Hansen 1954) although they first went astray on autumn migration three months or more before. Dr. D. L. Serventy has pointed out to me that in Australia long-distance vagrants appear to be much more likely to be wrecked on local beaches than the local birds when they do arrive in the area, presumably because they do not have innate reactions adapted to evade the hazards of the local meteorological and topographical situations. This may explain why vagrants are so often wrecked in situations avoided by local birds.

Among other trends revealed, one of the more interesting is surely transequatorial vagrancy, if only because it helps explain the present bipolar distribution of some groups supposedly characteristic of high latitudes, such as the albatrosses and fulmars. This has frequently been alleged in the past, and equally often denied, usually on the grounds that the birds involved have been transported on ships. It now seems clear that petrels can and do get transported on ships, both alive and dead, and both voluntarily and involuntarily. On the other hand, strong migrants such as the Black-browed Albatross Diomedea melanophris have now appeared on the "wrong" side of the equator much too often for all records to be explained so easily, and indeed that species must come very near to establishing a potential breeding population in the North Atlantic, in much the same way that Voous (1949) has suggested that an ancestral Fulmar once first colonized the Northern Hemisphere.

It now also seems clear, however, that a good many old records of transequatorial vagrancy are also invalid, especially those of such sedentary Antarctic species as the Grey-headed and Light-mantled Sooty Albatrosses Diomedea chrysostoma and Phoebetria palpebrata, and that it is normally only species that frequent low latitudes, and especially strong north-south migrants that regularly winter there, which are likely to find their way across the equator. These include, in addition to the Black-browed Albatross, the Yellow-nosed, Wandering and Shy Albatrosses Diomedea chlororhynchos, D. exulans and D. cauta, the Giant Petrel Macronectes giganteus, and more doubtfully the Cape Pigeon Daption capensis.

Where long-distance vagrancy occurs in other groups it can also often be explained as the result of a misplaced migration, as with the two Short-tailed Shearwaters Puffinus tenuirostris which migrated north in the Indian Ocean instead of the Pacific, or the Mottled and possible Kermadec Petrels Pterodroma inexpectata and P. neglecta that migrated north in the Atlantic instead of the Pacific; the Swinhoe's Storm-petrel Oceanodroma monorhis which appeared in Israel was presumably also trying to find its way back to Japan by the wrong route from the Arabian Sea. This type of displacement seems particularly likely to originate in the Southern Ocean, where from the evidence of birds which shower ashore after every gale, the native birds are clearly continuously drifted around the world by the strong winds of the far south. Such northern winter visitors as the Manx and Cory's Shearwaters Puffinus puffinus and Calonectris diomedea and the Leach's Storm-petrels Oceanodroma leucorhoa sometimes appear among them. These have travelled in the west wind zone from South Atlantic wintering places to Australia and New Zealand. Indeed, considering the local conditions, the mystery is not that some birds of the Southern Ocean wander so far, but that many do not do it more often; it is very strange, for example, that the Great Shearwater Puffinus gravis, so abundant in the South Atlantic, has not yet reached Australia and New Zealand with the other North Atlantic wanderers.

The manner in which rare vagrants appear is also often interesting. Many, especially young birds, are simply found dead, injured in landing, or exhausted along the coast and scattered at random inland, or, in a state of exhaustion, frequenting sheltered coastal or inland waters where they may eventually be found dead along the shore. There is a definite tendency for storm-driven birds to accumulate at the apex of inlets and estuaries, as occurred most notably with the great Leach's Storm-petrel Oceanodroma leucorhoa wrecks of 1892 around the northern approaches to the Irish Sea (Evans 1892) and in 1952 at the head of the Severn Estuary (Boyd 1954); more birds are often seen alive in such situations than are found dead there, because they are still not so exhausted that they are unable to fly on some distance inland before dying. It seems possible that records of rare sea-birds in such enclosed seas as the Mediterranean where they might not be expected, may result either from birds following coastlines through their narrow entrances, or through their being driven through a "funnel entrance" such as the Straits of Gibraltar by onshore gales.

The chances of vagrants being picked up if they disperse at random are low, and are growing less with the diminution of the numbers of rough shooters and amateur collectors in the more civilized parts of the world. This tendency for the numbers of casual reports to decline is offset by a growing understanding of the factors which lead to the accumulation of concentrations of sea-birds as of other birds, rarities with them. One of the more notable records of recent years was that of an apparent Kermadec Petrel Pterodroma neglecta in a passage of raptors at Hawk Mountain, Pennsylvania (Heintzelman 1961); clearly this soaring oceanic species is affected in much the same way by topographical features as terrestrial soarers. A growing number of records of rare sea-birds are also now being obtained in the British Isles by exploiting the "guiding line" effect of the shore where birds coast along it when they find it across their path, streaming around projecting headlands, instead of moving overland. Increased observations from many western headlands are now producing a growing harvest of records of passing rarities formerly only known from occasional records inland (it follows that there should be far more birds out at sea). There are now also a growing number of sight records of many species from the region which produced old specimen records of such birds as Bulwer's Petrel Bulweria bulwerii along the northeast coast. It seems likely that these are birds which have followed the Gulf Stream up to the coast of Norway trying to find their way back south through the North Sea, and perhaps in many cases beating back north up our northeast coast when they find their way barred by shallow water ahead of them.

A combination of birds coasting south through the North Sea and others blown up the English Channel by westerly gales might also help explain the large number of sea-bird rarities reported in the past around the Straits of Dover, although the absence of many recent records in this region despite greatly increased observations suggests that as with many old land-bird records from this vicinity a good many of the sea-bird records are due to unnatural causes. Reports of the occurrence of a whole series of Wandering Albatrosses Diomedea exulans in particular on the far side of the Channel seem rather strange when we consider how rare it is elsewhere in the North Atlantic.

Finally, we come to the two most extraordinary records on the British list, those of the Collared and Kermadec Petrels Pterodroma leucoptera and P. neglecta. I started this survey with the conviction that they were extremely improbable records. This was pointed out long ago by Iredale (1914) among others, who observed that such records would be remarkable even in their Pacific area of origin, let alone on the far side of the world. But, in the course of the survey as one encountered an increasing number of records of members of the genus Pterodroma found at vast and on the face of it improbable distances from their known range, it became increasingly difficult to doubt that the members of this genus must be capable of prodigious feats of dispersal, though the birds involved commonly turn up in such remote places and in such bizarre ways that their origin is often very hard to prove. Apparently their normal way of life requires constant cruising through the remotest parts of the ocean in search of widely dispersed food in the regions where storms are strongest and most frequent, so that the birds are liable to be swept far off course yet still carry much body-fat, sufficient for further long feats of endurance while seeking their way back to their normal range. Such birds might well occasionally end up thousands of miles away in another ocean or continent, and then get wrecked under the unfamiliar conditions found there.

While it does not seem impossible that a bird like the Kermadec Petrel Pterodroma neglecta might be blown, say, half way round the world in the Southern Ocean, and then wander north across the equator in the Atlantic instead of the Pacific to appear in eastern North America and Britain, it is a very different matter to prove such an event. It would be even more improbable with a tropical species like the Collared Petrel P. leucoptera, which appears to be more sedentary in tropical seas, and is still only reported once-in Wales-outside the tropical Pacific. There is normally no storm-track or lane of ecologically suitable surface water linking the two areas; moreover we do now know that in this case specimens similar to that found in Wales were being sent to Britain from the Pacific in the last century. This record verges on the incredible as an example of natural dispersal, and there is another explanation available. It is a matter of the utmost difficulty to decide how it should be treated.

Furthermore, as pointed out by Iredale (1914) and Bannerman (1959), it is illogical to accept some of these records, and yet not others. Why should we accept a gadfly petrel otherwise only known from Fiji, yet refuse a Cape Pigeon which is known to be common in the South Atlantic, when it is collected in much better-attested circumstances nearby, or a Red-billed Tropic-bird Phaethon aethereus which breeds commonly in the subtropical North Atlantic, which was picked up dead near Malvern in 1854 (Gurney 1894)? The fact must be accepted that some records of extreme extra-limital vagrancy occur, and probably more than are detected or adequately documented, though not all those accepted are genuine. These records both require very careful proof before complete faith can be placed in them, and yet from their very nature they are exceptionally difficult to prove, if only because human transportation of a bird at some early stage of its career as a vagrant is virtually impossible to investigate. It would be misleading to discount such records entirely, and inconsistent to accept some and not others, yet it is almost impossible to tell which of the older records especially can be accepted at their face value.

My own view is that they should not be accepted unreservedly in national lists at least until reliable eye-witnesses are available for their first occurrence in the country (which does not seem to be the case for the Collared and Kermadec Petrels in Britain, although it was for the Cape Pigeon); and that they certainly should not be described at length in accounts of the birds of a country; but that it could well be helpful if they were regularly noticed, if only in small print or brackets and with reservations expressed, as records of about the same standing as those of introduced species not yet fully acclimatized to the avifauna, extinct species and fossils. If such records are not borne in mind they may be overlooked when the time comes for us to remember them again.

SUMMARY

The records of long-distance vagrancy in the Procellariiformes are listed and re-examined. Some are clearly valid, more are obviously doubtful, many are difficult to confirm. Some records from the last century have failed to be repeated in this, but others have been repeated, sometimes more frequently. Some of them suggest hitherto unrecognized migrations or post-juvenile or post-breeding dispersal. Otherwise in general it appears that the more migratory albatrosses and much less often perhaps the southern fulmars may cross the equator into the opposite hemisphere, shearwaters and storm-petrels may go astray on migration into the wrong ocean, and the gadfly petrels of the genus Pterodromu, though rarely recorded anywhere near land away from the breeding stations, are occasionally capable of prodigious feats of wandering across several oceans and continents.

These last records are hard to explain, though some at least must be genuine; the best explanation appears to be that the birds are first displaced from their range by storms, and then have vast powers of endurance so that they are able to wander even further afield in attempting to return to it. They then often come to grief in circumstances where it may be very hard indeed to prove the manner of the appearance beyond all reasonable doubt, and there will always be a large element of uncertainty about many older records, including a number on current national lists, although many cannot be ignored entirely.