Primary Molt and Transequatorial Migration of the Sootyu Shearwater

Home , Great shearwater, Sooty shearwater

The Condor93:724-730 0 The CooperOmithological Society I99 1

PRIMARY MOLT AND TRANSEQUATORIAL MIGRATION OF THE SOOTY SHEARWATER ’

JOHN COOPER Percy Fitzpatrick Institute of African Ornithology, Universityof Cape Town, Rondebosch7700, South Africa

LESLIE G. UNDERHILL Department of Statistical Sciences,University of Cape Town, Rondebosch7700, South Africa

GRAHAM AVERY South African Museum, Box 61, Cape Town 8000, South Africa

Abstract. Examination of beachedcorpses of 244 Sooty Shearwaters(Pu@inus griseus) from the southwesternCape, South Africa, showed that simple descendentprimary molt occursfrom December to June, with most recordsbeing in the monthsof Februaryand March. Basedon percentagefeather mass grown, primary molt takesa mean of 159 days for completion,with mean estimatedstarting and completiondates of 2 Januaryand 10 June.Because the completionof breedingextends from late March to early May, the birds studiedcannot have bred successfullyin the previousaustral summer and were probably prebreeders.Unlike the situationin theNorth PacificOcean, Sooty Shearwaters in theNorth Atlantic Oceando not undergoprimary molt. The hypothesisthat SootyShearwaters of breedingage migrate preferentially into the North Pacific,whereas young birds first molt in the SouthAtlantic Oceanbefore migrating into the North Atlantic, needsto be testedby extensivebanding of birds at the breedinglocalities. Key words: SootyShearwater; Puffinus griseus; beached specimens; primary molt; transequatorial migration: South Africa.

INTRODUCTION Loomis 192 1, Marshall and Serventy 1956a, Beck Seabirds of the Order Procellariiformes are, with and Brown 1972, Brooke 1990). Such a delay has few exceptions, wholly pelagic outside their been attributed to the need for a complete set of breeding seasons when they cannot easily be flight feathers, with no increase in wing-loading studied in the hand. For this reason, there have and an assumed reduction in flight efficiency, been relatively few detailed studies of molt in while rapidly migrating long distances between procellariiforms, since for birds there is generally breeding and wintering areas (e.g., Stresemann little temporal overlap between the energetically and Stresemann 1970, Brown 1988). In contrast, costly activities of breeding and molting (e.g., closelyrelated specieswhich do not undergo such Payne 1972). Most procellariiforms undertake a long migrations may not experience such a delay post-nuptial molt of their primary feathers(Stre- in commencement of molt (e.g., Marshall and semann and Stresemann 1966), although excep- Serventy 1956a, Beck 1970, Beck and Brown tions to the lack of an overlap have been recorded 1972, Cramp and Simmons 1977, Brooke 1990). in the fulmarine petrels of the genera Daption, The difficulty of obtaining procellariiforms to Pagodroma, Fulmarus and Macronectes and in study their molt in the hand has led some work- the British Storm Petrel (Hydrobates pelagicus) ers to base their conclusions on observations (e.g., Harris 1966; Beck 1969, 1970; Hunter made from vessels of birds at sea seen to be 1984). actively molting their flight feathers (Watson It has long been known that procellariiforms 1971; Brown 1988, 1990). While useful, such which undertake long-distance migrations after studies do not allow the onset, completion and breeding delay molting their flight feathers until duration of molt to be accurately calculated, be- they have reached their wintering grounds (e.g., cause only the presence of obvious gaps in the wing (as illustrated by Watson 197 1 and Ogi et al. 198 1) can be used to denote active molt. Birds ’ ’ Received18 December1990. Final acceptance20 commencing and completing primary molt will February1991. not be so easily discerned and those with inter-

VW MOLT OF SOOTY SHEARWATERS 725

rupted molt (perhaps as many as half) will also by Underhill and Zucchini (1988). PFMG takes be overlooked (Brown 1988). The usual alter- into account the differences in the relative size native is to collect specimens at sea, either de- of the primaries and is used in preferenceto sum- liberately (e.g.,Stresemann and Stresemann 1970) ming the individual primary feather molt scores. or by examining birds caught incidentally during PFMG was assumed to increase linearly with fishing operations (e.g., Ogi et al. 198 1). For var- time and was used as a molt index in the Un- ious reasons, it is not always desirable or prac- derhill-Zucchini molt model. This model esti- tical to collect birds at sea. However, pelagic sea- mates the following parameters (and their stan- birds are prone to being ‘wrecked’ under adverse dard errors): duration of primary molt 7, mean conditions, when their corpses may be washed starting date K and the variability of starting dates ashore in numbers (e.g., Ryan et al. 1989). u. The mean starting date and the variability of Beachedspecimens offer an opportunity to study starting date are modeled as a normal distribu- the molt of procellariiform species which oth- tion with mean I.Land standard deviation u. Un- erwise are not readily available outside their der the assumptions of the model the distribution breeding seasons. of completion dates is also a normal distribution, The Sooty Shearwater (Pufinus griseus) is a with mean p + T and the same standard deviation procellariiform seabird which undergoeslong and u. The model makes it possible to estimate in- rapid transequatorial migrations between its tervals in which, say, 95% of the population com- Australasian and South American breeding menced or completed primary molt. grounds in the southern hemisphere and winter- The algorithm given by Underhill and Zuc- ing areasin the North Pacific and North Atlantic chini (1988) uses three types of data to estimate Oceans (e.g., Bourne 1956, Phillips 1963, Briggs theseparameters, data types 2 and 3 are relevant and Chu 1986, Brown 1988). Some birds may here. With data type 2, it is assumed that the winter south of the equator, including in south- observed molts consist of a random sample of em African waters (e.g., Stanford 1953, Phillips all birds on each sampling date, including birds 1963,Richdale 1963,RyanandRose 1989). Here, that have not yet commenced molt, thosein molt, we discussthe primary molt of Sooty Shearwa- and those that have completed molt. All three ters found dead on South African beachesin re- kinds of data are used to estimate the parameters lation to transequatorial migration. of molt. With data type 3, only birds in molt are used in parameter estimation, and the only as- METHODS sumption is that the birds observed in molt form Regular monthly patrols of selectedbeaches for a random sample of that fraction of the popu- dead seabirdshave been carried out in the south- lation which is actively molting on each sampling western Cape, South Africa since 1977 (Avery data: birds with no feathersin molt may be either 1989 and referencestherein). Sooty Shearwaters over-representedor under-represented in the ob- in fresh condition, and considered to have been served sample relative to the population without dead for no more than one week, had the molt altering the parameters estimated (Underhill et of their primary feathers scored on one wing, al. 1990; Underhill et al., in press).If the results following the standard O-5 scale of Ginn and for data types 2 and 3 coincide, the implication Melville (1983). Birds were not sexed or aged. is that the number of birds observed with com- Primary molt scores were also obtained from plete primaries (either old or new) is represen- nine birds shot at sea off the southwestern Cape tative of the proportions of such birds in the for dietary studies during March 1983 and Feb- population. For a fuller discussion of this point ruary 1984 (Jackson 1988, Fitzpatrick Institute, see Underhill et al. (1990) and Underhill et al. unpubl. data). (in press). Fully grown primary feathers were removed from eight birds, dried to constant mass at 50°C RESULTS in a forced-draught oven, and weighed to the A total of 653 Sooty Shearwaterswas found on nearest 0.1 mg. The dried masses for each pri- a cumulative total of 8,271 km of patrolled mary were averaged and expressedas a percent- shoreline of the southwesternCape, South Africa age of the total mass of all ten primaries. The from 1977 to 1990. Most birds (83.5%) were primary molt scoreswere then converted to per- found beached near Yzerfontein (33”21’S, centagefeather massgrown (PFMG) as described 18”08’E) and within False Bay (34”05’S, 18”10’E), 726 J. COOPER, L. G. UNDERHILL AND G. AVERY

TABLE 1. Seasonalvariation in the numbers and proportions of Sooty Shearwatersfound ashore and molting their primaries in the southwesternCape, South Africa, 1977-1990.

Jul Aug SOP Ott Nov Dee Jan Feb Mar APT May Jun TOti No. found on patrols 12 30 5 15 23 29 56 179 163 87 31 23 653 Total no. examined for molt 5 5 24 85 71 244 No. in molt 209 % molting i.0 2A.o ii.5 85.7

particularly in the months of January to April (3-l 15), no differencesin the seasonaloccurrence (74.3%, Table 1). of molting were noticeable between years. Primary molt scoreswere recorded from 244 Up to five primaries were found molting specimensduring the years 1977 to 1990, ofwhich (feathers incompletely grown or missing) at one 209 (85.7%) were actively molting their prima- time (Table 2). The mean number of primaries ries. Primary molt was simple and descendent in molt simultaneously decreased significantly and only one caseof interrupted molt (both worn with increasing molt score(x*~ = 52.8, P < 0.00 1, and fresh primary feathers present, but no miss- Table 2). ing or incompletely grown ones) was observed The dried mass of the tenth and outer primary (20 February 1986; molt score20). Twenty spec- was 5.3 times that of the innermost feather, rep- imens examined had not commenced and 15 had resenting nearly 19% of the total dried mass of completed primary molt. Nearly all those in pri- primaries (Table 3). Basedon the Underhill-Zuc- mary molt were found during the months of De- chini model using data type 2, the mean indi- cember to June, with three finishing their molts vidual duration of primary molt was estimated in August (molt scores46, 49 and 49). 72.7% of to be 159 days, with a 95% confidence interval molting birds were found in February and March of 143 to 175 days (Table 4). Mean estimated (Table 1). The seasonaldistribution of all beached starting and completion dates were 2 January specimensand those in primary molt are broadly and 4 June. An estimated 95% commenced pri- similar (Table 1). The proportions of examined mary molt between 13 November and 21 Feb- birds in active primary molt varied monthly from ruary and completed molt between 2 1 April and O-100% (Table 1). Perhaps becausethe numbers 30 July (Table 4). Nearly identical results, with of specimensfound varied greatly between years larger standard errors due to smaller sample size,

TABLE 2. Numbers and percentagesof Sooty Shearwaterssimultaneously molting a given number of primaries at three stagesof primary molt and mean numbers of primaries molting simultaneously at each stage in the southwesternCape, South Africa, 1977-1990.

Molt score NO. % l-15 NO. % 16-35 NO. % 3643

Number of primaries in molt simultaneously 0 0 0.0 1 0.7 0 0.0 1 2 8.3 13 9.4 10 22.7 2 2 8.3 92 6.1 25 56.9 3 12 50.0 31 22.0 7 15.9 4 7 29.2 : 0.7 ; 4.5 5 1 4.2 2.1 0.0 Mean number of primaries molting simultaneously 3.1 2.2 2.0 Sample size 24 141 44 MOLT OF SOOTY SHEARWATERS 727

TABLE 3. Mean massesof the primaries of the Sooty Unit, pers.comm.) and becausethere is no known Shearwater. taxonomic or morphological variation in the various breeding populations of the species Averag; IIII (SE) Relative mass (Murphy 1930, Cramp and Simmons 1977, Primary no. Sample size m (%) Jouanin and Mougin 1979) it cannot be stated 1 8 91.0 (4.4) 3.55 where the Sooty Shearwaters occurring in 2 105.6 (7.5) 4.11 3 ! 130.7 (9.8) 5.09 South African waters breed. Fledging of chicks 4 6 167.8 (10.9) 6.54 and departureof successfullybreeding adults takes 4 213.5 (4.5) 8.32 place from late March to early May and pre- 2 2 264.3 10.30 breeding birds depart from early February on- 4 312.4 (9.1) 12.17 wards at breeding localities in Australasia and ; 8 367.2 (7.4) 14.31 9 8 428.9 (23.4) 16.71 the Falkland Islands (Richdale 1944, 1963; 10 7 485.5 (21.2) 18.91 Woods 1975; Warham et al. 1982) Becausemost Sooty Shearwaters in South African waters undertake primary molt from mid-November to were obtained using data type 3 (Table 4). The mid-July, the visiting population must presum- fact that the parameters of molt estimated using ably contain a large proportion of prebreeders both data types 2 and 3 were essentially identical and possibly some failed breeders. This cannot indicates that mortality is independent of the be proven without recoveries of birds of known state of molt. It also indicates that there is no age, because Sooty Shearwaterscannot be aged tendency for Sooty Shearwatersto arrive in the on external morphological grounds (Cramp and southwestern Cape to commence primary molt, Simmons 1977), but the presenceofbirds in South or to leave as soon as molt is completed. There- African waters in all months of the year supports fore, movements of birds in and out of the south- this assumption. Age at first breeding is not ac- western Cape are apparently independent of the curately known for the Sooty Shearwater (Rich- state of primary molt. The three birds molting dale 1963, Warham et al. 1982) but is most com- in August (Table 1) were diagnosedby the model monly five to eight years in the closely related to be outliers relative to the remainder of the Short-tailed Shearwater(P. tenuirostris) (Serven- observations (Underhill and Zucchini 1988), and ty 1957, Wooller et al. 1988), which is also a’ therefore have not been included in the calcu- transequatorial migrant (Serventy 1953, Mar- lations. shall and Serventy 1956b). Loomis (19 18) reported molting Sooty Shear- DISCUSSION waters from late February off California and it Sooty Shearwatersare present all year round in must also be assumedthat thesebirds were either southern African waters, but are most abundant prebreeders or had failed breeding early in the during the austral winter (Richdale 1963, Jack- season. That young, nonbreeding and failed son 1988, Ryan and Rose 1989). Because no breeding procellariiform seabirds molt earlier banded birds have been recovered off southern (and sometimes faster) than do successfulbreed- Africa (T. B. Oatley, South African Bird Ringing ers has been shown variously for several species

TABLE 4. The parameters of molt for Sooty Shearwaters, estimated using the Underhill-Zucchini Model. Standard errors (in days) of estimates given in parentheses.

Standard deviation Dates between which Dates between Mean starting of starting dates 95% of birds start which 95% of date WYS) primary molt Duration (days) Mean completion date birds finish molt Data type 2 2 January (3.4) 25.7 (1.7) 13 November- 159.4 (8.0) 10 June (5.4) 21 April- 21 February 30 July Data type 3 2 January (5.6) 26.4 (2.2) 11 November- 159.0 (13.2) 10 June (8.3) 20 April- 17 February 31 July 728 J. COOPER, L. G. UNDERHILL AND G. AVERY

of fulmarine petrels and for the Greater Shear- lantic. A few individuals, assumed to be non- water (P. gravis) (Wynne-Edwards 1939; Carrick breeders,arrive off Newfoundland in late March and Dunnet 1956; Beck 1969, 1970; Stresemann (Rees 1964). and Stresemann 1970; Hunter 1984). Peak num- A newly fledged“juvenile” banded in the Falk- bers of Sooty Shearwatersoff California do not lands on 3 May 1962 and recovered at St. Philip, occur until April-May (Briggsand Chu 1986), at Barbados in the North Atlantic on 1 June the the end of the breeding season.Thus, most Sooty same year (Woods 1975) represents the only Shearwaters off California, and probably transequatorial movement of a banded prebreed- throughout the North Pacific, are likely to be of ing Sooty Shearwater. Transequatorial recover- breeding age. ies of adults banded in their burrows at Aus- In the North Pacific Ocean Sooty Shearwaters tralasian breeding localities (n = 10) are all from have been recorded in active primary molt from the North Pacific Ocean (Richdale 1957, Hitch- 22 June-7 August (Ogi et al. 198 1; but deduced cock 1961, Warham 1964, Hitchcock 1966, from a reduction in wing measurements, there- Warham et al. 1982, R. 0. Cossee,New Zealand fore resulting in an artificially delayed starting National Banding Scheme, pers. comm.; K. M. date). Brown (1988) observed the presence of Lowe, Australian Bird and Bat Banding Schemes, active primary molt in the first week of May off pers. comm.). A chick banded in New Zealand California (up to 72% of birds observed) and and recovered 10 years later in Oregon, U.S.A. from the end of April as soon as birds had arrived (Cossee 1989) also represents an adult move- off British Columbia, where by the end of May ment acrossthe equator. 90% showed signs of active primary molt. Chu The fact that many more Sooty Shearwaters (1984) recorded primary and secondary molt of migrate into the North Pacific than into the North 166 Sooty Shearwaterscollected at sea in Mon- Atlantic (Phillips 1963, Brown 1988) is not at terey Bay, California from May to September. variance with a migratory separation in two By the latter month over half those examined oceans of prebreedersand birds of breeding age, had completed molt. Loomis (19 18) recordedthe becausein an assumedlong-lived speciesthe lat- presence of primary molt off California as late ter category is certain to be the larger one. The as October. closely related Short-tailed Shearwater has been In complete contrast to the situation in the recorded as breeding over more than a 30-year North Pacific, 603 Sooty Shearwaters observed period (Wooller et al. 1988). at sea in the North Atlantic by Brown (1988, There are no previous published estimates of 1990) during May to November (with one ex- the duration of primary molt in the Sooty Shear- ception of a 30 May individual with a single water. Based on 112 examined specimens of the primary feather missing from one wing) had all Greater Shearwater, Stresemann and Strese- completed primary molt and were in fresh plum- mann (1970) estimated that primary molt took age. Mayaud (1949-1950) did not observe pri- about 90 days in adults of this species,which is mary molt in 13 specimens collected off France also a transequatorial migrant breeding in the between August and January which were in fresh southern hemisphere (Cramp and Simmons plumage, but did record primary molt in an early 1977). Stresemann and Stresemann (1970) also November specimen. Based on the state of their estimated that primary molt of adult Short-tailed gonads, Mayaud (1949-1950) considered his Shearwaters,while in the North Pacific (Marshall North Atlantic specimensto be mostly juveniles. andserventy 1956a, Palmer 1962) takes 82-100 On the above evidence and that of the present days. Brown (1988) considered that Greater study, it appears that Sooty Shearwatersmigrat- Shearwaterscompleted primary molt in only 40 ing into the South and North Atlantic Oceans days off Newfoundland, but this is likely to be are either prebreeders (as already postulated by an underestimate due to the observation tech- Brown 1988) and/or failed breeders, whereas nique used. Based on banding returns (Morant birds migrating into and molting within North 1977) it is clear that, apparently unlike Sooty Pacific waters are mainly adults of breeding age. Shearwaters, both juvenile and adult Greater Sooty Shearwatersthat arrive off the eastern sea- Shearwatersmigrate to the North Atlantic, where board of Canada in early to mid-May (Phillips they arrive in late May and early June and adults 1963, Brown et al. 198 1, Brown 1988) have pre- depart by the end of August (Stresemann and sumably completed their molt in the South At- Stresemann 1970, Brown 1988). The relatively MOLT OF SOOTY SHEARWATERS 129 short period spent by adults of this species on through the South African National Antarctic the wintering grounds (approximately two Research Programme. LGU acknowledges sup- months) means that they have to complete pri- port from the University of Cape Town and the mary molt quickly. Stresemann and Stresemann Foundation for Research Development. (1970) reported up to five primaries molting simultaneously in the Greater Shearwater. Mei- nertzhagen (1956) observed Greater Shearwaters LITERATURE CITED in “full wing moult, many unable to fly” in June AVERY,G. 1989. Results of patrols for beached sea- and August off Greenland, and Voous (1970) ob- birds conducted in southern Africa in 1984 and served a near-flightless bird with “all primaries 1985. Cormorant 1757-71. BECK,J. R. 1969. Food, moult and age of first breed- being short” in August in the Bay of Biscay. ingin the Cape Pigeon,Daption capensisLinnaeus. Greater Shearwaters,presumably young birds or Br. Antarct. Surv. Bull. 21:33-44. failed breeders, can also molt in the southern BECK, J. R. 1970. Breeding seasonsand moult in hemisphere as early as January (Watson 197 1). some smaller Antarctic petrels, p. 542-550. In M. W. Holdgate [ed.], Antarctic ecology. Vol 1. Ac- Stresemann and Stresemann (1970) found that ademic Press, London. juvenile Greater Shearwaters commenced pri- BECK,J. R., AND D. W. BROWN. 1972. The biology mary molt earlier than did adults in the North of Wilson’s Storm Petrel, Oceanites oceanicus Atlantic. (Kuhl), at Signy Island, South Orkney Islands. Br. It is assumed that Sooty Shearwatersmolting kntarct. Su&.Sci. Rpts. 69:1-54. BOURNE.W. R. P. 1956. Miarations of the Sootv in southern African waters during December to Shearwater. Sea Swallow 9:??3-25. June are birds of prebreeding age or are failed BRIGGS,K. T., AND E. W. CHU. 1986. Sooty Shear- breeders which may then migrate across the waters off California: distribution, abundanceand equator into the North Atlantic. Their appar- habitat use. Condor 88:355-364. BROOKE,M. 1990. The Manx Shearwater. T. & A. ently leisurely molt contrasts with that known D. Poyser, London. for adult Greater and Short-tailed Shearwaters. BROWN,R. G. B. 1988. The wing-mot& of Fulmars It remains to be investigated whether success- and shearwaters(Procellariidae) in Canadian wa- fully breeding Sooty Shearwatersmolt their pri- ters. Can. Field-Nat. 102:203-208. mary feathers more quickly, in a manner akin to BROWN,R. G. B. 1990. The wing-mot& of Cory’s Shearwater, Calonectrisdiomedea, off Nova Sco- that of Greater and Short-tailed Shearwatersand tia. Can. Field-Nat. 104:306-307. successfullybreeding fulmarine petrels. To this BROWN,R. G. B., S. P. BARKER,D. E. GASKIN,AND end, beached corpsesof Sooty Shearwatersin the M. R. SANDEMAN.198 1. The foods of Great and North Pacific should be examined quantitatively Sooty ShearwatersPujinus gravis and P. griseus in eastern Canadian waters. Ibis 123:19-30. for primary molt. The three Sooty Shearwaters CARRICK,R., ANDG. M. DUNNET. 1956. Breeding of molting in South African waters in August may the Fulmar Fulmarus glacialis. Ibis 96:356-370. have belonged to a different subpopulation to CHU. E. W. 1984. Sootv Shearwatersoff California: that of the remainder of the molting individuals diet and energygain, p. 64-84. In D. N. Nettleship, and could have been birds which had bred suc- G. A. Sanger, and P. F. Springer. [eds.], Marine birds: their feeding ecology and commercial fish- cessfully the previous austral summer. eries relationships. Can. Wild. Serv. Spec. Publ., To elucidate further whether prebreedingSooty Ottawa. Shearwaterspreferentially head north in the At- COSSEE,R. 0. 1989. Report on bird banding in New lantic Ocean whereasadults migrate to the North Zealand 1988-1989. New Zealand Dept. Conserv. Sci. Res. Ser. 19:l-24. Pacific requires concerted efforts to band large &, S., AND K. E. L. SIMMONS[EDS.] 1977. The numbers of both fledglings and breeding adults birds ofthe westernPalearctic. Vol. 1. Oxford Uni- at different breeding colonies in the southern versity Press, Oxford. hemisphere. G~NN,H. B., AND D. S. MELMLLE. 1983. Moult in Patrols for beachedseabirds in southern Africa birds. Br. Trust Om. Guide 19:l-l 12. HARRIS,M. P. 1966. Breeding biology of the Manx are supported by the African Seabird Group and ShearwaterPuJinus pufinus. Ibis 108:17-33. the South African Museum. We thank R. K. HITCHCOCK,W. B. 1961. Sixth annual report of the Brooke, R. G. B. Brown, R. 0. Cossee, K. M. Australianbird-banding scheme, July 1959 to June Lowe, T. B. Oatley, P. G. Ryan and J. Warham 1960. CSIRO Wildl. Res. 6:56-87. HITCHCOCK,W. B. 1966. Tenth annual report of the for their records and/or for comments on this Australianbird-banding scheme,July 1963 to June paper. JC acknowledgesthe support of the South 1964. Div. Wildl. Res. Tech. Pap. 11:l-49. African Department of Environment Affairs HUNTER,S. 1984. Moult of the giant petrels Macro- 730 J. COOPER, L. G. UNDERHILL AND G. AVERY

necteshalli and M. giganteusat South Georgia. water Pujinusgriseus.Proc. Zool. Sot., Lond. 141: Ibis 126:119-132. 1-117. JACKSON,S. 1988. Diets of the White-chinned Petrel RYAN, P. G., G. AVERY,B. ROSE,G. J. B. Ross, J. C. and Sooty Shearwater in the southern Benguela SIN-, ANDC. J. VERNON. 1989. The Southern region, South Africa. Condor 90:20-28. Ocean seabird irruption to South African waters JOUANIN,C., ANDJ.-L. MOUGIN. 1979. Order Procel- during winter 1984. Cormorant 17:41-55. lariiformes, p. 48-121. In E. Mayr, and G. W. RYAN, P. G., AND B. ROSE. 1989. Migrant seabirds, Cottrell [eds.],Check-list of the birds of the world. p. 274-287.ZnA.I.L. Payne,andR.J.M. Crawford Vol. 1, second edition. Museum of Comparative [eds.], Oceansof life off southernAfrica. Vlaeberg, Zooloav. Cambridge. MA. Cape Town. _- I LOOMIS,L. M. 19 18. Expedition of the Academy of SERVENTY, D. L. 1953. Movements of pelagic sea- Sciencesto the GalanaaosIslands. 1905-1906. XII. birds in the Indo-Pacific region. Proc. 7th Pacific A review of the aiba%osses,petrels and diving Sci. Congr. 4:394-407. petrels Proc. Calif. Acad. Sci, Ser. 4, Vol. 11(12): SERVENTY, D. L. 1957. Duration of immaturity in 1-187. the Short-tailed Shearwater, Pufinus tenuirostris LOOMIS,L. M. 1921. Remarks on the migration of (Temminck). CSIRO Wildl. Res. 2:60-62. southernhemisphere albatrossesand petrels. Auk STANFORD,W. P. 1953. Some sea birds in winter off 38:527-531. the S. W. Cape. Ostrich 24:17-26. MARSHALL,A. J., ANDD. L. SER~ENTY.1956a. Moult STRESEMANN,E., AND V. STRESEMANN. 1966. Die adaptation in relation to long-distance migration Mauser der Vijgel. J. Om 107 Sonderheft:1445. in oetrels. Nature 177:943. STRESEMANN. E., AND V. STRESEMANN.1970. Uber MARSHALL,A. J., AND D. L. SERVENTY. 1956b. The Mauser und Zug von Pufinus gravis.J. Om. 111: breeding cycle of the Short-tailed Shearwater,Puf- 378-393. jnus tenuirostris(Temminck), in relation to trans- UNDERHILL,L. G., R. P. Pays-JONES,R. J. DOWSETT, equatorial migration and its environment. Proc. M. R. LAWN, P. HERROELEN,D. N. JOHNSON,S. Zool. Sot., Land. 127:489-510. C. NORMAN,D. J. PEARSON, AND A. J. TREE. In MAYAUD, N. 1949-1950. Nouvelles precisionssur la press. The biannual moult of Willow Warblers mue des Procellariens. Alauda 17/18:144-155, Phylloscopustrochilus in Europe and Africa. Ibis. 222-233, UNDERHILL,L. G., ANDW. ZUCCHINI. 1988. A model MEINERTZHAGEN,R. M. 1956. Birds in Greenland. for avian primary moult. Ibis 130:358-372. Bull. Br. Om. Club 76: 17-22. UNDERHILL,L. G., W. ZUCCHINI,AND R. W. SUMMERS. MORANT,P. D. 1977. Bird ringing at Tristan da Cu- 1990. A model for avian primary moult-data types nha and Gough Islands. S. Afr. J. Antarct. Res. 7: based on migration strategies, illustrated by an 23-26. example using the Redshank Tringa totatus. Ibis MURPHY, R. C. 1930. Birds collected during the 132:<18-123.- ’ Whitney South Sea Expedition. XI. Amer. Mus. Voous. K. H. 1970. Moultine Great Shearwater(Puf- Novit. 419:1-15. j&s gravis) in Bay of Biscay. Ardea 58:265-266. GGI, H., K. SHIMAZAKI,AND K. NAKAMURA. 1981. WARHAM,J. 1964. Marked Sooty ShearwatersPuf- Sooty Shearwatersin the subarcticNorth Pacific: jnus griseusin the northern hemisphere. Ibis 106: seasonalchanges in body weight and molt. Res. 390-391. Inst. N. Pac. Fish., Hokkaido Univ. Spec. Vol.: WARHAM,J., G. J. WILSON,AND B. R. KEELEY. 1982. 207-215. The annual cycle of the Sooty ShearwaterPuffinus PALMER,R. S. [ED.] 1962. Handbook of North Amer- griseusat the Snares Islands, New Zealand. No- ican birds. Vol. 1. Loons through flamingos. Yale tomis 29:269-292. University Press, New Haven and London. WATSON,G. E. 1971. Molting Greater Shearwaters PAYNE, R. B. 1972. Mechanisms and control of molt, (Pufinus~avis) off Tierra de1Fuego. Auk 88:440- p.‘104-155. In D. S. Famer, and J. R. King [eds.], 44% - Avian biology. Vol. II. Academic Press,New York WOODS, R. W. 1975. The birds of the Falkland Is- and London. lands, Anthony Nelson, Oswestry, U.K. PHILLIPS,J. H. 1963. The pelagic distribution of the WOOLLER,R. D., J. S. BRADLEY,D. L. SERVENTY, AND Sootv Shearwater.Procellaria prisea. Ibis 105:34& I. J. SKIRA. 1988. Factors contributing to reuro- 353: ductive successin Short-tailed Shearwiters @uf- REES,E. I. S. 1964. The Sooty ShearwaterProcellaria jnus tenuirostris),p. 848-856. In H. Ouellet [ed.], griseaon the Newfoundland Banks. Ibis 106:118- Proceedingsof the 19th International Omitholog- 119. ical Congress.Univ. Ottawa Press, Ottawa. RICHDALE,L. E. 1944. The Sooty Shearwaterin New WYNNE-EDWARDS,V. C. 1939. Intermittent breeding Zealand. Condor 46:93-107. of the Fulmar (Fulmarus glacialis(L.)), with some RICHDALE,L. E. 1957. Recovery of Sooty Shearwater generalobservations of non-breedingin sea-birds. in northern hemisphere. Ibis 99:116. Proc. Zool. Sot., Lond. Ser. A, 109:127-132. RICHDALE,L. E. 1963. Biology of the Sooty Shear-