The Condor 90:220-227 0 The Cooper Ornithological Society 1988

STEPWISE MOLT OF REMIGES IN BLUE-EYED AND KING SHAGS

PAMELA C. RASMUSSEN Museumof Natural Historyand Departmentof Systematicsand Ecology, Universityof Kansas,Lawrence, KS 66045

Abstract. In Blue-eyed and King shags( atriceps and P. albiventer),the first prebasic molt of the primaries begins at Pl and proceedsdistally. Continual stepwise molt of the primaries begins with commencement of PB2 after replacement of P5 to P8 by the first prebasic molt. Adults had from one to four concurrent distally moving waves of molt in primaries. In secondariesof these ,the first prebasicmolt beginsperipherally at S 1 and S 15 and proceeds centrally. The second prebasic molt then begins peripherally before the central secondariesare replaced by PB 1. Secondariesof adults’ are molted by up to six concurrent waves that originate peripherally; I term this bidirectional stepwisemolt of secondaries.Stepwise molt tends to be asymmetrical both in adults and subadults.Sub- adults of these specieshave more molting per wave in primaries and secondaries than do adults, but the age classesdo not differ in number of molting feathers per wing. Molt of remiges in subadultsinitially resemblesthe condition in the Procellariiformes, the nearest outgroup,while the stepwisemolt of adults is synapomorphicfor Pelecanifonnes. Key words: Blue-eyedShag; King Shag;Phalacrocorax atriceps; Phalacrocorax albiven- ter; molt.

INTRODUCTION eventually reachesthe wing tip (Ashmole 1968) Although the wing molts of (Phala- but not during the year in which the wave began crocoracidae) are highly complex and unusual, (at least not in the Masked , Sula dacty- detailed information on molts of primaries in lutru, Dorward 1962; or the ,Potts this cosmopolitan is available only for the 197 1). Each wave probably continuesfrom where European Shag (Phalacrocorax aristotelis; Potts it paused; primaries are replaced once per year 197 1). All speciesof Pelecaniformes studied thus or less in the (Dorward 1962). far have shown stepwise molt (called “Stajd- The other known type of stepwisemolt is “per- nz2iuser”by Stresemann and Stresemann 1966, iodic stepwise molt” (Stresemann and Strese- and “serially descendent molt” by Ginn and mann 1966) in which successivewaves begin at Melville 1983) of the primaries (Dorward 1962, the innermost primary, but only one of every Potts 1971, Berry 1976, DeKorte and DeVries two, or two of three waves reach the wing tip. A 1978, Bernstein and Maxson 198 1, Crawford et variable number of the outermost primaries are al. 1982, Ginn and Melville 1983, Cooper 1985, thus molted less frequently than the innermost, Rasmussen 1987). However, very little is known although every primary is molted at least once concerning the ontogeny and mode of progres- yearly. Periodic stepwisemolt is known only in sion of stepwise molt in primaries. In stepwise some terns of the Sterna (Stresemann and molt of primaries, a molt wave begins at a certain Stresemann 1966). point (in most speciesPrimary 1 [Pl]) before the The molt of secondariesin Pelecaniformes has previous wave, which originated at the same been virtually ignored; published data on molt point, is completed, thus there is more than one patterns of secondaries in members of this di- concurrent, overlapping molt wave (seeFigs. 1A, verse exist only for the (P. ma- B). “Continual stepwise molt” occurs in Pele- gellanicus;Rasmussen 1987). Few data have been caniformes and a few other (Stresemann published on molts of the Blue-eyed Shag and and Stresemann 1966); in this type of stepwise the King Shag morph (P. atriceps and P. “albi- molt successivewaves begin at Pl. Each wave venter”; Watson 1975, Bernstein and Maxson 1981). In the Antarctic Blue-eyed Shag (P. atri- ceps bransfieldensis)molt of flight feathers ap- I Received15 June 1987. Final acceptance2 October pearsirregular, with no obvious orderly sequence 1987. of molt (Bernstein and Maxson 198 1).

VW MOLT OF SHAGS 221

IIIIII 111 11 II 11 1 II 11 1 I u 52-c (n +3) - 2 (n t-2) ____-__ g I -- (nt I) __-___..__-_* I I I I I I I I II I I$ (n) -__--______c 12345678910 Proximal Distal Illllrlllll Primaries 12345678910 Proximal Distal Primaries FIGURE 1. Stepwisemolt of primaries. Arrows indicate apparent direction and extent of wave progression; dashed portions of arrows indicate feathers of that round presumed replaced by later round(s). (A) Subadult King Shag, YPM 82275: commencement of stepwise molt. (B) Adult Blue-eyed Shag, AMNH 443237; four- round stepwisemolt.

In this paper I document the ontogeny of step- sultsin remigesof varying agesthat are not readi- wise molt of the remiges in Blue-eyed and King ly classifiableas “new” or “old” (pers. observ.); shags,and describe and compare characteristics for this reason I did not use Ashmole’s (1962) of continual stepwisemolt of primaries and bi- standard molt scoring system. directional stepwise molt of the secondaries in these cormorants. TABLE 1. Number of specimens of Blue-eyed and Ring shagsexamined for each season,age class, geo- METHODS graphic region, and specimen nature. I categorizedspecimens in first prebasic molt as BILE subadults, and those that had molted all juvenal eyed King feathers as adults. Total numbers and nature of Shag Shag specimens examined for each season, age class, Season and geographic region are given in Table 1. Winter (Jun-Aug) 4 9 Eighty-five (88%) of the 97 Blue-eyed Shag spec- Spring (Sep-Nov) 10 10 imens examined were molting; 49 (72%) of the Summer (Dee-Feb) 60 39 23 10 68 King Shagswere molting. Fall (Mar-May) Primaries (excluding the remicle, Pll) were Age classes numbered Pl to PlO from the carpal joint dis- Juveniles 14 17 Subadults 33 tally, and secondariesS 1 to S 15 (in two casesto Adults 50 :: S17) from the carpal joint to the humero-ulnar joint. Molt data were taken from all remigeswhen Geographicregion Southern 51 48 possible. Each remex was assignedto an age cat- 0 15 egory O-10 (1-5 are also growth categories):(0) South Georgia Island 10 0 juvenal; (1) missing; (2) new just visible; Shag Rocks (50”33’S, 42”02’W) 1 0 (3) less than ‘/4 grown; (4) l/4 to % grown; (5) South Shetland Islands 17 0 Antarctic Peninsula 18 0 greater than % to not quite fully grown; (6) full- Macquarie Island 0 5 length, no wear or fading; (7) very slight wear and/or fading; (8) light but obvious wear and Specimen nature fading; (9) moderate wear and fading; ( 10) heavy Museum skin 70 60 Freshly-killed 27 8 wear and fading. This system reduces error due Totals 97 68 to the extended molt of cormorants, which re- 222 PAMELA C. RASMUSSEN

To determine whether the first prebasic molt Shags, PB2 had not yet begun by the time P7 (PB 1) of primaries or of secondariescommenced was replaced by PB 1. In subadult King Shagsin first, and whether in the secondaries PBl first which PB2 had begun, PB2 started when juvenal commenced distally or proximally, I assumed P6 (n = 1) and P7 (n = 1) had been replaced by that all secondaries and the primaries grow at PB 1. The beginning of PB2 marks the onset of approximately equal rates (as primaries appar- stepwise molt, because from that point on PBl ently do in the Masked Booby, Dorward 1962; and PB2 continue simultaneously (Fig. 1A). and in the Great [P. C&JO] and Eu- In most cases,juvenal primaries were replaced ropean Shag, Ginn and Melville 1983). sequentially, with PI0 the last to be replaced. I considered molt of remiges symmetrical if However, omissive molt (Cannel1 et al. 1983) feathers of each bilateral pair (L5, R5) were in occurred occasionally; in one Blue-eyed Shag, the same age category or were molting simulta- juvenal P8 and P9 were retained on both wings neously. Symmetry of remiges was expressedas after PI0 had been replaced, and in another ju- the percent of individuals examined with com- venal P7 and P8 were retained on one wing and pletely symmetrical molt. juvenal P7 on the other after more distal pri- A “molt cycle” is the period from the begin- maries had been replaced. In the King Shag,ju- ning of one molt to the beginning of the next venal P6 was retained on both wings of one spec- homologous molt (Humphrey and Parkes 1959). imen, juvenal P7 on both wings of another, and Becausein stepwise molt homologous molts of juvenal P6 on one wing and P8 on the other wing remiges overlap temporally, I use the term of a third, all after more distal primaries had been “round” for the complete replacement of a feath- replaced. er row by one unidirectional wave (as in pri- Adults of both morphs showed stepwisemolt maries), or by two waves that begin more or less of primaries from proximal to distal (Fig. 1B). simultaneously distally and proximally (as in sec- From zero to four waves occurred in primaries ondaries); one round is thus the replacement of of each wing in adults of both morphs (Fig. 2). all feathers of a feather row (whether primaries In 46 adult Blue-eyed Shags, x = 1.93 -t 1.12 or secondaries)by one molt cycle. A “molt wave waves per wing, and in 3 1 adult King ShagsK = front” is the most recently molted feather in a 2.39 f 0.88 waves per wing. series of feathers of decreasing age (of a single Subadult Blue-eyed Shags had significantly wave of molt). The number of molt waves in more primaries molting per molting wave than remiges of each wing equals the number of wave did adults (Table 2). However, adults and sub- fronts (PlO was considered a wave front when it adults did not differ statistically in number of was very new or molting). primaries molting per wing (Table 2). Symmetry in molt of primaries between wings of individuals tended to be more common in RESULTS subadults than it was in adults, although the dif- Data on molt of remiges (coded in numerical age ferenceswere not statistically significant (Table categories)for all individuals used in this study 3). are available on request from the author. SECONDARIES PRIMARIES PBl of secondariesin both morphs began at the In 24 casesin subadult Blue-eyed Shags (both most distal and most proximal secondaries,and wings in 10 birds, one wing in four birds), single- continued to the central secondaries(for an ex- wave molt occurred in which the most distal ample, seeFig. 3A). In subadult Blue-eyed Shags nonjuvenal primary was the newest and the most in single-round (one-wave) molt of secondaries, proximal (PI) was the oldest; this occurred in 18 two individuals had molt waves only in the distal casesin subadult King Shags(both wings of nine secondaries(of both wings); four had molt waves birds). This shows that in subadults of both only in the proximal secondaries (two in both morphs PB 1 (the first prebasicmolt) commences wings; two in one wing). In those in single-round at Pl and proceedsdistally. (two-wave) molt of secondaries, 10 had molt In subadult Blue-eyed Shagsin which the sec- waves in both the distal and proximal second- ond prebasic molt (PB2) had begun, PB2 started aries. In single-round subadult King Shags,one at Pl when PBl had reachedjuvenal P5 (n = I), individual was in single-wave molt of the distal P6 (n = I), or P7 (n = 1). In seven Blue-eyed secondaries(both wings); five were in two-wave MOLT OF SHAGS 223 molt, with waves in both the distal and proximal 50 Primaries secondaries.In all of the waves, the wave front 40 (centralmost) feather was either newer than or as 30 new as any feather in the wave. 2 20 Of the six subadult Blue-eyed Shagsin single- E” IO wave molt of secondaries,the highestnumber of ‘b Q 0 I 2 3 4 feathers replaced by PB 1 was five distal second- aries; in the only King Shagin this category,only Secondaries two distal secondarieshad been replaced. Assuming that all secondariesgrew at equal rates, in 11 Blue-eyed Shags in two-wave PBl the distal wave started first in five individuals (both wings); the proximal wave first in both 0123456 wings of two; and both waves simultaneously in No. waves both wings of two. In one specimen, the distal wave started first in one wing, the proximal in FIGURE 2. Number of waves in primaries and sec- another; in another specimen the proximal wave ondariesper wing in right wingsof adult Blue-eyedand Ring shags.Left wings were used only if the right wing started first in one wing and both waves started of that specimencould not be used.Cross-hatched bars simultaneously in the other wing. In four King represent Ring Shags;solid bars represent Blue-eyed Shagspecimens in two-wave PB 1, the distal wave Shags.Numbers above bars indicate n for each cate- started first in both wings, the proximal started gory. first in both wings of two others; and in one specimen the proximal started first in one wing, and both waves started simultaneously in the secondariesper wing and no juvenal secondar- other wing. ies), eachsecondary occasionally appeared to molt In three casesin each morph, from one to four first (of all the secondaries)after molt pauses(Fig. more peripheraljuvenal secondaries(SlS5, SlO, 4). The observation that any secondary was oc- S 13) were retained after more central secondaries casionally the first to be molted after molt pauses, had already been replaced by PB 1. and no secondarywas overwhelmingly the most I interpreted data on relative ages of feathers common to molt first, showsthat in adults molt as showing the beginning of PB2 in one Blue- centers do not lie between Sl and S15. eyed and one King shag specimen (for an ex- In adult Blue-eyed Shags,there were nine cases ample, see Fig. 3B). In the Blue-eyed Shag spec- in which a new wave had just begun at the most imen, age categories(distal to proximal) of the distal secondary,and two at the most proximal, secondarieswere: (left wing) 2 1 8 8 2 0 0 0 0 0 while more central wave fronts (which appar- 0 0 0 0 2; (right wing) 7 2 8 8 1 0 0 0 0 1 8 8 8 ently had distal points of origin) were molting. 8 1. In the King Shag specimen, the categories In adult King Shags,there were five casesof this were: (left wing) 7 6 6 8 7 7 5 0 0 0 0 5 6 6 6. in the most distal secondaryand two in the most In adult Blue-eyed and King shagswhich had proximal. This shows that simultaneous waves apparently just resumed molt of secondariesafter occur in adults of both morphs, and that waves a pause (with at least 12 relatively new or older frequently start at S 1 and S 15.

TABLE 2. Comparisonsbetween molts of remiges of adult and subadult Blue-eyed Shags.

Subadults Adults x k SD n Range R I? SD n Range P=

No. primaries molting/waveb 1.4 * 0.1 63 l-3 1.1 * 0.3 101 l-2 *** No. primaries molting/wingc 1.6k0.7 63 l-3 1.9 f 0.9 56 l-4 n.s. No. secondariesmolting/waveb 1.4kO.6 31 l-3 1.1 + 0.3 93 l-2 * No. secondariesmolting/wingc 2.2 f 1.0 25 l-4 1.9 2 1.1 49 l-6 n.s.

*Mann-WhitneyU-tests between subadult and adult Blue-eyed Shags. b For waves in which at least one feather was moltmg; n = number of waves examined. EFor feather rows in which at least one feather was molting; n = number of wings examined. n.s. = P > 0.05, * = P < 0.05, ***p < 0.001. 224 PAMELA C. RASMUSSEN

$1 * - (n +3) F e 2 , < (n+2)- - Illlllllllrll 2 (n +I) ---- * c_-___ 151413121110987654321 (n) ____-_-_____--_)y

~~IIIIIIIIIIII~ 2222OOOcl 1514131211 109 8 7 6 5 4 3 2 I h -0-o .- .- .- ._ 5 ~-0U-o Proximal Distal ~g.paaEaPgfgg

I I I I I I I I I I I I I I1 I5 141312 II IO 9 8 7 6 5 4 3 2 I Distal Proximal Secondaries FIGURE 3. Stepwise molt of secondariesin Blue-eyed Shags.For explanation of conventions, see Figure 1. (A) subadult, KUMNH 8 1222, single-round molt. (B) subadult, KUMNH 8 1221; commencement of stepwise molt. (C) adult, AMNH (RHB 27 14); multiple-round molt (exact number of roundsand direction of centralmost wave cannot be determined).

Adults of both morphs had from one to six Symmetry between wings of each specimenap- waves per wing in secondaries(Figs. 2,3C; mean peared more common in molt of secondariesin number of waves in right wings of Blue-eyed subadult Blue-eyed and King shagsthan in adults, Shagswas 2.26 k 1.42 [n = 501; mean number but the differences were not statistically signifi- in King Shagswas 2.56 k 1.73 [n = 231). cant (Table 3). Although symmetry appearedless Subadult Blue-eyed Shags had significantly common in molt of secondariesthan in primaries more molting feathers per wave in secondaries for each age class and morph, none of the dif- than did adults (Table 2). However, numbers of ferenceswere statistically significant (Table 3). secondariesmolting per wing did not differ be- tween subadult and adult Blue-eyed Shags(Table MOLT DURING NESTING 2); this implies that subadults had marginally Five adult Blue-eyed Shags collected while nest- fewer waves per wing than adults. ing were molting from O-4 (X = 2.2 4 2.5) pri-

TABLE 3. Percent (n) of Blue-eyed and King shagswith complete symmetry in molt of remiges.

Blue-eyed Shag King Shag Subadults Adults P Subadults Adults P

Primaries 13 (20) 6 (38) 2.9 n.s. 4 (9) 2 (17) 0.6 n.s. Secondaries 6 (15) 3 (38) 0.9 ns. 2 (9) 0 (12) 0.2 n.s. P 0.5 n.s. 0.1 n.s. 0.2 ns. 0.1 n.s.

axi tests,df= I; n.s. = P > 0.05. MOLT OF SHAGS 225 maries and from O-3 (X = 1.4 + 1.5) secondaries. 20 One adult King Shag was molting two primaries 1 and another was molting two secondarieswhile nesting. One subadult King Shagwas molting one ,[;t primary and three secondarieswhile nesting.

DISCUSSION l23456789101112131415 PRIMARIES 5 Stepwisemolt of primaries of Blue-eyed and King King shags is similar to that of the European Shag (Potts 1971). The second prebasic molt begins IT? when the first has replaced P8 in the European Shag (n not given, Potts 1971) the Great Frig- atebird (Fveguta minor, n = 3, DeKorte and 0 123456789101112131415 DeVries 1978), the (Sulu leuco- guster,n = 1, Dorward 1962) or P7 in the Masked Secondary No Booby (n not given, Dorward 1962). The Blue- FIGURE 4. Frequenciesin which each secondaryis eyed Shagis the only continual stepwise-molting very new or molting (in adults in which at least 12 speciesyet studied for which variation is known secondariesper wing are relatively new or older). in timing of the beginning of PB2 in primaries in relation to progressof PBl, but it probably varies in other species as well. Blue-eyed and stepwisemolt of the remiges in adult cormorants King shagsoccasionally have more molt waves cannot be consideredcompleted at any time, since in primaries than occur in the the Great Cor- new waves begin before completion of the pre- morant, the (P. neglectus),and vious waves. the Rock Shag (the latter three have one to two waves, three rarely; Ginn and Melville 1983, SECONDARIES Cooper 1985, Rasmussen 1987). The Red-legged Although Ashmole (1968) suspected that the Shag (P. gaimardi; unpubl. data) and the Euro- Fairy Tern (Gygis alba) has stepwisemolt of sec- pean Shag(Potts 197 1) also have four molt waves ondaries, this has been confirmed only for the in the primaries. Rock Shag (Rasmussen 1987) and bidirectional Data from Blue-eyed and King shagsconfirm stepwise molt has not been described. On the Stresemann and Stresemann’s (1966) assertion basis of the following evidence from the molts that continual, not periodic, stepwisemolt occurs of Blue-eyed and King shags,I extend the defi- in cormorants. I did not find specimens with nition of stepwisemolt to include “bidirectional growth sequencesin primaries in which a more stepwise molt of secondaries:” (1) molt centers proximal, new wave front was about to overgrow (points from which molt waves regularly begin) an older, more distal wave front, as would be occur only at Sl and S15, and thus molt must predicted for periodic stepwise molt. Also, age be bidirectional; (2) up to six simultaneous waves discontinuities frequently occurredwithin waves occurwithin secondariesof one wing (a new wave (due to molt pauses)in both morphs; thesewould often had just begun in the most proximal or not occur in periodic stepwisemolt. most distal secondary, while growth was occur- Murphy (1936) was incorrect in stating that ring at more central wave fronts which had distal the new primaries of the King Shag do not reach points of origin; in the absence of central molt maximum length until July or August; in this centers,this demonstratesthe occurrenceof step- morph, as in the Blue-eyed Shag,primaries reach wise molt); (3) the ontogeny of molt of second- maximal length during all seasonsof the year. aries is similar to that of stepwise molt of pri- Bernstein and Maxson (198 1) noted that an Ant- maries, in that the second round begins before arctic Blue-eyed Shag had some old remiges in completion of the first round, thus initiating two- July, “probably indicating that a complete moult round molt (the first prebasicmolt of secondaries had not occurred in June.” This may be ex- consists of one round of two waves, instead of plained by the fact that, although pausesoccur, one wave as in the primaries). 226 PAMELA C. RASMUSSEN

CONSEQUENCES OF STEPWISE MOLT but do not differ from them in number ofmolting In the Fairy Tern (a continual stepwise-molting feathersper wing; this implies that subadultshave species)each primary is apparently replacedevery fewer waves per wing than do adults (Table 2). year, each molt wave pausesfor the seasonat the Subadultsalso commence molt with a singlewave point where the next more distal wave began that of descendentmolt. Therefore their molt is more season,and molt wavescontinue where they leave similar to that of nonstepwise molting birds, in- off (Ashmole 1968). It is thought that there are cluding the Procellariiformes, which have single- more waves per wing in older individuals of step- wave descendent molt from Pl to PlO (Strese- wise-molting species (DeKorte and DeVries mann and Stresemann 1966, Hunter 1984). 1978), as Dorward (1962) found for the Masked Single-wave descendent molt of primaries is Booby. However, the number ofwaves must stay thought to be the primitive mode of molt of pri- the same if every feather is replaced exactly once maries in birds (Stresemann and Stresemann per year and all waves start at Pl. This is the 1966), and Procellariiformes are often consid- case in the European Shag, in which not all pri- ered the most appropriate outgroup to the Pel- maries are replaced each year (Potts 197 l), and ecaniformes (Cracraft 1985). This and the on- the occasional occurrence of very old primaries togeny of primary molt in shags shows that in adult Blue-eyed and King shags shows that continual stepwisemolt in the Pelecaniformes is they also do not replace all primaries yearly. The derived. Since all pelecaniforms in which molt ontogeny of stepwise molt in the Fairy Tern is has been studied show continual stepwise molt not known, but it must differ from that of these of primaries, this type of molt is a synapomorphy cormorants if the molt features listed above are (shared derived character) for this order. Early invariant in this tern. Pelecaniformes are known from the Eocene of Berry (1976) statedthat a disadvantageof step- Wyoming and Britain (Harrison and Walker wise molt in cormorants is that outer primaries 1977, Olson 1977), and tropical conditions such are retained longer, but failed to note that this as existed at that time (Martin 1983) seem to disadvantageis restrictedto subadults.However, favor stepwise molt, because several Recent in adult cormorants wing tip primaries are usu- tropical groups (e.g., some tropical Tinami- ally replaced each year by a different round, so fonnes, Galliformes, and Columbiformes; Opis- at any time there are few or no very worn pri- thocomus, and Colius; Stresemann and Strese- maries, and these are not restricted to the wing mann 1966) have evolved it independently. tip. The Pelecaniformes are the only group with Re- Potts (197 1) assertedthat the disadvantage of cent nontropical representatives that have con- subadults retaining worn outer primaries is pre- tinual stepwise molt. Stepwise molt thus likely sumably outweighed by the advantages of their evolved in a pelecaniform ancestor during the having a slow molt that is energetically relatively early tropical period of their history, and as far nondemanding. However, becausemost species as known has been retained by all descendent of cormorant molt to some extent during breed- species. ing and winter (Turbott 1956, Dorward 1962, ACKNOWLEDGMENTS Stonehouse1962, Nelson 1964, Potts 197 1, Ber- ry 1976, DeKorte and DeVries 1978, Bernstein I thank P. S. Humphrey and D. Siegel-Causeyfor as- and Maxson 198 l), stepwisemolt appears to be sistanceand advice in various aspectsof this study. For accessto specimensI thank G. E. Barrowclouah. advantageous in adults as well in that it allows W. E. Lanyon, &d M. LeCroy (American Museum-oi slow molt throughout the year. Stepwise molt of Natural History, AMNH); K. C. Parkes(Carnegie Mu- remiges in cormorants and other pelecaniforms seum of Natural History); S. M. Lanyon (Field Mu- may also have the adaptive advantage of allow- seum of Natural History); R. Schreiber (Los Angeles County Museum); J. V. Remsen (Louisiana State &i- ing full flight capabilities year-round with min- versity Museum of Zoology): R. Pavnter (Museum of imal disruption. In stepwisemolt of adults small- Comparative Zoology); J. Navas (Mbseo Argentina de er gaps of molting feathers occur at more sites Ciencias Naturales); R. L. Zusi (National Museum of in the wings, rather than larger gapsat fewer sites Natural History); A. Rea (San Diego Natural History (Table 2) as occur in subadults. Museum); R. B. Payne (University of Michigan Mu- seum of Zoology); and E. Stickney (Yale Peabody Mu- Subadult Blue-eyed Shags have marginally seum,YPM). For helpful discussionI thank L. D. Mar- more molting feathers per wave than do adults tin, N. P. Bernstein,and M. A. Jenkinson.K. C. Parkes, MOLT OF SHAGS 221

R. M. Mengel, W. Hoffman, S. A. Rohwer, W. A. DORWARD,D. F. 1962. Comparative biology of the deGraw, and M. D. Gottfried improved versions of White Booby and the Brown Booby Sula spp. at the manuscript. I thank the followins for assistance:J. Ascension. Ibis 103b:174-220. Vinuesa, Consejo National de Inv&tigaciones Cien- GINN, H. B., AND D. S. MELVILLE. 1983. Moult in tificas v Tecnicas (CONICET): R. G. Wilson. Puerto birds. B.T.O. Guide 19. Br. Trust Omithol.. Beech Deseado, Santa Cruz Province; and R. Clarke, Dpto. Grove, Tring, Hertfordshire, England. Conservation de la Fauna, Santa Cruz Province, Ar- HARRISON,C.J.O., AND C. A. WALKER. 1977. Birds gentina. This study was supported by grants to the of the British Lower Eocene. Tertiary Res. Spec. author from the Frank M. Chapman Memorial Fund Pap. No. 3. Tertiary Res. Group, London. and the American Museum of Natural History Col- HUMPHREY,P. S., AND K. C. PARKES. 1959. An ap- lections Study Program; the Humphrey family; the proach to the study of molts and plumages. Auk Museum of Natural History (KUMNH), Department 76:1-31. of Systematicsand Ecology, and the Endowment As- HUNTER,S. 1984. Moult of the giant petrels Macro- sociation,the University of Kansas;and National Sci- necteshalli and M. giganteusat South Georgia. ence Foundation Grant BSR 84-07365 to Humphrey Ibis 126:119-132. and Siegel-Causey. MARTIN, L. D. 1983. The origin and early radiation of birds, p. 29 l-353. In A. H. Brush and G. A. LITERATURE CITED Clark, Jr. [eds.], Perspectives in ornithology. American Ornithologists’ Union, New York. ASHMOLE,N. P. 1962. The Black Noddy Anous te- MURPHY,R. C. 1936. Oceanic birds of South Amer- nuirostris on Ascension Island. Part 1. General ica. Vol. 2. American Museum Natural History, biology. Ibis 103b:253-273. New York. ASHMOLE,N. P. 1968. Breedingand molt in the White NELSON,J. B. 1964. Factors influencing clutch size Tern (Gygis alba) on Christmas Island, Pacific and chick growth in the North Atlantic . Ocean. Condor 70:35-55. Ibis 106:63-77. BERNSTEIN,N. P., AND S. J. MAXSON. 1981. Notes OLSON,S. L. 1977. A Lower Eocene from on moult and seasonablyvariable characters of the Green River Formation of Wyoming (Pele- the Antarctic Blue-eyed Shag.Notomis 28:35-39. caniformes: Fregatidae). Smithsonian Contribu- BERRY.H. H. 1976. Phvsioloaical and behavioural tion PaleobiologyNo. 35. SmithsonianInstitution ecologyof the Cape Cormor&t Phdacrocoruxca- Press,Washington, DC. pensis.Madoqua 9:5-55. POTTS,G. R. 1971. Moult in the ShagPhalacrocorax CANNELL,P. F., J. D. CHERRY,AND K. C. PARKES. aristotelis,and the ontogeny of the “StaJ&niiu- 1983. Variation and migration overlap in flight ser.” Ibis 113:298-305. feather molt of the Rose-breastedGrosbeak. Wil- RASMUSSEN,P. C. 1987. Molts of the Rock Shag son Bull. 951621-627. (Phalacrocoraxmagellanicus) and new interpre- COOPER,J. 1985. Biology of the Bank Cormorant, tations of the plumage sequence.Condor 89:760- Part 2. Morphometrics, plumage, bare parts and 766. moult. Ostrich 56:79-85. STONEHOUSE,B. 1962. The tropic birds (genusPhiie- CRACRAFT,J. 1985. Monophyly and phylogeneticre- thon) of Ascension Island. Ibis 103b:124-16 1. lationshipsof the Pelecaniformes:a numerical cla- STRESEMANN,E., AND V. STRESEMANN.1966. Die distic analysis.Auk 102:834-853. Mauser der Vogel. J. Omithol. 107:Sonderheft. CRAWFORD,R.J.H., P. A. SHELTON,R. K. BROOKE, TURBOTT,E. G. 1956. Notes on the plumages and AND J. CRAWFORD.1982. , distribu- breedingcycle of the ,Phalacrocorax tion, population size and conservation of the (Stictocarbo) punctatus punctatus (Sparrman, Phalacrocorax coronatus. 1786). Rec. Auckl. Inst. Mus. 4:343-363. Gerfaut 72:3-30. WATSON,G. E. 1915. Birds of the Antarctic and sub- DEKORTE,J., AND T. DEVRIES. 1978. Moult of pri- antarctic. American Geophysical Union, Wash- maries and rectrices in the Greater Frigatebird, ington, DC. Freguta minor, on Genovesa, Galapagos.Bijd. tot de Dierk. 48:81-88.