OY

HARVARD UNIVERSITY

LIBRARY

OF THE

Museum of Comparative Zoology

mus. COMP. ZOOL
LIBRARY

FEB 21 1964

HARVARD
: / / UNIVERSITY.
oslilla

YALE PEABODY MUSEUM

oF Natura History

Number 76 October 30, 1968 New Haven, Conn.

MOLT AND BREEDING IN POPULATIONS

OF THE SOOTY TERN STERNA FUSCATA

N. Poitier AsHMOLE

Epwarp Grey INsrirute, Oxrorp, ENGLAND*

I have recently discussed (Ashmole, 1963) the breeding cy-
cles of the Sooty Tern Sterna fuscata in all parts of its range.
Although there are many areas for which adequate informa-
tion is lacking, it appears that this species in different locali-
ties shows three different types of breeding cycle: namely, every
twelve months, every nine and one half months, and every six
months. In the places where breeding occurs every six months
it has not yet been shown whether the same individuals breed
in successive breeding periods, but I am at present carrying
out work on Christmas Island (Pacific Ocean) designed to
determine this.

During study of a large number of Sooty Tern skins in
United States and British museums I cbserved an unexpected
difference in the state of the primaries between samples of
birds from populations where breeding is annual and from
populations where breeding occurs every six months. This dif-
ference, described in the present paper, suggests that Sooty
Terns in populations where six-monthly breeding occurs have

* Present address: B. P. Bishop Museum, Honolulu 17, Hawaii.

2 Postilla Yale Peabody Museum No. 76

evolved special modifications of the species’ normal molt pro-
gram, which enable them to breed at approximately six-month
intervals and yet to replace their remiges cften enough to
maintain reasonable flying efficiency.

Sooty Terns in most populations appear to have a pattern
of molt similar to that in a number of other tern species: that
is, they undergo a postnuptial cr “basic”' molt after breeding
in which all feathers are replaced, and a prenuptial or “alter-
92

nate” molt shortly before the onset of breeding which does not

involve the primaries or secondaries. Although Dwight (1901)
says “The Terns undergo two complete msults in a year...,”
I know of no tern species for which there is adequate evidence
that all the remiges are replaced twice each year. In some
species none of the remiges are replaced more than once, while
in others the inner primaries are replaced twice, the outer ones
only once (Ashmole, in prep.). In the Sooty Tern I have found
no indication that any of the primaries are replaced more than
once between one breeding period and the next.

As in other terns, molt and breeding in the Sooty Tern are
more or less mutually exclusive. (However, Brown Noddies
Anous stolidus on Ascension Island and perhaps elsewhere
sometimes breed and molt at the same time (Dorward and Ash-
mole, 1963).) Few museum specimens are accompanied by infor-
mation as to whether the individuals were involved in breeding
activities when collected, but Table 1 shows that most Sooty
Terns collected on breeding grounds in all parts of the world
have complete sets of primaries and rectrices. From some locali-
ties there are a few birds just completing the replacement of
their primaries (primary molt scores 98 and 99), while from
some breeding stations there are birds which have recently
started a molt (primary molt scores nearly all below 30).°
I have already shown (Ashmole, 1968) that on Ascension Is-
land individuals complete a molt before starting to breed, but
some at least start their postnuptial molt before their chicks

'.°'These terms are those advocated by Humphrey and Parkes (1959).

*“Primary molt scores” are stages on the scale from 1 (molt of pri-
maries just started) to 99 (—molt of primaries almost completed); for
details of the method of scoring see Ashmole (1962).

LIBRArY.

FEB 21 1964

Oct. 30, 1963 Sooty Tern Sterna Fuscata HARVARD
UNIVERSITY.

Taste 1. Molt of primaries and rectrices of Sooty Terns collected
on breeding grounds in different areas.’

-——_ Primaries ——, —Rectrices—

Pro- Range in Pro-
Number portion scores of portion
of birds — of birds molting of birds
Locality examined molting birds? molting
GULF oF Mexico
AND W. InpIEs
Corpus Christi, Texas ..... 17 0 —- 0
Wit IIS: oo aeodcesaboodao 12 0 = 0
EavuatTorIAL ATLANTIC
Fernando Noronha ........ 33 Zi 2-28, 98 ad
PASCENSION els 2 2.4 sticlaxeie wiarernr 107 20 1-20 All
SoutnH ATLANTIC
Trinidade /Martin Vaz .... 11 0 -- 0
InpIAN OcEAN
Waeccadiverlss <2-..05 6+ sce 11 18 Qa 09
Nortu Pactric
ILENSEIN con oboocnasopdeDedr 39 03 98 05
Ie sabela, Mexico .......-- 11 18 213 .09
TESS OCOTTOM sere serene cee 15 .20 4, 24, 48 20
@lipperton) Peers. eer ao 24 O04 99 alg
EauatortaL Pactiric
(GIS Haas se rare sa oy ou scetersi.ayei usr 16 19 2, 98, 98 .06
Culpepper /Wenman,
Galapagos -oa..c02s 00s 6 10 30 All 98 10
Sourtn Pacrric
WordsHowevlbec--sn.2 eee. 13 0 — 23
INomfolkolepan cristo ccicres 10 10 4, 10
IKiermadeGess mee eters cnicicicie 18 0 = 0
INGEN Aeon ovo GumoDOGo ane 11 0 —
SUR Gio | Gomicanaeoe ee doas 40 .30 1-24. a5
Kauehi, Tuamotu Arch. ... 26 .O4 98 04
Marquesasilisiis. rei siete 19 21 All 98 05
MCN OMA ota aie es calor 13 0 _ 0

Notes. 1. Only breeding localities from which I have examined at least ten
birds are included in this table.

2. See footnote in text for explanation of “primary molt scores.”

+ Postilla Yale Peabody Museum No. 76

become independent. Both on Ascension and elsewhere body
molt of Sooty Terns occurs almost entirely while the birds are
absent from the breeding grounds; I have so little information
about it that I shall not consider it further in this paper.

PRIMARIES

Examination of molting birds on Ascension, and of skins of
molting individuals, shows that in the Sooty Tern molt of the
primaries normally starts with the first (innermost) feather
and progresses outwards to the tenth (outermost long pri-
mary). Sooty Terns when breeding should thus have the inner-
most primaries oldest and the outer ones progressively newer,
the whole series forming a smooth sequence. This was found to be
the case in 86 per cent of all skins of birds collected on the breed-
ing grounds and not in process of molt. However, in some birds
there are striking differences in the condition of adjacent pri-
maries; one finds a sudden break in the normal age-sequence
part way through the series. This I have called a ‘‘discon-
tinuity.”* It should be emphasized that the discontinuities were
not caused merely by molt in progress when the bird was col-
lected; most individuals were not molting at all, and in the few
which had recently started a molt the arrangement of old and
new feathers at the discontinuity could not be explained as a
result of the molt then in progress. It was evidently the re-
sult of an unusual molt sequence in the past, followed by a ces-
sation of molt prior to breeding.

* Like many other species (Dwight, 1901) Sooty Terns have a pale “frost-
ing” or “silvering” on the dark primaries and secondaries, which gradually
wears off, thus making it easy to detect large differences in the age of
adjacent feathers. I have recorded discontinuities only when the difference
in the condition of adjacent feathers was sufficiently striking for there to
be no doubt that they were of very different age. In badly set specimens
it is difficult to assess the relative ages of the small inner primaries,
especially as they tend to be protected from wear by the overlying sec-
ondaries. I may therefore have overlooked relatively new innermost pri-
maries in some birds, and the figures for the occurrence of discontinuities
between primaries 1 and 2 must be considered as minimum ones. Discon-
tinuities further out in the series are not likely to have been overlooked,
and there were few birds in which I was doubtful whether the feathers
had been molted in regular sequence.

Oct. 30, 1968 Sooty Tern Sterna Fuscata 5

Discontinuities are found at all points in the primary series,
but Table 2 (from which molting birds are excluded) shows
that they are not distributed at random. Nearly all popula-
tions (see Table 3 for details) contain a small proportion of
birds with first primaries much newer (occasionally much
older) than the second, but in most populations (grouped in
the bottom row of Table 2) discontinuities at other points in
the primaries are rare. The sample from Ascension is separated
since not only does it contain an especially large proportion of
birds with discontinuities between primaries 1 and 2, but it also
has a number with discontinuities between primaries 2 and 3;

Taste 2. Distribution of discontinuities in the primaries of Sooty Terns

collected on Ascension, on the Phoenix and Line Islands, on Bedout Island,
and in the other localities mentioned in Table 3.

No. and %
Number (above) and percentages — of wings
Number (below) of wings with discontinu- with no
of wings —ities at each point in the primaries— discon-
Locality examined % % % 4% % & RR % Wo tinuities
Ascension 164 NoteZ Ose On ele? Oe JOOP Oe 50 133
Island (82
birds) Gp NG By ONG OD) OOO 81
Phoenix 64. ING 2) (WD 2 Bs (De El 34.
and Line (=32
Islands birds) Ger 3 OL Be Se OW ABs all Ale 53
Bedout 10 Nag 2 PFO. OO MW O 4) 2 2
Island (=5
birds)
Other 946 No, Bo “C @C bh i ft O Bs 854.
areas (=473 j
birds) a 6 037, 0:6 0:5" Ol5, O04. 0 (013, 0:5 90

Notes. 1. Molting birds are not included.

bo

For each locality, the upper figures are the numbers of wings which
show discontinuities at each point in the series of primaries. Wings
are used rather than birds since the two wings on a single bird
sometimes have discontinuities in different places.

3. The lower figures show the number of discontinuities at each posi-
tion as percentages of the number of wings examined. Since there
are sometimes two discontinuities in one wing, and both are in-
cluded in the figures given, the percentages total more than 100
in some cases.

6 Postilla Yale Peabody Museum No. 76

these are rare in other populations. Samples from two areas
only contain an appreciable number of birds with discontinui-
ties further out in the series than the second primary. One of
these areas includes the Phoenix Islands and Line Islands in
the central equatorial Pacific, while the other is represented by
Bedout Island off the northwest coast of Australia. The distri-
bution of discontinuities in the wings of birds from these areas
(Table 2) seems certainly to indicate that many birds in these
populations have a very different molt program from the birds
breeding on Ascension and in the other localities from which
specimens were examined.

Table 3 shows that, from population to population, there is
no correlation between the frequency of occurrence of disconti-
nuities between primaries 1 and 2 and that of discontinuities at
other points in the series. Thus while 21 out of 38 birds from
the Phoenix and Line Islands have discontinuities among pri-

Taste 3. Occurrence of discontinuities among the primaries in adult Sooty Terns
from different breeding areas.

Occurrence of Occurrence of
discontinuities only discontinuities
between primaries among

———I and 2———_—__ -——primaries 2-10—
Number Number
Number — (and pro- Number (and pro-
of portion) of portion )
specimens — with dis- specimens _ with dis-
Geographical area available continuities available continuitie:
Gur or Mexico anp W. I[npigEs ...... 74 13 (.18) 80 4 (.05)
(inel. Corpus Christi (Texas),
Virgin Is.)
EQuaroriIaAL ATLANTIC
Fernando Noronha, Rocas Reef ... 34 2 (.06) 38 4 (.11)
ASCENSION RT: ed cactayttane oes stores eer 82 10 (.12) 106 6 (.06)
Sounma Aavanries seep eine: Ge 19 OmG—) 19 @ (=)
(Trinidade/Martin Vaz,
St. Helena)
TNDrAN GG OCEAN (14). sie chcteimerte sie creer iets 21 2 (.10) 24 1 (.04)

(incl. Gloriosa, Seychelles,
Rodriguez, Laccadive Is., Chagos)

NorTHWEST AUSTRALIA ...........--. 5 1 (.20) 5 3 (.60)
(Bedout I.)

Oct. 30, 1968

Sooty Tern Sterna Fuscata

Taste 3 (Continued)

-t

Geographical area

Norruwest Paciric
Paracel Is., Philippines, Ryu
Kyu Is., Bonin Is., Marianas )

NORTH OENTRAG LCAGCIFIG ....------

(Wake, Hawaiian chain incl.
Laysan, Johnston)

INORTEEAST) OACTEIC) 5 code 6 22 <0 2 = as

(Revilla Gigedo Is. incl. Socorro,
Clipperton, Lower California,
west coast of Mexico)

EavatortaL Pacrric
MarsuHatt Is.
Puoenix Is. (Howland, Baker,

Canton, Enderbury, Phoenix)
Line Is. (Palmyra, Christmas,
Jarvis, Malden, Starbuck)

Garapacos Is. (Culpepper/Wenman) 10 1

Souruwest PAactric

TOLL 1

5 ati: 4

Occurrence of
discontinuities only
between primaries

= 1 and 2
Number
Number — (and pro-
of portion )

specimens — with dis-
available continuities

me wile Bi (23)

~
oe
ho

(.03)

(.02)

(Lord Howe, Norfolk, Kermadec Is.,

Fiji, Tonga, Samoa)

SouTHEAST PACIFIC
Cook Is., Society Is., Tubuai Is.,
Tuamotu Is. incl. Kauehi,

Marquesas Is., Oeno, Henderson,

Ducie, Easter, San Felix

Swoorov, Tongareva ...........-

.- 34 0

=~
bo
ws)

(.04)

Ca)

Occurrence of

discontinuities
among
——primaries 2-10
Number
Number (and pro-
of portion )
specimens — with dis-
available continuities
13 C08)
82 De (>
63 2 (.03)
5 1 (.20)
11 9 (.82)
27 12 (44)
10 1 G-10))
55 1 (.02)
73 17.01)
42 3m (.07)

Motes. ar

them are included in this table.

ho

Only adult birds collected on the breeding grounds or within a few miles of

. Birds which were in process of primary molt are excluded, with two exceptions:

(a) birds whose tenth primaries only were growing have been included, and
(b) birds whose first and/or second primaries only were growing, have been
used in the right-hand section but not in the left-hand section of the table: this
accounts for the differences between the columns showing “Number of speci-
mens available” in the two sections.

3. Where ten or more birds were available for examination of primaries 2-10,
from one island or from a group of islands within a circle of radius 25 miles,
the name of the island or group is italicized.

8 Postilla Yale Peabody Museum No. 76

maries 2 to 10, only 1 out of 32—less than the average propor-
tion—have discontinuities between primaries 1 and 2. Only on
Ascension, as already mentioned, are there an appreciable num-
ber of birds in which primary 2, together with 1, is strikingly
different in age from the rest. Probably in this case birds with
the first and second primaries very different in age from the
next outwards should be classed with those in which only the
first primary 1s affected.

I suspect that discontinuities far out in the series are nor-
mally produced under quite different circumstances from those
between primaries 1 and 2. I have already suggested with re-
spect to the Ascension population (Ashmole, 1963) that birds
with first (or first and second) primaries newer than the next
outwards may be young birds breeding for the first time; in im-
mature Sooty Terns successive sequences of primary replace-
ment often overlap, so that as one sequence is nearing comple-
tion with the growth of the outermost primaries, another se-
quence is starting with the innermost ones. If molt stops for
breeding at the completion of one sequence, the outermost
feathers will be new, but so may be the innermost ones, with
a discontinuity outside them.

It is not possible to explain in this way the extremely high
incidence of discontinuities (Table 3) among the outer pri-
maries in the samples from the Phoenix and Line Islands, while
the small sample from Bedout Island may also require a dif-
ferent explanation. It can be no coincidence that it is on cer-
tain of the Phoenix Islands and Line Islands, alone of the
places from which I have examined an appreciable number of
specimens, that Sooty Terns are known to have two breeding
periods each year (Ashmole, 1963).

It has been argued (Ashmole, 1963) that the Sooty Terns on
Ascension are breeding as often as they can—that breeding,
followed by a complete molt, occupies about nine and a half
months. But if this is the minimum time needed by the Ascen-
sion birds, it is difficult to see how the birds in the Phoenix and
Line Islands (where there are two breeding periods each year)
could breed and undergo a complete molt in a period of only
six months. It is therefore not surprising that it has been
tentatively suggested in the past (Richardson and Fisher,

Oct. 30, 1963 Sooty Tern Sterna Fuscata 9

1950; Hutchinson, 1950; Chapin, 1954) that in the areas
where breeding occurs every six months, different populations
of birds might be involved in successive breeding periods, so
that each individual would breed only once a year. I also felt
that this must be the explanation, until I examined specimens of
the Sooty Terns from the islands concerned, and found that
many of the breeding birds had some old and some new pri-
maries, and had evidently not undergone a complete molt be-
tween breeding periods.

As it has been pointed out, Sooty Terns in other parts of
their range, where breeding occurs only once each year, have a
straightforward complete replacement of all their remiges after
breeding. If the individuals on the Phoenix and Line Islands
were also breeding only once a year, how could one explain the
fact that many of them do not have a complete set of new re-
miges when they start breeding? The most reasonable hypoth-
esis seems to be that at least some individuals breed in succes-
sive breeding periods and replace only some of their wing
feathers in the short interval in between. It is likely, however,
that no one individual breeds in every breeding period.

The sample from Bedout Island (N.W. Australia) is very
small, but three cut of the five birds show discontinuities among
primaries 2-10. Sooty Terns have so far only been recorded as
breeding in autumn on Bedout. However, it would not be sur-
prising if they were found to have two breeding periods there
each year, since Serventy (1952) has shown that they breed in
autumn on islands to the north of Bedout, but in spring on
islands further south (see Appendix in Ashmole, 1963, for
details). Bedout is at about the latitude at which several other
species of sea birds change from autumn to spring breeding,
and some of them are already known to breed in both seasons,
on Bedout and certain other islands (Serventy, 1952).

Of the other localities from which there are reasonably large
samples, Fernando Noronha (where the interval between suc-
cessive breeding periods has not yet been determined) has most
birds with discontinuities among primaries 2-10, but the pro-
portion is far lower than in the samples from the Line Islands
and the Phoenix Group. Of the latter, the Line Islands have the
lower proportion, but even this is significantly higher than

10 Postilla Yale Peabody Museum No. 76

among the Fernando birds (P <.01). It is evident, therefore,
that the populations from the Line and Phoenix Islands show
peculiarities which are almost entirely absent in the other popu-
tions sampled. This is expected on my hypothesis that peculiari-
ties in the molt normally occur only in individuals which breed
in successive breeding periods about six months apart.

Table 3 shows that although the proportion of birds with
discontinuities among primaries 2-10 is far higher in both the
Phoenix and Line Islands than in any other area apart from
Bedout Island, it is appreciably higher in the small sample
from the Phoenix Group than in that from the Line Islands.
This difference as it stands is statistically significant (at the
5 per cent level), but both samples are heterogeneous (birds col-
lected from several different islands, in different years and at
different stages in the breeding periods), so I doubt whether :t
is meaningful.

Examination of the precise arrangement of new and old
feathers in the primaries of birds from the Phoenix and Line
Islands and from Bedout Island, may help towards an under-
standing of the molt program in these populations and of the
way in which breeding and molt are interrelated; some actual
examples are therefore given below. In the examples, N repre-
sents an apparently brand new feather, (N) a newish one, (O)
an oldish feather, and O an old, worn feather. Where a series
of adjacent feathers have been replaced in a regular sequence,
they may grade from one category to another; in such cases
the two terminal members of the series are joined by an arrow
headed towards the newest feather. The figures 1, 2, 3 and 4
represent successive stages in the growth of a feather (see Ash-
mole, 1962 :255

1. Suvoroy LN NN OO) (OO), OO Ome
Re IN No IN, 70) O20) OO One®

2. Fernando Noronha I 1 (N) (N) () (CN) WW) (WW) (WW) ()_O
Re 1 (N) GN) GS). GQ) GA) Or SoRe

3. Jarvis I. L: 1 °(N)-G)2Q) ()) C20) 700 ae
(Line Is.) RAG WY) Gy) OSFOFtO O70

Oct. 30, 1963 Sooty Tern Sterna Fuscata 11

1 2 3 1 5 6 7 8 $ 10

4. Enderbury I. | OD) —>N
(Phoenix Is.) 1h (OS)) $= ——— Ne ©
peeHernandosNoronha, i. OF ———_____.____» ()) (GN) (ND) GN) GN)
1S) (10) (Ost) (ON) (ESO) GND)

6. Baker I. i, (0): (0) (O) NN -N’ (CO). () (0),C) 7)
(Phoenix Is.) R21 (O)1@O)TO) NNO) OO) Or)

7. Jarvis I. ee Ae ONS) ENDO) CO) O)TQONGN) GN)
(Line Is.) R. 3 (N) (N) (N) (O) (O) (O) (N) (N) (N)

8. Palmyra I. fe NNT NS NEN (CO) (Oy NNN
(Line Is.) EUS aNG aN Ney ONG Ne Ni SNe CNY SIN Ny

9. Christmas I. a (O) (NOR On JO Os GN) GND)
(Line Is.) R20) iN) OC, Ol Oe Oy) 0
10. Enderbury I. EN) (0) (0) ONO)

(Phoenix Is.) Fa ONG) NE (O)) 0 (O) 2 (O) Na)

In most birds with discontinuities among primaries 2-10, both
from areas where breeding occurs annually and from those
where it occurs every six months, the feathers inside the dis-
continuity are newer than those outside it (see examples 1-+).
This condition is the one which would arise if a normal sequence
of primary replacement stopped part way through the series.
Patterns of the type shown in example 5, in which there is a
definite discontinuity between old feathers on the inside and
newer ones on the outside, would result if, after primary re-
placement had stopped part way through the series for breed-
ing, it later started again where it had left off, and then con-
tinued outwards.’ However, the condition shown in example 5
is uncommon in all Sooty Tern populations. This suggests that
normally, when a bird with the inner feathers newest (as in
example 1) starts to molt again, the innermost primaries,
rather than those immediately outside the discontinuity, are
replaced first; this is in fact what is occurring in examples 2

and 3.

° Something equivalent to this certainly occurred in the replacement of the
secondaries of some Black Noddies Anous tenuirostris on Ascension Is-
land (Ashmole, 1962).

11 Postilla Yale Peabody Museum No. 76

The other examples shown are of the more complex situation
in which there is more than one discontinuity among the pri-
maries in one or both wings. Patterns of this type, which are
found in a significant proportion of the birds from areas where
breeding occurs every six months, have not yet been found
among birds from annual-breeding populations. They could not
be produced during a molt program in which each primary was
always replaced in regular sequence outwards through the
series, but they could arise if an incomplete primary replace-
ment was succeeded, after the breeding period, by another
incomplete molt, and if feathers replaced late in the first of
these molts tended to be skipped during the next. In this case
the bird in example 1, if it underwent another incomplete molt,
might in the next breeding period be in a condition similar to
example 6, for instance:

1 2 3 a 5 6 7 8 9 10

kL '(O)'(O) (O) N "NIN “N70, OF20
R2(O)NCO) CO) IN ENE INS NOs HOw tO

Another partial molt could lead to conditions comparable to
those in examples 7-10, for instance:

7,
r
4

ls Ne seN. 4 Ny (0), (©) 000) 2 (OO) AN
RN, N Nec). ©), (©). @), N

7,
:
Z

or
us NN ON 960) (O)n(O)m(O)P NE ENE BO
Ro oN ON Ni C(@)n(O)(O). (On NN sO

It will be clear from what has been said that it is not neces-
sary to postulate a random molt sequence to account for the
complex patterns of old and new primaries found in the wings
of some Sooty Terns. The indications are that in these birds
as in all other terns molt in the primaries proceeds from the
inside outwards, but that many individuals in populations
where breeding occurs every six months fail to replace all their
primaries between successive breeding periods, and that subse-
quent molts are modified by the presence in their wings of a
mixture of old and new feathers.

Oct. 30, 1963 Sooty Tern Sterna Fuscata 13

SECONDARIES

As in other tern species, replacement of the secondaries in
Sooty Terns starts much later than that of the primaries, but
is completed at the same time or only a little later. Replace-
ment normally starts at the two ends of the series of sec-
ondaries, and it is some of the middle feathers (often numbers
12 and 13 counting from the carpal joint inwards) which are
the last to be replaced. After the complete replacement of the
secondaries in this manner there should be no appreciable dis-
continuities within the series, but the feathers at the two ends
of the series will be oldest, and the middle ones newest. This is
in fact the situation found in nearly all specimens from most
Sooty Tern breeding colonies, including Ascension Island (Ta-
ble 4). This must imply that all the secondaries are replaced
once between breeding periods. However, in several populations

Tasie 4. Numbers of Sooty Terns with discontinuities among the secondaries, in relation
to the incidence of discontinuities among primaries 2-10, in different populations.

Birds without Birds with

discontinuities discontinuities
among primaries 2-10— —among primaries 2-10
Number with Number with
discontinuities discontinuities
Geographical Number among Number among
area examined secondaries examined — secondaries
Gur or Mexico anp W. INpIeEs 72 0 (+2?) 2 1
PASGENISTONG Ls - fe/a= ayaa teveisiousisye 7A 1 6 1
IGspihos; (Oxawosp Saacocassoaganc 22 Oo” (Sele) if 1
TOE NIX ph S5) oy. rove ores sree eiete fe 6.54 2 2 8 6(EE 2?)
GENIE LSSars speesis ccs. Setere ccavetels oncrets 13 Zn (teal) 11 11
SOU Hap yA CHET Cress syaiaye paces este = 116 A (-F 1?) 1 0
(incl. Marquesas,
but not Suvorov)
SUMOR OV meyattis nis srre bite) Pate aha 26 SY) (SL) 3 2 (+1?)

Notes. 1. Birds undergoing molt of primaries or secondaries are excluded, except for
those in which only the first primaries have so far been shed.

i)

. Additional figures in brackets, followed by question-marks, represent birds
whose secondaries show probable, but not striking, discontinuities.

14 Postilla Yale Peabody Museum No. 76

birds are found which have some secondaries much older than
the rest. I did not examine the secondaries of all specimens, and
there were a good many doubtful cases, so I cannot give reli-
able figures for the frequency of this peculiarity in all popula-
tions. However, Table 4 gives the information for those popula-
ticns from which the secondaries of a fair proportion of the
available specimens were examined.

As might be expected, the populations in which many birds
show discontinuities in the primaries (Phoenix and Line Is-
lands) also contain many birds with discontinuities among the
secondaries. However, in these populations some of the birds
without discontinuities in the primaries, nevertheless have sec-
ondaries which do not seem to have been replaced in a smooth
sequence: evidently the molt cycle is not entirely normal even
in these birds.

Among localities where discontinuities in the primaries are
rare, the island of Suvoroy, south of the equator in the central
Pacific, is the only one from which I have a fairly large sample,
in which many birds have discontinuities among the secondaries
(Table 4). In these birds some of the middle secondaries tend
to be much older than the rest, suggesting that the secondary
molt has stopped before completion. This situation invites
comparison with the Black Noddies on Ascension Island (Ash-
mole, 1962), where the primary molt was apparently never cut
short at the start of breeding, but some of the old middle sec-
ondaries, which are normally molted slightly later than the
last primaries, were sometimes retained through the breeding
period and replaced immediately after it. The occurrence of a
similar phenomenon among the Sooty Terns on Suvoroy sug-
gests that the breeding cycle there may be abnormal in some
respect, but there is very little information on the times of
breeding (Ashmole, 1963).

RECTRICES

Replacement of the tail feathers of Sooty Terns normally
starts with the outermost feathers (number 6 on each side),
the central pair (number 1) being molted next; molt probably
then continues in the sequence 2, 3, 5 and 4. It is possible that

Oct. 30, 1965 Sooty Tern Sterna Fuscata 1a)

in the annual-breeding populations all the rectrices are then
replaced again before the next breeding season, but in the
Ascension population there was evidence that the outer pair
alone are replaced twice (Ashmole, 1963). In the Ascension
birds the outer webs of the outermost feathers are normally
white in the breeding period, but are more often, if not always,
dark in the non-breeding period. In other populations there is
much variation in the color of these feathers, and samples from
different populations sometimes also differ markedly. In some
areas nearly all the birds taken on the breeding grounds have
entirely white outer webs to the outer rectrices (69 out of 82
birds from the North Central Pacific), but in other places (e.g.
Fernando Noronha, the Southeast Pacific, the Phoenix and
Line Islands, and Suvorov) the proportion is much lower. It is
likely that in the populations in which breeding occurs every
six months not even the outermost rectrices are always replaced
twice between successive breeding periods. I cannot suggest any
explanation of the different frequencies with which dark color
is present in the outer webs of these feathers in other popula-
tions.

DISCUSSION

The data presented in this paper, together with the informa-
tion on the times of breeding of various Sooty Tern popula-
tions given by Ashmole (1963), show that the schedule of
breeding and molt evolved among the Sooty Terns of the Phoe-
nix and Line Islands is remarkable both in that breeding occurs
every six months, and in that the program of molt is flexible to
a unique degree. It appears that an individual sometimes under-
goes a complete molt without interruption but at other times
replaces only some of its primaries and secondaries between one
breeding cycle and the next.

I have already mentioned that in the populations where
breeding occurs every six months, individuals which have under-
gone only a partial molt before breeding generally have the
outer primaries older than the inner ones. Since the outer pri-
maries are also more subject to wear, it is not surprising to
find some individuals with outer primaries in extremely poor

16 Postilla Yale Peabody Museum No. 76

condition while breeding. These birds are doubtless below their
maximum flying efficiency, but if the curtailed molt has enabled
them to breed in a breeding period which they would otherwise
miss, the disadvantage may on balance be outweighed. However,
it is clear that the molt program in the Phoenix and Line Island
populations, in which the inner primaries are on an average
replaced more often than the outer ones, although the latter
get more wear, is not the most efficient that might be evolved.
More birds would be close to maximum flying efficiency for
more of the time if molt always started from where it had left
off, so that the primaries were always replaced in order of age.
This evidently happens sometimes but cannot be common.

Although I have argued that the presence of discontinuities
among the primaries of breeding individuals from the Phoenix
and Line Islands implies that these birds were involved also in
the previous breeding period only six months before, I am not
suggesting that individuals breed every six months. It seems
unlikely that a pair could raise a chick successfully in one
breeding period and yet be ready to breed again in the next
breeding period only six months later. With the time required
for courtship plus incubation for a month and feeding the
young for two to three months (longer if the juveniles are not
independent as soon as they leave the colony), very little time
would be left before the next breeding period. I would guess
that birds which raise a chick in one breeding period may then
undergo a complete molt, missing the next breeding period;
this would account for the proportion of birds from the Phoe-
nix and Line Islands which appear to have undergone a per-
fectly normal and complete molt. Many of the birds, however,
losing their eggs or their chicks while fairly young, could be
ready to try again in the next breeding period, after only a
partial molt. .

This reasoning is of course highly speculative, and further
discussion cf the factors controlling the schedule of breeding
and molt in the Sooty Terns of the Phoenix and Line Islands
will be profitable only when we have more information on the
sequence of events in individual birds from one of these popula-
tions.

Oct. 30, 1963 Sooty Tern Sterna Fuscata ee
ACKNOWLEDGMENTS

This account of molt in populations of Sterna fuscata forms
a complement to my study of the species on Ascension Island
during the course of the British Ornithologists’ Union Cente-
nary Expedition (Ashmole, 1963). The present paper is a
product of my examination of a large number of skins of
Sterna fuscata during the summer of 1960, while I was the
holder of a Seessel fellowship at Yale University. I am most
grateful to Professor G. E. Hutchinson who suggested a study
of specimens of the species and who with Dr. S. Dillon Ripley
enabled me to visit Yale. I must also thank Dr. Philip S.
Humphrey, who gave me much of his time during my work in
the Yale Peabody Museum, and whose wide knowledge of molt
was a constant stimulant. My visit to Yale was made while I
was a member of the Edward Grey Institute of Field Orni-
thology, Oxford; it is a pleasure to thank the Director, Dr.
David Lack, for encouraging me to go to Yale, for helpful
discussion of my work and for reading a draft of this paper.
My wife has also made many helpful criticisms of the manu-
script.

Tern specimens were studied in the collections of the follow-
ing institutions: Peabody Museum of Natural History, New
Haven (which contains the bulk of the collections made by the
“Blossom” South Atlantic expedition, sponsored by the Cleve-
land Museum of Natural History) ; American Museum of Nat-
ural History, New York (which has fine series of Sooty Tern
skins collected by the Whitney South Sea Expedition) ; United
States National Museum, Washington, D.C.; Museum of Com-
parative Zoology, Cambridge, Mass.; Museum of Zoology, Ann
Arbor, Michigan; Department of Conservation, Cornell Univer-
sity; British Museum (Natural History), London. I am grate-
ful to the curators of all these collections for the facilities they
offered me. In addition, skins were generously loaned by the
authorities of the California Academy of Sciences, the Los
Angeles County Museum, the Carnegie Institution, Pittsburgh,
the Academy of Natural Sciences of Philadelphia, the Museum
of Zoology, Ann Arbor, Michigan, the Kansas University Mu-

18 Postilla Yale Peabody Museum No. 76

seum of Natural History, the Chicago Natural History Mu-
seum, the United States National Museum and the Bernice P.
Bishop Museum, Hawau.

SUMMARY

The Sooty Tern Sterna fuscata in most parts of its range
breeds at the same season in each year, and study of museum
specimens shows that the individuals replace all their remiges
and rectrices between breeding seasons. On Ascension Island,
where breeding occurs every nine and one half months, there
is also a complete molt between successive breeding periods.
However, among birds from the Phoenix Islands and Line
Islands in the central equatorial Pacific, where breeding occurs
every six months, many individuals have ‘‘discontinuities”
among the primaries and secondaries, indicating that they
have not undergone a complete molt between successive breeding
periods. These populations have apparently evolved a uniquely
flexible molt program, such that under certain circumstances
(perhaps the successful rearing of a chick) breeding is followed
by a complete molt, but often molt stops and the bird breeds
again before all of the primaries and secondaries have been
replaced. It is suggested that because of this flexibility in the
molt, individuals are sometimes able to take part in successive
breeding pericds only six months apart.

LITERATURE CITED

Ashmole, N. P., 1962. The Black Noddy Anous tenuirostris on Ascension
Island. Part 1. General biology. Ibis, 103b: 235-273.

Ashmole, N. P., 1963. The biology of the Wideawake or Sooty Tern Sterna
fuscata on Ascension Island. Ibis, 103b: 297-364.

Chapin, J. P., 1954. The calendar of Wideawake Fair. Auk, 71: 1-15.

Dorward, D. F., and N. P. Ashmole, 1963. Notes on the biology of the
Brown Noddy Anous stolidus on Ascension Island. Ibis, 103b: 447-457.

Dwight, J. 1901. The sequence of moults and plumages of the Laridae
(gulls and terns). Auk, 18: 49-63.

Humphrey, P. S., and K. C. Parkes, 1959. An approach to the study of
molts and plumages. Auk, 76: 1-31.

Hutchinson, G. E., 1950. Marginalia: Wideawake Fair. Amer. Scientist,
38: 613-616.

Richardson, F., and H. I. Fisher, 1950. Birds of Moku Manu and Manana
Islands off Oahu, Hawaii. Auk, 67: 285-306.

Serventy, D. L.., 1952. The bird islands of the Sahul Shelf. Emu, 52: 33-59.

ne

3 2044 066 305 236

Date Due

A Hs

Se :
ee TT en
a ae oo
ot ee aa
ate
Pn ap te
Po re
os petra

ae

’ ae PPP PAP P LO fate Ay Pe
EPP EPCOT OES ae Par? eee
re ip t lef eS Ret PA EINE PRL he
Tn —. Ried tS a