Content uploaded by Renata Medeiros
Author content
All content in this area was uploaded by Renata Medeiros on Jun 28, 2018
Content may be subject to copyright.
Ibis
(2008),
150
, 717–727
© 2008 The Authors
Journal compilation © 2008 British Ornithologists’ Union
Blackwell Publishing Ltd
Monteiro’s Storm-petrel
Oceanodroma monteiroi
:
a new species from the Azores
MARK BOLTON
1,2
*, ANDREA L. SMITH
3
, ELENA GÓMEZ-DÍAZ
4
, VICKI L. FRIESEN
3
, RENATA
MEDEIROS
1
, JOËL BRIED
1
, JOSE L. ROSCALES
4
& ROBERT W. FURNESS
5
1
Department of Oceanography and Fisheries, University of the Azores, 9901-862 Horta, Portugal
2
Royal Society for the Protection of Birds, UK Headquarters, The Lodge, Sandy, Beds, SG19 2DL, UK
3
Department of Biology, Queen’s University, Kingston, Ontario K7L 3N6, Canada
4
Departament de Biologia Animal, Universitat de Barcelona, Avenida Diagonal 645, 08028 Barcelona, Spain
5
Institute of Biomedical and Life Sciences, Graham Kerr Building, University of Glasgow, Glasgow G12 8QQ, UK
The existence of two seasonally distinct breeding populations of
Oceanodroma
storm-petrels
in the Azores islands was first documented in 1996. The discovery of morphological differences
between the populations led to the suggestion that they may represent cryptic sibling species.
Recent mtDNA and microsatellite analysis from storm-petrel populations has considerably
advanced our understanding of their taxonomic relationships. Here we present new
information on the timing of breeding and moult of the two Azores populations, the extent
of exchange of individuals between seasons, and diet from feather isotopes. We conclude
that the hot-season Azores population should be considered a new species for which we
propose the name
Oceanodroma monteiroi
, Monteiro’s Storm-petrel. The species is both
genetically distinct and genetically isolated from the sympatric cool-season population of
Madeiran Storm-petrel
Oceanodroma castro
, and from all other populations of
Oceanodroma
castro
in the Atlantic and Pacific Oceans examined to date. Differences in the vocalizations
permit species recognition, and the extent of primary feather wear and stage of moult aids
separation of the two species in the Azores, which is especially valuable during August when
both attend the breeding colonies in large numbers. Feather carbon and nitrogen isotopes
reveal that the diet of Monteiro’s Storm-petrel differs from that of the sympatric Madeiran
Storm-petrel during both breeding and non-breeding seasons, and unlike the Madeiran
Storm-petrel, Monteiro’s Storm-petrel appears to maintain the same foraging environment
during the summer and winter months, though it shows a dietary shift to higher trophic levels
during the non-breeding season. Monteiro’s Storm-petrel is thought to be confined to the
Azores archipelago, where it is currently known to nest on just two small neighbouring islets.
The total population size was estimated at 250–300 pairs in 1999.
Keywords:
genetics, isotopes, moult, taxonomy, vocalizations.
Records from the Azores of a small, white-rumped
nocturnal burrowing seabird date back to the 16th
century (Fructuoso 1561), when thousands were
attracted nightly to fires and killed for food and
oil. The description of the species closely matches
the Madeiran Storm-petrel
Oceanodroma castro
(Monteiro
et al
. 1996) and it was still considered to
be extremely abundant in the 17th century (Chagas
1645–1650). Following the formal description of the
Madeiran Storm-petrel from Madeira in the 19th
century (Harcourt 1851), the first record from the
Azores islands was in 1887 (Bannerman & Bannerman
1966). However, breeding was not confirmed in the
Azores until almost 100 years later (Le Grand
et al
.
1984). Subsequently, Luís Monteiro, carrying out
research for his doctoral thesis on eco-toxicology of
the seabird assemblage in the Azores in the 1990s,
*Corresponding author.
Email: Mark.Bolton@rspb.org.uk
718
M. Bolton
et al.
© 2008 The Authors
Journal compilation © 2008 British Ornithologists’ Union
discovered two seasonally segregated breeding
populations of storm-petrels (Monteiro
et al
. 1996,
Monteiro & Furness 1998), similar to those previously
described from the Galápagos islands (Harris 1969).
Data on the occurrence and development of brood
patches among mist-netted birds, incubation periods
and chick body size suggested breeding populations
were out of phase by 4–5 months, and overlapped at
the breeding colony only in August and early September.
Substantial differences in morphometrics and the
quantities of mercury in both regurgitated food items
and in the plumage of these seasonal populations
suggested differences in their feeding ecology and
the possibility of ecological isolation of the two groups
(Monteiro & Furness 1998, Monteiro
et al
. 1998).
More recently, Nunes (2003) and Bolton (2007)
have found differences in the structure of burrow
calls of the seasonal populations. Playback experiments
have shown that hot-season birds prospecting for
mates and nest sites do not respond to calls of
cool-season individuals (Bolton 2007), although the
response of cool-season birds to hot-season playback
has not been examined in the Azores. Storm-petrels
rely heavily on acoustic signals to locate potential
mates when prospecting the colony at night (Allan
1962, Harris 1969), so the absence of vocal recognition
between seasonally separated storm-petrel popula-
tions would act as a pre-mating isolating mechanism,
preventing gene flow. The differences in calls between
seasonal populations are clear to the human ear.
Analysis of mitochondrial DNA (comprising 266 bp
of Domain I and 180 bp of Domain II of the control
region) showed that the level of gene flow (number
of females exchanged per generation) between sym-
patric seasonal populations of storm-petrels in the
Azores did not differ significantly from zero (all val-
ues less than one female per generation between
colony pairs) and there were significant levels of net
sequence divergence (1.61–2.28%, Smith
et al
. 2007).
Assuming a sequence divergence rate of 21% per
million years (Wenink
et al
. 1993, though see Nunn
and Stanley (1998) who suggest slower rates of
mtDNA divergence for storm-petrels), seasonal
populations in the Azores were estimated to have
diverged 70 000–154 000 years ago. In contrast,
much higher levels of gene flow (0.96 to 8.40
females per generation) and low levels of net
sequence divergence (0.005–0.10%) were found
between cool-season populations in the Azores and the
Desertas islands, from where
O. castro
was first
described (Smith
et al
. 2007). These results are also
generally supported by microsatellite analyses (Friesen
et al.
2007) and indicate that whilst the cool-season
storm-petrels breeding in Azores are conspecific
with the nominate population from the Desertas
islands, storm-petrels nesting in the Azores during
the hot season are genetically distinct and genetically
isolated from sympatric cool-season birds and those
breeding in both the hot and cool seasons in the
Desertas islands.
Here, we present previously unpublished data on
the timing of breeding of the seasonal populations of
the Azores and on the timing of moult of primary
feathers. We also present updated information on
the rate of exchange of individuals among seasonal
populations and examine stable isotopes of nitrogen
and carbon in storm-petrel flight feathers to compare
foraging environments and trophic levels of the two
populations. We conclude that the seasonal popula-
tions should be considered two distinct species and
propose a new specific name for the summer-breeding
population. We evaluate its current conservation status
in the light of its taxonomic affinity to populations of
Madeiran Storm-petrels in the both the Atlantic and
Pacific Oceans.
METHODS
Fieldwork was carried out on Praia Islet, Graciosa,
Azores (39
°
03
′
N, 27
°
57
′
W), between 20 February
2000 and 10 September 2001 to establish the timing
of breeding of the seasonal populations. Nest sites
were located by searching the islet at night for storm-
petrels singing within nest crevices, by inspection
of potential nest holes and by playback of sound
recordings during daylight. In addition, artificial nest
boxes were installed during 2000 and natural nest
sites and boxes were inspected daily to determine
occupancy, and the timing of laying, hatching and
fledging (see Bolton
et al
. 2004 for full details).
Adults occupying nest sites were ringed and so could
be individually recognized if trapped subsequently
away from the nest. Mist-netting was carried out in
the colony on 71 nights covering all months of the year.
Individuals were examined to determine whether
the cloaca was enlarged (indicating recent egg-laying),
the stage of development of the brood patch (Monteiro
& Furness 1998) and the amount of wear and stage of
moult of the primary feathers. The extent of moult
of the 10 primaries was recorded according to Ginn
and Melville (1983), where a value of zero represents
a bird with worn primaries that has yet to commence
moult and a value of 50 represents a bird that has
completed moult of all primaries. Intermediate values
© 2008 The Authors
Journal compilation © 2008 British Ornithologists’ Union
New species of storm-petrel
719
refer to birds in active primary moult. The first author
also measured body mass, wing, tarsus, combined
head and bill, bill length (to feathering), bill depth
(at the gonys), tail length and tail fork, following
methods detailed in Monteiro and Furness (1998).
Data were examined for homoscedasticity and nor-
mality. Where significant departures from normality
occurred, data were analysed using the nonparametric
Wilcoxon two-sample test, approximated to a normal
distribution, due to large sample size. Procedures were
implemented in SAS® v9.1 (SAS Institute 2007).
Further fieldwork was carried out on several days
and nights during each breeding season from June
2002 to December 2007 to ring nestlings and adults
and to recapture individuals ringed in previous seasons
to assess patterns of colony attendance and exchange
of individuals between seasons. Mist netting and
ringing were also carried out on Vila Islet (36
°
55
′
N,
25
°
10
′
W), also situated in the Azores some 300 km
from Praia Islet, to assess the extent of movements
between colonies within the archipelago; Vila only
holds cool-season Madeiran Storm-petrels (Monteiro
et al
. 1999). When later recaptures of birds examined
previously allowed a more reliable assessment of
population membership (for example if individuals
were later retrapped with a developed brood patch
or in a nest with an egg or chick), the latter assessment
of breeding season was used in all analyses.
Information on the foraging environment and trophic
level of seasonal populations of storm-petrels was
obtained from analysis of stable isotopes of carbon
and nitrogen in flight feathers. Inshore and benthic
habitats are characterized by relatively higher ratios
of
13
C to
12
C compared with offshore and pelagic
environments (France 1995, Kelly 2000) and
15
N
enriches relative to
14
N at higher trophic levels due
to preferential incorporation into body tissues
(Minagawa & Wada 1984, Owens 1987). We exam-
ined the
13
C/
12
C (
δ
13
C ) and
15
N/
14
N (
δ
15
N) ratios
in the innermost primary feather, which will reflect
the diet when it is grown at the close of the breeding
season (Cherel
et al
. 2000, Bearhop
et al
. 2002), and
the eighth secondary feather, which is generally
grown at the end of the moult cycle (Scott 1970,
Arroyo
et al
. 2004) and will reflect the diet in the
latter part of the non-breeding season. Feather samples
were obtained, under licence, from the hot- and cool-
season populations on Praia Islet (
n
= 92 and 34
respectively) and the cool-season population on Vila
Islet (
n
= 70) between 2002 and 2004. Feathers were
cleaned in a solution of NaOH (0.25
M
), oven dried
at 40 ºC, and homogenized. Isotopic analyses were
carried out by EA-IRMS (elemental analysis-isotope
ratio mass spectrometry) by means of a Ther-
moFinnigan Flash 1112 elemental analyser coupled
to a Delta isotope ratio mass spectrometer via a
CONFLOIII interface (Serveis Científico-Tècnics,
University of Barcelona). Data were tested for nor-
mality and examined using multivariate analysis of
variance (
MANOVA
) for repeated measures. Wilks’
lambda was calculated to compare values for each
isotope (
δ
13
C and
δ
15
N) between primary and
secondary feathers and between sampled colonies
(two cool colonies and one hot). Significant effects
were further investigated using
post hoc
pairwise
comparisons with Bonferroni corrections to control
for Type I errors.
RESULTS
Timing of breeding
Daily inspection of both natural nest sites and nest
boxes confirmed that egg-laying occurred in two
discrete periods: May to early July (‘hot season’) and
October to early-December (‘cool season’; Fig. 1).
Many nest cavities were used for breeding in both
seasons. Incubation extended until early August for
the summer population and to early January for the
winter birds. The first hot-season chicks hatched in
early June and the last fledged by early October,
whereas the chick rearing period was far more pro-
tracted during the cool season, extending from early
November until the second half of April. The pro-
tracted nature of the winter fledging period was due
to very slow growth by some chicks that endured
considerable periods without feeds (up to 10 consecu-
tive days on occasion).
Timing of primary moult
Among cool season birds, the earliest occurrence of
primary moult was noted on 29 January, when two
of 12 individuals examined had begun to moult the
two innermost primaries (Fig. 2). Most cool-season
chicks were well-grown at this time. Among 36
cool-season birds returning to the breeding colony in
August and examined for moult, eight were still in
the final stages of re-growing their outermost primaries.
The latest occurrence of primary moult among cool-
season birds was noted on 17 October (two individuals
among 31 examined on that night). It is difficult to
obtain information on the progression of moult of
cool-season birds between April and July as they do
720
M. Bolton
et al.
© 2008 The Authors
Journal compilation © 2008 British Ornithologists’ Union
not attend the colony during these months. How-
ever, Madeiran Storm-petrels are seen off the east
coast of the USA at this time (Sangster 1999) and a
cool-season individual with a fully developed
brood patch ringed on Praia Islet on 25 September
1993 was subsequently captured alive on a fishing
boat in the Gulf of Mexico on 27 April 1998, when
it was in active primary moult (score
c
. 33, Woolfenden
et al
. 2001; Fig. 2). A further four individuals collected
from the Gulf of Mexico between May and July were
Figure 1. Timing of egg laying (bars show the proportion of nests at which an egg is laid in each time period); incubation (nests where
incubation is in progress) and chick-rearing periods of Madeiran Storm-petrels on Praia Islet, Azores, during the hot and cool seasons
(n = 120 and 103 nests, respectively).
Figure 2. Timing of primary moult of individuals from seasonal populations of Madeiran Storm-petrels in Azores. Note the x-axis covers
19 months to embrace the complete moult cycle of both populations. Data on primary moult of five specimens collected from the Gulf
of Mexico between April and July presented by Woolfenden et al. (2001) are also plotted.
© 2008 The Authors
Journal compilation © 2008 British Ornithologists’ Union
New species of storm-petrel
721
all in active primary moult (Woolfenden
et al
. 2001;
Fig. 2), indicating they were also cool-season, not
hot-season, breeders.
Among hot-season birds, commencement of primary
moult was first noted on 5 August, when hot-season
chicks were well grown (Fig. 2). The last date on which
a hot-season bird was still in primary moult was
17 May (one individual of 14 examined on that date).
Examination of the timing of moult of individuals
that can be reliably ascribed to a seasonal population
indicates that moult of primary feathers extends
from the latter stages of one breeding season to the
start of the next and can therefore be used to determine
population membership of adults at any time of year.
For example, two individuals trapped in late October
partway through primary moult (Fig. 2) were judged
to belong to the hot season population, whose
members would be expected to be at this stage
of primary moult in late autumn, in contrast to the
many cool-season individuals caught at this time,
which had all recently completed primary moult.
This identification was confirmed from biometrics
for both individuals.
Exchange of individuals between
seasons
A fuller understanding of the timing of breeding
seasons and the use of moult patterns to establish the
population membership of individuals, have revealed
a greater period of overlap in colony attendance of
the seasonal populations than was previously recog-
nized. The earliest attendance of a cool-season bird
(assessed on the state of the brood patch, primary wear
and moult) was 13 August and the latest attendance
of a hot-season bird (similarly assessed) was 30 October.
Between late August and early October, cool-season
birds may share nest sites during daylight with well-
grown hot-season chicks. However, despite colony
attendance by both populations during the summer
months, we have found only one instance of switch-
ing seasonal colony attendance among 1339 recap-
tures (1122 on Praia, 217 on Vila) involving 728
individuals (568 on Praia, 160 on Vila) from a total
of 4447 individuals ringed between 1990 and 2007.
The single occurrence relates to an individual ringed
as a hot-season nestling on Praia Islet in July 2003,
recaptured by mist-net on Vila Islet during the cool
season (18 November 2005). However, there was no
evidence that this individual was breeding during the
cool season. Additionally, all 1050 (900 on Praia,
150 on Vila) recaptures of 458 (353 on Praia, 105 on
Vila) birds known to have made at least one breeding
attempt, including 27 birds ringed as nestlings (25
on Praia, 2 on Vila) and later found breeding, have
been within the same season.
Biometrics
We found highly significant differences between
seasonal populations in all body measures examined,
except wing length (Table 1). On average, hot-season
Storm-petrels are smaller in terms of body mass, and
have longer, more deeply forked tails, smaller heads
and shorter thinner bills than Storm-petrels nesting
in the cool season.
Foraging environment and trophic level
There were clear differences in both
δ
13
C and
δ
15
N
between Storm-petrels breeding in the hot-season
and those breeding in the cool-season (Fig. 3, Wilks’
λ
,
F
2,196
= 33.38,
P
< 0.001 and
F
2,196
= 37.27,
P
< 0.001,
respectively). Additionally, among the cool-season
Table 1. Biometrics of Storm-petrels breeding in the cool and hot seasons on Praia Islet, Graciosa, Azores. Mean ± sd (min – max, n).
Significance was determined either with two-sample t-tests (where † indicates samples with unequal variances), or non-parametric
Wilcoxon two-sample tests for those variables whose distribution departed from normality (tail length, tail fork, bill depth and tarsus
length).
Variable Cool season Hot season Significance
Wing length (mm) 158.44 ± 3.54 (149–166, 129) 159.21 ± 3.63 (150–169, 209) t336 = −1.91, P = 0.057
Weight (g) 47.28 ± 3.06 (38.5–55.1, 115) 44.57 ± 4.01 (34.8–56.1, 227) t289 = 6.93†, P < 0.0001
Tail length (mm) 72.35 ± 2.53 (66–79, 124) 74.71 ± 3.34 (62–84, 206) Z = 6.73, P < 0.0001
Tail fork (mm) 3.53 ± 1.47 (0–7, 124) 7.81 ± 2.30 (1–14, 205) Z = 13.56, P < 0.0001
Bill length (mm) 15.14 ± 0.49 (13.8–16.4, 117) 14.90 ± 0.55 (13.7–16.2, 207) t322 = 4.03, P < 0.0001
Bill depth (mm) 5.20 ± 0.27 (4.6–6.2, 117) 5.09 ± 0.23 (4.6–5.8, 207) Z = 3.76, P = 0.0002
Head and bill length (mm) 41.62 ± 0.81 (39.4–43.4, 125) 40.31 ± 0.87 (37.6–42.5, 208) t331 = 13.63, P < 0.0001
Tarsus length (mm) 23.57 ± 0.76 (21.5–25.2, 112) 23.27 ± 0.77 (21.4–24.7, 205) Z = 3.06, P = 0.002
722
M. Bolton
et al.
© 2008 The Authors
Journal compilation © 2008 British Ornithologists’ Union
breeders there were differences between primary
and secondary feathers in carbon (F1,104 = 12.59,
P < 0.05) but not nitrogen signatures. Among hot-
season breeders the reverse pattern was found:
primary and secondary feathers differed in nitrogen
(F1,92 = 38.31, P < 0.001) but not in carbon signatures.
Taxonomic recommendation
All available evidence indicates that the seasonally
segregated populations of storm-petrels breeding
in the Azores should be regarded as distinct species
by either the biological or phylogenetic species
concepts (Helbig et al. 2002). Specifically, the taxa
are both diagnosable (see below) and likely to retain
their genetic and phenotypic integrity in the future.
As the genetic evidence shows that the Azores cool-
season population is conspecific with the nominate
O. castro from the Desertas, for the Azores hot-season
population we propose the name:
Oceanodroma monteiroi, sp. nov.
Monteiro’s Storm-petrel
Holotype: adult male, American Museum of
Natural History AMNH 528602 collected 25 April
1903 from a rock crevice on ‘Praya’ Islet, Graciosa,
procured by W.R. Ogilvie-Grant during his expedition
to the Azores islands (Hartert & Ogilvie-Grant 1905).
Previously identified as O. castro.
Description of the holotype
Measurements: Head and bill length 39.5 mm; bill
length to feathering 13.9 mm; bill depth at gonys
5.2 mm; wing length 157 mm; tarsus 22.5 mm;
outer tail feather length 73 mm; central tail feather
length 60 mm; tail fork 13 mm (all measurements
taken by M.B.).
Bare parts: Iris dark brown; bill, legs and feet black.
Plumage: Colour descriptions follow Smithe (1975).
Plumage all one generation, no feather moult in progress.
Primaries fresh with no sun-bleached ‘shadows’.
Upperparts: Entire head, mantle, scapulars, back
and upper rump and upper tail sooty brown-black
(Sepia, Colour 119). Nape, mantle and scapulars,
back and upper rump with grey wash in good light.
Longest scapulars thinly edged very pale grey. Upper
tail coverts white (white feather shafts) forming a
white rump band c. 15 mm wide in the centre and
c. 20 mm wide at the sides. Lower 10 mm of upper
tail coverts tipped black, concolorous with upper tail.
Bases of outer three tail feathers white, extending
33 mm from feather base on outmost tail feather.
Underparts: Chin and throat slightly paler than
upper headparts. Breast, belly, upper flanks, central
undertail coverts noticeably browner than upperparts
(Fuscous, Colour 21). Lower flanks and bases of
outer undertail coverts white, forming a contiguous
white band with rump, so the white rump ‘wraps
around’ the flanks, but not joining ventrally. Underside
of tail feathers brown-black (very similar to underparts)
with silvery sheen. Undertail coverts extending almost
to tail tip centrally.
Wing: Lesser coverts, carpal coverts, primary coverts,
primaries and secondaries uniformly sooty brown-
black (Sepia, Colour 119). Noticeable Light-Drab
(Colour 119C) diagonal wing bar extending from
tertials (and inner secondary), across the greater and
median coverts, to the outer lesser coverts, almost
reaching the leading edge of the wing. The outer three
greater coverts have outer lower edge concolorous
with the primary coverts. It was not possible to carry
out a detailed examination of the underwing, but it
appeared wholly dark, similar in colour to underparts.
Diagnosis
Monteiro’s Storm-petrel O. monteiroi is some 5–10%
smaller in terms of body mass than the Madeiran
Storm-petrel. Monteiro’s Storm-petrel is propor-
tionately longer-winged and has on average a longer,
more deeply forked tail (Monteiro & Furness 1998;
Figure 3. Ratios of stable isotopes of carbon and nitrogen in
primary and secondary feathers of seasonal populations of
storm-petrels nesting on Praia Islet and Vila Islet in the Azores.
© 2008 The Authors
Journal compilation © 2008 British Ornithologists’ Union
New species of storm-petrel 723
Table 1 and Fig. 4) and a shorter thinner bill,
although there is overlap in all these measurements.
Vocalizations are the single most useful diagnostic
feature of Monteiro’s Storm-petrel, which enables
separation from all populations of Madeiran Storm-
petrel examined so far. Burrow calls of Monteiro’s
Storm-petrel are diagnostically different from those
of Madeiran Storm-petrels nesting in the Azores,
Cape Verde and the Galapagos Islands, in both the
duration and structure of the breath notes that punctuate
the purr phrases (see Bolton 2007 for sonograms).
The breath note of Monteiro’s Storm-petrel is
shorter and contains fewer syllables than the corres-
ponding phrase of the Madeiran Storm-petrel, a
difference which is easily discernible to the human ear.
Due to the differences in the timing of breeding
of Monteiro’s and Madeiran Storm-petrels in the
Azores islands, there are similar differences in the
timing of wing moult, which commences towards
the end of the breeding season and is completed
towards the onset of the following breeding season.
The difference in the timing of primary moult is
currently the most useful single diagnostic criterion
outside the breeding season, and may assist the
identification of individuals at times of the year
when both species attend the colony (principally
August, with some Monteiro’s Storm-petrels remain-
ing as late as October; Figs 1 & 2). At this time,
Monteiro’s Storm-petrels have worn primary feathers
and are in active primary moult, beginning with the
innermost feathers. In contrast, Madeiran Storm-
petrels nesting in the Azores complete primary moult
in August, and consequently the outermost feathers
are very fresh, having been replaced relatively recently
(Fig. 5).
The use of primary moult as an identification
feature for individuals away from colonies will be
complicated by the existence of summer-breeding
Madeiran Storm-petrels elsewhere in their Atlantic
range (e.g. the Desertas Islands) which must presum-
ably share a similar moult cycle to Monteiro’s Storm-
petrel. Immature individuals and failed breeders
may also commence moult earlier than successful
breeders. Away from the breeding colonies, field
identification of Monteiro’s Storm-petrel is likely to
present similar challenges to the separation of Zino’s
Petrel Pterodroma madeira and Fea’s Petrel Ptero-
droma feae. The biometric differences noted above
are slight and unlikely to be perceptible under field
conditions.
Etymology
The specific name lugubris is available for any new
taxon split from O. castro, following misidentifica-
tion of a holotype specimen Procellaria lugubris
Bonaparte, 1845 now held at the Naturhistorisches
Museum Wien (NMW 40.989, Bauernfeind &
Schifter 2003). The specimen is obviously not
Hydrobates pelagicus, with which it was previously
synonymized, and agrees well with O. castro
(Bauernfeind & Schifter 2003). If this specimen
were found to belong to the taxon newly described
here, the specific name lugubris would take prece-
dence. Although the precise location and date of
collection are unknown, body measurements show
that it does not belong to the taxon we describe: the
exposed culmen of the lugubris holotype (16.4 mm)
lies outside the range for the Azores species
Figure 4. Difference in the extent of the tail fork between the
Madeiran Storm-petrel (a) and Monteiro’s Storm-petrel (b).
Photographs taken in August, when both species are present at
the breeding colony. At this time, Monteiro’s Storm-petrels are
tending large chicks and have worn plumage (note brown tinge
to tail feathers), whereas Madeiran Storm-petrels are prospecting
for nest sites and have fresh plumage having recently completed
the annual moult.
724 M. Bolton et al.
© 2008 The Authors
Journal compilation © 2008 British Ornithologists’ Union
Figure 5. Monteiro’s Storm-petrel Oceanodroma monteiroi (right) and Madeiran Storm-petrel Oceanodroma castro in flight off the only
two known breeding islets of the former: Praia Islet (foreground) and Baixo Islet, Graciosa, Azores. The figure illustrates the differences
in plumage wear and moult that are apparent in August, when both species attend the colonies in large numbers. Monteiro’s Storm-petrels
show faded plumage and have commenced moult of the inner primaries, whereas Madeiran Storm-petrels have relatively fresh plumage
and have recently completed moult.
© 2008 The Authors
Journal compilation © 2008 British Ornithologists’ Union
New species of storm-petrel 725
(Table 1). Additionally, the tail of the lugubris holotype
is not forked (central tail feathers equal in length to
the outer tail feathers, which may account for the
original misidentification as H. pelagicus). The nom-
inate O. castro can exceptionally show an unforked
tail (three individuals among 124 examined by M.B.,
Table 1), but no instance of an entirely unforked tail
was found among 205 individuals of the proposed
new species. We therefore propose the specific name
monteiroi, which acknowledges the huge contribution
to the discovery of this species made by the late
Dr Luís Monteiro, who first described the seasonal
populations of storm-petrels in the Azores and who
worked tirelessly for their conservation. We follow
the recommendations of ICZN (1999) in the formation
of the specific name.
DISCUSSION
Population size, distribution and foraging
On current knowledge, Monteiro’s Storm-petrel is
known only from the Azores islands and appears to
have diverged between 125 000 and 300 000 years
ago from populations in the Pacific and between
70 000 and 350 000 years ago from other popula-
tions within the North Atlantic (Friesen et al. 2007,
Smith et al. 2007). Within the Azores, the species is
known to breed only on two small (c. 12-ha) islets,
some 5 km apart, lying off the inhabited island of
Graciosa. The combined breeding population at
these two sites has been estimated at 200 pairs
(Monteiro et al. 1996, 1999). Further small colonies,
possibly holding a few tens of pairs, are suspected on
other islets lying off Graciosa, Flores and Corvo
(Monteiro et al. 1999). The total breeding popula-
tion is therefore estimated at between 250 and
300 pairs, but there is a clear need for more accurate
and up-to-date information on population size and
demography.
The distribution of Monteiro’s Storm-petrel outside
the breeding season is currently unknown, although
the capture of two individuals on Praia Islet in late
October and a further individual on Vila Islet in mid
November suggests that it remains in the vicinity of
the breeding grounds, rather than dispersing into the
western Atlantic, as does the Madeiran Storm-petrel.
The analysis of carbon isotopes supports this suggestion,
as for Monteiro’s Storm-petrel there were no differ-
ences in δ13C between feathers grown in the summer
and winter months, suggesting a similar geographic
origin of these feathers, whereas for the Madeiran
Storm-petrel there were significant differences in
carbon signatures between the primary and secondary
feathers. Whilst the diets of the Madeiran Storm-petrel
and Monteiro’s Storm-petrel are currently poorly
known, Monteiro et al. (1998) found considerably
higher levels of mercury in both the regurgitated
prey and body feathers of cool-season Madeiran Storm-
petrels than hot-season storm-petrels, and suggested
that the cool-season birds had a higher dependence
on mesopelagic prey. Analysis of δ13C data presented
here indicates that the diet of Monteiro’s Storm-petrel
has much less of a benthic component than that of
the Madeiran Storm-petrel. Even during the late winter
(February–March) when both species are present in
waters around the Azores, they appear to show dietary
differences: we found considerable differences in
both δ15N and δ13C of Madeiran Storm-petrel primary
feathers and Monteiro’s Storm-petrel secondary
feathers, which would both be growing around this
time. Analysis of δ15N suggests that Monteiro’s
Storm-petrel generally feeds at a higher trophic level
than the Madeiran Storm-petrel and shows a dietary
shift to higher trophic levels during the winter months.
Conservation threats
In common with other members of the Order Pro-
cellariiformes, Monteiro’s Storm-petrel lays a single
egg and annual productivity is low due to interspecific
competition with other, larger, burrowing Procellar-
iiformes. Productivity averaged just 0.16 fledglings
per pair over 2 years (2000–2001) on Praia Islet
(Bolton et al. 2004). The species therefore has limited
potential for population growth and consequently
low resilience to adverse effects. Identified threats
include high predation rates by Long-eared Owls
Asio otus that are resident in the Azores. Up to 40
adult Storm-petrels are killed in some seasons, including
some known from ring recoveries to be breeders
(M. Bolton pers. obs.). The two known breeding islets
are currently free of ground predators, but their close
proximity (c. 1 km offshore) to the inhabited island
of Graciosa, and the large numbers of visitors to one
islet during summer means that the introduction of
rodents is a possibility. Both islets lie within 2 km of
the main shipping route for large passenger ferries
and container ships docking on Graciosa. A cargo
ship containing livestock ran aground on one of the
breeding islets in 2000, leading to concerns over
pollution and rodents escaping ashore.
Both breeding islets are designated under European
legislation as Special Protection Areas and are assigned
726 M. Bolton et al.
© 2008 The Authors
Journal compilation © 2008 British Ornithologists’ Union
a full-time warden based on the island of Graciosa.
Recent work to reduce interspecific competition for
nest cavities with other larger procellariiform species
through the installation of nest boxes has met with
considerable success, leading to large increases in
annual productivity (Bolton et al. 2004). Nest boxes
and natural sites are monitored annually to determine
occupancy and breeding productivity. In 2007, 22
young fledged from 56 breeding attempts in nest
boxes and a further 12 pairs reared four young in
monitored natural nest sites.
The early 16th and 17th century accounts from
the Azores of a small black and white seabird that
was killed nightly by the thousands, relate that it had
young in the nest in September–October (Fructuoso
1561, Chagas 1645–1650), which accords more
closely with the breeding phenology of Monteiro’s
Storm-petrel than the Madeiran Storm-petrel. It
seems likely therefore that these early descriptions of
a seabird so abundant it was knocked from the air
using sticks, almost certainly relate to the species
newly described here. In comparison with its former,
virtually unimaginable, abundance, Monteiro’s Storm-
petrel currently survives by a very slender margin
indeed, and work must now be undertaken to establish
an action plan to ensure its continued existence.
We thank the many field assistants and research students
who have assisted with data collection and analysis, especially
Verónica Neves, Luís Aguiar, Ana Campos, Becky Hothersall,
Luís Dias and Maria Carvalho. Fieldwork on Praia Islet was
funded by grants from the Portuguese Fundação para a
Ciência e a Tecnologia (PRAXIS/C/BIA/13194/98 and
POCTI-BIA-13194/98), a post-doctoral research grant
(SFRH/BPD/20291/2004) awarded to J.B., and the Pro-
grammes ‘OGAMP’ (Planning and Management of Marine
Protected Areas, Interreg IIIB-MAC/4.2/A2), and ‘MAR-
MAC’ (Knowledge, Promotion and Valorization for a
Sustainable Utilization of Marine Protected Areas
in Macaronesia, Interreg IIIB/FEDER/MARMAC/003-4/
2005-6 and Interreg IIIB-05/MAC/4.2/A4). IMAR-DOP/
University of Azores is funded through the pluri-annual
and programmatic funding schemes of FCT (Portugal) and
DRCT (Azores, Portugal) as Research Unit #531 and Associ-
ate Laboratory #9. We thank Dr Jacob González-Solís for
his supervision of the work on stable isotopes and micro-
satellites, and for comments on an earlier draft of this paper.
We thank the Azorean Direcção Regional do Ambiente, the
Câmara Municipal de Graciosa and the Junta de Freguesia
de Santa Cruz for access permissions, provision of neces-
sary licences and logistic support for fieldwork. Consider-
able assistance in locating the holotype specimen, and
verifying its identification was provided by Dr Robert
Prys-Jones of the Natural History Museum and Paul Sweet
of the American Museum of Natural History. Mary LeCroy
of the AMNH made helpful comments on an earlier draft
and Dr Ernst Bauernfeind of the Vienna Natural History
Museum kindly provided data and photographs on the
lugubris holotype.
REFERENCES
Allan, R.G. 1962. The Madeiran Storm Petrel Oceanodroma
castro. Ibis 103b: 274–294.
Arroyo, B., Mínguez, E., Palomares, L. & Pinilla, J. 2004. The
timing and pattern of moult of flight feathers of European
Storm-petrel Hydrobates pelagicus in Atlantic and Mediterranean
breeding areas. Ardeola 51: 365–373.
Bannerman, D. & Bannerman, W.M. 1966. Birds of the Atlantic
Islands, Vol. III. A History of the Birds of the Azores. Edin-
burgh: Oliver & Boyd.
Bauernfeind, E. & Schifter, H. 2003. Oceanodroma castro
(Harcourt, 1851) – nomen protectum. Bull. Brit. Orn. Club
123: 280–283.
Bearhop, S., Waldron, S., Votier, S.C. & Furness, R.W. 2002.
Factors that influence assimilation rates and fractionation of
nitrogen and carbon stable isotopes in avian blood and
feathers. Physiol. Biochem. Zool. 75: 451–458.
Bolton, M. 2007. Playback experiments indicate absence of
vocal recognition among temporally and geographically
separated populations of Madeiran Storm-petrels Oceano-
droma castro. Ibis 149: 255–263.
Bolton, M., Medeiros, R., Hothersall, B. & Campos, A. 2004.
The use of artificial breeding chambers as a conservation
measure for cavity-nesting procellariiform seabirds: a case
study of the Madeiran Storm Petrel (Oceanodroma castro).
Biol. Conserv. 116: 73– 80.
Chagas, D. 1645–1650. Espelho Crystalino em Jardim de Varias
Flores, published in 1989, ed. A.T. Matos. Angra do Heroísmo,
Azores: Secretaria Regional da Educação e Cultura.
Cherel, Y., Hobson, K.A. & Weimerskirch, H. 2000. Using
stable-isotope analysis of feathers to distinguish moulting
and breeding origins of seabirds. Oecologia 122: 155–
162.
France, R.L. 1995. Carbon-13 enrichment in benthic compared
to planktonic algae: foodweb implications. Mar. Ecol. Progr.
Ser. 124:307–312.
Friesen, V.L., Smith, A.L., Gómez-Díaz, E., Bolton, M.,
Furness, R.W., González-Solís, J. & Monteiro, L.R. 2007.
Sympatric speciation by allochrony in a seabird. Proc. Natl
Acad. Sci. U S A 104: 18589–18594.
Fructuoso, G. 1561. Saudades da Terra, 2nd edn published in
6 volumes from 1978 to 1983, ed. J.B.O. Rodrigues. Ponta
Delgada, Azores: Instituto Cultural de Ponta Delgada.
Ginn, H.B. & Melville, D.S. 1983. Moult in Birds. BTO Guide 19.
Tring: British Trust for Ornithology.
Harcourt, E.V. 1851. A Sketch of Madeira. London: John Murray.
Harris, M.P. 1969. The biology of storm petrels in the Galapagos
Islands. Proc. Calif. Acad. Sci. 37: 95–165.
Hartert, E. & Ogilvie-Grant, W. 1905. On the birds of the
Azores. Novit. Zool. 12: 80–128.
Helbig, A.J., Knox, A.G., Parkin, D.T., Sangster, G. &
Collinson, M. 2002. Guidelines for assigning species rank.
Ibis 144: 518–525.
ICZN 1999. International Code of Zoological Nomenclature, 4th
edn. London: International Trust for Zoological Nomenclature.
© 2008 The Authors
Journal compilation © 2008 British Ornithologists’ Union
New species of storm-petrel 727
Kelly, J.F. 2000. Stable isotopes of carbon and nitrogen in the
study of avian and mammalian trophic ecology. Canadian
Journal of Zoology. 78: 1–27.
Le Grand, G., Emmerson, K. & Martin, A. 1984. The status
and conservation of seabirds in the Macaronesian Islands.
In Croxall, J.P., Evans, P.G.H. & Schreiber, R.W. (eds) Status
and Conservation of the World’s Seabirds: 377–391.
Cambridge, UK: International Council for Bird Preservation,
Tech. Publ. No. 2.
Minagawa, M. & Wada, E. 1984. Stepwise enrichment of δ15N
along food chains: further evidence and relation between δ15N
and animal age. Geochim. Cosmochim. Acta 48: 1135–1140.
Monteiro, L.R. & Furness, R.W. 1998. Speciation through
temporal segregation of Madeiran Storm-petrel (Oceanodroma
castro) populations in Azores? Phil. Trans. R. Soc. Lond. B
353: 845–953.
Monteiro, L.R. Granadeiro, J.P. & Furness, R.W. 1998.
Relationship between mercury levels and diet in Azores seabirds.
Mar. Ecol. Prog. Ser. 166: 259–265.
Monteiro, L.R., Ramos, J.A., Pereira, J.C., Monteiro, P.R.,
Feio, R.S., Thompson, D.R., Bearhop, S., Furness, R.W.,
Laranjo, M., Hilton, G., Neves, V.C., Groz, M.P. & Thompson,
K.R. 1999. Status and distribution of Fea’s Petrel, Bulwer’s
Petrel, Manx Shearwater, Little Shearwater and Band-rumped
Storm-petrel in the Azores archipelago. Waterbirds 22: 358–
366.
Monteiro, L.R., Ramos, R.A. & Furness, R.W. 1996. Past and
present status and conservation of the seabirds breeding in
the Azores archipelago. Biol. Conserv. 78: 319–328.
Nunes, M. 2003. Contributo das Vocalizações de Ninho para a
Distinção das Populaçoes de Período-frio e Período-quente
de Angelito Oceanodroma castro (Harcourt 1851). Masters
Thesis, Portugal: University of Coimbra.
Nunn, G.B. & Stanley, S.E. 1998. Body size effects and rates
of cytochrome b evolution in tube-nosed seabirds. Mol. Biol.
Evol. 15: 1360–1371.
Owens, N.J.P. 1987. Natural variations in 15N in the marine
environment. Adv. Mar. Biol. 24: 389– 451.
Sangster, G. 1999. Cryptic species of storm-petrels in the
Azores? Dutch Birding 21: 101–106.
SAS Institute. 2007. SAS for Windows 9.1.3. Cary, NC: SAS
Institute Inc.
Scott, D.A. 1970. The Breeding Biology of the Storm Petrel
Hydrobates pelagicus. DPhil Thesis, University of Oxford.
Smith, A.L., Monteiro, L.R., Hasegawa, O. & Friesen, V.L.
2007. Global phylogeography of the Band-rumped Storm-
petrel (Oceanodroma castro; Procellariiformes: Hydrobatidae).
Mol. Phyl. Evol. 43: 755–773.
Smithe, F.B. 1975. Naturalist’s Colour Guide. New York: American
Museum of Natural History.
Wenink, P.W., Baker, A.J. & Tilanus, M.G.J. 1993. Hypervariable-
control-region sequences reveal global population structuring
in a long-distance migrant shorebird, the Dunlin (Calidris
alpina). Proc. Natl Acad. Sci. U S A 90: 94–98.
Woolfenden, G.E., Monteiro, L.R. & Duncan, R.A. 2001.
Recovery from the northeastern Gulf of Mexico of a
Band-rumped Storm-petrel banded in the Azores. J. Field
Orn. 72: 62–65.
Received 7 November 2007;
revision accepted 8 May 2008.