Bulletin of The British Ornithologists' Club 2010 130(4):286-301
47499910

A new taxon of Collared Petrel Pterodroma brevipes from the Banks Islands, Vanuatu

Vincent Bretagnolle, Hadoram Shirihai
Vincent
Bretagnolle
&
Hadoram
Shirihai
297
Bull.
B.O.C.
2010
130(4)
magnificens
is
a
scavenger
that
remains
behind
the
flock,
commencing
to
feed
only
as
the
flock
starts
to
disperse
or
moves
to
the
next
location.
To
our
knowledge,
these
are
the
first
observations
of
Pterodroma
petrels
forming
constant
feeding
associations
with
inshore
terns
and
boobies.
Occasionally
other
gadfly
petrels
feed
similarly,
especially
White-necked
/
Vanuatu
and
Juan
Fernandez
Petrels
P.
externa,
as
well
as
several
Cookilaria
(De
Filippi's
P.
defilippiana,
Stejneger's
P.
longirostris,
Gould's
and
Collared
Petrels)
and
Black-winged
Petrel
P.
nigripennis
(pers.
obs.),
but
such
behaviour
is
never
as
habitually
observed
as
in
P.
b.
magnificens
off
the
Banks.
Interestingly,
it
is
also
the
only
gadfly
petrel
to
our
knowledge
feeding
so
close
to
islands,
which
may
be
related
to
its
very
dark
coloration
(Bretagnolle
1993).
However,
it
remains
to
be
seen
if
such
behaviour
is
seasonal,
and
to
what
degree
it
depends
on
factors
such
as
oceanographic
features.
Identification
at
sea.
P.
b.
magnificens
mirrors
several
other
subspecies
and
'cryptic
species'
of
petrels
in
lacking
absolute
diagnostic
characters
to
separate
it
at
sea
from
the
dark
morph
of
other
P.
brevipes
populations.
Especially
away
from
the
breeding
islands,
P.
b.
magnificens
will
not
be
separable
with
certainty:
according
to
our
observations
c.60%
of
birds
are
identical
to
the
darkest
examples
in
other
P.
brevipes
populations.
Nonetheless,
c.40%
of
P.
b.
magnificens
are
even
darker,
and
thus
referred
to
the
'extreme
dark
grey'
type,
which
has
uniformly
very
dark
underparts
and
much
of
the
underwing,
with
a
limited
area
of
pure
white
on
the
innermost
coverts
and
the
larger
row
of
median
coverts,
and
a
small
white
throat
(see
Figs.
9-10).
In
such
birds
the
dark
underwing
bar
almost
covers
half
the
width
of
the
wing,
i.e.
the
broadest
of
any
gadfly
petrel,
even
broader
than
Chatham
Petrel
Pterodroma
axillaris.
Unfortunately
these
differences
are
of
limited
use
for
field
identification,
because
their
correct
appreciation
at
sea
(and
in
photographs)
is
often
hampered
by
the
influence
of
light
and
angle
of
view,
or
it
changes
in
relation
to
the
way
the
wing
is
held.
Much
experience
is
also
required
to
evaluate
such
differences
correctly.
Furthermore,
P.
b.
magnificens
is,
apparently,
very
distinctive
in
its
feeding
flight
behaviour,
but
also
to
some
extent
at
other
times.
It
habitually
flies
very
low,
with
short
and
low
(steep)
banks
and
arcs,
long
and
very
low
glides,
often
with
sudden
sharp
turns.
Many
times
the
birds
adopt
the
flight
mode
and
path
of
Bulwer's
Petrel
or
prions
(Pachyptila),
a
similarity
heightened
by
their
distinctive
structure,
with
relatively
short
wings
but
a
proportionately
long
pointed
tail.
Our
impression
is
that
it
will
prove
possible
to
distinguish
these
birds
on
the
basis
of
their
distinctive
flight
behaviour
and
shape
alone,
given
prolonged
views
and
previous
experience
with
P.
brevipes.
Furthermore,
they
are
often
notably
slight-bodied
and
smaller
with
a
shorter
bill.
Aw^ay
from
their
breeding
islands,
P.
b.
magnificens
compared
to
dark-morph
individuals
of
other
P.
brevipes
populations
cannot
be
reliably
identified,
but
this
should
not
prevent
observers
recording
and
photographing
them.
Only
by
such
means
might
it
eventually
be
possible
to
identify
areas
preferred
by
birds
with
extremely
dark
underparts
and
underwings,
generally
slighter
/
smaller
size
and
the
flight
and
feeding
behaviours
of
P.
b.
magnificens.
We
should
emphasise
that
the
plumage
variation
described
here,
based
on
our
pelagic
observations
and
photographs
(from
Fiji,
Vanuatu
and
elsewhere
in
the
tropical
Pacific)
of
P.
b.
magnificens
and
other
P.
brevipes
populations,
is
completely
new,
i.e.
contra
all
previous
handbooks
and
field
guides
(e.g.
Harrison
1985,
Carboneras
in
del
Hoyo
et
al
1992,
Onley
&
Scofield
2007),
which
wrongly
illustrate
the
underwing
pattern
and
different
morphs.
This
emphasises
the
need
to
combine
field
observations
with
a
comprehensive
review
of
museum
material.
Taxonomic
rank.
Available
evidence
suggests
that
P.
b.
magnificens
should
not
be
viewed
within
monotypic
brevipes.
It
is
diagnosable
using
the
following
combined
features.