Pollination in alpine Norway: flowering phenology, insect visitors, and visitation rates in two plant communities


Totland, O.

Canadian Journal of Botany 71(8): 1072-1079

1993


Pollination studies in European alpine communities are few. The objective of this study was to describe the pollination ecology in two alpine plant communities at Finse, southwestern Norway. Because of late snowmelt and early winter at Finse, the time available for flowering and seed maturation is restricted. Flowering was concentrated at the beginning of the season in both communities, and large overlaps in flowering time were found for most species. In one of the communities, flowering peaks were significantly clumped, whereas in the other they were randomly distributed through the season. However, in this community, five insect-pollinated species flowered simultaneously early in the season. Diptera almost exclusively dominated the visitor assemblage. Most plant species pairs had high overlaps in flower visitor species. Species flowering simultaneously attracted the same visitor species. In one community, eight species pairs flowered sequentially and shared visitors. Visitation rates were highest at the lowest elevated site. The results are compared with those obtained in other alpine areas. It is argued that selection for an early flowering is probably stronger than selection pressures resulting from interspecific interactions.

1072
Pollination
in
alpine
Norway:
fl
owering
phenology,
insect
visitors,
and
visitation
rates
in
two
plant
communities
ORJAN
TOTLAND
Botanical
Institute,
University
of
Bergen,
Allegaten
41,
N-5007
Bergen,
Norway
Received
February
5,
1993
TOTLAND,
0.
1993.
Pollination
in
alpine
Norway:
fl
owering
phenology,
insect
visitors,
and
visitation
rates
in
two
plant
communities.
Can.
J.
Bot.
71:
1072—
1079.
Pollination
studies
in
European
alpine
communities
are
few.
The
objective
of
this
study
was
to
describe
the
pollination
ecology
in
two
alpine
plant
communities
at
Finse,
southwestern
Norway.
Because
of
late
snowmelt
and
early
winter
at
Finse,
the
time
available
for
flowering
and
seed
maturation
is
restricted.
Flowering
was
concentrated
at
the
beginning
of
the
season
in
both
communities,
and
large
overlaps
in
fl
owering
time
were
found
for
most
species.
In
one
of
the
communities,
fl
owering
peaks
were
significantly
clumped,
whereas
in
the
other
they
were
randomly
distributed
through
the
season.
However,
in
this
community,
five
insect
-pollinated
species
fl
owered
simultaneously
early
in
the
season.
Diptera
almost
exclusively
dominated
the
visitor
assemblage.
Most
plant
species
pairs
had
high
overlaps
in
fl
ower
visitor
species.
Species
fl
owering
simultaneously
attracted
the
same
visitor
species.
In
one
community,
eight
species
pairs
fl
owered
sequentially
and
shared
visitors.
Visitation
rates
were
highest
at
the
lowest
elevated
site.
The
results
are
compared
with
those
obtained
in
other
alpine
areas.
It
is
argued
that
selection
for
an
early
flowering
is
probably
stronger
than
selection
pressures
resulting
from
interspecific
interactions.
Key
words:
alpine,
Diptera,
fl
owering
phenology,
fl
ower
visitors,
season
length,
visitation
rate.
TOTLAND,
0.
1993.
Pollination
in
alpine
Norway:
fl
owering
phenology,
insect
visitors,
and
visitation
rates
in
two
plant
communities.
Can.
J.
Bot.
71
:
1072
—1079.
Il
n'existe
que
peu
d'etudes
sur
la
pollinisation
dans
les
communautes
alpines
de
l'Europe.
L'auteur
s'est
propose
de
decrire
Pecologie
de
la
pollinisation
dans
deux
communautes
de
plantes
alpines
de
la
region
de
Finse,
dans
le
sud-ouest
de
la
Norvege.
Parce
que
la
fonte
de
la
neige
est
tardive
et
que
l'hiver
est
hatif
a
Finse,
le
temps
disponible
pour
la
fl
oraison
et
la
maturation
des
graines
est
limite.
La
fl
oraison
survient
rapidement
au
debut
de
la
saison
dans
les
deux
communautes
et
on
retrouve
un
important
recouvrement
des
periodes
de
fl
oraison
pour
la
plupart
des
especes.
Dans
une
des
communautes,
les
periodes
d'anthese
sont
significativement
regroupees,
alors
que
dans
l'autre
ces
periodes
sont
distribuees
au
hasard
tout
au
long
de
la
saison.
Cependant,
dans
cette
communaute,
cinq
especes
pollinisees
par
les
insectes
fl
eurissent
simultanement,
et
tot
dans
la
saison.
Ce
sont
exclusivement
des
dipteres
qui
dominent
le
cortege
des
pollinisateurs.
Pour
la
plupart
des
paires
de
plantes,
on
note
de
forts
recouvrements
des
especes
visitant
leurs
fl
eurs.
Les
especes
fleurissant
simultanement
attirent
les
memes
especes
de
pollinisateurs.
Dans
une
des
communautes,
huit
paires
d'especes
fl
eurissant
en
sequence
partagent
les
memes
polli-
nisateurs.
Les
taux
de
visites
sont
plus
eleves
aux
basses
altitudes.
L'auteur
compare
ses
resultats
avec
ceux
obtenus
dans
d'autres
regions
alpines,
et
it
propose
que
la
selection
en
faveur
d'une
fl
oraison
hative
serait
probablement
plus
forte
que
les
pressions
selectives
provenant
d'interactions
interspecifiques.
Mots
des
:
alpin,
diptere,
phenologie
fl
orale,
pollinisateurs,
longueur
de
la
saison,
taux
de
visites.
[Traduit
par
la
redaction]
Introduction
The
fl
owering
periods
of
the
animal
-pollinated
plants
in
a
community
may
be
influenced
by
numerous
factors.
Interspecific
competition
for
pollination
is
the
most
frequently
cited
expla-
nation
for
the
distribution
pattern
of
fl
owering
times
(Mosquin
1971;
Pleasants
1980,
1983;
Waser
1983).
Competitive
inter-
actions
between
species
favor
divergence
in
fl
owering
times
or
fl
ower
morphology,
such
that
negative
effects
of
pollinator
sharing
are
minimized.
Therefore,
if
competition
for
pollina-
tion
was
a
sufficiently
strong
selection
pressure
in
the
past,
the
fl
owering
times
within
a
community
should
be
regularly
dis-
tributed
through
the
season
(Poole
and
Rathcke
1979;
Rathcke
1983,
1984).
Conversely,
positive
interaction
for
pollination
occurs
if
species
experience
a
higher
visitation
rate
when
fl
ower-
ing
simultaneously
than
when
fl
owering
sequentially
(Thomson
1982, 1983;
Rathcke
1983;
Laverty
and
Plowright
1988;
Laverty
1992).
Such
interactions
may
select
for
convergence
in
fl
owering
time
or
fl
ower
morphology.
Length
of
season
may
influence
flowering
times
(Rathcke
and
Lacey
1985;
Primack
1987).
In
many
alpine
areas
the
time
available
for
shoot
develop-
ment,
fl
owering,
pollination,
and
seed
maturation
is
short.
Here,
an
early
season
fl
owering
period
may
be
a
prerequisite
for
com-
pletion
of
seed
production
before
winter,
particularly
when
fruits
and
seeds
are
relatively
large
in
size
(Bliss
1956,
1971;
Billings
and
Mooney
1968),
and
interspecific
interactions
for
pollina-
tion
may
be
less
important
in
explaining
the
distribution
of
fl
owering
times
within
alpine
plant
communities.
Autogamous
pollination,
frequent
in
alpine
populations
(Crawford
1989),
may
suppress
interactions
between
species
because
individuals
are
independent
of
pollinating
agents
for
seed
production.
Therefore,
autogamous
populations
may
have
weak,
or
no,
interactions
for
pollination
with
other
populations,
which
consequently
leads
to
lower
selection
pressure
for
diver-
gence
or
convergence
in
fl
owering
time.
Autogamous
pollination
does
not
obviate
early
fl
owering
in
short
-season
environments.
Regardless
of
pollination
strategy,
seed
must
mature
before
winter
conditions
appear.
Diptera
are
frequent
fl
ower
visitors
in
alpine
and
arctic
environments
(Hocking
1968;
Kevan
1972;
Arroyo
et
al.
1982;
Inouye
and
Pyke
1988;
McCall
and
Primack
1992).
The
short
season
in
these
environments
may
restrict
the
establishment
of
social
bee
colonies,
such
as
bumblebees.
Also,
butterflies
are
rarely
recorded
as
flower
visitors
in
alpine
areas.
Besides
being
unselective
fl
ower
feeders
(Faegri
and
van
der
Pijl
1979)
and
inefficient
pollen
vectors
(Kendall
and
Solomon
1973;
Grace
and
Nelson
1981),
a
fl
y
-dominated
pollinator
assemblage
has
low
morphological
and
behavioral
diversity.
Consequently,
the
potential
for
specialization,
in
either
fl
owering
time
or
flower
Printed
in
Canada
/
Imprimd
au
Canada
1UILAND
1073
TABLE
1.
Species
studied
at
Mount
Kvannjolnut
(K)
and
Mount
Sandalsnut
(S)
Species
Abbr.
Ab
K
Ab
s
Flower
color
Breeding
system
References
on
breeding
system
Cerastium
alpinum
L.
Ca
10
10
White
Self
-compatible
Fryxell
1957
Geranium
sylvaticum
L.
Gs
37
0
Purple
Self
-incompatible
Hagerup
1951,
Fryxell
1957
Leontodon
autumnalis
L.
La
41
71
Yellow
Self
-compatible
Fryxell
1957
Parnassia
palustris
L.
Pp
50
41
White
Self
-compatible
Fryxell
1957
Potentilla
crantzii
(Cr.)
Beck
Pc
8
14
Yellow
Apomictic
Fryxell
1957
Ranunculus
acris
L.
Ra
124
128
Yellow
Self
-incompatible
Lundqvist
et
al.
1973
Silene
acaulis
(L.)
Jacq.
Sa
0
808
Pink
—red
Self
-compatible
Shykoff
1992
Taraxacum
croceum
Dahlst.
Tc
2
42
Yellow
Apomictic
Richards
1972
NOTE:
Abbreviations
of
species
names
are
those
used
in
Table
2
and
Figs.
1
and
2.
Ab
s
and
Ab
s
,
fl
ower
abundance
at
K
and
S
in
1991,
respectively,
calculated
as
the
total
number
of
fl
owers
counted
in
the
phenology
plots
per
number
of
census
days.
morphology
towards
a
specific
pollinator
species,
may
be
restricted.
The
objective
of
this
work
was
to
describe
the
pollination
ecology
in
a
geographical
area
that
was
not
previously
inves-
tigated.
I
asked
the
following
questions:
(i)
What
patterns
of
fl
owering
phenology
are
found
in
the
two
communities,
and
which
external
factors
give
the
most
reasonable
explanation
of
these
patterns?
(ii)
Which
fl
ower
visitors
are
the
most
frequent
in
the
two
communities
examined,
and
to
what
extent
do the
flowering
species
share
visitor
species?
(iii)
In
what
aspects
is
the
pollination
ecology
at
Finse
similar
to,
or
different
from,
that
of
other
comparable
areas?
Study
areas
The
study
was
conducted
at
two
alpine
sites
at
Finse,
Hardangervidda,
southwestern
Norway.
Both
were
south
facing
and
of
equal
size
(60
x
60
m).
They
are
2.5
km
apart.
Site
1
was
sampled
in
1991
from
June
16
to
September
9.
It
is
situ-
ated
on
the
south
-facing
slope
of
Mount
Kvannjolnut
(K)
(1300
m
asl;
60°35'N,7°33'E)
and
was
exposed
to
the
sun
between
ca.
07:00
and
21:00
h
during
the
study
period.
Snow
melted
away
in
early
June
at
about
the
same
time
all
over
the
site.
During
June
the
site
occasion-
ally
received
some
snowfall.
The
site
had
a
species
composition
characteristic
of
the
low
alpine
zone
at
Finse.
Most
species
at
the
site
fl
owered
abundantly.
Site
2
was
sampled
in
1990
from
July
6
to
September
9
and
in
1991
from
July
4
to
September
12.
It
is
situated
on
the
south
-facing
hillside
of
Mount
Sandalsnut
(S)
(1500
m
asl;
60°37'N,7°32'E).
I
consider
it
representative
of
middle
alpine
areas
at
Finse
with
high
species
and
fl
ower
densities.
The
snow
melted
gradually
away
down
the
hillside
and
in
both
years
the
last
snow
patches
disappeared
in
early
July.
The
site
occasionally
received
snowfall
in
late
August.
Study
methods
Field
methods
I
measured
flower
preferences
of
insect
visitors
by
observing
a
1
x
1
m
square
for
10
-min
periods
in
1990
and
20
-min
periods
in
1991.
The
duration
of
the
periods
was
doubled
to
avoid
large
numbers
of
zero
visit
periods.
I
positioned
the
square
subjectively
over
plants
before
each
observation
period
started.
I
regarded
flower
-visiting
insects
as
potential
pollinators
if
I
observed
them
to
forage
for
nectar
and
(or)
pollen
and
caught
each
insect
when
it
met
the
requirements
for
being
classified
as
a
pollinator.
I
noted
the
plant
species
visited.
Visitors
escaping
were
classified
to
family
level.
Prior
to
each
observation
period
I
counted
the
numbers
of
fl
owers
of
each
species
present
in
the
square.
For
species
of
Asteraceae
I
counted
the
number
of
inflores-
cences.
I
made
no
effort
to
collect
data
on
visitation
activity
during
periods
with
strong
wind
or
precipitation.
Measurements
were
carried
out
mainly
from
10:00
to
18:00.
I
measured
fl
owering
phenology
at
both
sites
every
2
—4
days
in
1991
by
counting
the
number
of
open
fl
owers
or
inflorescences
for
each
species
in
22
permanent
2
x
2
m
squares.
I
positioned
the
squares
at
random
before
any
species
came
into
flower
at
both
sites.
At
S
in
1990
I
measured
the
fl
owering
phenology
only
qualitatively
by
noting
every
2nd
day
which
species
were
in
bloom
along
seven
permanent
transects,
each
ca.
60
m
long
and
3
m
wide.
Data
analysis
I
used
Horn's
(1966)
niche
overlap
index
to
examine
differences
in
plant
species
niche
relationships.
Smith
and
Zaret
(1982)
showed
that
this
index
is
little
affected
by
differences
in
the
number
of
samples,
number
of
resource
categories,
or
evenness
in
the
data
set.
I
used
Levins'
(1968)
breadth
index
to
quantify
niche
breadth
in
visitor
use.
Bootstrap
analysis
(Mueller
and
Altenberg
1985)
was
used
to
examine
the
robustness
and
bias
of
the
measures.
In
the
niche
-breadth
and
overlap
calculations
there
were
large
differences
in
percent
bias
between
some
of
the
species
and species
pairs
(breadth:
0-19.48%;
overlap:
1.92-15.15%).
Therefore,
I
present
and
use
bias
-corrected
values
in
the
analysis.
In
the
analysis
of
overlap
in
fl
owering
time
in
1991
I
summed
all
counts
of
each
species
in
the
squares
from
1
day.
Because
fl
owering
phenology
was
not
quantified
in
1990,
no
overlap
data
are
presented
for
this
year.
Only
visitor
species
with
a
total
abundance
of
5
individuals
observed
were
used
in
the
niche
analysis
of
overlap
and
breadth
in
visitor
species.
All
niche
measurements
were
calculated
using
the
computer
pro-
gram
NICHE
written
by
D.
Schluter
(University
of
British
Columbia,
Vancouver)
(see
Schluter
1988).
Dispersion
of
fl
owering
peaks
was
calculated
using
the
program
PEAK
written
by
J.M.
Line
(University
of
Cambridge,
Cambridge,
England).
Study
species
The
species
studied
are
presented
in
Table
1.
I
concentrated
sampling
on
fl
ower
preference
of
insect
visitors
and
fl
owering
phenology
to
species
that
were
abundant
at
the
study
sites.
Only
species
that
received
a
sufficient
number
of
visits
for
statistical
analysis
are
presented.
All
species
had
a
fl
ower
morphology
that
allowed
fl
ies
to
forage
in
them.
No
detailed
examination
was
done
of
breeding
systems
of
the
flower-
ing
species.
The
information
given
in
Table
I
should
therefore
be
considered
suggestive.
Results
Flowering
phenology
Figure
1
shows
the
fl
owering
phenologies.
In
1991
the
fl
owering
season
at
K
started
2
weeks
earlier
than
that
at
S.
The
fl
owering
season
terminated
at
ca.
September
10
in
both
communities.
Thus,
the
fl
owering
season
at
K
was
ca.
2
weeks
longer
than
at
S.
The
fl
owering
season
at
K
lasted
for
87
days.
Peak
fl
owering
for
five
species
occurred
before
midseason,
and
the
mean
peak
for
all
species
occurred
at
day
37.
At
S
the
1074
CAN.
J.
BOT.
VOL.
71,
1993
Flowering
frequency
Flowering
frequency
(A)
•=Pc
1.0
-
L=Fia
0=Ca
0.8
-
C=Tc
0.6
-
0=La
•=Pp
0.4
-
0.2
-
0-
0
(B)
1.0
-
0.8
-
0.6
-
0.4
-
0.2
-
0
-
Pc
Ra
Sa
Ca
Tc
Pp
15
30
45
60
75
Number
of
days
after
16
June
•=Pc
L =
Ra
0=Ca
❑=Tc
■=Sa
0=La
•=Pp
90
0
15
30
45
60
Number
of
days
after
1
July
(C)
75
La
1
10
20
30
40
50
60
70
Number
of
days
after
6
July
FIG.
1.
Flowering
phenology
at
(A)
Mount
Kvannjolnut
in
1991,
(B)
Mount
Sandalsnut
in
1991,
and
(C)
Mount
Sandalsnut
in
1990.
In
A
and
B,
y-axis
values
on
a
given
day
represent
the
number
of
fl
owers
on
that
day
divided
by
the
maximum
daily
number
of
fl
owers
for
each
species.
Species
names
are
abbreviated
as
in
Table
1.
TABLE
2.
Percentage
of
fl
ower
visitors
to
the
plant
species
at
Mount
Kvannjolnut
in
1991
(K
1
)
and
at
Mount
Sandalsnut
in
1990
(S
o
)
and
1991
(S
i
)
Ra
Tc
Ca
Pc
Pp
La
Gs
Sa
Total
no.
of
visits
observed
K
1
S
I
S
o
K
1
S
i
S
o
K
1
S
i
S
o
K
1
S
i
S
o
K
i
S
i
S
o
K
1
S
i
S
o
K
1
S
i
S
o
K
i
S
i
S
o
Anthomyiidae
Diptera
Heterostylodes
pilifera
Zett.
1
2
0
1
0 0
3 3
0
Pegoplata
aestiva
Meigen
29
6
8
10
1
11
16
4
15
22
4
0
7
1
0
17 17
6
8
146
25
21
Pegoplata
tundrica
Schnabl
1
2
0
3
0 0
2
1
0
4
0 0
12
5
0
Zaphne
barbiventris
Zett.
2
11
2
0 0
1
2
3
0
5
18
12
0
3
0
71
17
10
74
7
Zaphne
frontata
Zett.
1
1
2
0
13
23
0
4
12
2
1
4
2
23
59
8
23
44
Dolichopodidae
Dolichopus
plumipes
Scop.
0
1
0 0
1
0 0
6
3
0
2
0 0
6
4
1
1
13
2
Dolichopus
rupestris
Haliday
0
1
0 0
4
0 0
20
3
0
1
6
0
41
4
0
3
0
4
6
0
54
5
Empididae
Empis
lucida
Zett.
0
1
0 0
1
1
2
1
3
2
0 0
3 3
1
Rhamphomyia
morio
Zett.
0
1
0
3
8
1
5
22
15
0
1
0
5
14
4
1
6
0
4
11
11
39
11
Muscidae
Phaonia
alpicola
Zett.
8
4
5
19
4
0 0 0
3
11
4
0 0 0
4
2
0 0
36
14
5
Phaonia
lugubris
Meigen
3
1
0
10
4
0 0
3
0
2
0 0
6
0 0
21
7
0
Spilogona
alpica
Zett.
0
0
7
0
1
11
0
3
15
0 0
6
1
0
17
0
6 6
1
5
28
Spilogona
nitidicauda
Schnabl
0
1
3
0 0
2
0
8 8
0
6
0
10
8
Thricops
aculeipes
Zett.
8
2
0
16
0 0
8
0 0
8
0 0
4
0 0
12
0 0
73
189
3
0
Thricops
cunctans
Meigen
1
1
0
2
1
3
2 2
0
2
0 0
1
0 0
1
9
5
1
a
r-
Thricops
furcatus
Stein
0
1
7
0
7
16
0
1
3
5
18
38
1
11
16
0
11
6
25
34
Thricops
hirtulus
Zett.
6
63
65
10
46
27 27
30
26
19
63
71
4 4
17
6 6 6
1
20
44
60
248
96
z
>
izi
Thricops
longipes
Zett.
2
0
0
3
0 0
1
0 0
3
0 0
10
0 0
Thricops
nigritellus
Zett.
35
1
0
29
0 0
19
0 0
23
0 0
46
0 0
43
0 0
5
237
2
0
Scathophagidae
Scathophaga
stercoraria
L.
0
0
2
0
1 1
0
1
0
2
1
6
0
9
3
1
3
6
4
Syrphidae
Cheilosia
sahlbergi
Zett.
0
1
0
3
1
0
3 3
0
Eristalis
spp.°
0
4
1
0
11
3
0
7
2
Melanostoma
dubium
Zett.
0
1
0
3
0 0
5
3
0
4 4
0
Platycheirus
manicatus
Meigen
2
1
0
7
14
1
0
Platycheirus
subordinatus
Becker
3
0 0
5
0 0
3
0 0
1
0 0
13
0 0
Hymenoptera
Apidae
Bombus
alpinus
L.
0
9
0 0
5
0
Bombus
lapponicus
Fabr.
3
0 0
3
10
0 0
Lepidoptera
Lycaenidae
Albulina
orbitulus
de
Prunner
11
0 0
12
0 0
1
19
0 0
Total
no.
of
visits
to
plant
species
248
176
60
31
72
81
63
79
34
64
106
17
84
71
24
157
35
32
181
145
18
Niche
breadth
(bias
corrected)
4.6
2.4 2.4
6.9
3.9
5.8
7.0
4.6
7.3
6.7
2.3
1.9
4.1
4.5
4.7
4.1
8.5
2.6
1.8 1.8
3.7
SE
0.4
0.3
0.4
1.2
0.9
0.5
0.7
0.6
0.9
0.8
0.3
0.4
0.7
0.8
0.9
0.6
1.0
0.6
0.2
0.4
0.8
NOTE:
The
values
in
the
table
are
the
percentage
of
visits
by
the
insect
taxa
to
each
plant
species.
Percentage
values
do
not
always
add
up
to 100
because
of
rounding
errors.
Only
visitor
taxa
observed
on
five
or
more
occasions
are
included.
The
niche
breadth
values
(Levin's
index)
for
each
plant
species
are
bias
corrected
by
bootstrapping
(1000
resamples).
Plant species
names
are
abbreviated
as
in
Table
1.
"Includes
Eristalis
tenax
L.
and
Eristalis
pertinax
Scop.
Ui
to
1076
CAN.
J.
BOT.
VOL
71,
1993
1.0
-
(1)
0.8
Mean
overlap
0.6
-
0.4
0.2
-
0
1
7
K
1991
S
1991
S
1990
Ra
Tc
Ca
Pc
Pp
La
Sa
FIG.
2.
Mean
bias
corrected
niche
overlap
in
fl
ower
visitor
species
for
each
plant
species.
Bootstrapping
was
repeated
1000
times.
Vertical
bars
are
the
standard
errors
of
the
means.
Species
names
are
abbreviated
as
in
Table
1.
fl
owering
season
lasted
for
74
days.
Here,
four
species
peaked
fl
owering
before
midseason,
and
also
here,
the
mean
peak
for
all
species,
day
33,
occurred
before
midseason.
In
both
com-
munities,
all
species
had
started
to
flower
before
midseason.
The
two
species
first
in
bloom
at
K,
Ranunculus
acris
and
Potentilla
crantzii,
flowered
only
very
sparsely
during
the
first
2
weeks
when
temperatures
and
snowfall
were
low.
Visual
inspection
of
Fig.
lA
suggests
that
the
species
at
K
are
divided
into
two
groups
that
fl
owered
at
different
times
through
the
season;
no
such
separation
of
flowering
times
was
evident
at
S
in
1991
(Fig.
1B).
An
analysis
of
the
dispersion
of
flowering
peaks
in
the
two
communities
reveals
that
the
peaks
at
K
were
randomly
dispersed
along
the
time
axis
(expected
variance
/
sample
variance
=
1.01,
)(
2
=
7.07,
df
=
7,
P
=
0.42;
see
Poole
and
Rathcke
(1979)
and
Rabinowitz
et
al.
(1981),
whereas
the
flowering
peaks
at
S
were
significantly
clumped
(expected
variance
/
sample
variance
=
2.40,
)(
2
=
16.78,
df
=
7,
P
=
0.02).
Figure
1C
shows
the
flowering
phenology
at
S
in
1990.
The
initiation
and
sequence
of
fl
owering
were
about
equal
for
all
species
both
years.
At
S,
the
species
were
in
flower
for
extensive
periods
(Fig.
1).
Ranunculus
acris
and
Silene
acaulis
flowered
for
nearly
the
whole
season
in
both
years,
and
all
species
at
S
fl
owered
for
more
than
half
the
season.
At
K,
only
R.
acris
and
Parnassia
palustris
were
in
flower
for
more
than
half
the
season.
Flower
visitors
Table
2
summarizes
the
percentage
of
visits
by
insect
taxa
to
each
plant
species.
Dipterans
constituted
almost
the
entire
visitor
assemblage
at
both
sites
and
in
both
years.
Most
of
the
plants
were
visited
by
many
insect
species.
However,
only
a
few
taxa
were
common
on
each
plant
species.
At
K,
the
most
common
flower
visitor
was
Thricops
nigri-
tellus.
Flies
of
this
species
showed
preference
for
R.
acris
early
in
the
fl
owering
season
and
later
switched
to
Leontodon
autumnalis
and
Parnassia
palustris.
The
abundance
of
Syr-
phidae
was
low;
only
4.1%
of
all
visits
were
by
this
family.
Bombus
lapponicus
(10
visits)
was
the
only
bumblebee
recorded
at
K.
Towards
the
end
of
the
fl
owering
season,
workers
of
B.
lapponicus,
although
in
small
numbers,
were
very
active
on
Geranium
sylvaticum
and
L.
autumnalis.
The
only
butterfly
observed
was
Albulina
orbitulus,
which
only
occurred
at
K.
Niche
overlap
Niche
overlap
(visitor
speci
10
-
(A)
0
.
0.8
-
u)
0
0.6
-
0.4
-
0.2
-
0
Y=0.68X+0.
11
R
2
=0.50
P=0.0003
0
0
0
0.8
-
0
0.6
-
0.4
-
0.2
-
0
0.2
0
0.4
0
0
0.6
0.8
1.0
0
0
R
2
=0.00
P
=0.90
0
0.2
0.4
0.6
0.8
1.0
Niche
overlap
(flowering
time)
FIG.
3.
Regressions
between
niche
overlap
in
fl
owering
time
and
niche
overlap
in
visitor
species
for
the
plant
species
pairs
(0)
at
(A)
Mount
Sandalsnut
in
1991
and
(B)
Mount
Kvannjolnut
in
1991.
Most
visits
by
this
species
were
to
Parnassia
palustris
and
Cerastium
alpinum.
At
S
in
1991,
Thricops
hirtulus
was
the
most
frequent
visitor.
Bombus
alpinus
visited
L.
autumnalis
in
low
numbers.
Syrphids,
Eristalis
spp.,
were
more
frequent
on
this
species
compared
with
the
previous
year.
Some
of
the
species
(e.g.,
Thricops
aculeipes
and
Thricops
nigritellus)
differed
greatly
in
abun-
dance
between
the
two
sites
sampled
in
1991.
At
S
in
1990,
Thricops
hirtulus
was
more
than
twice
as
frequent
as
any
other
visitor.
It
was
the
most
frequent
fl
ower
visitor
on
all
species,
except
for
late
-flowering
L.
autumnalis
and
Parnassia
palustris.
The
second
most
frequent
visitor
was
Zaphne
frontata,
especially
on
L.
autumnalis.
Earlier
in
the
season
it
frequently
visited
Taraxacum
croceum.
Eristalis
spp.
were
very
active
on
flowers
late
in
the
season
but
in
low
numbers.
No
bumble
bees
were
recorded
as
visitors
in
1990.
Table
2
also
shows
the
visitor
niche
breadths
for
each
plant
species.
For
some
of
the
plants,
the
niche
breadth
showed
large
differences
both
between
areas
and
between
years.
In
1991,
populations
of
the
same
species
had
wider
breadth
at
K
than
at
S,
with
the
exception
of
Parnassia
palustris
and
L.
autumnalis.
From
1990
to
1991
at
S,
the
niche
breadth for
L.
autumnalis
increased
markedly,
whereas
it
decreased
sub-
Rif
LAND
1077
TABLE
3.
Visitation
rates
(VR)
to
the
species
at
Mount
Sandalsnut
(S)
in
1990
and
1991
and
at
Mount
Kvannjolnut
(K)
in
1991
S
1990
S
1991
K
1991
VR
SD
n
E
vis
E
fl
VR
SD
n
E
vis
E
fl
VR
SD
n
E
vis
E
fl
Cerastium
alpinum
0.33
0.69
60
42
1024
0.42
0.48
41
109 655
0.90
1.02
31
82
386
Leontodon
autumnalis
0.36
0.57
46
43
741
0.15
0.24
52
45
1040
2.82 2.82
27
192
280
Parnassia
palustris
0.51
0.78
32
31
372
0.57
0.69
33
86
388
0.87
0.96
31
98
386
Potentilla
crantzii
0.18
0.39
36
20
504
0.69
0.99
53
134
568
0.51
0.81
51
94
524
Ranunculus
acris
0.45
0.99
115
97
1436
0.72
0.90
77
232
1144
1.44
2.04
76
311
952
Taraxacum croceum
1.20
1.62
92
106
681
1.05
1.44
64
96
510
5.16
7.56
13
43
39
Silene
acaulis'
0.03
0.06
47
49
3695
Geranium
sylvaticum
1.83 1.53
34
207
375
NOTE:
VR
is
presented
as
the
mean
±SD
of
the
number
of
visits
per
flower
per
hour
in
all
periods.
The
number
of
visits
observed
during
sampling
periods
from
1990
at
S
and
from
1991
at
both
sites
was
multiplied
by
six
and
three,
respectively,
because
observation
periods
lasted
for
10
min
in
1990
and
for
20
min
in
1991.
n,
number
of
10-
or
20
-min
observation
periods;
E
vis,
total
number
of
visits
observed;
E
fl,
total
number
of
flowers
observed
in
all
periods.
"Flower
density
within
squares
not
quantified
for
S
1990.
stantially
for
Taraxacum
croceum
and
C.
alpinum.
Niche
overlaps
in
visitor
species
Figure
2
shows
the
plants'
mean
niche
overlaps
in
visitor
species.
In
general,
large
overlaps
were
found.
All
species
sampled
in
both
sites
in
1991
had
the
highest
mean
overlap
at
K.
At
S,
all
species
had
the
highest
mean
overlap
in
1990.
In
1991,
L.
autumnalis
had
the
highest
mean
overlap
of
all
species
at
K,
whereas
it
had
the
lowest
at
S.
The
correlation
between
overlap
for
species
pairs
at
S
in
1990
and
1991
was
positive
and
highly
significant
(df
=
19,
r
=
0.69,
t
=
4.16,
p
=
0.0005).
This
close
correlation
did
not
exist
between
the
species
pairs
sampled
at
S
and
K
in
1991
(df
=
13,
r
=
0.34,
t
=
1.30,
p
=
0.22).
Figure
3
shows
the
relationship
between
species
pair
overlap
in
fl
owering
time
and
visitor
species
at
the
two
sites
sampled
in
1991.
At
S,
overlap
in
fl
owering
time
was
a
good
predictor
of
overlap
in
visitor
species.
At
K,
overlap
in
fl
owering
time
did
not
explain
the
variation
in
overlap
in
visitor
species.
Here,
eight
species
pairs
with
a
low
overlap
in
fl
owering
time
had
a
high
overlap
in
visitor
species.
Visitation
rates
Table
3
shows
the
pollinator
visitation
rates
to
the
fl
owers
of
each
plants
species.
The
visitation
rate
to
some
of
the
spe-
cies,
especially
L.
autumnalis,
differed
greatly
both
between
years
at
S
and
between
sites
in
1991.
In
1991
the
visitation
rates
were
highest
at
K,
with
the
exception
of
that
to
Potentilla
crantzii.
The
most
attractive
species
at
both
sites
was
Taraxacum
croceum.
In
1990,
this
species
was
more
than
twice
as
attrac-
tive
as
any
of
the
other
species
at
S.
The
average
visitation
rate
for
all
species
was
0.51, 0.60,
and
1.95
visits
per
fl
ower
per
hour
at
S
in
1990,
S
in
1991,
and
K
in
1991,
respectively
(S.
acaulis
and
G.
sylvaticum
excluded
because
these
species
were
not
measured
in
both
years
and
at
both
sites).
Discussion
Flowering
phenology
The
synchronous
early
fl
owering
is
probably
best
explained
by
the
short
time
available
for
shoot
and
fl
ower
development
and
seed
maturation.
Early
fl
owering
is
well
known
from
other
studies
in
short
-season
environments.
Bliss
(1956,
1971)
and
Billings
and
Mooney
(1968),
in
their
reviews
of
plant
develop-
ment
in
arctic
and
alpine
environments,
stated
that
fl
owering
occurred
shortly
after
the
occurrence
of
favorable
temperatures
and
that
fl
owering
was
synchronous
among
species.
Holway
and
Ward
(1965)
showed
that
snow
cover
was
the
primary
factor
influencing
phenology
and
that
species
initiated
fl
ower-
ing
early
after
snowmelt
in
the
Rocky
Mountains.
Inouye
and
Pyke
(1988),
in
an
investigation
of
the
pollination
biology
in
alpine
Australia,
found
that
most
species
fl
owered
early
in
the
season.
Helenurm
and
Barrett
(1987),
working
with
boreal
forest
herbs
in
Canada,
also
found
fl
owering
to
be
more
prominent
early
in
the
season
and
argued
that
abiotic
factors
such
as
climate
and
season
length
may
have
influenced
the
fl
owering
phenology.
Kudo
(1991)
showed
that
rapid
-flowering
alpine
species
are
more
successful
seed
producers
in
Japan.
Finally,
Galen
and
Stanton
(1991)
showed
that
differential
snow
accumu-
lation
influenced
fitness
in
Ranunculus
adoneus
A.
Gray.
Plants
with
delayed
fl
owering,
because
of
a
thicker
snow
pack,
suffered
a
reduction
in
seed
set
and
seed
weight.
They
concluded
that
the
time
available
for
seed
development
was
too
short
for
these
plants.
Thus,
in
short
season
environments,
the
timing
of
repro-
ductive
events
is
probably
most
influenced
by
the
limited
time
available
for
seed
maturation,
which
selects
for
early
fl
owering
individuals.
I
found
(unpublished
data)
that
the
fl
ower
visiting
activity
at
S
in
1990
is
highest
early
in
the
fl
owering
season.
This
may
also
favor
early
fl
owering
because
of
the
increased
possibilities
for
pollination
at
this
time.
In
the
two
communities
examined,
fl
owering
peaks
were
either
clumped
(S)
or
randomly
distributed
(K).
Therefore,
on
the
basis
of
Poole
and
Rathcke's
(1979)
method,
there
are
no
indications
that
past
interspecific
competition
influenced
fl
ower-
ing
times
to
minimize
overlap.
Competitive
selection
pressure
was
probably
too
infrequent,
nondirectional,
or
weak
to
produce
character
displacement
in
fl
owering
times
because
of
the
prob-
able
high
degree
of
environmental
stocasity
in
these
areas
and
the
reliance
on
autogamous
pollination
or
asexual
reproduc-
tion.
However,
care
should
be
taken
when
interpreting
data
for
species
interactions
because
no
experimental
testing
or
measurements
on
fitness
were
done.
Furthermore,
it
must
be
pointed
out
that
fl
owering
curves
for
only
seven
species
were
analyzed.
These
species
represent
only
a
proportion
of
all
species
in
the
communities,
and
inclusion
of
more
species
into
the
analyses
could
give
different
results.
Rabinowitz
et
al.
(1981)
found
that
the
dispersion
of
fl
owering
peaks
of
wind-
and
insect
-pollinated
prairie
plants
did
not
deviate
from
a
random
assemblage.
These
authors
suggested
several
causal
scenarios
to
explain
the
lack
of
a
community
-wide
pattern
but
pointed
out
the
difficulties
of
explaining
properties
of
fl
owering
times
distribution
from
purely
observational
studies
like
this
one.
1078
CAN.
1.
BOT.
VOL.
71,
1993
Flower
visitors
The
dominance
of
fl
ies
in
the
visitor
assemblage
agrees
with
the
open
flower
morphology
of
the
plants.
Zygomorphic
flowers,
often
associated
with
bumblebees,
are
rare
in
the
Finse
area.
A
survey
of
Faegri's
(1967)
plant
list
from
Finse
reveals
that
only
10%
of
the
species
have
this
flower
form,
and
visits
by
Bombus
spp.
were
observed
to
only
18
of
119
flowering
species
growing
naturally
at
Finse
(wind
-pollinated
species
excluded).
No
other
bees
have
been
observed
in
the
Finse
area
(personal
observation).
The
dominance
of
Diptera
as
flower
visi-
tors
has
also
been
recognized
in
other
alpine
areas
(Australia:
Inouye
and
Pyke
1988;
New
Zealand:
Primack
1983;
White
Mountains,
United
States:
McCall
and
Primack
1992;
Andes:
Arroyo
et
al.
1982)
and
in
arctic
Canada
(Kevan
1972).
Com-
pared
with
these
areas,
the
Finse
area
is
different
in
its
complete
lack
of
bees
others
than Bombus
spp.,
in
the
very
low
abun-
dance
of
Bombus
spp.,
and
in
the
scarcity
of
syrphids.
Each
plant
species
was
visited
by
many
insect
species.
This
observation
corresponds
with
the
fl
oral
morphology
of
the
species
studied.
All
of
them
had
a
fl
ower
morphology
permit-
ting
visits
by
a
large
range
of
species.
A
generalist
fl
ower
form,
e.g.,
bowl
shaped,
is
interpreted
as
an
adaptation
to
unpredictable
pollination
conditions
(e.g.,
Rathcke
1988),
such
as
those
at
Finse.
The
potential
for
pollinator
-restrictive
fl
ower
morphologies,
e.g.,
zygomorphic
fl
owers,
would
seem
to
be
low
because
of
the
scarcity
of
specialized
pollinators.
Niche
overlaps
in
visitor
species
In
general,
the
fl
owering
species
shared
visitor
species
to
a
large
extent.
This
reflects
the
large
similarity
between
the
spe-
cies,
both
in
fl
ower
morphology
and
fl
owering
time.
At
S,
species
pairs
with
a
high
overlap
one
year
also
had
a
high
overlap
the
next
year,
indicating
that
the
same
species
may
be
pollinated
by
the
same
insects
in
consecutive
years.
Such
similarity
between
species
pairs
was
not
found
on
a
spatial
scale.
Species
pairs
having
a
high
overlap
in
visitor
species
at
one
site
in
1991
did
not
necessarily
have
a
high
overlap
at
the
other
site.
This
indicates
that
the
visitor
fauna
is
more
variable
in
space
than
in
time
and
that
a
population's
visitor
assemblage
is
deter-
mined
by
its
location
more
than
by
its
fl
ower
morphology.
At
K
in
1991,
overlap
in
fl
owering
time
was
not
a
good
predictor
of
overlap
in
visitor
species
because
of
sequential
sharing
of
visitors
in
eight
species
pairs.
Such
sharing
of
visitors
might
benefit
individuals
in
both
species
because
presence
of
one
species
results
in
an
increase
in
visitation
rate
to
the
other
spe-
cies
(Waser
and
Real
1979;
McGuire
and
Armbruster
1991).
The
positive
correlation
between
overlap
in
flowering
time
and
visitor
species
at
S
shows
that
species
fl
owering
simul-
taneously
are
visited
by
the
same
insect
species
available
in
that
period.
Visitation
rates
The
average
visitation
rate
values
for
all
species
at
the
two
sites
correspond
to
those
found
in
other
alpine
areas.
McCall
and
Primack
(1992)
found
a
visitation
rate
of
1.32
visits
per
flower
per
hour
in
alpine
North
America.
In
alpine
Australia,
Inouye
and
Pyke
(1988)
measured
a
visitation
rate
of
0.87
visits
per
fl
ower
per
hour.
Arroyo
et
al.
(1985)
measured
a
decrease
in
visitation
rate
with
increasing
altitude
in
Andes.
They found
visitation
rates
of
0.41,
0.25,
and
0.19
visits
per
fl
ower
per
hour
at
the
low-,
middle-,
and
high
-elevation
sites,
respectively.
Also
at
Finse,
a
decrease
in
visitation
rate
with
altitude
was
found.
The
large
difference
between
the
two
sites
is
a
result
of
a
lower
density
of
flower
visitors
at
S,
coupled
with
a
higher
flower
density
at
this
site.
The
differences
in
visitation
rates
at
S
between
years
is
most
easily
explained
by
differences
in
climatic
conditions
during
time
of
sampling
and
fluctuations
in
the
insects
population
densities
between
years.
Taraxacum
croceum
was
much
more
attractive
to
the
insects
than
the
other
species
studied,
probably
because
of
its
showy
fl
oral
display
and
the
large
quantities
of
nectar
and
pollen
it
produces
(personal
observation).
The
species
assemblage
studied
showed
no
diver-
gence
away
from
this
species,
either
in
time
of
fl
owering
or
in
visitor
species.
Community
studies
on
pollination
ecology
have
usually
been
carried
out
in
species
-rich
tropical
and
temperate
floras
and
often
on
species
with
highly
specialized
pollination
modes.
More
work
is
needed
in
communities
in
which
unspecialized
pollinators
dominate
and
in
which
the
influence
of
abiotic
fac-
tors
like
length
of
flowering
season
and
climate
is
likely
to
govern
community
-wide
patterns
of
flowering
times,
pollina-
tion
mode,
and
reproductive
success.
Acknowledgements
I
thank
E.
Ostbye
and
E.A.
Leslie
for
living
facilities
at
the
High
Mountain
Research
Station
at
Finse.
A.C.
Pont
and
K.
Rognes
(Muscidae),
V.
Michelsen
(Anthomyiidae),
T.
Nielsen
(Syrphidae),
A.
LOken
(Bombus),
L.G.
Jensen
(Empididae,
Dolichopodidae),
T.
Jonassen
(Empididae,
Dolichopodidae),
and
G.E.
Ellingsen
(Lepidoptera)
kindly
identified
the
insects.
C.C.
Berg,
H.J.B.
Birks,
K.
Faegri,
F.
Gilbert,
C.
Ihlebaek,
H.
Monsen,
I.
Nordal,
P.H.
Salvesen,
and
an
anonymous
reviewer
read
the
manuscript
and
gave
useful
comments.
H.J.B.
Birks
also
kindly
provided
statistical
programs
and
sug-
gestions
and
corrected
the
language.
This
work
was
financed
in
part
by
the
Faculty
of
Science,
University
of
Bergen,
Norway.
Arroyo,
M.T.K.,
Primack,
R.,
and
Armesto,
J.J.
1982.
Community
studies
in
pollination
ecology
in
the
high
temperate
Andes
of
central
Chile.
I.
Pollination
mechanisms
and
altitudinal
variation.
Am.
J.
Bot.
69:
82-97.
Arroyo,
M.T.K.,
Armesto,
J.J.,
and
Primack,
R.B.
1985.
Commu-
nity
studies
in
pollination
ecology
in
the
high
temperate
Andes
of
central
Chile.
II.
Effects
of
temperature
on
visitation
rates
and
pollination
possibilities.
Plant.
Syst.
Evol.
149:
187-203.
Billings,
W.D.,
and
Mooney,
H.A.
1968.
The
ecology
of
arctic
and
alpine
plants.
Biol.
Rev.
43:
481-529.
Bliss,
L.C.
1956.
A
comparison
of
plant
development
in
micro
-
environments
of
arctic
and
alpine
tundras.
Ecol.
Monogr.
26:
303-337.
Bliss,
L.C.
1971.
Arctic
and
alpine
plant
life
cycles.
Annu.
Rev.
Ecol.
Syst.
2:
405
—438.
Crawford,
R.M.M.
1989.
Studies
in
plant
survival.
Blackwell
Scien-
tific
Publications,
Oxford.
Fryxell,
P.A.
1957.
Mode
of
reproduction
in
higher
plants.
Bot.
Rev.
23:
135
—233.
Faegri,
K.
1967.
The
plant
world
at
Finse,
Norway.
Available
from
Botanical
Institute,
University
of
Bergen,
Bergen,
Norway.
Faegri,
K.,
and
van
der
Pijl,
L.
1979.
The
principles
of
pollination
ecology.
Pergamon
Press,
Oxford.
Galen,
C.,
and
Stanton,
M.L.
1991.
Consequences
of
emergence
phenology
for
reproductive
success
in
Ranunculus
adoneus
(Ranun-
culaceae).
Am.
J.
Bot.
78:
978-988.
Grace,
J.,
and
Nelson,
M.
1981.
Insects
and
their
pollen
loads
at
a
hybrid
Heracleum
site.
New
Phytol.
87:
413
—423.
Hagerup,
0.
1951.
Pollination
in
the
Faroes
in
spite
of
rain
and
poverty
in
insects.
Dan.
Biol.
Medd.
18:
1-48.
Helenurm,
K.,
and
Barrett,
S.C.H.
1987.
The
reproductive
biology
of
boreal
forest
herbs.
II.
Phenology
of
flowering
and
fruiting.
TafLAND
1079
Hocking,
B.
1968.
Insect
-flower
associations
in
the
High
Arctic
with
special
reference
to
nectar.
Oikos,
19:
359
-387.
Holway,
J.G.,
and
Ward,
R.T.
1965.
Phenology
of
alpine
plants
in
northern
Colorado.
Ecology,
46:
73
-83.
Horn,
H.S.
1966.
Measurements
of
"overlap"
in
comparative
eco-
logical
studies.
Am.
Nat.
100:
419-423.
Inouye,
D.W.,
and
Pyke,
G.H.
1988.
Pollination
biology
in
the
snowy
mountains
of
Australia:
comparisons
with
montane
Colorado,
USA.
Aust.
J.
Ecol.
13:
191-210.
Kendall,
D.A.,
and
Solomon,
M.E.
1973.
Quantities
of
pollen
on
the
bodies
of
insects
visiting
apple
blossoms.
J.
Appl.
Ecol.
10:
627
-634.
Kevan,
P.G.
1972.
Insect
pollination
of
high
arctic
fl
owers.
J.
Ecol.
60:
831
-847.
Kudo,
G.
1991.
Effects
of
snow
-free
period
on
the
phenology
of
alpine
plants
inhabiting
snow
patches.
Arct.
Alp.
Res.
23:
436-443.
Laverty,
T.M.
1992.
Plant
interactions
for
pollinator
visits:
a
test
of
the
magnet
species
effect.
Oecologia,
89:
502-508.
Laverty,
T.M.,
and
Plowright,
R.C.
1988.
Fruit
and
seed
set
in
Mayapple
(Podophyllum
peltatum):
influence
of
intraspecific
factors
and
local
enhancement
near
Pedicularis
canadensis.
Can.
J.
Bot.
66:
173
-178.
Levins,
R.
1968.
Evolution
in
changing
environments.
Princeton
University
Press,
Princeton,
N.J.
Lundqvist,
A.,
Osterby,
U.,
Larsen,
K.,
and
Linde-Laursen,
I.
1973.
Complex
self
incompatibility
systems
in
Ranunculus
acris
L.
and
Beta
vulgaris
L.
Hereditas,
74:
161-168.
McCall,
C.,
and
Primack,
R.B.
1992.
Influence
of
fl
ower
charac-
teristics,
weather,
time
of
day,
and
season
on
insect
visitation
rates
in
three
plant
communities.
Am.
J.
Bot.
79:
434
-442.
McGuire,
A.D.,
and
Armbruster,
W.S.
1991.
An
experimental
test
for
reproductive
interactions
between
two
sequentially
blooming
Saxifraga
species
(Saxifragaceae).
Am.
J.
Bot.
78:
214
-219.
Mosquin,
T.
1971.
Competition
for
pollinators
as
a
stimulus
for
the
evolution
of
fl
owering
time.
Oikos,
22:
398-402.
Mueller,
L.D.,
and
Altenberg,
L.
1985.
Statistical
inference
on
measures
of
niche
overlap.
Ecology,
66:
1204
-1210.
Pleasants,
J.M.
1980.
Competition
for
bumblebee
pollinators
in
Rocky
Mountain
plant
communities.
Ecology,
61:
1446
-1459.
Pleasants,
J.M.
1983.
Structure
of
plant
and
pollinator
communities.
In
Handbook
of
experimental
pollination
biology.
Edited
by
C.E.
Jones
and
R.J.
Little.
Van
Nostrand
Reinhold,
New
York.
pp.
375
-393.
Poole,
R.W.,
and
Rathcke,
B.J.
1979.
Regularity,
randomness,
and
aggregation
in
fl
owering
phenologies.
Science
(Washington,
D.C.),
203:
470-471.
Primack,
R.B.
1983.
Insect
pollination
in
the
New
Zealand
mountain
fl
ora.
N.Z.
J.
Bot.
21:
317
-333.
Primack,
R.B.
1987.
Relationships
among
fl
owers,
fruits,
and
seeds.
Annu.
Rev.
Ecol.
Syst.
18:
409-430.
Rabinowitz,
D.,
Rapp,
J.K.
,
Sork,
V.L.,
et
al.
1981.
Phenological
properties
of
wind-
and
insect
-pollinated
prairie
plants.
Ecology,
62:
49
-56.
Rathcke,
B.J.
1983.
Competition
and
facilitation
among
plants
for
pollination.
In
Pollination
biology.
Edited
by
L.A.
Real.
Academic
Press,
New
York.
pp.
205
-329.
Rathcke,
B.J.
1984.
Pattern
of
fl
owering
phenologies:
testability
and
casual
inference
using
a
random
model.
In
Ecological
commu-
nities:
conceptual
issues
and
the
evidence.
Edited
by
D.R.
Strong,
L.G.
Simberloff,
L.G.
Abele,
and
A.B.
Thistle.
Princeton
Univer-
sity
Press,
Princeton,
N.J.
pp.
383
-393.
Rathcke,
B.J.
1988.
Flowering
phenology
in
a
shrub
community:
competition
and
constraints.
J.
Ecol.
76:
975
-994.
Rathcke,
B.J.,
and
Lacey,
E.P.
1985.
Phenological
patterns
of
terres-
trial
plants.
Annu.
Rev.
Ecol.
Syst.
16:
179-214.
Richards,
A.J.
1972.
The
Taraxacum
of
the
British
Isles.
Watsonia,
9(Suppl.):
1-141.
Schluter,
D.
1988.
The
evolution
of
finch
communities
on
islands
and
continents:
Kenya
vs.
Galapagos.
Ecol.
Monogr.
58:
229
-249.
Shykoff,
J.A.
1992.
Sex
polymorphism
in
Silene
acaulis
(Caryo-
phyllaceae)
and
the
possible
role
in
maintaining
females.
Am.
J.
Bot.
79:
138
-143.
Smith,
E.P.,
and
Zaret,
T.M.
1982.
Bias
in
estimating
niche
overlap.
Ecology,
63:
1248
-1253.
Thomson,
J.D.
1982.
Patterns
of
visitation
by
animal
pollinators.
Oikos,
39:
241
-250.
Thomson,
J.D.
1983.
Component
analysis
of
community
-level
inter-
actions
in
pollination
systems.
In
Handbook
of
experimental
polli-
nation
biology.
Edited
by
C.E.
Jones
and
R.J.
Little.
Van
Nostrand
Reinhold,
New
York.
pp.
451-460.
Waser,
N.M.
1983.
Competition
for
pollination
and
fl
oral
character
differences
among
sympatric
plant
species:
a
review
of
evidence.
In
Handbook
of
experimental
pollination
biology.
Edited
by
C.E.
Jones
and
R.J.
Little.
Van
Nostrand
Reinhold,
New
York.
pp.
277
-293.
Waser,
N.M.,
and
Real,
L.A.
1979.
Effective
mutualism
between
sequential
fl
owering
plant
species.
Nature
(London),
281:
670-672.