(a)
(b)
(c)
Figure 2 The percentages of species in Australia that are not
threatened (all), endangered or extinct for (a) habit, (b) sex system
and (c) fruit type. Percentages add to 100 for each extinction risk.
0
10
20
30
40
50
Herb
Vine
Liane
Shr
ub
Tree
Intro
Native
0
20
40
60
80
100
Short-lived
Long-lived
0
20
40
60
80
100
Bise
xual
Dioecious
Monoecious
0
20
40
60
80
Dry-indeh
Dry-deh
Fleshy
% of species
% of species
% of species
% of species
(a)
(b)
(c)
(d)
Figure 1 The percentages of species in Australia that are native or
introduced for (a) habit, (b) life span, (c) sex system and (d) fruit
type. Percentages add to 100 for total species.
Life-history characters in the Australian flora
Journal of Biogeography 33, 271–290
283
ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd
flora. In each of three trees (Asterids 550–577 genera; Core
Eudicots 454–529 genera; Commelinids 534 genera), character
mapping showed that the non-threatened status is the ancestral
condition and that the conditions of endangered or extinct are
derived characters. The analysis of character-state change in
the three trees, compared with the randomized trees, showed
that, for each tree, threat status (i.e., endangered or extinct)
was significantly non-randomly clustered (Table 4); in other
words, threat status occurs in some genera more often than
expected by chance. To explore this further a concentrated
changes test was used to look at the randomness of
co-occurrence of traits (correlated evolution) such as threat
status and genus size and threat status and sex system.
Correlated evolution was found in the three trees and within
them in several clades (Table 5). Threat status was significantly
phylogenetically concentrated with genus size in some of the
resolved polytomies but not conclusively so (Table 5). Threat
status was significantly phylogenetically concentrated with
unisexual sex systems (monoecy plus dioecy) in the Asteraceae,
Thymelaeaceae, and within a clade comprising the Cyperaceae,
Eriocaulaceae and Hydatellaceae.
D I S C U S S I O N
Australia is renowned for its unusual flora (e.g. Banksia,
Grevillea, Eucalyptus), but less is known about whether this
unusual biodiversity is concomitant with unusual life-history
strategies. The evaluation of the life-history characters in the
Australian flora shows that there are some interesting patterns.
For example, monoecy is the predominant unisexual system
(Table 6) – unlike the general trends found on islands (e.g. New
Zealand and Hawaii, Table 6) and in worldwide assessments
(Table 6). Renner & Ricklefs (1995) point out that non-random
distributions of sex systems in floras may reflect phylogenetic
influences rather than local selective factors. We found support
for this, as the floras of South Africa and India, which have many
botanical affinities with Australia, also have higher levels of
0
10
20
30
40
50
60
70
80
Ex
Herb
Vine
Liane
Shrub
Tree
Herb
Vine
Liane
Shrub
Tree
En All Ex En All Ex En All Ex En All Ex En All
Ex En All Ex En All Ex En All Ex En All Ex En All
Bisexual
Monoecious
Dioecious
0
10
20
30
40
50
60
70
80
Fleshy
Dry-indeh
Dry-deh
(b)
(a)
% of species
% of species
Figure 3 The percentages of species in Australia that are not
threatened (all), endangered or extinct for habit with different (a)
sex systems and (b) fruit types. Percentages add to 100 for each
extinction risk.
Table 2 Minimum adequate model describing the probability of
occurrence of character traits in endangered species. The intercept
refers to the baseline category of species with a shrub habit, flowers
bisexual and fruit dry-dehiscent. Sign, coefficient and significance
of the effects are all relative to levels included in the intercept
Step
Coefficient
estimate
SD
Wald
P
(Intercept)
)3.33
0.09
36.906
< 0.001***
Herbs
0.26
0.12
2.112
0.035*
Trees
)0.27
0.20
)1.347
0.178
Monoecious
)1.15
0.31
)3.680
0.002***
Dioecious
)0.18
0.37
)0.484
0.628
Fleshy
)0.65
0.26
)2.482
0.013*
Dry-indehiscent
)0.16
0.18
)0.914
0.361
Herbs : monoecious
0.58
0.47
1.235
0.217
Trees : monoecious
1.43
0.41
3.514
< 0.001***
Herbs : dioecious
)0.48
0.63
)0.761
0.447
Trees : dioecious
)0.65
0.71
)0.916
0.360
Herbs : fleshy
)0.77
0.76
)1.008
0.313
Trees : fleshy
0.44
0.39
1.140
0.254
Herbs : dry-indehiscent
)1.38
0.30
)4.681
< 0.001***
Trees : dry-indehiscent
0.02
0.63
0.031
0.976
Deviance explained (%)
84
Values are in the transformed (logit) scale.
Table 3 The analysis of deviance table for the minimum adequate
model for endangered species
Term
d.f.
Deviance
Residual
d.f.
Residual
deviance
P
(> |v|)
Null
41
143.036
–
–
–
Habit
2
0.383
39
142.653
0.826
Sex
2
26.874
37
115.779
< 0.001
Fruit
2
48.444
35
67.336
< 0.001
Habitat : sex
4
16.667
31
50.669
0.002
Habitat : fruit
4
27.354
27
23.315
< 0.001
A. Sjo¨stro¨m and C. L. Gross
284
Journal of Biogeography 33, 271–290
ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd
monoecy than of dioecy (Table 6). Gross (2005) identified high
levels of monoecy in the tree flora of tropical Australia and
hypothesized that inefficient insect pollination systems may be
instrumental in maintaining monoecy.
Hegde & Ellstrand (1999) point out that generalizations
about rare and endangered species are important for the
development
of conservation
management
policy,
and
moreover for understanding the nature of rarity. Our results,
the first to incorporate all of the Australian flowering plants are
complex – they show that the condition of endangered and
extinction in Australian flowering species can be associated
with a combination of life-history characters that can be
correlated with phylogeny. We explore here the evolutionary
and ecological significance of these findings; we are left with
more questions than answers but hope that we generate
interest to test our hypotheses.
The association of habit with extinction
Trees are absent from the Australian extinct list. Recently
presumed extinct taxa in some locations outside Australia also
show patterns among their life-history strategies. For example, a
bias in the habit of species that have become extinct is evident in
an isolated fragment of lowland rainforest in the Singapore
Botanic Gardens (Turner et al., 1996). In over a century of
isolation, proportionately fewer trees than other life forms have
become extinct there. In another study, Robinson et al. (1994)
investigated plant species losses and invasions during a period of
112 years (1879–1991) on Staten Island, New York. They found
that trees exhibit higher levels of persistence than shrubs or vines,
and woody plants in general appear to be less vulnerable to
extirpation than herbaceous ones. From Robinson et al. (1994)
and Turner et al. (1996), it appears that trees tend to persist in a
variety of locations around the world, and in the face of possibly
quite different threatening processes. This tendency, however,
must be interpreted within the context of the relative longevity of
Table 4 Results of the test for trait clustering (rare or not), where
the observed number of steps of character-state change (trait
gains) in each tree was compared with the random distribution of
characters from 1000 tree resolutions
Trait
examined
Observed
number
of steps
Range of
steps
in 1000
randomizations
Rank of
observed
relative to
randomized
Rarity – Asterid clade
21
46–50
1
Rarity – Core Eudicot clade 19
56–63
1
Rarity – Commelinid clade
16
42–47
1
Figure 4 The proportion of (a) endangered species (odds
ratio
¼ 1.131 with 95% confidence interval 1.099–1.163, t ¼ 4.37,
P < 0.001; G
2
¼ 19.17, d.f. ¼ 1, P < 0.001) or (b) extinct species
(t
¼
)0.18; G
2
¼ 0.61, d.f. ¼ 1, P ¼ 0.43) within a genus versus
log(genus size) for all of the Australian genera (n
¼ 1978), and
the proportion of endangered species or extinct species within a
genus versus log(genus size) for only those genera with (c)
endangered species (odds ratio
¼ 0.619 with 95% confidence
interval 0.596–0.643, t
¼
)7.17; G
2
¼ 169.53, d.f. ¼ 1, P < 0.001)
(n
¼ 179) or (d) extinct species (odds ratio ¼ 0.365 with 95%
confidence interval 0.318–0.420, t
¼
)7.24; G
2
¼
)61.54,
d.f.
¼ 1, P < 0.001) (n ¼ 26). t: Wald statistic. The outside lines
are the 95% confidence intervals for the fitted curve (central line).
Life-history characters in the Australian flora
Journal of Biogeography 33, 271–290
285
ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd
Table 5 Examples of correlated evolution in two traits in clades from the angiosperm supertree
Clade
Trait comparison
Number of taxa
P values
Asterids
Rarity and genus size
550
0.00–0.00
Core Eudicots
Rarity and genus size
454
0.00–0.006
Commelinids (monocots)
Rarity and genus size
534
0.00–0.147
Asterales
Rarity and genus size
168
0.00–0.514
Apiales
Rarity and genus size
44
0.00–0.551
Solanales
Rarity and genus size
13
0.00–1.00
Lamiales
Rarity and genus size
127
0.00–0.359
Asteraceae
Rarity and genus size
140
0.00–0.508
Apiaceae
Rarity and genus size
27
0.125–0.154
Asterids
Rarity and sex system
577
0.00–0.00
Core Eudicots
Rarity and sex system
529
0.00–0.00
Commelinids (monocots)
Rarity and sex system
534
0.00–0.00
Asterales
Rarity and sex system
162
0.001–0.006
Solanales
Rarity and sex system
14
0.63–0.09
Asteraceae
Rarity and sex system
145
0.001–0.010
Apiaceae
Rarity and sex system
27
0.125–0.191
Thymelaeaceae
Rarity and sex system
10
0.00–0.00
Sapindaceae
Rarity and sex system
45
0.486–0.672
Sapindaceae
Rarity and sex system (bisexual or not)
45
0.00–0.563
Euphorbiaceae
Rarity and sex system
45
Unisexual or not
0.574–0.672
Bisexual or not
0.486–0.671
Cyperaceae, Eriocaulaceae and Hydatellaceae
Rarity and sex system (unisexual or not)
50
0.00–0.02
Asterids
Rarity and fruit type
564
Fleshy or not
0.00–0.00
Dry-indehiscent or not
0.00–0.00
Dry-dehiscent or not
0.00–0.00
Core Eudicots
Rarity and fruit type
467
Fleshy or not
0.00–0.00
Dry-indehiscent or not
0.00–0.00
Dry-dehiscent or not
>0.05
Commelinids (monocots)
Rarity and fruit type
534
Fleshy or not
0.00–0.00
Dry-indehiscent or not
0.00–0.00
Dry-dehiscent or not
0.00–0.00
Liliaceae
Rarity and fruit type (fleshy or not)
35
0.062–0.084
Zingiberaceae
Rarity and fruit type (fleshy or not)
9
0.559–0.704
Triuridaceae, Stemonaceae and Pandanaceae
Rarity and fruit type (fleshy or not)
4
0.00–0.00
Cyperaceae, Eriocaulaceae and Hydatellaceae
Rarity and fruit type
44
Dry-indehiscent or not
0.98–0.993
Dry-dehiscent or not
0.007–0.014
Asterids
Rarity and fruit type
564
Fleshy or not
0.00–0.00
Dry-indehiscent or not
0.00–0.00
Dry-dehiscent or not
0.00–0.00
Apocynaceae
Rarity and fruit type (fleshy or not)
31
0.007–0.025
Goodeniaceae
Rarity and fruit type (dry-dehiscent or not)
13
0.045–0.087
Euphorbiaceae
Rarity and fruit type (fleshy or not)
44
0.00–0.00
Myrtaceae
Rarity and fruit type
76
Fleshy or not
0.00–0.00
Dry-indehiscent or not
0.00–0.001
Rutaceae
Rarity and fruit type (fleshy or not)
44
0.00–0.00
Sapindaceae
Rarity and fruit type (dry-indehiscent or not)
31
0.013–0.020
P values are based on 10 random trees with resolved polytomies. No other clades gave significant P values.
A. Sjo¨stro¨m and C. L. Gross
286
Journal of Biogeography 33, 271–290
ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd
Table
6
The
distribution
of
life-history
characters
(%)
in
20
studies
Study
Bio.
region
Number
of
species
Habit
Life
span
Sex
Fruit
Woody
N
on-wdy
Sht
L
ng
Mix
H
erm
Dio
Mono
M
ix
Flshy
Dry
Tr
Sh
Li
He
Vi
Dry-indeh
Dry-deh
This
study
A
ustralia
16,824
(total),
16,704
(habit),
16,702
(lfspan),
16,655
(sex),
16,802
(fruit)
17.3
40.8
2.1
38.5
1.2
10.1
89.9
76.3
5.5
18.1
13.0
28.0
59.0
Parsons
(1958)
S.
Australia
2102
89.0
4.2
5.9
1.0
McComb
(1966)
W.
Australia
3886
90.0
4.4
2.5
3.0
Godley
(1979)
New
Z
ealand
1813
76.0
15.0
9.0
Rogers
&
W
alker
(2002)
New
Z
ealand
2715
23.6
2.1
65.6
Steiner
(1988)
South
Africa
565
(habit,
fruit),
8497
(sex)
37.3
62.7
79.7
6.7
13.0
68.3
31.5
Flores
&
Schemske
(1984)
Caribbean
2036
(fruit),
2037
(habit,
sex)
22.7
15.7
10.7
50.9
78.9
6.2
14.9
35.3
64.7
McComb
(1966)
British
Isles
1594
80.7
3.1*
8.7
7.5
Hegde
and
Ellstrand
(1999)
British
Isles
860
12.4
87.6
30.9
69.1
6.2
4
4.8
49.1
Khedr
et
al.
(2002)
à
Egypt
2340
(habit),
2376
(lfspan)
1.4
18.5
0.8
79.3
47.8
47.4
4.8
Fox
(1985)
Alaska
1471
8.7
91.3
4.3
9
5.7
Fox
(1985)
California
5421
14.2
85.8
6.6
9
3.4
Hegde
&
Ellstrand
(1999)
California
718
17.7
82.3
35.1
64.9
1.9
4
2.8
55.4
Conn
et
al.
(1980)
N
+
S
C
arolina
3274§
5.3
8.6
83.0
3.1
20.6
79.4
Kaye
et
al.
(1997)
Oregon
338
–
0.3
7.1
92.6
16.9
83.1
Carlquist
(1966)
Hawaii
1484
55.4
30.7
13.9
Sakai
et
al.
(1995)
Hawaii
844
(habit),
966
(fruit)
28.9
40.8
30.1
44.1
55.9
Sakai
et
al.
(2002)
Hawaii**
1159
(habit),
1045
(lfspan),
1040
(sex)
25.5
42.4
1.5
25.4
4.9
2.1
97.9
62.9
21.0
16.1
42.4
57.7
Robinson
et
al.
(1994)
Statten
Island-current
639
2.7
5.0
59.3
10.3
56.7
17.1
dng
dng
Robinson
et
al.
(1994)
Statten
Island-historic
1082
8.7
8.9
79.8
15.5
81.8
26.5
dng
dng
Turner
et
al.
(1996)
Singapore-current
314
38.5
8.9
20.1
3.8
Turner
et
al.
(1996)
Singapore-historic
448
63.8
8.5
23.2
3.1
Roy
(1974)
India
13,988
79.4
6.7
13.7
Yampolsky
&
Yampolsky
(1922)
World
121,492
71.4
3.5
14.0
10.5
Tr:
tree,
Sh:
shrub,
Li:
liane,
He:
herb,
Vi:
vine,
Non-wdy:
non-woody;
lfspan:
life
span,
Sht:
short
duration,
Lng:
long
duration;
Herm:
hermaphrodi
te,
Dio:
dioecious,
Mono:
monoecious;
Flshy:
fleshy
fruits,
Dry-indeh:
dry-indehiscent
fruits,
Dry-deh:
dry-dehiscent
fruits;
dng:
data
not
given.
*Kay
&
Stevens
(1986)
report
dioecy
levels
of
4%
(n
¼
59).
Data
also
from
S.
Hegde,
pers.
comm.
à
Includes
introduced
species;
parasites
(n
¼
36)
omitted
from
habit.
§Includes
gymnosperms
and
introduced
species.
–
Threatened
species
only.
**Endangered
species
only.
Life
span
and
sex
are
endangered
endemics
only.
Life-history characters in the Australian flora
Journal of Biogeography 33, 271–290
287
ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd
trees and the observed low recruitment levels of trees in altered
landscapes. There are, however, many tree species on the
endangered list so complacency about this life form is unwar-
ranted.
The association of sex system, fruit type and habit
with rarity
For the first time we have an overall picture of what the
Dostları ilə paylaş: |