全 文 ：Received 6 Jun. 2003 Accepted 30 Aug. 2003
Supported by the Knowledge Innovation Program of The Chinese Academy of Sciences (KSCX-2-1-06B), Scib Directors Fund of South China
Institute of Botany, The Chinese Academy of Sciences, Special Fund for Biological Systematics and Floristics, Fund for Scholars and Students
Returned from Overseas and a grant from the Bureau of Environmental Protection of Guangdong Province (970165).
* Author for correspondence. E-mail: .
http://www.chineseplantscience.com
Pollination Ecology of Caesalpinia crista (Leguminosae: Caesalpinioideae)
LI Shi-Jin1, ZHANG Dian-Xiang1* , LI Lin2, CHEN Zhong-Yi1
(1. South China Institute of Botany, The Chinese Academy of Sciences, Guangzhou 510650, China;
2. No.1 Middle School of Guangzhou Railway Company, Guangzhou 510600, China)
Abstract: The flowering phenology, pollination ecology and breeding system of Caesalpinia crista L.
were studied in Dinghushan Nature Reserve, Guangdong, China. The species started blooming in February
or March, then last till late April. It took about one week from first flower appearance to its full blooming,
which lasted for 2-4 d. The pollen-ovule ratio was 18 000±500. The breeding system was self-incompatible,
and protogynous xenogamy. Hymenoptera constituted the major group of pollinators. The pollination type
is ambophily, the species could be pollinated by wind if the pollinators were unavailable: this is the first
record of ambophily in the genus Caesalpinia. The floral structure adaptation to the pollinating behavior of
carpenter bees was described. The influences of artificial treatments in pollination biological studies on
the flowering and fruiting of the plants were also discussed.
Key words: pollination; ambophily; breeding system; flowering phenology; Caesalpinioideae; Caesalpinia crista
Plant reproductive ecology is a vital and highly produc-
tive field within modern biology (Lovett Doust and Lovett
Doust, 1988; Richards, 1997; Geber et al., 1999). Caesalpinia
is a pantropical genus with 120-150 species of trees, shrubs,
and lianas, but the study of pollination ecology of the ge-
nus is so far limited to the New World species (Lewis, 1998).
Only a few species of Caesalpinia have been studied with
regard to their pollination biology: C. pulcherrima (L.) Sw.
(Ali, 1932; Cruden and Heaman-Parker, 1979; Bullock, 1985);
C. (Erythrostemon) gilliesii (Wall. ex Hook.) Dietr. (Coccuci
et al., 1992); C. calycina Benth. and C. pluviosa DC. var.
sanfranciscana G. P. Lewis (Lewis and Gibbs, 1999). Vogel
(1990) reported that four flower types were found in
Caesalpinia, and each of which adapted to a different pol-
linator (bee, butterfly, moth and hummingbird), suggesting
that the floral character suits in the segregated genera of
Caesalpinia may be closely related to the type of pollinators.
Jones and Buchmann (1974) stated that Caesalpinia relied
heavily on Centris species for pollination.
Twenty species of Caesalpinia have been recorded from
China (Chen, 1988; Li et al., 2001). C. crista is a species
widely distributed from India to Polynesia. In China, it is
found in Sichuan and Hubei provinces in the north and
covers all the provinces south of the Yangtzi River (Larsen
et al., 1984; Chen, 1988).
1 Materials and Methods
Caesalpinia crista L. is a liana climbing to 10 m, with
glossy branchlets more or less armed with recurved prickles.
Flower buds are glabrous. The lowest calyx lobe brown,
cucullate, covers the flower bud before blossom. The other
four calyx lobes are yellow. Petals are yellow, standard blade
limb suborbicular and with a red spot (Fig.1). The other four
petals have oblong-elliptic blades. Stamens 10, filaments
unequal, ca. 12 mm long, are woolly from the base to the
middle. The filaments form a pseudotube at the base and
remain only two splits (Fig.2). The splits are covered by the
standard’s claw. Style ca. 10 mm long; stigma truncated,
wider than the style, with a mucilage cavity at the top and
ciliate around the rim (Figs.7, 8).
Data were obtained throughout three flowering seasons
(2000-2002), from March to April, in three separate natural
populations. The site is located in Dinghushan Nature Re-
serve (Guangdong, China) at approximately 23o10 N, 112o34
E. The first population is located at an altitude of 20 m.a.s.l,
with about 10 individuals along a small stream in front of the
Reserve’s headquarters building. The other two populations
are both situated near the Baodingyuan Sight at an eleva-
tion of 70 m.a.s.l, with 12 and three individuals respectively.
The last two populations are distant of 500 m and are sepa-
rated by a forest of Pinus massoniana.
1.1 Floral phenology
To determine the flowering period, flower lifespan and
anther dehiscence, two individuals were randomly chosen
from each population and two large panicles were taken
from each plant. Stigmatic receptivity was determined
by hand-pollination in 2000 and 2001. Stigma viability was
determined by the changes in color and moisture of the stigma
Acta Botanica Sinica
植 物 学 报 2004, 46 (3): 271-278
Acta Botanica Sinica 植物学报 Vol.46 No.3 2004272
in 2000 and 2001 and tested by using the MTT solution
(Rodriguez-Riano and Dafni, 2000) in 2002. Pollen viability
was also tested by the method of MTT. The volume and
sucrose concentration of the floral and extrafloral nectar
were measured by microcaps (Sigma Co.) and a hand re-
fractometer (Atago) respectively. Floral nectar was collected
from full blooming flowers that were bagged with nylon
nets prior to their anthesis, and the extrafloral nectar from
one rachis at nine o’clock in fine weather.
1.2 Breeding system and pollination success
Emasculation was performed one day before the flow-
ers open. Hand-pollination treatments were carried out twice
a day (morning and afternoon). Fruit set was calculated as
number of young fruits (two weeks after flowering) to num-
ber of pollinated flowers. Counting out immature rather than
mature pods was a precaution against herbivora, as insects
were eating young pods at different stages. Several treat-
ments were given to test the breeding system (Table 4).
The nets and bags were made of fine nylon fabric and thick
waterproof paper respectively. Bags were used to prevent
pollination by insects and air-disseminated pollens, whereas
nets served to stop only insects.
The pollen-ovule ratio (P/O rate) was estimated follow-
ing the method of Cruden (1976) and Dafni (1992). Ten flow-
ers in each population were chosen at random, and the
number of pollen grains and ovules in each flower were
calculated and finally an average and standard deviation
was estimated.
1.3 Visitors
Flower visitors were observed in the field over the peak
of flowering period between 07:00 am and 19:00 pm. The
voucher specimens were collected and taken to the labora-
tory for identification. To determine if the main visitors are
pollinators, the pollen on the stigmas was recorded after
being first visited.
1.4 The flower-pollinator interaction test
Virgin flowers were bagged prior to anthesis and the
pollen grain numbers on the stigma were counted after the
first visit. In order to test the nectar volume consumed by
visitors in each visit, the nectar volume in unvisited and
once visited flowers were recorded and compared.
To determine the function and importance of each floral
organ (standard, anthers and other petals) in attracting the
pollinators, 600 flowers on three large panicles were chosen.
The three manipulations are: standard removed only, an-
thers removed only and all the petals removed. The visiting
frequency of the two principal pollinators on these ma-
nipulated flowers were recorded.
1.5 The influences of the treatments to flowering
Since bagging and netting may influence the function
of flowering and fruiting, their impact on the flowering of C.
crista was tested by control experiment. The influence of
flowers bagging and netting to the lasting time of the stigma
receptivity, the petals longevity, the extra-floral nectar vol-
ume on the rachis, and the nectar volume in the flowers
were all recorded and compared with untreated flowers.
2 Results
2.1 Floral phenology
C. crista started blooming from March to July across its
distribution range. The flowering period started from Feb-
ruary to March at Dinghushan. It usually takes about 10 d
for an inflorescence with fully developed buds to open all
its flowers, although it may take two to three weeks for one
inflorescence to develop its flower buds. The lifespan of a
single flower (from petals opening to stamens and petals
lost) was 4-5 d. More than five months were needed for
fruit to mature (Table 1).
2.2 Sexual and breeding system
The flowers of C. crista are bisexual and protogynous.
The protogyny in the species was revealed by the fact that
the stigma becomes viable and receptive, and stretched
out before corolla opened and anther dehisced (Fig.3).
Hand-pollinated one day before opening, the flowers could
be fertilized and produce fruits, which indicated the occur-
rence of protogyny in the species (Table 2). The flowers
were slightly but sweetly fragrant. There were about 200
flowers for one panicle and more than 2 000 flowers for one
plant. The petals were yellow. The style was 15 mm long
and the stigma light creamy green, ciliated around the rim
and chambered (Figs.7, 8), turning brownish in color soon
after being pollinated. One flower displayed 10 stamens.
Table 1 Floral phenology of Caesalpinia crista at Dinghushan Mountain, Guangdong, China
Year
Elevation Inflorescence First flower Full Flowering termination Fruit
(m.a.s.l.) appearing appearing blooming and first fruit appearing maturing
2000 20 10 Mar. 22 Mar. 27-29 Mar. 10-13 Apr. 5-10 Oct.
70 13 Mar. 28 Mar. 1-3 Apr. 15-18 Apr. 5-10 Oct.
2001 20 25 Feb. 6 Mar. 13-15 Mar. 1-4 Apr. 25-30 Sept.
70 27 Feb. 9 Mar. 15-17 Mar. 2-5 Apr. 25-30 Sept.
2002 20 10-11 Feb. 25-26 Feb. 3-7 Mar.
70 11-13 Feb. 27-28 Feb. 5-10 Mar.
273LI Shi-Jin et al.: Pollination Ecology of Caesalpinia crista (Leguminosae: Caesalpinioideae)
Figs.1-8. Flower and pollinators of Caesalpinia crista. 1. Flowers in blooming. 2. Base of the androecium, showing the two splits. 3.
Flowers before the corolla open, showing the stretched stigmas. 4. Xylocopa sinensis sucking nectar in the flower. 5. Apis cerana sucking
extraflora nectar along the rachis. 6. Episyrphus balteata eating pollen on the anthers. 7. A stigma, showing the cilia. 8. Stigma, showing
the cavity and the pollen in the cavity.
Acta Botanica Sinica 植物学报 Vol.46 No.3 2004274
The immature anthers exhibited approximately the same color
as the corolla, and became ochre after dehiscence. The pol-
len remained viable for 2-3 d (Table 3). Nectar was pro-
duced within the flowers at near the base of style and
filaments, and on the flower rachis by the extra-floral
nectariferous glands upon anthesis. Honeybees (Apis
cerana) have been observed sucking the nectar along the
rachis even after the flowers had withered (Fig.5). The
floral disc was stained red in neutral red solution, indicat-
ing that odor glands, which attracted pollinators were lo-
cated in the disc. Up to (1.2 ± 0.7) mL (range from 0.55 mL to
2.8 mL; N = 20) nectar with 29%± 5% (range from 17% to 35%;
N = 20) of sucrose were recorded per full blooming flower
from the natural populations at 09:00 am in fine weather.
Bagged by fine nylon net, the emasculated flowers set
fruit and the fruit set were 0.5%-3.0%, which were slightly
lower than the fruit production ratio under natural condi-
tions (Table 4; Fig.1). The result indicated that this species
could also be pollinated by wind if the pollinators are
unavailable. And this is the first report of ambophily (a
combination of both wind and insect pollination) in the
genus Caesalpinia.
Autogamy or apomixis played no role in fruit produc-
tion in the species: none of the flowers bagged or with
stigma cut produced fruit. All inflorescence bagged and
flowers pollinated with pollen from the same plants recorded
similar results (Table 4).
2.3 Pollinators
Seventeen species (included to four Orders) of insects
were captured while they were visiting the flowers during
our study period (Table 5). Five frequently encountered
visitors were proved to be pollinators by identifying the
pollen on the stigmas of the previously bagged virgin flow-
ers after they had been visited. Hymenoptera were the most
common visitors and covered more than 90% of the total
number of visitation. Carpenter bees (Xylocopa sinensis
Smith, Xylocopa sp.), with the largest body, highest visit-
ing frequency and bringing most pollen per visit, were the
most important pollinators (Table 6).
2.4 Visitation patterns of the main pollinators and the
flower-insect interaction
In the flower of C. crista, the ten filaments tightly closed
against each other by its widened pubescent base, while
only two narrow splits were formed (Fig.2). The splits faced
the upper red spotted petal (standard). The standard re-
flexed at the middle and closed the filaments, so the splits
were covered by the standard. Of all the pollinators only
the carpenter bees can easily force away the standard with
its strong head and thus suck the nectar from the splits.
The carpenter bees approached the flowers from above
Table 2 Stigma receptivity of Caesalpinia crista as evaluated
by fruit set under artificial hand-pollination
Date Flowers Fruit
Fruit set
(%)
One day before bloom 3/14/2001 100 3 3.0
The blooming day 3/28/2000 200 7 3.5
3/15/2001 200 8 4.0
Two days after blooming 3/29/2000 60 3 5.0
3/15/2001 60 2 3.3
Five days after blooming 3/19/2001 20 0 0
Nine days after blooming 3/22/2001 30 1 3.3
Table 3 Pollen viability during different stages of anthesis
Days after anther Total number of Viable Percentage of
dehiscence pollens tested pollen viable pollen (%)
0 500 500 100
2 753 361 48
3 687 55 8
4 552 0 0
Table 4 The fruit set of Caesalpinia crista under different treatments
Treatments Date Flowers Fruit Fruit set (%)
Emasculation and netting 3/28/2000 200 1 0.5
3/28/2000 250 7 2.8
3/14/2001 200 6 3.0
Emasculation and no sheathing 3/30/2000 100 2 2.0
3/16/2001 45 2 4.4
Hand cross pollination between different plants in same population 3/29/2000 200 7 3.5
3/15/2001 200 8 4.0
Hand pollination between different inflorescences in same plant 3/28/2000 100 0 0
3/14/2001 100 0 0
Emasculation and bagging 3/28/2000 200 0 0
3/14/2001 200 0 0
Stigma cutting 3/28/2000 200 0 0
3/14/2001 200 0 0
Flower bagging 3/28/2000 200 0 0
3/14/2001 200 0 0
Inflorescence bagging 3/28/2000 200 0 0
3/14/2001 200 0 0
Untreated 3/28/2000 177 7 4.0
3/14/2001 122 6 4.9
275LI Shi-Jin et al.: Pollination Ecology of Caesalpinia crista (Leguminosae: Caesalpinioideae)
and in front. Grasped by the carpenter bee, the flower turned
down, and so the carpenter bee dangled from the flower,
facing upward while sucking the nectar (and thus easier to
suck the nectar). The carpenter bee grasped the upper pair
of lateral petals (keel petals) and lower lateral petals (wing
petals) with its middle and hind legs respectively. The front
legs can help its head to push the standard away when
sucking the nectar and brush the pollen which stuck on the
head when it has a rest. Usually, carpenter bees are effi-
cient pollinators. Almost all the stigmas had pollen and
more than 30 pollen grains were recorded on each stigma
once visited by the carpenter bees (Table 6). These pollina-
tors could suck about 1 µL nectar in the flower at one visit
and almost no nectar was left when they quitted. The nec-
tar volume in the flower decreased about 83.3% after one
visit by the carpenter bees (Table 7).
Although the stamens were not equal in length, the an-
thers and the truncated stigma were located at the same
bevel. While a carpenter bee is sucking the nectar, its hairy
abdomen touches the anthers and the stigma.
The second most important visitor was A. cerana
species. They approached the flower from above or in front.
They grasped the filaments, collected the pollen and
touched the stigmas by their abdomens. Seventy-five per
cent stigmas had pollen and an average of ca. 14 pollen
grains was brought on one stigma at each visit on flowers
visited by A. cerana. The nectar volume in the flower de-
creased by 16.7% after once visited by the honey bees
(Table 7). Bringing pollen and taking away less nectar, A.
cerana was a very economic pollinator.
Episyrphus balteata was another economic pollinator.
Having small and light body, they could grasp and stay on
anthers when eating pollen. Eighty per cent of virgin stig-
mas had pollen and about eight pollen grains were brought
to one stigma after once visited by E. balteata.
A species of hover-fly (Episyrphus sp.) was also among
the floral visitors. The insect approached the flower from
above and landed on the flower to collect pollen grains
from the dehisced anthers by licking. Hover-flies were even
able to lick off all the pollen grains on all the opened an-
thers in a flower. As they also licked the stigmas for its
exudate, hover-flies were presumably helping pollination.
However, a direct observation of their pollination efficiency
needs further observation.
Two large panicles with a total of 200 flowers were se-
lected and all the standard petals with red spots cut off. It
seemed that the removal of the standard petal with nectar
guides did not influence the visitors’ activities. In the other
treatment, when all the anthers of the 200 flowers on an
inflorescence were removed, the insects with pollen as their
only reward never came again. In the third treatment, all the
petals of the 200 flowers on an inflorescence were removed,
Table 5 Floral visitors and their rewards
Order Family Species Frequency* Aim
Hymenoptera Xylocopidae Xylocopa sinensis Smith 55+10** nectar
Xylocopa sp. nectar
Syrphidae Episyrphus balteata De Geer 10+4 pollen
Syrphus sp. - pollen and nectar
Verpidae Vespa affinis - pollen
Verpidae sp. - pollen
Apidae Apis cerana Fabr. 35+5 pollen and nectar
Apis sp. - pollen and nectar
Formicidae Formicidae sp. - nectar
Coleoptera Cetonnidae Potosia brevitarsis Lewis 3+2 ?
Lepidoptera Lycaenidae Japonica lutea Hewitson 7+3 nectar
Antigius attilis Bremer 6+3 nectar
Pieridae Delias pasithoe Linnaeus 6+3 nectar
Nymphalidae Mimathyma chevana Moore 8+4 nectar
Uraniidae Uraniidae sp. 3+2 nectar
Diptera Muscidae Episyrphus sp. 12+4 pollen
Musca domestica - nectar
*, frequency (presented on one inflorescence per hour) of a daily average; **, very similar in outline, it is too difficult to record the frequency
of the two species of Xylocopa respectively, and the frequency is the summation of the two species; -, less than 5 or have not been recorded.
Table 6 Pollination efficiency of the flower visitors
Xylocopa sinensis Xylocopa sp. Apis cerana Episyrphus balteata Syrphus sp.
Virgin stigmas visited 20 20 20 20 20
Stigmas on which pollen grains 19 20 15 16 13
were detected after one visit 95% 100% 75% 80% 65%
Percentage
Average pollen on the stigma 32.7+7 34.0+6 13.9+5 8.4+6 5.2+4
Acta Botanica Sinica 植物学报 Vol.46 No.3 2004276
only bees with small bodies that could grasp the filaments
and collect the pollen would visit the flowers again. Car-
penter bees, the primary pollinator, with a larger body size,
never came again, as there were no petals on which to grasp
and to land on (Table 8).
2.5 Influences of the treatments on flowering
Under natural conditions the petals can keep themselves
on the flower for 4-5 d before withering. When caged by
fine nylon fabric nets, withering will defers 2-3 d. If caged
by thick waterproof paper, they will defer 5-7 d (Table 9).
Under natural conditions the volume of extra-floral nec-
tar on each rachis is (4.0 ± 2.0) mL (range from 2.2 mL to 8.0
mL; N = 20) with 30% ± 12% (range 17%-70%; N = 20)
sucrose each collecting (at nine o’clock in fine weather).
When the inflorescence caged by waterproof paper (12.1 ±
4.1) mL (range from 9.0 mL to 14.1mL; N = 10) of nectar with
34% ± 8% (range 25%-55%; N = 20) mean sucrose were
produced on one rachis each collecting (Table 9).
(1.2 ± 0.7) mL (range from 0.55 mL to 2.8 mL; N = 20) nectar
with a concentration of 29% ± 5% (range from 17% to 35%;
N =20) sucrose was recorded per untreated full blooming
flower at nine o’clock in fine weather. It increased to (2.0 ±
0.8) mL (range 1.6 mL-3.0 mL; N=20) if bagged for more than
24 h and to (1.8 ± 0.8) mL (range 1.5 mL-2.8 mL; N = 20) if
emasculated (Table 9).
Under natural conditions the stigmas could remain vi-
able for about 5 d. They could only be viable for 2-3 d if
they had been assisted with hand-pollination. But they can
remain viable for 7 and 9 d respectively when caged by fine
nylon fabric nets and waterproof paper (Fig.9).
3 Discussion
3.1 Sexual and breeding system
There are 36 000 ± 500 pollen grains and two ovules in
Table 7 The nectar volume left in the flower of Caesalpinia crista after once visited by different insects
Treatments No. of flowers Range (µL) Average (µL) Percentage of the decrease (%)
Natural for control 20 0.55-2.8 1.2+0.7
Visited by carpenter bees (2)* 20 0-1.2 0.2+0.2 83.3
Visited by Apis cerana 20 0.6-2.5 1.0+0.4 16.7
Visited by Syrphus sp. 20 0.6-2.4 0.8+0.5 33.3
*, Xylocopa sinensis and Xylocopa sp.
Table 8 The role of different floral organs in attracting pollinators
Treatment No. of flowers Visitors
Visit Percentage of the frequency
Remarksfrequency* changing (%)
Standard removed 200 Carpenter bees (2)** 54 + 10 1.8 No influence
Apis cerana 35 + 5 0 No influence
Episyrphus balteata 11 + 4 10.0
Anthers removed 200 Carpenter bees (2)** 54 + 8 1.8 No influence
A. cerana 5 + 3 85.7 No pollen
E. balteata 0 100.0
All petals removed 200 Carpenter bees (2)** 5 + 2 90.9 No holder
A. cerana 20 + 5 42.9 Some influence
E. balteata 10 + 5 0 No influence
Natural for control 200 Carpenter bees (2)** 55 + 10
A. cerana 35 + 5
E. balteata 10 + 4
*, frequency (presented on one inflorescence per hour) of daily average; **, Xylocopa sinensis and Xylocopa sp.
Table 9 The influence of artificial treatments to the floral lifespan and behavior
Inflorescence The petals keep themselves The extrafloral nectar on The nectar in the flower
(N=10) on the flower (d) the rachis (mL)* (mL)**
Bagged 10-13 12.1 + 4.1 2.0 + 0.8
Netted 6-8 - -
Emasculated 4-5 4.0 + 2.0 1.8 + 0.8
Naturally 4-5 4.0 + 2.0 1.2 + 0.7
*, one time collecting at 9 o’clock in fine weather; **, from one full blooming flower at 9 o’clock in fine weather.
Fig.9. Stigma viability of Caesalpinia crista as revealed by
staining with MTT solution.
277LI Shi-Jin et al.: Pollination Ecology of Caesalpinia crista (Leguminosae: Caesalpinioideae)
species of the genus Caesalpinia (Lewis, 1998) and it oc-
curs in C. crista also. But the standard with red colored
spots of C. crista seemed to have no function in attracting
the pollinators. The petals, sepals and anthers were yellow,
and could be perceived by bees (Percival, 1965). Hundreds
of yellow flowers were found on the plant during the flow-
ering period. A great deal of extrafloral nectar appeared along
the rachis of C. crista even after the flowers withered.
Consequently, C. crista very likely attracts pollinators by
combined factors, but not by a single one.
3.3 Visitation patterns of the main pollinators and the
flower-insect interaction
The size and behavior of the carpenter bees fit the flower
size and its structure. The nectar was protected by the
androecium, and only carpenter bees with powerful heads
could easily suck the nectar. Several Caesalpinia species
was reported to be pollinated by Xylocopa species (Lewis,
1998; Lewis and Gibbs, 1999).
3.4 Influences of treatments on flowering
The method of being bagged or netted can prolong the
lifetime of the stigmas and petals. It may be the result of
increased temperature and humidity in the cage that caused
the prolongation of the floral lifespan, or it is more likely
because bagging of the flowers prevent pollens from fall-
ing on the stigmas and thus the ovules kept un-fertilized.
The lifespan of the flowers are prolonged waiting for
pollination. However it is still unknown why emasculation
stimulates the secretion of floral nectar.
In most studies, pollination biologists routinely used
nets and bags to prevent insects and the pollen in the air.
Our study manifested that the artificial treatments could
bring a certain degree of influence to the flowering and
fruiting function of the plants, and this influence should be
taken into consideration.
Acknowledgements: We thank Prof. SHI Zhen-Ya (Henan
Agricultural University) and Dr. TONG Xiao-Li (South China
Agricultural University) for identifying insect specimens.
We are also grateful to Prof. Amots Dafni (Haifa University,
Israel) for critically reading the manuscript and giving valu-
able suggestions, and Prof. CHEN Zhong-Mei (Beijing
Aerographical College) for linguistic corrections.
References:
Adams D E, Perkins W E, Estes J R. 1981. Pollination systems in
Paspalum dilitatum Poir. (Poaceae): an example of insect pol-
lination in a temperate grass. Am J Bot, 68:389-394.
Ali S A. 1932. Flower birds and bird-flowers in India. J Bomb
Natur Hist Soc, 35:573-605.
Bullock S H. 1985. Breeding systems in the flora of a Tropical
Deciduous Forest in Mexico. Biotropica, 17:287-301.
one flower of C. crista. The P/O ratio is 18 000 ± 500, and it
belongs to obligate xenogamy according to Cruden (1976).
It has been postulated that few plant-pollinator relation-
ships are absolutely obligatory, and generalization appears
to be the rule rather than the exception (Waser, 1996). Wind-
pollinated species usually have much higher P/O ratio than
their insect-pollinated relatives (Tomlinson et al., 1979;
Melampyand Hayworth, 1980; Adams et al., 1981). The high
value of the P/O ratio may also indicate that anemophily
played a role in C. crista (Pohl, 1937; Cruden, 1977; Preston,
1986).
Revealed by the fruit set by wind and insect pollinated
flowers, the pollination pattern of C. crista is ambophily,
and this is the first report in family Leguminosae.
Ambophilous species are generally considered as charac-
terized by numerous and simple flowers with small and open
corollas (Linder, 1998), the more showy and intricate flower
of C. crista indicated that this may not be always the rule.
C. crista, as a species distributed widely in pantropical
regions, could have been dispersed comparatively recently
to some of its extend ranges, and ambophily, an efficacious
pollination mode, may have helped it survive in its newly
claimed habitat while the pollinators were not available.
Under this hypothesis (adaptation hypothesis), ambophily
would be an adaption to the new environment. The other
possibility is that ambophily in the species was inherited
from its ancestors (phylogenetic hypothesis). Further stud-
ies on more species in the genus with a close phylogenetic
relationship and species with similar distribution pattern
are needed to test these hypotheses. There probably ex-
isted a gradual evolution of adaptations to wind-pollina-
tion from biotic pollination, where ambophily may have
been an intermediate stage (Linder, 1998; Wallander, 2001).
Theresa et al. (2002) recently concluded that ambophily
might be more common than was previously presumed and
could represent either a stable or a transitional state.
Hairs on the stigma seemed to play an important role in
pollen reception in the species. When honeybees and car-
penter bees came to collect pollen or nectar, pollens ad-
hered on the abdomen of the insects were scraped off by
the stigmatic fringe hairs like “combs” and were forced into
the stigmatic chamber. The hairs could also present the
pollen dissipating by wind or insects. The “comb-like” stig-
matic fringe hairs were reported in several species of
Caesalpinia previously (Lewis and Gibbs, 1999).
3.2 Combined attraction
Very often, bee-pollinated plants have colored spots on
the corolla attracting pollinators which is called “nectar
guide”. Nectar guide occurs on the standard petals of many
Acta Botanica Sinica 植物学报 Vol.46 No.3 2004278
Cheng H-C. 1988. Caesalpinia. Chen T-C. Flora Reipulicae
Popularis Sinicae, Tomus 39. Beijing: Science Press. 102-
103. (in Chinese)
Cocuccis A A, Galetto L, Sérsic A. 1992. El Síndrome floral de
Caesalpinia gilliesii (Fabaceae – Caesalpinioideae).
Darwiniana, 31:111-135.
Cruden R W, Hermenn-Parker S M. 1979. Butterfly pollination
of Caesalpinia pulcherrima with observations on a
psychophilous syndrome. J Ecol, 67:155-168.
Cruden R W. 1976. Intraspecific variation in pollen-ovule ratios
and nectar secretion-preliminary evidence of ecotypic
adaptation. Ann Missouri Bot Gard, 63:277-289.
Cruden R W. 1976. Pollen-ovule Patios: A conservative indicator
of Breeding systems in flowering plant. Evolution, 31:32-46.
Culley T M, Weller S G, Sakai A K. 2002. The evolution of wind
pollination in angiosperms. Trends Ecol Evol, 17:361-369.
Dafni A. 1992. Pollination Ecology: a Practical Approach. Oxford:
Oxford University Press.
Geber M A, Dawson T E, Delph L F. 1999. Gender and Sexual
Dimorphism in Flowering Plants. Berlin: Springer-Verlag.
Hattink T A. 1974. A revision of Malesian Caesalpinia, including
Mezoneuron (Leguminosae: Caesalpiniaceae). Reinwardtia, 9:
1-69.
Jones E, Buchmann S L. 1974. Ultraviolet floral patterns as
tunction cues in Hymenopterous pollination systems. Anim
Behav, 22:481-485.
Larsen K, Larsen S S, Vidal J E. 1984. Flora of Thailand. Vol. 4.
Part 1. Bangkok: Tistr Press. 70-72.
Lewis G, Gibbs P. 1999. Reproductive biology of Caesalpinia
calycina and C. pluviosa (Leguminosae) of the Caatinga of
north-eastern Brazil. Pl Syst Evol, 217:43-53.
Lewis G P. 1998. Caesalpinia: a Revision of the Poincianella-
Erythorstemon Group. Richmond: Royal Botanic Gardens
Kew.
Li S-J , Zhang D-X, Chen Z-Y . 2001. A species of Caesalpinia
Linn. (Leguminosae) new to China. Guihaia, 21:165. (in Chi-
nese with English abstract)
Linder H P. 1998. Morphology and the evolution of wind
pollination. Owens S J, Rudall P J. Reproductive Biology.
Kew: Royal Botanic Garden. 123-135.
Lloyd D G. 1965. Evolution of self-compatibility and racial dif-
ferentiation in Leavenworthia (Cruciferae). Contr Gray Herb,
195:3-134.
Lovett Doust J, Lovett Doust L. 1988. Plant Reproductive
Ecology: Patterns and Strategies. Oxford: Oxford University
Press.
Melampy M N, Hayworth A M. 1980. Seed production and
pollen vectors in several nectarless plants. Evolution, 34:1144-
1154.
Percival M S. 1965. Floral Biology. Oxford: Oxford Press.
Pohl F. 1937. Die Pollenerzeugung der Windbluter. Beih Bot
Centraldb, 56:365-470.
Preston R E. Pollen-ovule ratios in the Cruciferae. Am J Bot, 73:
1732-1740.
Richards A J. 1997. Plant Breeding Systems. London: Chapman
and Hall.
Rodriguez-Riano T, Dafni A. 2000. A new procedure to assess
pollen viability. Sex Plant Reprod, 12:241-244
Tomlinson P B, Primack R B, Bunt J S. 1979. Preliminary obser-
vations on floral biology in mangrove Rhizophoraceae.
Biotropica, 11:256-277.
Vogel S. 1990. Radiación adaptive del síndrome floral en las familias
neotropicales. Bol Acad Nac Ci, 59:5-30.
Wallander E. 2001. Evolution of Wind-pollination in Fraxinus
(Oleaceae) — an Ecophylogenetic Approach. Sweden:
Goteborg University.
Waser N M, Chittke L, Price M V, Williams N M, Ollerton J.
1996. Generalization in pollination systems, and why it maters.
Ecology, 77:1043-1060.
(Managing editor: HAN Ya-Qin)