全 文 ：新西兰岩石百合中两种共存的繁育策略:
延迟自动自交和欺骗传粉吸引*
周摇 伟1,3, 李德铢1,2, 王摇 红1,2**
(1 中国科学院昆明植物研究所生物多样性和生物地理学重点实验室, 云南 昆明摇 650201; 2 中国科学院西南
野生生物种质资源库, 云南 昆明摇 650201; 3 中国科学院研究生院, 北京摇 100049)
摘要: 延迟自动自交和欺骗传粉吸引在被子植物多数类群中有相关报道, 但是至今没有发现两种繁育策略
在同一物种中共存现象。 本研究通过对照试验检测新西兰岩石百合雄蕊附属物和花冠闭合运动是否分别具
有欺骗吸引和延迟自交功能。 研究结果表明, 新西兰岩石百合黄色的雄蕊附属物拟态花粉 (或花药), 约
93%的昆虫访花行为源自黄色花药附属物的欺骗吸引, 雄蕊附属物的报酬拟态功能有效提高昆虫拜访频率
并促进异交。 同时, 研究发现新西兰岩石百合花期结束时花瓣闭合促使雌雄隔离距离的缩小, 花瓣自然闭
合的花朵平均结籽数 (20. 62) 显著高于闭合前去雄处理花朵 (11. 79)。 我们的结果表明延迟自动自交与
欺骗传粉吸引两种繁育策略在新西兰岩石百合中共存。
关键词: 延迟自动自交; 欺骗传粉吸引; 新西兰岩石百合
中图分类号: Q 948摇 摇 摇 摇 摇 摇 摇 文献标识码: A摇 摇 摇 摇 摇 摇 摇 文章编号: 2095-0845(2012)02-187-05
Coexistence of Delayed Autonomous Self鄄pollination and Deceptive
Pollination in Arthropodium cirratum (Asparagaceae)
ZHOU Wei1,3, LI De鄄Zhu1,2, WANG Hong1,2**
(1 Key Laboratory of Biodiversity and Plant Biogeography, Kunming Institute of Botany, Chinese Academy of Sciences,
Kunming 650201, China; 2 Plant Germplasm and Genomics Center, Germplasm Bank of Wild Species,
Kunming Institute of Botany, Chinese Academy of Sciences, Kunming 650201, China;
3 Graduate University of Chinese Academy of Sciences, Beijing 100049, China)
Abstract: Delayed autonomous pollination and deceptive pollination are assumed to be relatively common in flower鄄
ing plants, but no species have been reported to use both of these cunning reproductive strategies. In this study, we
examined whether delayed selfing and mimicry were used concurrently in Arthropodium cirratum. Flowers of A. cirra鄄
tum were manipulated to assess whether their stamen appendages and corolla closing movements were functional in
deceptive pollination and delayed selfing, respectively. Our results indicated that anther or pollen imitation of the
yellow stamen appendages contributed to 93 per cent of successful mimetic attraction, and this mimicry was an im鄄
portant driver of cross鄄pollination of the species. In addition, we observed closure of the perianth relocated the sta鄄
mens over the stigma at the end of anthesis, which significantly increased the average seed number per fruit of intact
flowers over flowers emasculated before closure (20. 62 versus 11. 79). Our findings confirmed the coexistence of
delayed autonomous self鄄pollination and deceptive pollination in A. cirratum. Our results also suggested that delayed
selfing could add fitness benefits for this mimetic attraction species.
Key words: Arthropodium cirratum; Delayed autonomous self鄄pollination; Deceptive pollination
植 物 分 类 与 资 源 学 报摇 2012, 34 (2): 187 ~ 191
Plant Diversity and Resources摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 DOI: 10. 3724 / SP. J. 1143. 2012. 11169
*
**
Foundation item: The National Basic Research Programme of China (973 Programme, 2007CB411600)
Author for correspondence; E鄄mail: wanghong@ mail. kib. ac. cn
Received date: 2011-11-28, Accepted date: 2012-02-08
作者简介: 周摇 伟 (1980-) 男, 博士, 主要从事植物繁育生物学与分子生物学方面研究。
摇 Autonomous self鄄pollination, which occurs in
flowers that are capable of outcrossing but in which
selfing occurs without the participation of an external
agent, is an important mating strategy to ensure sex鄄
ual reproduction in harsh environments where polli鄄
nator services are unreliable. Lloyd and Schoen
(1992 ) classified autonomous self鄄pollination into
three general categories, termed ‘ prior爷, ‘ compe鄄
ting爷 and ‘ delayed爷, according to the timing of
self鄄pollination relative to cross鄄pollination. Delayed
self鄄pollination occurs after the opportunity to out鄄
cross has passed, and it may be achieved by either a
partial overlap in the timing of male vs. female pha鄄
ses or by changes in the relative position of anthers
and stigma within a flower during corolla movements
(e. g. corolla abscission or closure). Delayed au鄄
tonomous self鄄pollination is usually considered to be
adaptive because it assures seed production when
pollinators are scarce, yet allows cross鄄pollination to
predominate when they are abundant (Cruden, 1977;
Wyatt, 1983; Cruden and Lyon, 1989; Sun et al.,
2005; Fenster and Mart佴n鄄Rodr侏guez, 2007).
Despite the selective advantage of delayed autono鄄
mous self鄄pollination, surprisingly few mechanisms are
involved. For entomophilous plants, variation in polli鄄
nator frequency may be a significant force in the main鄄
tenance of mechanisms that induce autogamy late in
floral life (Cruden and Lyon, 1989). Following this
general idea, we raise the hypotheses that plant using
mimetic attraction should be more likely to have
evolved delayed selfing as a “bet鄄hedging冶 strategy
since such plants may experience extremely unpredic鄄
table pollinator service due to their often rewardless,
imperfectly mimetic flowers and the resulting learning鄄
based cheating behavior of their pollinators (Mosquin,
1970; Menzel and Erber, 1978; Boyden, 1982; Ver鄄
eecken and Schiestl, 2008). However, a link be鄄
tween mimetic attraction and delayed autonomous
self鄄pollination in plants has not been described.
Preliminary observations of the “ renga renga
lily冶, Arthropodium cirratum (Asparagaceae), sug鄄
gested that this perennial herb may utilize both mi鄄
metic attraction from its pollen鄄like stamen appenda鄄
ges and delayed autonomous self鄄pollination promo鄄
ted by corolla closure. The flower of this species do
not offer nectar and was scentless, but the yellow
parts of the stamen appendages was always the target
for visitors while the purple parts of the stamen ap鄄
pendages and the inconspicuous anthers themselves
were mostly ignored. Late in flower development,
the petals closed, the motion of which brought an鄄
thers and stigma together. Therefore, we designed
this study to determine (1) whether the yellow sta鄄
men appendages attract pollinators to A. cirratum;
and (2) whether the petal closure of the species
triggers delayed autonomous self鄄pollination as a
means of reproductive success. We tested these
mechanistic hypotheses using detailed observations
and manipulating flowers to measure pollinator visit鄄
ation and seed production.
Materials and methods
Study system
Arthropodium cirratum is endemic to New Zeal鄄
and and distributed across the Northern part of the
island (Moore and Edgar, 1970). It was introduced
to China in 1990 as an ornamental. This study was
conducted at the Kunming Botanical Garden, Yun鄄
nan, China (KBG; 25毅07忆30义 N, 102毅44忆17义 E;
altitude, 1 950 m a. s. l), where the species flowers
from early May to late June, with the fruits maturing
more than one month later. A. cirratum produces 1 to
4 inflorescences that bear more than 50 flowers
each, and inflorescence are subtended by numerous
leafy bracts. The Latinate term “cirratum冶 refers to
the curled tendrils of the staminal appendage. These
hair鄄like appendages can be divided into two types,
the proximal yellow appendages and the distal purple
appendages (Fig. 1C).
Mimic attraction experiment
To investigate the role of anther hairs in polli鄄
nator attraction, we recorded pollinator visits to four
kinds of flowers: (1) intact flowers (n = 40); (2)
flowers with the yellow stamen appendages (YA) re鄄
881摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 植 物 分 类 与 资 源 学 报摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 第 34 卷
moved (n = 40); (3) flowers with the purple sta鄄
men appendages (PA) removed (n=48); and (4)
flowers with anthers entirely removed (n = 40). To
reduce the possibility that the flower modifications
would influence pollinator attraction at the inflores鄄
cence level, we combined three or four treatments
within a single inflorescence, with each treatment
applied to 2-5 flowers at randomized positions across
the inflorescence. Pollinator visits were recorded at
two time鄄intervals per day, coinciding with maximum
pollinator activity, from 0800 hours to 1000 hours
and from 1200 hours to 1400 hours during May 27 to
May 29 and June 3 to June 4 in 2008. Visits were
recorded each time a pollinator landed on any of the
flowers, and a total of 148 flowers on 15 inflores鄄
cences were observed.
Delayed selfing experiment
To detect whether delayed autonomous self鄄pol鄄
lination occurs during petal closure, we designed
four treatments: (1) natural pollination as a con鄄
trol, the seed set of which was expected to be high鄄
est; (2) open pollinated and emasculated before the
petals closed, the seed set of which would result
from pollinator鄄carried pollen; (3) pollinator exclu鄄
sion (using 1 mm伊1 mm nylon netting), the seed
set of which would be due to autonomous self鄄polli鄄
nation following floral closure (given the absence of
apomixis); and (4) apomixis, for which anthers
were removed upon anthesis and flowers were cov鄄
ered with impermeable bags. Treatments were isola鄄
ted on different inflorescences, and 20 -30 flowers
were manipulated in each treatment. One month lat鄄
er, we harvested 24 fruits in each treatment and
counted the number of seeds per fruit.
Results
Continuous observation indicated that A. cirra鄄
tum flowers opened between 0100 and 0200 hours,
and closed between 1900 and 2000 hours. Thus the
life鄄span of the flower was approximately 18 hours
and never reopened after closure. The events during
flower closure followed a rigid sequence with the in鄄
ner three petals closing before the outer three petals.
During two observation days in which rain fell con鄄
tinuously, flowers opened at the same time for sunny
days but petal closure was delayed by 2-3 hours.
Pollen鄄collecting bees and syrphid flies were
the major visitors and pollinators of A. cirratum flow鄄
ers, and no nocturnal visitor was observed when ob鄄
servation was made between 0100 and 0500. We re鄄
corded a total of 974 visits to flowers, of which bees
and syrphid flies accounted for 98. 5% (960 vis鄄
its), the remaining 1. 5% (14 visits) of visits were
performed by butterflies and beetles. Mean visits per
hour were 4. 05依1. 12 (mean依s. d. ), 0. 28依0. 45,
3. 64依1. 01 and 3. 86依0. 95 for treatments 1 to 4 re鄄
spectively ( Fig. 2A). Visits to intact ( treatment
1), PA鄄removed (treatment 3) and anther鄄removed
(treatment 4) flowers were not significantly different,
but all three treatments had significantly more visits
than for YA鄄removed (treatment 2) flowers (ANOVA,
F =150. 77, d. f. = 3, 144, P<<0. 001, using Dun鄄
can爷s multiple comparison test; see Fig. 2A).
Flowers from all treatments except treatment 4
produced fruit (24 in each treatment). Mean seed
number per fruit for treatment 1 ( natural pollina鄄
tion) was 20. 62依5. 28; treatment 2 ( flowers emas鄄
culated before the petals closed) was 11. 79依4. 61;
and treatment 3 ( automatic self鄄pollination) was
15. 91 依 5. 94 ( Fig. 2B). There were significant
differences in seed number per fruit among the three
treatments (treatment 1, treatment 2 and treatment 3;
ANOVA, F = 16. 61, d. f. = 2, 69, P<<0. 001).
The results for treatment 4 (20 flowers) indicated
that A. cirratum is incapable of apomixis.
Discussion
Because pollen grains of A. cirratum are sticky
and apparently not moved by wind, successful repro鄄
duction relies on pollen transfer by its insect visi鄄
tors. Our study showed that, although the flowers of鄄
fer no reward at least no nectar, they attract pollen鄄
collecting visitors by their conspicuous stamen ap鄄
pendages mimicking pollen鄄abundant anthers. How鄄
9812 期摇 摇 摇 ZHOU Wei et al. : Coexistence of Delayed Autonomous Self鄄pollination and Deceptive Pollination in …摇 摇 摇
ever, the attractiveness of the two kinds of the ap鄄
pendages is asymmetrical, with high levels of attrac鄄
tiveness from yellow appendages (93% ) but low at鄄
tractiveness from the purple appendages ( 10% )
(Fig. 2A). This result supports previous suggestions
that most anther鄄 and pollen鄄mimicking structures
copy the deep yellow color of pollen and anthers,
which represents a supernormal stimulus for most
visitors, especially for bees (Lunau, 2000, 2006).
In fact, we observed that bees and syrphid flies vis鄄
its (but for few Episyrphus balteatus targeted at an鄄
thers see Fig. 1B) were predominantly to the yellow
anther appendages, which brought the insects 爷
bodies (abdomens) contact with anthers and stigmas
(Fig. 1A). This foraging behavior facilitated pollen
transfer within or among flowers, with about 57. 0 per
cent of seed set (11. 79依4. 61 versus 20. 62依5. 28)
coming from pollinator鄄facilitated self鄄 and cross鄄pol鄄
lination. Moreover, the comparison of seed number
per fruit between pollinator鄄excluded and naturally
pollinated flowers showed that at least 22. 7 per cent
of seed set (1-77. 3% , 15. 91依5. 94 versus 20. 62
依5. 28) result from cross鄄pollination. Therefore, the
pollen mimicry of the yellow stamen appendages ap鄄
pears to be an important driver of cross鄄pollination of
the species.
The seed set of flowers emasculated before floral
closure (11. 79依4. 61) was significantly lower than
Fig. 1摇 Flower visitors and floral anatomy of Arthropodium cirratum. A, Mimetic attraction by yellow stamen appendges; B, Anther target visitation;
C, Stamen structures. PA, purple stamen appendage; YA, yellow stamen appendage; D, Positions of anther and stigma in closed flower
Fig. 2摇 Visit frequence and number of seed per fruit in Arthropodium cirratum. A, Numbers of bee visits per hour to flowers (n=148 flower)
subjected to four experimental treatments; B, Seed number per fruit from the four pollination treatments (n=72 fruit) . Different letter
indicate significant differences between means by Duncan爷s multiple comparison test (P<0. 01)
091摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 植 物 分 类 与 资 源 学 报摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 第 34 卷
for flowers allowed to close naturally with intact an鄄
thers (20. 62依5. 28 see Fig. 2B). This indicates that
the increase in seed set associated with corolla clo鄄
sure results from delayed autonomous self鄄pollination.
Selfed seeds were produced because closure of the
perianth pushes the stamens over the stigma at the
end of anthesis (Fig. 1D). This mechanism of de鄄
layed selfing is different from that of other species,
such as Kosteletzkya virginica, in which delayed sel鄄
fing results combination of stylar movements and co鄄
rolla closure (Ruan et al., 2005). Since at this study
site, the species was cultivated in clusters and the
pollinator visiting frequency was high (4. 05 依1. 12
per hour), the delayed selfing mechanism may be
more crucial for naturally occurring A. cirratum grow鄄
ing in small, isolated populations (Fisher et al., 1970;
Parris et al., 1971) where pollinator service may be
limited and reproductive assurance may be required.
Delayed autonomous self鄄pollination and decep鄄
tive pollination are both assumed to be relatively
common reproductive strategies in flowering plants,
especially for species with unpredictable pollinator
services and for orchid species, respectively (Schoen
and Brown, 1991; Lloyd, 1992; Barrett, 2003).
However, to our knowledge, this is the first report of
delayed autonomous self鄄pollination also active in
plants utilizing deceptive pollination. The delayed
selfing could lend increased fitness benefits in this
mimetic species. It would be interesting to know
whether delayed selfing also contributes to reproduc鄄
tive success in other deceptive pollination species,
since such plants may be more likely to suffer from
reproductive failure and risk population loss or even
species extinction.
Acknowledgements: The authors thank Dr. Yang Niu (Kunming
Institute of Botany, Chinese Academy of Sciences) and Hua鄄ying Sun
(Kunming Institute of Zoology, Chinese Academy of Sciences) for
their kind assistance with pollinator observation.
References:
Barrett SCH, 2003. Mating strategies in flowering plants: the out鄄
crossing鄄selfing paradigm and beyond [ J] . Philosophical Trans鄄
actions of the Royal Society Series B: Biological Sciences, 358:
991—1004
Boyden TC, 1982. The pollination biology of Calypso bulbosa var.
americana (Orchidaceae): initial deception of bumblebee visi鄄
tors [J] . Oecologia, 55: 178—184
Cruden RW, 1977. Pollen鄄ovule ratios: a conservative indicator of
breeding systems in flowering plants [J]. Evolution, 31: 32—46
Cruden RW, Lyon DL, 1989. Facultative xenogamy: Examination of
a mixed mating system [A]. In: Bock JH, Linhart YB eds.,
The Evolutionary Ecology of Plants [M]. Colorado: Westview
Press, 171—208
Fenster CB, Mart佴n鄄Rodr侏guez S, 2007. Reproductive assurance and
the evolution of pollination specialization [ J ] . International
Journal of Plant Sciences, 168: 215—228
Fisher ME, Satchell E, Watkins JM, 1970. Gardening with New Zea鄄
land Plants, Shrubs and Trees [M]. Auckland: Collins
Lloyd DG, 1992. Self鄄 and cross鄄fertilization in plants. II. The selec鄄
tion of self鄄fertilization [ J] . International Journal of Plant Sci鄄
ences, 153: 370—380
Lloyd DG, Schoen DJ, 1992. Self鄄and cross鄄fertilization in plants. I.
Functional dimensions [ J] . International Journal of Plant Sci鄄
ences, 153: 358—369
Lunau K, 2000. The ecology and evolution of visual pollen signals
[J] . Plant Systematics and Evolution, 222: 89—111
Lunau K, 2006. Stamens and mimic stamens as components of floral
colour patterns [J] . Botanische Jahrb俟cher, 127: 13—41
Menzel R, Erber J, 1978. Learning and memory in bees [J] . Scien鄄
tific American, 239: 102—111
Moore LB, Edgar E, 1870. Floral of New Zealand, Vol. II. Indige鄄
nous Tracheophyta, Monocotyledones Except Gramineae [ M].
Wellington: Government Printer
Mosquin T, 1970. The reproductive biology of Calypso bulbosa (Or鄄
chidaceae) [J] . Canadian Field鄄Naturalist, 84: 291—296
Parris BS, Lynch PA, Ferguson EJ, 1971. The vegetation of Whale
Island Part 1: the plant communities [J] . Tane, 17: 33—38
Ruan CJ, Qin P, Xi YG, 2005. Floral traits and pollination modes in
Kosteletzkya virginica ( Malvaceae) [ J] . Belgian Journal of
Botany, 138: 39—46
Schoen DJ, Brown AHD, 1991. Whole鄄and part鄄flower self鄄pollina鄄
tion in Glycine clandestina and G. argyrea and the evolution of au鄄
togamy [J] . Evolution, 45: 1651—1664
Sun SG, Guo YH, Gituru RW et al., 2005. Corolla wilting facilitates
delayed autonomous self鄄pollination in Pedicularis dunniana
(Orobanchaceae) [ J] . Plant Systematics and Evolution, 251:
229—237
Vereecken NJ, Schiestl FP, 2008. The evolution of imperfect floral
mimicry [J] . Proceedings of the National Academy of Sciences,
105: 7484—7488
Wyatt R, 1983. Pollinator鄄plant interactions and the evolution of
breeding systems [J] . Pollination Biology, 51: 403—418
1912 期摇 摇 摇 ZHOU Wei et al. : Coexistence of Delayed Autonomous Self鄄pollination and Deceptive Pollination in …摇 摇 摇