Annals of Botany 105: 999 –1014, 2010
doi:10.1093/aob/mcq041, available online at www.aob.oxfordjournals.org
PART OF A HIGHLIGHT SECTION ON SEEDS
Fruit and seed heteromorphism in the cold desert annual ephemeral
Diptychocarpus strictus (Brassicaceae) and possible adaptive significance
Juanjuan Lu 1, Dunyan Tan 1,*, Jerry M. Baskin 1,2 and Carol C. Baskin 1,2,3,*
1
Xinjiang Key Laboratory of Grassland Resources and Ecology, College of Grassland and Environment Sciences, Xinjiang
Agricultural University, Urümqi 830052, China, 2Department of Biology, University of Kentucky, Lexington, KY 40506, USA
and 3Department of Plant and Soil Sciences, University of Kentucky, Lexington, KY 40546, USA
* For correspondence. E-mail [email protected] or [email protected]
Received: 25 November 2009 Returned for revision: 4 January 2010 Accepted: 21 January 2010 Published electronically: 27 March 2010
† Background and Aims Diptychocarpus strictus is an annual ephemeral in the cold desert of northwest China that
produces heteromorphic fruits and seeds. The primary aims of this study were to characterize the morphology and
anatomy of fruits and seeds of this species and compare the role of fruit and seed hetermorphism in dispersal and
germination.
† Methods Shape, size, mass and dispersal of siliques and seeds and the thickness of the mucilage layer on seeds
were measured, and the anatomy of siliques and seeds, the role of seed mucilage in water absorption/dehydration,
germination and adherence of seeds to soil particles, the role of pericarp of lower siliques in seed dormancy and
seed after-ripening and germination phenology were studied using standard procedures.
† Key Results Plants produce dehiscent upper siliques with a thin pericarp containing seeds with large wings and
a thick mucilage layer and indehiscent lower siliques with a thick pericarp containing nearly wingless seeds with
a thin mucilage layer. The dispersal ability of seeds from the upper siliques was much greater than that of intact
lower siliques. Mucilage increased the amount of water absorbed by seeds and decreased the rate of dehydration.
Seeds with a thick mucilage layer adhered to soil particles much better than those with a thin mucilage layer or
those from which mucilage had been removed. Fresh seeds were physiologically dormant and after-ripened
during summer. Non-dormant seeds germinated to high percentages in light and in darkness. Germination of
seeds from upper siliques is delayed until spring primarily by drought in summer and autumn, whereas the
thick, indehiscent pericarp prevents germination for .1 year of seeds retained in lower siliques.
† Conclusions The life cycle of D. strictus is morphologically and physiologically adapted to the cold desert
environment in time and space via a combination of characters associated with fruit and seed heteromorphism.
Key words: Cold desert annual ephemeral, Diptychocarpus, fruit and seed heteromorphism, non-deep
physiological dormancy, seed dispersal, seed germination.
IN T RO DU C T IO N
Heterocarpy or fruit (seed) heteromorphism is the production
of seeds of different morphologies and/or behaviour on different parts of the same plant (Imbert, 2002) and is an adaptation
of species to the spatio-temporal variability of habitats
(Venable and Lawlor, 1980; Venable et al., 1998). It may be
viewed as a bet-hedging strategy in which organisms evolve
traits that reduce short-term reproductive success in favour of
long-term risk reduction (Venable, 1985, 2007) or that allow
escape from the negative effects of density (Levin et al.,
1984; Sadeh et al., 2009) or sib competition (Lloyd, 1984;
Cheplick, 1992). Heterocarpy appears to be confined to a
limited number of families of the phylogenetically advanced
angiosperms (i.e. 18 families and .200 species), in particular
the Asteraceae, Brassicaceae, Chenopodiaceae and Poaceae
(Imbert, 2002). Heteromorphic fruits or seeds may vary in
size, colour and morphology/anatomy, as well as in dispersal,
dormancy and germination (Baskin and Baskin, 1998).
In Brassicaceae, fruit and seed heteromorphism has been
studied most extensively in the genus Cakile. Fruit heteromorphism of Cakile may be associated with differences in
dispersal (Payne and Maun, 1981; Donohue, 1998;
Cordazzo, 2006), germination behaviour (Barbour, 1970a;
Maun and Payne, 1989; Zhang, 1993) and seedling emergence
and survival rates (Barbour, 1970b; Maun and Payne, 1989;
Zhang and Maun, 1992; Zhang, 1993, 1995). However, other
than for the genus Cakile, little is known about the role of
dimorphic fruits and seeds in the dispersal and germination
stages of the life cycle of Brassicaceae species.
Diptychocarpus strictus is an annual ephemeral
Brassicaceae species that occurs in Middle Asia, Iran,
Turkey, Caucasia and China (An and Zhou, 1995). This is
the only species in the genus, and in China it grows only in
the southern part of the Junggar Basin, Xinjiang province.
The species has two flower colour morphs ( purple and
white), and two morphologically distinct types of fruits
(upper and lower) are produced by each of them.
Additionally, seeds from upper and lower siliques of each
flower colour morph vary in width of the wing and in
amount of mucilage (Fig. 1). Diptychocarpus strictus is one
of the very common ephemeral (annual) species that germinates in autumn and early spring (mostly) in the Junggar
Desert. Flowering of this species occurs from mid-April to
# The Author 2010. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved.
For Permissions, please email: [email protected]
1000
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
A
Peduncle
C Peduncle
Beak
Beak
D
B
E
F
G
H
W
M
M
W
S
1 mm
1 mm
S
S
S
1 mm
1 mm
F I G . 1. Morphology of the two types of siliques and seeds in Diptychocarpus strictus. (A, B) Upper silique (A, ventral side; B, lateral side). (C, D) Lower silique
(C, ventral side; D, lateral side). (E, F) Seed from upper siliques (E, dry; F, imbibed). (G, H) Seed from lower siliques (G, dry; H, imbibed). M, mucilage; S, seed;
W, wing.
early May, and seeds mature in late May to early June. Siliques
and/or seeds are dispersed within a few days after maturity by
wind or rain; those at the bottom of the infructescence are shed
first. Pericarps of upper siliques first begin to dehisce at the end
of the fruit nearest to the peduncle, and the pericarp remains
attached to the mother plant via connection of the peduncle
to the membranous fruit partition. The seeds with a wide
wing from the upper siliques are units of dispersal and of germination. However, lower siliques with highly lignified pericarps do not dehisce, and usually whole (but sometimes
transversally broken segments) siliques are dispersed near
the mother plant. Thus, seeds are the dispersal units of upper
siliques, and both intact and transversally broken siliques are
the dispersal units of lower siliques.
We hypothesized that fruit and seed heteromorphism in
D. strictus aids the distribution of this species in time and
space in its cold desert habitat. To test this hypothesis, for
both flower colour morphs we (a) compared the morphology
and anatomy of upper and lower siliques and their seeds; (b)
compared dispersal of seeds after release from dehiscent
upper siliques with that of seeds retained inside indehiscent
lower siliques; (c) determined the role of mucilage on seeds
from upper and lower siliques that vary in amount of mucilage;
and (d) compared dormancy and germination characteristics of
seeds freed from the two silique types with each other and with
seeds retained in lower siliques. Since the results of the two
flower colour morphs were essentially identical, only those
for the purple flower morph will be presented herein.
The importance of fruit/seed heteromorphism in the life
cycle of D. strictus is not due to differences in the germination
ecophysiology between the two seed morphs per se, as has
been reported for many species, but to the release at maturity
of seeds from the upper dehiscent fruits and the long-term
retention and imposed dormancy by the lower indehiscent
fruits of the seeds within them.
M AT E R I A L S A N D M E T H O D S
Field site description and silique collection
Freshly matured siliques, i.e. with khaki coloured pericarps
and dispersing naturally, were collected from natural
Diptychocarpus strictus (Fisch. ex M. Bieb.) Trautv. populations
growing in the gravel desert in the vicinity of Urümqi on the
southern edge of the Junggar Basin of Xinjiang (438480 N,
878320 E, 850 m a.s.l.) on 1 June 2008. Siliques from the infructescenses were bulked into one collection and separated into upper
and lower fruits. In the laboratory, siliques were separated from
other plant material and stored in paper bags under ambient conditions [18–25 8C, 30 % relative humidity (RH)] until used.
This area of the Junggar Basin has typical desert vegetation
dominated by Chenopodiaceae and Asteraceae, gravelly grey
desert soil and a continental climate. Mean annual temperature
is 6.8 8C, and the extreme temperatures of the coldest
(January) and hottest (July) months are 232.8 and 40.5 8C,
respectively. Average annual precipitation (including rain and
snow) is 234 mm, 62.7 % of which occurs in spring and
summer, and the snow that falls in winter begins to melt in
March or April. The annual potential evaporation is
.2000 mm (Wei et al., 2003).
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
1001
Characteristics and dispersal of siliques and seeds
Seed mucilage
Colour, shape, size and mass. Colour, shape, size and mass were
determined for siliques of each type collected in June 2008 and
stored dry under laboratory conditions for 2 months. The
length and width of the silique, length of the peduncle,
length of the beak, length and width of the seed and width
of the seed wings (see Fig. 1) were measured for 100 randomly
chosen fruits or seeds using digital calipers. The number of
seeds was recorded in a sample of 50 upper siliques and 50
lower siliques. Fifty siliques of each silique type were
weighed individually using a Sartorius BS210S electronic
balance (0.0001 g), and then the mass of the pericarp and
seeds inside these siliques was determined. Ten replications
of 1000 seeds of each of the two morphs (10 1000 2) collected in June 2008 and stored dry under laboratory conditions
for 2 months were weighed to the nearest 0.0001g. Colour,
shape and size were determined for 50 dry siliques of each
of the two morphs and for 50 seeds (one seed per silique).
Mean and s.e. were calculated for mass and width.
When wetted, the mucilage on the seeds becomes gel like,
and the edges expand and form a sticky mucilaginous layer
that eventually surrounds the seeds (Fig. 1F, H). Thus,
various experiments were carried out to investigate the role
of mucilage on seeds collected in June 2008 and stored dry
under laboratory conditions for 2 months.
Anatomy. To compare the anatomy of the pericarps and seed
coats of each silique morph, standard anatomical sections
were prepared for 20 nearly mature siliques and seeds of the
two morphs (upper and lower). Siliques and seeds collected
in June 2008 were fixed at room temperature in FAA
(formalin – acetic acid, with 70 % ethanol, 0.5 : 0.5 : 9 v/v/v)
for 3 months; no shrinkage or distortion of tissues was
observed during the fixation period. The material was dehydrated in a graded 70 % ethanol series, cleared in a graded
dimethylbenzene series, infiltrated in paraffin (each step for
at least 1 h) and embedded in paraffin wax (Zheng and Gu,
1993). Transverse sections (8 – 12 mm) were cut by an exact
revolving microtome (HistoSTAT 820) and routinely stained
with safranin and fast green. Images were taken using a
camera mounted on an Olympus BH-2 Biology Microscope.
Based on the micrographs, characteristics of the pericarp and
seed coat were examined.
Dispersal. In July 2008, the dispersal ability of seeds (upper
siliques) and intact lower siliques was determined by the following procedures indoors. (a) To measure fall rate in still
air, a tube 120 cm tall and 15 cm in diameter was used.
Fifty individual seeds from upper siliques and 50 whole,
intact lower siliques were released individually from the top
of the tube, and the time required for them to fall from the
release point to the bottom of the tube was measured with a
digital stopwatch. Fall rate (cm s21) was calculated over this
height (Gravuer et al., 2003). The assumption in this method
is that seeds with a slower fall rate could be carried further
by the wind than those with a faster fall rate. (b) Dispersal distances were experimentally determined following the method
of Telenius and Torstensson (1989). Thus, 50 individual
intact seeds from upper siliques and 50 whole, intact lower siliques were exposed for 60 s to a constant stream of air ( produced by a fan) parallel to the flat seed-landing surface. The
seeds and siliques were released 1 cm from the front of the
fan at a height of 30 cm and exposed to wind velocities
of 1 and 4 m s21, and the distances they travelled were
measured to 0.001 m.
Change in width and mass of imbibed seeds. The widths of 30
intact dry and of 30 imbibed seeds, selected at random from
upper and lower siliques, were measured using a calibrated
Nikon SMZ1000 light microscope. Intact and mucilage-free
seeds (mucilage removed by scraping dry seeds with a
scalpel) were hydrated to constant mass. The mucilage-free
and intact seeds were weighed before and after imbibition.
Ten groups of 100 seeds were used for mass determination
of each morph. The increase in mucilage width (mm) (Wm)
of imbibed seeds was calculated as Wm ¼ (Wi 2 Wo)/2,
where Wi and Wo are the widths of imbibed seeds with and
without mucilage, respectively; the mucilage width ratio
(Wm0 ) of imbibed seeds as Wm0 ¼ (Wi 2 Wo)/2Wi; the
number of times the mass (Mt) increased after imbibition as
Mt ¼ Mi/Md, where Mi and Md are the mass of imbibed and
dry seeds, respectively; the mucilage mass of dry seeds
(Mmd) as Mmd ¼ Mwd 2 Mwod, where Mwd and Mwod are the
mass of dry seeds with and without mucilage, respectively;
the mucilage mass of imbibed seeds (Mmi) as Mmi ¼ Mwi 2
Mwoi, where Mwi and Mwoi are the mass of imbibed seeds
with and without mucilage, respectively; and the proportion
of mucilage to seed mass before (Mmd) and after (Mmi) imbibition as Mmb ¼ Mmd/Mwd and Mma ¼ Mmi/Mwi, respectively.
Water absorption and dehydration. Water absorption (water
uptake) was monitored for four replicates of 100 seeds in
each of the two seed morphs, with and without mucilage
(4 100 2 2), to compare the water-absorbing properties
of the mucilage. The mass of each group was recorded at time
0 and every 10 min until final constant mass.
Dehydration (water loss) was monitored in the same groups
of seeds under laboratory conditions (22 – 25 8C, 30 % RH).
The mass of each group of seeds was recorded at time 0 and
every 10 min until final constant mass.
Mass of seeds with and without mucilage plus adhered soil
particles.
(a) Adherence to wet soil particles by dry seeds with and
without mucilage. Four replicates of 100 dry seeds of
each of the two morphs, with and without mucilage,
were weighed. Seeds were distributed on water-saturated
desert soil in 9 cm diameter Petri dishes under room conditions (22 –25 8C, 30 % RH). After swelling to full water
capacity, the seeds were redried at room temperature. The
mass of the dry seeds including attached soil particles was
compared with the mass of dry seeds before adherence to
soil particles. During incubation, soil particles adhered to
the mucilage surrounding the seeds (¼ mass of imbibed
seeds with mucilage plus soil particles).
(b) Adherence to dry soil particles by imbibed seeds with and
without mucilage. Four replicates of 100 seeds of each of
the two morphs with and without mucilage were weighed
1002
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
and the seeds submerged in distilled water until they were
fully imbibed. Then, the seeds were placed on fine-particle
dry soil in 9 cm diameter Petri dishes and rolled until they
would not adhere to any more particles, after which the
seed – soil particle units were allowed to dry under room
conditions. The mass of the dry seeds with fully developed
mucilage including the soil particles that had adhered to
them was compared with the mass of dried seeds before
adherence to soil particles. The number of times seed
mass increased (Mt0 ) after adherence of sand particles
was calculated as Mt0 ¼ Ms0 /Md0 , where Ms0 and Md0 are
the mass of imbibed seeds with and without mucilage
plus sand particles and dry seeds, respectively.
Germination ecophysiology
Effect of dry storage (after-ripening) on germination. To investigate the germination responses of seeds during dry storage,
seeds stored in laboratory conditions for 0 (fresh), 1, 2, 3
and 12 months were tested for germination. For each test,
three replicates of 50 seeds of each of the two morphs for
each of the five storage periods were incubated on two layers
of Whatman No.1 filter paper moistened with 2.5 mL of
distilled water in 9 cm diameter Petri dishes. Seeds were
incubated at daily (12/12 h) temperature regimes of 5/2,
15/2, 20/10, 25/15 and 30/15 8C in light (12 h of
100 mmol m22 s21, 400– 700 nm, cool white fluorescent
light each day) or in constant darkness (Petri dishes with
seeds in them placed in light-proof black bags) for 14 d. The
four higher temperature regimes represent the mean daily
maximum and minimum monthly temperatures in the vicinity
of Urümqi during the growing season: 15/2 8C (March, early
April and November), 20/10 8C (late April and October), 25/
15 8C (May and September) and 30/15 8C (June, July and
August). The 5/2 8C temperature regime was used to test the
effect of a cold-stratifying temperature regime on seed germination (Baskin and Baskin, 1998). A seed was considered to be
germinated when the radicle had emerged. Germination in
light was examined daily for 14 d; germinated seeds were
removed at each counting. Seeds incubated in darkness were
checked only after 14 d; therefore, they were not exposed to
any light during the incubation period.
After the germination trials were complete, the nongerminated seeds were tested for viability. Seeds were cut
open and the embryo observed. Seeds with white and firm
embryos were counted as viable, and those with tan, soft
embryos were considered non-viable and excluded from the
calculations of germination percentages. Only a very few
seeds were non-viable. The tests of fresh seeds (0 months
old) were initiated on 3 June 2008, using seeds collected on
1 June 2008.
Effect of storage in soil in the field on germination. After collection on 2 June 2008, approx. 1000 freshly mature seeds from
upper siliques and 300 whole intact lower siliques were
enclosed in 12 fine-mesh nylon bags. Each bag was buried at
a depth of 5 cm in plastic pots (18 cm deep, 21 cm in diameter) with drainage holes at the bottom and filled with a
mixture of 75 % habitat soil and 25 % sand. The pots were
buried with the top of the pot level with the soil surface in
the experimental garden on the campus of Xinjiang
Agricultural University. Seeds from upper siliques and the
whole intact lower siliques were subjected to natural temperature and soil moisture conditions; temperature and rainfall data
were recorded at a weather station about 30 m from the buried
seeds and siliques.
Germination tests of fresh seeds from upper siliques, seeds
mechanically removed from lower siliques (by carefully
cutting them open with a knife) and whole intact lower siliques
with seeds inside (0 months old) were tested on 3 June 2008.
Except for months with a snow cover on the ground
(December 2008 to February 2009), one buried pot containing
seeds from upper siliques and intact lower siliques was randomly selected and retrieved at monthly intervals, starting on
2 July 2008 (seeds buried for 1 month) and ending on 2
June 2009 (12 months). After pots were retrieved, seeds
from upper siliques, seeds mechanically removed from lower
siliques and those inside intact whole lower siliques were
tested for germination. However, only data for fresh seeds
and for those buried for 4 months (October) and 10 months
(April) are presented. These three sets of data are sufficient
to tell our ‘story’. Seeds and whole intact lower siliques
were incubated in 9 cm diameter plastic Petri dishes, and
three replicates of 50 seeds from the upper and lower siliques,
and five replicates of five whole intact lower siliques were
placed in light at each of the five temperature regimes, as previously described, for 2 weeks. Germination was examined
daily for 2 weeks, and germinated seeds were removed at
each counting. Seeds and whole, intact lower siliques incubated in darkness were checked only after 2 weeks, at the
end of the experiment; therefore, they were not exposed to
any light during the incubation period. After the germination
trials were complete, the non-germinated seeds were tested
for viability. Seeds were cut open and the embryo observed.
Seeds with white and firm embryos were counted as viable,
and those with tan, soft embryos were considered non-viable
and excluded from the calculations of germination percentages. More than 98 % of the seeds were viable.
Germination phenology. The purpose of this experiment was to
compare the germination phenology of seeds of D. strictus in
the field under supplemented and non-supplemented (natural)
soil moisture conditions. Seeds from upper and lower siliques
and intact lower siliques collected in June 2008 were sown on
bare soil in plots (1.5 0.8 m) on 23 August 2008. There were
six treatments, each consisting of three replications of 200
seeds from upper and lower siliques or 200 lower siliques,
for a total of 18 plots. Seeds from upper siliques, seeds from
lower siliques and whole, intact lower siliques were watered
and not watered.
The experiment was carried out in the experimental garden
on the campus of Xinjiang Agricultural University, Urümqi,
China. In the watered treatment, the soil was watered to field
capacity every 3 d throughout the experiment, except during
the winter when the soil was frozen, while in the non-watered
treatment the soil received water only via precipitation or
snowmelt. At 7 d intervals, from August 2008 to May 2009,
germinated seeds (seedlings) were counted and marked.
Survivorship of marked seedlings was monitored until plants
died, either before or after reproduction. Information on
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
temperature and precipitation at the study site was obtained
from data collected at the Xinjiang Agricultural University
weather station near the study plots.
To compare the spring germination phenology of D. strictus
seeds further, three replications of 50 seeds of each of the two
morphs collected in June 2008 and stored in the laboratory
were sown on 20 March 2009 in plastic pots 18 cm deep and
21 cm in diameter (with drainage holes in the bottom)
filled with soil from the natural habitat of D. strictus. Sown
seeds were exposed to near-natural temperatures in a
non-temperature-controlled metal framehouse (top covered,
with plastic, only when it rained) in the experimental garden
of Xinjiang Agricultural University and monitored for germination. Soil was watered daily, and therefore it remained at
or near field capacity throughout the experiment. Seeds were
monitored for germination (emerged radicle) and seedlings
removed from the pots daily until 29 April 2009, when no germinants had appeared for 2 weeks.
Effect of pericarp and mucilage on germination.
(a) Seeds from upper siliques.
(1) Control: seeds were left intact.
(2) Seeds without mucilage were used to determine whether
the mucilage inhibited germination, either mechanically
or chemically.
(3) Fifty seeds with mucilage removed, along with the mucilage removed from the 50 seeds, were placed in each of
three Petri dishes to test whether soluble chemicals that
would inhibit germination might be leached from the
mucilage.
(b) Seeds from lower siliques.
(1) Control: unmanipulated dispersal units (seeds with pericarps) were placed in Petri dishes.
(2) Segments of dispersal units were placed in Petri dishes.
(3) Mucilaginous seeds without pericarps were placed in Petri
dishes to determine whether the pericarps inhibited germination, either mechanically or chemically.
(4) Mucilaginous seeds and pericarps were placed together in
each Petri dish to test whether soluble chemicals that
would inhibit germination might be leached from the
pericarps.
(5) Seeds without mucilage were placed in Petri dishes to
determine whether the mucilage inhibited germination,
either mechanically or chemically.
(6) Seeds without mucilage and mucilage from the 50 seeds
were placed together in Petri dishes to test whether
soluble chemicals that would inhibit germination might
be leached from the mucilage.
Experiments with seeds from both upper and lower siliques
were conducted at 15/2 8C in light using fresh seeds (0 months
old) and seeds that had been stored dry in the laboratory for 1,
2 and 3 months after they were collected in June 2008. Three
replicates of 50 seeds for each morph were used for each treatment. Germinated seeds were counted daily for 14 d, and germination percentages were calculated from the data.
1003
Imbibition by the indehiscent lower siliques and of seeds within
them. Fresh, intact siliques (15 replications), segments of
fresh siliques (15 replications), 1-year-old intact siliques collected from field plots (five replications) and segments of
1-year-old siliques collected from field plots (15 replicates)
were placed on Whatman No.1 filter paper moistened with distilled water in 9 cm diameter Petri dishes and kept on a laboratory bench under room conditions (22 – 25 8C, 30 % RH). At
time 0 and after 1, 2, 3, 4 and 5 d, the siliques and segments
were removed from the dishes, blotted dry with filter paper,
weighed and returned to the Petri dishes (i.e. the moist conditions). The percentage increase in fresh mass (Wr) of siliques
and segments was calculated as Wr ¼ [(Wi 2 Wd)/Wd] 100,
where Wi and Wd are the mass of imbibed and air-dried silique
(or segment), respectively. After siliques and segments were
fully imbibed, they were cut open and seeds removed. Seeds
of each silique and segment were weighed, allowed to redry
at room temperature for 24 h and reweighed. Also, the dry
mass of non-imbibed seeds that had been stored dry in the laboratory for 1 year was determined (n ¼ 30) and compared with
the mass of seeds allowed to redry after siliques and segments
were fully imbibed.
Data analysis
All data were analysed for normality and homogeneity of
variance prior to analysis to fulfil requirements of t-tests and
one-way analysis of variance (ANOVA). Percentage data
were arcsine transformed before statistical analysis to ensure
homogeneity of variance (non-transformed data appear in all
tables and figures). If data were normal and homogeneous,
they were subjected to further analysis. If data exhibited nonnormal distribution or if variances were not homogeneous,
they were log10 or square root transformed before analysis to
ensure homogeneity of variance. In cases where the ANOVA
assumptions continued to be violated following data transformation, treatment differences were assessed by using the
more conservative Kruskal – Wallis non-parametric test.
Paired sample t-tests were used to compare the increase in
seed mass by water absorption and by adherence to soil particles. When variances of data were homogeneous, ANOVA
was used to determine differences among dimorphic siliques
and seeds in morphology and mass, mass of water absorbed
and adherence to soil particles, rates of hydration and dehydration, germination percentage, mass and time of imbibition
by the indehiscent lower siliques, mass of seeds within
imbibed lower siliques and mass of non-imbibed seeds
stored in the laboratory for 1 year. When the variance of logarithmically transformed data was still not homogenous,
differences among dimorphic siliques and seeds in these
characteristics were determined by the Kruskal –Wallis test.
Four-way ANOVA was used to test the significance of main
effects (light condition, storage time, temperature and seed
type) and their interaction on germination in the ‘Effect of
dry storage (after-ripening) on germination’ experiment and
to test the significance of main effects (light condition, retrieval time, temperature and seed type) and their interaction on
germination in the ‘Effect of storage in soil in the field on germination’ experiment. Tukey’s HSD test was performed for
multiple comparisons to determine significant (P , 0.05)
1004
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
differences among dimorphic seeds and treatments. Statistical
tests were conducted at P ¼ 0.05. Regression analyses were
used to determine the relationship between dispersal unit and
fall time and between dispersal unit and fall rate. All data analyses were performed with the software SPSS 13.0 (SPSS Inc,
Chicago, IlL, USA). Values are means + 1 s.e. (Sokal and
Rohlf, 1995).
TA B L E 1. Comparison of the morphology and mass of dimorphic
siliques and seeds of Diptychocarpus strictus (mean + s.e.)
Silique
Morphology (mm)
RES ULT S
The results of this study generally provide support for our
hypothesis that fruit and seed heteromorphism in D. strictus
aids in the distribution of this species in space and time in
its cold desert habitat. Thus, we have shown that there is a
high degree of difference in morphology and anatomy of
both seeds and siliques of this species that is correlated with
differences in timing of germination and in capacity for dispersal between seeds in upper and lower siliques. Namely, dehiscence of the thin-walled upper siliques and their winged seeds
allows for dispersal in space, whereas the thick-walled indehiscent lower siliques prevent the nearly wingless nonmucilaginous seeds inside them from germinating for .1
year, thus forming a persistent seed bank of unknown length.
In spite of the dimorphism in shape, size and amount of mucilage between seeds in upper and lower siliques, they do not
differ physiologically, i.e. they are not physiologically
dimorphic. The thick mucilage layer on the winged seeds in
the upper siliques aids in lodgment on the soil surface and
increases the time for which they are hydrated, thereby increasing the chances for germination once the seeds have been dispersed from the mother plant.
Characteristics and dispersal of siliques and seeds
Colour, shape, size and mass. The upper siliques are elongated
(i.e. 46.4 mm in length and 3.5 mm in width) and compressed
(Fig. 1; Table 1), while the lower siliques are relatively shorter,
about 33.0 mm in length and 3.0 mm in width, approximately
columnar and highly lignified (Fig. 1; Table 1). Moreover, the
mass of upper siliques, seeds and pericarps was significantly
greater than that of lower siliques, seeds and pericarps
(Table 1). Both upper and lower siliques are khaki coloured
and have a peduncle and a beak.
The colour and shape of seeds from the upper and lower siliques are similar, i.e. dark brown and oval to ovate and slightly
compressed. The mean size of seeds from the upper and lower
siliques is 2.7 2.0 mm and 2.7 1.8 mm, respectively. The
hilum is obvious at one end of the seed, and there is a transparent wing around the edge of the seed (Fig. 1). However, the
wing of seeds from upper siliques is obviously wider than
that of seeds from lower siliques (Fig. 1; Table 1). When
fully hydrated, a transparent, gelatinous coating of mucilage
develops within a few minutes (Fig. 1), but mucilage on
seeds from upper siliques is wider than that of mucilage
on seeds from lower siliques. Also, the mean mass of 1000
seeds from upper siliques (about 1.8 g) is greater than that of
1000 seeds from lower siliques (1.4 g) (Table 1).
Anatomy. The general structure of the pericarp in the two types
of siliques is similar and mainly consists of three layers (i.e.
Mass (g)
No. of seeds per
silique
Seed
Morphology (mm)
Mass (g) of 1000
seeds
Length of
peduncle
Length of beak
Length of
silique
Width of
silique
Per silique
Pericarp
Seeds per
silique
Length of seed
Width of seed
Width of wing
Upper silique
Lower silique
3.414 + 0.012b
2.908 + 0.006a
4.990 + 0.019b
46.400 + 0.132b
6.143 + 0.022a
32.977 + 0.147a
3.543 + 0.008b
3.002 + 0.013a
0.142 + 0.004b
0.087 + 0.003b
0.055 + 0.001b
0.064 + 0.002a
0.051 + 0.002a
0.013 + 0.001a
31.700 + 0.666b
11.267 + 0.783a
2.674 + 0.046a
2.047 + 0.042b
0.085 + 0.032b
1.844 + 0.024b
2.658 + 0.042a
1.763 + 0.046a
0.025 + 0.005a
1.370 + 0.018a
Different letters within a row indicate significant differences (Tukey’s
HSD, P ¼ 0.05).
epicarp, mesocarp and endocarp), which are fused together
(Fig. 2A –D). The epicarp consists of only a single layer of
small rectangular epidermal cells, which are closely arranged.
The outer periclinal wall of epidermal cells is thickened by a
covering of cutin. The mesocarp has two distinctive cell
layers, i.e. palisade cells and large parenchyma cells. Cells
in the endocarp are much smaller than those in the epicarp,
and they usually are irregularly polygonal in shape and
closely arranged.
The anatomical structure of the pericarp, however, differs
greatly between the two types of siliques. In cross-section,
the upper silique is radially elongated to oval, and its length
is several times greater than its width (Fig. 2A), while the
lower silique is approximately round in cross-section
(Fig. 2C). The pericarp of lower siliques is nearly 1.5 times
thicker than that of upper siliques. Moreover, the degree of
thickening of the cell wall in pericarp differs between upper
and lower siliques (Fig. 2B, D). The cell wall of the endocarp
of upper siliques is thicker than that of lower siliques, but the
cell wall of palisade tissue in the mesocarp of lower siliques is
thicker than that of upper siliques. In addition, orientation of
parenchyma cells has a tangential arrangement in mesocarp
of upper siliques, while it is radial in lower siliques
(Fig. 2B, D).
Episperm and endopleura are the outermost cell layers of the
seed coat in both silique morphs (Fig. 2E – H). The episperm is
surrounded by a mucilaginous layer, which adheres tightly to
the surface of the seed, and the endopleura, which ruptures,
has two layers of cells, the inner of which presumably contains
tannin and the outer contains mucilage. The thickness of the
seed coat is much greater for seeds from lower siliques than
seeds from upper siliques (Fig. 2F, H). Conversely, the mucilaginous layer on the episperm is much thicker and better
defined on seeds from upper siliques than on those from
lower siliques.
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
A
ep
me
B ep
me
en
en
C
D
pal
1005
me
ep
ep
en
pal
me
par
par
en
E
F
G
epi
epi
end
H
epi
end
end
end
mc
mc
ta
ta
emb
epi
emb
emb
emb
F I G . 2. Cross-section of the two types of siliques and seeds in Diptychocarpus strictus. (A, B) Upper silique (A, cross-section, scale bar ¼ 200 mm; B, crosssection of pericarp, scale bar ¼ 50 mm). (C, D) Lower silique (C, cross-section, scale bar ¼ 200 mm; D, cross-section of pericarp, scale bar ¼ 50 mm). (E, F)
Seed from upper silique [E, whole section, scale bar ¼ 200 mm; F, seed coat (labelled epi, end, mc, ta), scale bar ¼ 50 mm]. (G, H) Seed from lower silique
[G, whole section, scale bar ¼ 200 mm; H, seed coat (labelled epi, end, mc, ta), scale bar ¼ 50 mm]. emb, embryo; en, endocarp; end, endopleura; ep,
epicarp; epi, episperm; mc, mucilaginous cells in episperm; me, mesocarp; pal, palisade tissue in mesocarp; par, parenchyma in mesocarp; ta, (presumably)
tannin in endopleura. The cross-sections of siliques (A, C) include one seed sliced through its centre and just a sliver of the next seed in line.
TA B L E 2. Fall time (s, mean + s.e.) and effect of wind speed on
dispersal distance (cm, mean + s.e.) of dispersal units of
Diptychocarpus strictus
Dispersal units
Fall time (s)
Fall rate (cm s21)
Dispersal distance (cm)
Wind speed 1 m s21
Wind speed 4 m s21
Seeds from upper
siliques
Whole, intact lower
siliques
0.523 + 0.006b
230.376 + 4.327a
0.377 + 0.013a
334.720 + 19.793b
25.228 + 1.244b
57.772 + 2.486b
6.018 + 0.534a
16.842 + 1.367a
Different letters within a row indicate significant differences (Tukey’s
HSD, P ¼ 0.05).
width of seeds without mucilage (mucilage removed) from
upper siliques. In imbibed seeds from lower siliques, there
was no difference in seeds with mucilage vs. those without
mucilage vs. dry seeds (Table 3). After they were fully
imbibed, seeds with mucilage from upper siliques increased
in mass by 30.118 times (3011.8 %) in comparison with the
mass of the dry seeds before wetting. On the other hand, the
increase in mass of seeds without mucilage from upper siliques
was only 1.921 times (192.1 %). However, the imbibed seeds
with and without mucilage from lower siliques increased in
mass by only 1.9 – 2.2 times in comparison with the original
mass of dry seeds. In imbibed seeds from upper siliques, mucilage accounted for .50 % of the width and . 90 % of the total
seed mass. However, for seeds in the lower siliques mucilage
accounted for only 5 % of the width and about 20 % of the
total mass of the seeds (Table 3).
Dispersal
In still air, seeds from upper siliques and whole lower siliques differed significantly in landing time and fall rate
(Table 2; P , 0.01), the seeds from upper siliques taking
longer to fall a specific distance. This suggests that the dispersal ability of the seeds from upper siliques will be much
greater than that of the whole lower siliques. The dispersal
unit was correlated significantly with fall time (r ¼ 0.92,
P , 0.01) and fall rate (r ¼ 0.83, P , 0.01). Dispersal distances of seeds from upper siliques in a parallel stream of
air were usually considerably greater than those of whole
lower siliques at wind speeds of both 1 and 4 m s21 (Table 2).
Seed mucilage
Change in width and mass of imbibed seeds. After imbibition,
the width of seeds with mucilage from upper siliques increased
more than 2-fold, whereas there was no obvious increase in
Water absorption and dehydration. The process of water absorption by the two seed morphs with and without mucilage was
similar and can be divided into three stages: (1) the rate of
water absorption was rapid; (2) the rate of water absorption
decreased; and (3) water absorption stopped. Seeds were
fully imbibed after 240 – 300 min (Fig. 3). Intact seeds from
upper siliques had a high capacity to take up water, and
their mass increased 1800 % (from 0.2 to 3.6 g) within only
30 min from the start of imbibition, which differed significantly from seeds without mucilage (P , 0.01). Thus, water
absorbed by seeds alone contributed little to the total amount
of water absorbed by intact seeds from upper siliques. Since
seeds with mucilage in the lower fruits had little mucilage,
they absorbed only 0.14 g of water in 30 min.
The dehydration process was almost the reverse of the water
absorption process and can also be divided into three stages:
(1) the loss of water was rapid; (2) the rate of dehydration
decreased; and (3) seeds had reached their original mass
1006
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
TA B L E 3. Width (mm, mean + s.e.) of Diptychocarpus strictus
seeds with and without mucilage and mass (g, mean + s.e.) of
dry and of imbibed seeds with and without mucilage and of only
mucilage
Mucilage
Mass (g) of 100 seeds
Seeds with mucilage
Seeds without mucilage
Mucilage
Mucilage/seed
Wd
Wi
Wo
Wm
Wm0
2.047 + 0.042Ab
4.420 + 0.056Bb
2.055 + 0.023Ab
1.183 + 0.035b
0.532 + 0.010b
1.763 + 0.046Aa
1.890 + 0.034Ba
1.785 + 0.031ABa
0.053 + 0.009a
0.054 + 0.009a
Md
Mi
Mt
Md
Mi
Mt
Md
Mi
Mt
Mmb
Mma
Ab
0.175 + 0.001
5.278 + 0.084Bb
30.118 + 0.469b
0.125 + 0.001Aa
0.240 + 0.008Ba
1.921 + 0.070a
0.046 + 0.003Ab
4.057 + 0.054Bb
92.654 + 6.735b
0.248 + 0.016b
0.944 + 0.002b
Mass (g)
Lower silique
(3)
4·0
AW
DW
AWO
DWO
(2)
3·0
(1)
2·0
1·0
0·0
0
240
0.124 + 0.011
0.272 + 0.005Ba
2.193 + 0.042a
0.128 + 0.002Aa
0.239 + 0.012Ba
1.869 + 0.080a
0.006 + 0.001Aa
0.055 + 0.010Ba
9.446 + 3.193a
0.045 + 0.010a
0.185 + 0.030a
Different letters indicate significant differences. Lower case letters
compare the two morphs of seeds and upper case letters different treatments
within an index (Tukey’s HSD, P ¼ 0.05). Wd, width of dry seeds; Wi and
Wo, widths of imbibed seeds with and without mucilage, respectively; Wm,
mucilage width; Wm0 , mucilage width ratio of imbibed seeds; Md, dry mass
of seeds and mucilage; Mi, mass of imbibed seeds with and without
mucilage and mucilage; Mt, number of times mass increased after imbibition;
proportion of mucilage to seed in mass before (Mmb) and after (Mma)
imbibition.
(Fig. 3). Seeds with mucilage from upper siliques of both
morphs lost their absorbed water more slowly than those
without mucilage. Seeds with mucilage from upper siliques
of the two morphs retained .50 % of their water for
360 min and did not return to their original masses until
1200 min, whereas the seeds without mucilage became fully
dehydrated (i.e. had reached the same moisture level as that
of seeds at the beginning of absorption) after 240 min. After
240 min, however, seeds with and without mucilage from
lower siliques had almost reached their original masses.
Mass of seeds with and without mucilage and with adhered
soil particles.
(a) Adherence of wet soil particles to dry seeds with and
without mucilage. The original mean mass of seeds with
and without mucilage from upper siliques was 0.18 and
0.13 g, and that of seeds with and without mucilage
from lower siliques was 0.14 and 0.13 g, respectively.
When dry seeds with mucilage from upper and lower siliques were placed on wet soil particles, a mucilaginous
layer formed after a few minutes at the contact area
between the seed and the wet soil surface on the lower
side of the seed. However, mucilage had not yet covered
the upper part of the seed; thus, the upper side of the
seed was still dry. The mass of seeds from upper siliques
with mucilage and adhered soil particles was significantly
greater than that of seeds from upper siliques without
mucilage (and soil particles) and of those from lower
480
720
960
1200
Time since wetting (min)
Aa
0
240
480
720
960
1200
Time since drying (min)
B
0·30
0·25
Mass (g)
Width (mm)
Seeds
Upper silique
A
5·0
0·20
0·15
(1)
0·10
0·05
0·00
(2)
0
120
240
360
(3)
480
600 720
Time since wetting (min)
0
120
240
360
720
Time since drying (min)
F I G . 3. Time course for water absorption and dehydration of seeds with mucilage and for Diptychocarpus strictus seeds without mucilage. (A) Upper
silique. (B) Lower silique. AW, water absorption of seeds with mucilage;
DW, dehydration of seeds with mucilage; AWO, water absorption of seeds
without mucilage; DWO, dehydration of seeds without mucilage. (1), (2)
and (3) indicate three stages of the imbibition and dehydration process (see
text for explanation). Bars are + s.e.
siliques (P , 0.01). Compared with its original mass,
however, the mass of seeds of the lower morph increased
very little (Table 4).
(b) Adherence of dry soil particles to imbibed seeds with and
without mucilage. Fully hydrated seeds with mucilage
from upper siliques adhered to many soil particles, resulting in an increase of more than seven times the total mass
of dry seeds. On the other hand, soil particles did not
adhere to imbibed seeds without mucilage from upper siliques or to those from lower siliques with or without
mucilage (Table 4).
Germination ecophysiology
Effect of dry storage (after-ripening) on germination. A four-way
ANOVA showed that germination was significantly affected by
light condition (P , 0.01), storage time (P , 0.01), temperature (P , 0.01) and seed type (P , 0.01) (Table 5). No significant interactions were observed in germination percentage
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
1007
TA B L E 4. Comparison of mass (mean + s.e.) of Diptychocarpus strictus seeds with and without mucilage and with adhered soil
particles after maximum water absorption and of dry seeds adhered to wet soil
Upper silique
With mucilage
Lower silique
Without mucilage
Seed mass (g) with adhered soil particles after maximum water absorption
Md0
0.182 + 0.001Ac
0.131 + 0.010Aa
Ms0
1.291 + 0.035Bb
0.133 + 0.003Aa
Mt0
7.106 + 0.085b
1.023 + 0.036a
Mass (g) of dry seeds with adhered wet soil particles
Md0
0.173 + 0.004Ac
0.134 + 0.002Ab
Ms0
0.752 + 0.011Bc
0.137 + 0.001Ab
Mt0
4.353 + 0.083b
1.022 + 0.014a
With mucilage
Without mucilage
0.142 + 0.001Ab
0.143 + 0.019Aa
1.005 + 0.065a
0.134 + 0.006Aa
0.137 + 0.007Aa
1.021 + 0.013a
0.143 + 0.003Ab
0.144 + 0.003Ab
1.008 + 0.006a
0.121 + 0.002Aa
0.124 + 0.001Aa
1.026 + 0.013a
Different lower case letters within a row for before water absorption, after water absorption and times increased in mass indicate significant differences
(Tukey’s HSD, P ¼ 0.05), and differences between before water absorption and after water absorption are compared by t-test using the upper case letters. Md0 ,
mass of dry seeds; Ms0 , mass of imbibed seeds plus sand particles; Mt0 , number of times seed mass increased after adherence of sand particles.
TA B L E 5. Four-way ANOVA of effects of seed type, temperature, light condition, storage time and their interactions on germination
of Diptychocarpus strictus seeds stored dry under laboratory conditions
Source
d.f.
SS
MS
F-value
P-value
Light (L)
Storage time (T)
Temperature (T0 )
Seed type (S)
LT
L T0
LS
T T0
TS
T0 S
L T T0
LTS
L T0 S
T T0 S
L T T0 S
1
4
4
1
4
4
1
16
4
4
16
4
4
16
16
367.413
296308.453
42204.987
5737.813
255.653
196.453
12.000
30151.547
2061.520
1689.253
270.480
23.867
70.267
2597.413
226.533
367.413
74077.113
10551.247
5737.813
63.913
49.113
12.000
1884.472
515.380
422.313
16.905
5.967
17.567
162.338
14.158
32.496
6551.631
933.188
507.472
5.653
4.344
1.061
166.669
45.582
37.351
1.495
0.528
1.554
14.358
1.252
,0.01
,0.01
,0.01
,0.01
,0.01
0.002
0.304
,0.01
,0.01
,0.01
0.104
0.715
0.188
,0.01
0.232
between light and seed type (P ¼ 0.30), between light, storage
time and temperature (P ¼ 0.10), between light, storage time
and seed type (P ¼ 0.72) or between light, temperature and
seed type (P ¼ 0.19), or among the interaction of the four
factors (P ¼ 0.23) (Table 5). All freshly harvested seeds were
dormant, and thus no seeds germinated at any of the five temperature regimes in either light or constant darkness. Dormant
seeds from the two silique types gradually after-ripened during
storage. After 3 months of dry storage, germination was about
95 % in light and 90 % in darkness at 15/2 and 20/10 8C
(Fig. 4). Seeds after-ripened further between 3 and 12 months
of dry storage, and 12-month-old seeds germinated to .80 %
at all five temperature regimes in both light and darkness.
Effect of storage in soil in the field on germination. A four-way
ANOVA showed that germination was significantly affected
by light condition (P , 0.01), retrieval time (P , 0.01), temperature (P , 0.01) and seed type (P , 0.01) (Table 6). Also,
significant interactions were observed in germination percentage between retrieval time and temperature (P , 0.01),
between retrieval time and seed type (P , 0.01), between
temperature and seed type (P , 0.01) and between retrieval
time, temperature and seed type (P , 0.01) (Table 6). In
June 2008, all freshly harvested seeds were dormant, and
thus no seeds germinated at any of the five temperature
regimes in either light or constant darkness. Dormant seeds
from the two silique types gradually after-ripened during
storage in the field (Fig. 5). After 4 months of storage, i.e.
October 2008, germination was about 47– 87 % in light and
43– 85 % in darkness at 15/2 and 5/2 8C. Seeds did not go
back into dormancy during winter, and 10-month-old seeds
(October 2008) germinated to about 47– 85 % in light and
43– 87 % in darkness at 15/2 and 5/2 8C in spring (April
2009) (Fig. 5). The germination percentage of whole intact
lower siliques was 0 % in all test conditions at time 0 and at
the two retrieval times. They also did not germinate when
retrieved after 12 months of burial (J. J. Lu, unpubl. res.).
Germination phenology. Germination of seeds from upper and
lower siliques on wet soil began between 6 and 13 September,
when mean daily maximum and minimum air temperatures
were 24.3 and 11.7 8C, respectively (Fig. 6A, B). Essentially
all seeds that germinated in autumn did so in September. We
did not observe germination of any seeds from upper or lower
siliques on dry soil or in intact lower siliques on either wet or
dry soil in autumn. In spring 2009, the period in which the
1008
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
Germination (%)
A
Light/Dark 1 month of storage
60
50
40
B
Dark
5/2 °C
15/2 °C
b b b
c
30
20/10 °C
25/15 °C
30/15 °C
c
b
20
aa
10
b bb
c c
aa
aa
b
aa
Germination (%)
Germination (%)
0
80
70
60
50
40
30
20
10
0
100
90
80
70
60
50
40
30
20
10
0
2 months of storage
d
d
c
c
c
c
b
b
a
b
b
a
b
b
3 months of storage
c c
cc
a
c c
b
bb
a
a
bb
b
a
a
b
ab
a
12 months of storage
bbb
b
100 ab
b
a
a
90
aa
80
Germination (%)
c
aa
a
b bb
b
aa
bb
a
aa
a
70
60
50
40
30
20
10
0
Upper
Lower
Seed type
Upper
Lower
Seed type
F I G . 4. Final germination percentages (mean + s.e.) of the two types of seeds in Diptychocarpus strictus (A) incubated in the light/darkness regime (left panels)
and (B) in constant darkness (right panels) at five temperature regimes following 1, 2, 3 and 12 months of dry storage under laboratory conditions. Germination
percentages of fresh (0-months-old) seeds from upper and lower siliques were 0 % in both light/darkness and constant darkness at all temperature regimes. Bars
with different letters are significantly different in multiple range comparison within the upper and within lower silique (Tukey’s HSD, P ¼ 0.05).
greatest number of seeds from upper and lower siliques on wet
and dry soil germinated was 3 – 28 March, when mean daily
maximum and minimum air temperatures were 7.3 and 22.5
8C, respectively. The germination percentage from upper and
lower siliques on dry soil was lower than it was on wet soil.
For example, of a total of 600 seeds sown on wet and dry
soil, 210 and 117 of them from the upper silique germinated
on wet and dry soil, respectively, and 101 and 36 from lower
siliques germinated on wet and dry soil, respectively. There
was no further germination on either wet or dry soil after 25
April 2009. On the other hand, no seeds inside the intact
lower siliques germinated on either wet or dry soil. Plants
from 100 % of the 41 seeds that germinated in the six
treatments in autumn 2008 reproduced in spring 2009, and
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
1009
TA B L E 6. Four-way ANOVA of effects of seed type, temperature, light condition, retrieval time and their interactions on seed
germination of Diptychocarpus strictus stored in soil in the field
Source
d.f.
SS
MS
F-value
P-value
Light (L)
Retrieval time (R)
Temperature (T)
Seed type (S)
LR
LT
LS
RT
RS
TS
LRT
LRS
LTS
RTS
STLR
1
9
4
1
9
4
1
36
9
4
36
9
4
36
36
115.282
150107.535
129186.993
5871.882
69.668
6.727
3.375
72929.540
4880.402
8415.060
302.073
119.175
15.167
7995.073
213.367
115.282
16678.615
32296.748
5871.882
7.741
1.682
3.375
2025.821
542.267
2103.765
8.391
13.242
3.792
222.085
5.927
5.289
765.250
1481.842
269.414
0.355
0.077
0.155
92.949
24.880
96.525
0.385
0.608
0.174
10.190
0.272
0.022
,0.01
,0.01
,0.01
0.955
0.989
0.694
,0.01
,0.01
,0.01
1.000
0.791
0.952
,0.01
1.000
A
Germination (%)
100
80
60
40
B Dark
Light/Dark
4 months of storage (October)
c
5/2 °C
c
c c
15/2 °C
c
20/10 °C
b
25/15 °C
b
b
b
30/15 °C
20
aa
aa
c
b
b
aa
aa
0
10 months of storage (April)
c
c
c
80
Germination (%)
100
60
b
b
c c
c
b
b
b
b
40
20
a
aa
a
aa
aa
0
Upper
Lower
Seed type
Upper
(range 3.2– 15.5 8C) and 20.6 8C (range 25.8 to 2.9 8C),
respectively (Fig. 6A, C). Most of the germination of seeds
from upper siliques occurred from day 9, when mean
maximum and minimum daily temperatures in the metal framehouse averaged 18.9 and 5.8 8C, respectively. The highest
germination of seeds from lower siliques occurred from day
12, when the mean maximum and minimum daily temperatures in the metal framehouse averaged 15.3 and 6.1 8C,
respectively. There was no further germination of seeds from
upper siliques after 9 April (i.e. 20 d after sowing) and from
lower siliques after 14 April (i.e. 25 d after sowing). Seeds
from upper siliques germinated faster than those from lower
siliques, reaching 50 % germination after 10 and 15 d, respectively. Also, total germination was higher (88 %) for seeds from
upper siliques than for those from lower siliques (77 %).
Lower
Seed type
F I G . 5. Final germination percentages (mean + s.e.) of the two types of seeds
in Diptychocarpus strictus (A) incubated in the light/darkness regime (left
panels) and (B) in constant darkness (right panels) at five temperature
regimes following 4 and 10 months of storage in soil in field conditions.
Germination percentages of fresh (0-months-old) seeds from upper and
lower siliques were 0 % in both light/darkness and constant darkness at all
temperature regimes. No seeds inside fresh or buried whole intact lower siliques germinated in any of the tests. Bars with different letters are significantly
different in multiple range comparison within upper and within lower silique
(Tukey’s HSD, P ¼ 0.05).
97– 100 % (depending on treatment) of the plants from the 464
seeds that germinated in spring 2009 reproduced in spring
2009.
When 9-month-old laboratory-stored seeds were sown on
soil at natural temperatures in March 2009, very little or no
germination occurred during the first week after sowing,
even on soil kept at or near field capacity, when the mean
maximum and minimum daily temperatures averaged 9.1 8C
Effect of pericarp and mucilage on germination. None of the
untreated freshly matured seeds (with mucilage) from upper
siliques germinated. Furthermore, seeds without mucilage
and those without mucilage with added mucilage germinated
to ,10 % (Fig. 7). After storage for 2 months, seeds without
mucilage from upper siliques germinated to .95 %. When
seeds without mucilage were placed in close contact with
mucilage, their germination was not inhibited (Fig. 7).
Germination percentages of treated seeds from upper siliques
were significantly higher than that of those of complete dispersal units, i.e. seeds with mucilage (P , 0.01). However, there
was no difference between the two seed treatments. After
storage for 3 months, there was no significant difference
among the intact dispersal units and treated seeds (P ¼ 0.39).
None of the untreated freshly matured seeds from the lower
siliques germinated. Removing the fruit coat and mucilage did
not increase the germination percentage of seeds from lower
siliques, which was still 0 %. Storage influenced germination
of seeds from lower siliques in a similar way to seeds from
upper siliques. The germination of the dispersal unit (i.e.
intact silique) was still 0 % in 3-month-old intact dispersal
units. However, removing the fruit coat was very effective in
promoting germination, and there were no significant differences in germination among seeds with mucilage, seeds with
mucilage plus pericarp, seeds without mucilage or seeds
1010
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
Germination (%)
Temperature (°C)
60
20
50
10
40
30
0
20
Precipitation (mm)
Min. temperature
–10
100
90
80
70
60
50
40
30
20
10
0
100
90
80
70
60
50
40
30
20
10
0
10
0
–20
J J
2008
A
S
O
N
D
J F
2009
M
A
M J
Month
Seeds from upper silique on wet soil
Seeds from upper silique on dry soil
Seeds from lower silique on wet soil
Seeds from lower silique on dry soil
Intact lower siliques on wet soil
Intact lower siliques on dry soil
B
40
35
Germination (%)
100
90
80
70
60
50
40
30
20
10
0
70
Max. temperature
30
100
90
80
70
60
50
40
30
20
10
0
Germination (%)
Precipitation
Germination (%)
80
A
Germination (%)
40
30
25
20
15
10
5
0
23 30 6 13 20 27 3 3 14 21 28 4 11 18 25 2 9
Aug. Sep.
Oct. Mar.
Apr.
May
2008
2009
Date
100
C
90
Germination (%)
80
70
60
Seeds from upper silique
Seeds from lower silique
50
40
0 months of storage
Seed with mucilage
Seed without mucilage
Seed without mucilage + Mucilage
1 month of storage
2 months of storage
3 months of storage
0
30
20
2
4
6
8
10
12
14
Time (d)
10
0
0
4
8
12
16
20
24
28
32
36
40
44
Time (d)
F I G . 6. (A) Monthly total precipitation and mean minimum and mean
maximum monthly temperatures at Xinjiang Agricultural University weather
station from June 2008 to June 2009. (B) Germination percentages (mean +
s.e.) of seeds of Diptychocarpus strictus on wet and dry soil in the field
during autumn and spring of 2008–2009 (black circles are below the open
circles). (C) Germination percentage (mean + 1 s.e.) of 9-month-old
laboratory-stored seeds of D. strictus sown in pots in March 2009. Wet soil
means watered, and dry soil means not watered.
without mucilage plus mucilage (.90 % for all treatments,
data not shown).
Imbibition by indehiscent lower siliques and of seeds within them.
Imbibition of water by intact siliques, by silique segments of
F I G . 7. Effects of mucilage on germination percentages (mean + s.e.) of 0-,
1-, 2- and 3-months-old seeds from upper siliques of Diptychocarpus strictus
in the light (12 h photoperiod) at 15/2 8C.
fresh and 1-year-old siliques and by the seeds within them
did not differ in water uptake. Siliques in all treatments
imbibed water readily and followed a typical pattern of rapid
initial water uptake, with whole silique and segment mass
increasing by 136 % and 145 %, respectively, after 1 d; after
5 d their mass had increased by 180 –190 % (data not
shown). Seeds within fresh and 1-year-old siliques and
silique segments also became fully hydrated. There was no significant difference in dry mass of non-imbibed seeds from
fresh and 1-year-old siliques or silique segments. In addition,
after 5 d, when whole siliques and segments of siliques were
fully imbibed, there was no significant difference in mass of
seeds removed from imbibed siliques, increase in the mass
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
1011
TA B L E 7. Imbibition of water (mg, mean + 1 s.e.) by seeds of Diptychocarpus strictus within whole lower siliques and by segments
of lower siliques
Seed from fresh siliques
Treatment
Mass of seed removed from imbibed silique
Mass of redried seeds
Dry mass of non-imbibed seed
Increase in mass per seed (%)
Seed from 1-year-old siliques collected from
the field
Unmanipulated
Segments
Unmanipulated
Segments
2.218 + 0.020Ba
1.301 + 0.011Aa
1.333 + 0.013Aa
70.674 + 2.100a
2.237 + 0.041Ba
1.273 + 0.039Aa
1.267 + 0.039Aa
77.264 + 4.318a
2.143 + 0.070Ba
1.215 + 0.050Aa
1.242 + 0.042Aa
80.176 + 10.389a
2.255 + 0.059Ba
1.236 + 0.046Aa
1.327 + 0.033Aa
84.160 + 3.937a
Different lower case letters within a row and upper case letters within a column indicate significant differences (Tukey’s HSD, P ¼ 0.05).
per seed or mass of redried seeds. Also, there were no significant differences between the dry mass of non-imbibed seeds
and that of seeds from imbibed whole siliques or segments
of siliques after they were redried (Table 7).
D IS C US S IO N
Desert annuals cope with their harsh, variable environment by
escape in space and escape in time. They may escape in space
to new habitats via seed dispersal or escape in time by fractional or delayed germination (Venable and Lawlor, 1980;
Venable, 1985). Heteromorphic fruits or seeds from the
same plant species may represent a combination of the two
opposing strategies of escape in space and in time. The existence of both dimorphic siliques and dimorphic seeds in a
raceme of the inland cold desert annual ephemeral
D. strictus confers differences in the capability for dispersal
and germination of seeds of this species in time and space,
which may be viewed as adaptive bet-hedging in an unpredictable environment.
The passive mode and local dispersal by wind are the two
principal mechanisms of dispersal in D. strictus. After plants
of D. strictus complete their life cycle in late May to early
June, one of the most obvious differences between the two
silique morphs is their seed dispersal ability. The wide wing
of seeds from dehiscent upper siliques decreases their terminal
velocity and thus prolongs their fall time, which increases the
chance of these seeds being carried away from the mother
plant by wind currents. However, seeds in the indehiscent
lower siliques remain within the highly lignified pericarp and
usually are dispersed near the mother plants, which serve
only to leave descendants in safe sites previously occupied
by their parents ( primary dispersal). Apart from the seed
wing, the mass of dispersal units can explain a significant
part of the dispersal ability, an observation that reinforces
the better dispersal ability of the lighter dispersal units
(Meyer and Carlson, 2001). Our results showed that, in
D. strictus, the fall rate of whole lower siliques is faster than
that of seeds from upper siliques. Moreover, seeds from
upper siliques were dispersed a greater distance than whole
lower siliques at wind speeds of both 1 and 4 m s21.
During summer and autumn, secondary dispersal as affected
by mucilage may increase differences in dispersal of the two
types of dispersal units in D. strictus under some environmental conditions, for example rainfall. When rain occurs
and the soil is moist, seeds from upper siliques rapidly
absorb water and develop a thick layer of mucilage, which
aids adherence of the seeds to the soil surface and in settling
in local sites. Although lower siliques are indehiscent, a thin
mucilaginous layer forms on the seeds; the role of this mucilage on the seeds is not clear. Mucilage on seeds from upper
siliques significantly increased dehydration time, but it did
not affect dehydration time of seeds from lower siliques.
Gutterman and Shem-Tov (1997) found that the dehydration
time of mucilaginous seeds was much longer than that of nonmucilaginous seeds. Thus, on the soil surface mucilaginous
seeds from upper siliques should have an advantage over mucilaginous seeds from lower siliques (if they were able to escape
from the silique). Low or non-mucilaginous seeds would take
up less water from wet soil because of poor contact with the
soil particles (Huang et al., 2001).
Moreover, mucilaginous seeds could adhere to the soil
surface, and mucilage could hold soil particles around the
seed, thereby increasing seed mass (Gutterman and
Shem-Tov, 1996; Huang et al., 2001). The more mucilage
the seed has, the more sand particles that can adhere to it.
Since soil particles tenaciously adhere to mucilaginous seeds
from upper siliques, the great increase in mass of these seeds
may prevent them from being further dispersed by wind
from favourable microhabitats (Gutterman, 1993; Huang and
Gutterman, 2004), delay or prevent seed collection by
insects, thereby avoiding seed loss (Gutterman, 1993; Huang
and Gutterman, 2004), or promote settlement of seeds onto
the soil (Huang et al., 2001).
Freshly matured seeds from both silique morphs of
D. strictus are dormant and thus do not germinate over a
range of temperature regimes in either light or darkness, probably because of the low growth potential of the embryo
(Baskin and Baskin, 1998). After 3 – 4 months of after-ripening
under laboratory (Fig. 4) and field (Fig. 5) conditions during
summer, the seeds germinate to high percentages at cool but
not at warm temperatures. After 12 months of after-ripening
under laboratory conditions, seeds had high germination in
both light and darkness over the range of cool to warm temperature regimes at which they were tested. Seeds stored in the
field germinated to 46.7 – 87.3 % in light and 43.3 – 85.3 % in
darkness at 5/2 and 15/2 8C, but to ,14.0 % in light and darkness at 25/15 and 30/15 8C when tested in October. The
optimum temperatures for germination in light and darkness
of seeds stored in the field until April of the following year
were still 15/2 and 5/2 8C. Dormancy break occurs in
summer, but, unlike seeds of temperate zone obligate winter
1012
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
annuals, those of D. strictus are not induced back into dormancy (secondary dormancy) during winter. Although the
seeds can germinate to high percentages at low temperatures
in autumn, in most years they are prevented from doing so
by the low amount of rainfall in autumn (Fig. 6A, B). Thus,
germination of most seeds is delayed until spring when soil
moisture is high due to rainfall and snow melt. Seeds of
D. strictus have non-deep physiological dormancy (Baskin
and Baskin, 2004).
After-ripened seeds of D. strictus can germinate to high percentages in both light/darkness and darkness. Thus, it is likely
that should seeds become covered with soil or litter, they
would not be prevented from germinating. In fact, after a
spring rain soil below the surface or under litter would
remain moist for longer periods of time than that on the soil
surface, and thus burial of seeds in soil or litter would
favour germination. Based on these results, it seems unlikely
that D. strictus could form a large persistent seed bank from
seeds of the upper siliques, which are dispersed soon after
these dehiscent siliques mature.
The pericarp of dehiscent upper siliques of D. strictus is
relatively thin, whereas the indehiscent lower siliques have a
thick pericarp. Thus, the pericarp plays little or no role in controlling germination of seeds from upper siliques, but it physically restricts/prevents germination of seeds within the lower
siliques. However, even when fresh seeds are removed from
the lower silique the presence of some embryo dormancy (or
non-deep physiological dormancy) prevents germination.
Germination of the excised seeds from the upper and lower
siliques of D. strictus was not inhibited by mucilage, indicating that water-soluble germination inhibitors were not
leached from it. However, the presence of a thick mucilaginous
layer might play a regulating role in diffusion of water and
oxygen to the embryo and thus influence germination
(Witztum et al., 1969; Liu and Tan, 2007; Thapliyal et al.,
2008).
After dry storage at room temperature for several months,
most of the seeds from upper siliques and those removed
from lower siliques, but not those held within intact lower siliques, could germinate. In the field, after-ripening also occurs
in summer, and seeds are non-dormant in autumn. However,
before the first snowfall in Urümqi in November 2008, we
searched for seedlings of D. strictus in the field but found
none. This was probably due to soil being wet for a period
of only 1 – 2 d at a time, which did not allow the seeds
enough time to germinate even though temperatures were
favourable. However, many seedlings were found in the field
in spring (mid-March to early April), when soil moisture ( precipitation plus snow melt) and temperature were suitable for
germination. Presumably, these seedlings were from upper
silique seeds, or perhaps from lower siliques that had been produced several years earlier (seed bank). In the plots of the
experimental garden, only seeds from upper siliques and
those removed from lower siliques germinated. After 1 year
on the soil surface, pericarps of the lower siliques were still
hard, and seeds inside them were viable. The dormancy breaking requirements of this morph have not been identified, but it
is not due to lack of water uptake by seeds within the intact
siliques or silique segments (Table 7; Fig. 9). We suggest
that seeds within these indehiscent siliques and silique
segments are prevented from germinating during the first
spring germination season (and perhaps also in the second
and later germination seasons) by the mechanical restraint of
the thick, rigid pericarp. The time required for pericarps of
lower siliques to release their seeds or to become soft
enough for the seeds inside to germinate in the field is not
known.
Thus, the lower siliques can form a persistent seed bank
(sensu Thompson and Grime, 1979), which is an advantage
in arid zones with unpredictable rainfall, such as the Junggar
Desert. Buffering variation in reproductive success by
delayed germination in desert annuals is a classic example of
bet-hedging (Philippi, 1993; Venable, 2007). While the best
thing a non-germinating seed can do is to survive, a germinating seed may either die or produce many new offspring. Thus,
producing a fraction of dormant seeds is the hedge against the
risk associated with germination. The germination fraction
favoured by natural selection is lower when there is a high
risk of seedling death than when there is a low risk (Ellner,
1985; Evans et al., 2007).
Seeds from the upper siliques of D. strictus after-ripen
during summer and can germinate in autumn or the following
spring. Seedlings produced in autumn and in spring can
survive and produce seeds in late May to early June and, in
contrast to obligate winter annuals, the non-dormant seeds
do not re-enter dormancy (secondary dormancy) during
winter (Baskin and Baskin, 1998). Thus, the species has the
potential to behave both as a winter annual and as a springephemeral annual, i.e. facultative winter annuals. In the field,
D. strictus behaves mostly as a spring-ephemeral annual,
however, apparently because most seeds, which are nondormant in autumn, are prevented from germinating at that
time by low soil moisture content. The probable reason why
only a low percentage of seeds sown on the soil surface in
summer and watered at 3 d intervals during autumn germinated
in autumn is that the soil did not remain moist for a long
enough period of time for the seeds to germinate (Fig. 6B).
It also seems likely that seeds in the lower siliques would
have the same germination phenology as that described
above for seeds from upper siliques, once they are released
from the siliques or the pericarp becomes soft enough for
seeds to germinate within it. The dormancy breaking and germination characteristics of seeds extracted from these indehiscent lower siliques are, in fact, the same as those released from
the dehiscent upper siliques (see Figs 4 and 5).
In contrast to extreme hot desert habitats, such as those in
the Mojave Desert of North America (MacMahon and
Wagner, 1985) and the Negev Desert of Israel (Gutterman,
1993), the Junggar Desert is a cold desert characterized by
snow and very low temperatures in winter and by aridity in
summer (Wei et al., 2003). Differences in morphology of siliques and seeds, dispersal and time delay between dispersal and
germination enable D. strictus to spread its offspring in space
and time, which is one of the ecological adaptive strategies to
the cold desert of the Junggar Basin. Like the majority of
species with heteromorphic seeds (Imbert, 2002), our results
suggest that high dispersal ability and reduced dormancy of
seeds (from upper siliques) provide D. strictus with the
chance of rapidly colonizing new sites. Furthermore, lack of
seed dispersal and delay of germination of seeds from lower
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
siliques give the species a better chance of retaining the
mother site and of colonizing suitable local sites. This constitutes a very safe means of reproduction and an available seed
reserve on/in soil of the desert habitat, thus increasing the
probability of persistence of D. strictus populations.
These studies have revealed differences in the behaviour of
the heteromorphic fruits and seeds of D. strictus compared
with those of Cakile, another member of the Brassicaceae.
Cakile is a genus of sandy freshwater lakeshores and of
sandy seashores (Rodman, 1974; Maun et al., 1990) and its
fruit and seed biology have been studied extensively (see references in Introduction). Whereas heteromorphism in D. strictus
is between upper and lower fruits and their seeds, in Cakile
spp. it is between proximal and distal segments of a fruit
and their seeds. In Cakile, the distal segment of the fruit
abscises and is dispersed by water, whereas the proximal
segment remains attached to the mother plant. Further,
neither segment is dehiscent, and thus seeds (usually one
per segment) germinate within the segment. In both
Diptychocarpus and Cakile, fresh seeds have non-deep physiological dormancy, and non-dormant seeds germinate in both
light and darkness. However, seeds of D. strictus come out
of dormancy (after-ripen) in summer, and those of Cakile generally need cold stratification (winter) to do so. Thus, at least in
the northern parts of its range (in the Northern Hemisphere)
seeds of Cakile germinate in spring and fruits mature in
autumn. This summer annual life cycle is in contrast to the
winter or spring annual ephemeral life cycle of D. strictus.
In summary, then, in Cakile the distal segment of the heteromorphic fruit distributes the species in space and the proximal
segment remains near the mother plant. In D. strictus, on the
other hand, the winged and mucilageous seed of the upper
dehiscent fruits distributes the species in space and the lower
indehiscent fruit containing a seed with a small wing and
very little mucilage remains near the mother plant and distributes the species through time via imposed dormancy of the
seeds inside it.
ACK N OW L E DG E M E N T S
This work was supported in part by the Projects on the
Research and Development of High Technology of Xinjiang
Uygur Autonomous Region (200810102), the Open Projects
of Xinjiang Key Laboratory of Grassland Resources and
Ecology (XJDX0209-2008-01), China, and the Construction
of Scientific and Technological Platforms Project from the
Chinese
Ministry
of
Science
and
Technology
(2005DKA21006 and 2005DKA21403).
L I T E R AT U R E C I T E D
An ZX, Zhou GL. 1995. Diptychocarpus Trautv. In: Mao ZM ed. Flora
Xinjiangensis, Vol. 2 (2). Urümqi: Xinjiang Sciences, Technology &
Hygiene Publishing House171– 172 (in Chinese).
Barbour MG. 1970a. Germination and early growth of the sand plant Cakile
maritima. Bulletin of the Torrey Botanical Club 97: 13–22.
Barbour MG. 1970b. Seedling ecology of Cakile maritima along the
California coast. Bulletin of the Torrey Botanical Club 97: 280– 289.
Baskin CC, Baskin JM. 1998. Seeds: ecology, biogeography, and evolution of
dormancy and germination. San Diego: Academic Press.
Baskin JM, Baskin CC. 2004. A classification system for seed dormancy.
Seed Science Research 14: 1 –16.
1013
Cheplick GP. 1992. Sibling competition in plants. Journal of Ecology 80:
567–575.
Cordazzo CV. 2006. Seed characteristics and dispersal of dimorphic fruit segments of a Cakile maritima Scopoli (Brassicaceae) population of southern
Brazilian coastal dunes. Revista Brasileira de Botánica 29: 259– 265.
Donohue K. 1998. Maternal determinants of seed dispersal in Cakile edentula:
fruit, plant, and site traits. Ecology 79: 2771– 2788.
Ellner SP. 1985. ESS germination strategies in randomly varying
environments. I. Logistic type models. Theoretical Population Biology
28: 50–79.
Evans ME, Ferrière R, Kane MJ, Venable DL. 2007. Bet hedging via seed
banking in desert evening primroses (Oenothera, Onagraceae): demographic evidence from natural populations. American Naturalist 169:
184–194.
Gravuer K, Von Wettberg EJ, Schmitt J. 2003. Dispersal biology of Liatris
scariosa var. novae-angliae (Asteraceae), a rare New England grassland
perennial. American Journal of Botany 90: 1159– 1167.
Gutterman Y, Shem-Tov S. 1996. Structure and function of the mucilaginous
seed coats of Plantago coronopus inhabiting the Negev Desert of Israel.
Israel Journal of Plant Science 44: 125– 134.
Gutterman Y, Shem-Tov S. 1997. Mucilaginous seed coat structure of
Carrichtera annua and Anastatica hierochuntica from populations occurring in the Negev Desert highlands of Israel, and its adhesion to the soil
crust. Journal of Arid Environments 35: 697–705.
Gutterman Y. 1993. Seed germination in desert plants. Berlin: Springer-Verlag.
Huang ZY, Gutterman Y. 2004. Value of the mucilaginous pellicle to seeds
of the sand-stabilizing desert woody shrub Artemisia sphaerocephala
(Asteraceae). Trees 18: 669 –676.
Huang ZY, Gutterman Y, Hu ZH, Zhang XS. 2001. Seed germination in
Artemisia sphaerocephala. I. The structure and function of the mucilaginous achene. Acta Phytoecologica Sinica 25: 22– 28 (in Chinese with
English abstract).
Imbert E. 2002. Ecological consequences and ontogeny of seed heteromorphism. Perspectives in Plant Ecology, Evolution and Systematics 5: 13– 36.
Levin SD, Cohen D, Hastings A. 1984. Dispersal strategies in a patchy
environment. Theoretical Population Biology 26: 165–191.
Liu XF, Tan DY. 2007. Ecological significance of seed mucilage in desert
plants. Chinese Bulletin of Botany 24: 414 –424 (in Chinese with
English abstract).
Lloyd DG. 1984. Variation strategies of plants in heterogenous environments.
Biological Journal of the Linnean Society 21: 357–385.
MacMahon JA, Wagner FH. 1985. The Mojave, Sonoran and Chihuahuan
deserts of North America. In: Evenari M, Noy-Meir I, Goodall DW.
eds. Hot deserts and arid shrublands. Amsterdam: Elsevier, 105– 202.
Maun MA, Boyd RS, Olson L. 1990. The biological flora of coastal dunes
and wetlands. 1. Cakile edentula (Bigel.) Hook. Journal of Coastal
Research 6: 137–156.
Maun MA, Payne AM. 1989. Fruit and seed polymorphism and its relation to
seedling growth in the genus Cakile. Canadian Journal of Botany 67:
2743– 2750.
Meyer SE, Carlson SL. 2001. Achene mass variation in Ericameria nauseosus (Asteraceae) in relation to dispersal ability and seedling fitness.
Functional Ecology 15: 274– 281.
Payne AM, Maun MA. 1981. Dispersal and floating ability of dimorphic fruit
segments of Cakile edentula var. lacustris. Canadian Journal of Botany
59: 2595–2602.
Philippi TE. 1993. Bet-hedging germination of desert annuals – beyond the
1st year. American Naturalist 142: 474–487.
Rodman JE. 1974. Systematics and evolution of the genus Cakile (Cruciferae).
Contributions of the Gray Herbarium of Harvard University 254: 3 –146.
Sadeh A, Guterman H, Gersani M, Ovadia O. 2009. Plastic bet-hedging in
an amphicarpic annual: an integrated strategy under variable conditions.
Evolutionary Ecology 23: 373– 388.
Sokal RR, Rohlf FJ. 1995. Biometry: the principles and practice of statistics
in biological research, 3rd edn. San Francisco: Freeman.
Telenius A, Torstensson P. 1989. The seed dimorphism of Spergularia
marina in relation to dispersal by wind and water. Oecologia 80:
206–210.
Thapliyal RC, Phartyal SS, Baskin JM, Baskin CC. 2008. Role of mucilage
in germination of Dillenia indica (Dilleniaceae) seeds. Australian Journal
of Botany 56: 583– 589.
Thompson K, Grime JP. 1979. Seasonal variation in the seed banks of herbaceous species in ten contrasting habitats. Journal of Ecology 67: 893– 921.
1014
Lu et al. — Fruit and seed heteromorphism in Diptychocarpus
Venable DL. 1985. The evolutionary ecology of seed heteromorphism.
American Naturalist 126: 577–595.
Venable DL. 2007. Bet hedging in a guild of desert annuals. Ecology 88:
1086–1090.
Venable DL, Lawlor L. 1980. Delayed germination and dispersal in desert
annuals: escape in space and time. Oecologia 46: 272–282.
Venable DL, Dyreson E, Pinero D, Becerra JX. 1998. Seed morphometrics
and adaptive geographic differentiation. Evolution 52: 344–354.
Wei WS, He Q, Liu MZ, Gao WD. 2003. Climate change and the desert
environment in Junggar Basin, Xinjiang, China. Journal of Desert
Research 23: 101–105 (in Chinese).
Witztum A, Gutterman Y, Evenari M. 1969. Integumentary mucilage as an
oxygen barrier during germination of Blepharis persica (Burm.) Kuntze.
Botanical Gazette 130: 238– 241.
Zhang J. 1993. Seed dimorphism in relation to germination and growth of
Cakile edentula. Canadian Journal of Botany 71: 1231– 1235.
Zhang J. 1995. Differences in phenotypic plasticity between plants from
dimorphic seeds of Cakile edentula. Oecologia 102: 353–360.
Zhang J, Maun MA. 1992. Effects of burial in sand on the growth and reproduction of Cakile edentula. Ecography 15: 296–302.
Zheng GC, Gu ZP. 1993. Biological micro technology. Beijing: China Higher
Education Press (in Chinese).