B99-011.vp

Color profile: DisabledComposite Default screen The effects of achene dimorphism on the
dispersal in time and space in Crepis sancta
(Asteraceae)
Eric Imbert
Abstract: In the annual species, Crepis sancta (L.) Bornm., peripheral and central florets of the capitulum yield
achenes that differ in mass, colour, morphology, and in the presence–absence of pappus. To examine the consequences
of seed dimorphism on dispersal in time in this species, I compared the viability of central and peripheral achenes
stored for up to 5 years. Both the germination percentage and the early seedling survival of both morphs decreased
progressively with duration of storage. However, the loss in viability was lower for peripheral achenes than for central
ones. Furthermore, two independent experiments showed that peripheral achenes have reduced dispersal ability, whereas
the pappus unit of central achenes confers greater dispersal ability. These differences between achene types may
explain the persistence of this colonizing species in communities dominated by perennial species, which is a major
ecological feature of C. sancta.
Key words: dispersal, dormancy, mixed strategy, secondary succession.
Résumé : Chez l’espèce annuelle, Crepis sancta (L.) Bornm., les fleurons situés en périphérie du capitule et ceux du
centre donnent des akènes qui diffèrent pour leur masse, leur couleur, leur morphologie, et pour la présence–absence de
pappus. Afin d’examiner les conséquences de ce dimorphisme sur la dispersion dans le temps de l’espèce, j’ai comparé
la conservation de la viabilité des akènes centraux et périphériques, en utilisant un stock d’akènes échantillonnés à
intervalle régulier pendant 5 ans. Le pourcentage de germination et la viabilité des plantules décroît progressivement
avec le temps de stockage. Cependant, cette perte était plus faible pour les akènes périphériques que pour les akènes
centraux. Deux expériences indépendantes ont montré que les akènes périphériques avaient une dispersion anémochore
réduite, alors que le pappus permettait aux akènes centraux une plus grande dispersion. Ces différences entre les deux
types d’akènes expliqueraient la persistance de cette espèce colonisatrice dans les communautés post-pionnières
dominées par des espèces pérennes, ce qui est une caractéristique majeure de C. sancta.
Mots clés : dispersion, dormance, stratégie mixte, succession secondaire.
Introduction
space and time (Venable and Lawlor 1980). This kind of Achene (one-seeded fruit) dimorphism, i.e., the produc- bet-hedging strategy reduces the impact of spatiotemporal tion by a single individual of two achene types, is fairly variability of environment on reproductive success (Venable common in the Asteraceae (Harper 1977). Except for some 1985) and allows achene dimorphic species to colonize species such as Gymnarhena micrantha Dest. (Koller and highly unpredictable habitats such as deserts and anthro- Roth 1964), and Geigeria alata (DC.) Oliv & Hiern (Burke pogenic habitats (Zohary 1962; Harper 1977).
1995), differentiation occurs between the fruits produced in Dimorphic achenes have been reported for several species the periphery of the capitulum and those in the centre. The in the genus Crepis (Asteraceae) (Babcock 1947). The an- most common trends are larger achenes in the periphery than nual species Crepis sancta (L.) Bornm. exhibits important those in the centre. Furthermore, peripheral achenes often morphological differentiation between achenes produced in lack dispersal structures such as pappus, conversely to cen- the periphery of the capitulum and those in the centre. Pe- tral achenes (Zohary 1950; Kigel 1992). These morphologi- ripheral achenes (hereafter called PA) are heavier and less cal differentiation are generally associated with differences numerous than central achenes (hereafter called CA): in dispersal ability and seed dormancy. Indeed, most studies 0.27 mg versus 0.10 mg and 9.6 PA versus 60.1 CA per head on dimorphic Asteraceae have shown that peripheral (n = 1132 heads). Difference in mass is due to a thicker achenes usually have high dormancy but low dispersal abil- pericarp and a taller embryo (Imbert et al. 1996). Further- ity, whereas central achenes show low dormancy and high more, PA are light coloured and lack a pappus, whereas CA dispersal ability (Venable and Lawlor 1980; Olivieri and are dark coloured and have a pappus. Both achene types dis- Berger 1985; Venable 1985; Rocha 1996). Therefore, achene perse in spring and remain dormant during summer. Indeed,as several species of the Asteraceae, achenes have to un- dergo a period of additional embryo maturation after seeddispersal (Baskin and Baskin 1976; Banovetz and Scheiner E. Imbert. Centre d’Écologie Fonctionnelle et Évolutive –
1994). This inability to germinate during summer prevents Centre National de la Recherche Scientifique, UPR 9056,F-34293 Montpellier Cédex 5, France.
the consequent high mortality of seedlings due to summer droughts (Mott 1972; Joley et al. 1997).
Can. J. Bot. 77: (1999)
Color profile: DisabledComposite Default screen Contrasting with most studies about seed dimorphism in Nonpollinated florets produce a white and soft achenelike struc- the Asteraceae, Ellner (1986) and Imbert et al. (1996) re- ture. Furthermore, late abortion can occur during development; ported that, after the post-maturation phase, peripheral and “achenes” are then similar in size and in colour to normal fruits but central achenes in C. sancta germinate immediately on en- they lack an embryo and endosperm. Those two kinds of acheneshave been excluded from the following experiments.
countering favourable conditions. Defining dormancy as aphysiological process that prevents germination of matureseeds Dispersal in time
(Vleeshouwers et al. 1995), none of achene morphs of Hundreds of mature seed heads were collected during five con- C. sancta is dormant. Therefore, seed dimorphism in secutive springs (1992–1996). Each sample represented one cohort C. sancta is not associated with the dual germination strat- (cohort 1992, cohort 1993, and so on). Only the cohort 1992 was egy (high dormancy versus low dormancy), commonly ob- collected from plants growing in natural populations (described in served in seed dimorphic Asteraceae. While the species Imbert et al. 1996), while other cohorts were from plants cultivated cannot benefit from dormancy, which is a common strategy in the common garden of the Centre d’Ecologie Fonctionnelle etEvolutive (CEFE; Montpellier, France). To exclude among-year for spreading risk over time in fluctuating environments variation for seed quality due to maternal environment effects, (e.g., Venable and Lawlor 1980), C. sancta commonly colo- plants used as seed source were always cultivated in the same con- nizes and persists in unpredictable habitats, such as Mediter- ditions, i.e., each plant was watered daily and individually grown ranean old fields. In such habitats, perturbations due to in a 1-L pot containing a 1:1 mixture of sterile soil and compost.
mammals or ants are frequent (Lavorel et al. 1994), and Mature achenes were collected during the seed dispersal phase. To seeds could be buried for several years. Therefore, the suc- prevent predation and disease, achenes were stored in room condi- cess of the species, as with other plant species, also depends tions, using paper bags in dark at 19 ± 5°C. Germination tests were on the ability of embryos to remain viable with ageing (Co- carried out in autumn 1996. Therefore, achenes from cohort 1992 hen 1966). It was thus necessary to test whether peripheral spent 4 years in storage, cohort 1993 spent 3 years, and so on. For and central achenes of C. sancta differ in maintenance of vi- each cohort, I randomly chose 1000 CA and 1000 PA without tak-ing the family factor into account. Achenes were germinated in ability with ageing. Indeed, interspecific comparisons have 10-cm Petri dishes, containing one disk of Whatmann filter paper, shown that large seeds, which contain more food in storage moistened with distilled water. Each dish contained 50 achenes and a more developed seed coat, can remain viable for lon- from the same cohort and the same morph. Petri dishes were ger period of time than small seeds (Priestley 1986; Baker placed in an unheated greenhouse (temperature range 0–26.5°C) 1989). Although several studies have focused on germina- with natural lighting. Germination was counted daily and was con- tion behaviour of achene dimorphic species (see references sidered as successful when cotyledons were emergent and in Imbert et al. 1996 and Rocha 1996), few experiments green-coloured. For each dish, I considered that the final rate of have been carried out to compare the maintenance of viabil- germination had been reached when there was no germination for 20 consecutive days. The experiment started on October 8, 1996, While the dormancy and the survival ability of seeds, or lasted 65 days, and ended on December 11, 1996. After germina-tion, every seedling was transplanted individually into 50-cm3 pots escape in time, is of critical importance in the population containing a 1:1 mixture of sterile soil and compost, and seedling persistence, dispersal in space also represents a strategy to survival was recorded daily over 30 days after germination.
face environmental unpredictability (e.g., Cohen and Levin Because germination response and seedling survival are binary 1991). Furthermore, many theoretical studies have demon- (success or failure), data were first analysed using the logistic re- strated that there should be negative correlations between gression procedure (Sokal and Rohlf 1995; Joley et al. 1997). In a dispersal in time and space (e.g., Venable and Lawlor 1980; second step, the effects of years in storage within each achene Cohen and Levin 1987), which have been confirmed with morph (2 × 5 contingency tables for each achene morph) and of empirical data on achene dimorphic Asteraceae (see refer- achene morph for each cohort (2 × 2 contingency tables for each ences above). Therefore, the second topic of the present arti- cohort) were tested using the Fisher’s exact test, since the total cle is to investigate the dispersal ability of both achene number of achenes was defined by the design (Sokal and Rohlf Materials and methods
Dispersal in space
The horizontal distance covered by a seed depends on the re- The species
lease height and on the rate of falling. Because CA and PA are pro- Crepis sancta is commonly found in Mediterranean area of duced within the same head, the horizontal dispersal can only France. It occurs in various kinds of anthropogenic habitats such as differ for their respective rate of descent. Settling velocity was de- road verges. However, the species mainly occurs in recently aban- termined by measuring the time necessary to cover a fixed dis- doned old fields (Escarré et al. 1983). In contrast with most colo- tance, using the apparatus described in Askew et al. (1997). The nizing species, C. sancta can tolerate successional processes and principle of the apparatus is that the seed falls in a tunnel. The seed persists in post-pioneer communities, up to 40 years after abandon- passes a first electronic detector (the timer is thus turned on), and ment (Imbert et al. 1999). It is a winter annual that germinates af- 25 cm farther, a second detector (the timer is thus turned off). As ter autumn rainfall (September–October). Individuals overwinter at previously described, PA and CA differ for the presence–absence a rosette stage and flower in early spring (February–March). All of pappus, their mass, and their shape. To test the impact of the the florets are ligulate and hermaphroditic. The species is dispersal structure, i.e., the pappus, measurements have been done self-incompatible and insect pollinated. Achenes disperse about for PA, CA with pappus, and CA without pappus. Ten replicate 1 week after pollination. They vary in morphology within each measurements, using 10 different achenes, were taken for each head as described above. Individuals can also produce achenes achene type. Achenes used in this experiment were collected from showing intermediate morphological characteristics, but this morph plants grown in controlled conditions. Achenes were sampled is not common and was not considered in the present study.
when the pappus was fully extended and stored in Plexiglas tube to Color profile: DisabledComposite Default screen Table 1. Germination rate after 65 days and seedling survival rate after 30 days according to achene morph
and number of years in storage.
Germination
Seedling survival
Note: Comparisons among achene morphs within each year and among years within each achene morph were made with
the Fisher’s exact test (Sokal and Rohlf 1995).
keep the pappus intact. Settling velocities were compared with a Achenes sampled in the spring 1996 had the highest germi- nation rate and the germination rate decreased for both The capacity for wind dispersal was also measured in natural achene morphs when the number of years in storage in- conditions. This experiment was carried out in spring 1994 during creased (Table 1). Within each achene morph, the germina- the natural season of dispersal in the experimental station of the tion rate was significantly lower after 1 year of storage ( p < CEFE. The experimental area was a 15 × 50 m field where the 0.001 for both morphs). For the oldest achenes sampled on vegetation had been previously mechanically removed. In the mid-dle of the field, along a longitudinal transect, I transplanted four plants from the common garden (cohort 1993, three years in groups each containing 36 plants and covering 1 m2. Two consecu- storage), germination rate was below 0.25 (Table 1), and tive groups were separated by 11 m, and distance between the two germination of PA was twice that of CA (Table 1). Finally, external groups and limit of the experimental area was 7 m. Seed for the achenes sampled in 1992, nine embryos of 1000 suc- rain was sampled using Petri dishes (diameter 10 cm) containing ceeded in germinating for CA and 49 for PA.
one disk of Whatmann filter paper painted with glue (Solveurode®, As for the germination rate, the logistic regression analy- Rhône Poulenc). Preliminary tests have shown that the glue held sis revealed a significant interaction between the number of the achenes, even after rainfall. The seed traps were placed in four years in storage and the achene morph on the survival rate perpendicular directions and at eight distances: within the groups (G = 16.67, df = 4, p < 0.01). Indeed, although seedlings (distance 0 m from the centre), at the margin of the group (distance from CA had a lower survival than those from PA, except 0.5 m from the centre), and at 1, 1.5, 2.5, 3.5, 4.5, and 5.5 m fromthe centre of each group. This design allowed me to sample seed for achenes collected in 1994 (Table 1), the difference was rain within the groups and within circular zones, which were rings significant only for achenes collected in 1993 (3 years of placed at 0.5, 1, 1.5, 2.5, 3.5, 4.5, and 5.5 m from the centre of storage) and 1995 (1 year of storage; Table 1). For achenes each population. The width of one ring was the trap diameter, i.e., collected in 1992 (4 years of storage), the survival rate was 10 cm. Traps were checked four times throughout the fruiting pe- 0.11 for CA (one seedling out of nine), and 0.28 for PA (14 riod and disks of paper were removed and achenes counted at each out of 49), but the difference was not significant ( p = 0.67).
Note that if data from the cohort 1992 are excluded from In the data analyses, I summarized the dispersal pattern of each the analysis, the Fisher’s exact test for a year effect is still morph by the number of achenes sampled at each distance consid- significant for germination ( p < 0.001) and survival rate ( p ering four replicates (the four groups) without considering direc- tion and the sampling date, since previous analyses have shownthat there was no effect of both factors on the dispersal ability ofPA and CA. I estimated the total number of achenes per distance Dispersal in space
(TN) using the following equation: TN = SN × TA/SA, where SN The settling velocity was clearly different among the three is the number of sampled seeds, SA is the sampled area (i.e., the samples (F[2,27] = 156.87, P < 0.0001). CA with pappus had surface of one seed trap × number of traps), and TA is the total the lowest rate of falling (0.23 ± 0.01 m·s–1; mean ± SE) and area (i.e., the surface of one concentric ring for each distance) CA without pappus the highest rate (1.48 ± 0.08 m·s–1). For (Ronsheim 1994; Thiede and Auguspurger 1996). The mean distri- PA, the falling rate was 1.21 ± 0.03 m·s–1 (the means are sig- butions (over the four replicates) of achene morphs were compared nificantly different at p < 0.05 with a Tukey’s test).
with Kolmogorov–Smirnov tests (Sokal and Rohlf 1995).
From the field experiment, a total of 5816 CA and 964 PA were trapped over 4 weeks (ratio CA/PA = 6.03). Of the CA, 69.9% were trapped beyond 0.5 m, whereas78.4% of the PA were trapped within the groups or at the Dispersal in time
margin of the groups. The distributions of both achene The interaction between the number of years in storage types according to the distance from the seed source were and the achene morph significantly affected the germination significantly different (Kolmogorov–Smirnov test, D = response (G = 41.53, df = 4, p < 0.001). For achenes sam- 0.55, p < 0.001). Within the group (distance 0), the num- pled in 1996 and 1995, there was no difference between ber of CA and that of PA trapped were the same morphs (Table 1), whereas germination of PA was higher (Fig. 1A). At the margin (distance 0.5), some PA were than that of CA after 2, 3, and 4 years of storage (Table 1).
found, and beyond the margin, the seed rain was essen- Color profile: DisabledComposite Default screen Fig. 1. Dispersal pattern (mean and SE) of central achenes (CA;
Discussion
open bars) and peripheral achenes (PA; solid bars): (A) Numberof achenes sampled according to distance. (B) Expected number The present data show that, in C. sancta, germination rate of achenes using the ratio between the total area and the decreased when duration of storage increased, consistently sampled area, according to the following equation: total number of achenes = (number of sampled achenes × total area)/sampled Asteraceae (Priestley 1986). Since unfavourable environ- area. Total area represents the surface of the concentric ring for mental conditions could induce secondary dormancy (Baskin each distance, and sampled area is the surface of one seed trap × and Baskin 1989), one could argue that ungerminated seeds number of traps (for more details, see Ronsheim 1994; Thiede were viable but dormant. I did not perform biochemical via- and Auguspurger 1996). For distances 2.5, 3.5, 4.5, and 5.5 m, bility tests to distinguish between dormant and nonviable seeds, in particular because of the small size of embryo inC. sancta. However, in winter annuals such as C. sancta,secondary dormancy is usually induced by low winter tem- perature (Baskin and Baskin 1989). In the present study, stored seeds did not suffer such low temperatures. Further- more, several observations during the experiment (e.g., fungi growth on ungerminated seeds, uncoloured seedlings) sug- gest that ungerminated achenes were nonviable. For in- stance, fungi attacks and uncoloured seedlings reveal that imbibition was not accompanied by the repair of deleterious changes accumulate during ageing (Villiers 1973; Priestley1986; Bernal-Lugo and Leopold 1998). Finally, seedling via- bility also decreased when the number of years in storage in- Maternal environment effects and genetic differences among the plants sampled may also contribute to variation among cohorts for seed characteristics. Regarding the co-horts from the common garden (1993–1996), plants weregrown in similar conditions; thus, maternal effects are un-likely to explain among-year variation. Furthermore, every year, more than one thousand of individuals were cultivated on the common garden. Therefore, even though genetic vari- ation among plants used as seed sources was not controlled,variation within cohorts was likely to be as great as variationamong cohorts. In contrast, genetic and nongenetic maternal effects may contribute significantly to the differences be- tween seeds collected in 1992 and those collected later, since the former were sampled in natural populations. However,germination and survival clearly decreased after 1 year of storage for seeds from the common garden. Furthermore, a germination test was performed in spring 1993 (i.e., 1 year of storage) on the samples here defined as cohort 1992, and germination response was greater than 0.60 (Imbert et al.
1996). Therefore, it clearly appears that ageing affected ger- mination response and seedling survival.
Peripheral and central achenes produced in spring 1996 did not differ. In particular, both achene morphs massively germinated when favourable conditions appear (for detailson the germination dynamics, see Imbert et al. 1996). Theseobservations are similar to those reported by Imbert et al.
tially composed with CA. Using the adjusted data to take the ratio total area/sampled area into account, there were C. sancta, was nonsignificant for the strategy of dispersal in obvious differences between the distribution of PA and time (see also Ellner 1986). Yet, it appears that effects of CA (Fig. 1B). For PA, the distribution was highly skewed storage are different for central and peripheral achenes of and leptokurtic (skewness = 2.05 and kurtosis = 2.54; C. sancta. In particular, embryos from peripheral achenes re- Fig. 1B); 70.9% of PA were estimated to fall within the main viable longer than those from central achenes, consis- groups (distance 0; Fig. 1B). The asymmetry for CA dis- tently with a study performed on Bidens pilosa, another seed tribution was less important (skewness = 0.42, kurtosis = dimorphic Asteraceae (Rocha 1996). Peripheral achenes may –1.56; Fig. 1B), revealing a greater distance of dispersal benefit from the thickness of pericarp and embryo size as predicted by the velocity data. On average, 21.7% of (Priestley 1986; Baker 1989). However, some endogenous CA fell within the groups (Fig. 1B).
factors, such as hormone quantity, can also affect viability Color profile: DisabledComposite Default screen conservation (Taylorson and Hendricks 1977; Bernal-Lugo placement is more likely to occur because plants from pe- and Leopold 1998). The present study sheds light on one ad- ripheral achenes are better competitors than those from cen- vantage of achene dimorphism in C. sancta: when achenes encounter conditions inhibiting germination (e.g., burial, Several empirical studies have shown that achene dimor- lack of rainfall, germination inhibition by other species), pe- phism combines reduced dispersal with high dormancy and ripheral achenes are more advantageous than central ones.
high dispersal with reduced dormancy (Venable and Lawlor This consequence of achene dimorphism should increase the 1980; Olivieri and Berger 1985). Such observations are con- likelihood to persist in disturbed habitats, in particular in sistent with many theoretical studies that predicts a negative Mediterranean old fields where seed banks play a major role correlation between dispersal in time and space (Cohen and in community structure (Lavorel et al. 1994).
Levin 1987; Venable and Brown 1988; Cheplick 1996). Yet, In some species, the different morphs have different dis- in C. sancta, neither achene morphs do not diapause (Ellner persal agents (Sorensen 1978). In C. sancta, there is no pas- 1986; Imbert et al. 1996; this study). Crepis sancta is pres- sive zoochory (Shmida and Ellner 1983), since none of the ent in early stages of the secondary succession in Mediterra- morphs has epizoochorous structures such as trichomes, and nean area and can persist through the succession (Imbert et wind is the main agent of dispersal. The falling velocity of al. 1999). Therefore, the reproductive success of a genera- central achenes was of the same order of magnitude re- tion is expected to be positively correlated with the perfor- corded in other species with achene–pappus as dispersal unit. The rates of falling of peripheral achenes and central self-replacement strategy also leads to positive temporal cor- achenes without a pappus are identical to velocities recorded relation. Such positive temporal correlation is expected to for diaspores which are 10 or 100 times heavier (Matlack produce positive correlation between dispersal in space and 1987). This interspecific comparison illustrates that wind time (Cohen and Levin 1991) and should explain the uncon- dispersal mainly depends on the presence–absence of a ventional pattern of achene dimorphism in C. sancta.
Because central achenes fall to the ground less rapidly Acknowledgements
than peripheral achenes, they should have a greater horizon-tal dispersal. The seed rain samples confirm the difference in I am grateful to the staff of the common garden of the dispersal ability between central and peripheral achenes. In- CEFE for technical assistance; Andrew Askew and Ken deed, most of peripheral achenes fall close to the mother Thompson for measuring falling velocities; Roger Pradel for plant while central achenes show some dispersal. As for sev- statistical advice; and Gregory Cheplick, José Escarré, eral achene dimorphic Asteraceae (Venable and Lawlor Isabelle Olivieri, and two anonymous reviewers for their 1980; Olivieri and Berger 1985), the morphological differen- comments. This research was supported by the Centre na- tiation between peripheral and central achenes in C. sancta tional de la recherche scientifique and the Ministère de la re- are associated with a differentiation in wind-dispersal ability.
cherche et de l’enseignement supérieur.
This result was expected because of the morphological dif-ferentiation between morphs. However, it was necessary to References
demonstrate it because it leads to important ecological con-sequences. Crepis sancta is a thermophilous species coloniz- Askew, A.P., Corker, D., Hodkinson, D.J., and Thompson, K.
ing open habitats, such as recently abandoned old fields.
1997. A new apparatus to measure the rate fall of seeds. Funct.
During the secondary succession, perennial species invade Ecol. 11: 121–125.
the herbaceous strata, and light availability decreases. Thus, Babcock, E.B. 1947. The genus Crepis. University of California persistence of thermophilous species depends on the likeli- hood to settle seeds within an open microsite. As seed Baker, H.G. 1989. Some aspects of the natural history of seeds spreading is dominated by stochastic events, especially in banks. In Ecology of soil seed banks. Edited by M.A. Leck, V.T.
anemochorous species, most seeds fall in unfavourable sites.
Parker, and R.L. Simpson. Academic Press, San Diego, Calif.
Seed loss is therefore important and annual species are rap- idly excluded from the succession (Escarré et al. 1983). In Banovetz, S.J., and Scheiner, S.M. 1994. Secondary seed dor- contrast, perennial species, which allocate resources to vege- mancy in Coreopsis lanceolata. Am. Midl. Nat. 131: 75–83.
tative growth, can invade the habitat. Peripheral achenes of Baskin, J.M., and Baskin, C.C. 1976. Germination dimorphism in Heterotheca subaxillaris var subaxillaris. Bull. Torrey Bot.
C. sancta lack dispersal; thus, plants from these achenes Club, 103: 201–206.
should be able to colonize the suitable opening after the Baskin, J.M., and Baskin, C.C. 1989. Physiology of dormancy and death of the mother plant. In other words, a site previously germination in relation to seed bank ecology. In Ecology of soil occupied by a C. sancta individual may be recolonized in seed banks. Edited by M.A. Leck, V.T. Parker, and R.L.
the next generation by an offspring. This strategy, called the Simpson. Academic Press, San Diego, Calif. pp. 53–66.
self-replacement strategy (McEvoy 1984), decreases the im- Bernal-Lugo, I., and Leopold, A.C. 1998. The dynamics of seed pact of stochastic events during dispersal for peripheral mortality. J. Exp. Bot. 49: 1455–1461.
achenes. This self replacement is likely to occur in most Burke, A. 1995. Geigeria alata in the Namib desert: seed hetero- wind-dispersed species because seed distribution often is morphism in an extremely arid environment. J. Veg. Sci. 6:
leptokurtic, and the seed rain mainly falls under the mother plant (Willson 1993). For instance, 21% of the central Cheplick, G.P. 1996. Do seed germination patterns in cleistoga- achenes of C. sancta did not show any dispersal. However, mous annual reduce the risk of sibling competition? J. Ecol. 84:
in C. sancta, as in other seed dimorphic species, the self re- Color profile: DisabledComposite Default screen Cohen, D. 1966. Optimizing reproduction in a randomly varying Olivieri, I., and Berger, A. 1985. Seed dimorphism for dispersal: environment. J. Theor. Biol. 12: 119–129.
physiological, genetic and demographic aspects. In Genetic dif- Cohen, D., and Levin, S.A. 1987. The interaction between dis- ferentiation and dispersal in plants. Edited by P. Jacquard, G.
persal and dormancy strategies in varying and heterogeneous Heim, and J. Antonovics. Springer-Verlag, Berlin, Germany.
environments. In Mathematical topics in population biology.
Edited by E. Teramoto and M. Yamaguti. Springer-Verlag, Hei- Priestley, D.A. 1986. Seed aging, implications for seed storage and persistence in the soil. Cornell University Press, Ithaca, N.Y.
Cohen, D., and Levin, S.A. 1991. Dispersal in patchy environ- Rocha, O.J. 1996. The effects of achene heteromorphism on the ments: the effects of temporal and spatial structure. Theor.
dispersal capacity of Bidens pilosa L. Int. J. Plant Sci. 157:
Popul. Biol. 39: 63–99.
Ellner, S. 1986. Germination dimorphism and parent–offspring Ronsheim, M.L. 1994. Dispersal distance and predation rates of conflict in seed germination. J. Theor. Biol. 23: 173–185.
sexual and asexual propagules of Allium vineale L. Am. Midl.
Escarré, J., Houssard, C., Debussche, M., and Lepart, J. 1983. Evo- Nat. 131: 55–64.
lution de la végétation et du sol après abandon cultural en Sheldon, J.C., and Burrows, F.M. 1973. The dispersal effectivness région méditerranéenne : étude de succession dans les Garrigues of the achene–pappus units of selected Compositae in steady de Montpelliérais (France). Acta Oecol. 4: 221–239.
winds with convections. New Phytol. 72: 665–675.
Harper, J.L. 1977. Population biology of plants. Academic Press, Shmida, A., and Ellner, S. 1983. Seed dispersal on pastoral grazers in open Mediterranean chaparral. Isr. J. Bot. 32: 147–159.
Imbert, E., Escarré, J., and Lepart, J. 1996. Achene dimorphism Sokal, R.R., and Rohlf, F.J. 1995. Biometry. Freeman, New York.
and among-population variations in some biological traits in Sorensen, A.E. 1978. Somatic polymorphism and seed dispersal.
Crepis sancta (Asteraceae). Int. J. Plant Sci. 157: 309–315.
Nature (London), 276: 174–176.
Imbert, E., Escarré, J., and Lepart, J. 1997. Seed heteromorphism Taylorson, R.B, and Hendricks, S.B. 1977. Dormancy in seeds.
in Crepis sancta (Asteraceae): performance of two morphs in Annu. Rev. Plant Physiol. 28: 331–354.
different environments. Oikos, 79: 325–332.
Thiede, D.A, and Augspurger, C.K. 1996. Intraspecific variation in Imbert, E., Escarré, J., and Lepart, J. 1999. Local adaptation and seed dispersion of Lepidium campestre (Brassicaceae). Am. J.
non-genetic maternal effects among three populations of Crepis Bot. 83: 856–866.
sancta (Asteraceae). Ecoscience, 6. In press.
Venable, D.L. 1985. The evolutionary ecology of seed hetero- Joley, D.B., Maddox, D.M., Mackey, B.E., Schoenig, S.E., and morphism. Am. Nat. 126: 577–595.
Casanave, K.A. 1997. Effects of light and temperature on germi- Venable, D.L., and Brown, J.S. 1988. The selective interactions of nation of dimorphic achenes of Centaurea solstitialis in Califor- dispersal, dormancy and seed size as adaptations for reducing nia. Can. J. Bot. 75: 2131–2139.
risk in variable environments. Am. Nat. 131: 360–384.
Kigel, J. 1992. Diaspore heteromorphism and germination in popu- Venable, D.L., and Lawlor, L. 1980. Delayed germination and dis- lations of the ephemeral Hedypnois rhagadioloides (L.) F.W.
persal in desert annuals: escape in space and time. Oecologia, Schmidt (Asteraceae) inhabiting a geographic range of increas- 46: 272–282.
ing aridity. Acta Oecol. 13: 45–53.
Villiers, T.A. 1973. Ageing and the longevity of seeds in field con- Koller, D., and Roth, N. 1964. Studies of the ecological and physi- ditions. In Seed ecology. Edited by W. Heydecker. Butterworths, ological significance of amphicarpy in Gymnarrhena micrantha (Compositae). Am. J. Bot. 51: 26–35.
Vleeshouwers, L.M., Bouwmeester, H.J., and Karssen, C.M. 1995.
Lavorel, S., Lepart, J., Debussche, M., Lebreton, J.D., and Beffy, Redefining seed dormancy: an attempt to integrate physiology J.L. 1994. Small scale disturbances and the maintenance of spe- and ecology. J. Ecol. 83: 1031–1037.
cies diversity in Mediterranean old-fields. Oikos, 70: 455–473.
Willson, M.F. 1993. Dispersal mode, seed shadows and coloniza- Matlack, G. 1987. Diaspore size, shape, and fall behavior in tion patterns. Vegetatio, 107/108: 261–280.
wind-dispersed plant species. Am. J. Bot. 74: 1150–1160.
Zohary, M. 1950. Evolutionary trends in the fruiting head of McEvoy, P.B. 1984. Seedling dispersion and the persistence of Compositae. Evolution, 4: 103–109.
ragwort Senecio jacobaea (Compositae) in a grassland domi- Zohary, M. 1962. Plant life of Palestine, Israel, and Jordan. Ronald nated by perennial species. Oikos, 42: 138–143.
Mott, J.J. 1972. Germination studies on some annual species from an arid region of Western Australia. J. Ecol. 60: 293–304.

Source: http://eric.imbert6.free.fr/articles/article5.pdf