EXPLOSIVE SEED DISPERSAL IN TWO
2Departamento de Ciencias Ambientales, Universidad Pablo de Olavide, Ctra. Utrera Km 1, 41013 Sevilla, Spain; and 3Departamento de Biologı´a Vegetal y Ecologı´a, Universidad de Sevilla, Apdo. 1095, 41080 Sevilla, Spain
The distance of explosive dispersal, its pattern in time, and the relative importance of autochory have been studied in two diplo-chorous species: Euphorbia boetica and E. nicaeensis. The seeds of E. boetica released by explosive dispersal reached a median distance of 156 cm and a maximum of almost 8 m, while the distances reached by the seeds of E. nicaeensis were lower: a median of 132 cm and a maximum of 5 m. The differences in explosive dispersal distance between species seem to depend on both seed mass and caruncle retention. The seeds of both species present a caruncle, but in E. boetica this is tiny, and in most cases is shed during the explosion of the capsules. The distances reached by the seeds of these species, dispersed just by capsule explosion, were similar to or greater than the distances to which ants disperse seeds in the Mediterranean sclerophyllous vegetation. Diplochorous plants may maximize either the distance of primary dispersal or that of secondary dispersal. Given that the seeds of E. boetica, that lose their caruncles, are not gathered by myrmecochorous ants, the results suggest that E. boetica maximizes its primary dispersal distance, whereas E. nicaeensis favors its secondary dispersal.
Key words: caruncle; diplochory; Euphorbia nicaeensis; Euphorbia boetica; explosive dispersal.
In plants with explosive dispersal, the seeds are discharged by the rupture or explosive dehiscence of the fruit, produced by the elastic contraction of its tissues (Garrison et al., 2000). The distance reached by seeds that are dispersed explosively is usually quite short compared with the distance reached by other means (Bullock and Primack, 1977; Willson, 1993). However, species presenting a primary explosive dispersal commonly have some type of secondary dispersal that increas-es the distance achieved by autochory (Wincreas-estoby and Rice, 1981; Andersen, 1988; Stamp and Lucas, 1990). Secondary dispersal is often via myrmecochory, because in many of these species, the seeds have an elaiosome that attracts ants (Beattie and Lyons, 1975; Berg, 1975a, b; Culver and Beattie, 1980; Beattie and Culver, 1981; Stamp and Lucas, 1983; Go´mez and Espadaler, 1994; Ohkawara and Higashi, 1994; Espadaler and Go´mez, 1996, 2000).
The Euphorbiaceae family presents a great diversity of dis-persal systems (Webster, 1994). Autochory is the primitive sit-uation in the family (Berg, 1975b; Webster, 1994) and is pre-sent in numerous genera; however, those prepre-senting caruncu-late seeds dispersed by ants and those having fleshy fruit dis-persed by birds are also frequent (Webster, 1994). The genus
Euphorbia has fruits with explosive dehiscence produced by
the differing orientation of the cells of the mechanical wall (Berg, 1990). In addition, in most cases the seeds have a lipid-rich caruncle that functions as an elaiosome, attracting ants and initiating a myrmecochorous secondary dispersal (Berg, 1975b). It has been reported that the distance to which the ants move the seeds is greater than that achieved by explosive dis-persal (Stamp and Lucas, 1990). The presence of an elaiosome for myrmecochorous dispersal may alter the seed
aerodynam-1Manuscript received 16 April 2004; revision accepted 18 November 2004. This work has been supported by the Natural Park Sierra de Grazalema (Proyecto Pinsapar) and by a grant of the Programa de Ayuda a los Grupos de Investigacio´n (Junta de Andalucı´a).
4Author for correspondence (E-mail: firstname.lastname@example.org) fax: (1) 95.434.9151.
ics, reducing the distance of explosive dispersal (Beattie and Lyons, 1975). In some species of Euphorbia, however, the caruncle is readily shed from the seeds (Benedı´ et al., 1997). This could increase the explosive dispersal distance, but the seeds would lose their attractiveness to the disperser ants. In fact, due to constraints of the explosive mechanism, plant spe-cies that use explosive dispersal are postulated to maximize either explosive distances or secondary dispersal (Beattie and Lyons, 1975; Stamp and Lucas, 1983).
Various studies have evaluated the relative importance of explosive and myrmecochorous dispersal in closely related di-plochorous species (Culver and Beattie, 1980; Stamp and Lu-cas, 1983; Ohkawara and Higashi, 1994). In the genus
Eu-phorbia, myrmecochorous dispersal has been extensively
stud-ied (Pemberton, 1988; Go´mez and Espadaler, 1994, 1995, 1998; Espadaler and Go´mez, 1996, 2000; Wolff and Debus-sche, 1999), but primary dispersal by autochory has been re-ported only in Euphorbia characias (Espadaler and Go´mez, 2000).
In this paper, we have studied explosive seed dispersal in two phylogenetically close Mediterranean perennial spurges:
Euphorbia boetica Boiss. and E. nicaeensis All. (Croizat,
1972; Benedı´ et al., 1997). The two species are quite similar in inflorescence architecture, and they produce three-seeded capsules. Their seeds present a caruncle, but in some cases we have observed that this is shed during the process of autocho-rous dispersal. In both species, only carunculated seeds are actively collected by myrmecochorous ants (Narbona, 2002). The aims of the present work are (1) to know the temporal pattern of seed dispersal, (2) to characterize the curves of ex-plosive dispersal, and (3) to determine whether the loss of the caruncle is a frequent phenomenon in the two species and if that loss affects dispersal distance.
MATERIAL AND METHODS
Study sites and species—Euphorbia boetica is endemic to the southern
is a circum-Mediterranean species that occurs above 600 m altitude in south Spain. Thus, the two species do not occur in the same stands. The study sites are in western Andalusia (SW Spain) and present Mediterranean sclerophyl-lous vegetation. Euphorbia boetica was studied in a population at Hinojos (Huelva Province) located on a peneplane at 80–90 m altitude and ca. 30 km from the sea; the vegetation consists of a mixed woodland of Pinus pinea L. and Quercus suber L., with a scrub layer comprising mainly Cistaceae, Lam-iaceae, and Leguminosae. The density of adult individuals of E. boetica in this population is 0.56 individuals/m2, and the mean neighbor distance is 1.44 m. Euphorbia nicaeensis was studied in a population at Grazalema (Ca´diz Province), situated in a mountainous calcareous zone at about 800 m altitude; the vegetation consists of scattered individuals of Quercus rotundifolia Lam. and Ceratonia siliqua L. and a scrub comprising Leguminosae, Anacardiacae, Cupressaceae, Oleaceae, and Lamiaceae. In this population, the density of adult individuals of E. nicaeensis is 0.32 individuals/m2, and the mean neigh-bor distance is 1.90 m. Additionally, capsules of Euphorbia boetica were collected in a population at El Gandul (Seville Province), and those of E.
nicaeensis at Aracena (Huelva Province).
Euphorbia boetica and E. nicaeensis are perennial shrubs that branch at the
base and reach heights of 50 and 60 cm, respectively. They produce numerous floral stems bearing pleiochasia. Each pleiochasial branch forms several plei-ochasia or dichasia, which bloom sequentially in the spring. The fruits are fully ripe at the beginning of the summer. In both species, the seeds are dispersed during the dry period: from mid-June to late July in E. boetica and from mid-July to late August in E. nicaeensis. In the fruit, each seed has a caruncle, but some seeds lose the caruncle during explosive dispersal (there-after ecarunculate seeds). In the studied populations, carunculate seeds are collected by the myrmecochorous ants Tapinoma nigerrimun, Pheidolle
pal-lidula, and Plagiolepis pygmaea; the granivorous ants Messor bouveri and M. marocanus collect both carunculate and ecarunculate seeds (Narbona, 2002).
Hourly rate of autochorous dispersal—Plants with a large number of ripe
capsules were selected in the population at Hinojos for E. boetica and at Grazalema for E. nicaeensis. These capsules were counted and marked indi-vidually. Each hour, the marked capsules were enumerated, and those that had disappeared were considered to have dispersed their seeds. The censuses were begun at 1000 hours and finished at 2000 hours, when dispersal was estab-lished to have ended. During each census, the shade temperature beside one of the observed plants was measured. In E. boetica, the rate of explosion was studied on 3 d in June (two in 1999 and one in 2000), and in E. nicaeensis, on 3 d in August (two in 1999 and one in 2000). The number of plants of each species observed each day was between 13 and 20. The capsules that remained undispersed in the last census of any day were examined first thing the next morning (0800 hours) to discover whether any seeds had scattered during the night.
Dispersal distance of seeds—Entire branches bearing ripe fruits were
col-lected in the populations of Hinojos and Grazalema, then placed in containers of water. These containers were then placed in the center of an open space, and the ground spread with canvas sheeting (203 20 m) to prevent any dispersed seeds from rolling after landing (Stamp and Lucas, 1983; Berg, 2000). The mean height of the capsules over the canvas sheeting was 46.7 cm for E. boetica and 56.3 cm for E. nicaeensis. When most of the capsules had exploded, the distance reached by each of the dispersed seeds was mea-sured, and the seed mass was noted. This experiment was performed four times for each species, always on sunny, windless days.
Frequency of caruncle loss in the two species—Ten ripe capsules were
collected from 30 to 50 individuals in each of the four populations. These capsules were taken to the laboratory and left to discharge separately inside paper envelopes. For each seed, the presence or absence of caruncle was then noted. Ecarunculate seeds and caruncles were weighed separately for each population, and the two largest axes of the seeds and the caruncles (length and breadth) were measured.
Data analysis—Skewness and kurtosis of dispersal curves were tested
us-ing the D’Agostino test (Zar, 1999). The median distances reached by the seeds after autochorous dispersal were compared between species using the Mann-Whitney test (Clewer and Scarisbrick, 2001) (data normalization by the usual transformations was not achieved). The same test was used to reveal any differences in the percentage of seeds with a caruncle between the pop-ulations of each species. Differences in the proportions of carunculate and ecarunculate seeds in each population were tested using chi-square tests. A Spearman correlation was used to reveal whether the dispersal distance of the seeds was dependent on seed mass. Interspecific differences in seed length and breadth were tested with MANOVA (with population nested within spe-cies), whereas those for seed mass were tested with ANOVA (with population nested within species).
The capsules of E. boetica and E. nicaeensis are pendulous during ripening and become erect as dispersal nears, so that at the moment of explosion the stalks bearing them are almost vertical (Fig. 1A–C). Under natural conditions, seed dispersal in E. boetica took place from 1300 hours until 2000 hours (Fig. 2). The peak dispersal was different on the 3 d studied: at 1400 hours the first day, 1700 hours the second, and 1500 hours the third. In no case did the peak coincide with the maximum temperature (Fig. 2). The capsules that had not ex-ploded by the end of the afternoon on the third day of the census were counted first thing the next day (at 0800 hours). Seven capsules (7.9% of the total for the day) had discharged during this period. In E. nicaeensis, dispersal began around 1200 hours and ended at 1900 hours (Fig. 2). As in E. boetica, the peaks of dispersal differed depending on the day studied: at 1800 hours the first day, 1300 hours the second, and 1700 hours the third. Again, none of the peaks coincided with the maximum temperature (Fig. 2). In E. nicaeensis, only five cap-sules (3.7% of the day’s total) discharged during the night.
The distribution of the distances reached by the seeds of E.
boetica was similar to a leptokurtic distribution (Z5 5.07, P
, 0.0001) and right-skewed (Z5 7.82, P, 0.0001; Fig. 3), with a peak at 100 cm from the source plant and a long tail that reached 800 cm. Some 75% of the seeds were found in a radius less than 2.5 m from the dispersing plant, and the me-dian distance reached by the seeds was 156 cm (mean 5 174.9; N5475). No correlation was found between seed dis-persal distance and seed mass (r5 20.02, t1275 20.21, P5
0.833). In the case of E. nicaeensis, the distribution of the distances reached by the seeds was similar to a mesokurtic distribution (Z 5 1.54, P 5 0.062), also right-skewed (Z 5 5.09, P , 0.0001; Fig. 3). In this species, the maximum dis-tance reached was 500 cm and the median disdis-tance was 132 cm (mean 5 148.8; N 5 305). In Euphorbia nicaeensis, a slightly significant negative correlation was found between seed dispersal distance and seed mass (r 5 20.13, t248 5
22.11, P 5 0.036). The distance of explosive dispersal was statistically different between species by the Mann-Whitney test (U 5 63 098, P 5 0.0023, N1 5 305, N2 5 475). The
coefficient of variation of the seed dispersal distance for E.
nicaeensis was greater than that for E. boetica (75.4 vs. 68.3,
The caruncle of E. nicaeensis seeds was considerably longer and broader than that of E. boetica (Fig. 1D–E). In E.
Fig. 1. Fruits and seeds of Euphorbia boetica and E. nicaeensis. (A) Aspect of E. boetica plant with numerous ripe capsules. (B) Brown, ripe and green, unripe capsules of E. boetica. (C) Ripe capsules with erect peduncles and unripe capsules with hanging peduncles of E. nicaeensis. (D) Carunculate seed of E.
Fig. 2. Hourly rates of explosive seed dispersal of Euphorbia boetica and E. nicaeensis (bars) and hourly temperatures (lines) during 3 d of the dispersal period.
of the caruncle of E. nicaeensis was 0.25 mg in Grazalema and 0.18 mg in Aracena, while that of E. boetica was 0.03 mg in Hinojos and 0.04 mg in El Gandul (N 5 20 for each population). In E. boetica, the proportions of seeds without a caruncle after dispersal was significantly higher than those of carunculate seeds (x25164.65, P,0.0001 at El Gandul and
x2 5 59.34, P , 0.0001 at Hinojos). In fact, only 20.0% of
those at Hinojos and 12.4% of those at El Gandul retained the
caruncle (Table 1; Fig. 1E–F); this difference between popu-lations was not significant by the Mann-Whitney test (U 5 696, P50.591, N1550, N2530). The loss or conservation
Fig. 3. Distances of explosive dispersal of Euphorbia boetica and E.
TABLE 1. Percentage of seeds with caruncle in Euphorbia boetica and E. nicaeensis and percentage of plants producing carunculate seeds, ecarunculate seeds, or a mixture of carunculate and ecarunculate seeds.
Seeds with caruncle (%) Mean6SE Median
Plants with carunculate seeds
Plants with ecarunculate seeds
Plants with carunculate and ecarunculate
E. boetica (Hinojos) 20.067.4 0 20 80 0
E. boetica (El Gandul) 12.464.5 0 8 86 6
E. nicaeensis (Grazalema) 98.361.1 100 90 0 10
E. nicaeensis (Aracena) 97.761.5 100 90 0 10
exclusively ecarunculate seeds was even greater, but in this case, a small percentage of plants produced seeds of both types (Table 1). In E. nicaeensis, the proportions of carunculate seeds after dispersal were significantly higher than those of ecarunculate seeds (x2 5 182.89, P , 0.0001 at Grazalema
andx25176.41, P,0.0001 at Aracena). In both populations,
practically all the seeds conserved their caruncle after dispersal (98.3% at Grazalema and 97.7% at Aracena; Table 1); there was no significant difference between the two populations by the Mann-Whitney test (U5 448.5, P5 0.982, N15 30, N2
5 30). In this species, most of the plants conserved the car-uncle in all the seeds, and only in some individuals was it lost in some seeds (Table 1). The seeds without a caruncle of E.
nicaeensis were significantly bigger (F1,112 5 18.8, P ,
0.0001) and significantly heavier (F1,11945242.9, P,0.0001)
than those of E. boetica (Table 2). DISCUSSION
In both species, the hourly pattern of capsule explosion was not regular through the day, there were marked peaks of dis-persal. Moreover, the hour at which the peaks were observed varied considerably between days. The irregular dispersal pat-tern observed in the species of Euphorbia contrasts markedly with the regularity found in other autochorous species that disperse most of their seeds during the morning (Turbull and Culver, 1983; Ohkawara et al., 1996).
The species of Euphorbia studied produced male cyathia at the lower levels of the inflorescence and the fruit-producing hermaphrodites at the upper levels (Narbona et al., 2000, 2002). The capsules adopt a vertical position just before dis-persal, maximizing explosive seed disdis-persal, as shown for oth-er species (Harpoth-er, 1977; Swaine et al., 1979; Stamp and Lu-cas, 1983; Thiede and Augspurger, 1996; Garrison et al., 2000). Beside intrinsic factors, dispersal distance is also af-fected by extrinsic factors, such as the vegetation and other obstacles, and wind speed (Garrison et al., 2000). Thus, the seed dispersal distances for the two Euphorbia species in nat-ural populations could differ from those found here as envi-ronmental conditions vary.
The species studied did not show a defined pattern. In E.
boetica, the distance reached by the seeds was independent of
seed mass, whereas in E. nicaeensis, there was a weakly sig-nificant negative correlation. The heaviest seeds might be ex-pected to have a shorter dispersal distance, as a result of a differential seed-fall pattern in which the heaviest, with greater resources and higher probability of survival, would remain closest to the parent plant (Morse and Schmitt, 1985). Nev-ertheless, no obvious relationship between seed mass and ex-plosive dispersal distance has been reported (Schmitt et al., 1985; Lisci and Pacini, 1997; Berg, 2000; Espadaler and Go´-mez, 2000).
myr-TABLE2. Characteristics of ecarunculate seeds of Euphorbia boetica and E. nicaeensis in the studied populations.
Seed length (mm) Mean6SE N Range
Seed breadth (mm) Mean6SE N Range
Seed mass (mg)
Mean6SE N Range
E. boetica (Hinojos) 2.760.03 30 2.2–3.1 1.860.02 30 1.6–2.1 5.0060.04 341 3.52–6.72
E. boetica (El Gandul) 2.660.03 30 2.3–2.9 1.860.02 30 1.6–2.0 4.6960.04 262 3.54–7.31
E. nicaeensis (Grazalema) 2.760.04 26 2.3–3.0 2.160.03 26 1.8–2.4 5.8160.05 284 4.12–8.02
E. nicaeensis (Aracena) 2.560.07 30 2.2–2.8 2.160.02 30 1.7–2.3 5.3660.05 189 4.07–7.25
mecochorous ants transport the seeds to a mean distance of 0.87 m (Go´mez and Espadaler, 1994). The dispersal by the ants could increase the seed dispersal distance in E. nicaeensis, but not in E. boetica whose seeds lose the caruncle.
Seeds of Euphorbia boetica were dispersed farther than those of E. nicaeensis. The fruit-bearing stems of E. nicaeensis were slightly higher than those of E. boetica (56.3 cm and 46.7 cm, respectively), so the height of the fruits on the plant did not appear to be responsible for the differences found be-tween both species. The seeds of E. nicaeensis were larger and heavier than those of E. boetica; the difference is even greater when the caruncle is included. Differences in seed mass should cause a difference in dispersal distances (Garrison et al., 2000). However, given that in this study, no obvious relationship be-tween seed mass and dispersal distance was found at intraspe-cific level, seed mass should not be the only factor causing the differences between species. Additionally, in diplochorous plants, a trade-off between the morphological requirements for long-distance explosive dispersal and the use of secondary dis-persal agents has been suggested; these plants may maximize either the primary or the secondary dispersal distance (Beattie and Lyons, 1975; Stamp and Lucas, 1983; Ohkawara and Hi-gashi, 1994). In fact, the presence of the elaiosome alters the seed aerodynamics and reduces the explosive dispersal dis-tance (Beattie and Lyons, 1975), but it is indispensable for myrmecochorous dispersal. The caruncle of E. boetica was much smaller than that of E. nicaeensis and was shed from most seeds of E. boetica during the explosion of the capsule. Thus, the interspecific differences in dispersal distance most likely were also due to the loss or conservation of the caruncle.
Euphorbia boetica would be maximizing its primary dispersal
distance, while E. nicaeensis would favor its myrmecochorous secondary dispersal. The conclusion that E. boetica maximizes its primary dispersal was supported by the fact that the seed dispersal distance for E. boetica varied less than the distance for E. nicaeensis, as expected for distance maximizers (Stamp and Lucas, 1983). A similar situation has been found in other diplochorous species-pairs of the same genus: in two species of Geranium, the species with the more effective explosive dispersal has seeds that are less appetizing to ants (Culver and Beattie, 1980; Stamp and Lucas, 1983), and in two of Viola, the species with greater explosive dispersal distance has a markedly smaller elaiosome (Ohkawara and Higashi, 1994).
In conclusion, the two species of Euphorbia presented an efficient mechanism of explosive dispersal, the seeds of E.
boetica were dispersed farther. The differences in explosive
dispersal distance between the species seemed to depend on both the seed mass and retention of the caruncle. The seeds of E. boetica shed the caruncle at the time of the explosion, favoring primary dispersal, while the seeds of E. nicaeensis had a larger, strongly attached caruncle, favoring their second-ary dispersal by ants.
ANDERSEN, A. N. 1988. Dispersal distance as a benefit of myrmecochory.
Oecologia 75: 507–511.
BEATTIE, A. J.,ANDD. C. CULVER. 1981. The guild of myrmecochores in the herbaceous flora of West Virginia forests. Ecology 62: 107–115. BEATTIE, A. J.,ANDN. LYONS. 1975. Seed dispersal in Viola (Violaceae):
adaptations and strategies. American Journal of Botany 62: 714–722. BENEDI´, C., J. MOLERO, J. SIMO´ N,ANDJ. VICENS. 1997. Euphorbia. In S.
Castroviejo et al. [eds.], Flora Iberica, vol. VIII, 210–285. Real Jardı´n Bota´nico de Madrid, CSIC, Madrid, Spain.
BERG, H. 2000. Differential seed dispersal in Oxalis acetosella, a cleistoga-mous perennial herb. Acta Oecologica 21: 109–118.
BERG, R. Y. 1975a. Fruit, seed, and myrmecochorous dispersal in
Micran-theum (Euphorbiaceae). Norwegian Journal of Botany 22: 173–194.
BERG, R. Y. 1975b. Myrmecochorous plants in Australia and their dispersal by ants. Australian Journal of Botany 23: 475–508.
BERG, R. Y. 1990. Seed dispersal relative to population structure, reproductive capacity, seed predation, and distribution in Euphorbia balsamifera (Eu-phorbiaceae), with a note on sclerendochory. Somerfeltia 11: 35–63. BULLOCK, S. H.,ANDR. B. PRIMACK. 1977. Comparative experimental study
of seed dispersal on animals. Ecology 58: 681–686.
CLEWER, A. G.,ANDD. H. SCARISBRICK. 2001. Practical statistics and ex-perimental design for plant crop science. John Wiley, Sussex, UK. CROIZAT, L. 1972. An introduction of the subgeneric classification of
Eu-phorbia L. with stress on the South African and Malagasy species. III. Webbia 27: 912.
CULVER, D. C.,ANDA. J. BEATTIE. 1980. The fate of Viola seeds dispersed by ants. American Journal of Botany 67: 710–714.
ESPADALER, X.,ANDC. GO´ MEZ. 1996. Seed production, predation and dis-persal in the Mediterranean myrmecochore Euphorbia characias (Eu-phorbiaceae). Ecography 19: 7–15.
ESPADALER, X.,ANDC. GO´ MEZ. 2000. Fı´sica i biologı´a en la dispersio´ de la lleteresa gran, Euphorbia characias L. (Euphorbiaceae). I. Jornades sobre la Recerca en els sistemes naturals de Collserola: aplicaciones a la gestio´ del Parc: 65–69.
EVANS, R. A., H. H. BISWELL,ANDD. E. PALMQUIST. 1987. Seed dispersal in Ceanothus cueatus and C. leucodermis in a Sierran oak-woodland savanna. Madron˜o 34: 283–293.
GARRISON, W. J., G. L. MILLER,ANDR. RASPET. 2000. Ballistic seed pro-jection in two herbaceous species. American Journal of Botany 87: 1257– 1264.
GO´ MEZ, C.,ANDX. ESPADALER. 1994. Curva de dispersio´n de semillas por hormigas en Euphorbia characias L. y Euphorbia nicaeensis All. (Eu-phorbiaceae). Ecologia Mediterra´nea 20: 51–59.
GO´ MEZ, C.,ANDX. ESPADALER. 1995. Variabilidad en la respuesta de
Phei-dole pallidula (Nyl.) como dispersante de semillas de especies del ge´nero Euphorbia L. Scientia Gerundensis 21: 49–57.
GO´ MEZ, C.,ANDX. ESPADALER. 1998. Seed dispersal curve of a Mediter-ranean myrmecochore: influence of ant size and the distance to nests.
Ecological Research 13: 347–354.
HARPER, J. L. 1977. Population biology of plants. Academic Press, New York, USA.
LISCI, M.,ANDE. PACINI. 1997. Fruit and seed structural characteristics and seed dispersal in Mercurialis annua L. (Euphorbiaceae). Acta Societatis
Botanicorum Poloniae 66: 379–386.
MORSE, D. H.,ANDJ. SCHMITT. 1985. Propagule size, dispersal ability, and seedling performance in Asclepias siryaca. Oecologia 67: 372–379. NARBONA, E. 2002. Estrategias reproductivas de dos especies perennes de
Euphorbia. Ph.D. thesis, Universidad de Sevilla, Sevilla, Spain.
especies perennes de Euphorbia (Euphorbiaceae): E. boetica Boiss. y E.
nicaeensis All. Anales del Jardı´n Bota´nico de Madrid 58: 183.
NARBONA, E., P. L. ORTIZ,ANDM. ARISTA. 2002. Functional andromonoecy in Euphorbia (Euphorbiaceae). Annals of Botany 89: 571–577. OHKAWARA, K.,ANDS. HIGASHI. 1994. Relative importance of ballistic and
ant dispersal in two diplochorous Viola species (Violaceae). Oecologia 100: 135–140.
OHKAWARA, K., S. HIGASHI,ANDM. OHARA. 1996. Effects of ants, ground beetles and the seed-fall patterns on myrmecochory of Erythronium
ja-ponicum Decne. (Liliaceae). Oecologia 106: 500–506.
PEMBERTON, R. W. 1988. Myrmecochory in the introduced range weed, leafy
spurge (Euphorbia esula L.). American Midland Naturalist 119: 431– 435.
SCHMITT, J., D. EHRHARDT,ANDD. SWARTZ. 1985. Differential dispersal of self-fertilized and outcrossed progeny in jewelweed (Impatiens capensis).
American Naturalist 126: 570–575.
STAMP, N. E.,ANDJ. R. LUCAS. 1983. Ecological correlates of explosive seed dispersal. Oecologia 39: 272–278.
STAMP, N. E.,ANDJ. R. LUCAS. 1990. Spatial patterns and dispersal distances
of explosively dispersing plants in Florida sandhill vegetation. Journal
of Ecology 78: 589–600.
SWAINE, M. D., T. DAKUBU,ANDT. BEER. 1979. On the theory of explosively dispersed seeds: a correction. New Phytologist 82: 777–781.
THIEDE, D. A.,ANDC. K. AUGSPURGER. 1996. Intraspecific variation in seed dispersion of Lepidium campestre (Brassicaceae). American Journal of
Botany 83: 856–866.
TURBULL, C. L.,ANDD. C. CULVER. 1983. The timing of seed dispersal in
Viola nuttallii: attraction of dispersers and avoidance of predators. Oec-ologia 59: 360–365.
WEBSTER, G. L. 1994. Classification of the Euphorbiaceae. Annals of the
Missouri Botanical Garden 81: 3–32.
WESTOBY, M.,ANDB. RICE. 1981. A note on combining two methods of dispersal-for-distance. Australian Journal of Ecology 6: 189–192. WILLSON, M. F. 1993. Dispersal mode, seed shadows, and colonization
pat-terns. Vegetatio 107/108: 261–280.
WOLFF, A.,ANDM. DEBUSSCHE. 1999. Ants as seed dispersers in a Medi-terranean old-field succession. Oikos 84: 443–452.