of Euphorbia characias seeds by ants

Acta CEcologica, 1997, 18 (1), 39-46

Soil surface searching and transport of Euphorbia characias seeds by ants Xavier Espadaler (*) and Crisanto G6mez (**) (*) CREAF, Universitat A utdnoma de Barcelona, 08193 BeUaterra, Spain. (**) Departament de Ci~nciesAmbientals, Universitat de Girona, Plafa Hospital, 6, 17001 Girona, Spain. Received: 10.6.96

Accepted: 25.10.96

Abstract The intensity of exploring the soil surface by ants was studied for the four species involved in the dispersal and predation of seeds of the West-Mediterranean myrmecochorous plant Euphorbia characias. During the dehiscence period (June) the whole soil surface is scanned in 43 minutes. Not all ants that find a seed take it to the nest. For the four ant species studied (Pheidole pallidula, Aphaenogaster senilis, Tapinoma nigerrimum, Messor barbarus) the proportion of ants that finally take the seed is 67.6%. In spite of this, the high level of soil surface searching explains the rather short time that seeds remain on the soil before being removed. The presence of an elaiosome is a key element in the outcome of the ant-seed interaction: a seed with elaiosome has a seven-fold increase in probability of being taken to the nest if found by a non-granivorous ant. The predator-avoidance hypothesis for myrmecochory is supported.

Keywords: Ants, seed dispersal, elaiosome, Euphorbia, myrmecochory, predation.

R6sm~6 L'intensit6 avec laquelle les fourmis explorent la surface du sol a 6t6 6tudi6e pour les quatre esp~ces qui participent ~ la dispersion et 5 la pr6dation des graines d'Euphorbia characias, une plante myrm6cochore de l'ouest m6diterran6en. Au cours de la p6riode de d6hiscence (juin), la surface enti~re du sol est pass6e au crible en 43 minutes. Toutes les fourmis qui trouvent une graine ne l'apportent pas dans le nid. Pour les quatre esp~ces 6tudi6es (Pheidole pallidula, Aphaenogaster senilis, Tapinoma nigerrimum, Messor barbarus), la proportion de fourmis qui emportent finalement la graine est de 67,6%. Malgr6 cela, l'intense exploration de la surface du sol explique le peu de temps pass6 par les graines sur le sol avant d'etre emport6es. La pr6sence d'un 61a'fosome est un 616ment-cl6 darts l'issue de 1'interaction fourmi-graine : une graine pourvue d'un 61a'fosome a sept lois plus de chances d'6tre emport6e dans le nid, si elle est trouv6e par une fourmi non-granivore. L'hypoth~se scion laquetle la myrm6cochorie permet d'6viter la pr6dation est 6tay6e.

Acta CEcologica 1146~509 X/97/0II$ 7.00/ 9 Gauthier-Vitlars

40

X. Espadaler and C. G6mez

INTRODUCTION

Myrmecochorous plants produce seeds that attract ants. The seeds are transported to the ants' nests. Seeds are provided with an elaiosome rich in lipids (BREW et al., 1989; KUSMENOGLUet al., 1989), whose high molecular weight suggests scant diffusion rates. Non-granivorous ants feed on the elaiosome and discard the seed inside the nest or in the refuse pile. Granivorous ants, however, eat seeds regardless of the presence of an elaiosome. If a seed does not have any chemical attractant this finding is also a matter of chance: the ant must pass nearby the seed and touch it. Vision may help but there are no data concerning detection of immobile objects in ants (HOIIDOBLER& WILSON, 1990). Several hypotheses have been proposed as explaining the advantages of myrmecochory (BEATTIE, 1985) but only a single one, predator avoidance, has the characteristic of being influenced by the time a seed is left on the soil, before it is found and recovered by a myrmecochorous ant. The sooner the seed is found by a dispersing ant, the better the opportunity for escaping predation. Seeds removed slowly are more vulnerable to predators (TURNBULL& CULVER, 1983; BOND & BREYTENBACH, 1985). Predators fall in three groups: rodents (REICHMAN, 1979; INOUYE et al., 1980), birds (MARES & ROSENZWEIG, 1978; KELRICKet al., 1986) and granivorous ants (ASHTON, 1979; ANDERSEN, 1987). Here we offer some evidence that the probability of escape from predation is rather strong in the case of the Mediterranean myrmecochore Euphorbia characias L. (Euphorbia hereafter). Seeds of this species are scattered ballistically all during the day and are transported by ants (BAIGES et al., 1991; GOMEZ, 1994). A low level of interspecific seed-robbing between dispersing ants has been detected (ESPADALER et al., 1996) though its significance for Euphorbia seedling recruitment has still to be evaluated. Seeds are not attractive from a distance and have to wait until a particular ant species finds it. This waiting time is short (mean 52.3-t-6.7 [s.e.] minutes; ESPADALER8Z; GOMEZ, 1996). The response of individual ants to different stimuli may depend on variables such as worker age, distance from the nest, prior experiences, resource availability, type of food or colony needs (Sul~D, 1987; TRANIELLO, 1987; TRANIELLOet al., 1992; FOURCASSII~& TRANIELLO, 1994). So, even if a seed is found, there is no guarantee that the ant will take it to the nest. We address the following questions: 1. What proportion of ants finding a seed removes it to the nest ? 2. Is the elaiosome a key element in seed transport ? 3. How intense is the searching of the soil surface by ants that retrieve Euphorbia seeds ?

METHODS Observations were made in a herbaceous community at the Collserola Park, an ecological preserve near Barcelona, NE Spain (41~ 2~ The site is an old field on former agricultural arable land, abandoned for 14 years. The site is located on an east-facing slope. The climate is Mediterranean, with 620 mm of annual rainfall. Mean monthly temperatures are highest in August with an average 22.6~ and a low of 7.1~ in January (data from 1914-1991). The site is a clearing bordered by a mixed wood. Vegetation in the clearing is mainly a herbaceous community of Inulo-Oryzopsietum miliaceae (BoLos, 1962). Though there are 17 ant species, only four remove Euphorbia seeds: Messor barbarus (granivorous), Pheidole pallidula (omnivorous), Aphaenogaster senilis (omnivorous) and Tapinoma nigerrimum (mainly nectarivorous). The total density of nests for the four species is 0.2 per m 2. During

Acta O~cologica

Soil surface searching by ants

41

the dehiscence period, in June, the soil surface is denuded, thus, ants can easily walk and retrieve items; there are no difficult areas to search such as steep or wet patches.

Data recording The proportion of seeds found and taken by ants was evaluated by offering intact single seeds (n = 590) to ants and noting if the seed was recovered after being found. The influence of the elaiosome was evaluated by offering single seeds (n = 361) with the elaiosome artificially removed. We recorded the behaviour of individual ants upon finding a seed as a) taking it to the nest or b) ignoring the seed. The intensity of surface soil searching by ants was evaluated by drawing the trajectory made by individual ants during a sampling period of 5 minutes on an acetate sheet with a square of 10 x 10 cm, on top of a wire frame at 5 cm from the soil. A single ant could be followed and traced even if more than one was present. If the ant we were tracing left the 10 x 10 area before the 5 minutes were over, the first new incoming ant was followed and so on. This constitutes a sample. A new sampling point was designated by randomly changing among six compass directions with a random distance from 1 to 6 meters from the previous point. Trajectories for each ant species were color coded. Data for Pheidole pallidula have been underestimated due to its small size and to its temporary disappearance from sight amid grasses or herbs. A total of 129 5-minutes samples was taken, beginning at 8 h in the morning and ending at 20 h in the evening. Data were obtained on different days (July 1991, June 1993, June and July 1994, May 1995), in the middle of the dehiscence period of Euphorbia.Trajectories were recorded only for those ant species involved in seed transport.

Data analysis The mean expected time (Te) a seed remains on soil surface after being released by the mature plant depends on a) the time needed for a complete searching of the surface (Tr) and, thus, for a seed to be found and b) the probability (/9) of a seed that has been found is transported to the nest. Not all seeds found are actually taken; a fraction of ballistically dispersed seeds will remain on soil and, in turn, will wait a new Tr and the process will iterate "a" times, as expressed in (1)

(t)

~=~

aTrrp(1-p)~-1 1

Tr was obtained by relating the surface to be scanned and the surface actually scanned by ants per time unit (Vr). The surface to be scanned is the sample of 10 x 10cm (S). The surface occupied by the seed was not considered since it is extremely small (7.43 x 10 - 4 cm 2, based on the estimated seed density of 2.29 seeds/m2/day and seed size of 3.65 mm length x 2.10 mm width; ESPADALER& GOMEZ, 1996). Vr, the surface searched per time unit, depends on the surface searched by each species. When foraging, ants walk with the funiculi directed somewhat outside. Since species differ in size, the antennal gap (= distance between tip of antennae) was measured on photographs of foraging ants for each species. Size of antennal gap of each species was used to construct a grid of small squares; thus, each species has a particular "searching grain". Each specific grid was superimposed on actual trajectories and the number of squares traversed by or touched by the trajectories were counted. From this count for all samples we estimated the surface searched by each species during 5 minutes. Interspecific overlaps of trajectories were calculated as rectangles of size "antennal gap sp. 1 • gap sp.2" and were discounted from Vr. Intraspecific overlaps were also discounted from Vr. Tr, the time needed for a complete coverage of the soil surface is:

(3~

Vol. 18, n ~ 1 - 1997

T~ = S / V r

42

x . Espadaler and C. G6mez

RESULTS Response of ants to seeds The proportion of ants taking a seed after finding it shows an overall value of 0.676 (table I). There is a continuum of probabilities of dispersing a seed upon finding it. The highest response comes from the granivorous M. barbarus, with p = 0.82 of taking an Euphorbia seed with elaiosome or p = 0.88 if the seed is deprived of the elaiosome though difference was not significant (Gaaj = 1.61; p = 0.199). The other three species show varying degrees of positive response to seeds: P. pallidula p = 0.72, T. nigerrimum p = 0.66 and A. senilis 19 = 0.49. If seeds lack the elaiosome the probability for a seed to be retrieved is sharply reduced and differences are highly significant for each species: Pheidole (Gadj = 60.9; p < 0.001), Tapinoma (Gaaj = 91.4; p < 0.001), Aphaenogaster (G~oj = 47.1; t9 < 0.001). The probability of a seed with elaiosome being taken to the nest after being found by a non-granivorous species is 0.63. When seeds are deprived of the elaiosome this probability falls to 0.09. TABLE I. -- Ant reaction to single seeds of E. characias. Number of positive (percentage) (Y; takes the seed to the nest) and negative (N," does not take the seed) responses when finding seeds, with (A) or without (B) the elaiosome. Differences between A and B for each species were evaluated with a G-test with Williams' correction (SOKAL & ROHLF, 1995, p. 729-732). ***: p < 0.001; n,s.: no difference. Since the number o f seeds offered are different for each species, the probability that a seed will be taken by any ant species was obtained from % of positive response and not from raw data. For seeds with elaiosomes, p = 0,68 for removal by any ant species; for seeds without elaiosomes, p = 0.09 for removal by any non-granivorous species. (A) Y N (B) Y N

T. nigerrimum

A. senilis

P. pallidula

M. barbarus

53 (66.25) 27 (37.75)

124 (49.6) 126 (50.4)

10t (72,t2) 39 (27.9)

99 (82.5) 21 (17.5)

T. nigerrimum

A. senilis

P. pallidula

M. barbarus

0 (0) 70 (100)

15 (13.39) 97 (86.61)

8 (13.56) 51 (86.44)

106 (88.33) 14 (11.67)

***

***

***

n.s.

Soil surface searching Vr and mean searching time. The distance between antennal tips (funiculi) was 4 mm (n = 5) for T. nigerrimum, 6.5 mm (n = 5) for A. senilis, 2.5 mm (n = 4) for P. pallidula and 6.5 mm (n - 6) for M. barbarus (medium workers). The number of differently sized squares traversed by each ant species varied during the day (table II). The mean number of squares and the actual surface scanned per sample of 5 minutes (from 129 samples) was: 32.64 squares o f 4 x 4 m m (5.222 cm 2) for T. nigerrimum, 8.7 squares of 6.5 • 6.5 m m (3.676 cm 2) for A. senilis, 10.45 squares of 2.5 x 2.5 m m (0.653 cm 2) for P. pallidula and 8.37 squares of 6.5 x 6.5 m m (3.536 cruZ) for M. barbarus. The total surface scanned per sample of 5 minutes was 13.087 cruZ. Since mean interspecific overlap of trajectories per 5 minute samples is 0.29 cm 2 and intraspecific overlap is 1.216 cm 2 Acta (Ecotogica


43

TABLE II. - Mean number (s.e.) of differently sized squares (see text) explored by individuals of four ant species followed through an acetate sheet of 100 cm 2 for 12 hours. Time, 5-minute samples (n) and species involved. Time

n

T. nigerrimum

8-9 9-10 10-11 11-12 12-13 13-14 14-15 t5-16 16-17 17-18 18-19 19-20

8 15 16 12 15 5 4 6 6 15 12 15

27.7 70.3 32.0 37.8 6.8

7.7 70.7 56.8 28.3 15.4

(9.8) (21.1) (12.2) (19.8) (4.9) 0 0 (6,5) (21.0) (15.6) (8.2) (9.4)

A. senilis

P. pallidula

3.2 6.6 18.0 15.4 1.2 10.4

18.9 4.2 4.4 5.1

(3.1) (3.3) (5.4) (6.6) (0.9) (6.0) 0 24.0 (11.4) 23.2 (13,8) 11.4 (4.2) 0 0

M. barbarus

(9.1) (3.6) (4.4) (3.6) 0 0 0 9.7 (9.7) t2.8 (12.8) 36.0 (15.3) 3.9 (3.1) 18.7 (12.8)

8.8 (8.8) 2.1 (1.5) 4.4 (3.0) 0 1.6 (1.6) 0 0 0 4.2 (4.2) 10.2 (3.8) 32.9 (10.2) 20.6 (6.8)

TABLE 1II. -- Number o f intra and interspeciflc overlaps among trajectories. Global data for 129 samples. T. nigerrimum

A. senilis

P. pallidula

M. barbarus

342 -

71 61 -

9 6 40 -

46 11 2 175

T. nigerrimum A. senilis P. pallidula 114, barbarus T0 ~

9

60

~" 5 0

~40 "E 30

20 10

0

1

2

3

4

5

6

7

8

a

Fl~. 1. - The mean expected time (Te) a seed of E. characias with an elaiosome remains on the soil surface after being released by the mature plant as a function of the number of times (a) it is found by individual ants. A seed has a p = 0.676 of being recovered after it is found by an ant. (Te) roughly saturates after five iterations at a value of 62.8 minutes.

(table III), Vr is reduced to 11.581 cm 2. Tr, the time needed for a complete searching of the surface would be, according to (2), 43.17 minutes. The first time this process occurs, a 67.6% of seeds are retrieved; after a second iteration, 67.6% of those still remaining are taken and so on; after five iterations 99.38% of seeds are recovered (fig. 1). Applying a = 5, Tr = 43.17 and p = .676 in (1) the mean expected time (Te) a seed remains on soil surface before being transported to a nest is 62.8 minutes.

Vol. 18, n ~ 1 - 1997

44

x. Espadaler and C. G6mez

DISCUSSION

Ants "...essentially vacuum the soil surface" (REICHMAN, 1979). Here we have shown that soil surface searching by myrmecochorous and granivorous ants is rather intense. Furthermore, our samples are also underestimates since only a single ant could be followed at a time. We do not give too much value to the similarity of Te (62.8 minutes) with the mean time of seed retrieval found by a different method (offering single seeds and waiting until an ant takes it: 52.3 minutes; ESPADALER & GOMEZ, 1996). Our reasoning depends critically on antennal gap at the tip of funiculi, a limit that may be somewhat variable and numbers would change accordingly. In spite of this, we feel confident that our measurements give a correct sense of actual soil surface searching by this group of ants. WEHNER (1987) showed very high search densities by desert ants and that the foraging activities of adjacent colonies evenly cover the entire ground surface. Ants are ubiquitous (WILSON, 1987) and are the main gatherers of animal remains (JEANNE, 1979; BARONI URBANI & AKTA~, 1981; FELLERS & FELLERS, 1982; RETANA etal., 1991). Ants are also assumed to be important seed predators in most plant communities in Australia (O'DOWD & GILL, 1984; ANDERSEN, 1987) and dry regions of Central Africa (Lt~VIEUX & DIOMANDE, 1978), South Africa (KERLEY, 1991), America (MARES & ROSENZWEIG, 1978; REICHMAN, 1979; HOBBS, 1985) and in Costa Rica lowland rain forest (LEvEY & BYRNE, 1993). As has been recently shown, elaiosomes can be regarded as insect analogues (HUGHES et al., 1994) and, from a nutritional standpoint, myrmecochorous seeds may be equated with insect remains for some ant species. Our results are similar to those found in other studies on removal rates (Lu & MESLER, 1981; DRAKE, 1981; MAJER & LAMONT, 1985; OHARA & HIGASHI, 1987; OOSTERMEIJER, 1989; IRELAND & ANDREW, 1995) and on the effect of the elaiosome (AULD, 1986; HUGHES & WESTOBY, 1992). The high degree of soil surface searching measured here and the strong effect of the presence of an elaiosome agree with previous knowledge. The presence of the elaiosome made no significant difference to removal rates by Messor barbarus, the only purely granivorous species, but made a large difference to the behaviour of the other three species: The estimates of retention times based on ant foraging behaviour and size are consistent with the hypothesis that one of the main selective advantages of elaiosomes is to ensure rapid removal of seeds. Seeds without elaiosome have a much lower probability of being dispersed by non-granivorous ants, implying a delicate balance between predation and dispersal. This gives support to the idea that elaiosomes increase the probability of seed removal by ant species that will leave at least some uneaten and so, for the hypothesis of predator avoidance as an advantage of myrmecochory, as suggested by HUGHES& WESTORY(1990), HUGHES et al. (1994) and COWliNG et al. (1994).

ACKNOWLEDGMENTS We are most grateful to Cesare BARONIURBANI,Johan BILLEN, Ram6n MARGALEF,Josep PIIqOL and James F. TRAN1ELLOfor helpful discussions and/or comments on earlier versions of the manuscript. Lesley HUGHESrevised and corrected two versions and we mostly acknowledge her patience. We thank the permission of the Sangrh family and Josep MORERAfor allowing our work inside their private properties;

Acta t~Ecologica


45

G, GENOVE, X. ROIG, Anna Bou and Gemma PASCUALhelped with field measurements. This work has been financed trough a grant from DGICYT (PB91-0482).

REFERENCES ANDERSEN A., 1987. - Effects of seed predation by ants on seedling densities at a woodland site in SE Australia. Oikos, 48, 171-174. ASHTON D. H., 1979. - Seed harvesting by ants in forests of Eucalyptus regnans F. Muell. in Central Victoria. Aust. J. Ecol., 4, 265-277. AULD Z. D., 1986. - Population dynamics of the shrub Acacia suaveolens (Sm.) Willd.: Dispersal and the dynamics of the soil seed bank. Aust. J. Ecol., 11, 235-254. BAIGES J. C., ESPADALERX. & BLANCHe.C., 1991. - Seed dispersal in W Mediterranean Euphorbia species. Bot. Chron., 10, 697-705. BARONI URBANl C. & AKTA~ N., 1981. - The competition for food and circadian successions in the ant fauna of a representative Anatolian semi-steppic arid environment. Mitt. Schweiz. ent. Ges., 54, 33-56. BEATrIE A. J., 1985. - The evolutionary ecology of ant-plant mutualisms. Cambridge University Press. Cambridge, 182 p. BOLOS O. de., 1962. - El paisaje vegetal barcelon~s. Universidad de Barcelona, Barcelona, 192 p. BOND W. J. & BREYTENBACHG. J., 1985. - Ants, rodents and seed predation in Proteaceae. South Afr. J. Zool., 20, 150-154. BREW C. R., O'DOWD D. J. & RA~ I. D., 1989. - Seed dispersal by ants: behaviour-releasing compounds in elaiosomes. Oecologia, 80, 490-497. CowuNc R. M., PIERCE S. M., STOCKW. D. & COCKS M., 1994. - W h y are there so many myrmecochorous species in the Cape fynbos ? In: ARIANOUTSOUM. & GROVES R. H., eds., Plant-Animal Interactions in Mediterranean-Type Ecosystems. Kluwer, The Netherlands, 159-168. DRAKEW. E., 1981. - Ant-seed interaction in dry sclerophyll forest on North Stradbroke Island, Queensland. Aust. J. Bot., 29, 293-309. ESPADALERX., GOMEZC. & SUIqERD., 1996. - Seed-robbing between ants intervenes in the myrmecochory of Euphorbia characias (Euphorbiaceae). Psyche, 102, 19-25. ESPADALER X. t~ GOMEZ C., 1996. - Seed production, predation and dispersal in the Mediterranean myrmecochore Euphorbia characias L. (Euphorbiaceae). Ecography, 19, 7-15. FELLERS G. M. & FELLERS J. n . , 1982. - Scavenging rates of invertebrates in an eastern deciduous forest. Amer. Midl. Nat., 107, 389-392. FOURCASSII~ V. t~ TRANIELLOJ. F. A., 1994. - Food searching behaviour in the ant Formica schaufussi (Hymenoptera, Formicidae): response of naive forages to protein and carbohydrate food. Anim. Behav., 48, 69-79. GOMEZ C., 1994. - Dispersi6 de llavors per formigues: llavors d'Euphorbia characias characias Linnaeus (Euphorbiaceae) a la Serra de Collserola (NE Espanya). Ses. Entom. ICHN-SCL, 8, 117-124. HOI3BS R. J., 1985. - Harvester ant foraging and plant species distribution in annual grassland. Oecologia, 67, 519-523. HOLLDOBEER B. ~e; WILSON E. O., 1990. - The ants. Springer Verlag, Berlin, 732 p. HUGHES L. & WESTOBY M., 1990. - Removal rates of seeds adapted for dispersal by ants. Ecology, 71, 138-148. HUGHES L. • WESTOBY M., 1992. - Effect of diaspore characteristics on removal of seeds adapted for dispersal by ants. Ecology, 73, 1300-1312. HUGHES L., WESTOBY M. t~; JURADO E., 1994. - Convergence of elalosomes and insect prey: evidence from ant foraging behaviour and fatty acid composition. Funct. Ecol_, 8, 358-365. INOUYE R. S., BYERS G. S. & BROWN J. H., 1980. - Effects of predation and competition on survivorship, fecundity and community structure of desert annuals. Ecology, 61, 1344-1351. IRELAND C. t~; ANDREW M. H., 1995. - Ants remove virtually all western myall (Acacia papyrocarpa Beuth.) seeds at Middleback, South Australia. Aust. J. Ecol., 20, 565-570. JEANNE R. L., 1979. - A latitudinal gradient in rates of ant predation. Ecology, 60, 1211-1224.

Vol. 18, n ~ 1 - 1997

46

X. Espadaler and C. G6mez

KELRICK M. I., MACMAHON J. A., PARMENTERR. R. & SISSON D. V , 1986. - Native seed preferences of shrub-steppe rodents, birds and ants: the relationship of seed attributes and seed use. Oecologia, 68, 327-337. KERLEY G. I. H., 199I. - Seed removal by rodents, birds and ants in the semiarid Karoo, South Africa. J. Arid Envir., 20, 63-69. KUSMENOGLUS., ROCKWOODL. L. & GRETZM. R., 1989. - Fatty acids and diacylglycerols from elaiosomes of some ant-dispersed seeds. Phytochemistry, 28, 2601-2602. LEVEY D. J. & BYRNE M. M., 1993. - Complex ant-plant interactions: rain forest ants as secondary dispersers and post-dispersal seed predators. Ecology, 74, 1802-1812. L~vIEux J. & DIOMANDE T., 1978. - La nutrition des fourmis granivores. I. Cycle d'activit6 et rrgime aIimentaire de Messor galla et de Messor (= Cratomyrmex) regalis (Hymenoptera, Formicidae). Insectes Sociaux, 25, I27-139. Lu K. L. & MESLER M. R., 1981. - Ant dispersal of a neotropical forest floor Gesneriad. Biotropica, 13, 159-160. MAJER J. D. & LAMONT B. B., 1985. - Removal of seed of Grevillea pteridi)blia (Proteaceae) by ants. Aust. J. Bot., 33, 611-618. MARES M. A. & ROSENZWEIG M. L., 1978. - Granivory in North and South American deserts: rodents, birds and ants. Ecology, 59, 235-241. O'DOWD D. J. & GILL A. M., 1984. - Predator satiation and site alteration following fire: mass reproduction of Alpine Ash (Eucalyptus detegatensis) in Southeastern Australia. Ecology, 65, 1052-1066. OHARA M. • HIGASHI S., 1987. - Interference by ground beetles with the dispersal by ants of seeds of Trillium species (Liliaceae). J. EcoL, 75, 1091-1098. OOSTERMEIJER J. G. B., 1989. - Myrmecochory in Polygala vulgaris L., Luzula campestris (L.) DC. and Vola curtisii Forster in a Dutch dune area. Oecologia, 78, 302-311. REICHMAN O. J., 1979. - Desert granivore foraging and its impact on seed densities and distributions. Ecology, 60, 1085-1092. RETANA J., CEP.DA X. & ESPADALERX., 1991. - Arthropod corpses in a temperate grassland: a limited supply ? Hol. EcoL, 14, 63-67. SOKAL R. B.