Invasion on Pollination of Euphorbia Characias (L.) (Euphorbiaceae)

Home , Euphorbia characias

Acta Oecologica 28 (2005) 49–55 www.elsevier.com/locate/actoec Original Article Consequences of the Argentine ant, Linepithema humile (Mayr), invasion on pollination of Euphorbia characias (L.) (Euphorbiaceae)

Xavier Blancafort *, Crisanto Gómez

Department of Environmental Sciences. University of Girona. Campus de Montilivi. 17071-Girona, Spain

Received 20 October 2004; accepted 15 February 2005 Available online 07 April 2005

Abstract

We have studied the influence of the Argentine ant, Linepithema humile, on the pollination of Euphorbia characias, a deciduous insect- pollinated shrub. The observations were made in two adjacent areas (invaded and non-invaded by L. humile) in a Mediterranean cork-oak forest. In the invaded area, L. humile has replaced most of the native ants that climb up this plant’s inflorescences. Five native ant species were detected in the non-invaded areas and only one in the invaded area. The number of visitors to infested inflorescences (1.54 ± 1.86 visitors/10 min observation) was lower than in non-infested inflorescences in the invaded area (3.74 ± 4.19 visitors/10′), and in the non-invaded areas (4.16 ± 5.00 visits/10′). For several species of flower-visiting insects, no differences were detected between the time spent in the flowers and the number of flowers visited in the two areas, except for Eristalis tenax, a dipteran which visited more flowers (15.2 ± 11.1 flowers visited/10′) and spent more time (9.4 ± 5.8 sec) in the non-invaded area than in the invaded area (7.8 ± 8.2 flowers visited/10′ and 5.3 ± 2.1 sec, respectively). The relative representation of insect orders in the two areas was not different. A significant reduction in fruit-set and seed-set was detected in the invaded area. These results suggest that the Argentine ant may greatly affect the reproductive success of components of the Mediterranean flora. © 2005 Elsevier SAS. All rights reserved.

Keywords: Argentine ant; biological invasion; Euphorbia characias; Linepithema humile; pollination

1. Introduction influences the seed-set and the quality of descendants (Her- rera, 2000). This is an important issue, if we consider the Pollination is one of the most important types of interac- possible variation in the assemblages of pollinators, due to tion between plants and animals in ecosystems because it is a the presence and spread of invasive introduced species. key process in the sexual reproduction of most angiosperms TheArgentine ant is one of the most invasive species docu- and can affect directly the plant reproduction success (Dafni, mented. Native to South America, it has invaded all Mediter- 1992; Kearns and Inouye, 1993). Most animal–plant interac- ranean ecosystems of the world (Suarez et al., 2001). Two tions are not species-specific relationships (Dafni and supercolonies of Argentine ants are present in a 6000-km O’Toole, 1994; Gomez and Zamora, 1999; Herrera, 1995; stretch of the Mediterranean and Atlantic coast of southern Zamora, 2000). For example, flowers of Mediterranean plant Europe (Espadaler and Gómez, 2003; Giraud et al., 2002). In species are usually visited by insects belonging to three or the invaded areas the Argentine ant’s presence usually exerts more orders (Herrera, 1996). Among plants visited by many a negative influence on invertebrate biodiversity (Cole et al., pollinator species, the relative contribution of each pollinator 1992; Holway, 1998; Human and Gordon, 1997). These to plant reproduction is determined by variation in both pol- changes can influence ecological processes such as pollina- linator and plant traits (Pellmyr and Thompson, 1996). The tion and seed dispersal, two key processes for the reproduc- presence of different pollinators with different efficiencies tive success of plants. Several studies examine the effect of the Argentine ant on seed dispersal by ants in invaded zones * Corresponding author. Fax: +34 972 41 81 50. (Christian, 2001; Carney et al., 2003; Gómez and Oliveras, E-mail addresses: [email protected] (X. Blancafort), 2003; Gómez et al., 2003) but few studies have examined the [email protected] (C. Gómez). impacts of this invader on the pollination. A reduction in the

1146-609X/$ - see front matter © 2005 Elsevier SAS. All rights reserved. doi:10.1016/j.actao.2005.02.004 50 X. Blancafort, C. Gómez / Acta Oecologica 28 (2005) 49–55 diversity and abundance of insects in Protea nitida inflores- ished at 19:00 h (local time). We alternated the different inflo- cences infested by L. humile has been observed in South Afri- rescences filmed (two from the invaded area and two from can fynbos (Visser et al., 1996), and Argentine ant workers the non-invaded zone) during the study day. The samples attack nests of pollinators (Buys, 1990). filmed were watched in the laboratory and the insect visitors The main objective of this work is to document and dis- that touched the flowers’ anthers or pistils were noted. The cuss the impact of the Argentine ant invasion on the pollina- number of visits per flower and the number of flowers visited tion of Euphorbia characias (L.) by comparing pollination every 10 min were compared between the two areas with in invaded and non-invaded areas. This work is also the first paired t-tests (each pair correspond to each study day). All to our knowledge to document which insects visits and poten- visitors were identified and for each visitor we noted the num- tially pollinate this plant species. Other studies of pollination ber of flowers visited, the time spent in each flower visited, in Euphorbia species have focussed on Euphorbia albomar- and the distance moved between visits to consecutive flowers ginata, E. capitellata, E. hyssopifolia in Arizona and New in the same inflorescence. As the data were obtained through Mexico (Ehrenfeld, 1979) and on the shrub E. dendroides on the filmed sample, the number of visits was for half of each the island of Cabrera (Balearic Islands) (Traveset and Sáez, inflorescence. 1997). Concerning the scan study, we also noted all visitors to the We studied (a) the number and diversity of visitors and plants (adult plants of E. characias can have from 1 to pollinators, (b) behavioral aspects of visitors and pollinators 50 inflorescences) during the video camera censuses. The dif- (time spent in each flower, number of flowers [cyathia] vis- ferent species of insect visitors were collected and catego- ited and distance moved between visited flowers), and (c) rized in to orders and the proportion of species belonging to fruit-set, i.e., the proportion of flowers that set fruit, and seed- each order was compared between the two areas using a chi- set, i.e., the number of viable seeds produced by fruit. square test. In order to check a possible “repellent” effect of the presence of the Argentine ant on the visitors, we counted the num- 2. Material and methods ber of visitors to the following groups of plants: • visitors in the non-invaded areas (96 inflorescences); 2.1. Study area • visitors to plants infested by the Argentine ant (47 inflorescences) and ; This study was carried out between February and July • visitors to plants in the invaded zone, but on which work- 2002 in the Serra Llonga, on the southern limit of Gavarres ers of the Argentine ant were absent (46 inflorescences). Massif (NE Spain) (41° 49′ N, 3° 00′ E). The study area was The number of visitors (log-transformed) was analyzed using 4 km from the Mediterranean coast. The climate in this region a two-way ANOVA with plant groups and days as fixed is Mediterranean subhumid, with a mean of 627 mm of annual effects. Differences in numbers of visitors found between rainfall. Invaded and non-invaded zones are situated at 250 m groups of plants could be due to variation in the density of E. elevation and the vegetation is open cork-oak secondary for- characias shrubs. To explore this possibility, we counted the est dominated by Quercus suber (L.), Quercus ilex (L.), and number of E. characias individuals in three 15 × 15 m plots Pinus pinaster (Ait.). at each zone and the density were compared with a one-way Euphorbia characias (L.) is a deciduous shrub 30–100 cm ANOVA. tall, growing in scrub, cork-oak and holm-oak forests (Bolòs et al., 1993). The genus Euphorbia has simple flowers gath- 2.3. Fruit-set and seed-set ered in a single structure: the cyathium. Each cyathium con- sists of 5 male flowers, each with only 1 stamen, a pedicel, To compare the fruiting efficiency and the seed-set medi- and no perianth, surrounding an apical female flower, also ated by different types of pollinators in the invaded and non- with pedicel and without perianth. This grouping of uni- invaded zones, we developed a pollinator exclosure experi- sexual flowers in a bisexual inflorescence surrounded by col- ment. Before the flowering period, we applied four different ored petaloid appendages and nectaries is functionally equiva- exclusion treatments to four inflorescences per plant. We chose lent to a hermaphrodite flower (Strasburger et al., 1994). 10 E. characias plants in each zone, invaded and non- invaded, and applied the following treatments: 2.2. Diversity and abundance of flower visitors • Exclosure of flying insects: the inflorescences were covered with a nylon bag. During the flowering season of E. characias, we per- • Exclosure of ants and other non-flying insects: the inflo- formed a focal study and a scan study in a zone with invaded rescence bases were coated with colorless and odorless and non-invaded areas next to each other. Concerning the focal non-poisonous glue (Ratimur®). study the observations were focused on the inflorescence. We • Exclosure of all insects: the inflorescences were covered chose 5 plants in each zone and filmed different inflores- with a nylon bag and coated with glue at the base. cences with a video camera for 10 min once a week (one day • Control: all insects could access the inflorescence. a week), for 8 weeks. The censuses started at 9:00 h and fin- The number of flowers in each inflorescence and the number X. Blancafort, C. Gómez / Acta Oecologica 28 (2005) 49–55 51 of fruits were counted weekly until the end of the flowering visited 141 flowers. The number of visited flowers/10′ con- period. Ripe fruits were collected and transported to the labo- sidering all observations was 2.84 ± 2.48 (mean ± sd) in the ratory. The fruits were cut with a scalpel and the number of non-invaded area and 1.42 ± 0.93 in the invaded area. The seeds in each fruit was counted (maximum number was three). number of flowers visited per day in the two areas during the Seed viability was tested using 2,3,5-triphenyltetrazolium flowering period was the same (paired t-test, t = 2.0, P = 0.08). chloride (TTZ) (Scharpf, 1970) on a varying number of seeds However, the species composition of visitors was different from each treatment type and zone (from a minimum of between the two areas. 27 seeds from the total exclusion treatment in the invaded For flower-visiting species present in both areas, no differ- zone to a maximum of 105 seeds from the ant exclusion treat- ences in the time spent in flowers and the number of flowers ment in the non-invaded zone; 549 seeds analyzed in all), and visited were detected between the two areas, except for Eri- compared with a test for difference between pairs of percent- stalis tenax, a dipteran which visited more flowers and spent ages. Fruit-set (data arc sin square-root transformed) and seed- more time in the non-invaded area (Table 1). The distance set were compared among treatments with a two-way moved between consecutive flower visits was similar for all ANOVA. All analyses were performed with the SPSS pack- species. We also compared the data for Camponotus cruen- age for Windows, v11.1.3 (©SPSS Inc). tatus and Linepithema humile, the most common ant species in each area. C. cruentatus moved greater distances between flowers searching for nectar. On the other hand, the Argen- 3. Results tine ant was more erratic. Workers of Camponotus cruentatus are larger (dry body-mass = 1.95–6.78 mg) than L. humile 3.1. Diversity and abundance of flower visitors (dry body-mass = 0.11 mg) and its body always touched the anthers or the pistils of flowers while, the Argentine ant work- The flowering period of E. characias starts in March and ers only touched them sometimes (data of ant body sizes ends in April, lasting seven weeks. We performed a total of obtained from the Iberian ant sizes data base [Gómez and 96 censuses in the non-invaded area and 93 in the invaded Espadaler., 2000] and calculated by applying the transforma- area. tion from Kaspari and Weiser, 1999). Concerning the focal study, the number of visits to the Concerning the scan study, 32 species visited E. charac- filmed inflorescences that we used to evaluate at floral-visit ias inflorescences in the non-invaded area, 28 species in the level the effect of Argentine ant invasion was 47 in the non- invaded area (39 species in total; 21 species were common to invaded area and 32 in the invaded area. The number of visi- both areas). Dipterans accounted for 58% of flower visitors tors per flower/10′ min considering all observations was in the non-invaded area and 40.3% in the invaded area. Ants 0.025 ± 0.011 (mean ± sd) in the non-invaded area and comprised 29.4% of the visitors in the non-invaded area and 0.014 ± 0.006 in the invaded area. No differences were 51.2% in the invaded area. The relative representation of insect detected between the number of daily visitors in the two areas orders in the two areas was not different (Chi-square = 2.66, during the flowering period (two paired t-test, t = 2.27, P = 0.10). The most common visitors were the same in both P = 0.58). The 47 visitors in the non-invaded area visited a areas. However, we detected an important difference in the total of 310 flowers, and the 32 visitors in the invaded area diversity of ant species visiting the inflorescences. Five ant

Table 1 Time spent in a flower (in sec), distance moved between flowers (in cm) and the number of visited flowers for the most common visitor species of Euphorbia characias inflorescences (nia: non-invaded area, ia: invaded area, t-Student test, * P < 0.05) Visited flowers Distance moved Time in flower Species N Mean ± sd TPNMean ± sd tPNMean ± sd tP Calliphora vomitoria nia 40 8.6 ± 8.1 0.26 0.79 32 2.5 ± 2.2 0.11 0.91 37 8.1 ± 5.0 1.09 0.27 (Diptera: Calliphoridae) ia 33 9.1 ± 6.3 31 2.5 ± 2.2 33 9.8 ± 7.7 Eristalis tenax nia 11 15.2 ± 11.1 2.07 0.04* 11 2.1 ± 1.2 0.66 0.51 11 9.4 ± 5.8 2.70 0.01* (Diptera: Syrphidae) ia 19 7.8 ± 8.2 15 2.6 ± 2.4 18 5.3 ± 2.1 Gonia divisa nia 6 13.3 ± 11.8 0.37 0.71 4 2.6 ± 2.1 0.11 0.91 5 5.3 ± 0.9 1.54 0.15 (Diptera: Tachinidae) ia 9 10.9 ± 12.7 7 2.4 ± 1.9 8 7.8 ± 3.5 Sarcophaga carnaria nia 3 15.3 ± 2.5 0.65 0.54 3 1.1 ± 0.4 0.78 0.46 3 9.0 ± 2.4 0.63 0.55 (Diptera: Sarcophagidae) ia 8 12.3 ± 7.7 6 1.8 ± 1.3 7 7.3 ± 4.2 Episyrphus balteatus nia 3 10.0 ± 14.7 0.20 0.84 2 3.4 ± 0.9 1.04 0.37 3 18.0 ± 4.6 0.60 0.18 (Diptera: Syrphidae) ia 4 8.3 ± 8.6 3 4.6 ± 1.4 3 12.9 ± 3.0 Camponotus cruentatus nia 20 11.1 ± 6.6 2.00 0.05 17 1.4 ± 0.6 2.15 0.04* 20 25.2 ± 27.8 0.77 0.44 Linepithema humile ia 8 6.1 ± 3.4 8 0.7 ± 1.1 8 16.8 ± 21.1 Camponotus cruentatus and Linepithema humile were compared because they were the most common ant species that visited flowers in the non-invaded and invaded area, respectively. 52 X. Blancafort, C. Gómez / Acta Oecologica 28 (2005) 49–55

Table 2 Number of visitors per plant during 10-min timed censuses in both areas (mean ± sd) Non-invaded area N Invaded area N Diptera 0.25 ± 0.16 229 0.22 ± 0.11 199 Family Syrphidae Eristalis tenax (L) 0.03 ± 0.02 31 0.03 ± 0.04 28 Episyrphus balteatus (De Geer) 0.01 ± 0.03 14 0.01 ± 0.02 9 Syrphidae A 0.01 ± 0.01 9 0.01 ± 0.01 2 Family Tachinidae Gonia divisa Meigen 0.02 ± 0.02 18 0.02 ± 0.02 23 Family Calliphoridae Calliphora vomitoria (L) 0.14 ± 0.09 127 0.11 ± 0.07 102 Lucilia caesar (L) – 0 0.002 ± 0.01 2 Family Sarcophagidae Sarcophaga carnaria (L) 0.01 ± 0.01 13 0.03 ± 0.02 25 Family Muscidae Fannia sp 0.02 ± 0.05 17 0.01 ± 0.02 8

Hymenoptera 0.13 ± 0.10 121 0.27 ± 0.17 254 Family Vespidae Dolichovespula sp 0.004 ± 0.008 4 0.001 ± 0.003 1 Family Eumenidae Pterochelis phaleratus (Panzer) 0.001 ± 0.002 1 – 0 Family Formicidae Linepithema humile (Mayr) – 0 0.24 ± 0.18 235 Camponotus cruentatus (Latreille) 0.07 ± 0.07 74 – 0 Plagiolepis pygmaea (Latreille) 0.01 ± 0.02 15 0.02 ± 0.03 18 Formica cunicularia Latreille 0.01 ± 0.02 10 – 0 Leptothorax nylanderi (Förster) 0.01 ± 0.01 8 – 0 Crematogaster scutellaris (Olivier) 0.01 ± 0.02 9 – 0

Other groups 0.05 ± 0.06 45 0.05 ± 0.11 41 All visitors were counted in all observed inflorescences (16 inflorescences in the invaded area and 16 in the non-invaded area). Other groups include less common visitors (4 Coleoptera, 1 Hemiptera and 1 Lepidoptera) and 6 individuals that were impossible to identify on the videos. species climbed the inflorescences in the non-invaded area: C. cruentatus, Plagiolepis pygmaea, Formica cunicularia, Leptothorax nylanderi and Crematogaster scutellaris. On the other hand, only 2 species climbed the inflorescences in the invaded area: L. humile and P. pygmaea, a tiny ant species that does not seem to be displaced byArgentine ants (Table 2). Concerning the repellent effect ofArgentine ants, the inter- day differences during the flowering period were only mar- ginally significant (two-way ANOVA; F = 1.99, P = 0.059) but differences were observed among groups of plants (non- invaded, invaded but non-infested, and invaded and infested: F = 7.26, P < 0.05). The post-hoc HSD Tukey test revealed a lower number of visitors to the infested plants in the invaded area (1.54 ± 1.85 visitors, mean ± sd) than in the other two non-infested groups of plants: 3.74 ± 4.18 visitors to non- infested plants in the invaded area and 4.1 ± 5.00 visitors to Fig. 1. Number of visitors per plant during 10 min in the three types of plant plants in the non-invaded area (Fig. 1). The number of visi- groups of plants (non-invaded, invaded but non-infested, and invaded and infested; see text) in the flowering period of Euphorbia characias (mean ± se). tors and the number of plant inflorescences were not corre- The different letters on the top of the bars indicate significant differences lated (Pearson correlation test, P > 0.05, n = 112). The den- according to the DHS Tukey post-hoc test (␣ < 0.05). sity of E. characias was similar in the two areas (one-way ANOVA; F = 1.19, P = 0.33). The mean (±sd) density in the invaded area was 0.18 ± 0.13 shrubs/m2 and of seeds was 3932 in the invaded area and 5983 in the non- 0.93 ± 0.03 shrubs/m2 in the non-invaded area. The infested invaded area. Fruit-set was significantly different among the and non-infested inflorescences were interspersed in the different treatments (two-way ANOVA; F = 53.23, P < 0.01) invaded area. and areas (F = 9.27, P < 0.05). The area × treatment interaction was not significant (F = 0.02, P = 0.99). In both areas, 3.2. Fruit-set and seed-set the ant exclosure treatments and the control resulted in larger fruit production than the flying-insect exclosure and all- The number of flowers per inflorescence was similar insect exclosure treatments. These results suggest that the (66.2 ± 40.0, mean ± sd) in the invaded area and non-invaded principal pollinators of E. characias are flying insects. Wind area (74.9 ± 41.8). We collected 2470 fruits in the invaded might also contribute to pollination when fruit-set is consid- area and 2932 in the non-invaded area, and the total number ered (Fig. 2A). X. Blancafort, C. Gómez / Acta Oecologica 28 (2005) 49–55 53

dipterans and ants. Dipterans represented 58% of the flower visitors in the non-invaded area and 40.3% in the invaded area. These results are similar to those observed for other Euphorbia species, in which the principal visitors are flying hymenopterans and dipterans (Ehrenfeld, 1979; Traveset and Sáez, 1997). Ants were a significant group of visitors to E. characias: 29.4% of the visitors in the non-invaded area and 51.2% in the invaded area. These proportions are similar to those observed in an herbaceous Mediterranean community where 58% of the flower visitors to all species of plants studied were ants (Bosch et al., 1997). The different pollinators could have different efficiencies, thereby influencing seed-set and the quality of descendants (Herrera, 2000). The pollination efficiency depends on variation in both pollinator and plant traits (Pellmyr and Thomp- son, 1996). In many cases, when a plant is consumed, dis- persed or pollinated by various organisms, only a few of the animal species involved have a significant impact on the plant’s reproductive success or population dynamics (Gómez and Zamora, 1999; Sanchez-Lafuente et al., 1999). In our case, the number of flowers visited per 10′, and the visitor groups, did not differ between invaded and non-invaded areas. The most common insect visitors spent the same time in the flowers and they visited the same number of flowers in both areas, except for E. tenax, a dipteran species which visited more flowers and spent more time in the non-invaded area. This species has been described as a good pollinator for Lavan- dula latifolia (Herrera, 1987) and for Hormathophylla spinosa (Gómez and Zamora, 1999). Seed production was higher in the plants with ant exclosure treatments than in those excluding flying insects, in both invaded and non-invaded areas. This result suggests that flying insects, in this case Fig. 2. Pollination exclosure experiment results. (A) Fruit-set, i.e., propor- dipterans, could be the most efficient pollinators of E. chara- tion of flowers (cyathia) that set fruit, for each treatment and area (mean ± se). cias. No significant differences between areas were detected. (B) Seed-set, i.e., Ants have traditionally been considered as poor or ineffec- number of viable seeds per fruit, for each treatment and area (mean ± se). tive pollinators (Beattie et al., 1984; Faegri and van der Pijl, The seed-set observed was significantly different between invaded and non- invaded areas, for all the treatments. 1971). Even though ants often exploit floral nectar opportu- nistically (Herrera et al., 1984; Schaffer et al., 1983), several The number of seeds per fruit varied between areas (two- plant species have been described as ant-pollinated (see review way ANOVA; F = 23.26, P < 0.05) and among treatments in Garcia et al., 1995; Gómez et al., 1996; Gómez, 2000; (F = 73.93, P < 0.001), and the area × treatment interaction Peakall et al., 1991). Several species of the ant genus Cam- was also significant (F = 5.61, P < 0.05). For all exclusion ponotus are good pollinators (Gómez, 2000; Gómez et al., treatments we detected significant differences in seed-set between areas. In all cases seed-set was higher in the non- 1996). In our case, the Argentine ant has displaced C. cruen- invaded area. This probably indicates that non-flying insects tatus, a potential pollinator, in the non-invaded area. C. cruen- play some role in pollination, and that flying insects have an tatus made up 25% of the flower visits and, for the flying- important role when considering the seed-set (Fig. 2B). Seed insect exclosure, the seed-set in the non-invaded area was viability ranged from 83% for the treatment excluding flying higher than in the invaded area. In addition, we only detected insects in the invaded zone to 100% for the treatment exclud- two ant species that climbed into the E. characias flowers in ing all insects in the invaded zone, and was not different the invaded area: the Argentine ant and Plagiolepis pyg- between zones in relation to each treatment, or from an over- maea. On the other hand, in the non-invaded area we observed all perspective. at least 5 species to do so: Leptothorax nylanderi, Cremato- gaster scutellaris, Formica cunicularia, Plagiolepis pygmaea and Camponotus cruentatus. A similar pattern has been 4. Discussion detected in California (Human and Gordon, 1997) and South The number of species that visited the inflorescences of Africa (Bond and Slingsby, 1984; Christian, 2001), where E. characias was 39, and the principal flower visitors were communities of native ants have disappeared or their popula- 54 X. Blancafort, C. Gómez / Acta Oecologica 28 (2005) 49–55 tions have dropped dramatically in the invaded areas. More- This study was financed by the Spanish Ministry of Educa- over, seeds of E. characias are transported by ants (Espad- tion and Science (CGL2004-05240-C02-02/BOS). aler and Gómez, 1997) and therefore the presence of this invasive ant also probably has a significant impact on seed dispersal (Gómez and Oliveras, 2003). However, the level of References fruit set observed in the treatments excluding all insects would mean that pollination is possible without insects, effected per- Buys, B., 1990. Relationships between Argentine ants and honeybees in haps by wind, as has been in other entomophilous species South Africa. In: Van der Meer, R.K., Jaffe, K., Cedeno, A. (Eds.), (e.g., Gómez and Zamora, 1996). Applied Myrmecology, a World Perspective. Westview Press, Boulder, Colorado, pp. 519–524. The number of visitors on plants where the Argentine ant Bond, W., Slingsby, P., 1984. Collapse of ant plant mutualism: the Argentine is present was lower than that on plants without L. humile, ant (Iridomyrmex humilis) and myrmecochorous Proteaceae. Ecology both in the invaded and the non-invaded area. In a previous 65, 1031–1037. study in South Africa, Protea nitida was found to be nega- Beattie, A.J., Turnbull, C., Knox, R.B., 1984. Ant inhibition of pollen tively affected by invasion; the Argentine ant foraged in the function: a possible reason why ant pollination is rare. American Journal of Botany 71, 421–426. flowers and had a repellent effect on the pollinator insects. Bolòs, O., Vigo, J., Masalles, R.M., Ninot, J.M., 1993. Flora manual dels There was also a drop in insect diversity and abundance in Països Catalans. Editorial Pòrtic., Barcelona. the infested inflorescences, and as a result, pollination was Bosch, J., Retana, J., Cerdà, X., 1997. Flowering phenology, floral traits and delayed, causing a negative effect on the reproductive capac- pollinator composition in an herbaceous Mediterranean plant commu- ity of the plant (Visser et al., 1996). The lower rate of insect nity. Oecologia 109, 583–591. visitation measured in the invaded area could be due to a lower Christian, C.E., 2001. Consequences of a biological invasion reveal the importance of mutualism for plant communities. Nature 413, 635–638. shrub density of E. characias (the mean density in non- Cole, F.R., Medeiros,A.C., Loope, L.L., Zuehlke, W.W., 1992. Effects of the invaded area was five times more important than in invaded Argentine ant on arthropod fauna of Hawaiian high-elevation shrubland. area), but if we consider only the data from invaded area we Ecology 73, 1313–1322. can see that the key factor was the presence of Argentine ant Carney, S.E., Byerley, M.B., Holway, D.A., 2003. Invasive Argentine ants workers on plants. The infested and non-infested inflores- (Linepithema humile) do not replace native ants as seed dispersers of cences were interspersed in the invaded area. Likewise, when Dendromecon rigida (Papaveraceae) in California. Oecologia 135, 576– 582. we measured the visitation rate of non-infested plants in both Dafni, A., 1992. Pollination Ecology, a Practical Approach. Oxford Univer- areas, any significant differences were detected between these sity Press, Oxford. two groups of plants. The difference in visitation rates was Dafni, A., O’Toole, C., 1994. Pollination syndromes in the Mediterranean: due to the lower rate observed in the infested plant group. On generalization and peculiarities. In: Arianoutsou, M., Groves, R.H. the other hand, these results were consistent with the total (Eds.), Plant–Animal Interactions in Mediterranean-type Ecosystems. number of visits observed in the both areas. The total number Kluwer Academic Publishers, Dordrecht, pp. 125–135. Ehrenfeld, E., 1979. Pollination of three species of Euphorbia subgenus of visits in the invaded-area without Argentine ants was Chamaesyce, with special reference to bees. The American Midland 159 and 395 in the non-invaded area. Our results confirm this Naturalist 101, 87–98. effect above commented of invasive ants; the infestation of Espadaler, X., Gómez, C., 1997. Soil surface searching and transport of inflorescences could lead to a significant reduction in the Euphorbia characias seed by ants. Acta Oecologica 18, 39–46. potential reproductive success of plants. Espadaler, X., Gómez, C., 2003. The Argentine ant, Linepithema humile,in the Iberian peninsula. Sociobiology 42, 187–192. In the invaded area in our study, the level of fruit produc- Faegri, K., van der Pijl, L., 1971. The Principles of Pollination Ecology. tion was uncharged but the number of viable seeds per fruit Pergamon Press, Oxford. was lower. This probably indicates better pollination in the Gómez, J.M., 2000. Effectiveness of ants as pollinators of Lobularia non-invaded area and a negative effect of the Argentine ant, maritima: effects on main sequential fitness components of the host as a result of (a) a repellent effect on visitors, (b) the modifi- plant. Oecologia 122, 90–97. cation of several behavioral aspects of E. tenax, the supposed Gómez, C., Espadaler, X., 2000. Species body-size distribution and spatial scale in Iberian ants. Vie et Milieu 50, 289–295. pollinator of E. characias and/or (c) the displacement of Gómez, C., Oliveras, J., 2003. Can the Argentine ants (Linepithema humile almost all native ants. The results obtained suggest that the Mayr) replace native ants in myrmecochory? Acta Oecologica 24, Argentine ant could interfere negatively with the pollination 47–53. of E. characias, with serious consequences for its reproduc- Gómez, J.M., Zamora, R., 1996. Wind pollination in high-mountain popu- tion success in the invaded areas. lations of Hormathophylla spinosa (Cruciferae). American Journal of Botany 83, 580–585. Gómez, J.M., Zamora, R., 1999. Generalization in the interaction between Hormathophylla spinosa (Cruciferae) and its pollinators. Ecology 80, Acknowledgements 796–805. García, M.B., Antor, R.J., Espadaler, X., 1995. Ant pollination of the pale- We are grateful to X. Espadaler, two anonymous referees oendemic dioecious Borderea pyrenaica (Dioscoreaceae). Plant System- atics and Evolution 198, 17–27. and the guest editors of the journal, especially to M. Hossaert- Giraud, T., Pedersen, J.S., Keller, L., 2002. Evolution of supercolonies: the McKey, for their help with the draft manuscript. We also thank Argentine ants of southern Europe. Proceedings of the National Acad- G. Vila and N. Roura for assistance with the laboratory work. emy of Sciences of the United States of America 99, 6075–6079. X. Blancafort, C. Gómez / Acta Oecologica 28 (2005) 49–55 55

Gómez, J.M., Zamora, R., Hódar, J.A., García, D., 1996. Experimental study Pellmyr, O., Thompson, J.N., 1996. Sources of variation in pollinator con- of pollination by ants in Mediterranean high mountain and arid habitats. tribution within a guild: the effects of plant and pollinator factor. Oeco- Oecologia 105, 236–242. logia 107, 595–604. Gómez, C., Pons, P., Bas, J.M., 2003. Effects of the Argentine ant (Linepi- Peakall, R., Handel, S.N., Beattie, A.J., 1991. The evidence for, and impor- thema humile) on seed dispersal and seedling emergence of Rhamnus tance of, ant pollination. In: Huxley, C.R., Cutler, D. (Eds.), Ant–Plant alaternus L. (Rhamnaceae). Ecography 26, 532–538. Interaction. Oxford University Press, Oxford, pp. 421–429. Herrera, C.M., 1987. Components of pollinator “quality”: comparative Scharpf, R.F., 1970. Seed viability, germination and radicle growth of dwarf analysis of a diverse insect assemblage. Oikos 50, 79–90. mistletoe in California. USDA. Forest Service Research, Berkeley, CA, Herrera, C.M., 1995. Plant–vertebrate seed dispersal systems in the USDA Paper PSW-59. Mediterranean: ecological, evolutionary, and historical determinants. Sanchez-Lafuente, A.M., Rey, P.J., Alcantara, J.M., Valera, F., 1999. Polli- Annual Review of Ecology and Systematics 26, 705–727. nation ecology and the role of floral visitors in seed production of a Herrera, C.M., 1996. Floral traits and plant adaptation to insect pollinators: “few-flowered” perennial herb, Paeonia broteroi Boss and Reut a devil’s advocate approach. In: Lloyd, D.G., Barrett, S.C.H. (Eds.), (Paeoniaceae). Ecoscience 6, 163–172. Schaffer, W.M., Zech, D.W., Buchmann, S.L., Kleinhans, S., Schaffer, V., Floral Biology, Studies on Floral Evolution in Animal-Pollinated Plants. Antrim, J., 1983. Competition for nectar between introduced honey bees Chapman and Hall, New York, pp. 65–87. and native North American bees and ants. Ecology 63, 564–577. Herrera, C.M., 2000. Flower-to-seedling consequences of different pollina- Strasburger, E., Noll, F., Schenck, H., Schimper, A.F.W., 1994. Tratado de tion regimes in an insect-pollinated shrub. Ecology 81, 15–29. botánica. Thirty-third German ed. Editorial Omega, Barcelona. Herrera, C.M., Herrera, J., Espadaler, X., 1984. Nectar thievery by ants from Suarez,A.V., Holway, D.A., Case, T.J., 2001. Patterns of spread in biological southern Spanish insect-pollinated flowers. Insectes Sociaux 31, 142– invasions dominated by long-distance jump dispersal: Insights from 154. Argentine ants. Proceedings of the National Academy of Sciences of the Holway, D.A., 1998. Effect of Argentine ant invasions on ground-dwelling United States of America 98, 1095–1100. arthropods in northern California riparian woodlands. Oecologia 116, Traveset, A., Sáez, E., 1997. Pollination of Euphorbia dendroides by lizards 252–258. and insects: spatio-temporal variation in patterns of flower visitation. Human, K.G., Gordon, D.M., 1997. Effects ofArgentine ants on invertebrate Oecologia 111, 241–248. biodiversity in northern California. Conservation Biology 11, 1242– Visser, D., Wright, M.G., Giliomee, J.H., 1996. The effect of the Argentine 1248. ant, Linepithema humile (Mayr) (Hymenoptera, Formicidae) on flower- Kaspari, M., Weiser, M.D., 1999. The size-grain hypothesis and interspecific visiting insects of Protea nitida Mill (Proteaceae). African Entomology scaling in ants. Functional Ecology 13, 530–538. 4, 285–287. Kearns, C.A., Inouye, D.W., 1993. Techniques for Pollination Biologists. Zamora, R., 2000. Functional equivalence in plant–animal interactions: University Press of Colorado, Colorado. ecological and evolutionary consequences. Oikos 88, 442–447.