Differential impact of two dominant Formica ant species ( Hymenoptera , Formicidae ) on subordinates in temperate Europe

Competition is one of the basic mechanisms shaping ant assemblages. Dominant territorial species are known to restrictively influence the traits of subordinates in various ways. However, there could be diffe­ rences in the effects of dominants due to differences in their colony structure, lifestyle and also behaviour. We tested this hypothesis in natural circumstances in an area where a Formica exsecta Nyl. supercolony neighbours a strong population of the European slave­maker F. sanguinea Latr. For the purpose of our study three different sites were selected: one dominated by F. exsecta, a neighbouring site dominated by F. sanguinea, and a third site where both species co­occurred. We analyzed the structure of subordinate ant communities based on nest counts, and by recording the activity of ants at baits. Based on our find­ ings the structure of ant communities differed significantly among the three sites. The F. exsecta site was characterized by the lowest nest density of subordinates, especially in the case of aggressive species as Lasius platythorax Seifert. At baits even the simple presence of the dominant limited the abundance and occur­ rence of subordinates. In contrast, no such negative effect could be revealed at the F. sanguinea site. The community of the mixed site showed intermediate features in many respects. Generally, the supercolonial JHR 50: 97–116 (2016) doi: 10.3897/JHR.50.8301 http://jhr.pensoft.net Copyright Gema Trigos Peral et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. RESEARCH ARTICLE Gema Trigos Peral et al. / Journal of Hymenoptera Research 50: 97–116 (2016) 98 F. exsecta had a bigger influence on the ant community than the slave­making F. sanguinea. On the other hand, our findings revealed a slight protective role of the territorial F. exsecta for potential slave species to F. sanguinea. The current study underlines the importance of differences between dominant ant species in shaping differentially ant communities even within the same restricted area.


Introduction
Most of the concepts regarding the assembly rules of animal communities point to competition as a crucial factor in shaping their composition and structure (MacArthur 1972, Wiens 1983, Farris et al. 2015, McFarlane Tranquilla et al. 2015, Sunarto et al. 2015).This is supported for ants as well by several case studies (e.g.Savolainen and Vepsäläinen 1988, Pisarski and Vepsäläinen 1989, Cerdá et al. 1998, Parr and Gibb 2010, Czechowski et al. 2013, Maák et al. 2014, Ślipiński et al. 2014, Dejean et al. 2015, Spotti et al. 2015) and the work reviewed quite recently by Cerdá et al. (2013).Specifically, ant communities in temperate zones are organized hierarchically and a crucial role in shaping them can be attributed to ecologically and behaviourally dominant species.Dominant species can restrictively structure local communities in various ways, from interference and exploitative competition to direct predation on lowerranked species, both quantitatively (decreasing nest densities and colony sizes of subordinates) and qualitatively (reducing species richness by competitive exclusion and altering the dominance structure), but also with regards to the spatial patterns of the nests (e.g. by forced resource partitioning, nest moving, etc.) (Higashi and Yamauchi 1979, Cherix and Bourne 1980, Pisarski 1982, Pisarski and Vepsäläinen 1989, Savolainen and Vepsäläinen 1988, 1989, Savolainen et al. 1989, Savolainen 1990, 1991, Vepsäläinen and Savolainen 1990, Hölldobler and Wilson 1990, Punttila et al. 1991, 1996, Sanders and Gordon 2003, Czechowski and Markó 2005, Adler et al. 2007, Czechowski et al. 2013, Markó et al. 2013, Rzeszowski et al. 2013, Ślipiński et al. 2014).
Typically, at least in EuroAsian boreal and temperate zones, the top dominants of ant communities are territorial species (sensu Pisarski 1982), i.e. those which actively protect the boundaries of their whole foraging areas against other territorials (Pisarski 1980, Vepsäläinen and Pisarski 1982, Pisarski and Vepsäläinen 1989).In temperate Europe, such territorial dominants, already wellknown for their competitive impact on subordinate ants, are the wood ant species of the Formica rufa group (subgenus Formica s. str.L.) (Savolainen and Vepsäläinen 1988, 1989, Savolainen et al. 1989, Savolainen 1990, 1991, Punttila et al. 1991, 1996, Czechowski and Markó 2005), species of the subgenus Coptoformica Müll.(genus Formica L.) (Czechowski 1975, Pisarski 1982, Pisarski and Vepsäläinen 1989), the dendrophilic Lasius (Dendrolasius) fuliginosus (Latr.)(Czechowski et al. 2013, Markó et al. 2013, Ślipiński et al. 2014), and Liometopum microcephalum (Panzer) (Petráková and Schlagamerský 2011).Colo nies of territorial species (either allo or conspecific) do not share the same territory, but nonterritorial species can nest and forage within the dominant's territory, how ever with diverse limitations.Their colonies are the centres of the spatial organization in ant communities through their differential effects on subordinate species, which are 'allowed' only to nest at certain distances from the nest of the territorial species -the farther away are the nests of the subordinates, the stronger could be the competition between the two species (Savolainen andVepsäläinen 1988, Pisarski andVepsäläinen 1989).The impact of territorials on subordinates may further be enhanced by specific social strategies, such as the formation of new colonies through budding, which leads to the development of a network of related nests over a larger area (socalled supercolo nies -see e.g.Helanterä et al. 2009, Robinson 2014 for reviews), a wellknown strategy in many territorial wood ants (Higashi and Yamauchi 1979, Cherix and Bourne 1980, Rosengren et al. 1985), but also in members of the subgenus Coptoformica (Pisarski 1982, Czechowski 1975, 1977, Markó et al. 2012).
Within subordinates, two competition levels are distinguished: (a) encounter spe cies, which are fairly aggressive and usually manage to monopolize (defend) single food sources, and (b) submissive species, which defend only their nests (Vepsäläinen and Pisarski 1982).Although they can manage to survive in the territory of dominants, subordinates are restrictively affected by territorials, and, in general, they fare bet ter outside than within territories of dominant species (Savolainen 1990(Savolainen , 1991)).On the other hand, in certain situations, nesting inside a dominant territory can even be good for certain species.Members of the subgenus Serviformica For.(genus Formica) are slave species for the facultative slavemaker Formica (Raptiformica) sanguinea Latr.Potential slave colonies may profit from the proximity of strong dominants since these can stop the raiding columns of F. sanguinea or force them to change their route when protecting their own territory.Such a positive effect of territorials has been reported for several species of wood ants of the F. rufa group (F.rufa L., F. polyctena Först., F. aquilonia Yarr., F. lugubris Zett.) for F. fusca L. and F. lemani Bondr., and known slave species F. sanguinea (Punttila et al. 1996, Czechowski 1999, 2000, Czechowski and Vepsäläinen 2001, Czechowski and Markó 2006, Väänänen et al. 2010).A similar mechanism may also work, at least to some extent, in the case of Serviformica species enslaved by the obligate slavemaker Polyergus rufescens (Latr.)(Czechowski 2006).
The differences among dominant species regarding their effect on subordinates may vary depending on their lifestyle (slavemaker vs nonslavemaker), colony struc ture (mono vs polydomous), and on their foraging strategies (narrow vs wide food spectrum).The differential effect of dominants on subordinates are best revealed when dominant species occur syntopically (in the same habitat), consequently, the basic species pool that could make up the ant community is the same.Therefore, if the dif ferences between the dominants are negligible, the differences between the ant com munities occurring in their territories/areas should also be minor.In order to test this hypothesis, an appropriate natural setup is needed, with the longterm cooccurrence of territorial allospecific neighbours (see e.g.Czechowski et al. 2013).Recent field studies performed on the largest European polydomous system of the territorial Formica exsecta (Markó et al. 2012) yielded the discovery of a large population of the slavemaker F. sanguinea, also a dominant species, neighbouring the abovementioned supercolony.This unique natural setup of two dominant neighbours from the same genus, yet with a different colony structure (supercoloniality vs mono or oligodomy) and lifestyle (non slavemaker vs slavemaker) offered the chance to perform a com parative study.In the framework of the current study, we proposed an analysis of the structure of ant communities in the area of these two dominants, and formulated the following questions: (1) are there any differences between the structure of ant commu nities of the two neighbouring dominant Formica species' territory/area, (2) are there any differential effects of these two dominants on the foraging strategies of subordi nate species, (3) do the abundance and foraging strategy of members of the subgenus Serviformica, known as a slave species of F. sanguinea, differ within and outside the F. sanguinea area?

Study species and sites
The main subjects of the study are two dominant ant species: Formica (Coptoformica) exsecta and Formica (Raptiformica) sanguinea.
Formica exsecta is a fairly common NorthPalaearctic moundbuilding territorial species, a typical forestecotone polytope dwelling in forest margins and clearings, thinned young growth, overgrowing meadows and glades, etc.Its colonies number several thousand to tens of thousands of workers and occur in two, mono and po lygynous social forms.Polygynous colonies may grow through nest splitting into poly domous systems, often with more than 100 nests (socalled supercolonies).Addition ally, colonies may be established through the temporary social parasitism of members of the subgenus Serviformica.The optimal home range of foraging workers from a single nest does not exceed a few meters.They are aggressive and predacious ants, which also feed on honeydew (Pisarski 1982, Seifert 2000, Hughes 2006, Erős et al. 2009, Csata et al. 2012).
Formica sanguinea is a common SouthPalaearctic species, a polytope of warm and dry habitats, both forested and open areas; generally its ecological preferences widely overlap with those of the former species.It nests in decaying tree stumps, partly cov ered with dry plant material and also in the ground, often under stones.In general, colonies, usually functionally monogynous, number up to several tens of thousands of workers.It is a facultative socially parasitic species which enslaves workers of the subgenus Serviformica; the range of its slave raids may reach 100 m (usually < 50 m).They are very aggressive and predatory ants; during their raiding period they do not respect the boundaries of other species' territories (Czechowski 2000, Hughes 2006).The territoriality of F. sanguinea is a moot point (see the Discussion).
As habitat niches of Formica exsecta and F. sanguinea overlap each other to a con siderable degree, colonies of these two species often cooccur.
Our study area was a vast semimoist meadow of the Molinion caeruleae W. Koch alliance [with, among others, Molinia caerulea (L.) Moench, Deschampsia caespitose (L.) P. B., Festuca pratensis Huds., Nardus stricta L. and Juncus sp.], densely overgrown with small birch saplings, located in the southern part of the Giurgeului depression (46°36'N; 25°36'E; 780 m a.s.l.) in the eastern Carpathians, Transylvania, Romania.The meadow is fairly intensely grazed by cows for most of the year.This is the location of the largest known European polydomous system of Formica exsecta.The supercolo ny contains 3,347 permanent nests over an area of nearly 22 ha (i.e. the nest density is ca.153/ha) (Markó et al. 2012).A population of F. sanguinea was recently detected, with a small number of F. cunicularia Latr.and F. rufibarbis F. slaves, within the same habitat near the area occupied by the supercolony.
Three sites were selected within the area for the purpose of our study (Fig. 1): (1) a site clearly dominated by Formica sanguinea right at the border of the F. exsecta poly domous system (referred to as the F. sanguinea site; 32 F. sanguinea nests on ca.3210 m 2 with 0.01 nests/m 2 , and four F. exsecta nests with 0.001 nests/m 2 ), (2) a neighbour ing patch within the F. exsecta territory with a single F. sanguinea nests (referred to as the F. exsecta site; 68 nests on ca.5780 m 2 with 0.012 F. exsecta nests/m 2 , and one F. sanguinea nest with 0.0002 nest/m 2 ) and (3) a site with interspersed nests of both F. sanguinea and F. exsecta (referred to as the mixed site; eight F. sanguinea on ca.2685 m 2 with and 0.003 nests/m 2 , and 80 F. exsecta nests with 0.03 nests/m 2 ).Within the whole study area, the meadow was homogeneous in respect of the vegetation, surface configuration and sun exposure.
The study area was not part of any nature protection area and no specific permits were required to carry out the field studies.

Sampling methods
Nest densities of different subordinate ant species were obtained using the biocoenom etric method, by carefully investigating 15 quadrats of 9 m 2 at each of the three study sites.The quadrats were randomly selected by a blindfolded person in the field, they did not border each other, and they were at least 2 m from any Formica exsecta nest (Fig. 1).A sample of workers was collected for precise identification from each nest found.Identification of ant species was carried out with the key of Czechowski et al. (2012).Field work was carried out between 17 and 21 June 2010.
Baiting is commonly used to study the foraging pattern of ants around their nests and species' foraging strategies.It is also an appropriate method for determining the hierarchical positions of particular species within an assemblage (e.g.Czechowski 1979, 1985, Czechowski and Pisarski 1988, Parr and Gibb 2010, Petráková and Schlaghamerský 2011, Markó and Czechowski 2012).As the food preferences of ants may change seasonally, we provided a mixture of two different kinds of bait: tuna fish flakes as a source of animal protein food and a water solution of mixed flower honey as a source of carbohydrates.Altogether, 10 observation plots were distributed along linear transects at each of the three study sites (Fig. 1), and the baits were put out in the centre of each plot, separated from the next plot by 2 m, and placed at least 2 m from any Formica exsecta or F. sanguinea nest and any previ ously surveyed plot.The bait portions of ca. 3 cm diameter were placed in the centre of a round green plastic plate of 10 cm diameter.Baiting observations were carried out on 8 July 2010.Observations were carried out in two periods: between 10 00 and 12 00 AM, and between 5 20 and 7 20 PM.Each period consisted of six observations of 1 min, separated by 20 minute breaks.During each observation, the number and species of ants appearing at the baits were recorded, with the exception of Myrmica spp., where species level identification could not be performed under field condi tions.The baits were put out 10 minutes before each period of observation (morning and afternoon) and left there until the end of the period when they were retrieved and the plates cleaned.

Data analyses
The differences in the abundance of different ant species among sites were revealed by the use of the KruskalWallis test, followed by Tukey and Kramer posthoc tests.The diversity of ant communities was calculated with the ShannonWiener entropy index (log 2 ), using data for individual quadrats and baits respectively for nest count and bait data, while differences among sites for these parameters were revealed by using the KruskalWallis test, followed by the Tukey and Kramer posthoc tests separately for nest count and bait data.In order to determine the differences among the ant assemblages of the three sites, permutational multivariate ANOVA (perMANOVA, 10000 permutations) was applied both for nest count and bait data, while Nonmetric Multidimensional Scaling (NMDS) was used to visualize the three assemblages, both in the case of nest and bait data.The SIMPER test was performed to determine the contribution of each species to dissimilarities between ant communities.
The Generalized Linear Mixed Model approach (GLMM, binomial, maximum likelihood) was applied to analyze the effect of dominant species on the presence and absence of subordinate species at baits.The abundance of Formica sanguinea and F. exsecta workers present at baits were introduced as variables, along with their pres ence vs absence data as factors, next to the type of the site (F.sanguinea, F. exsecta, mixed) as a factor.The period and time of the observation as well as bait ID were introduced as nested random factors.A similar approach (GLMM, Poisson error, maximum likelihood) was applied to separately test the effect of dominants on the abundance of subordinates.The number of all subordinates was pooled together, since they showed a low abundance separately (see the Results).The same model structure was used as above.
All statistical analyses were carried out using the R 3.1.2statistical package (R Core Team 2015).Multiple pairwise comparisons in the case of the KruskalWallis tests concerning diversity indices were performed using the Tukey and Kramer post hoc test after Nemenyi with the PMCMR Rpackage (Pohlert 2015).Permutational multivariate ANOVA, NMDS and the SIMPER test were carried out with the vegan Rpackage (Oksanen et al. 2013).GLMMs were performed using the glmer function in the lme4 Rpackage (Bates et al. 2014).Posthoc Tukeytests for sequential com parison among factor levels when performing GLMM were carried out with the glht function in the multcomp Rpackage (Hothorn et al. 2015).

Species composition and abundance
During the quadrat survey, 171 nests were found belonging to 11 subordinate ant species in addition to the two dominant species (i.e.Formica sanguinea and F. exsecta) (Table 1).Five subordinate species occurred at every site studied [Tetramorium cf.caespitum (L.), Myrmica scabrinodis Nyl., M. vandeli Bondr., M. schencki Viereck, and Lasius flavus (F.)].Total nest density of cooccurring ant species was 1.8 times higher at the F. sanguinea than at the F. exsecta site, whereas it was intermediate at the mixed site (Table 1).The most abundant subordinate species was the subterranean L. flavus followed by M. scabrinodis on every study site (Table 1).
Only three aggressive encounter species were found: Lasius platythorax, Tetramorium cf.caespitum and Formica rufibarbis, and none of them were present in high densities at the site (territory) of F. exsecta.The three sites differed mainly in the nest density of two encounter species: L. platythorax and T. cf.caespitum (Table 1).Nest density of the former was relatively high at the F. sanguinea site, whereas the latter had a higher density at the mixed site.Nest density of all submissive species [Myrmica spp., Leptothorax acervorum (F.) and Serviformica spp.] taken together was the highest at the F. sanguinea site, owing mainly to the trends noticeable in Myrmica species (Table 1).
The ant community of the Formica sanguinea site was significantly more diverse than the F. exsecta site, while the mixed site had an intermediate position (Kruskal Wallis χ 2 = 7.83, p = 0.02; Fig. 2).The structure of the ant assemblages also differed significantly among the three studied sites (perMANOVA F 1,40 = 3.51, p = 0.006), mostly with regards to the ant community of the F. sanguinea site, while seemingly the communities of the other two sites were quite similar (Fig. 3).According to the results of the SIMPER analysis, most of the dissimilarities (ca.70%) among the three sites could be contributed mostly to three common species: Lasius flavus, Myrmica scabrinodis and L. platythorax, except for the F. exsecta vs mixed site, where T. cf.caespitum took the place of L. platythorax in this respect (Table 2).

Foraging strategies and competitive effects
Besides Formica sanguinea and F. exsecta, foragers of other species were also present at the baits (Table 3).The highest abundance of ants at baits was observed at the F. exsecta site, exclusively due to the efficient recruitment of F. exsecta workers to the food sources (Table 3).As expected, F. sanguinea was present at baits in its site, though in a surprisingly low number, whereas it was totally absent from the F. exsecta site; in turn, F. exsecta was absent in the F. sanguinea site.Both species exploited baits, though, at the mixed site (Table 3).
Different trends were detected in the abundance of two aggressive (encounter) spe cies: Lasius platythorax and Tetramorium cf.caespitum.The former occurred in similar    numbers at baits in the Formica exsecta and F. sanguinea sites, whereas it was absent from the mixed site (Table 3).In turn, Tetramorium cf.caespitum was absent from the F. exsecta site, while it was the second most abundant ant in the two other sites (Table 3).The abundance of potential F. sanguinea slave species, such as F. cunicularia, gradu ally increased from the F. sanguinea site to the mixed one and then to the F. exsecta site (Table 3).The highest diversity was found at baits in the mixed site (KruskalWallis χ 2 = 9.11, p = 0.011) followed by the significantly lower diversities of Formica sanguinea and F. exsecta sites, which did not differ significantly from each other based on the posthoc test result (Fig. 4).There were no differences between the three sites in the qualitativequantitative composition of foraging ants (perMANOVA F 1,29 = 1.83, p = 0.12), however, baits in the F. exsecta territory seemed to stand apart from all other sites due to the marked presence of F. exsecta, while the presence of subordinate species was more characteristic for the F. sanguinea and the mixed sites (Fig. 5).
According to the results of the GLMM analysis, Formica exsecta's abundance (z = -3.09,p = 0.002) and presence (z = -2.32,p = 0.02) had a significant negative effect on the occurrence of subordinates at baits, while the abundance of F. sanguinea did not play a major role (z = -1.72,p = 0.08), but its presence did have a positive influ ence on the occurrence of subordinates (z = 2.13, p = 0.033).There were no differences between the F. sanguinea and the F. exsecta sites with regards to the frequency of the occurrence of subordinates (z = 1.99, p = 0.11).Also, no difference was found between the F. sanguinea and the mixed site in this respect (z = -1.8,p = 0.16), but subordinates were significantly less frequent at baits in the mixed site compared to the F. exsecta site (z = -3.18,p = 0.004).
Generally, a similar pattern with smaller adjustments was found when analyzing the effect of dominants on the abundance of subordinates at baits.In addition to the abundance of Formica exsecta (z = -8.33,p < 0.0001), even its simple presence (z = -2.01,p = 0.044) had a significant negative effect on the abundance of subordinates.On the other hand, F. sanguinea's abundance (z = 0.04, p = 0.96) and presence (z = 0.96, p = 0.33) had no significant effect.The abundance of subordinates also showed significant differences among the three study sites (z ≥ 9.92, p < 0.0001).

Discussion
The main factors which shape biodiversity and structure of animal communities are, besides habitat heterogeneity (see e.g.MacArthur and MacArthur 1961, Hölldobler and Wilson 1990, Tews et al. 2004, Ossola et al. 2015), intra and interspecific rela tionships, such as competition and social parasitism.The importance of the latter is well known in ants (Savolainen and Vepsäläinen 1988, Hölldobler and Wilson 1990, Andersen 1992, Ryti and Case 1992, Cerdá et al. 1997, Czechowski and Markó 2006, Scharf et al. 2011, Ślipiński et al. 2011, 2014).Colonies of dominant species, espe cially territorial ones, play a crucial role, as organizing centres of the ant community, by limiting the establishment of colonies of subordinate species within dominant territo ries or by managing their accessibility to various resources (Savolainen and Vepsäläinen 1988, Pisarski and Vepsäläinen 1989, Andersen 1992, Bestelmeyer 2000, Lester et al. 2010, Cerdá et al. 2012, 2013, Czechowski et al. 2013).In our study, we revealed con siderable differences among ant communities living in areas within the same habitat, but dominated by two dominant species with clearly different features.
The species composition of the studied ant communities was quite similar among all three sites, but differences were found in the abundance of different ant species, and also in the diversity of ant communities.In almost all respects the Formica sanguinea site was superior housing the most diverse ant community.Thus, despite the low chance of interaction with Lasius flavus due to its subterranean lifestyle, the lowest density of L. flavus nests was recorded in the Formica exsecta territory, which could be the result of the establishment of F. exsecta colonies on the top of their soil mounds, with the gradual extermination of L. flavus colonies (Pisarski 1982).This species was followed by Myrmica spp. with a low level of aggression compared to L. platythorax and Tetramorium cf.caespitum, while the lowest abundance was recorded for members of the subgenus Serviformica, being both a slave species to F. sanguinea and the host of temporary social parasitism caused by F. exsecta fundatrices (Czechowski et al. 2012).Generally speaking, there was an increase in the abundance of aggressive and submis sive species in the mixed territory where F. exsecta and the slave maker species showed lower abundance or were absent.
Usually, the exploitation patterns of food resources at the Formica exsecta site was in agreement with the general rules concerning hierarchy patterns, where the dominant species, especially a territorial one, monopolizes the area and its resources (Savolainen andVepsäläinen 1988, Savolainen et al. 1989).In the F. exsecta site the absence of T. cf.caespitum at baits, known as an aggressive species (e.g.Brian et al. 1966Brian et al. , 1979) ) with a quick reaction towards intruders (Dobrzański and Dobrzańska 1975) and with the abil ity to exploit habitat resources effectively (Brian et al. 1966), might have been caused by its general avoidance of the dominant species during foraging (see Cerdá et al. 2012).On the other hand, contrary to Formica exsecta, as other authors have already shown before (e.g.Czechowski 1999, 2000, Czechowski and Markó 2006, Ślipiński et al. 2011), F. sanguinea behaved as a considerably weaker competitor, exploiting the baits much less actively than e.g.Lasius platythorax and Tetramorium cf.caespitum, which succeeded in exploiting the baits within its area quite heavily.The strongest difference in the exploita tion pattern of baits is shown by the results obtained from the mixed territory, where the presence of almost all species at baits increased, with the exception of Formica exsecta, which exhibited the highest frequency and abundance in its own exclusive territory.
The low number of Formica sanguinea foragers at baits, especially in relation to the numbers of workers of other ant species, requires additional explanations.Recent studies on the foraging strategy of F. sanguinea have already shown, that this species seems to be less competitive towards subordinates to the south (e.g.Romania) than in the northern parts of the species range in Europe (e.g.Finland) (P.Ślipiński et al., in prep).One of the major signs of its weaker competitiveness is its reduced presence at artificial baits, as also confirmed by our present observations.However, it is possible that at least some of the F. cunicularia workers present at baits were individuals en slaved by F. sanguinea, and these de facto acted as foragers of the latter.Since, generally, slaves fulfill intranest tasks in colonies (Kharkiv 1979a,b, see also Czechowski 1996), the abundance of slaves foraging at baits could anyhow be quite low.
Territorial ants can hold off raiding columns of Formica sanguinea to some extent, and in this way directly protect possible slave species against their enslaver.This pro tective effect would manifest itself when there was a significantly greater abundance of slave species within, rather than outside of wood ant territories of the F. rufa group (Punttila et al. 1996, Czechowski 2000, Czechowski and Vepsäläinen 2001, Czechows ki and Markó 2006, Väänänen et al. 2010).However, this is not generally valid for all territorials, since, e.g., no measurable protective effect of the territorial Lasius fuliginosus towards the F. fusca slave species was detected by Ślipiński et al (2014), despite cases of direct interference of the latter during F. sanguinea raids (Czechowski 1999(Czechowski , 2000)).In our study area, two potential slave ant species (and, at the same time, potential hosts of the temporary social parasitism of F. exsecta) occurred, both members of the subgenus Serviformica: F. cunicularia and F. rufibarbis.The former is a typical submissive species, the latter is an aggressive encounter species.Due to this, the relations of each of them with both local dominant species, i.e.F. exsecta and F. sanguinea, might proceed differ ently.On one hand, the submissive, weakly competitive F. cunicularia should be less restrictively treated by the territorial F. exsecta than the more competitive F. rufibarbis.On the other hand, colonies of the aggressive F. rufibarbis (see Mori et al. 2001) should be much less prone to being destroyed by F. sanguinea, and they should also be more re sistant to being taken over by young F. exsecta queens than colonies of the submissive F. cunicularia.So, one could expect that these differences in behaviour might be reflected in differences in the distribution of the two subordinate species.However, both species were very scarce in the study area.In total, only three nests of F. cunicularia and one nest of F. rufibarbis were found.At the baits, only F. cunicularia foragers occurred, most abundantly within the F. exsecta site.At the remaining sites (both with F. sanguinea), they were generally scarce, and what is more, at least some of them may have been F. sanguinea slaves.This result may suggest some type of protective effect of the territorial F. exsecta on this slave species against the slavemaker.
While the decisive effect of Formica exsecta as a typically territorial species on the structure of ant communities is quite evident, this question still remains open regard ing F. sanguinea.In the literature, the latter is commonly handled in a similar manner as the territorial Formica s. str.and Coptoformica species (Vepsäläinen and Pisarski 1982, Savolainen et al. 1989, Punttila et al. 1996, Czechowski 2000, Czechowski and Markó 2006, Väänänen et al. 2010).However, our findings seem to contradict the similarly strong territorial (sensu Pisarski 1982) nature of F. sanguinea.The fact that some F. sanguinea nests can occur in the vicinity of F. exsecta nests (see the mixed site) also emphasizes the reduced territoriality of F. exsecta.Other field observations point in this direction as well.For example, in the Białowieża Forest (NE Poland), a very strong F. sanguinea colony was observed to peacefully coexist with a fairly large F. exsecta colony over several years, nesting just ca. 3 m from an edge of the mound of the latter (W.Czechowski, unpubl.).Also, the above mentioned recent case study on the competitive strategy of F. sanguinea (P.Ślipiński et al., in prep.)strongly suggests the nonterritoriality of this aggressive and undoubtedly influential species.
The results of this study underscore the importance of differences between dominant ant species in differentially shaping ant communities, even within the same limited area.

Figure 1 .
Figure 1.Map of the study area (F.exsecta nests outside the delineated sites are not marked out; grey area -compact area of the F. exsecta supercolony).

Figure 2 .
Figure 2. Diversity of the ant communities within the study sites based on nest surveys (medians, quartiles, min-max values and outliers).Survey performed on 9 m 2 quadrats (N = 15 in each site).Different letters indicate significant differences among groups.

Figure 3 .
Figure 3. NMDS graph of the ant communities within the study sites based on nest survey (stress = 0.172).

Figure 4 .
Figure 4. Diversity of the ant communities within the study sites based on bait observations (medians, quartiles, min-max values and outliers).Different letters indicate significant differences among groups.

Figure 5 .
Figure 5. NMDS graph of the ant communities within the study sites based on bait observations (stress = 0.114).

Table 1 .
Nest densities (mean number per 9 m 2 ± SD) of subordinate ant species and their proportions (%) in the three study sites.

Table 2 .
Results of the SIMPER analysis: the overall dissimilarity between study sites, the average contri bution of subordinated ant species to overall dissimilarity (Dissimilarity contribution), their mean abun dances in the compared sites (Mean 1 and Mean 2), and the cumulative percentages of contributions.

Table 3 .
Mean number of ant individuals per baits (± SD) for each study site and their relative propor tions (%) to other ant species at the same site.