Nesting ecology of Polistes gallicus (Hymenoptera: Vespidae) in South-Western Spain

Among the numerous factors that contribute to the suc cess of social wasps, colony productivity is the most important and depends on ecological and environmental factors, such as the nest site. Consequently, nest-site choice should be determined by nest-site characteristics. The objective of this study was to determine on which species of plants Polistes gallicus wasps build their nests and identify the possible preferred features of plants used during nest-site selection by the foundress. From February to October 2017–2019 we recorded details of the nests in natural areas in south-western Spain (at 38° latitude). Nests of P. gallicus were recorded on 49 species of plants belonging to 23 families, with Scirpoides holoschoenus, Epilobium hirsutum and Foeniculum vulgare the most common. Signifi cantly more wasps built their nests on herbaceous than on shrub and arboreal plants, and on “green” than “brown-grey” plants. Most of the nests were 21 to 189 cm above the ground, with a nest azimuth of about 102°. Nests were located at a mean distance of 43.51 ± 123.79 m from the nearest source of water. The length of the nesting season in the area studied was between 30 and 33 weeks, and more colonies were founded by a single foundress than several foundresses. Furthermore, a greater percentage of the single foundress colonies failed than of the multiple-foundress colonies.


INTRODUCTION
Among the numerous factors that contribute to the success of social wasps, colony productivity is the most important. Pro ductivity depends on ecological factors, including temperature, prey availability and number of founders (Gamboa et al., 2005). Colony success and productivity may also be related to environmental factors, for example, nesting site, as reported by Inagawa et al. (2001) and Nadeau & Stamp (2003) in Polistes snelleni (Saussure) and Polistes fuscatus (Fabricius), respectively. Nest-site selection is closely associated with fi tness because of its effect on offspring production. Consequently, choice of a nest-site should be determined by particular nest-site characteristics. Thus, habitat characteristics that infl uence the probability of nest predation may be particularly important because nest predation is often the primary cause of nest mortality in polistine wasps (Cervo & Turillazzi, 1985).
In bird assemblages, nests may be affected by habitat at two spatial scales (Martin & Roper, 1988): the nest site (characteristics of the immediate vicinity of the nest) and the nest patch (characteristics of the habitat surrounding the nest). Previous studies on polistine wasps have focused on the nest site, examining mainly nest site selection and 2010). All these authors mention that the nests of P. dominula are in hedges of prickly pear cactus (Opuntia sp.). There is no information on the species of plants utilised by P. gallicus. Thus, the aim of this study was to determine on which species of plants P. gallicus build their nests and identify the features of species of plants foundresses respond to when selecting a nest site. We also considered the height at which the nest was located and its orientation in relation to north (azimuth). In addition, we investigated the foundation of colonies of P. gallicus in natural environments in South-Western Spain.

Study site
Fieldwork was carried out in the Llanos de Olivenza area near the city of Olivenza (South-Western Spain, 38°41´0.7˝N / 7°06´0.1˝W, 267 m a.s.l.) from February to October 2017, 2018 and 2019. The climate is typically Continental-Mediterranean, with relatively cold wet winters and dry hot summers. During the period studied, mean annual temperatures were above the historical average (Fig. 1), while rainfall was below the historical average (mean temperature: 16.3°C, rainfall: 432 mm yr-1, Capel Molina, 1981). The fl at-to-gently undulating landscape is dominated by a mosaic of dry winter cereal crops, olive groves and vineyards. In this area, natural habitats only occur in fi eld margins, some patches of riparian vegetation and holm oak (Quercus ilex ssp. ballota) dehesas (an agro-silvo pastoral system).

Location of nests
We started the fi eld research in early February at twelve different sites near Olivenza (within a 15 km radius). We searched for nests in favourable habitats (Fig. 2) and marked each nest with an adhesive plastic band (the marks were placed on the stem nearest to the nest). The location was also recorded using GPS (Garmin). Next, we recorded the species of plant and measured the height at which the nest was located and its orientation in relation to north (azimuth). We measured nest height as the distance between the ground and the petiole of the nest. Using a digital compass we recorded the angle between a line oriented towards the magnetic north and the axis of the petiole of each wasp's nest. In the laboratory, we measured the distance from the nest to the nearest source of water (small streams in most cases) using SIGPAC (the Span-species richness. The relationship between plant species diversity and vegetation structure on invertebrate diversity has been studied, for example, in butterfl ies (Collinge et al., 2003), one of the preferred prey species (caterpillars) of wasps. These authors report that habitat quality positively affects butterfl y species richness and composition. Water can also be a limiting resource for wasps (Raveret, 2000). In addition to imbibing water or passing it to larvae (Raveret, 2000), social wasps mix water with masticated plant fi bre when processing material for nest construction and use it in conjunction with wing fanning in the evaporative cooling of the nest (Steiner, 1930;Heinrich, 1993;Höcherl et al., 2016). Kasuya (1982) demonstrate that in Polistes chinensis antennalis, the rate of water intake decreased as the distance to the water source increased and Horwood et al. (1993) point out that water is an important factor limiting the spread of Vespula germanica in Australia.
In natural environments, polistine wasps build their nests almost exclusively on plants. Cervo & Turillazi (1985) state that the nests of P. nimpha are built predominantly on shrubs of Inula viscosa (Asteraceae). Rusina (2006Rusina ( , 2009 Kozyra et al. (2016) report that foundresses of P. nimpha prefer to nest in particular species of plants, at a particular height above the ground with the combs at a particular geographic orientation (azimuth). The fi rst two factors are important in terms of maximizing the chances of colony survival in the pre-emergence phase, the most dangerous initial part of a colony's life cycle, while the azimuth is most likely important for nest thermoregulation when the surrounding vegetation is still low and the nest is directly exposed to sunlight. These authors report that nests are most often initiated on Hypericum, Tanacetum, Daucus and Achillea plants.
In the Old World, Polistes gallicus (L., 1767) is, without doubt, one of the most abundant species of Polistes. It occurs in the Western and Central Mediterranean area, eastwards to the Greek island of Corfu; in North-Western Africa, from Tunisia to Morocco; and northwards to the Italian Alps and Southern Switzerland (Schmid-Egger et al., 2017). It is particularly abundant in countries around the Mediterranean basin and in North African countries. According to Schmid-Egger et al. (2017), seven species of Polistes are present on the Iberian Peninsula, with P. gallicus and P. dominula the most abundant species in the South, where both species are common in natural and urban environments (Gayubo & Torres, 1990).
There is only information on the nesting sites of paper wasps on the Iberian Peninsula for P. dominula and records of the plants it nests on are scattered in various publications and mostly in the description of the area studied (Cant et al., 2006;Zanette & Field, 2008, 2011Leadbeater et al.,   Plants were identifi ed in situ or in the laboratory according to Flora iberica (Castroviejo, 1986(Castroviejo, -2012. For each plant, the following data were recorded: type (arboreal, shrubby, herbaceous), life form according to Raunkiaer (1934), stem colour (green, grey, brown), presence of defences (thorns, spines, hairs) and subjectively estimated the density of stems or leaves (low, medium, high). Wasps were identifi ed using the identifi cation key of Schmid-Egger et al. (2017).

Nest productivity
The number of foundresses was recorded when each nest was located and verifi ed during the following weeks until the fi rst workers emerged. At the end of the founding period, we recorded the number of intact nests and compared the percentage survival of those that were established by different numbers of foundresses. A nest was classifi ed as failed if all the foundresses disappeared during the founding period and no adults emerged from the nest (Tibbetts & Reeve, 2003).

Statistical analysis
The preferences for nesting on different plants and their characteristics were tested using X 2 tests. The differences in nest height and distance to a source of water between years were tested using the Mann-Whitney U test. We used Program ORIANA (Kovach, 2011) to calculate circular statistics for nest orientation: mean vector and standard error. Rayleigh's test of uniformity was used to test mean exposures for non-random orientations (Batschelet, 1981). The concentration (r) or length of the mean vector calcu- lated using circular methods, ranged from 0 to 1, and is affected by variation in circular data, sample size and grouping. Values of r near 1 indicate data points closely concentrated near the mean angle or, in our case, nest-sites with a particular direction of orientation. Differences in mean direction between years were analysed using the parametric Watson-Williams F-test in cases where the assumptions underlying this test were fulfi lled. In total, we found nests of paper wasps on 49 species of plants belonging to 23 families (Table 1)  In 2017, the orientation of the paper wasp nests was to the south-east (Fig. 4), with a mean vector of 97.092° ± 27.54° (circular standard deviation) and a length of mean vector of In 2017, paper wasp nests were located 73.97 ± 140.14 m (range: 0.5-552 m) from a water source (Fig. 5) At this study site, the number of foundresses per nest ranged between one to three. Exceptionally, we found a nest with seven foundresses. The overall ratio of multiple to single-foundress colonies was 47 : 314 for the three years studied. That is more single (104, 95 and 115 in 2017, 2018 and 2019, respectively) than multiple-foundress colonies (18, 11 and 18 in 2017, 2018 and 2019, respectively) were found.

We
Colony failure was higher in single than in multiplefoundress colonies: 35.03% of single-foundress colonies failed, whereas only 18.74% of the multiple-foundress colonies failed. The percentage survival of multiple-foundress nests was 83.33%, 54.54% and 94.44% in 2017, 2018 and 2019, respectively, whereas for single-foundress nests it was 67.30%, 50.52% and 74.78% in 2017, 2018 and 2019, respectively. None of the nests that had survived to the following year were reoccupied (n = 15). (Mainwaring et al., 2014), prior to constructing a nest, the foundress of social wasps must decide on the location in which to construct it. The selection of a suitable nest site is determined by a combination of fi ve main factors: the availability of food, risk of predation, presence and behaviour of conspecifi cs, availability of suitable nest material and a suitable ambient climate for raising offspring. Many studies have shown that wasps build nests in response to the environments they inhabit (Matsuura, 1990;Yamane, 1996;Hozymi & Kudô, 2012) since the development of the immature stages depends on the ambient temperature. In Polistes, there is a clear relation between the nesting season and temperature (or latitude). In Mediterranean species, such as P. dominula, the length of the nesting seasons is between 6.5 and 7 months (Yamane, 1996). However, Höcherl & Tautz (2015) report a mean nesting season of 4.6 months for this species in Germany (latitude 49°). The nesting season is even longer than that reported for this species in the United States (see Höcherl & Tautz, 2015). Thus, the duration of the nesting season at the site studied (7.9 months, latitude 38°) cannot be considered as unusual. Zanette & Field (2008) state that the nests of P. dominula appear in late winter (late February-March)   in the south of Spain. Obviously, the nesting season can differ at the same place from year to year depending on temperature (Höcherl & Tautz, 2015). These authors report nesting seasons ranging from 18 to 21 weeks over a period of three years for P. dominula. In P. gallicus, the nesting season ranged from 30 to 33 weeks from 2017 to 2019.

As in birds
Although we found colonies of P. gallicus in urban areas in the southwest of Spain, this study focused on natural populations. This study revealed that P. gallicus uses a wide variety of plants for nesting (48) in the south-west of Spain. There are no other similar studies for this species in this area or in its natural range. However, Kozyra et al. (2016) report P. nimpha constructs nests on 30 species of plants in Poland, which is similar to that reported for the same species in the Ukraine by Rusina & Orlova (2011). In Italy, Cervo & Turillazzi (1985 report that most of the nests (76%) of P. gallicus were on Ditrichia (= Inula) viscosa (Asteraceae).
Why a wasp selects a given plant is unclear, and several factors may affect the choice: (a) The availability of plants when foundresses emerge in late winter. Kozyra et al. (2016) state that at the end of winter the only available nesting sites for wasps are young trees, shrubs, lignifi ed parts of herbaceous plants, or grass culms remaining after the winter period. However, in early winter in the south of Spain (with warmer winters than in Poland), some shrubs and herbaceous plants can be utilised by wasps to build their nests, as is the case of Diplotaxis plants or some Asteraceae plants. (b) Protection. At the site studied, approximately one third (33.33%) of the plants selected by wasps had some form of passive defence (spines, thorns). The most selected species of plant in this study was Scirpoides holoschoenus, a perennial plant with pointed stems. In addition, we did not fi nd nests on isolated plants. This means that plants surrounding nests may provide protection against high temperatures by shading and thus prevent overheating in the warmest weeks of the colony cycle. (c) Occultation /camoufl age. As in P. dominula and P. nimpha, the colour of the nests of P. gallicus is beige and grey, with dark grey linings (Bagriaҫik, 2012). This pattern may serve to camoufl age the nest on most of the plants selected by wasps, especially in late spring and summer, when most plants dry out because of high temperatures. (d) Nest consistency. The nest of Polistes is made of chewed plant fi bres from weathered wood and other sources (Jeanne, 1975). It is also known that species of Polistes prefer long vegetative fi bres and plant hairs for nest material (Bagriaҫik, 2012). This author reports that the nests of P. gallicus are less consistent, smaller and lighter than those of P. dominula and P. nimpha. Cole at al. (2001) report differences in the nests of the three vespine wasps (Vespa vulgaris, Dolichovespula norwegica, and D. sylvestris), namely in the types of fi bres used to build the comb and envelope of their nests. These authors attribute these differences to the use of different sources of pulp to build their nests. (e) Protection against parasitoids. Parasitoids are a major cause of mortality in many insects (Godfray, 1994), including wasps (Rusina, 2008).
In this study, nest height (21-181 cm) was higher than that reported for P. nimpha in Poland (15-25 cm, Kozyra et al., 2016) andItaly (4-60 cm, Cervo &Turillazzi, 1985). Plant growth is infl uenced by a variety of external and internal factors. The former include light, temperature, water, nutrients, etc., while the latter include genetic and hormonal factors. Hence, it is not surprising that there are differences between Polish and Spanish plants. We even found signifi cant differences in plant height from year to year, probably caused by differences in rainfall. Nest height does not depend on plant species and is not accidental. However, we agree with Kozyra et al. (2016) that nest height plays an important role in protecting the nest from predators. We suppose, as occurs in birds (Mainwaring et al., 2014), they build their nests at varying heights in response to predators, higher from the ground in response to mammalian and ant predators and lower in response to avian predators.
In many social insects, such as ants and termites, the geographic orientation of the nests is not accidental and is of great signifi cance for their thermoregulation (Jones & Oldroyd, 2007). Nest orientation often infl uences the amount of solar radiation absorbed by a nest and the time of day that the highest level of radiation is received. Many species orientate their nests so that it is warmed by solar radiation in the cool of the morning (Jones & Oldroyd, 2007). The nest orientation of P. gallicus in our study (102.11°) is very similar to that reported for P. nimpha in Poland (110°, Kozyra et al., 2016) and Italy (162°, Cervo & Turillazzi, 1985) and agrees with that reported by Yamane (1996) for P. snellini and P. biglumis (on southern and south-eastern slopes).
Water is a limiting factor for the development of wasps (Horwood et al., 1993) and is required for temperature control (Greene, 1991), nest construction (Jeanne, 1996) and metabolism (Richter, 2000). To overcome the desiccation associated with their habitats, insects have adopted several physiological and behavioural practices (Kovac et al., 2009), such as living in more favourable microhabitats (Benoit et al., 2007), foraging in more protected areas (Hu et al., 2012) and making frequent trips to sources of moisture (Willmer, 1986). Building a nest near a water source may be advantageous for colony development if we consider that the collection of liquids may make up 80% of the material collected by some vespids. In this study, the mean distance of nests from sources of water (small temporary rivers in most cases) was 43.51 m. Assuming a fl ight range of 250 m for species of Polistes (see Prezoto & Gobbi, 2005), only 9.61% of the nests analysed were located at distances greater than 250 m. This indicates that in the south-west of Spain, P. gallicus nest close to sources of water, confi rming the importance of this resource for these wasps. Another advantage of building a nest near a source of water is related to thermoregulation. Polistes wasps build exposed combs, without any cover and protection. Because of this, the thermoregulation of their nests depends mainly on environmental conditions (Höcherl et al., 2016). The passive mechanisms of thermoregulation include all those that help to optimise the internal temperatures, such as site selection. Nest-site selection plays a major role in the ability of social insect colonies to maintain stable nest temperatures. Nest-site choice falls into two broad categories. For many species, the main criterion when selecting a nest site is physical protection against environmental perturbations; others select sites where the microclimate provides a relatively stable temperature (Jones & Oldroyd, 2007). In the south of Spain, where the summers are hot, the patches of vegetation surrounding small rivers or streams may function as temperature islands. Thus, by nesting in these patches of vegetation the wasps may reduce the heating load on their nests and avoid extreme temperatures in July and August.
As a primitive species of eusocial wasp, the cooperation of queens in nest building and brood care in Polistes is common (Reeve, 1991). We recorded a maximum of seven foundresses per nest in this study, which is similar to that recorded in other studies. For instance, Zanette & Field (2009) records P. dominula wasp nests built by 1-10 foundresses in Southern Spain. In this study, we recorded more single than multiple foundation colonies for P. gallicus. For P. dominula, Turillazzi et al. (1982) report 36% single and 64% multiple-foundress colonies, respectively and Zanette & Field (2011) 7% single and 93% multiple-founded nests, respectively. However, in Germany (Höcherl & Tautz, 2015), the percentage of the colonies established by hibernating queens of P. dominula, multiple and single founded are similar (46% and 54%, respectively). One of the advantages of multiple founding is that the percentage survival of these nests is greater than that of single founded nests. Our results confi rm this and agree with the reports of Höcherl & Tautz (2015). Another benefi t of multiple-foundress nests is increased productivity, measured in terms of the number of cells per nest (Queller et al., 2000;Tibbetts & Reeve, 2003). However, Höcherl & Tautz (2015) report that the mean number of cells per nest is not signifi cantly different between single and multiple-founded colonies in Germany.