EUROPEAN JOURNAL OF ENTOMOLOGY EUROPEAN JOURNAL OF ENTOMOLOGY Overwintering of ladybirds (Coleoptera: Coccinellidae) on Scots pine in Central Europe

. We surveyed ladybirds (Coleoptera: Coccinellidae) in 10 stands of Scots pine ( Pinus sylvestris ), all monoculture stands 5–100 years old, in western Slovakia, Central Europe, over two successive periods, October 2013 – March 2014 and October 2014 – March 2015. The winter in each period was exceptionally mild. Ladybirds were collected from the lower branches of pine trees using beating trays and were present in 61% of the 1040 samples (one sample containing ladybirds from 20 branches, 1 m long each). In total 3965 individuals of 20 species were recorded. Non-conifer dwelling species associated with broadleaved trees or herbaceous plants prevailed (45% of species), followed by conifer specialists (40%) and generalists (15%). Although 13 species were found at least in one winter month, December, January or February, only four of them, Exochomus quadripustulatus , Coccinella septempunctata , Harmonia axyridis and Hippodamia variegata , were recorded continually during both winters. The number of species, the abundance of all ladybirds and the abundance of dominant species ( E. quadripustulatus , C. septempunctata and H. axyridis ) decreased from late autumn towards winter and remained lowest during this most adverse time of the year for ladybirds. Overwintering species assemblages of ladybirds changed over time and varied with age of pine stand. Our results suggest that Scots pine in Central Europe supports species rich assemblages of ladybirds from late autumn to early spring and, being widely distributed, it could be suited to winter surveying of ladybirds at large spatial scales to reveal behavioural and ecological responses of species to changing weather or different climates.


INTRODUCTION
Overwintering is an important phase of an insect's life in temperate areas. For ladybirds (Coleoptera: Coccinellidae) it is passed in the adult stage. As for other periods of the life-cycle of ladybirds, the ecology of overwintering has received much attention (Majerus, 1994;Hodek & Honěk, 1996;Hodek, 2012a, b;Hodek et al., 2012). To overcome

Study area
Our study was conducted in the western part of Borská nížina lowland in western Slovakia in the localities of Lakšárska Nová Ves, Studienka, Lozorno, Pernek and Moravský Svätý Ján (48.333-48.566N, 16.983-17.166E). The area (approximately 30 × 20 km) is covered by approximately 44,000 ha of monoculture of Scots pine on sand dunes (Šomšák & Kubíček, 1994) that replaced the former oak-pine forests (Pino-Quercetum; Kollár et al., 2011). It is characterized by a sub-continental climate, with mean annual temperatures varying from 9.0 to 9.6°C and mean temperatures in January between -1.0 and -2.0°C (Šťastný et al., 2015). The two successive winters (December-February) of the study were both exceptionally mild, with mean temperature in January of +2.1°C in 2014 and +2.4°C in 2015. There were 53 subzero days in the winter of 2013/2014 and 46 subzero days in the winter of 2014/2015. Each winter had only a single day with more severe air temperatures of between -10 and -15°C (Fig. 1). In both winters snow cover was reduced to a few days only.

Sampling of ladybirds
Ladybirds were collected in margins of pine stands (all monoculture stands 5 to 100 years old) with a south-eastern aspect at altitudes between 159 and 222 m (negligible altitudinal range relative to the size of the area, absence of prominent features of landscape such as hills). The stands bordered forest as well as non-forest habitats -a mosaic of meadows and arable fi elds. To get insights into changes in abundance and composition of species assemblages over time, ladybirds were repeatedly collected in these stands between October 2013 and March 2014 (the fi rst period) and between October 2014 and March 2015 (the second period), in the middle of each month except for November 2013 when, due to logistical diffi culties, ladybirds were not obtained. Seven stands in October and 10 stands in other months were sampled in each period on days without rain, snow or strong wind, between 10:00 and 14:00 h using beating-trays. One sample con-dense shrubs, bark crevices, peeled bark as well as buildings and other artifi cial constructions (Hodek, 1960(Hodek, , 1973Majerus, 1994;Hodek et al., 2012). Certain species may utilize different types of overwintering sites but often prefer one or a few. Earlier studies mostly focused on overwintering sites of frequent and abundant species such as Coccinella septempunctata Linnaeus (Hemptinne, 1988;Honěk, 1989;Ceryngier, 2000;Honěk et al., 2007), Adalia bipunctata (Linnaeus) (Hemptinne, 1985), Ceratomegilla undecimnotata (Schneider) (Hodek et al., 2012), Myrrha octodecimguttata (Linnaeus) (Puliainen, 1966) and Tytthaspis sedecimpunctata (Linnaeus) (Nedvěd, 2006). The preferred overwintering sites of ladybirds within Europe are provided in Hodek (1973), Majerus (1994, Roy et al. (2011), Hodek et al. (2012) and Nedvěd (2015. The fast spreading non-native harlequin ladybird Harmonia axyridis (Pallas), infl uencing biodiversity over large areas (Koch, 2003;Roy et al., 2012), dominates many ladybird communities within Europe (Brown et al., 2008Viglášová et al., 2017) where it has been intensively studied from various aspects (Raak-van den Berg et al., 2012;Roy & Brown, 2015;Honěk et al., 2016;Roy et al., 2016;Sloggett, 2017). This eurytopic invader has been reported from many habitats but with most information from the active periods of its life-cycle, specifi cally spring, summer and autumn Roy & Brown, 2015;Brown & Roy, 2018). As most of the winter records of H. axyridis come from urban areas and/or anthropogenic habitats (Panigaj et al., 2014;Viglášová et al., 2017) there is a need to understand the overwintering ecology of this species in rural habitats.
Herein, we surveyed ladybirds on Scots pine (Pinus sylvestris Linnaeus, Pinaceae), a conifer having a very wide distribution extending the breadth of Eurasia (Rehfeldt et al., 2002). Emphasis was placed on the lower branches of pine trees as an overwintering habitat of ladybirds in the fi fth and sixth year after the establishment of H. axyridis in Slovakia, Central Europe Panigaj et al., 2014). Earlier studies from Western, Central and North-Eastern Europe revealed that Scots pine hosts ladybird species specialised on conifers as well as non-specialist species (Höregott, 1960;Bielawski, 1961;Klausnitzer, 1965Klausnitzer, , 1967Hodek, 1973;Bastian, 1982Bastian, , 1984Majerus, 1994;Czechowska, 1995a, b;Roy et al., 2011;Nedvěd, 2015). Despite a reasonable understanding of the associations of ladybirds and Scots pine within Europe during spring and summer, there is little information from systematic winter surveys. Indeed not much is known about overwintering of native ladybird species or H. axyridis in coniferous forests (Pendleton & Pendleton, 1997-2016Hodek, 2012a).
We assessed variation in species richness and abundance of ladybirds in a Scots pine forest in Central Europe from late autumn to early spring. Focusing on ladybirds on the lower branches of pine trees we predicted: (1) species specifi c assemblages during winter, (2) decrease in abundance of ladybirds from late autumn to winter, and (3) temporal and/or spatial variation in the composition of overwintering species assemblages of ladybirds. stituted ladybirds from 20 lower branches (up to 3 m above the ground) of several close-growing trees beaten over a circular canvas beating tray of 1.0 m in diameter. Young pines (approximately 5 years old) were sampled as whole trees. In each stand samples were always obtained from different marked trees to avoid sampling of branches beaten on previous sampling dates. Visual observations (not quantifi ed) were made to determine the distribution of predominant ladybird species on branches. Ladybirds collected were placed in 70% ethanol and identifi ed and counted later in the laboratory. Species were classifi ed according to their trophic and habitat preferences following Bielawski (1961), Majerus (1994) and Nedvěd (2015).

Data processing
(1) The relative occurrence of ladybirds in pine forest was assessed for each species, month and time period as the ratio of the number of samples containing ladybirds and the total number of samples (October: n = 70, other months: n = 100). The ratio enabled an assessment of the probability of ladybird species recording and the effort needed to obtain particular species over time.
(2) The number of species (species richness), the number of individuals of all species and the number of the three predominant species, Exochomus quadripustulatus (Linnaeus), C. septempunctata and H. axyridis (response variables), were modelled as a function of month (categorical explanatory variable with particular months as factor levels). A generalized linear mixed effects model (GLMM, Bolker et al., 2009) with the Poisson distribution and the log link function, was fi tted on species richness data (the number of different species on 200 branches). As the variances of ladybird numbers far exceeded the mean values (overdispersion), a GLMM with the negative-binomial distribution and the log link function was fi tted on abundance data (the number of individuals on 200 branches). The samples collected monthly in each stand were merged to avoid spatial pseudoreplication, and the temporal dependence in data was considered in the mixed model. In the GLMMs, the factor month was specifi ed as a fi xed effect, the factor stand (the 10 repeatedly sampled stands) was considered as a random effect. The post-hoc multiple comparisons of the mean number of species and individuals at the 0.05 signifi cance level were performed using the Tukey's contrasts test, the P-values being adjusted by the Holm's method. Mean values along with 95% confi dence intervals were plotted. Diversity of species assemblages in Scots pine forest was characterized by the Simpson's index of diversity calculated for each collecting time. Separate analyses were made for each time period.
(3) To assess the general variability in species data, unconstrained detrended correspondence analysis (DCA) was employed. The species data (abundances) were log(x+1) transformed prior to the analysis. The longest DCA axis was the second one with the length of 3.59 units. The rare species were down-weighted as some of them had a strong effect on the ordination. Species data were then fi tted in the constrained correspondence analysis (CCA). Species assemblages in October and March (in both periods) differed from the species assemblages in December, January and February (10,000 permutations of model residuals; P < 0.01). Hence, only the samples from the three winter months were analysed to account for variation in species data. In the CCA overwintering species assemblages were modelled as a function of the month (categorical variable: three winter months), age of forest stand (numerical variable: ages of ten different stands) and type of habitat adjacent to an overwintering habitat (categorical variable: forest/no forest as factor levels).
Statistical analyses were performed in R (R Development Core Team, 2016). The Simpson's index of diversity was calculated using PAST (Hammer et al., 2001).

RESULTS
Approximately 60% of the samples from the fi rst (n = 470) and 62% from the second period (n = 570) contained at least one ladybird, hence, ladybirds were present in 61% of the 1040 samples obtained in both periods. Three species, E. quadripustulatus (Linnaeus), C. septempunctata and H. axyridis, occurred most frequently in samples. They were observed to spend winter directly on the exposed bark Table 1. Relative occurrence (%) and trophic groups of ladybird species on Scots pine in Central Europe. Continually recorded species during both winters are marked with an asterisk (*). APH -aphidophage, COC -coccidophage, PAL -pallinophage, PSO -psocopterophage, PSY -psyllophage, PHYL -phyllophage. The relative occurrence of most of the dominant species indicated lower probability of detection during winter than during late autumn. Judged from trophic preferences in the literature, a total of 16 species (80%) mainly feed on aphids, three species (15%) prefer coccids and one species (5%) is a herbivore (Table 1). In total, 3965 individuals of 20 species belonging to the tribes Chilocorini, Coccidulini, Coccinellini, Epilachnini and Scymnini were recorded. A total of 1440 individuals and 17 species were obtained in the fi rst period, while 2525 individuals and 18 species were recorded in the second pe-riod. Exochomus quadripustulatus and C. septempunctata dominated species assemblages in both periods, and H. axyridis and Hippodamia variegata (Goeze) were characteristic predominant species in only the second period ( Table 2 ).
The composition of overwintering species assemblages of ladybirds showed a signifi cant temporal variation (pseudo-F = 3.5, P < 0.001) and also varied with the age of pine stand (pseudo-F = 4.1, P = 0. 003). The temporal factor (month) explained 25% and the spatial factor (stand) explained 5.8% of the variability in species data (total inertia = 1.3726). The number of overwintering ladybirds (December -February) varied greatly between the stands of different ages (range/1st period: 6-171, n = 10; range/2nd period: 6-299 individuals, n = 10). It was consistently (over both periods) lowest in the youngest of the stands (fi ve years old). For overwintering, ladybirds mostly utilized the stands that were at least 15 years old. The type of habitat adjacent to an overwintering habitat of ladybirds (forest, no forest) did not infl uence composition of overwintering species assemblages (pseudo-F = 0.9, P = 0.651).

DISCUSSION
Our results provide the fi rst insights into temporal (phenological) changes in species assemblages of ladybirds on Scots pine in Central Europe from late autumn to early spring. They reveal the importance of the lower branches of pines as a habitat for at least 20 ladybird species (of them 13 overwintering, 4 continually), that is, more species than Bastian (1982) recorded on young Scots pines during the growing season (17 species) or Czechowska (1995a) reported from the canopies of Scots pine stands of different ages during the growing season and in early autumn (14 species).
Of the conifer specialists, only E. quadripustulatus was frequently recorded during the whole overwintering period. This species overwinters on evergreen trees and shrubs, particularly in pine foliage (Roy et al., 2011), bark crevices and litter (Bielawski, 1961;Hodek, 2012a;Nedvěd, 2015), or on pine branches, as shown in this study. The reproductive activity of E. quadripustulatus starts very early, an increasing number of individuals on trees can be observed in February (Nedvěd, 2015), in agreement with our results.
Adult C. septempunctata often rest at the ends of twigs of young pines before they drop and shelter on the ground (Hodek, 1960;Bielawski, 1961;Klausnitzer, 1967). We (O. Nedvěd, unpubl.) observed this behaviour between 1995 and 2010 in a forest margin in South Bohemia where adult C. septempunctata were often present on pines until November and moved to the litter in December. Herein, for the fi rst time, we demonstrated the continual overwintering of adult C. septempunctata on the lower branches of Scots pine until spring, proving that at least some of them avoid dropping to low herbage, litter, small stones or dead standing herbs -the preferred overwintering sites of this species (Honěk, 1989;Honěk et al., 2007). Compared to Central Europe, in Britain the known range of overwintering sites of C. septempunctata is broader, and overwintering individuals are known to utilize buildings, crevices in tree trunks and the foliage of gorse (Ulex spp.) and conifers, in some cases as high as 15 m above ground level (Majerus, 1994). As C. septempunctata was intensively studied in Central Europe (Hodek, 1973;Honěk & Martinková, 2005; Hodek, 2012a) it is unlikely that its overwintering on conifers prior to this study was overlooked. The contrast in overwintering site choices of C. septempunctata between Britain and Central Europe suggests that the mild British maritime climate favours the overwintering of this species on trees compared to the colder climate of continental Europe (Majerus, 1994). The winters of 2013/2014 and 2014/2015 both were mild for much of the time and among the fi ve mildest in Slovakia since 1950 (data source: Slovak Hydrometeorological Institute, 2016). The mild winter weather may have supported the unexpected persistence of both C. septempunctata and H. variegata on exposed branches of pine trees. Adult H. variegata typically require leaf litter and/or dry grass as an overwintering habitat (Majerus, 1994;Roy et al., 2011;Nedvěd, 2015) but they may be overwintering on Scots pine trees as well.
Of the factors infl uencing recording of ladybirds on branches of Scots pine trees, fl uctuations of abundance in species such as C. septempunctata (cf. Honěk & Martinková, 2005;Selyemová et al., 2007;Honěk et al., 2016;) or H. axyridis (low abundance in the fi rst period but high in the second) need special consideration in the winter surveying of ladybirds. Some ladybirds may have fallen off the sampled branches and moved to other trees and other branches as a result of beating. The probability that this, or the movement of ladybirds to other habitats, could infl uence the temporal and/or spatial variation in species assemblages was very low, judged from the negligible number of branches and trees sampled in comparison to the number of trees across all investigated sites.
The overwintering of H. axyridis in coniferous forests, and specifi cally on Scots pine, has not previously been considered using systematically collected data. The known overwintering habitats of H. axyridis in nature include crevices in rocks and the bark of standing or fallen trees, including Scots pine, and overwintering adults are also found under the bark or underneath fallen logs (Hodek, 2012a;Pendleton & Pendleton, 1997-2016. In the Czech Republic, we (O. Nedvěd, unpubl.) have rarely recorded adult H. axyridis overwintering in an empty wasp nest and in a bee hive (a single record in both cases). Six years after its arrival and establishment in Slovakia (Brown et al., 2011;Panigaj et al., 2014), over the period 2014-2015, H. axyridis became the third most abundant ladybird species on Scots pine. Adults remained active on pine trees until mid-November, following the main dispersal time in October (Panigaj et al., 2014). The decreasing abundance of H. axyridis on pines later in the winter parallels the trends in Britain (Roy et al., 2011;Roy & Brown, 2015) where adults are extremely diffi cult to fi nd openly on Scots pine and their outdoor winter records are rare (Pendleton & Pendleton, 1997-2016. Other ladybird species were collected less frequently from branches. Adult M. octodecimguttata overwinter on the lowest portion of tree trunk (the tree butt) of Scots pine (Pulliainen, 1966) or in the litter (Nedvěd, 2015), and adult C. magnifi ca spend the winter near Formica ant nests, in leaf litter, dead bracken fronds, heather and pine foliage (Majerus, 1994). Adult S. suturalis overwinter under the bark of Scots pine near the tree base (Majerus, 1994;Nedvěd, 2015) or in moss (Roy et al., 2011), and we recorded them on branches in mid-March after leaving hibernacula. In the pine forest S. suturalis possibly entered overwintering sites before mid-October, unlike in Britain where this species may be found on pines until at least mid-October (P.M.J. Brown, pers. observ.). Adult A. ocellata spend the winter in soil (Roy et al., 2011) or leaf litter (Kučera, 1955;Nedvěd, 2015), and adult M. oblongoguttata choose moss pads or leaf litter as overwintering habitats (Roy et al., 2011;Nedvěd, 2015). Accordingly, we did not record these conifer specialists spending winter on pine branches. Another conifer specialist, H. quadripunctata, was only rarely found during winter. This species overwinters in the bark crevices or under the bark of Scots pine, rarely amongst pine foliage (Majerus, 1994;Roy et al., 2011;Nedvěd, 2015).
Our study revealed that species assemblages of ladybirds changed in the mild winter weather and were affected by the age of pine stand. The signifi cant effect of this site factor could refl ect speciesʼ preferences for selecting particular stands earlier during the growing season (cf. Bastian, 1982Bastian, , 1984Czechowska, 1995a) or during autumn dispersal (Galecka, 1991). The preference for stands of Scots pine at least 15 years old could be explained by better protection of adults against abiotic factors (wind, rain or snow) in these stands.
All of the ladybirds collected during winter were alive. This demonstrates successful survival of native species and H. axyridis on pine branches, although mortality was not assessed. According to McClure (1987) only 10% of caged adult H. axyridis survived a cool winter in North America with a mean temperature in January of -5.8°C and a minimum temperature in January of -17.8°C. The lethal temperatures for adult H. axyridis in outdoor hibernacula vary from -17.1 to -16. 3°C (Berkvens et al., 2010). In our study the lowest air temperatures in the winter of 2013/2014 (-11.5°C in January) and 2014/2015 (-13.3°C in December) were above the lethal values, hence, the mild winter weather probably reduced the mortality risk of this species, which is considered eurythermic (Barahona-Segovia et al., 2016). Resistance of ladybirds to subzero temperatures is high in the middle period of dormancy, however, early or late frosts can be very damaging (Hodek, 1973). An exception is E. quadripustulatus which can quickly acclimitise to low temperatures after activation from dormancy (Nedvěd, 1995).
Uncertainties about the effects of climate change on ecosystems highlight the need for thorough studies of species and communities (Campbell et al., 2005;Bale & Hayward, 2010;Bellard et al., 2012). The predictions of climate change within Europe for 2080 and 2100 show a robust and signifi cant increase in both January and annual mean temperatures (Murphy, 2000;Jacob et al., 2014). In the continental climate of Central Europe ladybirds are expected to experience increasing fl uctuation in winter weather both within and between years. As the winters of 2013/2014 and 2014/2015 were both exceptionally mild, the recorded species assemblages could differ from those in harsh winters. Therefore long term data, spanning various winter weather conditions, are needed to better predict ladybird responses to changing weather or different climates.
Our results suggest that Scots pine, being widely distributed (Rehfeldt et al., 2002), could be suited to winter surveying of ladybirds at large spatial scales. However, more research is required to assess the effect of spatial factors such as the forest type (monospecifi c pine stands vs. heterospecifi c stands with pine), forest margin aspect, sampling heights from trees (lower vs. upper branches) to optimize the surveys with regard to sampling effort, comparability of data and reliability of results. As the probability of recording ladybird species on the branches of Scots pine trees can be low during winter, judged from low values of relative occurrence for most recorded species in our study, high numbers of samples are required to enable robust quantitative data analyses.
The results verifi ed our predictions. (1) Ladybird assemblages on Scots pine during winter constituted fewer species than were recorded in late autumn or early spring.
(2) Overall abundance of ladybirds decreased from autumn to winter and remained lowest during this most hazardous time of the year for ladybirds, this temporal pattern being common to most species. (3) Species assemblages of overwintering ladybirds in pine forest showed a signifi cant temporal and spatial variation.
Further studies are required on the ultimate causes of temporal changes in species assemblages of ladybirds on Scots pine, such as photoperiod and temperature (Hodek, 2012a, b), and their trends and synergistic effects, as determinants of behavioural responses of ladybird species.