Salix transect of Europe: latitudinal patterns in willow diversity from Greece to arctic Norway

Abstract Background Willows (Salix spp.) are ecosystem "foundation species" that are hosts to large numbers of associated insects. Determining their patterns of distribution across Europe is therefore of interest for understanding the spatial distribution of associated fauna. The aim of this study was to record species composition at multiple sites on a long latitudinal gradient (megatransect) across Europe as a baseline for the future detailed analysis of insect fauna at these sites. In this way we used willow stands as comparable mesocosms in which to study floristic and faunistic changes with latitude across Europe. New information To determine spatial patterning of an ecologically important group on a latitudinal gradient across Europe, we sampled willows at the stand level in 42 sites, approximately 100 km apart, from the Aegean (38.8°N) to the Arctic Ocean (70.6°N), but at a similar longitude (21.2 to 26.1°E). The sites were predominantly lowland (elevations 1 to 556 metres amsl, median = 95 m) and wet (associated with rivers, lakes, drainage ditches or wet meadows). The median number of willow taxa (species and hybrids) per stand was four, and varied from one to nine. There is a progressive increase in willow diversity from south to north with the median number of taxa per stand in southern Europe being three, and in northern Europe six. A total of 20 willow species were recorded, along with 12 hybrids. The most widespread willow in the transect was Salix alba L. (occurring in 20 sites out of 42) followed by S. triandra L. (15 sites), S. caprea L., S. phylicifolia L. (14 sites) and S. myrsinifolia Salisb., Salix ×fragilis L. (13 sites). Voucher specimens from this study are deposited in the herbaria of the Natural History Museum (BM) and the Royal Botanic Gardens Kew (K). These samples provide a "snapshot" of willow diversity along a latitudinal gradient and an indication of the geographically changing taxonomic diversity that is presented to willow-feeding herbivores across Europe. It is anticipated that further papers will examine the insect fauna collected from these sites as part of this study.


Introduction
The ecological significance of the genus Salix Willows (the genus Salix L.) are "foundation species" (Ellison et al. 2005) in many wet habitats in the north temperate region. By providing an abundant food-source for many willow-feeding animals (generalists and specialists) they provide the basis for characteristic ecosystems (Brändle andBrandl 2001, Nyman et al. 2007). Willow leaves frequently show signs of leaf damage resulting from herbivore feeding. Herbivores include mammals: rodents (Tahvanainen et al. 1985b), deer, elk and, in the arctic, reindeer (den Herder andNiemelä 2003) and also phytophagous insects, notably Lepidoptera, Coleoptera and Symphyta Hymenoptera (sawflies) (Volf et al. 2015).
The abundant herbivores further support a predator trophic level, from birds (Sipura 1999), ants and predatory beetles, as well as large numbers of parasitic wasps (Callan 1940). The diversity of willow-feeding herbivores suggests that willows can be considered a "superhost". The concept of superhost is usually applied to hosts of galling insects (de Araújo et al. 2013). Willows do indeed host many galling insects, but also act as a superhost more generally for many guilds of herbivorous insects. In a survey of 25 European tree species, willows had both the greatest number of phytophages and the greatest number of specialist herbivores (Brändle and Brandl 2001).

Taxonomy of willow
The genus Salix in Europe is usually considered to be a difficult and confusing group for classification and identification (Karrenberg et al. 2002, Meikle 1992, Rechinger 1992, Skvortsov 1999. The main reasons for this are: (1) genetically-based morphological polymorphism, (2) phenotypic plasticity (3) the prevalence of hybridization (4) differences in taxonomic opinion. Although some willows (such as S. pentandra L.) are rather uniform, other species are highly variable. Salix myrsinifolia is a good example of a species that shows extensive polymorphism: notably in leaf indumentum (hairy to glabrous) and leaf shape (narrowly to broadly elliptical). Although willow polymorphism is rarely formally tested in common garden experiments it is likely that much of this polymorphism is genetically based as different morphs can be found mixed in populations, having developed under the same environmental conditions. Willows also exhibit phenotypic plasticity, such that even different plants of the same clone can look quite different, particularly if coppiced. Coppice shoots and their leaves can be very different from those of normal branches. However, probably the most remarkable and problematic aspect of willow taxonomy is the great ability for willows to hybridize. Crosses between quite unrelated species occur and many hybrids have a high degree of fertility. A recent study has shown that widespread hybridization is sufficient for chloroplast capture to occur even when species boundaries are maintained (Percy et al. 2014).
Coupled with this is the fact that many hybrids are of economic importance, due to their fast growth, and are widely planted. An example is the widespread hybrid Salix ×rubra Huds. (S. purpurea L. × S. viminalis L.). Another case is S. ×meyeriana Rostk. ex Willd. (S. euxina I.V.Belyaeva × S. pentandra) frequently planted as a more easily propagatable alternative to S. pentandra (Zinovjev 2011). Sometimes hybrids are so widespread they behave effectively as homoploid hybrid species. An example is the crack willow (S. × fragilis) which is a hybrid between S. alba and S. euxina (Belyaeva 2009) but which constitutes a characteristic landscape feature over much of Europe and which authors have in the past considered a species (Meikle 1984). Another case where taxonomic treatment varies is S. bebbiana Sarg. Here, we follow Skvortsov (Skvortsov 1999) in recognizing S. bebbiana as an Eurasian as well as a North American species, despite considerable variation across the range. However, many European authors (e.g. Bennett et al. 1991, Rechinger andAkeroyd 1993) consider the European S. bebbiana to represent a separate species, S. xerophila Flod. Salix bebbiana (=S. xerophila) is closely related to the glabrescent Pale Willow (S. starkeana Willd.). However, S. starkeana is a comparatively rare willow.

Geography of willow and stand level sampling
Species of the genus Salix have a long history of being mapped in Europe starting with the monumental Atlas Florae Europaeae project (Jalas and Suominen 1976). In turn, these data have been used for detailed analyses of geographic distribution using numerical methods at a continent-wide (Myklestad and Birks 1993) and regional (Myklestad 1993) scale. A more recent resource at the country level (with more up-to-date taxonomy) is that of the Euro+Med Plantbase (Uotila 2011). However, stand composition cannot be easily predicted from occurrences in large grid squares or whole countries. Willows in natural stands across Europe provide a distinctive ecospace for the willow-feeding organisms and understanding the changing stand-level taxonomic composition of the Salix species is important for understanding the host choice and distribution of willow herbivores. It is the stand that provides the landscape unit and the ecospace within which host choice operates. Also large-scale mapping projects often exclude hybrids, which may be an important part of natural stands and particularly important as they may possibly form "hybrid bridges" (Floate and Whitham 1993) for herbivorous insects to move between hosts. Furthermore, direct observation of natural willow stands, as in this study, allows the co-collection of herbivores with the collection of voucher herbarium specimens.
The collection of data over a long geogrphical distance falls into the category of studies now dubbed 'megatransects'. The power and utility of this technique has been amply demonstrated by numerous recent studies. Some recent examples include: Anstett et al. 2014, Barrios-Garcia et al. 2014, Hernández et al. 2007, Huber 2015, Senterre et al. 2004.

Stand selection
Willow stands were examined during two journeys by road across Europe: Greece to Poland in April 2015 and Poland to Norway in June 2015 (Fig. 1). Sites were selected by driving approximately 100 km north of the previous site and opportunistically locating a suitable habitat in which to find willows, generally a river or low-lying ground. The spacing of sites varied according to logistic and travel constraints. In southern Europe, willows are largely restricted to riparian habitats, but northwards they become commoner in many more habitats. The sampling unit was the willow stand (willow dominated local area). The requirement for a site was that it had a stand of willows that met certain minimum size requirements (at least 100 m in longest linear dimension). A stand of willows is defined as a contiguous area where willows are the dominant vegetation for at least 100 m in linear dimension (as for instance along a river bank). Some stands are very extensive, in which case our sampling was limited to approximately 200 m in largest linear dimension.
Because willow stands differ so much in size, shape and density, it was not practical to impose equal area or grid sampling. Time constraints limited entomological and botanical sampling to 1-2 hours per site. A total of 42 sites were sampled across Europe from south to north (38.8 to 70.6°N) while minimising east-west deviation to between approximately 21.2 and 26.1°E. In addition to the 42 sites, a series of "Supplemental sites" are recorded at which additional insect collections were made but the full site recording process was not carried out.

Data collection
At each site latitude, longitude and altitude data were collected using a hand held Garmin Etrex global positioning system, accurate to within 3 m. Basic notes on the immediate environment were taken to provide a habitat profile of the sites. At each site, the willow diversity was determined and voucher specimens were made in order to validate the species present and to capture variation in species that exhibited considerable phenotypic variation. If the willows were flowering, an attempt was made to collect both male and female individuals. Willow abundance relative to abundance of other woody plants was estimated on a four-point scale: 1) abundant -30% of individuals or more; 2) common -10-30%; 3) occasional <10%; 4) rare -one or two individuals only were detected. Samples were processed using standard herbarium techniques and specimens are deposited at the Natural History Museum, London (BM) or at the Royal Botanic Gardens Kew (K). Field identifications were made by QC and DMP. Confirmation, and critical determination of all vouchers, including hybrids, was done by IB. In addition to herbarium samples, samples of leaf tissue were dried using silica gel to permit future DNA-based studies and retained at NHM. Salix transect of Europe: latitudinal patterns in willow diversity from ...

Climate data
As background information, climate data for three contrasting individual sites is given from publically available data sources (Table 2). These use a dataset of mean historical monthly temperature (°Celsius) and rainfall (mm), computed globally for the period 1990-2009 by the Climatic Research Unit (CRU) of University of East Anglia (UEA) and available through the World Bank Climate Portal (World Bank Group 2015).   Table 2.

SITE Lat °N
Summary climate variables taken from publically available resources (see Methods) for three contrasting sites on the transect: 1 and 42, the most southerly and most northery sites on map ( Fig.  1) together with a middle site, 20, indicated by an arrow on map. Mean monthly temperature (°C) and mean monthly precipitation (mm) are given here. This table is provided as background information on the climatic gradient represented by the megatransect.

Results
Sites Table 1 shows the 42 sites recorded in this study as well as details of further "supplemental sites" where insect collections were made but without the level of sampling accorded to the main sites. The geographical distribution of sites is shown in Fig. 1. The supplemental sites will not be discussed here but their basic details are given, as subsequent papers on the insects sampled along the transect may refer to them. The latitudinal variation provides an enormous variation in climate. Table 2 shows summary climatic statistics for three sites: the most southerly, the most northerly and a central site (Poland).
As can be seen from Table 1, site elevations varied from 1 m to 556 m above mean sea level (amsl), with a median of 95 m. Because the sampling was predominantly in lowlands, the diversity of mountain or upland willows was not captured in this study, nor was it intended to be. Instead we capture the diversity of large stands of willow found in wet lowlying areas, which from an "insect eye view" or "insect chemosensory perspective" represent the largest areas of willow resource in the landscape, generally associated with landscape features such as rivers, lakes, drainage ditches or poorly drained meadows. Table 3 lists the total of 20 willow species that were recorded, together with the 12 hybrids.

Willows
For each taxon the total number of sites (out of 42) is given. In this transect, the most widespread willow is Salix alba, which occurs in 20 sites (out of 42). This species is followed by S. triandra (with 15 sites), S. caprea, S. phylicifolia (with 14 sites each) and by S. myrsinifolia, S. ×fragilis (with 13 sites).       has undergone a major radiation in boreal regions which may go some way towards explaining this inversion of the norm.
Finally, it should be noted that these willows formed the background for a major sampling of insects and it is anticipated that further papers forming part of this study will examine the insect fauna collected.