A Key for the Microhistological Determination of Plant Fragments Consumed by Carpathian Forest Cervids

: We present a microhistological key for identiﬁcation of plant fragments consumed and partially digested by free-roaming, forest cervids based on collection of 92 plant species representing forage availability of the Western Carpathian forests. The key represents a determination tool to facilitate microhistological analyses of faecal and ruminal material. We summarized, integrated, and developed current knowledge on microstructures of plants consumed by Cervidae using speciﬁc diagnostic features of plant fragments including type, shape, orientation, and arrangement of cells and stomata, type of venation, presence, and type of trichomes and crystalline inclusions. Since most plant species of the same taxa show common patterns in morphology of the different epidermal traits, we categorized collected material into seven functional botanical groups, i.e., grasses and sedges, herbs and leaves of broadleaved trees, needles, ferns and mosses, seeds and fruits, and genera Rubus , Rosa , Vaccinium . The key is consistent with classiﬁcations used in the majority of studies on diet of wild cervids and is supported with photographs of the main diagnostics features. The key has the potential to decrease amount of time needed for processing of the reference material, and to improve consistency between users studying feeding behaviour of forest cervids in central Europe.

Despite being a low cost and relatively accurate method, microhistology is limited by processing time, equipment quality and training of a user [29]. However, the most limiting factor for a wider use is the availability of the area-specific reference material [30]. Several region-specific keys have been published worldwide, such as for the region of the Great plains in the USA [31], (south-eastern USA [32], Yellowstone National Park in the USA [33], tropical regions of Northern India [34], rangelands of Argentina [35], arctic parts of Canada [30], and for the Ladakh, India [36]. Although several microhistological studies on wild cervids have been published throughout the Europe [22][23][24][25][37][38][39] a microhistological key is available only for the area of the Gödöllő Hills in Hungary [40], and to our knowledge, there is no key from the area of the Western Carpathians. In this study, we fill this gap by creating an identification key for the determination of a botanical association of partially digested plant fragments by wild cervids from this area. Our goal was to summarize, integrate and extend available data on microscopic characteristics of plant species consumed by wild cervids in the Western Carpathians and to create a tool for faster identification of undigested plant fragments. Our key is based on description of the arrangement, type and size of basic plant structures such as epidermal cells, stomata, inclusions (phytoliths, starch, etc.) and outer structures (trichomes). To minimize concerns about differential digestibility and the number of fragments that cannot be identified, we designed the key at the level of functional botanical groups. We believe that our key has the potential to greatly decrease the amount of time needed for collection and processing of the reference plant material, and that it will improve consistency between users studying feeding behaviour of wild cervids within the Western Carpathians as well as in areas with similar environmental conditions.

Reference Material
We collected reference material primarily from plant species that are known to be consumed by wild cervids. The preference of plant species by wild cervids was determined either by using published studies from Central Europe [25,26,41,42] or by direct observations during foraging and/or by signs of consumption by cervids. We also included maize (Zea mays L.) and wheat (Triticum spp. L.); plant species commonly used for supplementary feeding in winter (Table S1). We collected above-ground parts of each specimen, including stems and leaves. We excluded flowers because they are usually fully digested. For trees, we collected samples of twigs (<3 mm thick) including wood and bark and buds in winter and leaves during a vegetation season. Fruits of beech, oak, and grains of maize, wheat (Triticum spp.) and barley (Hordeum vulgare L.) were collected in early autumn. In total, we collected 92 plant species as a reference material, of which 19 species were trees (15 deciduous, 4 coniferous), 20 species were monocotyledons, 40 species were forbs, 9 species were shrubs, 3 species were ferns and 1 species of moss (Table S1).
All collected samples were dried at room temperature for 72 h. To soften plant material for the microscopic analysis, we soaked samples in hot water for 6-72 h, depending on the hardness of a structure. Next, we cut approximately 1 cm 2 of the material from all parts of each collected specimen, such as stem, petiole, node, internode, leaf tip, leaf blade, bud, needle, flower, seed, fruit and grain [30], or the biggest possible fragment in case of buds and needles. In parts with bilateral surfaces (leaves and needles), we extracted both abaxial (abx) and adaxial (adx) surfaces. To expose the cutinized epidermis, the plant tissue was scraped away from the inside of the examined surface by a scalpel [43]. In cases when fragments were very dark or from a hard material, we used bleach to remove additional pigments and to break down rigid structures inside fragments [44]. Each sample was then mounted in a drop of glycerine upside down on a slide and was observed and photographed under a microscope MOTIC BA 210 (Motic Electric, Linz, Austria) with 4 × 10, 10 × 10 and 40 × 10 magnification.

Morphology Diagnostics and Categorization
Description of the plant microstructures was performed following Metcalfe and Chalk [45], Metcalfe [46], Cutler [47] Esau [48] and Evert [49]. Our identification criteria focused on characteristics of structures that remain preserved after digestion ( Figure S1). We considered diagnostic features such as the shape, orientation, and arrangement of cells, the thickness of cell walls, the type of venation, the type, position and orientation of stomata, the shape of guard cells of stomata, the presence, type and arrangement of trichomes and inclusions, the presence and form of sclerenchyma, the presence and type of cork cells, libriform fibres and parts of tracheary elements such as tracheids and vessels. We categorized the reference material based on similar anatomical features into seven functional groups (Table 1), consistent with classifications used in the majority of studies on ungulate diet [25,42,50,51]: • Grasses and sedges-monocotyledons of grass-like appearance, such as the Order of Poales, Family Poaceae, Juncaceae, and Cyperaceae. • Herbaceous plants and leaves of deciduous trees-dicotyledons, including herbs (except for ferns and genera Rubus, Rosa, Vaccinium) and assimilation organs of deciduous trees and shrubs including buds. We grouped herbs and leaves of trees together due to their similar micromorphology, including shape and arrangement of epidermal cells, type and arrangement of stomata, type of venation and presence of variable types of trichomes. • Genera Rubus, Rosa, Vaccinium. We defined this category because of specific morphological structure regarding amount and type of trichomes, and presence of crystalic inclusions, making them well distinguishable from other herbaceous, and because this category has a high nutritional value as a forage source [52], and thus, a specific importance in evaluating forage of wild cervids [1,25,42]. We excluded anthers and awns from analyses because they spread uncontrollably through vegetation and they often contaminated samples for analysis. We also excluded an internal tissue of stems, due to lack of distinguishable characteristics to classify. Further, we excluded family Liliaceae and Asparagaceae from the key because they can be easily mistaken with grasses or sedges, they have microcharacteristics of monocotyledonous plants although botanists classify these families into herbs. Table 1. List of seven functional groups based on similar anatomical features as arrangement, type and size of basic plant microstructures consistent with classifications used in the majority of studies on ungulate diet with a description of detailed diagnostic features for each botanical category.

Herbaceous plants and leaves of deciduous trees
Epidermal cells of variable, mostly irregular-shape (hexagonal, round, elongated) [53], cells of elongated organs have elongated shape. Epidermal cells of leaves arranged diffusely, rarely parallel to the veins [53]. Usually, reticulated venation. Kidney-shaped guard cells of stomata (Amaryllis type) arranged irregularly or/and randomly [48]. Great morphological variability of trichomes, which occur in different taxa [45].
Presence of different types of crystalline inclusions found in almost all taxa [45].
Genera Rubus, Rosa, Vaccinium Venation, shape, and arrangement of epidermal cells identical to category Herbaceous plants and leaves of deciduous trees, except Vaccinium spp. that have also polygonal to rectangular epidermal cells irregularly beaded with straight, thickened, pitted walls [54]. Typical presence of kidney-shaped guard cells of stomata (Amaryllis type) arranged irregularly or/and randomly [45]. Rubus and Rosa: presence of hollow trichomes (simple-unicellular or stellate) sharply pointed, conical with smooth surface. Simple trichomes are twisted or convoluted (length to width ratio min. 10:1), or they are almost straight, narrowed gradually towards to the apex, bases of trichomes have rectangular shape to hexagonal shape with a rosette of epidermal cells; stellate trichomes (with a smooth surface, with pointed apexes, filiform and hollow arms almost equally long) are smaller compared to simple trichomes; sporadic presence of glandular trichomes with multicellular head or presence of multicellular tongue-shaped trichomes Numerous presences of randomly arranged crystals as druses or styloids, especially typical for Rosa [40,45,[54][55][56].
Vaccinium: presence of multicellular, glandular tongue-shaped trichomes bent over parallel to the midrib when situated to the apex, also presence of glandular trichomes consist of biseriate or /and multiseriate stalks bearing heads of variable size and small conical with wart-like surface [45].

Needles
Elongated rectangular shape of epidermal cells. Epidermal cells arranged in rows [57] parallel to vein. Leaves generally have simple venation and have only one or two long veins running down their centre [58]. Stomata arranged in longitudinal rows forming dense stomatal bands [58]. Stomatal guard cells sunken below the surface of the epidermis [49]. Presence of resin canals in the mesophyll [59].

Wood and bark
Presence of parts of tracheary elements and/or vessels elements, fiber tracheids and cells with accompanied libriform fibers or parenchymal cells, crystal-containing cells and sclereids or their combination [49] Presence of sclereids of various forms such as unbranched (brachysklereids) or sclereids of irregular shapes with a different number of protrusions, as well as large heterogeneous assemblage of sclereids (asterosclerides) [49]. Typical presence of soft cork cells (thin-walled cells of variable shapes) and/or hard cork cells (thick-walled cells in the shape of a square) [60]. Presence of phytoliths; calcium oxalate crystals are common in the secondary phloem of conifers [49].  [61]. Right (true) vascular tissue contains phloem and xylem absent in mosses [61]. Exclusively presence of Mnium type of stomata [63]. Presence of sporangia on the abaxial side of the fern leaves form clusters, sometimes covered by sharps [62].

Seeds and fruits
Various type of thick wall cells: sclereids, variable oriented layers or clusters of sclereids, irregularly arranged parenchymal cells with a strongly thickened cell wall, cells with pitted walls, parenchymatic cells and/or irregular polyhedrons with internal filling [48], cells usually arranged diffusely. Presence of cross cells with thick lignified walls or/and tube cells (lignified cells elongated parallel with the long axis of the grains). Sporadic presence of stomata randomly arrangement [49]. Presence of starch, starch grains, trichomes.

Discussion
Our key is a simple and effective tool for identification and categorization of plant fragments in faeces of wild cervids of the Western Carpathian region. Roots (except Solanum tuberosum L.), rhizomes, and inflorescence were not included in the key. Each plant part (leaves and stem) of a single species usually includes different diagnostic features. For this reason, microhistological keys cannot be purely dichotomous [30]. Our key allows identification of plant fragments via several steps. First, the key divides possible choices into two groups (coded 2 and 5) based on the arrangement of cells (regular vs. irregular), which narrows possible choices and leads to further categorization based on the shape of cells (for example, 1 → 2 → 3 or 4 → final category) or thickness of the cell walls (1 → 5 → final category). Further identification is based on additional diagnostic features such as trichomes, stomata, sclereids etc., and leads to the final choice. This key helps to identify

Discussion
Our key is a simple and effective tool for identification and categorization of plant fragments in faeces of wild cervids of the Western Carpathian region. Roots (except Solanum tuberosum L.), rhizomes, and inflorescence were not included in the key. Each plant part (leaves and stem) of a single species usually includes different diagnostic features. For this reason, microhistological keys cannot be purely dichotomous [30]. Our key allows identification of plant fragments via several steps. First, the key divides possible choices into two groups (coded 2 and 5) based on the arrangement of cells (regular vs. irregular), which narrows possible choices and leads to further categorization based on the shape of cells (for example, 1 → 2 → 3 or 4 → final category) or thickness of the cell walls (1 → 5 → final category). Further identification is based on additional diagnostic features such as trichomes, stomata, sclereids etc., and leads to the final choice. This key helps to identify fragments, not whole plants or plant species; and not all parts of a species are identifiable using epidermal features. For example, some species produced fragments that proved difficult to differentiate using microhistological features, such as Liliaceae and Asparagaceae from monocotyledons. Genera Rubus, Rosa, Vaccinium can also be mistaken with other species of the family Rosaceae which are categorized as herbaceous plants and leaves of deciduous trees. However, based on detailed observation of stomata and type of trichomes can be clearly distinguished. Certain features of some plants that were included in the key are not apparent on the photos, such as the gloss of epidermis on needles. For deeper identification, we recommend using the literature from Metcalfe and Chalk [45,46], resp. the key from the Gödöllő Hills [40] that allow identification at the species level. However, species level identification requires concominant presence of several specific diagnostic features which are often missing and thus should be approached with caution.
On the other hand, certain microscopic features allow to distinguish between some categories with high certainty. For example, regular brick-like arrangement of cells is typical for grasses and needles, while irregular-shaped cells arrangement is typical for herbaceous plants, leaves of deciduous trees, ferns and mosses. Other highly distinctive characteristics is the arrangement of venation, which can be used for distinguishing between monocotyledons from dicotyledons. Further, majority of monocotyledons and dicotyledons, but also ferns and mosses have thin-walled cells, while cells of stems, tree bark, cuticle of seeds and fruits have mostly thickened cell wall. Type and arrangement of stomata, shape of guard cells and subsidiary cells is another important diagnostic feature. For example, the Mnium type of stomata occurs exclusively in the category of mosses and ferns, the dumbbell shape of guard cells of stoma (type Graminae) is typical for grasses and sedges and Amaryllis type of guard cells of stomata occur in herbaceous plants and leaves of deciduous trees, including genera Rubus, Rosa, Vaccinium, and also needles, fruits, and seeds. Moreover, stomata may be at the same level as other epidermal cells in plants but deeply sunken below the level of the surrounding cells in needles, and classifications of stomata according to type of subsidiary cells (anomocytic or brachyparacitic) and amount and type of trichomes may be used to distinguish genera Rubus, Rosa, Vaccinium from other herbaceous plants and leaves of deciduous trees. Finally, "short cells and long cells" occur only in grasses and sedges, sporangia occur exclusively in the category of ferns and mosses, and cross cells and tube cells occur only in the group of seeds and fruits. Cork cells with tracheary elements, vessel elements and libriform fibres are typical for wood and bark and very fine structured fragments are typical for mosses and ferns.

Conclusions
The purpose of this study is to provide a tool to identify plant fragments in the faecal remains of forest cervids inhabiting Western Carpathians, in order to improve knowledge on feeding behaviour of free-roaming cervids. Rapid growth of large herbivore populations in Slovakia and across Europe increases intra-and inter-specific competition for food which results in ever increasing damages on forest ecosystems. Since the rough terrain and dense-canopied mixed forests prevent direct observations of food habits, microhistology is a suitable procedure to study feeding behaviour through indirect observations. Using this key will allow researchers to investigate feeding behaviour of cervids with relative ease, and it is expected to save time that can be devoted to other management goals.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/f12091229/s1, Table S1: A list of vascular plant species (except for moss) and their parts (leaf, stem, bud, bark and wood, fruit, grain) collected within Western Carpathians, Slovakia and used as a reference material for the microhistological key. Figure S1: Comparison of undigested plant fragments from reference material and digested plant fragments from faecal matter within the same category.
Funding: This research was funded by the projects VEGA 1/0797/19 and APVV-14-0637 supported by the Ministry of Education, Science, Research and Sport of the Slovak Republic and the projects "Centre of Excellence: Adaptive Forest Ecosystems" (ITMS 26220120006) and "Completing the Centre of Excellence: Adaptive Forest Ecosystems" (ITMS 26220120049) supported by the Operational Programme Research and Development within the European Regional Development Fund.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.