Relationships between age, time of emergence, and leaf traits of individuals were investigated in a population of Erythronium dens-canis L. in a hilly woodland area named Farneto-C, near Bologna, Italy. In 2015, 591 individuals were counted, 19 of which were flowering (FLO), 442 were mature non-flowering (MNF) and 130 were juveniles (JUV). FLO emerged at the end of February, whereas most MNF and JUV appeared at the middle and end of March, respectively. The mean aboveground survivorship of MNF was 24 days. Most MNF had large, oval to shield-shaped leaves with red-brown mottling, whereas most JUV leaves were smaller, usually oblong or lanceolate with a rough maculation or none. These results suggest that both timing of emergence and leaf shape are related to the age of the bulb. Based on leaf background, plants were classified into three major types with a likely genetic basis in the 2015 and 2016 surveys (the latter limited to FLO): a dominant silvery type (SLV, 62–74%), silvery-and-green type (S&G, 23–32%), and a less frequent vivid-green type (GRN, 3–5%). Several subtypes were also identified, but only one was dominant within each type. The three basic patterns appear to be phenotypically stable and no differences between MNF and FLO were found; once the juvenile stage has passed, each plant produces the same leaf type year after year. In addition, our results on the discoloration time-course of red-brown spots suggest that the functional role of leaf mottling is not related to pollinator attraction. Instead, leaf mottling could play a role in camouflage against herbivores.The observed massive grazing on flowers, more than leaves, could explain why the frequency of mature individuals was biased towards the non-flowering ones.

demography, discoloration, leaf traits, mottling, spring ephemerals, survival rate

Spring ephemerals can be considered as model species for addressing plant responses to environmental changes (Lapointe 2001), especially since climate change has become a major issue for ecologists and conservationists (e.g. Dormann and Woodin 2002, Totland and Alatalo 2002, Parmesan 2006, Pfeifer et al. 2006, De Frenne et al. 2011, Marchin et al. 2014). As seasonality determines changes in soil and atmospheric temperature as well as in nutrient, light and pollinator availability, responses displayed by spring ephemerals are of various kinds and involve adaptation in many physiological, morphological, and reproductive traits, implying a fine control of phenological transitions (e.g. Lapointe 2001, Yamagishi et al. 2005, Augspurger and Salk 2017). For example, species within the genus Erythronium L. (Liliaceae, Tulipeae) are typical spring ephemerals that have been the object of several studies dealing with responses to seasonal changes (Lambert et al. 2010, Gandin et al. 2011, Kim et al. 2015, Augspurger and Salk 2017), especially those involving effects of temperature on phenology of leaves and bulbs as well as on seed germination.

All Erythronium species are bulbous geophytes with the phenological characteristics of spring ephemerals (Schemske et al. 1978, Lapointe 2001).They play a relevant role in some forest ecosystems due to their high biomass and effectiveness in preserving soil nutrients (Muller and Bormann 1976, Muller 1978). This genus is distributed in the northern hemisphere with five Eurasian species (Bartha et al. 2015) and about 24 species in North America (Mathew 1992, Allen and Robertson 2002, Allen et al. 2003). The numerous American Erythronium species (trout lilies) have life cycles that are quite different between high-mountain and forest taxa, but all are strictly connected with snow and snowmelt (Vézina and Grandtner 1965, Lapointe 2001, Yamagishi et al. 2005), so these plants are particularly exposed to the effects of climate change. In fact, mean bulb size decreases with increasing soil temperatures (Gandin et al. 2011) and, in recent years, there is concern for a growing pollination mismatch (Lambert et al. 2010, Thomson 2010). For the Japanese species E. japonicum, too, several contributions have been devoted to the phenological and ecophysiological responses to temperature variation in seeds, bulbs, and leaves (Yoshie and Fukuda 1994, Sawada et al. 1997, Kondo et al. 2002).

On the contrary, few papers have considered the life cycle and ecology of E. dens-canis (dog’s tooth violet) (Guitiàn et al. 1999, Mondoni et al. 2012). This is the only European species of the genus Erythronium; it is distributed in the southern part of the continent, ranging from the Iberian Peninsula through southern France and northern Italy to the whole Balkan Peninsula down to Greece, with some populations pushing north of the Alps and some others east of the Balkans (Govaerts 2017). The species’ distribution range is apparently limited by the January 0°C isotherm to the north and by long summer drought to the south. In Italy, E. dens-canis is widely distributed in deciduous woods at the southern edge of the Alps and in the northern and central Apennines, with frequent gaps (Pignatti 1982, Kleih 2010). The plants flush from bulbs in late winter and carry out their yearly photosynthetic and reproductive cycle within a couple of months by exploiting full sunlight in the understory; then, they rapidly senesce and die at the time of tree leafing out and closure of the forest canopy.

The single flower of E. dens-canis is a small lily, white to rosy in colour, nodding at the top of a red stalk and with a basal pair of elongated, finely mottled leaves. However, many mature plants in natural populations do not flower and are characterized by a single, large leaf. These non-flowering individuals most probably have not yet attained the sexually competent stage (La Rocca et al. 2014). Very young individuals and seedlings have a small leaf with a tenuous, unapparent design of light-brown spots, or a green leaf with no pattern at all. The dominant leaf mottling type of mature E. dens-canis plants consists of red-brown spots or patterns on a grey-glaucous (“silvery”) background, the latter originated through detachment of the epidermis from mesophyll cells resulting in light reflection (La Rocca et al. 2014). The red-brown motif is produced by a single layer of mesophyll cells containing a red vacuolar pigment (anthocyanin; Esteban et al. 2008), which, however, will vanish within a few weeks leaving bright-green areas (La Rocca et al. 2014). Concerning the function of mottling, three alternative hypotheses have been proposed: 1) a specific function in pollinator attraction to flowers (La Rocca et al. 2014), 2) a possible role in photoprotection (Esteban et al. 2008), and 3) as camouflage against herbivores (Givnish 1990).

Due to the importance of obtaining demographic data for investigating plant responses to environmental change, the present work addresses a population of E. dens-canis occurring in an area of the Nature Park of Gessi Bolognesi (Farneto). In particular, we aimed to answer the following questions: 1) Is there any relationship between age and time of emergence of individuals? 2) Are leaf shape and leaf mottling related to the phenological stage of individuals? 3) What is the functional role of mottling in the species? By answering these questions, we can provide insight into the relationships between individual leaf traits and population structure (i.e., age classes) and dynamics (i.e., time of emergence and survival) in E. dens-canis.

The Erythronium dens-canis population examined at Farneto-C was mainly formed by MNF, mostly emerged at mid-March, and characterized by a single, large, usually OV or SH leaf. In February 2015, a small number of FLO preceded the outburst of MNF, as also occurred in the 2016 sampling of FLO. The early emergence of Erythronium flowers is due to the relatively long time required for flowering, fruit maturation and bulb renewal (about two months) before tree canopy closure and possible dry periods set in. Since FLO are credited with an age of 6-7 years (La Rocca et al. 2014), most MNF individuals are probably 3- to 6-years old. JUV amounted to ca. 22% of the total, and included seedlings and other very young specimens mainly with small EL or LA leaves. They had an estimated age of 1 to 3 years and tended to emerge later than MNF. In other words, both leaf shape and time of emergence seem to be a function of the physiological age of the bulb. Hence, our data suggest that physiological ripeness and timing of spring emergence are inversely correlated: the younger the bulb, the later it will emerge.

Concerning leaf background, the three basic types found at Farneto-C appear to be phenotypically stable: once the juvenile stage is over, every year each plant produces the same leaf type. Therefore, a S&G or GRN-MO leaf does not represent a developmental stage leading to the widespread SLV type: SLV, S&G, and GRN-MO plants coexist in what appears to be a genetically balanced polymorphism. However, the higher percentage of SLV leaves and lower percentage of S&G leaves in FLO compared with MNF might suggest a transition of some S&G plants to the widespread SLV-PC type. This point needs confirmation on a longer time scale. Within each of the three principal background types, only one subtype is likely to represent the final, stable form: for SLV plants it is the PC subtype, for S&G plants it is the CH subtype, and for the GRN plants it is MO. Other variants are probably juvenile or transitional characters, or they may represent a local response to a harsh environment (see below).

The main leaf traits of dog’s tooth violet observed in other localities (La Rocca et al. 2014; personal observations) appear similar to those of Farneto, whereas the frequency of minor pattern variants seems to fluctuate. For example, rusty specimens (silvery with nearly complete red-brown coverage) occur with some frequency on Apennine highlands, e.g., in the Contrafforte Reserve (personal observations; Fig. 4F). Online photographs of E. dens-canis from official websites (e.g., University of Trieste, project Dryades http://dryades.units.it/euganei/index.php?procedure=taxon_page&id=6797; Forum Acta Plantarum http://www.actaplantarum.org/floraitaliae/viewtopic.php?t=2723) support the general prevalence of the PC subtype, although they occasionally show a leaf design of an infrequent type, e.g., rusty. In any case, some phenotypes often found at Farneto (S&G-CH and GRN-MO) seem to be rare or possibly absent in other Italian populations (personal observations). This issue deserves further investigations on the variability and genetics of E. dens-canis throughout its range.

The discoloration time-course of the red-brown spots on E. dens-canis leaves was similar for all age groups, with some exceptions. The fact that adult plants (flowering or not) exhibited similar leaf patterns (background and spots) and underwent parallel discoloration processes is inconsistent with a specific function of mottling in pollinator attraction to the flowers, as hypothesized by La Rocca et al. (2014). On the other hand, a possible role of leaf mottling in photoprotection (Esteban et al. 2008) seems consistent, at least, with the presence of a rusty highland variant. However, one of the most likely explanations of Erythronium’s mottling remains camouflage against herbivores, as proposed by Givnish (1990). In our case, discoloration timing is consistent with mimicry: the elaborate red-brown pattern could mimic soil colour and dead leaves, and when it finally disappears, to be replaced by a vivid green drawing, the time is ripe for tree leafing out, increased shadowing, and herb growth in the understory. At Farneto, roe deer actively fed on E. dens-canis flowers rather than on leaves, with a selective preference for still closed, nutrient-rich blossoms in late winter, and fruits in spring (Pupillo and Marconi 2016). Guitiàn et al. (1999) also reported on fruit nibbling in Spain. The average aboveground survival rate of MNF was fairly long (24 days) testifying that leaves were less heavily grazed; when it was short, this had often to do with trampling and digging by wild boars rather than grazing. Although data are not available on the survival rate of FLO, results similar to MNF were expected and confirmed by our observations. Interestingly, we noticed that flowers damaged in 2014 failed to bloom in 2015, whilst some flowered in 2016, suggesting lengthy recovery times. Thus, ungulates may limit the reproductive success of E. dens-canis and contribute to the biased frequency of mature individuals towards the MNF. Indeed, in the study site, only flowers growing on slopes or under tree cover seem to have a chance of bringing their capsules to maturation. Similarly, Muller (1978) found that about 99% of the E. americanum plants crowding in the Brookhaven Forest were MNF, so that our results may be more common than expected. In conclusion, leaf mottling could play a role in camouflage against leaf-eaters, but it appears inefficient against flower-eaters.