Temperature and storage time strongly affect the germination success of perennial Euphorbia species in Mediterranean regions

Abstract This study aims to explore the effect of environmental factors (temperature, light, storage time) on germination response and dormancy patterns in eight Mediterranean native wildplants, belonging to the Euphorbia L. genus. In detail, we considered E. amygdaloides subsp. arbuscula, E. bivonae subsp. bivonae, E. ceratocarpa, E. characias, E. dendroides, E. melapetala, E. myrsinites, and E. rigida. We collected seeds from natural plant populations and performed germination assays in climatic chambers at seven constant temperatures (from 5 to 35°C, with 5°C increments), and four fluctuating temperature regimes (8/15, 8/20, 8/25, and 8/30°C, with a 12/12 hr thermoperiod). Germination assays were set up both in dark (D) and in light/dark conditions (L/D, 12/12 hr photoperiod), after short and long seed storage (SS around 30 days and LS around 150 days). For all these species, except E. amygdaloides subsp. arbuscula, results show that the final germinated proportions were improved by a long storage period (>150 days), which supports the existence of nondeep physiological dormancy. Optimal temperature levels ranged from 14.3 to 21.3°C and base temperatures ranged from 5.6 to 12.1°C, while ceiling temperatures from 25.6 to 34.7°C. For none of these species, germinations were favored by an alternating daily temperature regime, while in several instances, germinations were quicker and more complete in darkness, than in an alternating light/dark regime. In some instances, extreme temperature levels (5 and 30°C) induced dormancy and germinations did not resume when seeds were exposed at optimal temperature levels. Results are discussed in terms of the dynamics of emergences and how this might be affected by climate changes.


| INTRODUC TI ON
Seed germination behavior is a fundamental trait to determine the ability of plants to successfully establish in a given habitat or geographical region. Indeed, seeds need to be able to germinate in the right timing and location, so that the chances of seedling survival and establishment are maximized. A key role is played by several environmental factors, including temperature (T), light, and water availability. In particular, the effect of temperature is related to threshold levels for seed germination, such as base temperature (T b ), optimum temperature (T o ), and maximum or ceiling temperature (T c ), which form the basis for the concept of thermal time.
In some cases, high (and low) temperatures can also induce secondary dormancy, according to a phenomenon known as thermodormancy (Bewley & Black, 1985). This was observed especially in annual plants of deserts as well as in some Mediterranean species.
Light is also important, and, in this respect, previous research has shown that some plants give better germination capabilities when exposed to an alternating light/dark regime .
In addition to the climatic variables, germination is also controlled by endogenous factors, such as primary dormancy. Although this is a potentially expensive trait (Willis et al., 2014), it has been recognized as an adaptive mechanism, by which the plant avoids germination in unfavorable conditions for seedling establishment (e.g., excess heat and drought) and postpones them until the season is more favorable .
Among the endogenous factors, several studies have shown that seed germination is genetically determined and the phylogenetic signal may be a significant constraint to interspecific variation within a genus (Zhang, Du, & Chen, 2004). This was found, for example, for some Romulea species in Mediterranean habitats (Carta, Hanson, & Müller, 2016), even though a contrasting behavior was observed with other genera, such as Stellaria and Nothofagus, which showed considerable interspecific variability in germination behavior (Arana et al., 2016;Vandelook, Van de Moer, & Van Assche, 2008).
Morphological traits, such as seed size, seed mass, seed shape, and seed dispersal, can also play a role in seed germination behavior as shown, for example, in Verbascum sp. pl., Euphorbia humifusa and Solanum nigrum Wang et al., 2016).
Studying the relationship between environmental factors and plant physiology is fundamental to understand and predict ecological dynamics, depending on the climate characteristics of a certain location, such as latitude, altitude, soil moisture, temperature and rainfall patterns, light, and photoperiod Cristaudo, Gresta, Restuccia, Catara, & Onofri, 2016;Gresta, Cristaudo, Onofri, Restuccia, & Avola, 2010;Zhang et al., 2017). This is particularly important for Mediterranean regions, which are characterized by a high variability of environmental conditions, where wet and mild periods are alternated with dry and hot periods, which make seedling survival very difficult. In this respect, climate change and global warming should be carefully considered, as they can pose additional problems for seed germination and plant recruitment Mondoni, Rossi, Orsenigo, & Probert, 2012;Walck, Hidayati, Dixon, Thompson, & Poschlod, 2011), which can modify the geographical distribution of plant species (Poschlod et al., 2013).
Climate change has been shown to produce a more erratic rainfall pattern, with higher temperatures and more frequent conditions of water shortage. A few studies have addressed the effects of water stress on germination in Mediterranean species. Conifers have shown high tolerance to water stress (Boydak, Dirik, Tilki, & Çalikoğlu, 2003;Thanos & Skordilis, 1987), whereas shrub species have shown variable responses, from high (e.g., Antyllis cytisoides; Ibanez & Passera, 1997) to low or moderate tolerance (e.g., Genista scorpius, Cistus monspeliensis, C. salviifolius, Calicotome villosa; Bochet, García-Fayos, Alborch, & Tormo, 2007;Chamorro, Luna, & Moreno, 2017;Pérez-Fernández, Calvo-Magro, & Ferrer-Castán, 2006). Negative effects of water stress can persist for several years, through maternal effects which lead to less viable seeds with lowest germination capacity. Unfortunately, the effects of climate change on seed germination seem to be highly variable, depending on plant species, populations, and sites, which makes it difficult to make predictions for a given species at a specific site. This motivates research on the germination responses to climate factors, such as temperature, light, and water availability.
In this study, we present a novel assessment of seed germination strategies, focusing on some taxa of the Euphorbia L. genus (Euphorbiaceae Juss.). This is the largest genera of the Euphorbiaceae family, also known as spurge family, and it is one of the five most species-rich genera of the angiosperm group, with around 2,000 species (Frodin, 2004;Govaerts, Frodin, & Radcliffe-Smith, 2000), that occur in all temperate and tropical regions. Some of them have a considerable economic importance in medicine (Rahman & Akter, 2013) or for revegetation, landscaping, and xero-gardening in semiarid environments (Benvenuti, 2014;Franco, Martínez-Sánchez, Fernández, & Bañón, 2005).
( Figure 1). The investigated species are phylogenetically related and belong to the monophyletic subgenus Esula Pers., one of the four major clades within the genus Euphorbia (Geltman, 2015;Riina et al., 2013). Some of these species are widespread across the Mediterranean region, in a wide range of altitudes, while three of them are narrowly distributed (endemic).
The seeds of all species in the group possess a caruncle, but differences in the size and shape of the caruncle are pronounced.
Considering those eight species, the objectives of this study were as follows: (a) evaluate the relationships between temperature, light, storage time, and germination behavior; (b) estimate threshold temperatures for seed germination and thermal time; and (c) verify the existence of thermo-inhibition and/or thermo-dormancy.  (Table 2).

| Fruit collection and seed storage
Fruits were collected from randomly chosen individuals (>50 individuals per species) in a relatively big area (200-500 m 2 ) to obtain an adequate representation of genetic diversity. Plants were randomly selected from the middle of each population, in order to avoid any edge effects and any disturbances to the natural spread of the population (sustainable harvesting).
In the laboratory, the fruits were air-dried at room temperature (25°C ± 2°C), for about 3 weeks, in plastic box perforated for ventilation with bottom surfaces covered with blotting paper, to absorb humidity; mesh lids were used, to ensure air circulation and prevent seed losses, due to the "explosion" of fruits. After their release, seeds were cleaned from the remaining fruit tissues using a stack of sieves. For each taxon, the weight of five replicates of 20 randomly chosen seeds was determined ( Table 1). The average weight of one dry seed, expressed in mg, of each species was determined using a precision balance with an accuracy of 0.0001 g (Mettler AE 50). A preliminary cut test on a subsample of seeds showed that the viability was nearly 100%, for all species. The collected seeds were stored in paper bags in laboratory conditions (22 ± 2°C, 50% RH), until the beginning of the germination experiments, starting approximately 1 month from the collection, for all species, regardless of the sampling year.

| Experimental design
Germination experiments were carried out in automatic temperature-, humidity-, and light-controlled growth chambers (Sanyo  For each germination assay, four replicates of 25 randomly selected seeds were used for each treatment and studied species. Seeds were placed in 9 cm diameter Petri dishes, on top of three layers of filter papers (Whatmann No. 1), previously moistened with 5 ml of distilled water and incubated in growth chambers at the different temperature regimes.
For darkness treatments, the Petri dishes were wrapped in two layers of aluminum foil and kept in the same growth chamber. All  Table 2 Petri dishes were sealed using Parafilm ® to avoid moisture losses; water was added to the dishes as needed to keep an adequate moisture level.
Seed germination was monitored daily, and germinated seeds were counted and removed from Petri dishes. Seeds were considered as germinated when the radicle protrusion was about 2 mm.
In continuous darkness treatments, counts were made under green-filtered light (Philips PF710E), which had not shown any effect on seed germination, as assessed by preliminary assays. In L/D, germination counts were performed during the light period.
Experiments were continued for 30 days. At the end of incubation period, the viability of the remaining seeds was estimated by cut test; seeds with white, hard embryos were considered to be alive.
Empty and dead seeds were excluded from the calculation of final germination percentages.
After the end of assays, ungerminated seeds exposed at 5 and optimal for all species), to assess whether they could recover germination or whether they had become dormant.

| Data analyses
For each Petri dish, the observed counts were used to parameterize a log-logistic germination model, by using a time-to-event modeling platform (Onofri, Benincasa, Mesgaran, & Ritz, 2018;Ritz, Pipper, & Streibig, 2013). The fitted model was used to derive the final percentage of germinated seeds (FPG) and the time to 50% germination (T 50 ), which were submitted to ANOVA. FPGs were arcsine-squareroot transformed and T 50 were log-transformed prior to analyses, in order to meet the basic assumptions for linear models; back-transformed means were used for tables and graphs.
Considering the assays at constant temperature, the number of days to achieve 10% and 30% germination was also derived from the fitted germination models and their reciprocal values, together with the reciprocal of T 50 values, were taken as the germination rates respectively for the 10th percentile (GR 10 ), 30th percentile (GR 30 ), and 50th percentile (GR 50 ) (Bierhuizen & Wagenvoort, 1974).
The observed GRs for each species, light regime, and storage time were used to parameterize the following thermal-time model, derived from Mesgaran, Onofri, Mashhadi, and Cousens (2017): where g is the percentile (10, 30, or 50th), T is the temperature, Θ T (g) is the thermal time to germination for the g-th percentile, T b is the base temperature, and k relates to the decrease of germination velocity when temperature exceeds the optimal level. Considering fluctuating temperatures, the effects were rather small, although higher FGPs were noted with E. amygdaloides subsp.

| Thermo-dormancy and/or thermo-inhibition
For all species, no germination was observed at 5°C for 30 days.  (Figure 7).

| Germination velocity
Considering the germinated fraction, the time (days) taken for 50% of seeds to germinate (T 50 ) generally decreased as temperature increased from 5 to 20°C (Figures 8 and 9

| Thermal-time parameters
Threshold temperatures (T b , T o, and T c ) are reported in  Thermal times (°Cd) calculated for different germination percentiles (Θ 10 , Θ 30, and Θ 50 ) are shown in Table 3. Values for Θ 50 ranged between 20 and 100°Cd. The extremes were the E. myrsinites, which had a high T b (ca. 13.0°C) and low thermal time (ca. 21 degree days), and E. characias which had a low estimated T b (ca. 0.5°C) and a high thermal time (100 degree days). We tried to classify the abovementioned dormancy, according to Soltani, Baskin, and Baskin (2017). In this respect, it is possible to note that storage time did not only influence germination capability, and it also increased ceiling temperatures for E. bivonae, E. ceratocarpa, E. characias, E. dendroides, E. melapetala, and E. rigida.

| D ISCUSS I ON
Accordingly, these species should be classified in the Type 1 nondeep PD category. On the contrary, storage time did not affect ceiling temperature in E. myrsinites, at least for germination in complete darkness. This behavior might support the idea that this species should be included in the Type 5 nondeep PD category (Soltani et al., 2017).
Once primary dormancy has been released, in the absence of other limiting factors (e.g., water), seed germination of the eight Euphorbia species seemed to be mainly driven by temperature. The highest germination capabilities and speeds were reached at optimal temperatures levels, ranging from 14.3 to 21.3°C. Base temperatures ranged from 5.6 to 12.1°C, while ceiling temperatures ranged from 25.6 to 34.7°C. In Mediterranean regions, under present climatic conditions, these results support a germination peak from early autumn to early winter, while mid-winter germination is prevented by temperatures below the base level requirements. Early spring germination is also possible, although germinations may progressively become more difficult as the season progresses, because of water shortage (Fenner & Thompson, 2005). In summer, high temperatures (above the T c level) may prevent germinations, until the unfavorable F I G U R E 6 Germination of nongerminated seeds of six Euphorbia species after 30 days of exposure at the benefit temperature (20°C). Seeds were first exposed at 5°C for 30 days, where no germinations were observed (see Figures 3 and 4). Immediately afterward, seeds were reincubated at 20°C for 30 days conditions are overcome. These results confirm that cardinal temperatures for germination represent a mechanism of adaptation, by which a given species can match its germination timing to favorable conditions for seedling recruitment (Huang, Liu, Bradford, Huxman, & Venable, 2016;Qiu, Bai, BiFu, & Wilmshurst, 2010;Thompson, 1970). Furthermore, the evident reduction in the germinative performance at both high (≥30°C) and low (5°C) constant temperatures may also dictate the altitudinal and latitudinal limits for the geographical distribution of these species.
During the unfavorable periods (winter and summer), the germination of these eight Euphorbia species may be obstacled either by thermo-inhibition or by thermo-dormancy. Our results show that seeds of E. bivonae, E. characias, and E. dendroides were thermo-inhibited at high temperatures (Horowitz & Taylorson, 1983), while seeds of E. amygdaloides subsp. arbuscula, E. ceratocarpa, and E. rigida showed the existence of some mechanisms of thermo-dormancy.
We also tested the effect of fluctuating temperatures, and we  (Best, Bowes, Thomas, & Maw, 1980;Brecke, 1995;Narbona, Ortiz, & Arista, 2006). It is possible that this phenomenon can be related to their large seeds or processes of seed dispersal. Indeed, many Euphorbia species are diplochorous and they are characterized by a primary explosive dispersal system and a secondary ant dispersal system, aided by the presence of the caruncle. Ants may bury the seeds inside their nest or may abandon them outside, on waste piles.
In these conditions, the capacity to germinate in darkness can be ad-  Priority Axis III, Specific Objective 3.1, Project code C1-3.1-16.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

AUTH O R CO NTR I B UTI O N S
AC, SC, AM, and AR conceived the ideas and designed methodology; AC, SC, and AR collected the data; AO and AM analyzed the data; AC and AO led the writing of the manuscript. All authors contributed critically to the drafts and gave final approval for publication.

DATA AVA I L A B I L I T Y S TAT E M E N T
The main dataset is available at https ://doi.org/10.5061/dryad.

m1b1k35.
TA B L E 3 Cardinal temperatures for germination for seven Euphorbia species exposed to a constant temperature regime (T b : base temperature; T c : ceiling temperature; T o : optimal temperature), together with thermal times to germination for the 10th (Θ 10 ), 30th (Θ 30 ), and 50th (Θ 50 ) percentile for the germinated fraction (SE: standard errors)