Comparative study of seed germinations and phenology of population of the invasive weed Solanum rostratum Dunal in Israel

Abstract

Phenotypic variation in germination and phenological traits can enhance the invasive potential of alien weeds by promoting establishment across heterogeneous environments. Solanum rostratum Dunal (buffalobur), a prickly annual species native to Mexico, has become invasive in several regions worldwide, yet little is known about the mechanisms underlying its spread in the Mediterranean Basin. We evaluated variation among seven Israeli populations in seed morphology, germination under different temperature regimes, emergence from burial depths, and subsequent plant growth and flowering phenology. Seed traits varied significantly among populations, with heavier seeds (e.g. SL, MM) or lighter seeds (KL, KM), reflecting trade-offs between establishment potential and dispersal. Germination was generally highest at 32/38°C, but populations differed in optimal conditions; KSH and SL reached 100% germination at 22/28°C, whereas others required higher temperatures. Emergence declined with burial depth, yet the KL population maintained up to 30% emergence from 15 cm, suggesting local adaptation to soil disturbance. Phenological comparisons revealed modest variation in flowering onset, but pronounced differences in flowering duration (4.35–7.81 days), with extended flowering windows potentially enhancing pollination opportunities in this buzz-pollinated species. Collectively, these results highlight substantial intra-specific variation in key life-history traits of S. rostratum. Such variation likely facilitates persistence across diverse Israeli habitats and contributes to invasion success by combining broad environmental tolerance with population-specific adaptive strategies. Our findings emphasize the need for regionally tailored weed management approaches and advance understanding of how phenotypic differentiation supports weed invasions.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-29564-4.

Introduction

Variation in germination and phenological traits among populations of invasive weeds plays a central role in their ability to colonize diverse environments and evade control efforts^1,2. Such inter- and intra-population differences allow invasive species to synchronize germination and flowering with local resource availability, avoid peak management interventions, and exploit ecological niches inaccessible to less plastic conspecifics. Consequently, variation in germination behavior and phenological patterns constitutes a critical component of invasion success, enhancing adaptive potential and complicating the implementation of uniform control strategies.

Phenotypic differentiation within the invaded range often reflects local adaptation. In Europe, Ambrosia artemisiifolia exhibits pronounced phenotypic plasticity in vegetative and reproductive traits across altitudinal and temperature gradients³. Leiblein-Wild and Tackenberg⁴ found significant variation across 38 A. artemisiifolia populations in base germination temperature (0.6–2.7 °C), base water potential (− 1.44 to − 0.78 MPa), and seed size, all of which influence establishment under heterogeneous climatic conditions. Similarly, Malka et al.⁵ reported significant differences among Israeli populations of Parthenium hysterophorus in seed size, germination cardinal temperatures, and osmotic stress responses, indicative of population-specific adaptation in germination timing and drought resilience. Bar et al.⁶ showed divergent phenology and emergence responses of native vs. invasive Cynanchum acutum populations originating from temperate versus hyper-arid regions of Israel. Rhizomes originated from the native region sprouted more readily at 15–30 °C, while southern rhizomes had higher sprouting at 35 °C. Moreover, while northern-origin plants flowered earlier at both test sites, flowering was triggered approximately 20 days sooner at the southern site, reflecting both genetic adaptation and phenotypic plasticity to temperature regimes. These findings reinforce the notion that intra-specific trait variation, including germination behavior and flowering phenology, underpins the ecological breadth and invasion success of alien weeds, and highlights the importance of population-specific management approaches.

Solanum rostratum Dunal (buffalobur) is native to the Mexican highlands⁷. It is considered a noxious weed, as it grows aggressively following habitat disturbance and has invaded several countries worldwide, including Canada, China, Russia, and Australia⁸. In Israel, S. rostratum was first documented in the Jezreel Valley in 1953⁹. Since then, populations have been recorded in the Jordan Valley, the Golan Heights, the Hula Valley, and along the Mediterranean coastline¹⁰. Although abundant, S. rostratum is usually restricted to open, disturbed habitats such as roadsides, fallow fields, and field margins⁸. In the United States, it has been reported as an agricultural pest in Oklahoma¹¹, Nebraska, and Wyoming¹².

S. rostratum combines high drought resilience with aggressive reproductive strategies. After seed ripening, berries dry and split open, the main stem breaks near the ground, and the plant tumbles across the soil surface, scattering thousands of seeds¹³. Freshly harvested seeds are dormant, yet they may germinate under certain conditions¹⁴. A study from China showed that S. rostratum seeds are non-dormant, water-permeable, and form a short-lived persistent seed bank, with ~ 55% germinating after field cold stratification¹⁵. Beyond prolific reproduction, the species also exhibits competitive superiority in resource acquisition: in comparative studies, it outperformed co-occurring native grasses such as Leymus chinensis and Agropyron cristatum through greater nitrogen uptake and biomass production¹⁶.

In this study, we evaluated phenotypic variation among seven populations of S. rostratum collected from different regions of Israel. We assessed seed morphology (area and weight), germination responses under varying environmental conditions (temperature and burial depth), and phenological development (plant height and flowering time). These findings provide critical insights into the ecological flexibility and competitive abilities of S. rostratum, contributing to a better understanding of its invasive potential and informing future management strategies.

Materials and methods

Plant material

Seeds of S. rostratum were collected in the summer of 2023 in agriculturally cultivated fields at different locations across Israel: Ginegar (GO; 32.6543271802, 35.2488696828) Givat Yoav (GY; 32.8012266548, 35.6977272346), 34.9184483341), Kfar Masaryk (KM; 32.8851355473, 35.109928859), Kibbutz Lavi (KL; 32.7874152565, 35.4297385362), Kibbutz Sha’alvim (KSH; 31.8722352816, 34.9683375942), Ma’agan Michael (MM; 32.5472419936, and Kfar Szold (SL; 33.194514, 35.642185). The formal identification of the plant was conducted by Dr. Maor Matzrafi. As we are still working on these populations, seed specimens have not yet been deposited in a public herbarium; however, seeds are available upon reasonable request. From each population, berries were collected haphazardly from 30 to 40 plants. Fruits were dried at room temperature for 5–10 days, after which seeds were extracted, passed through a 0.8-mm sieve, cleaned, and dried on absorbent paper for 48 h at room temperature before storage under dry conditions.

Because S. rostratum seeds exhibit strong physiological dormancy¹³, all seeds used in experiments were pretreated. Dormancy was broken by immersion in H₂SO₄ (1.4 M) for 10 min, followed by rinsing under running water for 10 min. Seeds were then pre-soaked in 2.4 mM GA₃ for 24 h at room temperature in darkness.

Seed spatial parameters and weight

Fifty seeds from each population were photographed with a 3D digital microscope (Hirox^®, equipped with a CMOS high-definition sensor and MXB-050Z 50–400× zoom lens). Seed surface area was measured using the built-in image analysis software (RH 2000). For seed weight, 500 seeds per population were assessed in 10 replicates of 50 seeds, weighed on a micro-scale (Mettler-Toledo^® GmbH, WXTE).

Germination under different temperatures

To determine the optimum germination temperature, four replicates of 15 seeds each were incubated in 90-mm Petri dishes on two layers of filter paper moistened with 3 ml distilled water supplemented with 0.1% Signom^® (ADAMA Israel Agan) to prevent fungal infections. Seeds were exposed to six alternating temperature regimes: 12/18, 17/23, 22/28, 27/33, 32/38, and 35/42 °C (night/day). These regimes were chosen to simulate field conditions during the natural germination season of S. rostratum in Israel and were consistent with previous studies on the invasive Solanum elaeagnifolium¹⁷. In total, 168 Petri dishes were tested (7 populations × 4 replicates × 6 temperatures). Germination was monitored at 24-h intervals for 20 days.

Cumulative germination data were fitted with a three-parameter log-logistic model using Eq. (1):

1

where d is the upper asymptote (maximum germination), e is the inflection point corresponding to the time required for 50% germination (t₅₀), and b is the slope at the inflection point. The lower limit was constrained to zero.

Effect of burial depth on seedling emergence

Pretreated seeds were buried in plastic pots filled with Newe Ya’ar soil (57% clay, 28.2% silt, 8.1% sand, 1.63% organic matter) at depths of 0, 5, 10, and 15 cm. Fifteen seeds were evenly distributed per pot at each depth. The experiment included four replicates per population and was conducted under a 25/30°C (night/day) regime. Seedlings were recorded as emerged once cotyledons were visible, and emergence was monitored daily for 29 days.

Comparative phenology of S. rostratum

Seeds were treated for dormancy break and germinated under optimal conditions as described above. At the two-three leaf stage, seedlings were transplanted into 3-L pots filled with commercial potting mix (Tuff Marom Golan^®, supplemented with Osmocote^® slow-release fertilizer). For each population, 25 plants were used.

Phenological traits recorded included flowering onset and plant height (measured weekly for seven weeks and at the end of the study). At harvest, above-ground shoots were cut, dried at 65 °C for 48 h, and weighed. Flowering dynamics were modeled with a four-parameter Weibull function with a lag phase (W1.4) as described in Eq. (2).

2

Parameters shown in this equation are, c is the lower limit, d is the upper limit (maximum flowering), e is the inflection point corresponding to the time required for 50% flowering (t₅₀), and b is the slope at the inflection point. The lower limit was constrained to zero.

Statistical analyses

Differences among populations in seed surface area, seed weight, and phenological traits (stem elongation, plant height, above-ground biomass) were tested using one-way analysis of variance (ANOVA). When significant (p ≤ 0.05), differences among means were identified with Tukey’s Honest Significant Difference (HSD) test.

Germination data were analyzed in R (v3.3.2, i386)¹⁸ with the drc package using the functions drm() and LL.3(). Emergence data (burial depth) were analyzed with drmte() and LL.3(), and flowering time was analyzed with the W1.4() function. Model fit was assessed with root mean square deviation (RMSE). Flowering duration was calculated as the interval between the onset and completion of flowering.

Results

Seed size and average weight

Analysis of seed area across the seven populations revealed significant differences. Specifically, seeds from the MM population exhibited a significantly larger area compared to all other populations except for seeds of GY (MM − 4.54 ± 0.49 mm², GY − 4.40 ± 0.41 mm²; Fig. 1A; Table S1). No other significant differences in seed area were observed among the remaining populations.

Fig. 1.

Seed spatial parameters for seven Solanum rostratum populations: GO, GY, KL, KM, KSH, MM, and SL. (A) surface area (n = 50) and (B) seed weight (10 replicates of 50 seeds each). Statistical differences between populations were evaluated using one-way ANOVA, followed by Tukey’s HSD test at a significance level of p ≤ 0.05.

Differences in average seed weight were more pronounced. MM and SL exhibited the highest seed weights (328.89 ± 18.69 mg and 328.28 ± 6.83 mg, respectively), followed by GY and GO (310.91 ± 10.31 mg and 295.73 ± 17.05 mg; Fig. 1B; Table S1). In contrast, KL, KM, and KSH had the lowest average seed weights, with no statistically significant differences among them.

Germination under different temperatures

We investigated the effect of temperature on germination capacity in seeds from seven S. rostratum populations. All populations germinated across the tested temperature regimes (Fig. 2); however, germination at 12/18 °C did not exceed 25% in any population. In general, total germination for each population was higher at the high temperature regimes, the optimal germination temperature regime varied among populations (Table S2). In population KSH and MM, total germination has reached 100% at 22/28 °C and even at higher temperature regimes. In most populations, it appears that the 32/38 is the temperature regime were highest germination percentage was reached. The shortest time in days for germination, pronounced as GR50, was different among populations; GO and KL at 22/28°C (13.97 and 14.22, respectively), GY at 27/33 °C (12.82), SL and MM at 37/43°C (10.65 and 13.30, respectively) and KSH at 32/38°C (14.54).

Fig. 2.

Non-linear regression temperature-response model estimating the germination rate across seven S. rostratum populations. The model was fitted using Eq. (1) in "Materials and methods" section. Each data point represents the mean of four replicates (n = 4), with 15 seeds per replicate.

Emergence from different burial depths

Seedling emergence was strongly inhibited by increasing burial depth (Fig. 3). In some cases, emergence at 5 cm depth exceeded that at the soil surface (0 cm), as observed in the MM (0.9 ± 0.15 at 5 cm vs. 0.85 ± 0.11 at 0 cm) and SL (1 ± 0.17 at 5 cm vs. 0.66 ± 0.14 at 0 cm) populations (Table S3). However, emergence from 10 cm was low for both populations (0.08 ± 0.05 for MM and 0.19 ± 0.11 for SL), and nearly no seedlings emerged from 15 cm. Notably, the KL population had the highest emergence from 15 cm (0.3 ± 0.13), though this value was still relatively low. The KSH population demonstrated consistently high emergence across all depths, with rates of 1 ± 0.03, 1 ± 0.03, and 0.98 ± 0.11 at 0, 5, and 10 cm, respectively.

Fig. 3.

Non-linear regression model estimating the emergence rate across seven S. rostratum populations seeded at different burial depths (0, 5, 10, and 15 cm). Seed burial depth treatments were conducted under alternating night/day temperatures of 25 °C/30 °C. The model was fitted using Eq. (1) in "Materials and methods" section. Each data point represents the mean of four replicates (n = 4), with 15 seeds per replicate.

Phenology, development and reproductive output of different S. rostratum populations

Plant height measurements across time

Plant height increased over time in all populations, with most elongation occurring between weeks 3 and 4 (Fig. 4). While some populations showed a near-linear increase throughout the experiment, others displayed reduced elongation at later stages. For example, the MM population grew from 51.92 ± 6.89 cm in week 5 to 70.96 ± 7.13 cm in week 6, and 82.28 ± 6.72 cm in week 7. In contrast, the KSH population plateaued, with height measurements of 56.6 ± 3.77, 56.2 ± 6.3, and 57.92 ± 6.08 cm for weeks 5, 6, and 7, respectively.

Fig. 4.

Shoot elongation over seven weeks from emergence for seven S. rostratum populations (GO, GY, KL, KM, KSH, MM, and SL; n = 25 plants per population). Significant differences between weekly measurements within each population are indicated by different letters (p ≤ 0.05). Differences among populations were evaluated using one-way ANOVA followed by Tukey’s HSD test (p ≤ 0.05).

Analysis of weekly shoot elongation (Table S4) revealed that the greatest growth occurred between weeks 3 and 4, ranging from 13.08 ± 5.69 cm (GO) to 20.60 ± 2.45 cm (KSH). Between weeks 4 and 5, elongation rates declined drastically across all populations except for GO, which maintained an increase of 12.40 ± 4.79 cm. MM showed a growth spurt between weeks 5 and 6, averaging 19.04 ± 4.60 cm. Additionally, populations KL and KM experienced a second peak in growth between weeks 6 and 7, with elongation of 18.60 ± 5.84 cm and 16.88 ± 6.06 cm, respectively.

Reproductive biology

Plants from all populations eventually reached the flowering stage (Fig. 5). Five populations exhibited similar flowering patterns, with FT₅₀ (time to 50% flowering) ranging from 24.44 to 25.52 days (Table S5). KM had the earliest flowering onset (FT₅₀ = 22.34 days), while SL flowered the latest (FT₅₀ = 36.65 days). Flowering duration (from 5% to 95% flowering) was longest in KM (7.81 days) and KSH (7.26 days), and shortest in GY (4.35 days). By 49 days after transplanting, all plants from all populations had flowered.

Fig. 5.

Flowering dynamics of seven S. rostratum populations (GO, GY, KL, KM, KSH, MM, and SL) monitored over 42 days (n = 25 plants per population). Flowering progression was modeled using a four-parameter Weibull model with a lag phase to capture differences in flowering onset, rate, and duration among populations (Eq. (2) in “Materials and methods” section).

Plant productivity

Final shoot dry weight and plant height were recorded for all populations (Fig. 6). While average dry weight across populations was ~ 34 g with no statistically significant differences (Table S6), final plant height varied both within and among populations. MM produced the tallest plants (82.28 ± 6.72 cm), followed by KM (74.88 ± 9.49 cm). The lowest average heights were observed in GO (48.88 ± 6.43 cm) and GY (53.92 ± 7.34 cm) populations.

Fig. 6.

Shoot dry weight (expressed as a percentage of the control) (a) and plant height (b) of S. rostratum plants from seven populations (GO, GY, KL, KM, KSH, MM, and SL). Statistical differences between populations were evaluated using one-way ANOVA, followed by Tukey’s HSD test at a significance level of p ≤ 0.05 (n = 25 plants per population).

Discussion

Our study revealed substantial phenotypic variation among S. rostratum populations in Israel in traits spanning seed morphology, germination under contrasting environmental conditions, emergence, shoot elongation, and flowering phenology. Such variation has important implications for the species’ adaptive potential and invasion success. The interplay between trait plasticity and population differentiation we observed supports the view that invasive plants often exhibit both high environmental tolerance and adaptive divergence^19,20.

The observed differences in seed size, mass, and coat morphology between populations are consistent with earlier reports showing that seed traits strongly influence establishment potential in novel habitats²¹. Populations with larger seeds, such as GY (Table S1), may have a competitive advantage in resource-poor or stressful microsites due to greater seed reserves supporting early seedling growth²². However, smaller seeds with lower mass, as in the KL and KM populations, may disperse farther and persist longer in the soil seed bank²³, enhancing colonization capacity in disturbed landscapes. Such a trade-off aligns with Baker’s²⁴ “ideal weed” concept, which emphasizes traits that balance dispersal and establishment potential to maximize invasion success.

Germination responses varied considerably across populations and environmental treatments, reflecting both plasticity and possible local adaptation. While seeds from most populations germinated best at 32/38 °C, germination at suboptimal temperatures differed strongly (Fig. 2; Table S2). The ability of some populations to maintain high germination rates under suboptimal temperature regimes is consistent with observations in other invasive Solanum species^17,25 and supports the idea that broad germination niche breadth facilitates colonization of heterogeneous habitats²⁶. The fact that certain populations performed better under specific environmental regimes suggests that S. rostratum may be undergoing local adaptation, a pattern often seen in invasive plants with multiple introduction events and spatially variable selection pressures²⁷.

Emergence experiments further emphasize the adaptive flexibility of S. rostratum. For example, while emergence generally declined with increasing burial depth, the KL population still exhibited 30% emergence from 15 cm burial. For the same burial depth, all other populations dropped below 10%, apart from KSH that showed 12% emergence (Fig. 3; Table S3). This tolerance to deeper burial is likely advantageous in arid and semi-arid habitats, where seeds may be incorporated into the soil by tillage. As seed area and weight in this population were relatively low compared to other populations, it is unlikely that the higher emergence potential is related to larger seed size or mass. However, if the KL population historically experienced management practices or disturbances that buried seeds (e.g., deep tillage), selection could have favored genotypes capable of germinating and emerging from greater depths. While direct evidence of selection for deep-emergence in weed populations under intensive tillage is lacking, tillage clearly alters burial depth and emergence abilities²⁸, which creates the potential selective environment for adaptation. In contrast, populations with emergence from shallow depths but higher emergence percentage may be favored in more competitive habitats. For instance, the KSH populations showed high emergence rate at 0 (100%), 5 (100%) and 10 (98%) cm depth. These divergent strategies illustrate how intraspecific variation can generate ecological breadth, a feature strongly linked to invasion success²⁹.

Differences in shoot elongation and flowering timing among populations may also contribute to invasion success by enabling temporal niche partitioning and facilitating escape from environmental stress or competition. Early-flowering populations could complete reproduction before the onset of summer drought, while later-flowering populations may exploit extended resource availability in wetter habitats. Such phenological variation has been recognized as a key mechanism underlying invasion in a heterogeneous landscape³⁰. Our research showed little variation in the onset of flowering, except for the SL population, which flowered much later (38.65 days). However, greater variation was recorded in the overall flowering period within each population (from 4.35 to 7.81 days; Table S5). A prolonged and staggered flowering period can increase opportunities for successful pollination by extending temporal overlap with potential pollinators^31,32. This is particularly relevant for S. rostratum, which relies on buzz-pollination by a subset of bee species; in such systems, extended flowering windows may enhance both the likelihood of attracting effective pollinators and the probability of cross-pollination among plants³³. Thus, phenological variation among and within populations likely represents an important component of the species’ invasion strategy, increasing reproductive success under diverse environmental conditions.

Overall, our findings align with invasion biology theory predicting that successful invaders combine general-purpose genotypes with the capacity for local adaptation^24,34. S. rostratum populations in Israel display trait combinations that could confer both high colonization potential and persistence across contrasting habitats. The coexistence of plasticity and population differentiation suggests that management strategies should be tailored to specific regions, accounting for both environmental conditions and the dominant trait profiles of local populations. From a broader perspective, our results emphasize that phenotypic variation is not merely a by-product of invasion but may be a central driver of invasion dynamics, shaping both spread rates and ecological impacts.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

J.A.-N. and M.M. designed the experiments. J.A.-N. conducted the experiments. J.A.-N. and M.M. analyzed the data. J.A.-N. and M.M. wrote and reviewed the manuscript. All authors have read and approved the manuscript.

Funding

This work was supported by the chief scientist of the Israeli Ministry of Agriculture and Food Security, Grant Number 20-02-0195.

Data availability

Data will be made available upon request, for more details please contact Dr. Maor Matzrafi.

Declarations

Competing interests

The authors declare no competing interests.