Introduction

Phenotypic plasticity refers to the alteration of an individual’s morphology, physiology, development, and/or life history in response to environmental heterogeneity1,2. For plant species, it is the means by which an individual adjusts to environmental changes and optimizes resource acquisition. Likely to have ecological, horticultural, evolutionary, and fitness implications3,4,5, phenotypic plasticity plays a key role in the ecological expansion of invasive species6,7. In horticultural crops, it can be exploited to improve productivity8 and is an important consideration for breeders wrestling with genetic correlation among important morphological traits9.

Historically, light, water, nutrient, temperature, and wind were the common environmental factors for which phenotypic plasticity was evaluated in plant species4,10,11,12,13; and the degree of plasticity was consistently found to vary, depending on both the species and the traits under consideration4,13,14,15,16,17. For instance, Pistacia lentiscus L., and P. terebinthus L. were more plastic than Quercus coccifera L. and Q. faginea Lam. in their response to irradiance, whereas Quercus spp. were more plastic than Pistacia spp. in their response to water availability18. Likewise, it has been suggested that light-demanding species exhibit greater plasticity in growth, morphology, and physiological traits than shade-tolerant species19; and in liana species (e.g. Rosa longicuspis Bertol., Embelia procumbens Hemsl.), biomass and growth traits were found to be more plastic in response to altered light than morphological traits20. In Populus deltoides Bartr. ex Marsh., altered nutrient availability was associated with greater plasticity in shoot-specific traits than in leaf-specific traits, whereas for the same species the relative magnitude of the plasticity in the two sets of traits was similar in response to changes in water availability21. Likewise, a recent study comparing seven domesticated crops to their respective wild relatives highlighted a differential plasticity pattern to nutrient and water availability for a suite of traits, including maximum height, total leaf area, plant-level photosynthetic rate, and growth performance traits14. In another study, altered nutrient availability induced plasticity in both leaf number and total leaf area in Pelargonium australe J. Jacq.22. Such findings suggest that factors like light, water, and nutrient availability do not trigger a predictable plasticity pattern across all plant species. In general, understanding such species-specific plasticity patterns is necessary to inform production techniques and optimize trait selection. Thus, for the miracle plant Synsepalum dulcificum (Schumach & Thonn.) Daniell, an endangered orphan fruit species in which fast growth, early fruiting, and higher yields are desired, basic investigation of these issues is needed.

Synsepalum dulcificum is an evergreen West African native species belonging to the Sapotaceae family. This tropical species is the only known natural source of “miraculin”, a sweetening glycoprotein23. The fruit (miracle berry) has been reported as a promising economic alternative to synthetic sugar24,25 and is recommended for diabetic patients26. Currently, S. dulcificum is utilised in cosmetics and food, though it is most extensively used by the pharmaceutical industry27. Interest in the species has grown tremendously in the last decade, with one kilogram of pure powder of the fruit fetching prices of up to $2,500 (http://miraclefruitfarm.com/shop/). Developing early fruiting and highly productive ecotypes for large-scale production will be crucial for meeting this increasing demand. To date, however, S. dulcificum is a notoriously slow-growing species28, often encountered in home gardens and less frequently in farms29. Empirical observations have suggested that seedlings growing in home, tree-based gardens, are more vigorous than those exposed to full sunlight on open farms. It is therefore important to know whether or not irradiance reduction is beneficial to the species and, if so, to what extent.

Wilkie et al.30 indicated environmental induction as an appropriate means of inducing flowering in various tropical horticultural crops. Meilan31 suggested that photoperiod, nutrient, availability, and water were the main environmental factors that trigger flowering in woody species. For example, light exposure consistently accelerated flowering in the blueberry (Vaccinium corymbosum L.)32, while in the white birch (Betula platyphylla Suk.) a balanced NPK fertilization was reported not only to promote the transition from the juvenile stage to maturity, but also to increase flower production33. In addition, in the case of S. dulcificum, irrigation and inorganic fertilization have previously been reported to accelerate and increase flowering34, though whether or not the source of the fertilization mattered for growth and development remains unknown. While a number of studies dealing with organic fertilization in tree species report beneficial effects on soil conditions, effects on the trees’ intrinsic performances (e.g. survival, growth, and reproduction) are less clear35,36. In S. dulcificum, no study has yet evaluated the response of seedlings to light exposure and organic nutrient supply. Nutrient availability and light have also been reported as important factors influencing biomass allocation, thus playing a key role in the timing of growth and reproduction37; however, biomass allocation pattern in S. dulcifucm has also yet to be investigated.

The objective of this study is to evaluate the response of S. dulcificum to changes in light exposure and organic nutrient supply (i.e. compost application) in an effort specifically to identify the potential of these factors to induce rapid growth and early fruiting in the species. Since S. dulcificum is a tropical species, we hypothesize that the plasticity of its phenotype to light and compost application will be enough to realize significant gains in both vegetative growth and early flowering.

Results

Effect of light exposure and compost application on seedling survival

The results indicate that shading significantly improves seedling survival in S. dulcificum (Fig. 1, p = 0.002). While no main effect of compost application on survival was detected (p > 0.05), we did observe a significant interaction between light exposure and compost application (Fig. 1, p = 0.002), such that a significant effect of compost application was detected under shade conditions. Likewise, for the same compost dosage, survival was greater in shaded seedlings than in full sun-exposed seedlings.

Figure 1
figure 1

Effect of light exposure and compost application on seedling survival in Synsepalum dulcificum. Barplots (Treatments) with the same letter are not statistically different at 5% (Least Significant Difference post-hoc test). n = 20 seedlings and error bars represent standard error.

Full size image

Vegetative growth in response to light exposure and compost application

Shading significantly increased nearly all measured aspects of seedling growth, including stem diameter (Fig. 2a, p < 0.001), plant height (Fig. 2b, p < 0.001), branching (Fig. 2c, p = 0.03), leaf area (Fig. 2e, p < 0.001), and specific leaf area (Fig. 2f, p < 0.001); but there was no significant effect of light exposure on leaf production (Fig. 2d, p > 0.05). Likewise, compost application positively affected stem diameter (Fig. 2a, p = 0.006), plant height (Fig. 2b, p < 0.001), branching (Fig. 2c, p = 0.002), and specific leaf area (Fig. 2f, p = 0.03); however, increased dosage did not induce a significant increased effect on those growth parameters (p > 0.05). Contrary to light exposure, compost application did not affect leaf area (Fig. 2e, p = 0.19), though it did affect leaf production (Fig. 2d, p = 0.003). Here also, an increase in compost dosage did not translate into significantly higher leaf production (p = 0.87). The interaction between light exposure and compost application was significant for all parameters except leaf area (Fig. 2). Overall, the beneficial effect of compost application on vegetative growth traits was more prominent under shade versus full sunlight.

Figure 2
figure 2

Vegetative growth of Synsepalum dulcificum in response to light exposure and compost application. (a) Diameter growth. (b) Height growth. (c) Branches production. (d) Leaf production. (e) Leaf area. (f) Specific leaf area. Barplots (Treatments) with the same letter are not statistically different at 5% (Least Significant Difference post-hoc test). n = 20 seedlings and error bars represent standard error.

Full size image

Reproductive growth in response to light exposure and compost application

Table 1 presents the percentages of budding, flowering, and fruit-bearing seedlings within each of the six treatments. Shading significantly enhanced budding (p < 0.001), flowering (p = 0.006), and fruiting (p = 0.02). Likewise, compost application also significantly improved budding (p < 0.002), flowering (p = 0.02), and fruiting (p = 0.04). However, increasing compost application beyond 4.5 g per plant had no significant effect on plant phenological transition (p > 0.05). At the end of the experiment (17 months days after transplanting), only shaded and fertilized seedlings bore fruit (p < 0.01).

Table 1 Effect of light exposure and compost application on budding, flowering, and fruit bearing in Synsepalum dulcificum seedlings.
Full size table

The times to first budding, flowering, and fruiting are presented in Table 2. While compost application was found to accelerate the transition to reproductive growth, no significant difference was observed in the times to first budding and flowering for plants that received different compost doses (p > 0.05). In contrast, the time to fruiting was significantly lower in plants that received 4.5 g compost compared with those fertilized with 9.0 g compost. Overall, treated plants (shaded and fertilized) entered reproductive phase at an average age of 23 months (19 months for early-maturing individuals); and fruit-bearing was observed at an average age of 26 months (24 months for early-maturing individuals).

Table 2 Effect of light exposure and compost application on the timing of reproduction in Synsepalum dulcificum seedlings.
Full size table

Biomass allocation in response to light exposure and compost application

The observed pattern of biomass partitioning within the S. dulcificum seedlings indicates a significant effect of light on all parameters (Tables 3 and 4). Leaf mass fraction (LMF) (p = 0.007) and stem mass fraction (SMF) (p = 0.01) were significantly higher in shaded plants than in those grown in full sunlight. In contrast, root mass fraction (RMF) (p = 0.001) and the root-to-shoot ratio (R/S) (p < 0.001) were significantly higher in seedlings grown in full sunlight. None of the tested biomass allocation parameters was significantly affected by compost application (p > 0.05). Similarly, no significant interaction between light exposure and compost application was detected for any of the biomass allocation parameters (RMF, SMF, LMF, or R/S) (p > 0.05) (Table 4).

Table 3 Biomass allocation in Synsepalum dulcificum seedling growing under shade and full sunlight and different compost application doses.
Full size table
Table 4 ANOVA results for each biomass allocation parameter (LMF, SMF, RMF, and R/S).
Full size table

Phenotypic plasticity in S. dulcificum

As reported in Table 5, phenotypic plasticity indices (PPI) in S. dulcificum in response to light exposure and nutrient availability (compost application) ranged from 0.03 to 0.72. Of the twelve traits measured, ten were found to respond to light exposure, whereas seven were responsive to nutrient availability. The variation intensity differed greatly among quantitative traits (p = 0.02). The top three varying quantitative traits to light exposure included plant height, leaf area, and stem diameter, whereas the three most varying quantitative traits to nutrient availability were plant height, number of leaves, and stem diameter. Overall, S. dulcificum showed a greater plasticity to changes in light intensity than to nutrient availability. From a functional point of view, S. dulcificum exhibited allocational plasticity only to light exposure; and in both morphological and fitness functional groups, the mean PPI was higher to light exposure than to nutrient availability (Fig. 3). The light-induced phenotypic plasticity also induced variation in leaf color, with shaded seedlings being greener than those grown under full sunlight. An illustration of the response of leaf colour to varying light exposure is shown in Supplementary Fig. S1.

Table 5 Phenotypic plasticity index (based on Valladares et al.81) in Synsepalum dulcificum measured traits for light exposure and nutrient availability factors.
Full size table
Figure 3
figure 3

Mean phenotypic plasticity indices, by functional groups, for Synsepalum dulcificum.

Full size image

Discussion

Our study highlighted the beneficial effect of light exposure and compost application in S. dulcificum, a tropical tree of the Sapotaceae family. Such investigations are scanty in that family7, particularly in the Synsepalum genus, although previous studies have documented the growth of perennial species in response to shading and fertilization23,38.

Synsepalum dulcificum exhibited a relatively higher survival rate under shade conditions, as well as with ample water and nutrient supply. The beneficial effect of continuous water supply on the tree was previously known34. Our findings indicate that moderate shade is favourable for the survival of transplanted seedlings. The survival and growth rates obtained under shade were indicative of the sensitivity of S. dulcificum to heat. This observation partly explains the scarcity of the species, especially seedlings, in open field (e.g. farms, fallows). According to Grubb39, a shortage of light in plant species may compromise the survival of plants. However, in S. dulcificum the survival of seedlings was improved with a substantial reduction of sunlight. Higher juvenile survival rates under shade conditions were reported to characterize either intermediate- shade tolerant or shade tolerant species40. A similar phenomenon was also reported in Ilex aquifolium L.41 (another shrub species from Northern Africa, Western Asia, and Europe reported to exhibit shade tolerance).

Our findings indicate that shade also improved reproductive performance by reducing the time to first fruiting and increasing the proportion of seedlings entering reproductive phase. This finding is new in S. dulcificum and complements existing knowledge on tree species. Previous studies that assessed the effect of shading on tree species focused on morphological growth, and biomass allocation pattern42,43,44,45,46 or the non-beneficial effect of shading47. This non-beneficial effect of shading was illustrated in Plukenetia volubilis L.48, Phlox drummondii Hook.49 and Opuntia humifusa Raf.50 with a delay in flowering, and in Olea europaea L. that exhibited a reduction of individual fruit weight and oil concentration51.

In some species [e.g. Nothofagus glauca (Phil.) Krasser (hualo), Picea sitchensis (Bong.) Carr., Thuja plicata, Donn ex D. Donn], the response to shading varied from one trait to the other44,52. But our findings revealed a consistent beneficial effect of shade on most vegetative and growth traits. For instance, we found that height in shaded seedlings was at least twofold higher than those in sun-exposed seedlings. In addition, shaded seedlings presented larger leaf blade areas than sun-exposed seedlings. Overall, we found that S. dulcificum is an intermediate shade tolerant species. Synsepalum dulcificum successfully survived, grew and reproduced under moderate shading. According to the definition of Martínez-García et al.47, S. dulcificum can be deemed shade tolerant.

Our study also demonstrated a positive effect of compost application on the reproductive performance of the tree, implying that organic fertilization is as good as inorganic fertilization34 in accelerating bud, flower and fruit induction. This offers farmers the possibility to utilise locally available resources for the production of healthy fruits for a growing organic fruit market53. Compost application also consistently improved vegetative growth in S. dulcificum. This indicates that the species may be more responsive to fertilization than other fruit tree species such as Uapaca kirkiana (Muell. Arg.), and Sclerocarya birrea (Hochst.)35 where the effect of compost supply was not conclusive.

We observed that shoot biomass was higher in shaded seedlings than in seedlings that grew under full sunlight. Similarly, in non-fertilized seedlings, more biomass was allocated to roots, probably to better explore the soil for nutrients34. Biomass partitioning in plants is species-specific and dependent on the environment54. For instance, woody species increased their leaf mass fraction in response to waterlogging, while herbaceous species did not. According to the balanced growth hypothesis55, plants allocate their biomass in an optimal pattern if the required above- and below-ground resources limits growth to an equal extent. As such, resource allocation will then favour above-ground organs when light is limiting, whereas below-ground organs will be favoured when nutrient or water are the limiting factors.

Valladares and Niinemets56 indicated that in intermediate shade-tolerant species, stem elongation was often found to be the most plastic trait. Our observations also revealed seedling height as the most plastic trait in S. dulcificum. This supports the hypothesis that S. dulcificum is an intermediate shade-tolerant species. We observed that morphological traits in S. dulcificum were more plastic than fitness and biomass allocation traits irrespective of the factor into play. This plasticity pattern differed from what was reported on other species in the literature. For example, in Quercus ilex and Phillyrea latifolia L., the most plastic traits to light were the physiological traits57. Similarly, physiological traits were found more plastic than morphological traits in Betula alleghaniensis Britt58.

Our findings illustrate phenotypic plasticity of S. dulcificum to light exposure and to nutrient availability with potential implications in terms of evolution, adaptation and development in new environments59, especially in the context of climate change. The adaptive plasticity of S. dulcificum under reduced light exposure and nutrient-rich conditions highlights the potential of the species to successfully colonize new environments. This adaptive plasticity might partly explain the current distribution of the species. Indeed, the species naturally occurs in six African countries, of which four are humid forest countries (Congo, Cameroon, Nigeria, Ghana) and two are low forest cover countries (Benin and Togo)60,61. Synsepalum dulcifucum was recorded among the most common understorey species in the Littoral Congo forest in Congo (a humid forest country)62. In Benin and Togo (low forest cover countries), the species was rather characterized by a rarity index of 0.9863.

Synsepalum dulcificum is commercially important and has the potential to contribute to income generation, empowerment of smallholder farmers29, and overall economic growth of occurrence countries. For these reasons, the potential of the tree to withstand climate change is worthy of consideration. Climate change is undoubtedly a major threat to terrestrial ecosystems64, and it is postulated that in response to changing environmental conditions plants will either migrate to maintain their adaptive optimum, or respond through phenotypic plasticity65 and or adaptive evolution66. While phenotypic plasticity was often viewed as an immediate response to very rapid environmental changes, adaptive evolution was considered important over the long term67. Synsepalum dulcificum occurs naturally in West Africa and it is predicted that temperatures in that region will increase by up to 1 °C by 206667, and conditions will be drier with the retraction of moist and wet zones68, as well as longer and more frequent heat waves69. Under these projections, the phenotypic plasticity to light exposure observed in the current experiment may become maladaptive due to the increased heat stress that will constrain the fitness of the species. It has been suggested that phenotypic plasticity holds the potential to help plant species to adapt to climate change only when the plasticity is adaptive under the newly experienced environment70,71. Previous findings in S. dulcificum indicated the crucial importance of water availability for growth and fitness in the species34. However, since it is predicted that the species will undergo higher water stress in its natural environment, we suggest that climate change will adversely affect the future distribution of S. dulcificum in at least West Africa, and more intensely in Benin where the species is currently distributed in the Guineo-Congolian region only (an area predicted to shift from wet to semi-arid conditions)68. The consequence is that adaptive evolution would be the most reliable mechanism for the species to withstand future climate change, while its success will hinge on the existence of an adequate level of genetic variation in the species. Moreover, adaptive evolution will need to be rapid enough to respond to climate change66,72. Therefore, the future development of S. dulcificum will require a change in the current cultivation system, from plain field production to a more appropriate future-climate resilient system. In this regard, our findings imply some agronomic and horticultural pathways including possible development of agroforestry systems, commercial propagation nursery establishment, and greenhouse orchard promotion.

Agroforestry systems are viewed as a climate-smart agriculture practice that enhances food security, while serving adaptation and mitigation objectives73, and are particularly suited to a context of increasing pressure on the land. For species such as Theobroma cacao L. and Coffea arabica L. that have been mostly cultivated as agroforests74,75, the use of intermediate shade to optimize the incomes of smallholder farmers and biodiversity services are now recommended76,77. This is relevant to S. dulcificum, which exhibits a high growth rate and reproductive performance under moderate shading and organic fertilization. As in T. cacao75, S. dulcificum is able to survive under either “service” legume shade trees (e.g. Gliricidia sepium (Jacq.) Walp) or “productive” shade tree crops (e.g. Musa spp, timber or other perennial tree crop species). Synsepalum dulcificum was known as a slow-growing species, in which growth can be improved with adequate water supply28,34. Our findings suggest that moderate shading and compost application are also favourable for the growth conditions of the species. All these characteristics combined together constitute a promising package to promote the greenhouse orchard development in the species, especially in a context of the emerging organic greenhouse production. Indeed some of the major constraints affecting profitability in fruit tree greenhouse promotion includes costs required to maintain light sources, ensure adequate light distribution, and intensive fertilization78. For a number of fruit tree species (e.g. Pyrus communis auct., Acca selowiana (O.Berg) Burret), shading created by either the competing individuals or the greenhouse structure often negatively affected growth79. For such species, external lighting source is vital for ensuring production and profit. In the case of S. dulcificum, there is minimal competition between individuals, and the species can tolerate moderate shading while requesting a limited amount of organic nutrient for optimal growth. Therefore, there is potential to cultivate the species in marginal areas of its distribution, provided a greenhouse technology exists.

Overall, we conclude that seedlings of S. dulcificum responded positively to shading and compost application. Shade and compost supply consistently improved vegetative growth, whereas the interaction between shade and compost supply accelerated transition to reproductive phase. Shaded and 4.5 g compost-fertilized seedlings started fruiting at 24-month-old. Biomass allocation in the species supported the balanced growth hypothesis and morphological traits exhibited higher plasticity. Shade induced higher plasticity than nutrient availability in the species. In both cases, plant height and stem diameter were among the most plastic traits. Based on the plasticity trends observed, we suggest S. dulcificum is an intermediate shade tolerant species and recommend the use of 4.5 g per seedling every two months, as the basic compost application dose.

Methods

Plant materials

In March 2015, mature and ripe fruits of S. dulcificum were hand–harvested from a single tree in the locality of Sèhouè, Benin (06°55′09.5″N, 002°16′23.3″E), and processed by removing their red outer skins and seeds. A total of 250 seeds were thus obtained and subsequently germinated in black polystyrene nursery bags (754.2 ml) filled with sterilized sowing substrate (see below) and using one seed per bag. The physico-chemical characteristics of the soil used as sowing substrate were as follows: pH(KCl) = 5.48, pH(H2O) = 5.88, silt = 25.75%, clay = 12.27%, sand = 61.98%, organic carbon = 1.03%, N = 0.06%, Mg = 2.37 meq/100 g, Ca = 0.63 meq/100 g, phosphorus = 2.08 meq/100 g, and assimilable phosphorus = 23.06 ppm. The germination process was conducted in the Laboratory of Genetics, Horticulture and Seed Science at the University of Abomey-Calavi, Benin (06°25′00.8″N, 002°20′24.5″E). Seeds germinated within 21–45 days after sowing, and the seedlings were grown in a nursery for 12 months before being used for this experiment. Seedlings were watered once a day to ensure an adequate supply of water to the growing medium.

Experimental system

Vigorous, similarly-sized 12-month-old seedlings were selected and transplanted into black polystyrene pots (25 cm diameter; 15.26 L) filled with the same substrate used for germination, again with one seedling per pot. The experiment started in May 2016 after all seedlings were planted in their new containers.

Over the following 15-month period (May 2016–August 2017), we used a factorial split-plot to evaluate the effects of light exposure and compost application on the survival, vegetative growth, reproductive performance, biomass allocation, and plasticity of the transplanted S. dulcificum seedlings. Light exposure was the main plot factor, and seedlings were subjected to either full sunlight (Sun) or moderate shading (Shade). Compost application was the subplot factor with three levels: 0 g, 4.5 g, and 9.0 g of compost/seedling. To the best of our knowledge, there has been no study on the effect of organic nutrient supply (i.e. compost) on the performance of S. dulcificum. Furthermore, there has been no reliable study of the effect of organic fertilization in the Sapotaceae more broadly. Because there was no previous study that could be used as a reference for this study, we used compost dosages that would allow us to examine the plasticity of S. dulcificum in response to varying nutrient availability. Doses of 0 g, 4.5 g, and 9.0 g were chosen to represent a gradient of nutrient availability and were respectively defined as low, medium and high nutrient availability.

Each of the six treatment combinations was replicated three times, with each replicate consisting of 6–7 seedlings (i.e. 20 seedlings per treatment). Full sunlight represented the conditions of an open field, while the moderate shading simulated the light conditions in a tree-based system. The moderate shade was obtained using a shade-house built at 3.1 m height with an aluminium roof. The compost, applied per pot every two months from May 2016 to January 2017, was based on poultry organic manure with the following properties: pH(H2O) = 6.9, nitrogen = 0.77%, phosphorus = 0.2%, potassium = 0.12, organic matter = 16.46%, organic carbon = 8.23%, magnesium = 0.42%, Calcium = 1.16%, and C/N ratio = 11. In total, five separate compost applications were made during the course of the study.

The climatic data during the experiment (May 2016–August 2017) are presented in Table S1. During the experiment, the monthly average temperature was 27.44 °C, relative humidity was 83.53%, total rainfall was 117.6 mm, and total solar radiation was 450 Mj/m2. These data were obtained from a weather station installed by the Trans-African Hydro-Meteorological Observatory (TAHMO) on the experimental site at the University of Abomey-Calavi and from the meteorological station of the International Institute of Tropical Agriculture (IITA), Abomey-Calavi, Benin, located less than 1 km from the experimental site.

Measurements

We measured survival rate, growth (vegetative and reproductive), and biomass allocation in the treated seedlings. For survival rate, we counted the number of living seedlings in each replicate at the end of the experiment. For vegetative growth parameters, stem diameter, plant height, number of branches, and number of leaves were counted at the beginning and at the end of the experiment. Stem diameter was measured at the soil surface using a digital vernier caliper of 0.01 mm precision, and plant height was measured with a ruler from the soil surface to the tallest seedling point. For each seedling, we scanned three differently sized leaves (large, medium, and small) and transferred the images into the software Mesurim Pro Version 3.4 (Académie d’Amiens, Amiens, France) for leaf area (cm2) determination. The final leaf area for each seedling was the average value from the three harvested leaves80. The scanned leaves were afterwards oven-dried at 65 °C and then weighed to the nearest 0.01 g using a digital balance to obtain leaf dry mass. Specific leaf area (SLA) was afterwards calculated for each seedling as the leaf area divided by leaf dry mass.

For reproductive growth, we monitored the development of each seedling daily throughout the duration of the experiment. The number of plants bearing buds, flowers, and fruits, as well as the dates of first budding, flowering, and fruiting (where applicable) were noted. Proportions of budding, flowering, and fruiting seedlings were determined as well as the times to first budding, flowering, and fruiting.

At the end of the experiment, we selected three plants per treatment (one per replicate) for biomass allocation determination. Each selected plant was labelled and partitioned into leaves, stem, and roots; and each portion was oven-dried at 65 °C. We determined for each plant the root mass fraction (RMF = root dry mass/total plant dry mass), the stem mass fraction (SMF = stem dry mass/total plant dry mass), the leaf dry mass (LMF = leaf dry mass/total plant dry mass), and the root-to-shoot ratio [R/S = RMF/(LMF + SMF)].

Statistical analysis

Growth rate, determined as the difference between measurements of growth parameters at the beginning and at the end of experiment, was used as a dependent variable. The effects of light exposure, compost application, and their interaction on seedling survival and on the proportion of seedlings bearing buds, flowers, and fruits were tested using generalized linear models fitted with a binomial/quasi-binomial error structure to account for over-dispersion. We used a two-factor ANOVA to analyse the effect of the interaction between light exposure and compost application on stem diameter, plant height, and biomass allocation (LMF, STM, RMF, and S/R). When compost application or the interaction between light exposure and compost application was significant, we used a contrast analysis to depict the effect of the increase in compost application. Based on significant factors, we determined the phenotypic plasticity index (PPI) per trait following the formula given by Valladares et al.81:

$${\rm{PPI}}=({{\rm{X}}}_{{\rm{\max }}}-{{\rm{X}}}_{{\rm{\min }}})/{{\rm{X}}}_{{\rm{\max }}}$$

with Xmax and Xmin corresponding respectively to the highest and the lowest values among the mean values of each factor’s modalities. PPI theoretically ranged from 0 (no plasticity of the trait to the considered factor/no response variation to the environmental gradient) to 1 (extreme plasticity and high sensitivity to the environmental factors). We used a one-way ANOVA or a t-test (where appropriate) to compare the plasticity among measured traits and functional groups (fitness, morphology, and biomass allocation) to the tested factors (light exposure and nutrient availability). All analyses were performed in R version 3.5.082.