The plant sample collection site is located in Jinchuan peatland (42°20′56′′N, 126°22′51′′E) of Longwan National Nature Reserve in Northeast China. The climate of the study site is temperate continental monsoon climate. The average annual temperature is 4.1°C, the average annual precipitation is 704.2 mm, which mostly concentrated in July and August (Zhang et al., 2016). We randomly selected Lilium concolor var. megalanthum plants in Jinchuan peatland and labeled 200 male plants and 200 hermaphrodite plants as experimental bulb materials in 2020. After the growth of the above-ground part of the plant is completed, the bulbs are excavated and brought back to the laboratory. The bulbs were stored in a refrigerator at 4°C. The bulbs of Lilium concolor var. megalanthum with good growth and no pests and diseases were selected (from male plants and hermaphrodite plants respectively) and weighed. Bulbs were soaked in carbendazim wettable powder diluent (1: 500) with active ingredient content of 50% for 5 min, disinfected and dried for pot planting. The soil in the pot is turfy soil mixed with farmland soil (volume ratio of 1: 1). The bulb was buried upward in the soil substrate, and the top of the bulb was buried 5 cm below the soil surface and slightly compacted to make the bulb in close contact with the soil. The bulbs overwinter and dormancy in the ground, then sprout and gradually emerge into seedlings in 2021.

This experiment was conducted in 2021. We measured soil nutrient status in different habitats ( Table S1 ). Combined with the soil matrix configuration ratio, it was calculated that the application of nitrogen (1.8 g/per pot), phosphorus (120 mg/per pot), and potassium (300 mg/per pot) fertilizers to the Lilium concolor var. megalanthum plants derived from the hermaphrodite flower bulbs made the soil bottom value level in the basin equal to and lower (CK: no fertilizer application) than the soil bottom value level in the field habitat. The application of fertilizer (3.6 g/per pot), phosphorus (240 mg/per pot) and potassium (600 mg/per pot) to the lily plants from the bulbs of male flowers made the soil bottom value in the pot higher than and equal to (CK: nitrogen: 1.8 g/per pot; phosphorus: 120 mg/per pot; potassium: 300 mg/per pot) the soil base value in the field habitat. Single (N, P, K) and combined (NP, NK, PN, NPK) fertilization methods were used. Urea was selected as nitrogen fertilizer, sodium dihydrogen phosphate as phosphorus fertilizer and potassium sulfate as potassium fertilizer. There were 16 treatments for male flower bulbs and hermaphrodite flower bulbs, and 10 plants for each treatment (single plant in single pot). Bulbs of about 2 g were selected in different treatment combinations ( Table S2 ). The bulbs from hermaphrodite plants were larger than those from male plants, but there was no significant difference in bulbs between different treatments.

Fertilization was carried out at the seedling stage (seedling height 20 cm) of Lilium concolor var. megalanthum. Nitrogen, phosphorus and potassium fertilizers were applied twice before flowering, with the interval of 10 days. The fertilization method was to evenly apply the dilute solution to the root of the plant. The experiment was carried out in Longwan Ecological Experimental Station of Key Laboratory of Wetland Ecology and Vegetation Restoration. The climatic conditions of the site were the same as those of the plant sample collection site (about 500 m away).

The growth and reproduction indexes were measured at the flowering stage. The number of flowers per plant and the sex of flowering plants were recorded, and the biomass of plants was measured after the growth of the aboveground part finished. All plants bloomed one flower in 2021. Sexual fecundity index = flower dry weight/total biomass, asexual fecundity index = bulb dry weight/total biomass. The plant height of single plant was used as an index to measure the plant size. To avoid damaging sampling of bulbs, bulb biomass (dry weight) was calculated by linear fitting function using field survey sampling data. Thirty plants of two sex phenotypes were randomly selected from Jinchuan peatland in mid-July 2020 (full flowering stage). The bulbs were placed in an oven at 105°C for 0.5 h, then dried at 80°C until constant weight. The fresh weight and dry weight (biomass) of bulbs were weighed using a balance (accuracy is 0.001 g). The obtained data are fitted by variables, and the fitting functions are as follows: The regression relationship between dry weight and fresh weight of hermaphrodite flower bulbs is: y=0.3379x-0.0661, R² = 0.91, P<0.001; the regression relationship between dry weight and fresh weight of male flower bulbs is: y=0.271x+0.0241, R² = 0.87, P<0.001. y is dry weight, x is fresh weight.

The independent sample T test was used to analyze the difference in size of bulbs of Lilium concolor var. megalanthum from hermaphrodite flowers and male flowers. One-way analysis of variance and Duncan multiple comparison test (data conforming to normal distribution and homogeneity of variance) or rank sum test in non-parametric test and Kruskal-Wallis method (data still did not conform to normal distribution and homogeneity of variance after data conversion) were used to test whether there were significant differences in the growth and reproduction indexes of Lilium concolor var. megalanthum among the different nutrient treatments. The significance was set at 0.05 (α = 0.05). All statistical analyses were performed using SPSS (25.0). All data were plotted using Origin 9.2 software. The data are expressed as mean ± standard error.

Under the conditions that the nutrient level is equal to or lower than that of natural habitats, the plant size of most plants (91.25%) in 2021 is less than that in 2020 ( Table 5 ), but the bulbs of most plant (96.25%) are larger than the previous year ( Table S3 ). Among them, 60% of the plants continued to produce hermaphrodite flowers, 11.25% of the plants produce male flowers, and 28.75% of the plants did not bloom ( Table 5 ). In the treatments of single nutrient element, when the content of phosphorus, potassium, and nitrogen and phosphorus is low, the proportion of plants not flowering or produce male flowers is higher. In the treatments of compound nutrient addition, plants are more inclined to produce hermaphrodite flowers. In 2022, the morphological size and bulb size of all plants were larger than those in 2021, and all plants produced hermaphrodite flowers.

Under the conditions that the nutrient level is equal to or higher than that of the natural habitat, only 36.25% of the plants with male flowers in 2020 have increased in shape and size in 2021 compared with the previous year ( Table 6 ), but the bulbs of all plants have increased compared with the previous year ( Table S3 ). Among them, 67.5% of the plants produce hermaphrodite flowers, 10% of the plants continued to produce male flowers, and 22.5% of the plants did not bloom ( Table 6 ). Under higher nutrient addition treatment, plants are more inclined to produce hermaphrodite flowers. The promotion effect of single nutrient element addition was more significant than that of compound nutrient addition, and no plants produce male flowers. In 2022, the morphological size and bulb size of all plants increased compared with 2021. In the nitrogen and potassium addition treatment group, only one Lilium concolor var. megalanthum continued to produce male flowers, and all the other plants produce hermaphrodite flowers ( Table 6 ).

The phenotypic plasticity of plants reflects their adaptability to the environment (Solé-Medina et al., 2022). Lilium concolor var. megalanthum has a variety of reproductive patterns, so its phenotypic plasticity is more diversified than other plants. The plant height of Lilium concolor var. megalanthum show strong morphological plasticity. In the soil with low nitrogen and phosphorus content, the plant height of Lilium concolor var. megalanthum derived from bulbs of plants with hermaphrodite flower is significantly reduced. When nutrient resources are constrained, Lilium concolor var. megalanthum allocates more resources to the leaves to improve the efficiency of photosynthesis and meet the needs of its own growth. For Lilium concolor var. megalanthum plants derived from bulbs of male plants, higher levels of nutrients enable it to obtain sufficient nutrients, and the resource allocation of each organ changes accordingly during growth and development to achieve maximum fitness, resulting in no significant difference in each growth index between different treatments.This result indicated that there is a resource allocation trade-off for plants in the process of using limited resources, with increasing the resource allocation of some functions while reducing the resource allocation of other functions (Elle, 1999; Ronsheim and Bever, 2000), and this trade-off is variable with different soil nutrient levels.

The accumulation and allocation of plant biomass reflects the way of plant growth and metabolism and affects the functional attributes of plant organs (Hecht et al., 2019). Plants can better capture and preserve limited resources in soil by changing their aboveground and belowground biomass accumulation (Wang et al., 2019). As the main supporting and photosynthesis organs, the stems and leaves of Lilium concolor var. megalanthum occupy a large biomass allocation during the whole growth period, which is the guarantee for the transportation of flowering and fruiting resources (Russo et al., 2022). However, there was no significant differences in the accumulation of aboveground and belowground biomass of Lilium concolor var. megalanthum plants derived from bulbs of hermaphrodite flower and male flower plant among the different fertilization treatment groups (except potassium addition treatment). The results showed that Lilium concolor var. megalanthum considered the dual needs of belowground root space expansion and aboveground leaves to obtain light energy resources, and promoted the simultaneous growth of aboveground and belowground parts (Bai et al., 2020). In addition, the response of plants to the potential limitation of essential nutrients makes them reallocate resources based on the trade-off of important organ resource demand (Goodnoe and Hill, 2018). When the content of nitrogen and phosphorus in the soil was low (potassium addition treatment), the Lilium concolor var. megalanthum plants derived from bulbs of hermaphrodite flower plant reduced the input of aboveground biomass and increased the accumulation of belowground biomass. Lilium concolor var. megalanthum allocated most of the resources to the belowground part, mainly to promote the growth of bulb and accumulation of bulb biomass. The bulbs of the mother plant accumulate more resources for the growth of the next year, which may be a reproductive strategy to remedy for the low efficiency of sexual reproduction of aboveground parts. Secondly, it may also be that sexual reproduction requires a lot of resource input, and the resource input of asexual reproduction is lower than that of sexual reproduction. Therefore, the selection of asexual reproduction may make the population diffusion efficiency and reproductive success rate higher.

The uptake and accumulation of nutrients in different organs can indirectly reflect the nutritional needs of plants at a certain stage (Quiñones et al., 2012). At the same time, the lack or excess of one or more nutrients may cause physiological stress and hinder the growth and development of plants (Srivastava, 2013). We found that the flower diameter of Lilium concolor var. megalanthum derived from bulbs of hermaphrodite flower plant decreased significantly under low soil nitrogen and phosphorus (potassium addition treatment) nutrient levels; but the flower diameter increased significantly under the condition of nitrogen addition. This shows that nitrogen can promote the growth of flower of Lilium concolor var. megalanthum plants, and larger flowers production has the advantage of attracting pollinators and improving their fitness. However, the petal length and stamen length of Lilium concolor var. megalanthum derived from bulbs of male flower plants were significantly reduced at higher nitrogen levels. It can be thought that the flowers of the plant are sensitive to nutrient levels, and excessive nutrient levels will inhibit them. In addition, the flower diameter of Lilium concolor var. megalanthum decreased while the flower longevity was longer. This reflects the trade-off between flower longevity and flower size. Limited resource allocation restricts plants from either producing larger flowers or having longer flowering period. Adaptability under low resource conditions may optimize resource allocation by limiting the size of flowers at the expense of extending flower longevity. Our results only showed the trade-off between flower longevity and flower size, but many studies have demonstrated the trade-off between flower longevity and flower number, and between flower size and flower number (Caruso et al., 2012; Spigler and Woodard, 2019). This is a reproductive strategy formed by plants under adverse environmental conditions, which is the embodiment of their adaptation to environmental changes.

Lilium concolor var. megalanthum, as a perennial herb, can not only produce bulbs to form ramets through clonal growth, but also produce seeds through sexual reproduction for germinating and forming new seedlings. This may lead to trade-off between the organs of the plant itself when competing for limited resources, including the trade-off between asexual and sexual reproduction. We found that the higher asexual fecundity of Lilium concolor var. megalanthum, the lower sexual fecundity, which indicates that the fecundity of Lilium concolor var. megalanthum has great ecological plasticity. Long-term asexual reproduction will reduce the genotype of its population. Once the environment changes, Lilium concolor var. megalanthum population will face great risks. The Lilium concolor var. megalanthum population can survive well, may be dependent on the supplement of sexual reproduction to asexual reproduction (Arizaga and Ezcurra, 2002). A large number of flowering can increase the opportunity for gene exchange between populations and produce genotypes that are more adaptable to changing environments. This trade-off mechanism of Lilium concolor var. megalanthum makes its population easy to adapt to the environmental conditions in the region, which is conducive to the maintenance of its population.

Our results showed that when the availability of resources is limited, there is a strong trade-off between the different functions of Lilium concolor var. megalanthum. However, the effect of soil nutrient difference on resource allocation of different functions was not significant in the second year. In the first year of the growing season, all Lilium concolor var. megalanthum plants produce one flower, while in the second year, most plants produce two or three flowers, and some individuals even produce 5-6 flowers. The number of flowers proves that Lilium concolor var. megalanthum have sufficient nutrients for sexual reproduction (Henriot et al., 2019). We believed that the reason may be that the bulbs of Lilium concolor var. megalanthum provide sufficient resources for sexual reproduction, and the resource allocation of different organs is less regulated by soil nutrients.

We observed the sex expression of Lilium concolor var. megalanthum plants for three consecutive growing seasons, and found that under the nutrient levels lower than and equal to the natural habitat, the size of the majority of plants in the second growing season decreased, but bulbs increased after Lilium concolor var. megalanthum consumed more resources to produce hermaphrodite flowers, and some plants changed sex to produce male flowers or did not bloom. Zhang et al. (2014) also found that Lilium apertum (Liliaceae) can switch from one gender stage to another with plant size changes from year to year. This result is consistent with our findings. The availability of resources affects the size of plants, indicating that nutrient levels play a key role in determining plant sex expression. Some plants not expressed as hermaphrodite plants for two consecutive growing seasons may be related to certain hormone levels. Many studies have found that many hormones are involved in and regulate plant sex determination. For example, jasmonic acid can promote the differentiation of female flowers (Li et al., 2021), and cytokinin can transform Vitis vinifera from male to hermaphrodite flowers (Negi and Olmo, 1966). Studies have shown that high levels of zeatin in hermaphrodite flowers may be an important regulator of sex determination. The difference in zeatin between the two sexual phenotypes is partly due to the difference in the content of available nitrogen as an essential substrate for auxin synthesis (Huang et al., 2022). Auxin can be used as a remote signal to transmit plant rhizosphere nitrogen status and is positively correlated with available nitrogen (Kamada-Nobusada et al., 2013). Therefore, nitrogen deficiency may lead to insufficient auxin synthesis and induce the transformation of hermaphrodite plants to male or vegetative plants (non-flowering) during the second growing season (Huang et al., 2022). This indicates that nutrient levels can affect sex expression by regulating plant hormone synthesis. In our study, when the soil nitrogen and potassium or nitrogen and phosphorus content is low, the proportion of non-flowering Lilium concolor var. megalanthum plants is more than 50%. We speculate that various nutrient elements produce synergistic effects, jointly regulate hormone synthesis under the condition of limited resources, and regulate plant selective expression the most effective sexual function of reproduction.

The plants with male flowers mainly promote the growth of belowground parts and accumulate belowground resources when the nutrient level was higher than that of natural habitats. In the second growing season, the size of some plants increased, the bulbs of all plants increased, and most plants produce hermaphrodite flowers. Male flowers often have underdeveloped and nonfunctional female organs, which is caused by the stop of pistil development in the late stage of development (Astuti et al., 2020). Contrary to the role of auxin, jasmonic acid was found to play an important role in promoting pistil abortion. The related genes involved in jasmonic acid signal transduction are highly expressed in male plants, eliminating pistil primordia through cell death and promoting male development (Acosta et al., 2009). Our results showed that the increase of single nutrient was more likely to promote the occurrence of hermaphrodite flowers in Lilium concolor var. megalanthum compared with the increase of compound nutrients. Studies have shown that higher nutrient levels may promote the production of auxin. Auxin may act as an antagonist of jasmonic acid signal transduction related gene expression and play an important regulatory role in plant development and sex expression (male flower female organ abortion) (Nitschke et al., 2016).

In the third growing season, regardless of whether it was originally derived from bulbs of hermaphrodite plants or bulbs of male plants, all plant size and bulbs increased. Most of plants produce male flowers in the first two growing seasons produce hermaphrodite flowers in the third growing season. The results showed that the sex expression of Lilium concolor var. megalanthum was flexible and regulated by resource availability. When resources in the environment become constrained, the nutrient supply of the Lilium concolor var. megalanthum does not meet its hermaphrodite flowers, plants will produce male flowers or do not bloom. In this process, bulbs grow and accumulate resources. We found that the sex expression of Lilium concolor var. megalanthum changed for three consecutive years, and it can be defined as a gender diphasy system. The sex expression is dependent on the size of the bulb (Niu et al., 2017). We speculate that the bulb size is more closely related to the sex expression than the plant size. When the bulb grows to a certain size, and enough to maintain its sexual reproduction, especially female function, plant will always produce hermaphrodite flowers. At this time, the role of resource availability in the environment will decline. This may be a way for resource availability as an environmental factor to affect the sex expression of gender diphasy plants.