The 3-year-old C. chinensis used in the present study were collected from the cultivation base in Pengzhou, Sichuan, China (31° 10′ N, 103° 50′ E). The average daily sunshine duration, air temperature, and precipitation for the last 20 years (2000–2020) were obtained from the nearby meteorological station and are shown in Supplementary Figure S2 . During the period of leaf expansion (early spring), which had the highest number of daily sunshine hours throughout the year, the range of sunshine hours was between 4 h and 5.5 h ( Supplementary Figure S2 ). Before the experiment was conducted, the plants were transplanted (February 2022) from the Pengzhou to the Chinese Medicine plant garden of Chengdu University of Traditional Chinese Medicine (30° 68′ N, 103° 81′ E) for 1 month of acclimatization. The plants were grown individually in plastic pots (13 cm³ × 15 cm³). An ochric aquic cambosol soil was obtained from the surface layer (0–20 cm) of farmland in the plant garden. The soil was air-dried, sieved to 2 mm, and thoroughly mixed, and properties were as follows: pH (1:2.5, soil:water) was 6.31, organic matter content was 24.1 g kg⁻¹, total N was 1.6 g kg⁻¹, available N was 120 mg kg⁻¹, Olsen-P was 15.3 mg kg⁻¹, and available K was 110 mg kg⁻¹. Each pot was filled with 5 kg of air-dried soil. As the fertility of the soil was adequate for plant growth, no extra mineral fertilizers were added to the soil. All pots were watered daily until they were moved to the greenhouse.

The experiment was a randomized block design in an artificial climate chamber with 85% relative humidity and a day/night temperature of 20°C/10°C. According to the daily sunshine hours during the period of leaf expansion, a photoperiod of 8 h day/16 h night was employed. A light-emitting diode (LED), with wavelengths ranging from 380 nm to 800 nm, provides the irradiance, with a red-to-blue ratio of 3:1. The lamps were evenly distributed on the panel with a 4-cm row distance, a 5-cm lamp distance, and increased density around the panel to ensure uniformity and stability of light intensity under the lamp ( Supplementary Figure S3 ). Usually, 0%–4% and 4%–12% of the open-sky light are approximately closed forest understory and tree-fall gap light environments, respectively (Canham et al., 1990; Walters and Reich, 1999). In combination with the light intensity of the growing area (canopy light intensity at noon averages about 2,000 μmol m⁻² s⁻¹) of C. chinensis ( Supplementary Figure S4 ), two light intensities were employed: 50 μmol m⁻² s⁻¹ ± 5 μmol m⁻² s⁻¹ (low light) and 200 μmol m⁻² s⁻¹ ± 5 μmol m⁻² s⁻¹ (relatively high light), which represent 2.5% and 10% of the canopy light, respectively. Our previous study using a fast light response curve also found that a light intensity of 200 μmol m⁻² s⁻¹ is typical for the cultivation of C. chinensis in the laboratory ( Supplementary Tables S1 , S2 ), although it is much lower than natural light intensities found in field conditions. Light intensity was measured using a quantum sensor (UPRtek, PG100N, Taiwan, China).

When the new leaves were ready to appear (March 10, 2022), the uniformly growing and healthy plants in pots were transferred from the Chinese Medicine plant garden to the artificial climate chamber. The arrangement of pots under the LED panel is shown in Supplementary Figure S3 . The pots were arranged in a diamond shape, with a 20-cm distance between two pots to ensure that the plants were unobstructed from each other and received almost the same amount of light from the LED. A distance of 50 cm was maintained between the canopy and the light to eliminate the effect of temperature produced by the lamp in the two light conditions.

Plants from both light treatments were transplanted back to the Pengzhou base after the controlled experiment, and the field management after transplanting was the same as the local cultivation practice. The plants were dug up in October to determine the dry weight of the RO.

The chlorophyll fluorescence parameters of the ILs and MLs were measured at 15 d (ILs fully expanded and pale green; MLs grow normally dark green) and 35 d (ILs fully expanded and dark green; MLs with no obvious changes) after the start of the light treatment, respectively. The ILs of fresh plants were cut for the determination of pigment content. Plants were harvested 15 d and 35 d after the start of light treatment, frozen in liquid nitrogen, and stored at −80°C for subsequent index determination.

The chlorophyll fluorescence of ILs and MLs, respectively, was measured using a MINI-PAM-II Portable Fluorometer (Walz, Effeltrich, Germany) in situ in complete darkness in an artificial climate chamber on day 15 and day 35 of light treatment. The leaves were dark-adapted for 30 min, and then a 0.6-s saturating light pulse (1,100 μmol m⁻² s⁻¹) and an actinic light for 60 s (261 μmol m⁻² s⁻¹) were applied to obtain the minimal and maximal fluorescence yield (F ₀ and F_m ). The maximum light-adapted fluorescence (F_m’) and steady-state fluorescence (F_s ) were then measured with continuous actinic light. The PSII operating efficiency [Y(II)], relative electron transport rate (rETR), photochemical quenching (qP), the maximum photochemical efficiency PSII (F_v/F_m ), and the nonphotochemical dissipation of absorbed energy (NPQ) were calculated with the following formula (Kooten and Snel, 1990; Oquist and Chow, 1992):

After 15 d and 35 d of treatment, fresh ILs were cut and crushed immediately with liquid nitrogen, and 0.1 g (fresh weight (FW)) of powder was soaked in 10 mL of 80% v/v acetone (20% ethanol) at 25°C to extract chlorophyll for 24 h in the dark (Lichtenthaler and Buschmann, 2001). Powdering the leaves speeds up the leaching of chlorophyll and shortens the steeping time. The absorbance at 645 nm and 663 nm was measured with an enzyme-labeled instrument (SpectraMax iD3, Molecular Devices, USA), and the chlorophyll concentration was calculated with the following formula (Arnon, 1949; Porra et al., 1989).

Where V is the total volume of extract (mL) and W is the leaf fresh weight (g).

Soluble sugars from lyophilized (FDU-2110, EYELA, Japan) leaves and RO (dry weight (DW)) were extracted with 80% v/v ethanol at 80°C for 10 min, and the supernatants of the two extractions were combined to prepare soluble sugar samples for analysis. Soluble sugar was measured directly in the extract by a Waters Binary HPLC System (Waters 1525-2707, Milford, USA) equipped with a refractive index detector (2414, Waters, Milford, USA). The analytical conditions were as follows: column Agilent Hi-Plea Ca (8% crosslinked) 300 mm × 7.7 mm, 8 μm in diameter (Agilent Technologies Inc., USA); column temperature 85°C; mobile phase Milli-Q water; flow rate 0.6 mL min⁻¹ (Zhou et al., 2022). Data were collected and processed by the Waters Chromatography Station DataApex. Sugars were identified by comparison with retention times and coelution of authentic standard solutions.

The residue after ethanol extraction was washed several times with ultra-pure water and used for starch analysis. The precipitate was gelatinized at 100°C for 10 min, and the process was repeated once. The resulting supernatants were combined and treated with perchlorate before analysis. The starch content was determined by taking an appropriate amount of sample by the anthrone method (Setter and Ella, 1994).

Small squares of 1 mm² size were cut from the ILs after 35 d of growth in the light resource and immediately placed in 3% glutaraldehyde for prefixation. The tissue was then postfixed in 1% osmium tetroxide, dehydrated in series acetone, infiltrated in Epox 812 for a longer period, and embedded. The semithin sections were stained with methylene blue, and ultrathin sections (EMUC7, Leica, Germany) were cut with a diamond knife and stained with uranyl acetate and lead citrate. Sections were examined with TEM (JEM-1400FLASH, JEOL, Japan).

To explore the role of ILs and MLs in plant development under different light intensities, the ILs or MLs were labeled with ¹³CO₂. After 20 d of light treatment, six pots (three for IL labeling and three for ML labeling) in each light treatment were subjected to ¹³CO₂ labeling. Leaf labeling was performed 1 h after the start of illumination. Similar sizes of the ILs or MLs were carefully covered with a homemade sealing bag (sealed with foam glue). A syringe was used to inject 40 mL of ¹³CO₂ gas into the bag, and 30 min later, the marking device was disassembled.

Two weeks after labeling, the plants (ILs labeled, MLs labeled, and unlabeled) were harvested and oven-dried at 70°C for 3 d for an assay of ¹³C abundance in the ILs, MLs, and RO. The movement of ¹³C under different light conditions was monitored; we used the labeled part as the denominator and the unlabeled part as the numerator, which reflects the amount transferred from labeled to unlabeled. A 0.2-mg plant sample was precisely weighed, and the ¹³C/¹²C ratio was measured using a stable isotope mass spectrometer (Delta V Advantage, Thermo Fisher Scientific, Germany).

The four fractions of foliar P (lipid P, nucleic acid P, metabolic P (including Pi), and residual P) were extracted from 50 mg of freeze-dried ILs and MLs according to Hidaka and Kitayama (2013) with modifications. The plant samples were digested with a mixed acid solution (nitric acid: perchlorate = 3:1, v/v) to extract different P fractions. The RO of total P was extracted by sulfuric acid. P was determined by the molybdenum blue method using a spectrophotometer (UV-6100, Mapada, Shanghai, China) (AMSs, 1966).

The determination of the alkaloid content of C. chinensis in RO was performed after 35 d of light intensity treatment according to the method of Qi Qi et al. (2018) with high-performance liquid chromatography (HPLC) (LC-20AT, Shimadzu, Japan). The lyophilized RO was weighed and ultrasonically dissolved in a hydrochloric acid–methanol solution (1:100, v/v). An Xtimate C18 chromatographic column (250 mm × 4.6 mm, 5 μm) (Welch, China) was applied to separate different alkaloid compounds, and the column temperature was 25°C. The mobile phase included acetonitrile (A) and 30 mmol/L ammonium bicarbonate solution (consisting of 0.7% ammonia and 0.25% triethylamine) (B) at a flow rate of 1 mL/min. The gradient program was set as follows: 0–15 min, 90%–75% B; 15–25 min, 75%–70% B; 25–50 min, 70%–55% B. The injection volume was set as 10 μL, and the detection wavelength was set at 270 nm.

A one-way analysis of variance was performed to compare the data between the light-intensity experiments. The data was compared between two light intensity treatments or between the two sampling times of the same light intensity treatment using Tukey’s honestly significant difference test. All statistical analyses were performed using the software SPSS 25.0 (SPSS Inc., USA). The figures were drawn using Origin 2019 (Origin, USA).

Light intensity affected the photochemistry of C. chinensis leaves during the leaf expansion stage. The Y(II), ETR, and qP in ILs under 200 μmol m⁻² s⁻¹ were significantly higher than those exposed to 50 μmol m⁻² s⁻¹ light intensity ( Figures 1A–C ). However, in MLs, the Y(II) showed no significant difference between the two light treatments ( Figure 1F ), and there was no difference between the observed ETR and qP between the two sampling times ( Figures 1G, H ). Compared to relatively high light, the F_v/F_m in leaves was higher when the plants were under low light, but there was no change at the two sampling times ( Figures 1D, I ). In ILs, light intensity had little effect on NPQ ( Figure 1E ), while NPQ in MLs under relatively high light was significantly higher than that under low light ( Figure 1J ).

The plant phenotype observed that ILs turn green quickly under low light ( Figure 2A ). The chlorophyll concentration in ILs was consistent with our observations that decreasing light intensity increased the chlorophyll concentration, particularly Chl b concentration, and subsequently decreased the ratio of Chl a/Chl b ratio ( Figures 2B, C ).

Light intensity affected the sucrose, glucose, and starch concentrations in ILs, MLs, and RO at the two sampling times ( Figure 3 ). The sucrose concentration in ILs and MLs under 200 μmol m⁻² s⁻¹ light intensity was higher than that under 50 μmol m⁻² s⁻¹ light intensity, and it was higher at 15 d compared to 35 d ( Figures 3A, B ). Compared with plants under low light intensity, relatively high light intensity increased the glucose concentration at both sampling times, irrespective of the age of the leaves ( Figures 3C, D ). In ILs, the starch concentration showed no difference between the two light treatments at 15 d, but in relatively high light, it was significantly higher than that in the low light at 35 days ( Figures 3E, F ). In MLs, the starch concentration showed a decrease from 15d to 35 d under both two light treatments, which was decreased further under low light ( Figure 3F ). After 35 d light intensity treatment, the starch and glucose in the RO under 200 μmol m⁻² s⁻¹ were higher than those under 50 μmol m⁻² s⁻¹ ( Figure 3G ). In addition, the results of the electron microscope photography were consistent with the results obtained above. The chloroplasts were full of starch granules under 200 μmol m⁻² s⁻¹, but almost no starch granules were observed under 50 μmol m⁻² s⁻¹ ( Figures 3H–K ).

¹³C labeling plays an important role in monitoring carbon flow. The proportion of ¹³C in ILs under 200 μmol m⁻² s⁻¹ was significantly higher than that under 50 μmol m⁻² s⁻¹, regardless of whether ILs or MLs were labeled with ¹³C. Furthermore, under 200 μmol m⁻² s⁻¹ treatment, ¹³C atoms were detected in ILs and RO but not in MLs when ¹³C was labeled in ILs. Under 50 μmol m⁻² s⁻¹ light, the ratio of mature-to-immature leaf ¹³C concentration in the treatment of IL ¹³C labeling (12.2%) was higher than the ratio of immature-to-mature leaf ¹³C concentration in the treatment of ML ¹³C labeling (7.3%), ¹³C atoms were detected in ILs and MLs but were not in the RO ( Figure 4 ).

The total P concentration in leaves and RO under 200 μmol m⁻² s⁻¹ was lower than that under 50 μmol m⁻² s⁻¹ ( Figures 5A–C ). Under 50 μmol m⁻² s⁻¹ light intensity, the leaf P concentration of MLs showed no difference observed at both sampling times ( Figure 5B ). However, the leaf P concentration of ILs under 200 μmol m⁻² s⁻¹ light intensity was lower than that under 50 μmol m⁻² s⁻¹ light intensity at both sampling times ( Figure 5A ). Light not only affected the foliar total P concentration but also the foliar P fractions, especially the organic P ( Figures 5D–G ). In ILs, the lipid P and metabolite P under relatively high light were significantly lower than those under low light intensity at both sampling times ( Figures 5D, E ). At 15 d, the main sources of phosphorus in ILs were lipid P and metabolic P ( Figure 5D ), but metabolic P and nucleic P had a high proportion at 35 d ( Figure 5E ). In MLs, only nucleic P increased with decreasing light intensity at 15 d ( Figure 5F ). Nucleic P had a high proportion in MLs at 15 d, and the proportion of lipid P and metabolic P increased at 35 d compared to 15 d ( Figures 5F, G ). Furthermore, the residual P showed no difference in ILs and MLs under the two light intensity treatments at both measuring times ( Figures 5D–G ).

The findings demonstrated that under low light conditions, the alkaloid concentrations in the RO of berberine (23.0%), coptisine (19.9%), and palmatine (17.9%) were all greater than those under relatively high light intensity ( Figure 6 ). The light treatment at the leaf expansion stage has an effect on the biomass of RO at the mature stage. The RO dry weight of plants under relatively high light at the mature stage was significantly higher than that under low light ( Supplementary Figure S6 ).

The relationships between the starch concentration in RO and the concentration of the three main alkaloids in RO were investigated via regression analyses ( Figure 7 ). The concentrations of the three main alkaloids ( Figure 7 ) were negatively correlated with starch concentration in RO (n = 6; R ^2 = 0.67, 0.67, 0.66; p ≤ 0.05), indicating that the lower the starch concentration, the higher the alkaloid concentrations.

The initial growth stages of ILs primarily rely on imported reserves from source organs such as MLs, stems, and RO, but eventually become net exporters of carbon and rely on their own photosynthates (Miyazawa et al., 2003). High light intensity enhances local photosynthesis, which induces the transfer of produced carbohydrates to source organs, reducing the consumption of reserves in source organs to supply new leaf growth, making reserves more durable, and ultimately increasing their biomass (Rossi et al., 2015; Feng et al., 2019; Wimalasekera, 2019; Zhou et al., 2022). Our findings confirm that increased carbon assimilation by ILs and MLs under high light during leaf expansion reduces RO reserve consumption ( Figure 4 ), ultimately increasing RO biomass ( Supplementary Figure S6 ). Additionally, high light accelerates the conversion of ILs from sink to source, which reduces RO reserve consumption. However, we observed that MLs exacerbated the depletion of RO reserves as a sink under low-light conditions, which contrasts with prior studies.

In C. chinensis, a delayed-greening species, leaves from emergence to maturity take at least 45–60 d. In this process of leaf maturation, ILs, MLs, and RO are involved in a high intensity of reserve exchange. In most dicotyledonous plants, the leaf changes from sink to source when the leaf area reaches 10%–60% of the final area (Miyazawa et al., 2003; Pantin et al., 2012). In the present study, some ILs were not fully expanded 15 d after treatment, and all had reached the maximum area by 35 d, but the chlorophyll concentration in ILs was lower than MLs ( Figure 2 ). Furthermore, the F_v/F_m showed no difference between the two sampling times and reached a high level above 0.8 ( Figures 1D, I ). The amount of glucose and starch, the major carbohydrate forms for transport and storage, increased markedly from 15 d to 35 d in ILs ( Figures 3C, E ). These results are consistent with the typical characteristics of the transition from the sink to the source of ILs in the studies that have been reported (Turgeon, 1989; Czech et al., 2009). Indeed, the growth of ILs in C. chinensis at least partially depends on its photosynthesis from 15 d to 35 d after emergence of leaves. The glucose and starch concentrations in ILs increased markedly from 15 d to 35 d when plants were grown under 200 μmol m⁻² s⁻¹ ( Figures 3C, E ). In contrast, under 50 μmol m⁻² s⁻¹, the glucose and starch concentrations in ILs remained at very low levels and did not show significant changes at the two sampling times ( Figures 3C, E ). Interestingly, the glucose concentration in the RO was lower in plants grown under 200 μmol m⁻² s⁻¹ than those grown under 50 μmol m⁻² s⁻¹ at 35 d ( Figure 3G ), suggesting that starch breakdown was triggered to support shoot growth under low light conditions (Hajirezaei et al., 2003; Stitt and Zeeman, 2012). Additionally, ¹³C was detected in the RO of the ILs ¹³C label under 200 μmol m⁻² s⁻¹ but not under 50 μmol m⁻² s⁻¹ ( Figure 4 ). These findings indicate that increasing light intensity before leaf maturation in C. chinensis can significantly enhance the autotrophic capacity of ILs by providing carbohydrates to support their own growth and transporting carbohydrates underground to accumulate RO biomass. In contrast, under low light, the RO continues to supply the growth of ILs, leading to increased consumption of reserves in the RO.

The MLs of evergreen species can resume their functions in early spring to sustain themselves and provide partial resources to support the new leaf growth as the temperature rises (Walters and Reich, 1999; Ida and Kudo, 2008). Our findings reveal that the functional performance of MLs is mainly influenced by the light intensity to which they are exposed. Under high light intensity, a higher proportion of ¹³C was detected in the RO and ILs when MLs were labeled with ¹³C. This observation provided evidence to support the view that MLs provide partial resource support for IL growth (Turgeon, 1989). No difference in starch concentration was observed in the MLs between the two light intensity treatments at 15 d ( Figure 3F ). However, at 35 d, MLs under low light exhibited a lower starch concentration compared to those under high light ( Figure 3F ). These suggest that under 50 μmol m⁻² s⁻¹, MLs rely on stored starch to ensure their survival (Stitt & Zeeman, 2012). Furthermore, the ratio of ¹³C in mature-to-immature leaf in IL ¹³C labeling was 12.2%, and the ratio of ¹³C in immature-to-mature leaf in ML ¹³C labeling was 7.3% under 50 μmol m⁻² s⁻¹ light intensity ( Figure 4 ). Moreover, under 200 μmol m⁻² s⁻¹, the ¹³C was not even detected in MLs in IL ¹³C labeling ( Figure 4 ). Therefore, under low light, MLs did not have additional carbon to sustain IL growth, but instead received energy from the ILs.

Shade-tolerant plants have evolved a trade-off between growth potential and survival in low light, such as increasing their photosynthesis efficiency and decreasing the respiration rate to decrease the consumption of energy when the plants are grown in deep shade (Loach, 1967; Reich et al., 1998; Yao et al., 2017). However, in our study, MLs maintained the same respiration rate under low light intensity as under high light and remained at a high level (data not shown), which is inconsistent with previous observations in other species (Loach, 1967; Reich et al., 1998; Yao et al., 2017). One possible explanation for this discrepancy might be that the 50-μmol m⁻² s⁻¹ light intensity harmed the carbon fixation of C. chinensis but did not cause severe stress ( Supplementary Figure S5 ). For example, the F_v/F_m , a sensitive indicator of plant responses to environmental stresses (Baker, 2008), showed no difference between 15 d and 35 d in the two samples, regardless of leaf age ( Figures 1D, I ). Additionally, C. chinensis is typically found in mountainous regions at elevations of 1,000–2,000 m. During the new leaf growth period (from February to April), the average day/night temperature is around 15°C/5°C ( Supplementary Figure S2D ). Under low temperatures, C. chinensis, like other shade-tolerant plants, can survive under low light by reducing their respiration rate (Walters and Reich, 1999; Walters and Reich, 2000). However, the day/night temperatures set in this experiment were higher than those found in its natural habitat. To ensure survival, MLs may need to utilize energy from other organs. Nevertheless, the strategy of MLs of shade-tolerant perennial plants to adapt to low light stress at higher temperatures by consuming energy provided by other organs to sustain survival rather than rapid senescence needs more evidence.

The energy consumption in the source organ (especially RO) by flowering was not considered in this experiment. This may have potentially amplified the result indicating that light-mediated IL growth increases the consumption of reserves in the source organ (RO), particularly under low-light conditions. However, the amount of reserves in the RO did not change the source–sink relationship among ILs, MLs, and RO, which is mainly determined by the light-regulated carbon assimilation capacity of ILs and MLs.

The expansion of new leaves in understory plants during spring is mainly supported by stored P rather than uptake by roots in the current season (Schachtman et al., 1998; Rychter and Rao, 2005). The low P concentration in the new leaves suggests a low total of P export from the source organ. Our data are consistent with this observation, as the ILs of plants grown under 200 μmol m⁻² s⁻¹ had lower P concentration than those grown under 50 μmol m⁻² s⁻¹ ( Figure 5A ). In addition, the Y(II) of ILs under 200 μmol m⁻² s⁻¹ was higher than those under 50 μmol m⁻² s⁻¹ ( Figure 1A ).

Relatively high light increased photosynthesis efficiency in ILs associated with lower P concentrations in two ways. First, the level of enzymes involved in Rubisco and other Calvin–Benson cycles is mainly connected to the amount of rRNA, which is mainly determined by the amount of nucleic P (Rychter and Rao, 2005; Hussain et al., 2021). The high light intensity did not decrease the nucleic P concentration in the ILs compared to that under low light ( Figures 5D, F ). Second, under high light intensity, a lower number of chloroplasts and a low chlorophyll concentration will decrease the investment of P in protein synthesis and cellular growth (Elser et al., 2010; Veneklaas et al., 2012). Light intensity rather than chlorophyll concentration is the main factor limiting photosynthesis (Rychter and Rao, 2005; Hussain et al., 2021). Our data are consistent with this observation, as the chloroplast number and chlorophyll concentration in ILs of plants grown under 200 μmol m⁻² s⁻¹ were lower than those of ILs grown under 50 μmol m⁻² s⁻¹ ( Figures 2B , 3H–K ). Furthermore, when the plants were grown under 50 μmol m⁻² s⁻¹ light intensity, the increased chlorophyll concentration was mostly attributed to a high proportion of Chl b ( Figures 2B, C ), which is currently believed to improve light harvesting under low light conditions and has been demonstrated in numerous earlier investigations (Escoubas et al., 1995; Sun et al., 2003; Valladares and Niinemets, 2008; Melgar et al., 2009; Feng et al., 2019).

The P concentration in the RO of C. chinensis grown under 200 μmol m⁻² s⁻¹ was lower than that of plants grown under 50 μmol m⁻² s⁻¹, which may challenge the view that high light suppresses the amount of P exported from the RO. One possible explanation is due to the concentration effect ( Supplementary Figure S6 ) (Jarrell and Beverly, 1981). Under 50 μmol m⁻² s⁻¹, the amount of carbohydrates was much higher than P from RO to supply the growth of ILs, resulting in an increased P concentration in the RO. Indeed, relatively high light intensity reduced the demand for P to maintain the photosynthesis potential of ILs, thus inhibiting the export of P from the RO ( Figure 8 ).

Rather than directly mediating synthesis, low light promotes the increase in alkaloid concentration in C. chinensis RO in early spring by mediating substance translocation. However, low light significantly reduces RO biomass ( Supplementary Figure S6 ). In most evergreen plants, nonstructural carbohydrate concentrations typically reach their maximum in the fall, decline during the winter, and reach minimum levels immediately after growth initiation in the spring (Wyka, 1999). The alkaloid concentration in the RO of C. chinensis showed two peaks in March (new leaf expansion) and September (vigorous growth period), with the concentration in September being significantly higher than that in March (Ke, unpublished; Zhao and Du, 2002). RO alkaloids and nonstructural carbohydrates were negatively correlated in March and positively correlated in September. Secondary metabolism usually uses the end products of primary metabolism as substrates (Drew and Demain, 1977), suggesting that the increase in RO alkaloids in March may not be regulated by synthesis.

In the present study, RO under low light in the early spring had lower starch and higher alkaloid concentration than that of plants under high light ( Figure 6 ). Furthermore, there was a negative correlation between starch and alkaloid concentration ( Figure 7 ). The increase in alkaloid concentration mainly attributed to the concentration effect (Jarrell and Beverly, 1981), as the amount of carbohydrates in the RO supplying ILs growth was higher under low light than under high light, resulting in an increased alkaloid concentration in the RO. These findings support our view, but further exploration is needed to understand the effect of light on alkaloid synthesis in C. chinensis.