We determined the time needed to reach different coverage levels in a specific area to analyze plant growth rates by estimating the days based on the sigmoidal curves fitted to the projected canopy size (PCS) information. Plants growing under different lighting regimes needed different days to achieve specific sizes indicated by the coverage percentage of a reference area. For example, to cover 25% of the said area, plants growing with 0, 1, 2, 3, 4, and 5 days with low DLI following a day with high DLI needed an average of 11.46, 12.41, 11.81, 11.86, 11.6, and 11.54 days, respectively ( Figure 1A ). We did not find a significant interaction between the lettuce cultivars and lighting treatment (F_5,24 = 0.028, P=0.99). The time taken to achieve this level of coverage showed significant differences (F_5,29 = 3.91, P=0.007) between plants growing with one day of low DLI compared to the control treatment (P=0.007) (0 days with low DLI) and plants with 4 (P=0.029) and 5 (P=0.017) days with low DLI.

Likewise, to cover 50% of the reference area, plants subjected to 0, 1, 2, 3, 4 and 5 days with low DLI following a day with high DLI needed 14.25, 15.22, 14.58, 14.6, 14.26, and 14.24 days, respectively ( Figure 1B ). The interaction between cultivar and lighting treatment was not significant (F_5,24 = 0.26, P=0.92). Once again, the time needed to attain this coverage level exhibited significant differences (F_{5, 29} = 3.62, P=0.011) when comparing plants subjected to one day of low DLI when compared with the control (P=0.017) and plants with 4 (P=0.023) and 5 (P=0.018) days with low DLI.

To achieve 75% coverage, plants growing with 0, 1, 2, 3, 4 and 5 days with low DLI following a day with high DLI needed 16.75, 17.68, 17.1, 16.97, 16.76, and 16.96 days, respectively ( Figure 1C ). Here, significant differences were found (F_5,29 = 2.59, P=0.046) (Tukey’s test did not show differences when comparing means) with no interaction between lighting treatments and cultivars (F_5,24 = 0.9, P=0.49). Since Tukey’s test did not show differences when comparing means, we used Fisher’s LSD test. Here, plants growing with one day of low DLI needed significantly more days to achieve 75% of coverage when compared to plants growing under 0 (P=0.005), 3 (P=0.0087), 4 (P=0.0064) and 5 (P=0.028) days with low DLI.

Finally, to cover 100% of the reference area, plants growing under 0, 1, 2, 3, 4 and 5 days with low DLI following a day with high DLI needed 19.3, 19.96, 19.5, 19.72, 19.14, and 19.27 days, respectively ( Figure 1D ). There were no significant differences among treatments regarding covering the equivalent of the entire reference area (F_5,29 = 0.76, P=0.26) (F_5,29 = 1.35, P=0.26). Interaction between cultivar and lighting treatments was not found (F_5,24 = 1.35, P=0.29). (F_5,24 = 1.304, P=0.29).

We measured the final leaf area on three plants situated at the central transverse position of each tray, yielding a total count of six plants. The average leaf area for plants growing under the least amount of days with low DLI (T0) ascendingly to 5 days with DLI was as follows: 1436.09, 1276.86, 1342.12, 1388.37, 1409.01 and 1452.79 cm² respectively ( Figure 2 ). However, we did not find significant differences in this parameter (F_5,29 = 1.65, P=0.17). Furthermore, there was no interaction between light treatment and cultivar (F_5,24 = 0.36, P=0.88).

We did not find significant differences (F_5,29 = 0.79, P=0.56) in specific leaf area (SLA) among plants growing under different lighting treatments, and there was no interaction between SLA and lettuce cultivar (F_5,24 = 0.74, P=0.59). The average SLA per plant from treatment 0 to 5 days were as follows: 407.86, 371.18, 386.29, 370.17, 355.63, and 370.17 cm²·g^-1 ( Figure 3 ).

We conducted pigment content measurements on both lettuce varieties. For ‘Rouxai’, the leaf anthocyanin content varied from 7.36 to 7.46 anthocyanin content index (ACI) across treatments with varying low DLI days ( Figure 4 ), with no significant differences found (F_5,12 = 0.36, P=0.86). Additionally, the leaf chlorophyll content in sequence from treatment T0 to T5 were 8.36, 8.05, 8.06, 8.41, 8.11, and 7.98 chlorophyll content index (CCI) ( Figure 5 ). However, no significant differences were found (F_5,29 = 1.02, P=0.42) nor interaction between cultivar and treatment (F_5,29 = 0.74, P=0.59).

We averaged the final shoot dry weight for the different lighting treatments for both cultivars. The values in sequence from treatment T0 to T5 were recorded as follows: 3.91, 3.4, 3.6, 3.75, 3.87, and 3.83 g per plant ( Figure 6 ). The number of days of low DLI significantly affected the final shoot dry weight (F_5,29 = 3.29, P=0.017). Compared to the control treatment, plants growing with one day of low DLI showed a significant decrease in dry weight of 13% (P=0.022). The other significant differences were between plants growing under T1 and T4 (P=0.047). In addition, we did not find a significant interaction between light treatment and cultivar (F_5,24 = 0.33, P=0.88).

On the other hand, we also assessed the final fresh weight of shoots across both cultivars. The values were the following from the treatment with less amount of days with low DLI (T0 days) to the one with more days with low DLI (T5): 80, 73, 76.07, 78.85, 80.42, and 77.95 g per plant respectively ( Figure 7 ). Lighting treatments did not show significant differences (F_5,29 = 1.48, P=0.22), and there was no significant interaction between cultivar type and lighting treatment (F_5,24 = 0.49, P=0.77).

We calculated light use efficiency (LUE) for both cultivars for their whole growing cycle. The LUE for treatment from T0 to T5 were 0.56, 0.5, 0.51, 0.54, 0.55, and 0.54 g·mol^-1, respectively ( Figure 8 ). We did not find significant differences among lighting treatments on LUE (F_5,29 = 1.92, P=0.121) or interaction between cultivar and treatment (F_5,24 = 0.63, P=0.147). (F_5,24 = 0.69, P=0.63).

We counted the leaf number of three plants in the middle transversely of each tray, resulting in six plants counted (three per cultivar per treatment). The average leaf count per treatment, ranging from the lowest to the highest number of days with low DLI, was as follows: 19, 18.5, 18.44, 18.77, 19.72, and 19.55 leaves ( Figure 9 ). The analysis of variance (ANOVA) test showed a significant difference in leaf number (F_5,29 = 2.79, P=0.035). However, Tukey’s test did not show significant differences among the treatments. Additionally, we did not find an interaction between treatment and cultivar (F_5,24 = 0.45, P=0.8). Then, we used the Fisher DLS test since Tukey’s test did not show the significant differences announced by the ANOVA. Significant differences were found between plants growing under T1 in comparison to T4 and T5, among plants growing under T2 in comparison to T4 and T5. Finally, there was a significant difference in leaf number between plants growing under T3 and T4. However, no significant differences existed between the treatment and the control treatment or plants growing with zero days with low DLI or no DLI fluctuations.

In the first hypothetical scenario about energy needs, we assessed the energy required to generate the additional DLI needed to reach the target DLI of 15 mol·m^-2·d^-1 on days with low DLI, considering the number of low DLI days in our treatments over a 30-day period. The energy requirements were as follows: 113.6 MWh/ha when the additional DLI was 3.75 (1 low DLI days), 86.1 MWh/ha for an extra DLI of 2.5 mol·m^-²·d^-¹ (2 low DLI days), 68 MWh/ha for 1.87 mol·m^-²·d^-¹ (3 low DLI days), and 56.9 MWh/ha for 1.5 mol·m^-²·d^-¹ (1 low DLI days) ( Figure 10 ).

In the second hypothetical case of energy savings, we calculated the energy required to produce an additional DLI of 5 mol·m^-2·d^-1 (to reach the optimal 15 mol·m^-2·d^-1) on days with low DLI, assuming that sunlight provided a DLI of 10 mol·m^-2;·d^-1 over 30 days. The energy requirements were as follows: 152.7 MWh/ha for 2 days with low DLI after each day with high DLI, 171.8 MWh/ha for 3 days, 183.3 MWh/ha for 4 days, and 190.9 MWh/ha for 5 days of low DLI following a day with high DLI. Additionally, we assessed the energy needed to produce extra DLIs of 1.25, 2.5, 3.13, and 3.5 mol·m^-2;·d^-1 for 2, 3, 4, and 5 days of low DLI, respectively, following a day with high DLI. The resulting energy values were 38, 81.6, 114, and 133.3 MWh/ha for the additional DLI levels during the 30-day period ( Figure 11 ).

Plants growing under treatment1 day with high DLI followed only by 1 day with low DLI (T1) grew at a slower rate in comparison to plants growing with 4 (T4) and 5 days with low DLI (T5). We observed this while determining the time required for plants to reach 25%, 50%, and 75% coverage of a reference area. We assessed the plant growth rate using the PCS registered and calculated at various time points throughout the plant growth cycle. This means that when plants required more days to attain a specific coverage, they had a lower PCS, implying a slower growth rate. According to Klassen et al. (2003), plant growth is influenced by the quantity of light reaching the plant, which is directly tied to PCS and LUE (Legendre and van Iersel, 2021). Then, faster plant growth can be associated with a high PCS value. Plants with higher PCS intercept more light (Weaver and van Iersel, 2020), leading to more photosynthesis and biomass accumulation (Klassen et al., 2003). Said differences in biomass accumulation can be observed in Figure 6 , what explains the differences in canopy sizes and plant growth rate in our study ( Figures 1A–C ).

As indicated by Jayalath and van Iersel (2021), LUE was one factor that plays a role in plant growth. LUE is a measure of the plant’s efficiency in producing biomass with light reaching its canopy (Legendre and van Iersel, 2021). Then, differences in LUE between the plants growing under distinct light treatments are expected to contribute to differences in PCS or growing rates. However, those significant differences disappear when calculating the number of days needed to cover 100% ( Figure 1D ) of the reference area, a consequence of the overlapping canopies of the plants ( Figure 12 ).

LUE ( Figure 8 ) followed a similar trend to fresh and dry weight despite the absence of significant differences. This means that plants, regardless of the treatment, had the same efficiency in producing biomass with the incident light provided (Jayalath and van Iersel, 2021). The intensity of light may influence variations in LUE. Elevated light levels lead to a greater closure of the reaction centers within the photosystem II (PSII). With an increased closure of these centers, a higher proportion of absorbed light by the PSII light-harvesting complex remains unused for electron transport in PSII (van Iersel, 2017). Excess of absorbed light might be dissipated in different ways (Bassi and Dall’Osto, 2021). Consequently, when the photosynthetic machinery does not utilize light, more photons are being redirected to other routes and not used by photosynthesis, resulting in lower LUE values. On the contrary, reduced light levels produced an inverse effect and a higher LUE. The treatments in our study were made up of a combination of days with low DLI or low light intensity and days with high DLI or high light intensity, which would mean different LUE depending on the day. However, said variations in light intensities across treatments did not significantly affect the final cumulative utilization of absorbed light for electron transport or dissipation, as evidenced by the absence of significant differences in LUE.

Plants growing under 1 day with high DLI (22.53 mol·m^-2·d^-1) followed by 1 day with low DLI (7.53 mol·m^-2·d^-1) showed the lowest final dry shoot weight, meaning that the group of plants had a lower biomass accumulation ( Figure 6 ). Dry weight increased when the number of days with low DLI increased, and at the same time, the DLI for those days increased (11.25, 12.5, 13.17, and 13.52 mol·m^-2·d^-1, respectively). As far as we know, most studies that look into DLI and biomass accumulation have grown plants under fixed DLI, even if that DLI is composed of a different combination of PPFD and photoperiod (Elkins and van Iersel, 2020). Jayalath et al. (2024) recently experimented with DLI changes only for two consecutive days. In our study, our treatments had variable DLIs for more than two following days, making it more challenging to compare results. Yan et al. (2019b) showed an increasing direct relationship between leaf dry weight and DLIs between 5.04 to 15.12 mol·m^-2·d^-1. Furthermore, Pennisi et al. (2020) reported the highest value in shoot dry weight for plants growing under a DLI of 14.4 mol·m^-2·d^-1 made of a photoperiod of 16 hours and a PPFD of 250 µmol·m^-2·s^-1. Plants growing under lower DLIs (5.8, 8.6, 11.5 mol·m^-2·d^-1) and a higher DLI (17.3 mol·m^-2·d^-1) showed significantly lower final biomass values for lettuce. This relationship of higher dry weight with high light levels has been also observed on dwarf tomato (Lycopersicon esculentum) ‘Micro-Tom’ (Ke et al., 2023), common purslane (Portulaca oleracea) (Kudirka et al., 2023), petunia (Petunia × hybrida) ‘Wave Blue’, geranium (Pelargonium × hortorum) ‘Pinto Premium Orange Bicolor’, and coleus (Solenostemon scutellariodes) ‘Wizard Golden’ (Park and Runkle, 2018).

In addition, differences in photosynthesis rates and canopy size may explain differences in final shoot dry weight in our study. Zhou et al. (2020) observed an increase in photosynthesis rates of lettuce plants with increasing PPFD levels, ranging from 0 to 350 μmol·m^-2·s^-1 (12-hour photoperiod of about DLI 15·mol·m^-2·d^-1), reaching a plateau phase at approximately 500 µmol·m^-2·s^-1. In our study, plants exposed to 1 day of low DLI (T1) received around 104 µmol m^-2·s^-1, while those subjected to 5 days of low DLI (T5) received approximately 187 µmol·m^-2·s^-1. This difference in incident light could signify variations in photosynthesis rates, and greater photosynthesis is associated with increased biomass accumulation (Klassen et al., 2003). However, all treatments were also exposed to a day with a high DLI of approximately 14.4 mol·m^-2·d^-1, achieved by subjecting plants to PPFD of around 312 µmol·m^-2·s^-1. Specifically, plants growing with just 1 day with low DLI (T1) had a DLI of 22 mol·m^-2·d^-1 during half of the growing cycle. This might suggest that these plants could have exhibited a higher photosynthetic rate during days with elevated PPFDs than other treatments (Zhou et al., 2020). However, it is crucial to note that, for the other half of the growing cycle, plants under T1 experienced lower DLI values (approximately 7.53 mol·m^-2·d^-1 and PPFD of 104 µmol·m^-2·s^-1). The potential reduction in the photosynthetic rate at these lower light levels could account for a decrease in the final biomass. Zhou et al. (2020) showed reduced photosynthetic values when light intensity was low. Additionally, Pennisi et al. (2020) showed lower values of biomass accumulation for plants growing under a DLI of 8.6 mol·m^-2·d^-1 and PPFD of 150 µmol·m^-2·s^-1.

When the number of days with low DLI increased, the minimum DLI also increased. Plants growing for 4 and 5 days with low DLI had an average DLI of 13.17 and 13.52 mol·m^-2·d^-1, respectively. Hence, those plants were growing most of the days under a DLI that was close to the control treatment (14.88 mol·m^-2·d^-1), which is close to the ideal DLI for lettuce between 12 and 16mol·m^-2·d^-1 depending on the cultivar (Zhang et al., 2018; Yan et al., 2019a; Kelly et al., 2020; Pennisi et al., 2020; Modarelli et al., 2022). Consequently, the photosynthetic rates, LUE, and biomass accumulation of plants growing under T4 and T5 are higher (in comparison to T1) and similar to the control treatment. Jayalath et al. (2024) reported a decrease in final dry weight when the DLI fluctuation from one day to the next is above 15 mol·m^-2·d^-1. They also explained this partly due to a lower canopy rate expansion that decreased the incident light available for photosynthesis. In our study, this lower expansion rate on the canopy can be seen in the number of days plants needed to reach different sizes ( Figure 1 ).

These results corroborate the hypothesis the carryover effect exists, wherein higher photosynthesis rates during days with high DLI might compensate for the lower photosynthetic rates observed on days with lower DLI. They also indicate that plants experiencing more days with low DLI maintain similar photosynthetic rates because these lower DLI values are closely aligned with the optimal conditions required for lettuce growth.

Despite the significant differences observed in shoot dry weight results among the different lighting treatments, these differences were not found when measuring the final shoot fresh weight ( Figure 7 ). However, the consistent trend of increasing dry weight when the number of days with low DLI increased was also observed in the fresh weight. Changes in significant differences between shoot dry and fresh weight have been reported before. Khwankaew et al. (2018) reported mismatched significant differences between shoot dry and fresh weight on Ipomoea aquatica plants growing under LED light with different spectrum combinations. Fresh weight should be assessed promptly after sampling, or plant material should be stored in hermetically sealed recipients, as fresh plant material tends to lose water rapidly (Turner, 1981). Then, if fresh weight is not taken consistently across all the samples, some variation might be induced. In our study, we measured this parameter right after cutting the shoot from the root system. Therefore, we did not expect any variability in fresh weight induced by the sampling process. A possible reason for the differences in dry weight not being reflected in the final fresh weight could be related to the water status of the plants and their water retention capacity. For instance, plants growing under our control conditions showed a fresh weight of around 80 g per plant and a dry weight of approximately 4 g per plant. We can see that the final dry weight represented only 5% of the total fresh weight of the plants. This means that the significant differences observed in dry weight are lost due to the amount of water inside the leaf. Fresh weights have been reported in other studies to be more variable for showing differences among treatments compared to dry weights (Bashan and De-bashan, 2005; Huang et al., 2017).

No significant differences were found in final leaf area or SLA in our study ( Figures 2 , 3 ). Pennisi et al. (2020) observed an increase in leaf area for lettuce and basil with higher DLI. In contrast, the SLA for both species decreased as DLI increased. This reduction in SLA was attributed to the denser arrangement of mesophyll cells and the development of thicker and larger leaves. Similarly, Carotti et al. (2021) found that SLA decreased when lettuce plants grew under higher light intensities. However, these results were obtained for plants growing under constant DLI or light intensities. Bhuiyan and van Iersel (2021) grew lettuce plants under variable PPFD levels every 15 minutes (400/0, 360/40, 320/80, 280/120, 240/160, and 200/200 μmol·m^-2·s^-1), and observed that leaf area decreased when plants were subjected to high PPFD fluctuations but increased when the PPFD fluctuations were minimized between light levels. In contrast, SLA was higher under high PPFD fluctuations, whereas it was lower when the fluctuations were smaller at their respective light levels. They argue that extreme fluctuations in light prevent plants from reaching a steady state of photosynthesis. Consequently, this leads to reduced carbon gain by leaves, resulting in decreased dry weight and increased SLA. In our study, the DLI and PPFD fluctuations occurred over an extended period, suggesting that plants could reach a steady state of photosynthesis. This likely contributed to the production of similar biomass across treatments, resulting in comparable SLA values. Given the similar biomass production, we expected similar leaf area values regardless of our lighting treatments.

For plants growing under DLI fluctuations that happened during extended periods, specifically focusing on day-to-day variations, Jayalath et al. (2024) found that DLI fluctuations above 15 mol·m^-2·d^-1 significantly reduced leaf area. Lower PCS and LUE may explain these differences. A larger PCS captures more light, and higher LUE transforms that light more efficiently into biomass. While our study also observed differences in PCS initially, we found that PCSs became similar across all treatments once canopies started to overlap. This suggests that plants reached similar leaf area values with similar PCS and LUE. Nonetheless, our treatment involving a single day with low DLI exhibited a DLI fluctuation from one day to the next of approximately 15 mol·m^-2·d^-1, consistent with findings reported for lettuce (Jayalath et al., 2024). Despite this fluctuation, we did not observe significant differences in leaf area under the same lighting treatment.

Even when we found significant differences in leaf number among the treatments involving a DLI fluctuation, none showed a significant difference compared to the control treatment ( Figure 9 ). Changes in leaf number depending on different light levels have been reported before. Kang et al. (2013) found a higher leaf number for lettuce plants growing under high PPFD (290 μmol·m^-2·s^-1) levels when their photoperiod was 6/2 (light/dark) in 3 cycles per day. On the other hand, the least number of leaves was found when the plant grew at a low PPFD (200 μmol·m^-2·s^-1) with a photoperiod of 18/6 (light/dark). Also, Zervoudakis et al. (2012) found a significant reduction of leaf number in common sage (Salvia officinalis L.) when they grew under 25% ambient light compared to plants growing under full ambient light. These results suggest that plants growing under lower light levels generally produce fewer leaves. In our study, even plants that grew under different PPFD levels depending on the number of days with low DLI; they had, on average, the same light levels when taking into account the days with DLI and the days with low DLI. This might explain the lack of differences in leaf number compared to the control treatment (T0).

DLI fluctuations only from one day to another or “carryover” effect has been reported to potentially reduce energy cost related to supplemental lighting in green gouses set up (Jayalath et al., 2024). In this study, we tested how plants behave when said DLI fluctuations happen during multiple days. We proposed two hypothetical cases to assess the energy requirements for supplemental light when fluctuating DLI levels during various days. For the first case, we assumed that plants would be subjected to DLI of 11.25, 12.5, 13.13, and 13.5 mol·m^-2·d^-1 (provided by sunlight) during 2, 3, 4, and 5 days after a day with high DLI (22.5 mol·m^-2·d^-1) respectively. Under the conditions of our study, said plants did not require a DLI of 15 mol·m^-2·d^-1 after a day with high DLI to not show significant differences in plant growth compared to plants growing consistently under the optimal DLI of 15 mol·m^-2·d^-1. It has been showed that plants could be allowed to get a high DLI on sunny days, and this could compensate for a day with a low DLI immediately after, and this concept was called the carryover effect (Jayalath et al., 2024). Then, if we experience one day with a high daily light integral (DLI) of 22.5 mol·m^-2·d^-1, followed by subsequent days with lower DLIs of 11.25, 12.5, 13.13, or 13.5 mol·m^-2·d^-1 over the next 2, 3, 4, or 5 days respectively, all under natural light conditions (as described in our first hypothetical case), the use of energy for supplemental lighting ( Figure 10 ) to achieve an optimal DLI of 15 mol·m^-2·d^-1 may not be necessary. This is due to the carryover effect from the initial high DLI day.

In our second hypothetical case, we calculated the energy requirements assuming that days with a high DLI of 22.5 mol·m^-²·d^-¹ (from sunlight) would be followed by periods of 2, 3, 4, and 5 days with low DLI. Additionally, we assumed that these low DLI days would receive only 10 mol·m^-²·d^-¹ (also from sunlight). We estimated the energy needed to supplement an extra 5 mol·m^-2·d^-1 to achieve the optimal 15 mol·m^-²·d^-¹, as well as the energy required to generate additional DLI to reach 11.25, 12.5, 13.13, and 13.5 mol·m^-²·d^-¹ (depending on the number of consecutive days with low DLI) to take advantage of the ‘carryover’ effect. In our study, the contrast in energy requirements between achieving the optimal DLI and the energy to use the ‘carryover’ effect could indicate potential energy savings.

This research was conducted in a walk-in growth chamber (vertical farm) at the University of Georgia (College of Agriculture and Environmental Sciences, Department of Horticulture, Horticultural Physiology Laboratory), in Athens, GA, USA. The environmental conditions during the experiment, without distinguishing between light and dark periods, were: (mean ± standard deviation): temperature 24.35 ± 0.673°C, relative humidity 65.22% ± 7.56%, CO₂ concentration 847.64 ± 43.52 mg/L, and vapor pressure deficit 1.0201 ± 0.235 kPa.

Inside the growth chamber, there were three metal racks (2.4 m long × 0.6 m wide × 2.2 m high), each serving as a separate replication. Each rack had three horizontal shelves, and each shelf was divided into two equal parts vertically, resulting in six growing spaces per rack and 18 growing spaces in total, each with dimensions of 1.2 m long × 0.6 m wide × 0.6 m high. Each growing space was equipped with two LED fixtures (SPYDRx Plus with PhysioSpec indoor spectrum; Fluence Bioengineering, Austin, TX, USA) ( Supplementary Figure 1 ). Furthermore, four small fans (AD0412HB-C50; ADDA, Orange, CA, USA), evenly distributed, were positioned on the sides of each growing space to ensure proper lateral airflow.

We tested six lighting treatments randomly assigned to the growing spaces. These lighting treatments were controlled by a datalogger (CR6; Campbell Scientific, Logan, UT, USA) and six dimmable drivers (4009715; Intertek/Fluence, Arlington, VA, USA), with each driver responsible for controlling three growing spaces that shared the same lighting treatment, one space per rack or replication.

PPFD levels were assessed in the middle of every cultivation area using a quantum sensor (MQ-500; Apogee Instruments, Logan, UT, USA). Each treatment consisted of two DLI levels, called high and low, with a photoperiod of 20 hours. Plants were exposed to one day under high DLI, followed by varying numbers of days under low DLI, denoted as T0 (Control), T1, T2, T3, T4, and T5, indicating the respective number of days with low DLI in each treatment ( Figure 13 ). DLI and PPFD for each treatment are shown in Table 1 .

Ten-cm square plastic pots were filled with soilless substrate (Metro-Mix^® 830; SunGro Horticulture, Agawam, MA, USA) up to about 1 cm below the top rim. Three pelleted ‘Waldmann’s dark green’ or ‘Rouxai’ (Johnny’s Selected Seeds, Winslow, ME, USA) seeds were planted in each pot. The substrate was then covered with calcined clay or metakaolin (Turface MVP; Turface Athletics, Buffalo Grove, IL, USA) to avoid algae growth affecting PCS measurement. These pots were organized in trays, with fifteen pots arranged in a 5×3 configuration and placed in the designated growing spaces. Once the seeds germinated, a thinning process was carried out to keep only one seedling per pot. The plants received irrigation and nutrients through an ebb-and-flow subirrigation system, which delivered a 15N–2.2P–12.45K nutrient solution containing 100 mg·L^-1 of nitrogen using a water-soluble fertilizer (15-5-15 Ca-Mg Professional LX; J.R. Peters, Allentown, PA, USA).

The experimental setup followed a randomized split-block design with three blocks and six lighting treatments. Each experimental unit consisted of 15 plants. We used ANOVA by using an statistical software (R version 4.1.2; R Project for Statistical Computing, Vienna, Austria) to compare differences in the number of days plants need to reach specific sizes, differences in shoot-dry weight, pigment content, leaf area, relative leaf area, dry and fresh weights, leaf number and light use efficiency.