Green butterhead lettuce (Lactuca sativa ‘Rex’) and cucumber (Cucumis sativa ‘Straight Eight’) seeds were direct seeded then thinned for uniformity after emergence leaving four plants per root module. Two supplemental studies with lettuce cultivars ‘Waldmann’s Dark Green’ (replicated once) and ‘Green Salad Bowl’ (replicated three times) were conducted following the same methods as ‘Rex’ lettuce to determine the consistency of the responses across cultivars. Root modules measured 20 × 18 × 13 cm (4680 cm³) and contained a 1:1 ratio of peat and vermiculite by volume with 0.75 g per L wetting agent (AquaGro G), 1 g per L hydrated lime, and 1 g per L of uniformly mixed slow-release fertilizer (Nutricote total; 18-6-8, N-P-K, type 70). Planted root modules were randomly placed into the 12 treatment chambers. Each chamber had dimensions of 20 × 23 × 30 cm (L × W × H, 13800 cm³) with gloss white walls. Fans provided an air velocity of 0.5 m s^-1 at the top of the canopy. Root modules were watered to 10% excess as needed with dilute fertilizer solution at a rate of 100 mg N per L (21-5-20; Peters Excel; EC = 1 dS m^-1) and allowed to passively drain. Type-E Thermocouples connected to a data logger (CR1000, Campbell Scientific, Logan UT) continuously monitored ambient air temperature. Day/night temperature averaged 23.4 ± 1.2/20.9 ± 0.8 ˚C, with less than a 1˚C variation among chambers. CO₂ concentration was continuously monitored and was identical for all treatments and varied over time between 450 to 500 ppm.

The system included 12 chambers with four percentages of FR (Figures 1, 2) at three levels of ePPFD (4×3 = 12 treatments). Light was provided over a 16 h photoperiod. In the cucumber study the three levels of ePPFD were 50, 200 and 500 µmol m^-2 s^-1 (Daily light integral, DLI: 2.88, 11.52 and 28.8 mol m^-2 d^-1). In the lettuce studies, the lowest ePPFD treatment (50 µmol m^-2 s^-1) was increased to 100 µmol m^-2 s^-1 (DLI: 5.76 mol m^-2 d^-1) because 50 µmol m^-2 s^-1 was below the light compensation point.

Treatments were developed using white (2700 K and 6500 K), red (peak about 658 nm) and far-red (peak 726 nm) LEDs (Lumileds LLC, San Jose, CA) to output 20% blue, 40% green, and 40% red photons as a proportion of PAR. Treatments reduced traditionally defined PAR while increasing FR photon flux density. Percent FR, calculated by dividing the FR photon flux density by the ePPFD and multiplying by 100, was 2, 9, 17 and 33% (Figures 1, 2). Spectral photon distributions of the treatments were measured before each replicate study with a spectroradiometer (model PS-300; Apogee Instruments, Logan UT). ePPFD was measured with an ePAR sensor (MQ-610, Apogee Instruments, Logan UT) at the top of the plant canopy throughout the study, and chambers were adjusted to maintain ePPFD at ± 5% of the setpoint. Root modules were rotated every three days to increase the uniformity of treatments within the chamber.

Plants were harvested at canopy closure. This occurred 17 or 18 days after emergence in lettuce and 9, 10, or 11 days after emergence in cucumber. At harvest, stem length, leaf area, and fresh mass were measured. Leaf area was measured using a leaf area meter (LI-3000; LI-COR, Lincoln NE). Petiole length was measured in the cucumber study and leaf length and leaf width were measured in the lettuce study. Leaves and stems (and cotyledons in cucumber) were separated, and dry mass (DM) of each organ was measured after the tissue was dried at 80 ˚C for 48 hours. Percent stem mass was calculated by dividing stem dry mass by total shoot dry mass, multiplied by 100; roots were not included. Specific leaf area (SLA, m²_LA kg DM_leaf^-1) was calculated by dividing the leaf area by the leaf dry mass.

The lettuce study was replicated three times and the cucumber study was replicated five times. Every replicate in time contained four plants in each treatment. All data was analyzed using R statistical software (R Foundation for Statistical Computing; Vienna, Austria). ePPFD was treated as a categorical variable with three levels for all analysis. Percent FR was treated as either a continuous variable or a categorical variable depending on the general response of the parameter. Responses that were linear with increasing percent FR were analyzed with linear mixed-effects regression analysis (percent FR treated as a continuous variable) and non-linear responses were analyzed with two-way ANOVA analysis (percent FR treated as a categorical variable). Replicates in time were treated as random factors in all statistical analyses. Significant effects were determined at the p < 0.05 level. Detailed statistical analysis for specific parameters are provided below.

The effect of percent FR on total shoot dry mass was non-linear in both lettuce and cucumber (Figures 5A, B), thus two-way ANOVA analysis and Tukey’s post hoc test were used (with percent FR as a categorical variable) to separate the significant differences between treatments. The smaller values and variance of total dry mass in the low ePPFD treatment resulted in reduced statistical power to determine significant difference between percent FR treatments. Therefore, the data of each replicate in time was normalized to its respective 2% FR (no added FR) control, and this normalized response was analyzed with 1) two-way ANOVA analysis/Tukey’s post hoc test to determine significant differences between treatments and 2) a student’s t-test to determine if the normalized response was statistically different from one, which represented the response of no added FR (Figures 5C, D).

Higher ePPFD produced more biomass in both species. In lettuce, percent FR and ePPFD interacted to predict biomass accumulation, where increasing percent FR increased dry mass at high ePPFD, but decreased dry mass at low ePPFD. This decrease occurred in both the 100 and 200 µmol m^-2 s^-1 ePPFD treatments. There was also an interaction between percent FR and ePPFD in cucumber, where the highest percent FR at the lowest ePPFD had no effect compared to higher ePPFD (Figure 5D), but in general the interaction was far less pronounced when compared to lettuce. In general, the response of cucumber shoot dry mass showed no effect of increasing percent FR from 2 to 9%, an increase between 9 and 17% FR, and little response between 17 and 33% FR.

The effect of percent FR and ePPFD on total leaf area was similar to the effects on shoot dry mass (Figure 6). This indicates that the effects of percent FR on biomass accumulation in Figure 5 were likely driven by the ability of the plant to capture photons (leaf area is highly correlated with growth before canopy closure). Due to the similar non-linear response of leaf area to percent FR and ePPFD, this response was analyzed with the same methods as total dry mass.

Because leaf area/leaf expansion significantly contributes to the rate of biomass accumulation, it is valuable to better understand the interaction between the FR and ePPFD in controlling this response in lettuce. Leaf lengths increased with percent FR at all three levels of ePPFD, while the response of leaf width was similar to the response of leaf area (Figures S2A, B). This resulted in longer, narrower leaves rather than rounder leaves at higher percent FR (Figure S2C). Although an increase in leaf length is generally expected under shaded conditions, the change in leaf shape does not fully explain the interaction between percent FR and ePPFD in predicting total leaf area in lettuce.

Specific leaf area (SLA) is a measure of leaf area given the biomass partitioning to the leaves, and it is widely reported to increase with shade. Compared to the responses of dry mass and total leaf area, the response of SLA to percent FR appeared linear. Thus, this parameter was analyzed with linear mixed-effects regression with percent FR as a continuous variable and ePPFD as a categorical variable. Increasing the percent FR and decreasing the ePPFD increased SLA in both species (Figures 6A, B).

Because ePPFD had a large effect on SLA (the lowest ePPFD induced a 2.5 to 3 times higher SLA than the highest ePPFD in both species), the best way to determine if there is an interaction between percent FR and ePPFD was to normalize the response. This is because the non-normalized response could indicate an interaction even if the effect of percent FR induced the same fractional change at all three levels of ePPFD. Figures 6C, D normalizes the SLA response to the average 2% FR response at each ePPFD. Linear mixed-effects regression analysis was then performed on this normalized data, which found no interaction between ePPFD and percent FR for either species, although the interaction approached significance in cucumber (p = 0.11; Figure 6D).

Increases in SLA in response to both increasing percent FR and decreasing ePPFD has often been observed in previous studies. However, because there was no interaction between the two treatments to predict SLA, this parameter (SLA) does not explain the interaction between the two treatments to predict total leaf area in lettuce (Figures 7A, C).

Increased shade is often reported to increase stem lengths, and this is often associated with an increase in biomass partitioning to stems. Percent stem mass, a metric that assesses partitioning to stems, is calculated by dividing stem dry mass by total shoot dry mass and multiplying by 100. Subtracting percent stem mass from 100 is the shoot biomass allocation to leaves (and cotyledons for cucumber).

In lettuce, the effects of percent FR and ePPFD on percent stem mass were analyzed with linear mixed-effects regression following the same procedure as SLA (Figure 6A). In cucumber, the effect of percent FR on percent stem mass was non-linear. Using two-way ANOVA analysis and Tukey’s post hoc test, it was determined that there was no difference between 2 and 9% FR for the three levels of ePPFD (Figure 8B, dotted lines), but percent stem mass was significantly higher at the lowest ePPFD for both the 2% and 9% FR treatments (Figure 8B). Since there was no difference between 2 and 9% FR, to simplify the model, the 2% FR treatments were removed for analysis and the remaining data (9, 17 and 33% FR) were analyzed with linear mixed-effects regression analysis, similar to lettuce.

Increasing FR from 9 to 33% increased biomass partitioning to stems at all three levels of ePPFD in cucumber. Additionally, there was an effect of ePPFD on biomass partitioning to the stems with an increase in percent stem mass at lowest ePPFD. Finally, there was an interaction between the two treatments, meaning that there was a difference in the slopes of the lines among the levels of ePPFDs (p < 0.001). Unlike SLA, this interaction was determined for non-normalized data, although normalized data also showed an interaction (this is the case for both percent stem mass and stem length in both species). The interaction between percent FR and ePPFD resulted in a decreased response of percent stem mass to percent FR at the at the highest ePPFD (Figure 8B).

In lettuce, there was also an effect of percent FR, ePPFD and an interaction between the two treatments in the prediction of percent stem mass (p < 0.001). The most striking impact of the interaction between percent FR and ePPFD in lettuce was that increasing percent FR from 2 to 33% increased percent stem mass by 2.5-fold at the lowest ePPFD but had no effect at the highest ePPFD (Figure 8A). Similar to the effect of ePPFD at the lowest percent FR in cucumber (dotted lines in Figure 8B), percent stem mass in lettuce tended to be higher at 100 µmol m^-2 s^-1 compared to the higher two intensities at 2% FR. But, this was not statistically significant at the p = 0.05 level using Tukey’s post hoc test from a two-way ANOVA (p = 0.09 and 0.06 for 100 compared to 200 and 100 compared to 500 µmol m^-2 s^-1, respectively).

The striking interaction between percent FR and ePPFD in the prediction of percent stem mass in lettuce explains the interaction between the two treatments in the prediction of total plant leaf area. As percent stem mass (shoot biomass allocation to stems) increases, percent leaf mass (shoot biomass allocation to leaves) decreases. Because SLA increased with increasing percent FR at all levels of ePPFD (Figure 6C), and because biomass partitioning to leaves was unaffected by percent FR at the highest ePPFD, overall total leaf area was increased with increasing percent FR. However, at the lowest ePPFD, even though increasing FR increased SLA by about 30%, biomass accumulation was redirected away from the leaves to a significant enough degree to reduce total leaf area.

Due to thermal reversion of phytochrome, ePPFD and FR have been combined into one unifying model called the cellular model, which accounts for both increased thermal reversion at higher temperatures and the larger contribution of thermal reversion on phytochrome dynamics at lower ePPFD. This model was unable to explain the response of percent stem mass in cucumber (Figure S3), indicating a contribution of other factors (e.g. cryptochrome and/or photosynthate signals, see discussion).

The responses of lettuce and cucumber stem lengths to percent FR and ePPFD were similar to the response of percent stem mass, and thus the data for each species was analyzed using the same methods (Figure 8). There was an effect of percent FR, ePPFD and an interaction between the two (p < 0.001), in the prediction of stem length in both species. In cucumber, the tallest plants at the highest percent FR occurred at the middle ePPFD instead of the lowest ePPFD. A similar response was observed in longest petiole length (Figure S4). This is likely a direct result of lower photosynthesis resulting in reduced growth rate.

We investigated the same five parameters (biomass accumulation, total leaf area, specific leaf area, percent stem mass, and stem length) in two additional cultivars of lettuce: ‘Waldmann’s Dark Green’ and ‘Green Salad Bowl’ (Figures S5–S8). Because ‘Green salad bowl’ was replicated the same number of times as ‘Rex’ (n = 3), it was analyzed following the same statistical approach. ‘Waldmann’s dark green’ was only replicated once, and thus, no statistical analysis was performed, but the trends are consistent.

In all three cultivars, increasing percent FR resulted in an increase in biomass accumulation at the high ePPFD, but a decrease in accumulation at the low ePPFD. At the intermediate ePPFD, ‘Waldmann’s dark green’ showed a general downward trend with increasing percent FR, while ‘Green salad bowl’ showed an increase followed by a decrease (Figure S5). As expected, the response of total leaf area was similar to the response of total shoot dry mass (Figures S1, S6). In general, the responses of SLA were similar between all three cultivars (Figure S7), although FR appeared to have a larger effect at the high ePPFD in ‘Waldmann’s dark green’, but this was likely due to reduced growth in the no added FR treatment. Additionally, at the lowest ePPFD in ‘Green salad bowl’ increasing FR appeared to increase SLA up to a threshold, then decrease it. Finally, there was a clear interaction between percent FR and ePPFD in predicting percent stem mass and stem length in all three cultivars (Figure S8). One notable difference from ‘Rex’ is that increasing percent FR had a small effect on percent stem mass and stem length at the high ePPFD in ‘Green salad bowl’. These results indicate that the general interaction between ePPFD and percent FR on the prediction of leaf area and biomass accumulation may be observed across many lettuce cultivars.

Ballaré et al. (1987) demonstrated that in competitive light environments FR can increase in the horizontal direction prior to a decrease in PAR. This is notable because it means that in certain circumstances there can be increases in FR in the natural environment even at high PAR. These FR signals are the first indication of oncoming shade and they have been shown to increase internode elongation prior to the subsequent decrease in PAR in Sinapis alba and Datura ferox (Ballaré et al., 1987). These two species have an upright plant architecture, similar to cucumber, in comparison to the rosette architecture of lettuce and Arabidopsis. Perhaps plants with an upright architecture, like cucumber, Sinapis alba, and Datura ferox, will always elongate in response to increasing FR no matter the ePPFD. We did not investigate the highest intensities that occur in the middle of a summer day, above 2000 µmol m^-2 s^-1, but perhaps there would have been no effect of FR on stem elongation in cucumber at this intensity. It is interesting that cucumber increased leaf area at all three levels of ePPFD, meaning that in the natural environment FR would be likely to increase leaf expansion from both neighboring plants at high ePPFD and from true shade at low ePPFD.

For lettuce, and perhaps other species with a rosette architecture, the increase in FR from neighboring vegetation could actually increase leaf expansion in ecosystems with more light. This is further notable because in this study we substituted PAR for FR, but in the natural environment FR can increase with no change in PAR (Ballaré et al., 1987), and this could lead to additional increases in growth through photosynthesis. However, the increase in FR in most shaded environments would be much more likely induce decreases in leaf expansion, as the plants attempt to reach out of the shade.

The lowest two FR treatments (2 and 9%) are probably not entirely comparable to sunlight. Leaf expansion was not significantly different between 17 and 33% FR; thus, it could be argued that no such increases in leaf expansion would occur in the natural environment. However, FR LEDs have a peak that closely matches the peak of the Pfr form of phytochrome. Therefore, lower percentages of FR from FR LEDs may be comparable to higher percentages of FR in sunlight, but such comparisons are difficult.

Often, studies investigating plant responses to FR typically add FR photons on top of a constant intensity of PAR instead of decreasing the photon intensity within PAR while increasing FR. Maintaining a constant ePPFD removes the potential impact of photosynthesis on leaf expansion. Assuming PAR and FR photons both drive photosynthesis with equal efficiencies, then the studies performed here (substituting PAR with FR) would at most result in equal leaf-level photosynthesis. However, this 1:1 assumption is likely not the case. Zhen and Bugbee (2020a) noted that the photosynthetic relationship of PAR photons and FR photons began to deviate from a 1:1 equivalency when FR was added in excess of 40% of PAR (29% FR based on the definition used here). Furthermore, Zhen et al. (2019) showed that photons at 752 nm did not increase the quantum yield of PSII (the fraction of absorbed photons used for photochemistry), indicating that photons at 752 nm did not contribute to photosynthesis (the next lowest wavelength measured was 731 nm). These considerations indicate that 1) extending the definition of photosynthetic photons out to 750 nm may overestimate the photosynthetic response, and b) 33% FR used in this study would have deviated from a 1:1 photosynthetic relationship FR and PAR. Perhaps this also explains the saturating effect of FR on shoot dry mass and leaf area (Figures 5, 7).

Despite this limitation of our experiment, both of these factors indicate that the higher FR treatments in this study would have resulted in lower leaf level photosynthesis, which further highlights the effect increasing percent FR can have on leaf expansion and thus photon capture in cucumber and lettuce (at high photon fluxes in the latter).

The results presented here have substantial implications for the horticultural industry – especially in sole-source lighting environments, where electric lighting from LEDs has been rapidly expanding. Providing photons for crop production is expensive, especially when considering the electrical power required to operate the LEDs. FR LEDs are among the highest efficacy LEDs (µmol of photons per joule input electrical energy), primarily because of the lower energy of the photons (Kusuma et al., 2020). Interest in FR for horticultural environments has expanded in the past decade where studies have often found that adding FR increases leaf area or plant diameter in lettuce, resulting in an increase in fresh and dry mass (Lee et al., 2016; Meng and Runkle, 2019; Zou et al., 2019; Zhen and Bugbee, 2020b; Legendre and van Iersel, 2021). But, many of these studies supplemented FR rather than substituting PAR photons for FR. The addition of FR LEDs in fixtures increases costs for two reasons: 1) FR LEDs are in lower demand, meaning they are more expensive, and 2) increasing the total photon output of a fixture increases the power requirement. Thus, substitution of PAR with FR is much more reasonable for practical applications. Zhen and Bugbee (2020b) found that substituting 14% of PAR with FR at an ePPFD of 350 µmol m^-2 s^-1 resulted in similar quantum yields for CO₂ fixation in ‘Waldmann’s dark green’ lettuce, but increased biomass accumulation by about 30% via an increase in photon capture (leaf expansion). We observed a similar response with ‘Waldmann’s dark green’ in our experimental chambers; adding 9, 17 and 33% FR at an ePPFD of 500 µmol m^-2 s^-1 resulted in 31 to 33% increase in biomass accumulation. However, at lower ePPFD (100 and 200 µmol m^-2 s^-1), increasing percent FR resulted in a decrease in biomass accumulation (Figures S5-S8).

In another similar study, Legendre and van Iersel (2021) concluded that FR photons were between 57 to 183% more effective at increasing photon capture than PAR photons in ‘Green salad bowl’ lettuce and were thus 93 to 162% more effective at increasing biomass accumulation. Interestingly, they observed no interaction between PPFD and FR when PPFD varied between 111 and 245 µmol m^-2 s^-1. In this cultivar, we observed an increase in leaf expansion and shoot dry mass with an increase in percent FR at 500 µmol m^-2 s^-1, but a decrease at an ePPFD of 100 µmol m^-2 s^-1 and little effect at 200 µmol m^-2 s^-1 (Figures S5, S6). Interestingly at 200 µmol m^-2 s^-1 shoot dry mass was higher at 9% FR compared to 2% FR (Figure S5F), but at higher levels of FR, dry mass decreased. The contradictions between our study and Legendre and van Iersel (2021) are difficult to parse, as many of the environmental conditions were similar.

Producers can increase yields of lettuce by including FR LEDs in lighting fixtures because FR photons will increase leaf expansion and photon capture, while decreasing electrical input (depending on operating conditions and other LEDs in the fixture). However, caution must be taken as increases will only occur at high enough ePPFD. For cucumber, including more FR appears to be beneficial no matter the ePPFD, but the increase in leaf expansion appears to be accompanied by increases in stem elongation. This may be manageable in greenhouse environments but may be undesirable in sole-source lighting environments. Additionally, the added FR may only benefit cucumber plants when they are young.

Based on current definitions, FR photons are not considered in the calculation of photosynthetic photon efficacy, which is the flux of photosynthetic photons (µmol s^-1) divided by the input power (W), resulting in µmol per J. Currently, photosynthetic photons are considered to only be those with wavelengths between 400 to 700 nm based on studies by McCree (1972a); McCree (1972b). Because this definition excludes photons between 700 to 750 nm, the benefits of photons from FR LEDs, both on leaf expansion and photosynthesis are excluded. These definitions ought to be reconsidered (Zhen et al., 2021), but at the same time, FR must be used with caution, as we have shown it can be detrimental in some conditions.