**4. Materials and Methods**

The experimental work consisted of 2 replications over time. Cucumber (*Cucumis sativus* cv. Diva) seeds were germinated in total darkness at 32 ◦C. Once germinated, seeds were transplanted into 4 in. pots containing UC mix (1/<sup>3</sup> peat moss, <sup>1</sup>/<sup>3</sup> redwood sawdust, <sup>1</sup>/<sup>3</sup> fine sand), covered with an additional 100 cm<sup>3</sup> of UC mix, and randomly distributed into their light treatment chambers. Plants were irrigated with <sup>1</sup> 2 strength Hoagland's solution every third day for the first two weeks, then daily thereafter [54]. Leaf photosynthetic rate, stomatal conductance, and fluorescence measurements were obtained during both replications. Morphological measurements were made four weeks after transplant.

In both replications, plants were grown in chambers 61 cm wide, 122 cm long, and 90 cm tall. An 8 in. duct fan exhausted air from the chambers so that the average temperature was 23.0 ± 0.2 ◦C when the lights were on and 20.9 ± 0.2 ◦C when the lights were off.

Each chamber was illuminated with lamps consisting of various light-emitting diode (LED) bars (Demegrow, Inc., Sacramento, CA, USA) specifically designed to provide a custom spectrum in each chamber. Fixtures consisted of different combinations of diodes emitting far-red, red, green, or blue light, with peak intensities at wavelengths of 744 nm, 661 nm, 521 nm, and 460 nm, respectively. Spectra of the resulting lamp systems were measured with a JAZ spectrometer (model: JAZ spectrometer, Ocean Optics, Largo, FL, USA). The full width at half maximum for each peak was 21.8, 20.7, 33.6, and 21.6 nm for far-red, red, green, and blue peaks, respectively (Figure 6). Green light in the RGB and RGB + FR treatments came from 15,000 K white LEDs, which is why the green peak is broader in these treatments than other treatments containing green light. Each fixture installation was configured so that all had comparable photosynthetic photon flux density (PPFD) levels of 120 µmol photons m−<sup>2</sup> s −1 in each chamber. This was achieved by raising or lowering the lamp array in each chamber and averaging measurements over a 45-point grid.

**Figure 6.** Light treatment spectra: monochromatic blue (B), monochromatic green (G), monochromatic red (R), green–blue (GB), red–blue (RB), red–green (RG), red–green–blue (RGB), and red–green–blue with far red (RGB + FR). **Figure 6.** Light treatment spectra: monochromatic blue (B), monochromatic green (G), monochromatic red (R), green–blue (GB), red–blue (RB), red–green (RG), red–green–blue (RGB), and red–green–blue with far red (RGB + FR).

**Table 5.** Yield photon flux (YPF) and photostationary state of phytochrome (PSS) for each light treatment. **Treatment YPF PSS µmol m**<sup>−</sup><sup>2</sup> **s** <sup>−</sup><sup>1</sup> **Pfr:Ptotal** B 88 0.51 G 94 0.83 GB 90 0.62 R 114 0.89 RB 98 0.86 RG 106 0.88 Where both colors were present, the intensity of blue and red are roughly 1:1, blue and green are 2:1, and red and green are 2:1; actual percentages of total light as in Table 4. Since the energy of far-red light does not contribute to photosynthetic photon flux density (PPFD), the RGB and RGB + FR treatments have roughly the same PPFD and light ratios between 400 and 700 nm, but 18% of all incident irradiation between 400 to 800 nm in the RGB + FR treatment was in the far-red (700 to 800 nm) region. One percent of incident photons were in the far-red region in the RGB treatment, while all other treatments had negligible levels of far-red light. In addition to the traditional color quantification (red 600–700 nm, green 500–600 nm, and blue 400–500 nm), the light is reported based on the quantity from each 'color' of LED bar. This was determined by only powering LED bars of a given light color and measuring PPFD, then calculating the percentage of total PPFD from that bar color (Table 4).

RGB 102 0.86 RGB + FR 104 0.76 Yield photon flux (YPF) was calculated for each light treatment according to [9] by multiplying relative quantum efficiency at a given wavelength with the photon flux at that wavelength, then integrating from 300 to 800 nm (Table 5). The YPF model adjusts PPFD based on the likelihood that a photon of a given wavelength will be absorbed and the likelihood that the energy will be used for photosynthesis once absorbed.


**Table 5.** Yield photon flux (YPF) and photostationary state of phytochrome (PSS) for each light treatment.

Finally, photostationary state of phytochrome (PSS), an estimate of active phytochrome as a portion of total phytochrome, was calculated using

$$PSS = \left(\sum\_{300}^{800} \mathcal{N}\_{\lambda} \sigma\_{r\_{\lambda}}\right) / \left(\sum\_{300}^{800} \mathcal{N}\_{\lambda} \sigma\_{r\_{\lambda}} + \sum\_{300}^{800} \mathcal{N}\_{\lambda} \sigma\_{fr\_{\lambda}}\right) \tag{1}$$

as reported in Table 2 [9].

Equation (1) gives *PSS* where *N* is incident photon flux at a given wavelength (*λ*), *σ<sup>r</sup>* is the photochemical cross section of P<sup>r</sup> (the red-absorbing, inactive form of phytochrome) at *λ*, and *σfr* is the photochemical cross section of Pfr (the far-red-absorbing, active form of phytochrome) at *λ*.

Photosynthesis was measured in two ways. First, using a LI-6400 with a clear-top chamber (LI-COR Biosciences, Lincoln, NE, USA), net photosynthesis (A) and stomatal conductance (gs) were measured at an ambient CO<sup>2</sup> concentration (Ca) of 400 ppm and a leaf temperature of 25 ◦C, illuminated by treatment light spectra at ambient intensity.

Second, A vs. C<sup>c</sup> response curves, the net photosynthesis rate obtained under varying concentrations of CO<sup>2</sup> in the chloroplast (Cc) under saturating light, were measured using the LI-6400 portable photosynthesis system with a 6400-40 leaf chamber fluorometer attachment (LI-COR Biosciences, Lincoln, NE, USA) in order to gain insight about possible molecular adaptations to the light environment. Measurements were taken at external CO<sup>2</sup> concentrations of 400, 300, 250, 200, 150, 100, 50, 400, 500, 600, 850, and 1000 ppm in that order. Following initial fluorescence measurements, the plants had half an hour to adapt to light at 1000 µmol m−<sup>2</sup> s −1 . Each CO<sup>2</sup> concentration was held for two to four minutes at a flowrate of 300 µmol s−<sup>1</sup> with leaf temperature set to 25 ◦C while the plant was at room temperature (23 to 26 ◦C). The first true leaf, unshaded by neighboring leaves, was measured.

Plants were first dark adapted for half an hour before initial measurements. Fluorescence measurements were taken on the dark-adapted leaves, before acclimating to the light for a half hour. The minimum chlorophyll fluorescence for dark-adapted leaves (Fo), maximum light- and dark-adapted chlorophyll fluorescence (Fm' and Fm, respectively), and steady state light-adapted chlorophyll fluorescence (F') were measured. The maximum quantum efficiency of PSII (Fv/Fm), the relative PSII operating efficiency (ΦPSII), the coefficient of photochemical quenching (qp), the quantum yield of non-light-induced nonphotochemical quenching (ΦNPQ), and the quantum yield of light-induced nonphotochemical quenching (ΦNO) were calculated according to [37]. The fraction of oxidized plastoquinone, qL, was calculated according to [55]. Due to the difficulties of measuring Fo', the minimal fluorescence of a light-adapted leaf, it was calculated using the equation Fo' = Fo/[(Fv/Fm) + (Fo/Fm')] where F<sup>o</sup> is the minimal fluorescence of a dark-adapted leaf, F<sup>m</sup> is the maximal fluorescence from a dark-adapted leaf, Fm' is the maximal fluorescence from a light-adapted leaf, and F<sup>v</sup> is the difference between F<sup>m</sup> and F<sup>o</sup> [40].

A vs. C<sup>c</sup> curve fitting was done using SAS Studio 3.8 software via the NLIN procedure, a procedure for fitting nonlinear models, using Equations (2)–(4) [56,57]. Typically, these model fittings involve 3 segments representing photosynthesis as limited either by the maximum ribulose-1,5-bisphosphate (RuBP) carboxylation rate Equation (2), the RuBP regeneration rate Equation (3), or the triose phosphate utilization (TPU) rate. However, our data suggest that TPU was not a limiting factor and so we fitted to only the Rubisco-limiting (Equation (2)) and the RuBP-limiting curves (Equation (3)). The equation for calculating the concentration of CO<sup>2</sup> at Rubisco, Cc, has also been included (Equation (4)).

$$A = V\_{c\max} \left[ \frac{\mathcal{C}\_{\mathcal{C}} - \Gamma^\*}{\mathcal{C}\_{\mathcal{C}} + K\_{\mathcal{C}} \left( 1 + \frac{O}{K\_o} \right)} \right] - R\_d \tag{2}$$

$$A = J \left[ \frac{\mathbb{C}\_{\mathfrak{c}} - \Gamma^\*}{4\mathbb{C}\_{\mathfrak{c}} + 8\Gamma^\*} \right] - R\_d \tag{3}$$

$$\mathcal{C}\_{\mathcal{C}} = \mathcal{C}\_{i} - \frac{A}{g\_{m}P\_{atm}} \tag{4}$$

where *C<sup>i</sup>* is the intercellular concentration of CO2, *C<sup>c</sup>* is the concentration of CO<sup>2</sup> at Rubisco, *A* is net CO<sup>2</sup> assimilation, *Vcmax* is maximum carboxylation rate of Rubisco, Γ \* is the point at which oxygenation is twice the rate of carboxylation (CO<sup>2</sup> uptake equals CO<sup>2</sup> photorespiratory release), *K<sup>o</sup>* is the inhibition constant of Rubisco for oxygen, *K<sup>c</sup>* is the Michaelis–Menten constant of Rubisco for CO2, *O* is the partial pressure of O<sup>2</sup> at Rubisco, *R<sup>d</sup>* is non-photorespiratory CO<sup>2</sup> release, *J* is the rate of electron transport, *Patm* is atmospheric pressure, and *g<sup>m</sup>* is mesophyll conductance.

Due to the difficulty of accurately determining g<sup>m</sup> due to the method of data collection and initial fittings determining that g<sup>m</sup> did not significantly differ between any treatments, the overall average value of 2.12 µmol m−<sup>2</sup> s <sup>−</sup><sup>1</sup> Pa−<sup>1</sup> was used [58]. Likewise, since estimates of R<sup>d</sup> did not significantly differ between treatments, an average value of 2.71 µmol CO<sup>2</sup> m−<sup>2</sup> s <sup>−</sup><sup>1</sup> was used.

All plants' shoots were severed at the substrate surface and weighed for fresh weight, separated into leaf blades and all other material (stem, petioles, cotyledons, and leaves < 2 cm<sup>2</sup> ), oven dried for 72 h at 60 ◦C, and weighed to obtain dry weights. Stem diameter was measured with an electronic caliper just below the cotyledons with the caliper arm held parallel to the cotyledons to give a consistent measurement for seedlings with non-circular stem cross-sections. Stem height was measured from the point at which the shoot was severed to the base of the apical meristem to the nearest millimeter. The two largest leaves on each plant had length, width, petiole length, and leaf blade area measured. Additionally, total leaf area was measured using a LI-COR 3100 leaf area meter (LI-COR Biosciences, Lincoln, NE, USA). Finally, stomatal density was measured by taking a 1 cm by 2 cm section of leaf tissue adjacent to the midrib approximately halfway between leaf tip and leaf blade base and applying clear nail polish [59].

All means separations were determined using SAS Studio software 3.8 (SAS Institute Inc., Cary, NC, USA). Data from the two replications were treated as separate blocks with means separation analyzed by a Tukey–Kramer HSD (*p* = 0.05).
