Supplemental Lighting Orientation and Red-to-blue Ratio of Light-emitting Diodes for Greenhouse Tomato Production

Abstract

Current greenhouse supplemental lighting technology uses broad-spectrum high-pressure sodium lamps (HPS) that, despite being an excellent luminous source, are not the most efficient light source for plant production. Specific light frequencies in the 400- to 700-nm range have been shown to affect photosynthesis more directly than other wavelengths (especially in the red and blue ranges). Light-emitting diodes (LEDs) could diminish lighting costs as a result of their high efficiency, lower operating temperatures, and wavelength specificity. LEDs can be selected to target the wavelengths used by plants, enabling growers to customize the light produced, to enable maximum plant production and limit wavelengths that do not significantly impact plant growth. In our experiment, hydroponically grown tomato plants (Solanum lycopersicum L.) were grown using a full factorial design with three light intensities (high: 135 μmol·m-2·s-1, medium: 115 μmol·m-2·s-1, and low: 100 μmol·m-2·s-1) at three red (661 nm) to blue (449 nm) ratio levels (5:1, 10:1, and 19:1). Secondary treatments for comparison were 100% HPS, 100% red LED light supplied from above the plant, 100% red LED light supplied below the plant, a 50%:50%LED:HPS mixture, and a control (no supplemental lighting). Both runs of the experiment lasted 120 days during the Summer-Fall 2011 and the Winter-Spring 2011-12. The highest biomass production (excluding fruit) occurred with the 19:1 ratio (red to blue) with increasing intensity resulting in more growth, whereas a higher fruit production was obtained using the 5:1 ratio. The highestmarketable fruit production (fruit over 90 g) was obtained with the 50%:50%LED:HPS followed by 5:1 high and 19:1 high. Consistently the 5:1 high performed well in every category. LEDs have been shown to be superior in fruit production over HPS alone, and LEDs can improve tomato fruit production when mixed with HPS. LEDs provide a promising mechanism to enhance greenhouse artificial lighting systems.

Full text

HORTSCIENCE 49(4):448–452. 2014.
Supplemental Lighting Orientation
and Red-to-blue Ratio of Light-emitting
Diodes for Greenhouse Tomato
Production
Paul Deram
3
Department of Bioresource Engineering, Macdonald Campus of McGill
University, 21111 Lakeshore Road, Ste-Anne-de-Bellevue, Quebec, H9X
3V9, Canada
Mark G. Lefsrud
1
and Vale
´rie Orsat
1,2
Department of Bioresource Engineering, Macdonald Campus of McGill
University
Additional index words. high-pressure sodium, intercanopy lighting, photosynthesis, supple-
mental lighting
Abstract. Current greenhouse supplemental lighting technology uses broad-spectrum
high-pressure sodium lamps (HPS) that, despite being an excellent luminous source, are
not the most efficient light source for plant production. Specific light frequencies in the
400- to 700-nm range have been shown to affect photosynthesis more directly than other
wavelengths (especially in the red and blue ranges). Light-emitting diodes (LEDs) could
diminish lighting costs as a result of their high efficiency, lower operating temperatures,
and wavelength specificity. LEDs can be selected to target the wavelengths used by plants,
enabling growers to customize the light produced, to enable maximum plant production
and limit wavelengths that do not significantly impact plant growth. In our experiment,
hydroponically grown tomato plants (Solanum lycopersicum L.) were grown using
a full factorial design with three light intensities (high: 135 mmol·m
L2
·s
L1
,medium:
115 mmol·m
L2
·s
L1
, and low: 100 mmol·m
L2
·s
L1
) at three red (661 nm) to blue (449 nm)
ratio levels (5:1, 10:1, and 19:1). Secondary treatments for comparison were 100%
HPS, 100% red LED light supplied from above the plant, 100% red LED light supplied
below the plant, a 50%:50% LED:HPS mixture, and a control (no supplemental lighting).
Both runs of the experiment lasted 120 days during the Summer–Fall 2011 and the Winter–
Spring 2011–12. The highest biomass production (excluding fruit) occurred with the 19:1
ratio (red to blue) with increasing intensity resulting in more growth, whereas a higher
fruit production was obtained using the 5:1 ratio. The highest marketable fruit production
(fruit over 90 g) was obtained with the 50%:50% LED:HPS followed by 5:1 high and 19:1
high. Consistently the 5:1 high performed well in every category. LEDs have been shown
to be superior in fruit production over HPS alone, and LEDs can improve tomato fruit
production when mixed with HPS. LEDs provide a promising mechanism to enhance
greenhouse artificial lighting systems.
Tomatoes are one of the most important
horticultural crops in the world. According to
Statistics Canada, tomato sales accounted for
close to 50% of the total fruit and vegetable
sales in the country in 2011 (Statistics Canada,
2011). Tomato sales in Canada reached
$496 million, an increase of 5.5% since 2010
(Statistics Canada, 2011). Greenhouseproduc-
tion exceeded field-grown tomatoes in 1997,
and in 2003, the greenhouse production was
more than nine times that of field-grown
tomatoes (Cook and Calvin, 2005). Today
there are 23 million m
2
allocated to green-
house fruit and vegetable production in
Canada with sales exceeding $1.1 billion
annually (Statistics Canada, 2011). Accord-
ing to Brazaityt_
e et al. (2009), in countries
of high latitude (e.g., Canada), tomatoes are
almost exclusively grown in greenhouses.
Light irradiance is the limiting factor for in-
creasing production in greenhouses, when
all other factors (temperature, nutrient levels,
and water availability) are adequately main-
tained (Smith, 1982). Ohashi-Kaneko et al.
(2007) state that artificial lighting permits
stable vegetable crop production no matter
the environmental conditions (with a favor-
able temperature in the greenhouse). Green-
houses in northern latitudes must compensate
for the attenuation in total light availability
(from prolonged winter with short daylight
hours), and supplemental artificial lighting
is required to maintain a consistent crop
yield throughout the Canadian winters. Con-
ventional greenhouse lighting systems use
broad-spectrum light sources such as HPS or
fluorescent lamps. LEDs have been shown to
decrease artificial lighting costs to less than
25% of the cost of traditional artificial light-
ing as a result of their 75% higher electrical
conversion efficiency (Gomez et al., 2013).
HPS lamps were tailored for human vision
and therefore are not ideally suited for plant
growth (Bula et al., 1991).
Earlier research has shown that the most
important wavelengths for photosynthesis are
in the blue and red wavelengths; peaks in
photosynthetic efficiency are found at 440
(blue), 620 (red), and 670 (red) nm (± 10 nm)
(McCree, 1972). The rapid improvement in
LED technology has been driving research in
plant production in recent years (Brazaity_
e
et al., 2009). Xiaoying et al. (2012) reported
new LEDs with a bandwidth of ± 15 nm,
permitting for much better focus on the most
efficient wavelengths for photosynthesis. The
wavelength specificity, small mass (less than
1 g each), small volume, relatively cool emit-
ting temperature, longevity of over 100,000 h,
and linear photon output are all characteris-
tics that make LEDs better suited for crop
production than earlier greenhouse lighting
systems (Folta et al., 2005).
Different wavelengths of the light spec-
trum have been found to have specific effects
on plant morphology, physiology, photosyn-
thesis efficacy, and flowering capabilities
(Menard et al., 2006). Although light is the
primary source of energy for photosynthesis,
it is also a vital regulator of many photo-
sensory circuits in plants. Blue light and red
light trigger different circuits and gene ex-
pression, which can have both positive and
negative effects on the growth and develop-
ment of plants (O’Carrigan et al., 2014). Thus,
research is focusing more on proper combi-
nations of light (Massa et al., 2008). Okamoto
et al. (1996) reported that both red and blue
light can be used by chlorophyll during photo-
synthesis and explained that blue light is
beneficial to plant morphology and overall
health. Blue light (450 nm) has been shown to
heavily suppress stem elongation in multiple
plant species (Okamoto et al., 1996) but also
increases plant biomass and fruit yield in
tomato and cucumber plants (Menard et al.,
2006). Herna´ndez and Kubota (2012) reported
that at the tomato seedling stage, red light is
sufficient to grow tomato seedlings and that
the addition of blue light is unnecessary. Over-
all, blue light is not as effective for photosyn-
thesis as red light, because it inhibits leaf
growth by reducing cell expansion and re-
duces the total amount of chlorophyll in the
leaves (Goto, 2003). As a result of this lower
efficiency, researchers tend to undervalue the
use of blue light and not consider it in high
proportions for plant growth (Goto, 2003).
A lack of blue light has been shown to have
adverse effects on plant morphology: low
number of chloroplasts, lower thickness of
cell walls, and low spongy mesophyll tissues
Received for publication 30 Aug. 2013. Accepted
for publication 8 Jan. 2014.
This work was supported by General Electric Light-
ing Solutions Canada and the Natural Sciences and
Engineering Research Council of Canada (NSERC)
(CRDPJ project no. 418919-11). We thank Claire
Boivin for her tomato growing expertise.
1
Associate Professors.
2
Chair.
3
To whom reprint requests should be addressed;
e-mail paulderam@gmail.com.
448 HORTSCIENCE VOL. 49(4) APRIL 2014

(Goto, 2003). Boccalandro et al. (2012) show
that blue light promotes leaf expansion, sto-
matal opening, and chloroplast relocation
through phototropin activation in Arabidop-
sis and other higher plants. Blue light was
shown to stimulate stomata opening and in-
crease the rate of photosynthesis by up to
30% in some species (Menard et al., 2006).
Blue light from LED sources was also shown
to be highly efficient at protecting crop plants
from many pathogens as well as increasing
antioxidants and osmoprotectant produc-
tion (Kim et al., 2013). Plant response to
light from the red and the blue spectra has
been documented extensively (Menard et al.,
2006). The ratio of blue to red light is shown
to be the most important factor when using
LEDs, because having both blue light and red
light increases plant biomass growth and fruit
production by over 20% when compared with
having only one of the wavelengths (Brazaityt_
e
et al., 2009; Goto, 2003; Lefsrud et al., 2008;
Xiaoying et al., 2011, 2012).
This experiment used LED prototypes from
General Electric Lighting Solutions Canada
(Lachine, Quebec, Canada), where the over-
all goal of the project was to determine which
ratio and intensity combinations best suit to-
mato plants for fruit production under green-
house applications. Previous studies performed
at McGill University (Martineau et al., 2012;
Schwalb, 2013) have shown that the ratios
of red to blue light used for this experiment
increased plant photosynthetic rate at the
seedling stage and this research would con-
firm that these same ratios result in improved
fruit and biomass production for mature
plants.
Materials and Methods
Tomato plants (S. esculentum cv. Trust
were grafted onto S.hirsutum cv. Maxifort)
(Ontario Plant Propagation, St. Thomas,
Ontario, Canada) were obtained 55 d after
seeding and were hydroponically grown in
a greenhouse setting following similar cul-
ture methods to the industrial practices of
Savoura (Les Serres Du St-Laurent Inc.,
Portneuf, Quebec, Canada) using a modified
full-strength Hoagland nutrient solution
(Savoura proprietary information) (Deram,
2013).
The air temperature ranged from 21.8 ±
2.4 C during the day and 16.9 ± 3.4 C
during the night (average of both experimen-
tal runs). The LED light arrays were turned
on from 0600 to 2200 HR for a total photo-
period of 16 h. The nutrient solution was
supplied through drip irrigation (14257 Bio-
floral, Montreal, Quebec, Canada) for 3 min
out of every 20 min (19 L·h
–1
) for a total ir-
rigation of 6 L per plant per day (based
on 1 mL per 1 J of irradiance). The plants
were pollinated by hand, shaking the flowers
with a cotton-tipped swab three times a week.
A central computer in the greenhouse controlled
the lights, irrigation, and ventilation coupled
with a misting system for cooling. Relative
humidity was monitored but was not controlled.
Average daily relative humidity reached a
high of 82% ± 5% and a low of 30% ± 10%
(average of both experimental runs).
The LED light arrays were prototypes
(1.78 m ·8cm·2 cm) with wavelengths
of 449 nm for blue light and 665 nm for red
light. Twelve ratio and intensity-specific LED
prototypes were used, nine of which were
used to set up a full factorial with three ratios
of red to blue (5:1, 10:1, and 19:1) and three
intensities (high: 135 mmol·m
–2
·s
–1
, medium:
115 mmol·m
–2
·s
–1
, and low: 100 mmol·m
–2
·s
–1
)
measured and confirmed at the beginning and
end of each replicate with less than 2% drift
during replicates. The light sensor used was
the underwater spherical quantum sensor
(LI-193; Li-COR, Lincoln, NE); an explana-
tion on why this sensor was chosen can be
found in the ‘‘Results and Discussion’’ sec-
tion. The remaining prototypes, used for further
experimental treatments, were two 100% red
light arrays (red top and red bottom at the
high intensity of 135 mmol·m
–2
·s
–1
) and a 10:1
medium (for use in the 50%:50% LED:HPS,
combined irradiance at 135 mmol·m
–2
·s
–1
).
Three HPS lamps (Philips SON-T Master
600W, Amsterdam, The Netherlands) were
used, one for the 50%:50% LED:HPS and two
for the 100% HPS section (135 mmol·m
–2
·s
–1
).
The control section did not have any supple-
mental lighting. The red bottom treatment
lights were placed at the bottom of the section
(at the level of the first leaves) and shone
upward into the plant canopy, whereas the red
top LEDs were placed at the top, similar to all
other treatments.
A greenhouse section (7.6 m ·12 m set
in a north–south orientation) with long wire
mesh tables (1.2 m in height) was used. The
tables were separated into 14 sections (2 m ·
1 m), each one for a specific light treatment.
Three lamps were installed per section, mak-
ing space for two rows of plants between
them. The lamps were set at an intercanopy
level (10 cm below the top of the plants) and
raised weekly to maintain relative height,
except for the red bottom treatment, where
the lamps were kept at the height of the first
leaves.Eachsectionwassurroundedbya
double layer of 2.44 m high (8 feet) 80%
shadecloth (8MK808; Harnois, St-Thomas,
Quebec, Canada), preventing 96% of light
from passing through. A single layer of
2.44 m (8 feet) 60% shadecloth (8MK608;
Harnois) was installed above all of the sections
during the first replicate (summer months) to
simulate winter conditions.
Two experimental replicates, summer (July
to Oct. 2011) and winter (Jan. to Apr. 2012),
were performed with two harvest times, one
at 70 d and one at 120 d. Half the plants were
randomly selected and harvested at 70 d, and
the remaining plants continued the experi-
ment for the full 120 d. A 2-month period was
allotted between the experimental replicates
to minimize the risk of pathogens carrying
over. The location of each light treatment in
the greenhouse was randomly allotted at the
beginning of each experimental replicate.
Eight plants (six during the second experi-
mental replicate, because the Savoura plants
were larger and the same plant density was
kept) were placed in each section, in two
rows, between the lamps. The individual
plants were measured as individual replicates
between seasons. The change in plant number
(from eight to six) did not have any effect on
the statistical analysis of the reported values.
At harvest, fresh weight was determined
by separating and weighing both aerial plant
biomass (the rooting system was discarded)
and fruit biomass. All fruit greater than 2 g
were counted and weighed (fruit under 2 g
were included as plant biomass). Ripe fruit
was harvested throughout the experiment at
the first observed red pigmentation (consid-
ered ripe), counted, and individually weighed.
Plants were pruned according to Savoura’s
methods (every 2 weeks, Savoura unpublished
data) with fresh and dry biomass measured
for each plant. The marketable fruit value
was set at 90 g and is an internal standard for
Savoura signifying marketable fruit. The
fresh biomass harvested (aerial and fruit)
was dried according to the ASABE (2007)
standard with a temperature of 65 C for no
less than 72 h and subsequently weighed.
Main effects were analyzed by one-way
analysis of variance with the standard Tukey-
Kramer test (a#0.05) applied to determine
significant differences between treatments.
The relationship between the experiment’s
dependent variables and treatments was de-
termined by regression analysis. As a result
of the strong constraints of the experiment
(large plant variability, limited space, limited
time, large number of different factors need-
ing testing), the power of the statistical anal-
ysis was calculated post hoc to be between
10% and 38% depending on the quantity being
tested.
Results and Discussion
The point source nature of the LED arrays
makes it difficult to measure the light in-
tensity with conventional light-measuring
devices. Typical spectroradiometers and light-
measuring devices are built to measure light
levels from the sun or other light sources that
disperse light in every direction (such as HPS
lighting) and therefore are not as compatible
with the LED arrays. A series of irradiance
tests were performed with a range of different
light measuring devices to find the most ap-
propriate solution for measuring the LED
point source arrays. All light measurements
were taken at horizontal, vertical, and 45at
three different heights and three points along
the length of each light fixture. The first test
was performed using a pyranometer (total
solar radiation) (MP-100; Apogee Instruments,
Logan UT) and a quantum meter (measures
only the photosynthetically active radiation)
(MQ-100; Apogee Instruments). It was found
that the light measurement from both devices
was variable (30% differences in readings)
under weighing blue light and over weighing
red light (Apogee Instruments Inc., 2012a,
2012b); they may not properly report wave-
lengths outside the 460- to 660-nm range, and
therefore they could not be used for the exper-
iment. A spectroradiometer (BLACK-Comet
HORTSCIENCE VOL. 49(4) APRIL 2014 449

Concave Grating Spectrometer; StellarNet
Inc., Tampa, FL) was then tested. The spec-
troradiometer was much more reliable than
the quantum meter and pyranometer, but the
same problem with variability resulting from
spatial position was found. The spectroradi-
ometer was designed for conventional over-
head light sources and only records light
coming from directly above the sensor. The
field of view of a typical spectroradiometer
was shown to be 10(MacArthur et al., 2007).
With adjustments to the positioning (directly
facing the LED for the measurements), a light
map could be created using the spectroradi-
ometer, but it was still not as reliable as
expected for measurements of the light re-
ceived by the plants.
An underwater spherical quantum sensor
(LI-193; Li-COR) was tested. The spherical
quantum sensor was developed to capture
light dispersion in underwater biological ex-
periments by measuring the photon flux com-
ing from all directions. This sensor was chosen
as a result of its ability to record the photon
flux coming from all directions, which was
well suited for measuring the LED lighting in
this experiment (because the LED arrays do
not result in a wide dispersion of light). The
measurements of the light from all angles
above a theoretical horizontal surface resulted
in no significant reduction in reported values.
This underwater spherical light sensor was
well adapted to the LED fixtures and gave us
much more accurate measurements of the
lighting with less than 5% variability be-
tween measurements. However, it should be
noted that when the handle of the light sensor
was parallel to the light source (or above the
theoretical horizontal surface), the light mea-
surement was 20% lower than when placed
perpendicular to the light source. Despite the
20% loss in reading light perpendicular to the
sensor, the underwater quantum sensor was
found to be the most accurate sensor for this
research and a well-adapted sensor to LED
testing. No statistical difference was measured
between the initial and final light readings or
between the intensities of treatments at the
same intensity level. The average daily light
integral (solar and artificial) was 21 mol·d
–1
±
37 mol·d
–1
for the summer run and 42 mol·d
–1
±
20 mol·d
–1
for the winter run.
Results from the experiment revealed that
the five highest number of fruit-producing
light treatments were 5:1 high (385 fruit), 5:1
medium (358 fruit), 5:1 low (341 fruit), 19:1
high (315 fruit), and 100% LED (310 fruit).
The control was statistically significantly dif-
ferent from all five of the high producing
treatments. The average number of fruit per
plant is shown in Figure 1. The 5:1 ratio was
the highest producing ratio from this exper-
iment with all three intensities leading the
ranking for most fruit produced. The 5:1 ratio
being the highest fruit-producing ratio in this
experiment agrees with research from Goto
(2003) and Menard et al. (2006), who found
that blue light promoted flowering and fruit-
ing, but contradicts Miyashita et al. (1995),
who reported that red light, and not blue light,
would increase flowering. With regard to
fruit mass, 5:1 high (27.2 kg), 5:1 medium
(25.8 kg), 19:1 high (24 kg), 50%:50%
LED:HPS (23.3 kg), and red top (23.2 kg)
were the five highest treatments. No statisti-
cal differences were measured between any
of the treatments with the exception of the
control, which was statistically different from
all five of the high producing treatments. The
average fruit mass per plant is shown in
Figure 2. The highest five treatments for
marketable number of fruit were 50%:50%
LED:HPS (with a total of 118 fruit over 90 g),
red top (115 fruit), 5:1 medium (113 fruit),
19:1 high (109 fruit), and 5:1 high (104 fruit).
The highest five total marketable fruit mass
were the same treatments but in a different
order: 50%:50% LED:HPS with 17.4 kg fol-
lowed by 5:1 high with 16.3 kg, 19:1 high with
16.2 kg, red top with 15.8 kg, and 5:1 medium
with 15.6 kg. Fifty percent:50% LED:HPS
was one of the highest producing treatments
for marketable fruit and agrees with research
from Menard et al. (2006) that shows that
supplementing blue and red light to HPS
creates much higher production than with
HPS alone. The differences between the high-
est treatments were not statistically different,
but all highest producing treatments were
statistically different from the control. The
average marketable fruit mass per plant is
shown in Figure 3.
The high irradiance level was the highest
producer for all ratios, both for vegetative bio-
mass and for fruit production. As expected from
the literature, higher intensities bring forth
more production (McAvoy, 1984; Tennessen
et al., 1994) with all ratios producing more
under higher irradiance levels.
The five highest treatments for both fresh
biomass and dry biomass were 19:1 high and
red bottom and included the other 19:1 treat-
ments (medium and low) and the 100% red
top. For fresh weight, they were 19:1 high had
37.8 kg followed by red bottom with 37.4 kg,
50%:50% LED:HPS with 34.5 kg, and 10:1
medium with 33.9 kg. The average fresh
biomass per plant is shown in Figure 4. For
dry mass, 19:1 high had 2.8 kg followed by
Fig. 1. Average number of fruit per plant per light treatment. The data were separated into three categories.
Best: over 19 fruit (in black). Average: from seven to 19 fruit (in dark gray). Poor: under seven
marketable fruit (in light gray). Statistically significant differences were observed between the best
category and the poor category.
Fig. 2. Average fruit mass per plant (kg) per light treatment. The data were separated into three categories.
Best: over 1.9 kg of fruit (in black). Average: from 0.2 kg to 1.9 kg of fruit (in dark gray). Poor: under
0.2 kg of fruit (in light gray). Statistically significant differences were observed between the best
category and the poor category.
450 HORTSCIENCE VOL. 49(4) APRIL 2014

red bottom at 2.7 kg, 5:1 high at 2.5 kg, and
19:1 medium at 2.5 kg. This result is sup-
ported by the literature that shows that red
light produces more biomass (Brown et al.,
1995; Hoenecke et al., 1992), because all
treatments with high red light were found
consistently to be the highest. The largest
fresh and dry vegetative biomass occurred
with the 19:1 high LED.
A fruit mass-to-plant fresh biomass ratio
was calculated as shown in Figure 5. The five
highest ratio treatments were 5:1 low (0.87
fruit to biomass), 5:1 high (0.82), 5:1 medium
(0.79), red top (0.75), and 10:1 high (0.71).
The control ratio was 0.12. All 5:1 intensity
levels as well as 10:1 high, red top, and
50%:50% LED:HPS were found to be statis-
tically different from the control treatment.
The 5:1 ratio treatments (high, medium, and
low) were shown to be statistically different
from treatments 10:1 low and 19:1 low (the
two lowest after control).
For the marketable fruit mass-to-plant
fresh biomass ratio, the five highest treat-
ments were red top (0.49), 5:1 high (0.49),
50%:50% LED:HPS (0.48), 5:1 low (0.49),
and 5:1 medium (0.46). The control ratio
was measured at 0.09. Statistically significant
differences were observed between the red
top and 5:1 high treatments with the control
only. Fruit-to-biomass ratio provides infor-
mation on the amount of vegetative biomass
the plant produced relative to fruit produc-
tion. This ratio is important, because it shows
how efficient the plant was at producing fruit
from the resources given. A higher value is
more desirable, because it shows that a higher
proportion of the plant growth was turned
into the marketable product.
The red top and the red bottom treatments
were directly compared with each other. The
major difference between the two treatments
was the setup with the red bottom treatment
lights placed at the bottom of the section and
shone upward into the plant canopy, whereas
the red top LEDs were placed at the top,
similar to all other treatments. The red bottom
had slightly less light levels than the red top.
Overall, the red bottom section slightly under-
performedwhencomparedwiththeredtopin
fruiting categories but created more biomass.
No statistical differences were measured be-
tween these two treatments. The differences
between the two treatments are more pro-
nounced toward the end of the run. These
differences could be explained by the lower
light levels or because the plants in the red
bottom section grew more side stems and
suckers close to the LED arrays and did not
grow as tall (observed but not measured).
However, no conclusions can be drawn from
the statistics, and the production results were
similar for both treatments. This result could
suggest that the light coming from below was
at least as beneficial as the light coming from
above (Moss, 1964).
From the regression analysis of the facto-
rial experiment, it can be reported that higher
levels of light increased production, whereas
increased levels of red light (relative to blue)
resulted in less fruit:
Number of Fruit ¼23:9 + Intensity0:12
+ Light Ratio4:5ðÞR2¼0:68

[1]
where intensity is the amount of light in the
treatment (100 mmol·m
–2
·s
–1
for low, 115
mmol·m
–2
·s
–1
for medium, 135 mmol·m
–2
·s
–1
for high) and light ratio is the ratio of red light
to blue light (1 for 5:1, 2 for 10:1, and 3 for
19:1, chosen in sequence).
The regression analysis also indicated
that increased levels of light resulted
in increased biomass, and increased red
Fig. 3. Average marketable fruit mass per plant (kg) per light treatment. No statistically significant
differences were observed.
Fig. 4. Average fresh weight of plant biomass (excluding fruit) per light treatment (kg). No statistically
significant differences were observed.
Fig. 5. Total fruit mass-to-plant biomass ratio per light treatment. Statistically significant differences
between the best category and the poor category; also between the highest two (5:1 low and 5:1 high)
and the treatments under 0.4 (10:1 low and 19:1 low).
HORTSCIENCE VOL. 49(4) APRIL 2014 451

light (relative to blue) resulted in more
biomass:
Vegetative Fresh Biomass in gramsðÞ
¼1982:2 + Intensity5:1
+ Light Ratio426:3R
2¼0:64

[2]
where intensity is the amount of light in the
treatment, in mmol·m
–2
·s
–1
(100 for low, 115
for medium, 135 for high) and light ratio is
the ratio of red light to blue light (1 for 5:1, 2
for 10:1, and 3 for 19:1).
Conclusion
Overall, it was shown that the highest
producing LED treatments (5:1 high, 5:1 me-
dium, 19:1 high, and 5:1 low) and 50%:50%
LED:HPS consistently outperformed the HPS
treatment alone, and thus these treatments
can be considered an improvement over tra-
ditional HPS lighting for greenhouses. Al-
though no significant differences were found
between the higher performing treatments,
5:1 high consistently performed well in every
category. As expected from the literature, it
was found that an increase in light intensity
brought higher production of both fruit mass
and plant biomass, and an increase in red light
increased biomass production and slightly
lowered the amount of fruit production.
Literature Cited
Apogee Instruments Inc. 2012a. Apogee pyran-
ometer technical information. 1 Nov. 2012.
<http://www.apogeeinstruments.co.uk/apogee-
pyranometer-technical-information/>.
Apogee Instruments Inc. 2012b. Apogee quan-
tum sensor technical information. 1 Nov.
2012. <http://www.apogeeinstruments.co.uk/
apogee-instruments-quantum-sensor-technical-
information/>.
ASABE. 2007. Moisture measurement—Forages.
American Society of Agricultural and Biolog-
ical Engineers, Standard S358.2. ASABE, St.
Joseph, MI.
Boccalandro, H., C. Giordano, E. Ploschuk,
P. Piccoli, R. Bottini, and J. Casal. 2012. Photo-
tropins but not cryptochromes mediate the blue
light-specific promotion of stomatal conduc-
tance, while both enhance photosynthesis and
transpiration under full sunlight. Plant Physiol.
158:1475–1484.
Brazaityt_
e, A., P. Duchovskis, A. Urbonavicˇi
ut_
e,
S. Samuoliene, J. Jankauskien _
e, V. Kazenas,
A. Kasiulevicˇi
ut _
e-Bonak_
er _
e, Z. Bliznikas,
A. Novie`kovas, K. Breiv _
e, and A. Z
ˇukauskas.
2009. After-effect of light emitting diodes
lighting on tomato growth and yield in green-
house. Sodininkyst_
e ir Darz
ˇininkyst_
e 28:115–
126.
Brown, C., A. Schuerger, and J. Sager. 1995.
Growth and photomorphogenesis of pepper
plants under red light-emitting diodes with
supplemental blue or far-red lighting. J. Amer.
Soc. Hort. Sci. 120:808–813.
Bula, R., R. Morrow, T. Tibbitts, and D. Barta.
1991. Light-emitting diodes as a radiation
source for plants. HortScience 26:203–205.
Cook, R. and L. Calvin. 2005. Greenhouse
toma toes change the dynamics of the North
American fresh tomato industry. USDA ERR-
2:1–81.
Deram, P. 2013. Light-emitting-diode (LED) light-
ing for greenhouse tomato production. Master’s
thesis, McGill University, Montreal, Quebec,
Canada.
Folta,K.,L.Koss,R.McMorrow,H.Kim,
J. Kenitz, R. Wheeler, and J. Sager. 2005. Design
and fabrication of adjustable red-green-blue
LED light arrays for plant research. BMC Plant
Biol. 5.
Gomez, C., R. Morrow, C.B. Bourget, G. Massa,
and C. Mitchell. 2013. Comparison of interca-
nopy light-emitting diode towers and overhead
high pressure sodium lamps for supplemental
lighting of greenhouse grown tomatoes. Hort-
Technology 23:93–98.
Goto, E. 2003. Effects of light quality on growth of
crop plants under artificial lighting. Environ.
Control in Biol. 41:121–132.
Herna´ndez, R. and C. Kubota. 2012. Tomato
seedling growth and morphological responses
to supplemental LED lighting red:blue ratios
under varied daily solar light integrals. Acta
Hort. 956:187–194.
Hoenecke, M.E., R.J. Bula, and T.W. Tibbitts.
1992. Importance of ‘blue’ photon levels for
lettuce seedlings grown under red-light-emitting
diodes. HortScience 27:427–430.
Kim, K., H.S. Kook, Y.J. Jang, W.H. Lee, S.
Kamala-Kannan, J.C. Chae, and K.J. Lee.
2013. The effect of blue-light-emitting diodes
on antioxidant properties and resistance to
Botrytis cinerea in tomato. J. Plant Path. Micro-
bio. 4:203.
Lefsrud, M., D. Kopsell, and C. Sams. 2008.
Irradiance from distinct wavelength light-
emitting diodes affect secondary metabolites
in kale. HortScience 43:2243–2244.
MacArthur, A.A., C. MacLellan, and T.J. Malthus.
2007. Determining the FOV and directional
response of field spectroradiometers. The 5th
EARSeL Workshop on Imaging Spectroscopy.
Martineau, V., M. Lefsrud, M.T. Naznin, and D.A.
Kopsell. 2012. Comparison of light emitting
diode and high pressure sodium light treatments
for hydroponics growth of Boston lettuce.
HortScience 47:477–482.
Massa, G., H. Kim, R. Wheeler, and C. Mitchell.
2008. Plant productivity in response to LED
lighting. HortScience 43:1951–1956.
McAvoy, R. 1984. The use of high pressure sodium
lights in greenhouse tomato crop production.
Acta Hort. 148:877–888.
McCree, K. 1972. The action spectrum, absorption
and quantum yield of photosynthesis in crop
plants. Agr. Meteorol. 9:191–216.
Menard, C., T. Hovi, A. Gosselin, and M. Dorais.
2006. Developmental and physiological re-
sponses of tomato and cucumber to additional
blue light. Acta Hort. 711:291–296.
Miyashita, Y., Y. Kitaya, and T. Kozai. 1995.
Effects of red and far-red light on the growth
and morphology of potato plantlets in-vitro:
Using light emitting diode as a light source for
micropropagation. Acta Hort. 393:189–194.
Moss, D.N. 1964. Optimum lighting of leaves.
Crop Sci. 4:131–136.
O’Carrigan, A., E. Hinde, N. Lu, X.Q. Xu,
H. Duan, G. Huang, M. Mak, B. Bellotti, and
Z.H. Chen. 2014. Effects of light irradiance on
stomatal regulation and growth of tomato.
Environ. Exp. Bot. 98:65–73.
Ohashi-Kaneko, K., M. Takase, N. Kon, K. Fujiwara,
and K. Kurata. 2007. Effect of light quality on
growth and vegetable quality in leaf lettuce,
spinach and komatsuna. Envrion. Control Biol.
45:189–198.
Okamoto, K.T.Y. and S. Takita. 1996. Develop-
ment of plant growth apparatus using blue and
red LED as artificial light source. Acta Hort.
440:111–116.
Schwalb, M. 2013. Measuring the short-term plant
photosynthetic response to varying light quality
using light emitting diodes (LEDs). Master’s
thesis, McGill University, Montreal, Quebec,
Canada.
Smith, H. 1982. Light quality, photoperception,
and plant strategy. Annu. Rev. Plant Physiol.
33:481–518.
Statistics Canada. 2011. Greenhouse, sod and
nursery industries catalogue # 22-202-x.
Tennessen, D., E. Singsaas, and T. Sharkley. 1994.
Light-emitting diodes as a light source for photo-
synthesis research. Photosynth. Res. 39:85–92.
Xiaoying, L., G. ShiRong, X. ZhiGang, and
J. XueLei. 2011. Regulation of chloroplast
ultrastructure, cross-section anatomy of leaves,
and morphology of stomata of cherry tomato by
different light irradiations of light-emitting di-
odes. HortScience 46:217–221.
Xiaoying, L., G. ShiRong, C. TaoTao, X. ZhiGang,
and J. XueLei. 2012. Regulation of the growth
and photosynthesis of cherry tomato seedlings
by different light irradiations of light emitting
diodes (LED). Afr. J. Biotechnol. 11:6169–
6177.
452 HORTSCIENCE VOL. 49(4) APRIL 2014