Effects of Seeding Pattern and Cultivar on Productivity of Baby Spinach (Spinacia oleracea) Grown Hydroponically in Deep-Water Culture - MDPI
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
horticulturae
Article
Effects of Seeding Pattern and Cultivar on
Productivity of Baby Spinach (Spinacia oleracea)
Grown Hydroponically in Deep-Water Culture
Daniel B. Janeczko and Michael B. Timmons *
Department of Biological and Environmental Engineering, Cornell University, Ithaca, NY 14853, USA;
dbj42@cornell.edu
* Correspondence: mbt3@cornell.edu
Received: 1 December 2018; Accepted: 21 February 2019; Published: 28 February 2019
Abstract: Baby spinach (Spinacia oleracea) was grown in a bench-scale deep-water culture (DWC)
system in expanded polystyrene (EPS) plug trays. Two experiments were performed. In the first,
different seeding patterns, [1-2-1-2 . . . ] or [3-0-3-0 . . . ] seeds per sequential cell, at the same overall
density per tray, were compared to evaluate the potential of an EPS tray designed with fewer cells, but
sown with more seeds per cell (to preserve canopy density). Using such a flat would lower growing
substrate requirements. Seeding in the [3-0-3-0 . . . ] pattern reduced seed germination, but only by
5%. Harvested fresh weight was also less numerically in the [3-0-3-0 . . . ] pattern but not statistically.
The second experiment observed cultivars Carmel, Seaside and Space grown concurrently. Carmel
had the highest germination, nearly 100%, which was significantly greater than Seaside but not Space.
Germination for Space was not significantly different from that of Seaside. Carmel also had the
highest harvested fresh weight but was not significantly different from Space; both Carmel and Space
produced significantly more harvested fresh weight than Seaside.
Keywords: hydroponics; deep-water culture; baby spinach; Carmel; Seaside; pericarp; dibbler;
density; Spinacia oleracea
1. Introduction
Hydroponics is the soilless culture of plants grown in a nutrient solution. Within hydroponics,
several different techniques are employed to expose plant roots to a nutrient solution. The Deep-Water
Culture (DWC) method is characterized by use of a trough or tub filled with aerated nutrient solution
to which a floating raft is added that holds the plants with their roots suspended vertically in the water
column [1].
Hydroponic greenhouse crop production has the potential to reduce dependence on imported
crops and to increase food security and safety while simultaneously enhancing environmental quality
and energy/resource efficiency of food production systems [2,3]. Baby spinach is an attractive choice
for hydroponic production since it is popular as a fresh vegetable and has relatively high concentrations
of bioactive compounds including ascorbic acid, carotenoids, and flavonoids that contribute to its high
nutritional value, despite its low caloric content [4]. Growing this crop in the northeast is particularly
attractive since 86% of US fresh spinach production in 2017 (53,030 acres total) occurred in Arizona or
California [5].
Expanded polystyrene (EPS) flats with varying cell volume and spacing are commonly used for
baby leaf production in DWC systems. Typical cell volumes can range from 2 or 3 cm3 to 10 cm3 . Cell
volume impacted tomato, eggplant, and pepper seedling fresh weight when grown from seed [6].
Planting (canopy) density was less predictive of seedling fresh weight (within the range considered:
Horticulturae 2019, 5, 20; doi:10.3390/horticulturae5010020 www.mdpi.com/journal/horticulturae
Horticulturae 2019, 5, 20 2 of 12
Horticulturae 2019, 5, x FOR PEER REVIEW 2 of 13
had greater
94–1125 yield per
plants/m 2 ) [6].
unitBaby
area romaine
and resulted in plants
lettuce (Lactucawith longer
sativa) but fewer
seeded leavesdensities
at higher compared in to
EPSlower-
flats
density arrangements
had greater yield per unit [7]. This
area demonstrated
and resulted in that planting
plants withdensity
longer butand fewer
available cellcompared
leaves volume can to
impact the morphological development of baby leaf crops. However, the
lower-density arrangements [7]. This demonstrated that planting density and available cell volume effect of proportional cell
volume
can impactavailable to spinach seedlings
the morphological development and the higher
of baby leafdensity of plants in
crops. However, thethe early
effect of stages of growth
proportional cell
associated with sowing
volume available to spinachmultiple seeds and
seedlings per cell
the have
higher not been thoroughly
density of plants ininvestigated.
the early stages of growth
The soilless
associated medium
with sowing in which
multiple seeds
seeds perarecellsown
have is notoften
beenathoroughly
large fixed investigated.
material cost of each crop.
WhenThesowing an off-the-shelf
soilless medium in which EPS flat (generally
seeds re-used
are sown for 30
is often successive
a large fixedplantings),
material cost withofsphagnum
each crop.
peat moss germination mix and spinach seed at a density of 1.5 seeds per
When sowing an off-the-shelf EPS flat (generally re-used for 30 successive plantings), with sphagnum cell, the proportional cost
of growing
peat medium can
moss germination mixbeand
estimated
spinachtoseed be over 50% of of
at a density the1.5total
seedsmaterial
per cell,input costs (Figurecost
the proportional 1).
Employing
of growing amedium
flat withcan fewer cells, but to
be estimated sowing
be overmore50% seeds
of the in total
each material
cell (thusinput
preserving the same
costs (Figure 1).
canopy density), could meaningfully reduce material expenses. While
Employing a flat with fewer cells, but sowing more seeds in each cell (thus preserving the same the density of the canopy
would
canopytheoretically
density), could remain unchanged,
meaningfully reduceit ismaterial
unclearexpenses.
how packing Whilemore seeds into
the density of theeach cell affects
canopy would
germination and growth
theoretically remain throughout
unchanged, the short
it is unclear howcrop cycle.more
packing It is hypothesized
seeds into eachthat cellseeding more seeds
affects germination
per
andcell will negatively
growth throughoutaffect germination
the short crop cycle. andItgrowth as the seed-substrate
is hypothesized that seedingcontact needed
more seeds perfor seeds,
cell will
especially
negativelylarge
affectones, to imbibe
germination andand germinate
growth as thewill be disruptedcontact
seed-substrate in cellsneeded
with multiple
for seeds,seeds by the
especially
earliest germinating
large ones, to imbibeseeds and [8]. One of our
germinate willobjectives
be disrupted was into evaluate
cells withthis hypothesis.
multiple seeds by the earliest
germinating seeds [8]. One of our objectives was to evaluate this hypothesis.
Figure 1. Flat material input costs.
Figure 1. Flat material input costs.
Spinach is not considered to be a high-value crop by producers and generally not tailored to
hydroponic
Spinachproduction. A main
is not considered todriving force behind
be a high-value cropcultivar selection
by producers andduring seednot
generally production is
tailored to
resistance toproduction.
hydroponic blue mold aimedA main at driving
field-grown
forceapplications,
behind cultivarbut selection
this is notduring
of critical
seedimportance
productionfor is
hydroponictobaby
resistance blueleaf
moldproduction
aimed atwhen compared
field-grown to other cultivar
applications, characteristics
but this such importance
is not of critical as simultaneity
for
of germination,
hydroponic baby germination rate, and
leaf production whenplantcompared
morphology to during the shortcharacteristics
other cultivar crop cycle [9].suchRecent
as
research on hydroponic
simultaneity baby germination
of germination, leaf spinach rate,
production in DWC
and plant has focused
morphology on the
during twoshort
round-leaf, downy
crop cycle [9].
mildew-resistant
Recent research on cultivars:
hydroponicSpace andleaf
baby Carmel.
spinachSeaside is a smooth-leaf,
production in DWC has slow-bolting
focused oncultivar that has
two round-leaf,
not beenmildew-resistant
downy thoroughly tested in a DWC
cultivars: system
Space [2,10]. Comparative
and Carmel. testing of these
Seaside is a smooth-leaf, cultivars is
slow-bolting a next
cultivar
step has
that in determining which will
not been thoroughly growinvigorously
tested a DWC system in a DWC
[2,10].environment.
Comparative testing of these cultivars is
a nextThe
stepobjectives of this study
in determining whichwere: a) to vigorously
will grow determine effects
in a DWCof seeding methods on performance of
environment.
babyThe
spinach planted
objectives of at high
this densities
study were: and
a) tob)determine
to compare performance
effects of seedingcharacteristics
methods on from seeding of
performance to
harvest
baby for three
spinach cultivars
planted (Carmel,
at high Seaside,
densities andandb) toSpace) using
compare one of the high
performance density seeding
characteristics frommethods.
seeding
to harvest for three cultivars (Carmel, Seaside, and Space) using one of the high density seeding
2. Materials and Methods
methods.
Experiments were conducted to test seeding patterns (Experiment 1) and performance
2. Materials and
characteristics Methods 2). Germination, fresh weight production, and observations of canopy
(Experiment
characteristics
Experiments were considered
were conductedthe to
mosttestrelevant
seedingmetrics to evaluate
patterns (Experimenttreatments.
1) and Experiment
performance1
compared seeding
characteristics patterns with
(Experiment the Carmel variety
2). Germination, only and
fresh weight consisted and
production, of either seeding identically
observations of canopy
sized cells with a one-two-one alternating pattern [1-2-1-2] of seeds per sequential
characteristics were considered the most relevant metrics to evaluate treatments. Experiment cell or using three
1
seeds in a cell and then an empty cell and then three seeds in a cell etc. [3-0-3-0], giving
compared seeding patterns with the Carmel variety only and consisted of either seeding identically the same
sized cells with a one-two-one alternating pattern [1-2-1-2] of seeds per sequential cell or using three
Horticulturae 2019, 5, 20 3 of 12
average density of 1.5 seeds per cell in both treatments. Experiment 2 observed three cultivars that
were grown concurrently side by side, using the cultivars Carmel, Seaside, and Space seeded at a
fixed density of 1.5 seeds per cell in the [1-2-1-2] pattern. Data on both germination and fresh weight
was collected by counting seedlings in each treatment mid-way through the crop cycle and weighing
hand-harvested leaves after they reached marketable size.
Speedling (Ruskin, FL) 338 cell flats were used for all experiments (see Figures 3 and 9a for a
representative flat). The flats had dimensions of 34.3 cm by 57.2 cm by 4.4 cm. Each flat was divided
into two blocks by using two center rows (no seeds) to separate blocks. Thus, each block had 130 cells
for planting. Both seeding patterns resulted in 195 seeds being planted per block with a potential
to have 195 seedlings. Treatments were assigned randomly to the blocks. Cells are square on the
top, 4.4 cm deep, spaced on 2.54 cm centers, and tapered to a square open bottom, giving the cells a
pyramid-like shape. The blocks described above are replications of the treatments imposed in both
experiments. An advantage of this approach was that time or season was eliminated as a variable that
could impact response in an individual experiment.
Experiments took place in Ithaca, NY (42◦26056.2” N 76◦28008.3” W). Flats were filled and seeded,
then placed in a dark, climate-controlled growth chamber where germination occurred for two and
a half days after seeding. Subsequently, flats were moved to a conventional glass greenhouse and
transferred into tubs using a bench-scale deep-water culture (DWC) system, where they remained
until harvest. A summary of seeding, floating, and harvest dates are given in Table 1.
Table 1. Experiment descriptions and dates.
Experiment Number Experiment Description Seeding Date Float Date Harvest Date
1 Seeding Pattern 10/11/2017 10/13/2017 10/28/2017
2 Cultivar Comparison 10/28/2017 10/30/2017 11/17/2017
2.1. Germination Chamber and Greenhouse Description
A climate controlled chamber (0.93 m2 ) with no light and constant temperature of 24 ◦ C was used
for germination. The DWC system was located on a waist-high bench in the NE corner of a conventional
glass greenhouse. The greenhouse thermal characteristics are well-described by Vandam et al. [10] who
raised spinach in a greenhouse section within the same complex under similar operational protocols.
No supplemental light was provided.
2.2. Greenhouse Environmental Conditions
Greenhouse ambient temperature was maintained at 24 ◦ C from 8 am–8 pm and 19 ◦ C at night
by an automated Argus system (Argus Control Systems Ltd., Surrey, BC, Canada). The relatively
new glass structure was estimated to have a solar light transmissivity of 65% based upon previous
measurements by our research group.
Light values were calculated using W/m2 readings gathered from an outdoor sensor connected to
the greenhouse automation system that controlled the shade curtain and took readings at ten minute
intervals. The quantum units were converted to PAR and multiplied by the transmissivity of the
structure (65%). When the curtain was drawn over the bench on any time step, the PAR value was
multiplied by the transmissivity of the shade curtain (40%) to estimate PAR reaching the canopy. These
PAR values were integrated over 24 hour periods, yielding daily light integral (DLI) values.
For Experiment 1 (seeding pattern) in which the flats were in the DWC tubs for 16 days, the
average cumulative light received was 5.5 mol/m2 /day (Std. Dev. = 1.46) for a total of 88.4 mol per m2
over the entire crop cycle. For Experiment 2 (cultivar comparison), the average light received was
3.0 mol/m2 /day (Std. Dev. = 1.72) for a total of 56.2 mol per m2 received over the 19-day crop cycle.
In both experiments the light levels were below the 17 mol/m2 /day considered ideal for continuous
production [11].
Horticulturae 2019, 5, x FOR PEER REVIEW 4 of 13
both experiments the light levels were below the 17 mol/m2/day considered ideal for continuous
Horticulturae 2019, 5, 20 4 of 12
production [11].
2.3
2.3.Deep-Water
Deep-WaterCulture
CultureSystem
SystemDescription
Description
The production
productionperiod
periodfor forthe
thespinach
spinachplants
plantswaswas conducted
conducted using
using twotwo steel
steel tubs
tubs (61(61cm cm by 122
by 122 cm
cm by 30.5
by 30.5 cm deep)
cm deep) which which
allowedallowed five
five cut cut(described
flats flats (described
below)below) to be floated
to be floated when completely
when completely utilized.
utilized.
Nutrient Nutrient
solution solution
(described (described
in Sectionin2.3)
Section
flowed2.3)freely
flowed freely between
between the two channels
the two channels through through
a 1.9 cm
aID1.9 cm Tubs
hose. ID hose.
wereTubs were with
insulated insulated
1.9 cmwith 1.9 cm polystyrene
polystyrene boards on theirboards on their
exterior exterior heat
to minimize to minimize
transfer
heat transfercooling
and reduce and reduce
load.cooling
A 1/6 HP load. A 1/6 HP submersible
submersible pump was pump used towas pumpusednutrient
to pump nutrient
water water
through a
through a venturi injector (for aeration), 1
followed by a ¼ HP inline chiller that kept the water
venturi injector (for aeration), followed by a 4 HP inline chiller that kept the water temperature at
19 ◦ C. Water at
temperature was19 returned
°C. Watertowas thereturned to thea tubs
tubs through 2.5 cmthrough
ID PVC a 2.5
pipecmwith
ID PVC pipe
outlets with
into outlets
each tub. into
The
each tub.water
nutrient The nutrient
and rootwater and root
zone were zone were
maintained below 20 ◦ C andbelow
maintained highly20aerated,
°C andsince
highlythisaerated, since
temperature
this temperature
has been has been
to be effective to be effective
for mitigating spreadforofmitigating spread ofespecially
Pythium infection, Pythium P. infection, especiallyand
aphanadermatum, P.
aphanadermatum,
the high degree ofand the high
aeration alsodegree of aeration
contributes alsoresistance
to greater contributes to greater
to infection resistance
[12]. Oxygen to infection
levels were
[12]. Oxygen levels were also enhanced by using two air stones in each
also enhanced by using two air stones in each tub. Dissolved oxygen saturation was verified using a tub. Dissolved oxygen
saturation
calibrated YSIwasPro20
verified using
(Yellow a calibrated
Springs, OH, USA)YSI Pro20
meter (Yellow
when theSprings,
system wasOH,operating.
USA) meter when the
system was operating.
Sections of the tubs that were not occupied by flats were covered with StyrofoamTM to prevent
growthSections of the
of algae andtubs that were entering
contaminants not occupied by flats
the water. were covered
Between with Styrofoam
experiments, TM to prevent
the tubs were emptied,
growth
cleaned,ofandalgae and contaminants
sanitized with Greenshieldentering the water.
(BASF, Between experiments,
Ludwigshafen, Germany) tothe tubs were
minimize emptied,
risk of root
cleaned,
infection and
and sanitized with Greenshield
ensure consistent (BASF,conditions
nutrient solution Ludwigshafen,between Germany) to minimize
experiments. Figure 2risk of root
shows the
infection and ensure consistent nutrient solution
experimental setup and placement of tubs, chiller and air pump. conditions between experiments. Figure 2 shows
the experimental setup and placement of tubs, chiller and air pump.
Figure 2.
Figure Bench-scale deep-water
2. Bench-scale deep-water culture
culture (DWC)
(DWC) system
system at
at half-capacity.
half-capacity.
2.4. Nutrient Solution Conditions
2.4 Nutrient Solution Conditions
The stock solutions recommended by Sonneveld and Straver [13] were prepared per the recipe
The stock solutions recommended by Sonneveld and Straver [13] were prepared per the recipe
(Table 2). The day before the flats were removed from the germination chamber for floating, stock
(Table 2). The day before the flats were removed from the germination chamber for floating, stock
solutions were added to reverse osmosis (RO) water in the tubs in a 1:1 ratio (A:B) to achieve an
solutions were added to reverse osmosis (RO) water in the tubs in a 1:1 ratio (A:B) to achieve an
electrical conductivity (EC) of 1200–1400 µS/cm. This corresponds to a half strength solution as
electrical conductivity (EC) of 1200–1400 μS/cm. This corresponds to a half strength solution as
recommended by Sonneveld and Straver [13]. The half strength solution has been shown to be effective
recommended by Sonneveld and Straver [13]. The half strength solution has been shown to be
for lettuce and spinach [10,14]. The nutrient solution resulted in a pH between 5.7 and 6.3 without the
effective for lettuce and spinach [10,14]. The nutrient solution resulted in a pH between 5.7 and 6.3
addition of any additional acid or base over the course of each experiment.
without the addition of any additional acid or base over the course of each experiment.
Throughout each experiment’s growth period, RO water and stock solutions were added to
Throughout each experiment’s growth period, RO water and stock solutions were added to
maintain the EC and pH within the desired range as checked by handheld digital meters, calibrated
maintain the EC and pH within the desired range as checked by handheld digital meters, calibrated
at the beginning of every experiment and every two days thereafter. Water level in the tubs was
at the beginning of every experiment and every two days thereafter. Water level in the tubs was
maintained high enough so that growing flats were not shaded by the tub walls.
maintained high enough so that growing flats were not shaded by the tub walls.
Horticulturae 2019, 5, x FOR PEER REVIEW 5 of 13
Table 2. Stock solution recipes.
Table 2. Stock solution recipes.
Stock A – Chemicals/salts added to 30 L of RO Water
Stock A – Chemicals/salts added to 30 L of RO Water
Calcium Nitrate 2916.0 g
Horticulturae 2019, 5, 20 5 of 12
Calcium Nitrate 2916.0 g
Potassium Nitrate 613.2 g
Potassium Nitrate 613.2 g
Ammonium Nitrate 84.0 g
Ammonium Nitrate Table 2. Stock solution recipes.
84.0 g
Sprint 330 Iron – DTPA (10% Iron) 67.0 g
Sprint
Stock 330 Iron – DTPA
A—Chemicals/salts (10%toIron)
added 67.0Water
30 L of RO g
Calcium Nitrate 2916.0 g
Stock BNitrate
Potassium – Chemicals/salts added to 30 L Water 613.2 g
Stock
Ammonium B – Chemicals/salts added to 30 L Water
PotassiumNitrate
Nitrate 2037.8 g 84.0 g
Sprint 330 Iron—DTPA
Potassium Nitrate (10% Iron) 2037.8 g 67.0 g
Monopotassium Phosphate 816.0 g
Stock B—Chemicals/salts
Monopotassium added to 30 L Water
Phosphate 816.0 g
Potassium Sulfate 65.6 g
Potassium
PotassiumNitrate
Sulfate 65.6 g 2037.8 g
Magnesium Sulfate
Monopotassium Phosphate 1478.0 g 816.0 g
Magnesium
Potassium Sulfate
Sulfate 1478.0 g 65.6 g
Manganese Sulfate* H2O (25%Mn) 2.56 g
Magnesium
Manganese Sulfate
Sulfate* H2O (25%Mn) 2.56 g 1478.0 g
Zinc Sulfate*
Manganese H 2O
Sulfate* H2(35%Zn)
O (25%Mn) 3.44 g 2.56 g
Zinc
Zinc Sulfate*
Sulfate* H 2O (35%Zn) 3.44 g
Boric AcidH2 O (35%Zn) 5.58 g 3.44 g
Boric Acid
Boric Acid 5.58 g 5.58 g
Copper
Copper Sulfate*
Sulfate* 5H2 O 2O (25%Cu)
5H(25%Cu) 0.56 g 0.56 g
Copper
Ammonium Sulfate* 5H
Molybdate 2O (25%Cu) 0.56 g 0.26 g
Ammonium Molybdate 0.26 g
Ammonium Molybdate 0.26 g
2.5
2.5.Dibbling
DibblingTools
Tools
2.5 Dibbling Tools
A combination of conventionally built and 3D-printed dibbling dibbling tools was used for these
A combination
experiments (Figures of
(Figures3 andconventionally
3 and built
4). Dibblers
4). Dibblers and 3D-printed
developed
developed dibbling
by Cornell
by Cornell CEAtools
CEA specifically forwas used
specifically
spinach forfor these
spinach
production
experiments
production
create a divot (Figures
create
in which 3 and
a divotseeds 4).
in which Dibblers developed
seeds atop
are sown, are sown,
whichatop
moreby
whichCornell CEA
more can
substrate specifically
substrate can be
be added for
so added spinach
that thesolarge
that
production
the create
largeseeds
spinach spinach
were a adequately
divot
seeds in which
were deepseeds
adequately are sown,
and deep atop which
andincovered
covered the EPS more
inflats.
the EPS substrate
flats. can be added so that
the large spinach seeds were adequately deep and covered in the EPS flats.
Figure 3. Conventionally-built plate dibbler.
Figure 3.
Figure Conventionally-built plate
3. Conventionally-built plate dibbler.
dibbler.
(a) (b)
(a) (b)
Figure
Figure 4. (a)
(a) Single
Single rolling
rolling dibbler;
dibbler; (b) Whole flat dibble roller.
Figure 4. (a) Single rolling dibbler; (b) Whole flat dibble roller.
2.6. Seeds
Seeds were purchased from Jonny’s Seeds Inc. (Winslow, ME) in September 2017. When not in
use, they were stored in a cool, dark, dry environment. Germination percentages reported by the
Horticulturae 2019, 5, x FOR PEER REVIEW 6 of 13
2.6 Seeds
Seeds2019,
Horticulturae were
5, purchased
20 from Jonny’s Seeds Inc. (Winslow, ME) in September 2017. When not 6 of in
12
use, they were stored in a cool, dark, dry environment. Germination percentages reported by the
supplier were: Carmel—84%, Seaside— 97%, Space—99%. Carmel and Space are both semi-savoy,
supplier
meaning were: Carmel—84%,
the true leaves have aSeaside— 97%, Space—99%.
slightly crinkled appearance.Carmel
Seasideand Space are both
is a smooth-leaf semi-savoy,
variety.
meaning the true leaves have a slightly crinkled appearance. Seaside is a smooth-leaf variety.
2.7 Sowing, Germination, and Floating Procedure
2.7. Sowing, Germination, and Floating Procedure
Before seeding, flats were cleaned, disinfected with a Greenshield solution, thoroughly rinsed,
Before seeding, flats were cleaned, disinfected with a Greenshield solution, thoroughly rinsed,
and dried. Sungro (Vancouver, BC, Canada) Propagation mix, composed of sphagnum peat moss,
and dried. Sungro (Vancouver, BC, Canada) Propagation mix, composed of sphagnum peat moss,
horticultural vermiculite, pH adjustment lime, wetting agent, and starter nutrients was used to fill
horticultural vermiculite, pH adjustment lime, wetting agent, and starter nutrients was used to fill
flats. Reverse Osmosis (RO) water was added to the dry medium at a ratio of 0.6 kg RO water per 1
flats. Reverse Osmosis (RO) water was added to the dry medium at a ratio of 0.6 kg RO water per
kg dry medium and mixed in 20 L buckets. This moisture content was chosen from a trial run that
1 kg dry medium and mixed in 20 L buckets. This moisture content was chosen from a trial run that
showed comparable performance between flats germinated with either 60% or 100% medium wet
showed comparable performance between flats germinated with either 60% or 100% medium wet
weight to medium dry weight. The medium and water were well-mixed, and the lid was kept on the
weight to medium dry weight. The medium and water were well-mixed, and the lid was kept on the
buckets except when mixing or filling flats to minimize moisture loss.
buckets except when mixing or filling flats to minimize moisture loss.
When filling the flats for the first pass, a 0.6 cm metal mesh screen (Figure 5) was used as a sieve
When filling the flats for the first pass, a 0.6 cm metal mesh screen (Figure 5) was used as a sieve to
to remove any clumps from the medium and ensure homogenous, loosely-packed filling of cells.
remove any clumps from the medium and ensure homogenous, loosely-packed filling of cells. Excess
Excess medium was scraped off the top of the flat with a straight edge. The plate dibbler in Figure 3
medium was scraped off the top of the flat with a straight edge. The plate dibbler in Figure 3 was used
was used to create consistent dibbles in each cell and then flats were seeded at the appropriate density
to create consistent dibbles in each cell and then flats were seeded at the appropriate density using the
using the designated cultivar.
designated cultivar.
Figure 5.
Figure Metal mesh
5. Metal mesh sieve
sieve used
used to
to remove
remove clumps
clumps and
and ensure
ensure loose
loose filling.
filling.
After initial seeding, another layer of medium was added on top of the seeds, following the same
After initial seeding, another layer of medium was added on top of the seeds, following the same
sieving procedure described for the first pass. Finally, the rolling dibbler pictured in Figure 4 was used
sieving procedure described for the first pass. Finally, the rolling dibbler pictured in Figure 4 was
to compress the medium on top of the seeds in the second pass by rolling the dibbler across the flat
used to compress the medium on top of the seeds in the second pass by rolling the dibbler across the
surface back and forth several times.
flat surface back and forth several times.
All flats were stacked randomly after they had been seeded, filled, and dibbled. Non-seeded
All flats were stacked randomly after they had been seeded, filled, and dibbled. Non-seeded
guard flats were placed on the top and bottom of the stack. The entire stack of flats was wrapped in
guard flats were placed on the top and bottom of the stack. The entire stack of flats was wrapped in
plastic to prevent moisture loss. The stack was moved into a larger, rigid plastic bin for transportation
plastic to prevent moisture loss. The stack was moved into a larger, rigid plastic bin for transportation
to the germination chamber, where the flats were stored for 60 hours in the dark at 24 ◦ C before
to the germination chamber, where the flats were stored for 60 hours in the dark at 24 °C before
removal for inspection and floating.
removal for inspection and floating.
Upon removal from the germination chamber, flats were inspected for pests and abnormalities.
Upon removal from the germination chamber, flats were inspected for pests and abnormalities.
Throughout the experiments, no pests were found in any seedlings. Germination appeared successful
Throughout the experiments, no pests were found in any seedlings. Germination appeared successful
upon removal from the chamber, meaning there were some visible seedlings protruding from the
upon removal from the chamber, meaning there were some visible seedlings protruding from the
medium, as seen in Figure 6, and many more young roots protruding out of the bottoms of flats. After
medium, as seen in Figure 6, and many more young roots protruding out of the bottoms of flats. After
this step, the flats were floated in the tubs, where they remained until removal for harvest.
this step, the flats were floated in the tubs, where they remained until removal for harvest.
Horticulturae 2019, 5, x FOR PEER REVIEW 7 of 13
Horticulturae 2019, 5, 20 7 of 12
Flats just
Figure 6. Flats
Figure just after
after floating
floating with seedlings breaching the medium surface.
2.8. Seedling Count
2.8 Seedling Count
On the sixth day after floating in the tubs, the visible seedlings were counted for each treatment.
On the sixth day after floating in the tubs, the visible seedlings were counted for each treatment.
This count included “popups” (seedlings that withered because their roots did not reach the water),
This count included “popups” (seedlings that withered because their roots did not reach the water),
which were uncommon [9]. Counting seedlings earlier than six days was determined to be a poor
which were uncommon [9]. Counting seedlings earlier than six days was determined to be a poor
indication of germination and initial growth success.
indication of germination and initial growth success.
2.9. Harvest Protocol
2.9 Harvest Protocol
Flats were harvested when the fastest growing leaves of a treatment reached marketable size,
Flats were harvested when the fastest growing leaves of a treatment reached marketable size,
which occurred between 14 and 18 days after transfer to the tubs. This harvest date initiation was
which occurred between 14 and 18 days after transfer to the tubs. This harvest date initiation was
chosen so that no leaves exceeded 7.5 cm in length (when measured from the base of the leaf).
chosen so that no leaves exceeded 7.5 cm in length (when measured from the base of the leaf).
Harvest was performed by hand with scissors. A handful of leaves would be carefully grasped.
Harvest was performed by hand with scissors. A handful of leaves would be carefully grasped.
The stems would then be cut about halfway down the stems of the longest leaves. This meant some
The stems would then be cut about halfway down the stems of the longest leaves. This meant some
smaller leaves sitting lower in the canopy were harvested as well, but with shorter stems. A large
smaller leaves sitting lower in the canopy were harvested as well, but with shorter stems. A large
variance in leaf size was found for all treatments, which was due both to the non-simultaneous
variance in leaf size was found for all treatments, which was due both to the non-simultaneous
germination observed, and the presence of young, second true leaves in addition to the larger first true
germination observed, and the presence of young, second true leaves in addition to the larger first
leaves. Harvest weight was measured immediately after leaves were harvested using a digital scale
true leaves. Harvest weight was measured immediately after leaves were harvested using a digital
accurate to one gram. As can be seen in Figure 7, many small leaves sitting low in the canopy were
scale accurate to one gram. As can be seen in Figure 7, many small leaves sitting low in the canopy
not harvested.
were not harvested.
2.10. Statistical Analysis
In Experiment 1 (seeding pattern with single cultivar) there were four replications of each pattern
and in Experiment 2 (cultivar response to fixed seeding method), there were four replications of the
Carmel and Seaside cultivars and two replications of the Space cultivar. For both experiments, a
one-way analysis of variance using Tukey–Kramer paired t-tests was performed using JMP PRO 11
statistical analysis software [15]. The average and standard deviation of each treatment (pattern or
cultivar) except for the cultivar Space was calculated for both final seedling counts and fresh weight.
Data from a typical high-germinating flat is presented to show the intra-flat variation observed, even
when germination was high.
Horticulturae 2019, 5, 20 8 of 12
Horticulturae 2019, 5, x FOR PEER REVIEW 8 of 13
Figure 7. Before and mid-harvest of Carmel; some small second true leaves are left behind low in
Figure 7. Before and mid-harvest of Carmel; some small second true leaves are left behind low in the
the canopy.
canopy.
3. Results
2.10 Statistical Analysis
3.1. Experiment 1: Seeding
In Experiment Pattern pattern with single cultivar) there were four replications of each
1 (seeding
pattern and in Experiment 2 (cultivar response to fixed seeding method), there were four replications
Results
of fromand
the Carmel theSeaside
seeding patternand
cultivars experiment showed
two replications that
of the seeding
Space pattern
cultivar. [1-2-1-2]
For both or [3-0-3-0]
experiments,
negatively
a one-way analysis of variance using Tukey–Kramer paired t-tests was performed using JMP PROblock
affected germination. There were 157 (Std. Dev. = 4.2) seedlings per replicate 11 in
the [1-2-1-2]
statisticalpattern
analysisversus
software148[15].
(Std. Dev.
The = 4.5)
average seedlings
and standardper replicate
deviation block
of each in the [3-0-3-0]
treatment (pattern pattern
or
(Tablecultivar)
3) fromexcept
a totalforofthe
195cultivar
seeds Space
planted
wasper replicateforblock.
calculated The seedling
both final Carmel cultivar
counts andhad a manufacturer
fresh weight.
Datasprouting
reported from a typical high-germinating
percentage of 83%. Ourflat is presented
results to show
for both the intra-flat
treatments variation
achieved highobserved, evenof the
percentages
when germination was high.
manufacturer’s reported germination percentage although numerically the [1-2-1-2] pattern was 97%
of the reported value while the [3-0-3-0] pattern was 91% of the reported value. Fresh weight harvest
3. Results
yield was lower numerically but not statistically with the [3-0-3-0] sowing pattern as well. Since the
reductions in germination
3.1 Experiment 1: Seeding and yield were less than 10%, growers should consider and evaluate the
Pattern
cost of additional substrate versus reductions in performance based upon these results.
Results from the seeding pattern experiment showed that seeding pattern [1-2-1-2] or [3-0-3-0]
negatively affected germination. There were 157 (Std. Dev. = 4.2) seedlings per replicate block in the
Table 3. Effects of seeding pattern on seed germination from 195 seeds per replicate block and harvested
[1-2-1-2] pattern versus 148 (Std. Dev. = 4.5) seedlings per replicate block in the [3-0-3-0] pattern (Table
fresh weight yield of Carmel spinach. Cells were seeded according in [1-2-1-2] or [3-0-3-0] patterns
3) from a total of 195 seeds planted per replicate block. The Carmel cultivar had a manufacturer
giving average density of 1.5 seeds per cell. Different superscripts indicate significant differences
reported sprouting percentage of 83%. Our results for both treatments achieved high percentages of
between
the treatmentsreported
manufacturer’s by Tukey–Kramer
germinationpaired t-tests atalthough
percentage α = 5%. numerically the [1-2-1-2] pattern was
97% of the reported
Treatmentvalue while Replication
the [3-0-3-0] pattern was 91%
Seedling of the reported
Count value. (g)
Fresh Weight Fresh weight
harvest yield was lower numerically but not statistically with the [3-0-3-0] sowing pattern as well.
[1-2-1-2] 1 150 144
Since the reductions in germination and 2 yield were less159 than 10%, growers should
157 consider and
evaluate the cost of additional substrate3 versus reductions158 in performance based148
upon these results.
4 161 156
Mean 157 a 151.2 a
Std. Dev. 4.2 5.4
[3-0-3-0] 1 144 134
2 155 152
3 144 146
4 149 145
Mean 148 b 144.2 a
Std. Dev. 4.5 6.5
Horticulturae 2019, 5, 20 9 of 12
3.2. Cultivar Comparison
As shown in Table 4, Carmel and Space cultivars had a significantly higher germination percentage
than Seaside (α = 5%) and fresh weight yield production (α = 5%). Carmel had the highest germination
(81%) of the three cultivars tested. Carmel averaged 98% of its reported 83% germination rate. Final
germination percentage was stable for Carmel across both experiments; Experiment 1 was also 81%.
Space and Seaside cultivars germinated and produced seedlings at a relatively high rate, as well, but
at significantly lower rates than reported by the supplier.
Table 4. Effects of Carmel, Seaside, and Space cultivars on germination and harvested fresh weight
yield with all cells seeded in a [1, 2, 1, 2] seeding pattern (195 seeds planted). Different superscripts
indicate significant differences between treatments by Tukey–Kramer paired t-tests at α = 5%.
Reported Actual Seedling Harvested Fresh
Cultivar
Germination % Germination % Count Weight (g)
Carmel 83 81 157 a 150 a
Seaside 97 74 144 b 94 b
Space 99 76 147 a 146 a
Seaside baby spinach grown in the low-light conditions of November in Ithaca produced a short,
bushy canopy compared to that of Carmel and Space as can be seen in Figure 8. The latter two cultivars
Horticulturaelonger
produced 2019, 5, xstemmed
FOR PEERleaves
REVIEWthat sat higher above the cotyledons and were easier to harvest.
10 of 13
Figure 8. Carmel (left) and Seaside (right) at harvest illustrating the contrast between the long stems
Figure 8. Carmel (left) and Seaside (right) at harvest illustrating the contrast between the long stems
of Carmel and the short, bushy canopy of Seaside.
of Carmel and the short, bushy canopy of Seaside.
Pericarps are the hard outer coat of spinach seeds that are sometimes stuck to the tips of emerging
Pericarps are the hard outer coat of spinach seeds that are sometimes stuck to the tips of
cotyledons (Figure 9a). At time of harvest, they were infrequent and hard to spot as cotyledons become
emerging cotyledons (Figure 9a). At time of harvest, they were infrequent and hard to spot as
tangled in the lower leaves. However, it was rare to spot a pericarp on cotyledons after a week in the
cotyledons become tangled in the lower leaves. However, it was rare to spot a pericarp on cotyledons
tubs as most were pushed off by the growing cotyledons and true leaves. (Figure 9b).
after a week in the tubs as most were pushed off by the growing cotyledons and true leaves. (Figure
In the Carmel and Space treatments, fewer cotyledons ended up in the harvested spinach, in
9b).
general, due to the longer stems of the true leaves and the nature of the cotyledons to fold over
In the Carmel and Space treatments, fewer cotyledons ended up in the harvested spinach, in
under the true leaves compared to Seaside. Overall, this resulted in almost no pericarps in the
general, due to the longer stems of the true leaves and the nature of the cotyledons to fold over under
harvested leaves.
the true leaves compared to Seaside. Overall, this resulted in almost no pericarps in the harvested
leaves.
after a week in the tubs as most were pushed off by the growing cotyledons and true leaves. (Figure
9b).
In the Carmel and Space treatments, fewer cotyledons ended up in the harvested spinach, in
general, due to the longer stems of the true leaves and the nature of the cotyledons to fold over under
the true leaves compared to Seaside. Overall, this resulted in almost no pericarps in the harvested
Horticulturae 2019, 5, 20 10 of 12
leaves.
Horticulturae 2019, 5, x FOR PEER REVIEW 11 of 13
Figure 9. (top) Pericarp stuck on cotyledon late in the crop cycle; (bottom) A flat one week after floating
Figure 9. (top) Pericarp stuck on cotyledon late in the crop cycle; (bottom) A flat one week after
with few pericarps.
floating with few pericarps.
3.3. Intra-Flat Variation
3.3 Intra-Flat Variation
One flat of Carmel was evaluated to look at row to row variation in germination. Even for a
One flat of Carmel
high-germinating flat of was evaluated
Carmel, to lookin
the variation atseeds
row to row
that variation in
germinated germination.
between rows ofEven
a flatfor
wasa
high-germinating flat of Carmel, the variation in seeds that germinated between rows of
surprising: across 22 rows, germination was 80.4% (Std. Dev. = 2.32), and ranged from 56% to 97%. As a flat was
surprising:
the seedlingsacross
grew,22they
rows, germination
took advantagewas 80.4%
of all (Std. space
available Dev. =to2.32), and
create ranged from
a relatively 56% to
uniform 97%.
canopy
As the seedlings
which seemed togrew, theyeffects
mitigate took advantage
caused by of all available
intra-row spacein
variation toseed
create a relatively uniform canopy
germination.
which seemed to mitigate effects caused by intra-row variation in seed germination.
4. Discussion
4. Discussion
These experiments correspond to first steps in a comprehensive comparison of the cultivars
represented in these experiments.
These experiments Ideally,
correspond the steps
to first studyin
would have been conducted
a comprehensive in a growth
comparison chamber
of the cultivars
with consistent
represented DLI,experiments.
in these temperature,Ideally,
and relative humidity
the study wouldashave
in [2].
beenHowever,
conducted thein
germination results
a growth chamber
were consistent and largely independent of fluctuations in greenhouse light conditions. It
with consistent DLI, temperature, and relative humidity as in [2]. However, the germination results is peculiar
that the
were cultivarand
consistent withlargely
the lowest reportedof
independent germination,
fluctuationsCarmel, had thelight
in greenhouse highest overall It
conditions. germination
is peculiar
that the cultivar with the lowest reported germination, Carmel, had the highest overall germination
percentage (nearly 100% of that reported by the supplier) and the most vigorous growth during the
short 15–18 day in-tub cycles used in our experiments. Space also performed relatively well when
compared to Seaside. Seaside was the least suitable for DWC hydroponic production given its lower
germination, short stems, and reduced growth.Horticulturae 2019, 5, 20 11 of 12
percentage (nearly 100% of that reported by the supplier) and the most vigorous growth during the
short 15–18 day in-tub cycles used in our experiments. Space also performed relatively well when
compared to Seaside. Seaside was the least suitable for DWC hydroponic production given its lower
germination, short stems, and reduced growth.
It appears the current sowing and germination procedure is sub-optimal for Space and Seaside,
which germinated less frequently than their reported rate. In future tests, cells will be taken apart
and studied where no seedlings emerged to observe the seed condition. If seeds had germinated in
the cell but then had not developed properly, this would suggest altering the planting procedures.
For instance, if the seed germinated but the seedling withered and died before it could breach the
medium surface or its roots reached the nutrient solution, different cell dimensions, seed positions
in cells, or compaction/dibbling schemes should be tested. On the other hand, if cells were found
to have non-germinated seeds, then different germination temperatures, a wider range of medium
moisture contents, or duration spent in the germination chamber should be investigated, among other
germination procedure alterations.
As mentioned previously, seedling numbers earlier than six days after floating started were not
necessarily indicative of the final numbers. This suggests methods to imbibe or prime seeds could
be applied to synchronize germination timing and speed up the germination of the slowest seeds. If
the last seedlings are emerging several days after the first, there is a clear loss of productivity and
the maximum potential of the seeds is not being realized. Cornell CEA has done extensive work on
imbibing procedures [9,16,17].
The pattern experiment results showed that there was a small reduction (10%) in final yield of
Carmel when fewer cells were seeded but at a higher density per cell. This implies that proportional
cell volume available to each seedling is not predictive of the growth within the range considered,
which could be due to the significantly shorter cycle observed than considered by Romano et al. [6].
Plus, in a DWC system the large majority of the root mass is dangling in the nutrient solution below
the flat, so the cell volume available does not necessarily restrict root growth or nutrient uptake.
A custom-designed flat with half as many cells as the Speedling 338, but spaced twice as far
apart would cut medium costs roughly in half. The relative costs of medium and seeds are such that
compensating for the reduced germination associated with sowing more seeds in each cell (as was
observed in Experiment 1) with 10% more seeds while using half the medium would result in a large
proportional decrease in material input costs per flat. These preliminary tests corroborate similar
results observed with the spinach cultivar Alrite and highlight the potential for a flat custom designed
for baby spinach production [9,16,17].
Another promising observation from Experiment 1 was that the canopies of blocks in both
treatments were full and relatively uniform. This validates the experimental design assumption that
the canopy density effects between the treatments would be the same. It also suggests that seedlings
will not be shaded out by neighbor plants when more seeds are planted in each cell if the overall
canopy density is not too high and there is adequate room adjacent to the cell for seedlings to expand
into during growth.
Author Contributions: D.B.J. and M.B.T. conceived and designed the experiment; D.B.J. performed the
experiment; D.B.J. and M.B.T. analyzed the data; D.B.J. wrote the paper with significant contributions from
M.B.T.
Acknowledgments: This research was supported by the Cornell University Department of Biological and
Environmental Engineering as well as Element Farms, Inc. who supplied the seeds. We would like to thank
Francoise Vermeylen from the Cornell Statistical Consulting Unit for her guidance on experimental design and
statistical analysis. We express our gratitude to Olav Imsdahl (Cornell M.Eng 2017) for his work on 3D printed
dibbling tools. Additional thanks to Alan Taylor and David deVilliers for advice and guidance and Neil Mattson
for equipment and greenhouse space.
Conflicts of Interest: The authors declare no conflict of interest.Horticulturae 2019, 5, 20 12 of 12
References
1. Resh, H.M. Hydroponic Food Production: A Definitive Guidebook for the Advanced Home Gardener and the
Commercial Hydroponic Grower; CRC Press: Boca Raton, FL, USA, 2012.
2. Albright, L.; Brechner, M.; deVilliers, D. Technology Transfer, Education, and Applied Research from the Center
for Excellence for Controlled Environment Agriculture at Cornell University; Final Report; The New York State
Energy Research and Development Authority, Cornell University: Ithaca, NY, USA, 2016.
3. Albright, L. Energy Investments and CO2 Emissions for Fresh Produce Imported Into New York State Compared to
The Same Crops Grown Locally; The New York State Energy Research and Development Authority, Cornell
University: Ithaca, NY, USA, 2008.
4. Bergquist, S. Bioactive Compounds in Baby Spinach (Spinacia oleracea L.); Diss. (summary); Swedish University
of Agricultural Sciences, Acta Universitatis Agriculturae Sueciae: Uppsala, Sweden, 2006; ISSN 1652-6880.
ISBN 91-576-7111-7.
5. Vegetables 2017 Summary; National Agricultural Statistics Service, United States Department of Agriculture:
Washington, DC, USA, 2018. Available online: http://usda.mannlib.cornell.edu/usda/current/VegeSumm/
VegeSumm-02-13-2018.pdf (accessed on 26 February 2019).
6. Romano, D.; Paratore, A.; Rosi, A.L. Plant Density and Container Cell Volume on Solanaceous Seedling
Growth. Acta Hortic. 2003, 614, 247–253. [CrossRef]
7. Gonnella, M.; Serio, F.; Conversa, G.; Santamaria, P. Yield and Quality of Lettuce Grown in Floating System
Using Different Sowing Density and Spatial Arrangements. Acta Hortic. 2003, 614, 687–692. [CrossRef]
8. Harper, J.L.; Williams, W.T.; Sagar, G.R. The behaviour of seeds in soil. I. The heterogeneity of soil surfaces
and its role in determining the establishment of plants from seed. J. Ecol. 1965, 53, 273–286. [CrossRef]
9. A Commercially Viable Controlled Environment Agriculture (CEA) Spinach Production System; The New York
State Energy Research and Development Authority, Cornell University: Ithaca, NY, USA, 2005.
10. Vandam, D.A.; Anderson, T.S.; De Villiers, D.; Timmons, M. Growth and Tissue Elemental Composition
Response of Spinach (Spinacia oleracea) to Hydroponic and Aquaponic Water Quality Conditions.
Horticulturae 2017, 3, 32. [CrossRef]
11. Hydroponic Spinach Production Handbook; deVilliers, D.; Brechner, M. (Eds.) Cornell University Controlled
Environment Agriculture: Ithaca, NY, USA, 2013.
12. deVilliers, D.; Shelford, T. Root Disease Treatment Methods for the Commercial Production of Hydroponic Spinach;
Department of Biological and Environmental Engineering, Cornell University Controlled Environment
Agriculture: Ithaca, NY, USA, 2007.
13. Sonneveld, C.; Straver, N. Nutrient solutions for vegetables and flowers grown in water or substrates.
In Proefstation voor Tuinbouw onder Glas te Naaldwijk; Series: Voedingsoplossingen Glastuinbouw; Research
Station for Floriculture and Glasshouse Vegetables: Aalsmeer/Naaldwijk, The Netherlands, 1994; Volume 8,
pp. 1–45.
14. Both, A.J.; Albright, L.D.; Langhans, R.W.; Reiser, R.A.; Vinzant, B.G. Hydroponic lettuce production
influenced by integrated supplemental light levels in a controlled environment agriculture facility:
Experimental results. Acta Hortic. 1997, 418, 45–51. [CrossRef]
15. JMP Pro software: http://www.jmp.com/en_us/software/jmp-pro.html, JMP® , Version . SAS
Institute Inc.: Cary, NC, USA, 1989–2019.
16. Katzman, L.S.; Taylor, A.G.; Langhans, R.W. Seed Enhancements to Improve Spinach Germination.
Hortscience 2001, 36, 979–981. [CrossRef]
17. Katzman, L.S. Developing a System to Germinate and Grow Hydroponic Spinach, Spinacia Oleracea; Cornell
University: Ithaca, NY, USA, 1999.
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).You can also read
