Ro

Ta

b

c

a

ARRA

KDDPYA

1

H

scadmdop

h0

Scientia Horticulturae 179 (2014) 306–313

Contents lists available at ScienceDirect

Scientia Horticulturae

journa l h om epa ge: www.elsev ier .com/ locate /sc ihor t i

esponses of two Anthurium cultivars to high daily integralsf diffuse light

. Lia,∗, E. Heuvelinka, F. van Noortb, J. Kromdijkc, L.F.M. Marcelisa

Horticulture and Product Physiology Group, Wageningen University, P.O. Box 630, 6700AP Wageningen, The NetherlandsWageningen UR Greenhouse Horticulture, P.O. Box 644, 6700AP Wageningen, The NetherlandsInstitute for Genomic Biology, University of Illinois, 1206W Gregory Drive, Urbana, IL 61801, USA

r t i c l e i n f o

rticle history:eceived 31 May 2014eceived in revised form 22 August 2014ccepted 19 September 2014

eywords:iffuse glass coveraily light integrallant growthield component analysisnthurium andreanum cultivars

a b s t r a c t

Heavy shading is commonly applied during production of pot-plants in order to avoid damage causedby high light intensities; usually the daily light integral (DLI) is limited to 5–8 mol m−2 d−1 photosyn-thetically active radiation (PAR). However, shading carries a production penalty as light is the drivingforce for photosynthesis. Diffuse glass has been developed to scatter the incident light in greenhouses.This study aims at investigating the effect of diffuse glass cover and high DLI under diffuse glass coveron the growth of pot-plants; furthermore, to systematically identify and quantify the yield componentswhich are influenced by these treatments. Experiments were carried out with two Anthurium andreanumcultivars (Royal Champion and Pink Champion) in a conventional modern glasshouse compartment cov-ered by clear glass with DLI limited to 7.5 mol m−2 d−1 (average realized DLI was 7.2 mol m−2 d−1), andanother two glasshouse compartments covered by diffuse glass with DLI limited to 7.5 (average realizedDLI was 7.5 mol m−2 d−1) and 10 mol m−2 d−1 (average realized DLI was 8.9 mol m−2 d−1). Diffuse glasscover resulted in less variation of temporal photosynthetic photon flux density (PPFD) distribution com-pared with the clear glass cover. Under similar DLI conditions (DLI limited to 7.5 mol m−2 d−1), diffuseglass cover stimulated dry mass production per unit intercepted PPFD (RUE) in ‘Royal Champion’ by 8%;whilst this stimulating effect did not occur in ‘Pink Champion’. Under diffuse glass cover, biomass pro-duction was proportional to DLI in both cultivars (within the range 7.5–9 mol m−2 d−1). Consequentlyhigher DLI led to more flowers, leaves and stems. Furthermore, high DLI resulted in more compact plants

without light damage in leaves or flowers in both cultivars. ‘Pink Champion’ produced more biomassthan ‘Royal Champion’ in all treatments because of higher RUE which resulted from a more advanta-geous canopy architecture for light capture and more advantageous leaf photosynthetic properties. Weconclude that less shading under diffuse glass cover not only stimulates plant growth but also improvesplant ornamental quality (i.e. compactness).

. Introduction

Light is the most important factor in determining plant growth.owever, when excessive light energy is absorbed by the light

Abbreviations: DLI, daily light integral; PPFD, photosynthetic photon flux den-ity; PAR, photosynthetically active radiation; LAI, leaf area index; K, light extinctionoefficient; SLA, specific leaf area; VPD, vapour pressure deficit; TDM, total dry mass;FM, aboveground fresh mass; aDM, aboveground dry mass; aDMC, abovegroundry mass content (i.e. aboveground dry mass/fresh mass); aDMP, aboveground dryass partitioning (i.e. aboveground dry mass/total dry mass); RUE, dry mass pro-

uction per unit intercepted PPFD; Ii , cumulative intercepted PPFD; Io, PPFD at topf plants; I(L), PPFD at leaf area index L; I(L)/Io, fraction of intercepted PPFD; Pn, nethotosynthetic rate.∗ Corresponding author. Tel.: +31 0 317 483678.

E-mail address: tao.li@wur.nl (T. Li).

ttp://dx.doi.org/10.1016/j.scienta.2014.09.039304-4238/© 2014 Elsevier B.V. All rights reserved.

© 2014 Elsevier B.V. All rights reserved.

harvesting antennae at a rate which surpasses the capacity for pho-tochemical and non-photochemical energy dissipation, this maylead to photo-damage (Long et al., 1994). In the long term, thismay result in discolouring of leaves or even necrosis in the mostextreme case. Light damage occurs mostly as a result of prolongedexposure to excessive peaks in light intensity (Asada, 1999; Niyogi,1999; Kasahara et al., 2002). Consequently, growers apply shadingduring summer cultivation of many greenhouse crops by closing ascreen or having a white wash on the greenhouse cover in order toprevent damage under conditions of high light.

In greenhouses, the distribution of light over the different leavesof a canopy shows large variations. The greenhouse construction,

equipment and overstory leaves cast shade, resulting in shade-spots and lightflecks, of which the position continuously changesdepending on solar angle. Light damage may occur particularly inthose lightflecks (Way and Pearcy, 2012). It has been shown that

icultur

dc2ha2goa1r(ltgt

(r(hrpgli2awhconDq

r1giilsr22

gpcDoqtwadC

2

2

R

T. Li et al. / Scientia Hort

iffuse light is more homogeneously distributed over the cropanopy than direct light (Farquhar and Roderick, 2003; Gu et al.,002; Li et al., 2014; Mercado et al., 2009). Recently diffuse glassas become available that increases the diffuseness of light withoutffecting light transmission in the greenhouse (Baeza and López,012; Hemming et al., 2008). Li et al. (2014) observed that diffuselass cover result in a more homogeneous light distribution notnly in the vertical plane, but also in the horizontal plane within

tomato canopy, which compared with clear glass cover, lead to0% higher yield (Dueck et al., 2012). Additionally, diffuse light alsoesults in lower leaf or flower temperature and less photoinhibitionKempkes et al., 2011; Li et al., 2014; Urban et al., 2012), because ofess severe local peaks in light intensity. Considering these proper-ies, we speculate that diffuse glass cover may help stimulate plantrowth at higher daily light integral [DLI, mol m−2 d−1 photosyn-hetically active radiation (PAR)] without leading to light damage.

Increasing DLI increases plant growth and developmentMarcelis et al., 2006; Poorter et al., 2013). Fausey et al. (2005)eported a linear relationship between the amount of light5–20 mol m−2 d−1 PAR) and shoot dry mass in a number of green-ouse grown herbaceous perennial species. Similar findings wereeported by Faust et al. (2005) in a number of bedding plants. Pot-lants are often grown under very low DLI conditions in commercialreenhouse production. For instance in the Netherlands growersimit the DLI in many pot-plants to 5–8 mol m−2 d−1. However, its clear that low DLI can carry a production penalty (Scuderi et al.,012, 2013), since potential crop growth is positively related to themount of light that can be captured. Pot-plants could grow fasterhen less shading was applied in combination with moderatelyigh air humidity (Kromdijk et al., 2012). Furthermore, less shadingould increase plant compactness as indicated by a higher ratiof aboveground dry mass to plant height with increasing DLI in aumber of bedding plants (Faust et al., 2005). Therefore, increasingLI can improve not only plant growth but also plant ornamentaluality.

Yield component analysis has been valuable in many cropesearch programs (Higashide and Heuvelink, 2009; Jolliffe et al.,990; Plénet et al., 2000). Lawlor (1995) suggested that plantrowth and production is determined by component processesntegrated over the canopy, e.g. dry mass production per unitntercepted photosynthetic photon flux density (PPFD) (RUE),eaf photosynthesis, canopy architecture, biomass allocation (e.g.hoot/root ratio). These components vary across species and envi-onments (Barthelemy and Caraglio, 2007; Falster and Westoby,003; Sinclair and Muchow, 1999; Sultan, 2000; Sarlikioti et al.,011b), resulting in differences in crop production.

The objective of this study was to investigate the effect of diffuselass cover and high DLI under diffuse glass cover on the growth inot-plants. It aims to systematically identify and quantify the yieldomponents which are influenced by diffuse glass cover and highLI. Our hypothesis is that high daily integral of diffuse light notnly stimulates plant growth but also improves plant ornamentaluality (i.e. more compact plants without light damage). To testhis hypotheses, a study was conducted under diffuse glass coverith two levels of DLI. Two Anthurium cultivars (Pink Champion

nd Royal Champion) were used in this study; these two cultivarsiffered in light sensitivity based on grower’s experience that ‘Royalhampion’ is more sensitive to light than ‘Pink Champion’.

. Materials and methods

.1. Plant material and growth conditions

Two Anthurium andreanum cultivars (Pink Champion andoyal Champion, Anthura, Bleiswijk, The Netherlands) were

ae 179 (2014) 306–313 307

grown in three Venlo-type glasshouse compartments of 144 m2

(15 m × 9.6 m) with a gutter height of 5.5 m at WageningenUR Greenhouse horticulture in Bleiswijk (TheNetherlands, 52◦N,4.5◦E). The three compartments were covered by glass (GuardianAgro, Dudelange, Luxembourg) with 0% haze (clear glass; one com-partment) and 71% haze (diffuse glass; two compartments). Hazeis defined as the percentage of transmitted light that is scatteredsuch that it deviates more than 1.5◦ from the direction of the inci-dent beam. The hemispherical transmission of PPFD of the glasswas 84% for both glass types. The haze factor and hemispheri-cal transmission of the glass was measured in an optical sphereaccording to ASTM International (2007). The spectral propertiesof the two glass types are presented in Supplementary Fig. S1and Table S1. The DLI was limited to 7.5 mol m−2 d−1 in the clearglass treatment, and to 7.5 and 10 mol m−2 d−1 in the two diffuseglass treatments. The DLI treatment of 10 mol m−2 d−1 under clearglass cover was not included in this experiment, because a sim-ilar treatment in an earlier experiment resulted in leaf damage(Van Noort et al., 2011). The DLI treatments were realized by con-trolling a white sunscreen (XLS 16 F Revolux, transmission of 37%and haze factor of 10%, LudvigSvensson, Kinna, Sweden) and black-out screen (XLS obscural Revolux A/B + B/B, LudvigSvensson, Kinna,Sweden) which were placed in the top of the greenhouse (belowgutter height). The white sunscreen was fully closed in the lowDLI compartments (7.5 mol m−2 d−1) and 50% closed in the highDLI compartment (10 mol m−2 d−1) when global outside radiationreached 250 W m−2; it was fully closed in the high DLI compartmentwhen global outside radiation reached 450 W m−2. The blackoutscreen was closed when DLI reached the DLI limitation point in theafternoon in all compartments. Three quantum sensors (LI-190, LI-COR, USA) were installed in each of the greenhouse compartmentsto measure incident PPFD at 5 min intervals. Fogging systems wereused to maintain high air humidity (80%). A standard horticulturalcomputer (Hogendoorn-Economic, Hogendoorn, Vlaardingen, TheNetherlands) was used to control the greenhouse temperature, airhumidity, CO2 concentration, as well as opening and closing of thescreens.

Plants, propagated in vitro, were raised in a greenhouse by anursery. When the first flowers had appeared, the plants wererepotted and moved to the experimental greenhouses on 6 Apr2012. The experiment ended on 28 Aug 2012. Plants were grown onpotting soil (30% fine peat + 10% coarse peat + 43% coco peat + 10%bark + 7% perlite) in black plastic pots (12 cm diameter and 11 cmheight) on cultivation tables (4 m by 1.8 m) with an automaticebb/flood irrigation system. In each compartment, six cultivationtables were used and each table was equally divided into two partsfor the cultivation of two cultivars. The outer two rows of eachplot were considered as border plants. The starting plant densitywas 30 plants m−2; this was reduced to 20 plants m−2 three weeksafter the start of the experiment. After each destructive harvest,plants were moved to maintain the same plant density. Duringthe growing season, average daily outside global radiation was16 MJ m−2 d−1. Inside the greenhouse the average day/night tem-perature was 25/21 ◦C; relative air humidity was 75/78%; averagedaytime CO2 concentration was 754 �mol mol−1; average realizedDLI were 7.2 mol m−2 d−1 in the compartment of clear glass + lowDLI, 7.5 mol m−2 d−1 in the compartment of diffuse glass + low DLI,and 8.9 mol m−2 d−1 in the compartment of diffuse glass + high DLI.An overview of DLI during the growing season in the three com-partments is providedin Supplementary Fig. S2.

2.2. Plant measurements

Plants were destructively measured at 4, 10, 16, 18 and 21weeks after the start of the experiment (at 18 weeks one extra

3 iculturae 179 (2014) 306–313

miptoPLLlattottl

2e

(2diwjsccm

Li

I

((iwt

2

teeitfoglcoe

I

wt

o

Fig. 1. General scheme with a top-down analysis of aboveground fresh mass (aFM)

08 T. Li et al. / Scientia Hort

easurement was added in order to improve accuracy of the resultsn the later growth stages). For each destructive measurement, twolants per cultivar were randomly selected from each cultivationable, which resulted in 12 replicates. Fresh and dry weight of plantrgans (leaves, flowers, stems, petioles and roots) were determined.lant organs were dried for at least 48 h at 80 ◦C in a ventilated oven.eaf area was measured with a leaf area meter (LI-3100C, Li-Cor Inc.,incoln, USA). Specific leaf area (SLA) was calculated by dividing theeaf area by leaf dry weight. Number of flowers, leaves and stems,s well as plant height were determined. During the fourth destruc-ive harvest, in all the harvested plants, three to four images wereaken of the main stem from the side view of each plant, i.e. after allther leaves were removed (Supplementary Fig. S3), which showedhe leaf angle information of a single leaf. These images were usedo quantify the leaf angle which was determined as the angle of theeaf surface with the horizontal plane (0◦) (Supplementary Fig. S3).

.3. Canopy PPFD interception measurements and lightxtinction coefficient calculation

Canopy PPFD interception was measured on four overcast days11 May, 15 Jun, 17 Jul and 24 Aug) and three clear days (23 May,0 Jun and 25 Jul). These days were close to the period whenestructive measurements were taken in order to correlate light

nterception data with leaf area index (LAI). The measurementsere done with a line light probe, in relation to a reference sensor

ust above the crop (Sunscan, Delta-T, Cambridge, UK). Six mea-urements were done above as well as below the canopy for eachultivar on each cultivation table. Measurements at the top of theanopy were taken just above the highest leaf, while the bottomeasurements were done at pot height.The light extinction coefficient (K) was calculated according to

ambert–Beer law (Monsi and Saeki, 2005) by combining the PPFDnterception measurements on cloudy days and the measured LAI

(L) = I0e−KL (1)

Where, I0 is incident PPFD at the top of the canopy�mol m−2 s−1), L is LAI (m2 leaf m−2 ground), I (L) is PPFD�mol m−2 s−1) at L. Differences between treatments and cultivarsn K were considered significant when K of one treatment/cultivar

as out of the range of 95% confidence interval of the oppositereatment/cultivar.

.4. Canopy RUE determination

Canopy RUE was defined as the ratio between the accumulatedotal dry mass (TDM) and the sum of intercepted PPFD during thexperimental period, which was estimated by the slope of the lin-ar relationship between the accumulated TDM and the sum ofntercepted PPFD. For calculating the sum of intercepted PPFD, theime course of fraction of intercepted PPFD [I(L)/I0] was estimatedrom the four periodic canopy PPFD interception measurementsn cloudy days [Eq. (2)]. These data can represent I(L)/I0 over therowing season due to I(L)/I0 measured on clear days was simi-ar as on cloudy days (data not shown). In each treatment I(L)/I0ould be well fitted by a negative exponential curve with numberf days after start of the experiment and reaching a plateau in thend (r2 = 0.99 for all treatments)

(L)/I0 = 1 − e−ad (2)

here a is saturating coefficient, d is number of days after start ofhe experiment.

Daily canopy intercepted PPFD was calculated as the productf the interpolated dailyI(L)/I0 multiplied by the measured DLI.

into component variables. Brackets following each component indicates abbrevi-ations and units. Scheme is a modification of Fig. 2 presented by Higashide andHeuvelink (2009).

Integrating the daily canopy intercepted PPFD during the desig-nated growing period yields the sum of intercepted PPFD.

2.5. Net leaf photosynthesis measurements

At 17 weeks after start of the experiment, net leaf photosyn-thesis rates were measured with a portable gas exchange deviceequipped with a leaf chamber fluorometer (LI-6400; LI-COR, Lin-coln, USA). One fully expanded leaf of each cultivar was randomlyselected from each cultivation table in each treatment for measur-ing gas exchange at 500 followed by 100 �mol m−2 s−1 PPFD, themeasurements were taken when the photosynthesis rate reachedsteady state (after about 10 min). The light source (10% blueand 90% red) only illuminated the adaxial side of the leaf. Allmeasurements were carried out between 9:00 and 16:00. In themeasurement chamber, vapour pressure deficit (VPD) was main-tained in the range of 0.5–1 kPa, reference CO2 concentration wasset at 800 �mol mol−1, leaf temperature at 27 ◦C, these parameterswere close to that in the greenhouse compartments.

2.6. Plants ornamental quality determination

Plant compactness was determined based on the above groundfresh mass/plant height ratio at each destructive harvest. At 18weeks after start of the experiment, two plants of each cultivaron each cultivation table were randomly selected from each green-house compartment (12 plants) to determine the colour of flowersand leaves, and projected area of flowers and leaves. Top-viewimage of each plant was recorded with an image acquisition sys-tem consisting of a colour CCD-Camera (Hitachi HV C-20). Basedon the images, the CIE-L*a*b* colour space was used to quantifythe colour of flowers and leaves (Minolta, 1994); the CIE-L*a*b*space is defined as a sphere, it is typified by the lightness param-eter L* (lightness: black–white), and the colour co-ordinates a*(−a = green, +a = red) and b* (−b = blue, +b = yellow). Furthermore,the projected area of flowers and leaves were also quantified fromthese images.

2.7. Calculations and statistical analysis

Treatment or cultivar effects on yield of a plant can be analysedby breaking down the effect in different underlying components(Fig. 1). For example, higher aboveground fresh mass (aFM) can

T. Li et al. / Scientia Horticulturae 179 (2014) 306–313 309

Fig. 2. Photosynthetic photon flux density (PPFD) inside and outside of the green-house compartments on a typical clear day (25 Jul 2012) as measured by a pointsensor. Clear glass + Low DLI (daily light integral, mol m−2 d−1 PAR) represents 0%haze with DLI of 7.2 mol m−2 d−1; Diffuse glass + Low DLI represents 71% haze withDLI of 7.5 mol m−2 d−1; Diffuse glass + High DLI represents 71% haze with DLI of8.9 mol m−2 d−1. indicates the white sunscreen was fully closed in the low DLIaca

blmbgadctiiifegdg

alaobctovAbs

3

3

ai(w1it

Fig. 3. Relationship between relative PPFD (relative to the top of canopy) and LAIin ‘Royal Champion’ and ‘Pink Champion’ on cloudy days. Data presented in each

nd 50% closed in the high DLI treatment; indicates the white sunscreen was fullylosed in the high DLI treatment; indicates the DLI reached the limitation pointnd the blackout screen was closed.

e caused by higher aboveground dry mass (aDM) and/or byower aboveground dry mass content (aDMC, i.e. aboveground dry

ass/fresh mass). An increase in aboveground dry mass (aDM) cane explained by an increase in dry mass partitioning to above-round (aDMP, i.e. aboveground dry mass/total dry mass) and/orn increase in total dry mass (TDM). The latter results from a higherry mass production per unit intercepted PPFD (RUE) and/or higherumulative intercepted PPFD (Ii). An increase in dry mass produc-ion per unit intercepted PPFD (RUE) is determined by an increasen net photosynthetic rate (Pn) and/or by a decrease in fraction ofntercepted PPFD [I(L)/Io]. Higher cumulative intercepted PPFD (Ii)s directly linked with higher daily light integral (DLI) and/or higherraction of intercepted PPFD [I(L)/I0] which results from higher lightxtinction coefficient (K) and/or higher leaf area index (LAI). Above-round fresh mass (aFM), aboveground dry mass (aDM) and totalry mass (TDM) represent the accumulated biomass during therowing season.

Parameters determined from destructive harvests (aFM; aDM;DMC; TDM; aDMP; LAI; SLA; plant height; number of flowers,eaves and stems; plant compactness) were presented as the aver-ge of the last two measurements (18 and 21 weeks after the startf the experiment) in order to get more reliable information (num-er of replicates were doubled). The two destructive harvests wereonsidered as two blocks in statistical analysis. Treatment and cul-ivar effects on measured parameters were evaluated by analysisf variance (ANOVA). Differences between treatments and culti-ars in RUE were tested using a multiple linear regression model.ssuming replications in the same greenhouse compartment aseing independent. P-values smaller than 0.05 were regarded asignificantly different.

. Results

.1. Light distribution and interception

In the greenhouse with clear glass (0% haze) incident PPFD at given spot just above the plants fluctuated much stronger thann the greenhouse with diffuse glass (71% haze) on clear daysFig. 2). In all treatments, incident PPFD inside the greenhouse

−2 −1

as kept below 500 �mol m s whereas outside PPFD reached800 �mol m−2 s−1 in the middle of a clear day. On overcast days,

ncident PPFD at a given spot showed similar dynamic patterns inhe three compartments (data not shown).

cultivar were measurements from the three compartments. Each symbol representsthe average of six replicates measured in one day. Lines are fitted curves based onthe Lambert–Beer law [Eq. (1)].

Both on clear as well as cloudy days treatments did not affectcanopy PPFD interception, which was reflected by similar relativePPFD at the same LAI on clear days (Supplementary Fig. S4) andsimilar light extinction coefficient on cloudy days (SupplementaryTable S2). PPFD penetrated deeper into the canopy in ‘Pink Cham-pion’ than in ‘Royal Champion’ (Fig. 3), which was reflected by alower light extinction coefficient (K) in ‘Pink Champion’ [K = 1.01;95% confidence interval (0.99–1.05)] than in ‘Royal Champion’[K = 1.08; 95% confidence interval (1.03–1.14)].

3.2. Plant measurements

At low DLI, diffuse glass significantly increased the number ofleaves and stems in ‘Royal Champion’ compared with clear glasstreatment; while these effects did not occur in ‘Pink Champion’(Table 1). Under diffuse glass cover, high DLI significantly increasedthe number of leaves, flowers and stems in both cultivars comparedwith the low DLI (Table 1). Furthermore, high DLI increased plantheight in ‘Pink Champion’; while it decreased plant height in ‘RoyalChampion’ (Table 1).

In terms of plant ornamental quality (Table 2), high DLI resultedin more compact plants in both cultivars compared with low DLI.Furthermore, high DLI resulted in higher projected leaf area andfraction projected area flowers/area leaves in ‘Pink Champion’;while in ‘Royal Champion’, only higher projected leaf area wasobserved. The colour of leaves and flowers were not affected bythe diffuse glass cover as well as high DLI.

SLA and leaf angle were neither influenced by diffuse glass covernor high DLI. In all treatments, SLA in ‘Pink Champion’ was lowerthan in ‘Royal Champion’ (Supplementary Fig. S5). Moreover, ‘PinkChampion’ had positive leaf angle compared with the horizontalplane (0◦), while ‘Royal Champion’ had negative leaf angle (Sup-plementary Fig. S6).

Net leaf photosynthesis rates were not affected by the treat-ments when measured at the same conditions in leaf measurementchamber. ‘Pink Champion’ had significantly higher net photosyn-thesis rates than ‘Royal Champion’ at 500 �mol m−2 s−1 PPFD inall treatments; while this effect did not occur at 100 �mol m−2 s−1

PPFD (Supplementary Fig. S7).

3.3. Yield component analysis

Under low DLI, diffuse glass significantly increased abovegroundfresh mass (aFM), aboveground dry mass (aDM), total dry mass

(TDM), dry mass production per unit intercepted PPFD (RUE) andleaf area index (LAI) in ‘Royal Champion’ compared with clear glasstreatment. However, these effects did not occur in ‘Pink Champion’except aboveground dry mass content (aDMC) which was increased

310 T. Li et al. / Scientia Horticulturae 179 (2014) 306–313

Table 1The effect of diffuse glass cover and daily light integral (DLI, mol m−2 d−1 PAR) on plant growth parameters in two Anthurium cultivars (Royal Champion and Pink Champion)(n = 24). Data represent the average of the last two harvests (18 and 21 weeks after start of the experiment).

Treatment Number of leaves (no. plant−1) Number of flowers (no. plant−1) Number of stems (no. plant−1) Plant height (cm)

‘Royal Champion’Clear glass + Low DLI1 29.0a 9.0a 2.7a 29.6bDiffuse glass + Low DLI2 36.0b 9.3a 3.1b 28.2bDiffuse glass + High DLI3 41.1c 10.4b 3.6c 25.8aP-value <0.001 <0.001 <0.001 <0.001‘Pink Champion’Clear glass + Low DLI 35.6a 8.7a 3.9a 33.7abDiffuse glass + Low DLI 39.0a 9.3a 4.0a 32.2aDiffuse glass + High DLI 44.5b 11.6b 4.5b 35.0bP-value <0.001 <0.001 0.04 0.002

Means of each cultivar followed by different letters within one column differ significantly.

bc

adwdm

malmc

4

4l

baVf

TTlt

M

1 0% haze with DLI of 7.2 mol m−2 d−1.2 71% haze with DLI of 7.5 mol m−2 d−1.3 71% haze with DLI of 8.9 mol m−2 d−1.

y 5% in the diffuse glass treatment (Fig. 4). Absolute values of allomponent variables are presented in the Supplementary Table S3.

Under diffuse glass cover, high DLI significantly increasedboveground fresh mass (aFM), aboveground dry mass (aDM), totalry mass (TDM) and leaf area index (LAI) in both cultivars comparedith low DLI (Fig. 4). Additionally, high DLI decreased abovegroundry mass content (aDMC) by 3% and increased aboveground dryass partitioning (aDMP) by 2% in ‘Pink Champion’.In all treatments, ‘Pink Champion’ had higher aboveground fresh

ass (aFM), aboveground dry mass (aDM), total dry mass (TDM)nd dry mass production per unit intercepted PPFD (RUE), andower aboveground dry mass content (aDMC), aboveground dry

ass partitioning (aDMP) and light extinction coefficient (K) inomparison with ‘Royal Champion’ (Fig. 5).

. Discussion

.1. Diffuse glass cover improves temporal rather than verticalight distribution

Scattering the direct component of solar light in the greenhouse

y diffuse glass cover is an interesting way to improve plant growthnd production of greenhouse crops (Dueck et al., 2012; Garcíaictoria et al., 2011; Hemming et al., 2007; Li et al., 2014). Dif-

use glass cover strongly reduced short term fluctuations in PPFD

able 2he effect of diffuse glass cover and daily light integral (DLI, mol m−2 d−1 PAR) on colour veaves and flowers (n = 12), as well as the plant compactness (aboveground fresh mass/plahe experiment except compactness determination which was done based on the last two

Treatment L* (lightness)1 a* (green–red)2 b* (blue–ye

Flower Leaf Flower Leaf Flower

‘Royal Champion’Clear glass + Low DLI4 35.7 30.2 20.9 −5.3 15.7

Diffuse glass + Low DLI5 35.1 31.4 19.8 −4.8 15.4

Diffuse glass + High DLI6 36.1 31.2 19.9 −4.6 16.0

P-value 0.38 0.08 0.26 0.07 0.59

‘Pink Champion’Clear glass + Low DLI 50.6 28.3 30.4 −5.6 17.1

Diffuse glass + Low DLI 51.2 28.6 30.2 −4.9 17.7

Diffuse glass + High DLI 52.3 28.0 30.6 −5.1 17.8

P-value 0.20 0.29 0.89 0.09 0.14

eans of each cultivar followed by different letters within one column differ significantly1 Lightness parameter L* [lightness: black–white].2 Colour co-ordinates a* (−a = green, +a = red).3 Colour co-ordinates b* (−b = blue, +b = yellow).4 0% haze with DLI of 7.2 mol m−2 d−1.5 71% haze with DLI of 7.5 mol m−2 d−1.6 71% haze with DLI of 8.9 mol m−2 d−1.

at a given spot in the greenhouse on clear days (Fig. 2), becausethe diffuse glass minimized the effects of local shade by construc-tion parts, equipment and overstory leaves. In terms of spatial lightdistribution, we observed a similar relationship of canopy PPFDinterception in response to LAI between the clear and diffuse glasstreatments on clear days (Supplementary Fig. S4), which suggestsdiffuse glass treatment had no effect on vertical light distribu-tion in Anthurium. This contradicts with the finding in cucumber(Hemming et al., 2007) and tomato (Li et al., 2014) in which dif-fuse glass cover resulted in deeper PPFD penetration within thecanopies on clear days. This phenomenon could be explained bythe characteristics of canopy architecture which usually plays a piv-otal role for canopy PPFD interception (Falster and Westoby, 2003;Sarlikioti et al., 2011a; Valladares and Niinemets, 2007). Anthuriumpot-plants are characterised by short and compact canopies withrelatively large leaves (Supplementary Fig. S8), resulting in sub-stantial leaf overlap and self-shading especially when LAI is high.Self-shading decreases the net amount of leaf area exposed tolight and this leads to poor light distribution even under diffuselight condition (Falster and Westoby, 2003). Therefore, this typeof canopy structure is less responsive for scattering the light, and

limits the potential effect of diffuse glass cover on canopy lightdistribution. Furthermore, shading screens were applied in all thegreenhouse compartments, which already transformed a portionof direct PPFD into diffuse (10%). Therefore, we conclude that

alues of flowers and leaves in the CIE-L*a*b* colour space and the projected area ofnt height ratio) (n = 24). All measurements were done at 18 weeks after the start of

destructive harvest measurements (18 and 21 weeks after start of the experiment).

llow)3 Projected leaf area(cm2)

Fraction projectedarea flowers/arealeaves

Compactness(g cm−1)

Leaf

4.3 694a 0.26 5.7a3.8 710ab 0.28 6.4b4.6 766b 0.28 8.0c0.26 0.04 0.68 <0.001

5.8 729a 0.13a 5.8a5.6 767a 0.14ab 6.0a6.3 828b 0.16b 7.3b0.11 0.003 0.03 <0.001

.

T. Li et al. / Scientia Horticultur

Fig. 4. Effects of diffuse glass cover [diffuse glass vs. clear glass, both at low DLI (dailylight integral, mol m−2 d−1 PAR)] or high DLI (low DLI vs. high DLI, both under diffuseglass cover) on the yield components in ‘Royal Champion’ and ‘Pink Champion’. *

P < 0.05, ** P < 0.01 and *** P < np indicates statistical analysis was not possible. Relativedifference in net photosynthetic rate (Pn) was not determined. For explanation ofabbreviations of the variables see Fig. 1.

Fig. 5. Cultivar effects on the yield components in ‘Pink Champion’ and ‘Royal Championmol m−2 d−1 PAR) represents 0% haze with DLI of 7.2 mol m−2 d−1; Diffuse glass + Low DLI r71% haze with DLI of 8.9 mol m−2 d−1. * P < 0.05, ** P < 0.01 and *** P < 0.001. np indicates s(Pn) was not determined. For explanation of abbreviations of the variables see Fig. 1.

ae 179 (2014) 306–313 311

diffuse glass cover had negligible effect on vertical PPFD distribu-tion within the canopy of Anthurium pot-plants mainly because ofthe short and compact canopy structures, as well as the experimen-tal management practice.

4.2. Effect of diffuse glass cover on plant growth is cultivar specific

Total dry mass (TDM) was 14% higher in the diffuse glass than inthe clear glass treatment in ‘Royal Champion’, which lead to higheraboveground dry mass (aDM). Consequently the aboveground freshmass (aFM) was also higher as the aboveground dry mass content(aDMC) was not significantly affected by the treatments (Fig. 4).The increased biomass production was attributed to a higher (5%)cumulative intercepted PPFD (Ii) and a higher (8%) dry mass produc-tion per unit intercepted PPFD (RUE) (Fig. 4). The former resultedmainly from an unexpected higher (4%) average DLI in the dif-fuse glass treatment, which may have been caused by a higher(about 3%) light transmission during condensation under diffuseglass cover (unpublished data) or the side-effects of neighbouringgreenhouses. In spite of 5% more cumulative intercepted PPFD (Ii),diffuse glass still increased RUE by 8%. This could be explainedby leaf photosynthesis because the fraction of intercepted PPFD[I(L)/I0] played a negligible role (1% difference between treatments)(Fig. 4). However, under steady state measurement conditions, netleaf photosynthetic rates of the fully expanded leaves were simi-lar in both treatments (Supplementary Fig. S7). Maybe there weretreatment differences in dynamic leaf photosynthesis, as dynamicphotosynthesis can be affected by the dynamic incident light dis-tribution (Pearcy et al., 2004; Tinoco-Ojanguren and Pearcy, 1992).This needs to be confirmed by further study. In ‘Pink Champion’,however, diffuse glass treatment had no effect on dry mass produc-tion per unit intercepted PPFD (RUE), as well as fresh and dry massproduction (Fig. 4). The different response of these two cultivars todiffuse glass treatment needs to be further explored.

4.3. Increasing DLI under diffuse glass cover not only stimulatesplant growth, but also improves plant ornamental quality

DLI is an important variable to determine plant growth anddevelopment (Marcelis et al., 2006; Warner and Erwin, 2003). The

minimum DLI in this study was 7.2 mol m−2 d−1 which was alreadyhigher than the commercial practice (about 5 mol m−2 d−1), this isbecause air humidity was kept at moderately high level (75–80%) inall greenhouse compartments which could result in open stomata

’. Expressed relative to ‘Royal Champion’. Clear glass + Low DLI (daily light integral,epresents 71% haze with DLI of 7.5 mol m−2 d−1; Diffuse glass + High DLI representstatistical analysis was not possible. Relative difference in net photosynthetic rate

3 icultur

(madgtcma(

(im(fdDRrFaDt

traChphdcuissttp

4mi

aw(gmd

lhFwuCtlbccc

12 T. Li et al. / Scientia Hort

Aphalo and Jarvis, 1991), thereby avoiding overheating leaves andaintaining photosynthesis. Furthermore, an increase of DLI may

lso lead to increase in temperature which could promote plantevelopment and flowering (Oh et al., 2009); while in this studyreenhouse temperature in all treatments were maintained equal,herefore, the DLI and temperature interaction effect need not to beonsidered. Additionally, this experiment did not include the treat-ent of high DLI (10 mol m−2 d−1) under clear glass cover, because

similar treatment in an earlier experiment resulted in leaf damageVan Noort et al., 2011).

Increasing DLI by 19% resulted in 20–23% higher total dry massTDM) (comparing DLI under diffuse glass cover), this resulted in anncrease in aboveground dry mass (aDM) and aboveground fresh

ass (aFM) (Fig. 4), as well as more stems, leaves and flowersTable 1). The effect of high DLI on plant growth mainly resultedrom a higher cumulative intercepted PPFD (Ii), while dry mass pro-uction per unit intercepted PPFD (RUE) was not affected by highLI treatment (Fig. 4). In line with the absence of an effect of DLI onUE, there was also no effect of DLI on the net leaf photosynthesisates when measured at the same PPFD (Supplementary Fig. S7).urthermore, the fraction of intercepted PPFD [I(L)/I0] was also notffected by the treatments, although LAI was increased by the highLI. This is because LAI in both treatments was already high and

herefore not limiting light interception.Plant shape and colour are important parameters determining

he quality of pot-plants. Increasing DLI from 7.5 to 8.9 mol m−2 d−1

esulted in more compact plants in both cultivars as indicated by higher ratio of plant fresh mass to height (Table 2). In ‘Royalhampion’, the more compact plants as a result of a lower planteight and higher aboveground fresh mass (aFM). In ‘Pink Cham-ion’, plants were also more compact even though plant height wasigher at high DLI (Table 1). Increasing DLI under diffuse glass coverid not induce damage of flowers and leaves as indicated by theolour values (Table 2). When Van Noort et al. (2011) increased DLInder clear glass to a similar level as in our study, mild discolour-

ng of leaves occurred in Anthurium ‘Baby talk’ and severe necroticpots in leaves of Guzmania ‘Hilda’, even though plant growth wastimulated. The possible reason of this leaf damage was that theirreatments were conducted under a conventional glass type wherehe plants were exposed to a higher occurrence of more severe localeaks in light intensity.

.4. Differences in plant growth between the two cultivars isainly due to differences in dry mass production per unit

ntercepted PPFD (RUE)

In all the treatments, ‘Pink Champion’ had significantly higherboveground fresh mass (aFM) than ‘Royal Champion’ (Fig. 5),hich was explained by a lower aboveground dry mass content

aDMC, i.e. aboveground dry mass/fresh mass) and a higher above-round dry mass (aDM). The latter resulted from higher total dryass (TDM) which was determined by an increased dry mass pro-

uction per unit intercepted PPFD (RUE).RUE can be influenced by plant morphological and physio-

ogical properties (Sinclair and Muchow, 1999). ‘Pink Champion’ad leaves pointing upward (positive leaf angle; Supplementaryig. S6), while leaves in ‘Royal Champion’ were pointing down-ard which might generate more self-shading compared with thepward leaves (Falster and Westoby, 2003). Furthermore, ‘Royalhampion’ was shorter than ‘Pink Champion’ (Table 1), this leado more foliage packing, thereby, self-shading. The differences ineaf angle and plant height resulted in different spatial PPFD distri-

ution within the canopy as indicated by a lower light extinctionoefficient (K) in ‘Pink Champion’ than in ‘Royal Champion’ (Fig. 3),onsequently, lower fraction of intercepted PPFD [I(L)/I0] whichontributed to higher RUE in ‘Pink Champion’. Furthermore, these

ae 179 (2014) 306–313

two cultivars are characterized by relatively large area of flowerswhich are standing above the canopy (Supplementary Fig. S8), thiscould exert a negative effect on RUE because flowers can inter-cept a portion of PPFD which cannot be utilized for photosynthesis.This negative effect is likely to be smaller in ‘Pink Champion’ dueto less area of flowers (Table 2, Supplementary Fig. S8). Addition-ally, ‘Pink Champion’ had higher net leaf photosynthesis rates than‘Royal Champion’, though this difference occurred only at high PPFD(Supplementary Fig. S7). All these together resulted in the higherRUE in ‘Pink Champion’.

4.5. Implications for commercial production

Although shading is indispensable for production of pot-plantsin order to avoid light damage, we have shown that allowing morelight under diffuse glass cover (71% haze) in combination withmoderately high levels of air humidity (75–80%) not only stimu-lates plant growth but also improves plant ornamental quality (i.e.more compact plants without light damage) in Anthurium pot-plants. The realized DLI in the reference treatment (clear glass,7.2 mol m−2 d−1) in the present study was already higher than thecommercial practice (about 5 mol m−2 d−1). Growers regularly vis-ited the experiment, and estimated that plants even in the referencetreatment were growing 25% faster than their plants. A furtherincrease of DLI by 24% (8.9 mol m−2 d−1) under diffuse glass coverresulted in additional 27–37% increase in dry mass production.Therefore, it is clear that our strategy can reduce the cultivationtime to reach the marketable value.

5. Conclusions

Our conclusions are: (1) the stimulating effect of diffuse glasscover on dry mass production per unit intercepted PPFD (RUE) inAnthurium pot-plants is cultivar specific. (2) Increasing DLI underdiffuse glass cover not only stimulates biomass production, but alsoimproves plant ornamental quality (i.e. more compact plants with-out light damage). (3) Differences in plant growth between the twoAnthurium cultivars mainly resulted from difference in RUE.

Acknowledgements

The authors would like to thank Chinese Scholarship Councilfor awarding a scholarship to T. Li; Dr. LY. Chang for helping thedestructive harvest measurements; Elias Kaiser, Geert Van Geestand Craig Taylor for their comments on the manuscript. We alsothank for support from Powerhouse and the programmes TowardsBiosolar Cells and Greenhouse as energy source as funded by theMinistry of Economic Affairs and the Product Board for Horticulture.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.scienta.2014.09.039.

References

Aphalo, P.J., Jarvis, P.G., 1991. Do stomata respond to relative humidity? Plant, CellEnviron. 14, 127–132.

Asada, K., 1999. The water–water cycle in chloroplasts: scavenging of active oxygensand dissipation of excess photons. Annu. Rev. Plant Biol. 50, 601–639.

ASTM International. 2007. Standard Test Method for Haze and Luminous Transmit-tance of Transparent Plastics, D1003-07. West Conshohocken, USA.

Baeza, E., López, J.C., 2012. Light transmission through greenhouse covers. Acta Hor-tic. 956, 425–440.

Barthelemy, D., Caraglio, Y., 2007. Plant architecture: a dynamic, multilevel andcomprehensive approach to plant form, structure and ontogeny. Ann. Bot. 99,375–407.

icultur

D

F

F

F

F

G

G

H

H

H

J

K

K

K

L

L

L

M

M

T. Li et al. / Scientia Hort

ueck, T.A., Janse, J., Li, T., Kempkes, F.L.K., Eveleens, B., 2012. Influence of diffuseglass on the growth and production of tomato. Acta Hortic. 956, 75–82.

alster, D.S., Westoby, M., 2003. Leaf size and angle vary widely across species: whatconsequences for light interception? New Phytol. 158, 509–525.

arquhar, G.D., Roderick, M.L., 2003. Pinatubo, diffuse light, and the carbon cycle.Science 299, 1997–1998.

ausey, B.A., Heins, R.D., Cameron, A.C., 2005. Daily light integral affects floweringand quality of greenhouse-grown achillea, gaura, and lavandula. HortScience 40,114–118.

aust, J.E., Holcombe, V., Rajapakse, N.C., Layne, D.R., 2005. The effect of daily lightintegral on bedding plant growth and flowering. HortScience 40, 645–649.

arcía Victoria, N., Kempkes, F.L.K., Van Weel, P., Stanghellini, C., Dueck, T., Bruins, M.,2011. Effect of a diffuse glass greenhouse cover on rose production and quality.Acta Hortic. 952, 241–248.

u, L., Baldocchi, D., Verma, S.B., Black, T.A., Vesala, T., Falge, E.M., Dowty, P.R.,2002. Advantages of diffuse radiation for terrestrial ecosystem productivity. J.Geophys. Res. 107 (2-1-2-23).

emming, S., Dueck, T.A., Janse, J., van Noort, F., 2007. The effect of diffuse light oncrops. Acta Hortic. 801, 1293–1300.

emming, S., Mohammadkhani, V., Dueck, T.A., 2008. Diffuse greenhouse coveringmaterials-material technology, measurements and evaluation of optical prop-erties. Acta Hortic. 797, 469–475.

igashide, T., Heuvelink, E., 2009. Physiological and morphological changes overthe past 50 years in yield components in tomato. J. Am. Soc. Hortic. Sci. 134,460–465.

olliffe, P.A., Tarimo, A.J.P., Eaton, G.W., 1990. Plant growth analysis: growth andyield component responses to population density in forage maize. Ann. Bot. 65,139–147.

asahara, M., Kagawa, T., Oikawa, K., Suetsugu, N., Miyao, M., Wada, M., 2002.Chloroplast avoidance movement reduces photodamage in plants. Nature 420,829–832.

empkes, F.L.K., Stanghellini, C., García Victoria, N., Bruins, M., 2011. Effect of diffuseglass on climate and plant environment: first results from an experiment onroses. Acta Hortic. 952, 255–262.

romdijk, J., van Noort, F., Driever, S., Dueck, T.A., 2012. An enlightened view onprotected cultivation of shade-tolerant pot-plants: benefits of higher light levels.Acta Hortic. 956, 381–388.

awlor, D.W., 1995. Photosynthesis, productivity and environment. J. Exp. Bot. 46,1449–1461.

i, T., Heuvelink, E., Dueck, T.A., Janse, J., Gort, G., Marcelis, L.F.M., 2014. Enhancementof crop photosynthesis by diffuse light: quantifying the contributing factors.Ann. Bot. 114 (1), 145–156.

ong, S.P., Humphries, S., Falkowski, P.G., 1994. Photoinhibition of photosynthesisin nature. Annu. Rev. Plant Biol. 45, 633–662.

arcelis, L.F.M., Broekhuijsen, A.G.M., Meinen, E., Nijs, E.M.F.M., Raaphorst, M.G.M.,

2006. Quantification of the growth response to light quantity of greenhousegrown crops. Acta Hortic. 711, 97–103.

ercado, L.M., Bellouin, N., Sitch, S., Boucher, O., Huntingford, C., Wild, M., Cox, P.M.,2009. Impact of changes in diffuse radiation on the global land carbon sink.Nature 458, 1014–1017.

ae 179 (2014) 306–313 313

Minolta Co. 1994. Exakte farbkommunikation. vom farbgefuhl bis zur objektivenfarbmessung. Company brochure.

Monsi, M., Saeki, T., 2005. On the factor light in plant communities and its importancefor matter production, 1953. Ann. Bot. 95, 549–567.

Niyogi, K.K., 1999. Photoprotection revisited: genetic and molecular approaches.Annu. Rev. Plant Biol. 50, 333–359.

Oh, W., Cheon, I.H., Kim, K.S., Runkle, E.S., 2009. Photosynthetic daily light inte-gral influences flowering time and crop characteristics of Cyclamen persicum.HortScience 44, 341–344.

Pearcy, R.W., Krall, J.P., Sassenrath-Cole, G.F., 2004. Photosynthesis in fluctuatinglight environments. In: Neil R. Baker (Ed.), Photosynthesis and the Environment.Springer, The Netherlands.

Plénet, D., Mollier, A., Pellerin, S., 2000. Growth analysis of maize field crops underphosphorus deficiency. II. Radiation-use efficiency, biomass accumulation andyield Components. Plant Soil 224, 259–272.

Poorter, H., Anten, N.P., Marcelis, L.F.M., 2013. Physiological mechanisms in plantgrowth models: do we need a supra-cellular systems biology approach? PlantCell Environ. 36, 1673–1690.

Sarlikioti, V., de Visser, P.H.B., Marcelis, L.F.M., 2011a. Exploring the spatial dis-tribution of light interception and photosynthesis of canopies by means of afunctional–structural plant model. Ann. Bot. 107, 875.

Sarlikioti, V., de Visser, P.H.B., Buck-Sorlin, G.H., Marcelis, L.F.M., 2011b. How plantarchitecture affects light absorption and photosynthesis in tomato: towards anideotype for plant architecture using a functional–structural plant model. Ann.Bot. 108, 1065–1073.

Scuderi, D., Giuffrida, F., Leonardi, C., 2013. Growth and visual quality ofweeping fig in relation to shading conditions. J. Food Agric. Environ. 11,1828–1832.

Scuderi, D., Giuffrida, F., Toscano, S., Romano, D., 2012. Growth, physiologicalresponse, and quality characteristics of weeping fig in response to shading levelsand climatic conditions. HortScience 47, 1586–1592.

Sinclair, T.R., Muchow, R.C., 1999. Radiation use efficiency. Adv. Agron. 65,215–265.

Sultan, S.E., 2000. Phenotypic plasticity for plant development, function and lifehistory. Trends Plant Sci. 5, 537–542.

Tinoco-Ojanguren, C., Pearcy, R.W., 1992. Dynamic stomatal behavior and its role incarbon gain during lightflecks of a gap phase and an understory piper speciesacclimated to high and low light. Oecologia 92, 222–228.

Urban, O., Klem, K., Ac, A., Havrankova, K., Holisova, P., Navratil, M., Zitova, M.,Kozlova, K., Pokorny, R., Sprtova, M., 2012. Impact of clear and cloudy sky condi-tions on the vertical distribution of photosynthetic CO2 uptake within a sprucecanopy. Funct. Ecol. 26, 46–55.

Valladares, F., Niinemets, U., 2007. The architecture of plant crowns: from designrules to light capture and performance. Funct. Plant Ecol., 101–149.

Van Noort, F., Kempkes, FLK. and De Zwart, F.;1; 2011. New Cultivation Practices Pot-

Plants. Wageningen UR Greenhouse Horticulture. Report. GTB-1093 (in Dutch).

Warner, R.M., Erwin, J.E., 2003. Effect of photoperiod and daily light integral onflowering of five Hibiscus sp. Sci. Hortic. 97, 341–351.

Way, D.A., Pearcy, R.W., 2012. Sunflecks in trees and forests: from photosyntheticphysiology to global change biology. Tree Physiol. 32, 1066–1081.