K

Aa

Bb

c

a

ARR1A

KMPYKK

1

aanec1

w

(

0d

Field Crops Research 120 (2011) 360–369

Contents lists available at ScienceDirect

Field Crops Research

journa l homepage: www.e lsev ier .com/ locate / fc r

ernel number and kernel weight determination in dent and popcorn maize

lan D. Severinia,∗, Lucas Borrásb, Mark E. Westgatec, Alfredo G. Ciriloa

Grupo de Ecofisiología y Agrometeorología, Estación Experimental Agropecuaria Pergamino, Instituto Nacional de Tecnología Agropecuaria, Ruta 32, km 4.5, CP 2700 Pergamino,uenos Aires, ArgentinaDepartamento de Producción Vegetal, Facultad de Ciencias Agrarias, Universidad Nacional de Rosario, S2125ZAA Zavalla, Santa Fe, ArgentinaDepartment of Agronomy, Iowa State University, 1577 Agronomy Hall, Ames, IA 50011-1010, USA

r t i c l e i n f o

rticle history:eceived 22 June 2010eceived in revised form1 November 2010ccepted 12 November 2010

eywords:aize

opcornield componentsernel numberernel weight

a b s t r a c t

Yield formation in maize (Zea mays L.) dent hybrids has been directly linked to the rate of plant biomassaccumulation and partitioning of assimilates to the developing grain. Maize popcorn genotypes havebeen studied less extensively, but their kernels are known to differ in terms of endosperm structureand typical growth patterns. Our objective was to evaluate how variation in plant growth rate (PGR) atdifferent stages of kernel formation and development affected kernel number per plant (KNP), individualkernel weight (KW) and rate and duration of kernel growth in popcorn genotypes, relative to dent ones.We conducted three experiments (two in Ames, Iowa, and one in Pergamino, Argentina) in which PGRsaround flowering and during the linear phase of the grain-filling period of four dent and eight popcorngenotypes were altered by plant density, defoliations and thinning treatments.

Yield per plant, KNP, KW, rate and duration of kernel growth all showed significant kernel type (pop-corns vs. dents) effects (p < 0.01). KNP was highly correlated with ear biomass accumulated aroundflowering in dents and popcorns, and popcorns showed a higher efficiency for setting kernels per unitof ear biomass accumulated around flowering (p < 0.01). Popcorn inbred R18 in particular showed a sig-nificantly higher efficiency, consistent across experiments. Relationships between potential KW at earlygrain filling or kernel growth rate and the PGR per kernel around flowering were different for dent andpopcorn genotypes. Most popcorns established a lower potential KW compared to dent genotypes at sim-

ilar PGRs per kernel around flowering. Also, popcorn kernels were less prone to decrease KW in responseto severe reductions in plant growth during the linear phase of the grain-filling period as promoted bydefoliation treatments (significant kernel type × source manipulation treatment interaction, p < 0.001).Despite different patterns of KNP and KW determination, yield variation across dent and popcorn geno-types and environments corresponded closely to the potential sink capacity established by the end ofthe lag phase 14 days after anthesis. This result emphasizes the importance of the flowering period to

ss di

establish KN and KW acro

. Introduction

Maize (Zea mays L.) grain yield formation is often studied asfunction of harvested kernels per unit land area and the aver-

ge individual kernel weight. Trait dissection of maize kernel

umber and kernel weight determination has directed researchfforts towards understanding their relationship with plant androp growth at different crop developmental stages (Andrade et al.,999; Borrás and Gambín, 2010). Studies relating plant growth with

Abbreviations: PGR, plant growth rate; KNP, kernel number per plant; KW, kerneleight; WC, water content; MC, moisture concentration; DAA, days after anthesis.∗ Corresponding author. Tel.: +54 2477 439014.

E-mail addresses: aseverini@pergamino.inta.gov.ar, alanseverini@hotmail.comA.D. Severini).

378-4290/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.fcr.2010.11.013

fferent maize germplasm.© 2010 Elsevier B.V. All rights reserved.

reproductive development (i.e., number and size of kernels formed)have focused primarily on dent germplasm. In the present study,we expanded this analysis to include popcorn genotypes, focusingon possible differences in developmental patterns that determinegrain yield components across a more diverse germplasm.

It is well established that the number of kernels set per plantvaries with plant growth around flowering (Tollenaar et al., 1992;Andrade et al., 1999). The number of kernels per plant that a maizehybrid or inbred will set at a given plant growth rate dependsupon the biomass partitioning to the developing ear (Andrade et al.,1999). A slower rate of ear growth, associated with slow plant

growth or less partitioning of assimilates to the ear, results inreduced kernel set. Moreover, Echarte and Tollenaar (2006) andD’Andrea et al. (2009) documented genotypic differences in kernelset per unit of ear biomass accumulated around flowering. Infor-mation on biomass partitioning to the ear for dent germplasm

ps Re

hiif2mote

tcirpaots1balcwlhkoptdtWi

oBgidupi2irpaeeIo

aaoirtmbe(ps

1993) and ca. 20 days before flowering. Pests, weeds and plant dis-eases were controlled by standard agronomic practice throughout

A.D. Severini et al. / Field Cro

as increased dramatically in recent years due to its importancen understanding the physiological basis for the anthesis–silkingnterval and genotype × environment interactions affecting yieldormation (Edmeades et al., 1993; Echarte et al., 2004; Borrás et al.,007). Although there is a general understanding that popcorns areore prolific than dents (Ziegler, 2001), there is also a basic lack

f knowledge on how efficient they are for biomass partitioning tohe ear around flowering or how partitioning to the ear relates tofficiency of kernel set for popcorns.

Kernel weight at physiological maturity depends on the poten-ial kernel size established early in grain filling, and the plantapacity to provide assimilates needed to fulfill this potential dur-ng grain filling (Borrás and Westgate, 2006). Recent studies haveevealed that kernel water relations are strongly associated withotential and final kernel weight, and may mediate environmentalnd genetic effects on this yield component. During the first periodf grain filling, called the lag phase, the number of starch deposi-ion sites is established (i.e., the number of cells per endosperm andtarch granules per cell) (Reddy and Daynard, 1983; Jones et al.,996). Dry matter accumulation is almost nil during this period,ut water accumulation is rapid, driving endosperm expansionnd increasing potential sink size. Kernels continue to accumu-ate water until about mid grain fill, when kernel maximum waterontent (MWC) is achieved. Borrás et al. (2003) showed that MWCas a fairly reliable predictor of potential kernel weight and was

inearly related to the kernel growth rate in dent inbred lines andybrids grown under favorable conditions. As grain filling advances,ernels progressively desiccate and moisture concentration (MC;n a fresh weight basis) provides an accurate measure of therogress towards physiological maturity. Kernel moisture concen-ration at physiological maturity appears to be fairly stable amongent germplasm (Gambín et al., 2007). However, popcorn geno-ypes typically matured at lower MC values than dents (Borrás and

estgate, 2006; Borrás et al. 2009). The origins of these differencesn MC at maturity are unknown.

Plants adjust their seed number and potential seed size basedn growth conditions around flowering (Sadras, 2007; Gambín andorrás, 2010). The close relationships between MWC and kernelrowth rate with assimilate supply during the lag phase stronglymplicate source availability early in grain filling as an importanteterminant in establishing potential kernel sink capacity – andltimately KW (Borrás and Gambín, 2010). Plant growth rate (PGR)er kernel around flowering has been used as a surrogate of assim-

late availability per kernel during the lag phase (Gambín et al.,006, 2008). Kiniry et al. (1990), however, found no KW response

n popcorn genotypes when KNP was reduced and plant growthate per kernel was increased at flowering. Greater resistance toericarp expansion resulting from greater pericarp thickness (Tracynd Galinat, 1987) or a different content of extensin proteins (Hoodt al., 1991) could limit the responsiveness of popcorn kernelxpansion to changes in assimilate supply early in grain filling.f so, their relationship with plant growth might differ from thatbserved in rapidly expanding kernels of dent genotypes.

We evaluated how KNP, KW and kernel growth patterns areffected by changes in plant growth for several contrasting popcornnd dent inbreds and hybrids. We tested whether the allocationf biomass to reproductive structures follows the same patternn popcorn and dent genotypes through a series of expectedesponses: (i) the number of kernels set per plant will depend onhe biomass allocated to the growing ear around flowering, with

inimum differences in the number of kernels set per unit of eariomass when genotypes are compared, (ii) PGR per kernel at flow-ring will explain differences in potential KW for all genotypes, and

iii) dent and popcorn genotypes will reduce their KW to a similarroportion (relative reduction) when source strength is reduced toimilar levels during the linear phase of the grain-filling period.

search 120 (2011) 360–369 361

2. Materials and methods

Two field experiments (Exps. I and II) were carried outat Iowa State University, Ames, USA, at the Brunner Farmduring 2007, and a third study (Exp. III) was conducted atINTA, Pergamino, Argentina, during the 2007/2008 growing sea-son. In Exps. I and II, we used a set of public inbred linesthat were selected for diverse KW and kernel growth patternsfrom a previous screening (Borrás et al., 2009). Based on dif-ferences in above ground biomass per plant (small, mediumand large plant sizes) and in kernel sizes, selected genotypeswere: R18 (small plant type, highly prolific, ∼100 mg kernel−1,popcorn), IDS69 (small plant type, ∼120 mg kernel−1, popcorn),IDS91 (medium plant type, ∼120 mg kernel−1, popcorn), B73(large plant type, ∼270 mg kernel−1, dent), Mo17 (large planttype, ∼310 mg kernel−1, dent) and N209 (medium plant type,∼270 mg kernel−1, dent). In Exp. II genotypes consisted of six pop-corn (95:2, IDS69, IDS91, R18, R-28-2 and R-53-1) and two dent(B73 and N209) inbred lines. A full description of all these publicgenotypes can be found at www.ars-grin.gov (verified 1 October,2009). Genotypes R18, R-28-2 and R-53-1 are classified as popcorngenotypes, as they pop producing flakes when heated, althoughthey were not specifically developed as inbreds for producingcommercial popcorn hybrids. In Exp. III, we used three commer-cial maize hybrids. Two were popcorn genotypes (P625 and P802,Agricultural Alumni Seed Improvement Association Inc.) and onewas a dent (AW190, Monsanto Argentina) kernel type. Exps. I andIII were conducted for evaluating the effect of altered plant growthon KNP, KW and kernel growth patterns; Exp. II was designed tocompare genotypes at a single uniform plant density of 9 pl m−2.

Treatments in Exp. I were a factorial combination of (i) geno-types, (ii) two plant densities (3 and 9 pl m−2), (iii) a defoliationtreatment applied at the beginning of grain filling (17 d after 50%anthesis (DAA)), and (iv) control at each plant density. Exp. III wasplanted at 9 pl m−2 and consisted on five treatments designed toalter plant growth rate: (i) a defoliation treatment applied at ca.15 d before 50% anthesis (DBA), (ii) 50% plant thinning at ca. 15DBA, (iii) a defoliation at 17 DAA, (iv) 50% plant thinning applied at17 DAA, and (v) control treatment in which leaf area and plant den-sity were not altered. Defoliation treatments involved removing allleaves starting from the bottom of the plant leaving only the topthree or four leaves. Thinning treatments were done by removingevery other plant within the row. Both high plant density (Exp. I)and defoliation treatments (Exps. I and III) were intended to reducePGR. Low plant density (Exp. I) and thinning treatments (Exp. III)were imposed to increase PGR around flowering or during grainfilling. Exps. I and II were planted 11 May and 25 May, respectively,in a randomized complete block design with three replicates. Exp.III was planted 5 October in a split-plot design with three repli-cates, where treatments were the main plots and genotypes thesub-plots. In Exp. I, each plot consisted of 6 (high density) or 8 (lowdensity) rows, 5.5 m long and 0.76 m apart; in Exp. II, plots were6 rows wide, 5.5 m long and 0.76 m apart. In Exp. III, plots were5 (control and defoliation treatments) or 7 (thinning treatments)rows wide, 7 m long and 0.70 m apart. In all cases, plots were overplanted and thinned at the 3-leaf stage (Ritchie et al., 1993) to thedesired plant density. Exps. I and II were rain-fed, but timely rainsoccurred and plants showed no signs of water stress. Exp. III wasirrigated through a sprinkler system. Nitrogen was applied beforeplanting in Exps. I and II with 110 kg N ha−1. Nitrogen was appliedat 100 kg N ha−1 twice in Exp. III: at the 4-leaf stage (Ritchie et al.,

the growth cycle in all experiments.Individual kernel dry weight and water content were mea-

sured throughout kernel development beginning 10 days after 50%

3 ps Re

st(2wsearoiawa

mfi(

K

K

wprafi1onlfttS

tm

W

wactS

waS

K

K

whgier

nu

s(i

62 A.D. Severini et al. / Field Cro

ilking and continuing until kernels reached about 200 g kg−1 mois-ure concentration, following procedures described by Borrás et al.2003). Silking dates were recorded for 10 (Exp. I low density) or0 tagged plants (Exps. I high density, II and III) in each plot, whichere then used for subsequent sampling. One plant per plot was

ampled each 3–5 days between 0800 and 1000 h. At sampling thentire ear with surrounding husks was immediately enclosed in anir-tight plastic bag and transported to the lab. Fifteen kernels wereemoved from the ear at floret positions 10–15 from the bottomf the rachis within a humidified box. Fresh weight was measuredmmediately after sampling, and kernel dry weight was determinedfter drying samples at 65 ◦C for at least 96 h. Fresh weight and dryeight were used to calculate kernel water content (mg kernel−1)

nd kernel moisture concentration (MC, fresh weight basis).Total duration of grain filling and kernel growth rate was deter-

ined for each genotype × treatment × replicate combination bytting a bi-linear model (Eqs. (1) and (2)) following Borrás et al.2003):

W = a + bTT for 250 < TT ≤ c (1)

W = a + bc for TT > c (2)

here KW is kernel dry weight, TT is the number of heat units afterollination (◦Cd), a is the Y-intercept (mg), b is the kernel growthate during the linear phase of the grain-filling period (mg ◦Cd−1),nd c is the total duration of grain filling (◦Cd). Daily values for grainlling were calculated using 0 ◦C as base temperature (Muchow,990). Mean daily air temperature was calculated as the averagef hourly air temperatures registered at a weather station locatedear the experimental plots. Kernel dry matter samples during the

ag phase (less than 250 ◦Cd) were not taken into considerationor fitting the bi-linear model. The bi-linear model was fitted tohe genotype × treatment × replicate kernel dry weight data usinghe iterative optimization technique of Table Curve V 3.0 (Jandelcientific, 1991). The R2 values ranged from 0.55 to 0.98.

Maximum water content was determined for each geno-ype × treatment × replicate combination by fitting a curvilinear

odel (Eq. (3)) following Borrás et al. (2009):

C = d + eTT + f TT1.5 + gTT2 (3)

here WC is kernel water content, TT is the number of heat unitsfter pollination (◦Cd), and d, e, f and g are model parameters. Theurvilinear model was fitted to the kernel water content data usinghe iterative optimization technique of Table Curve V 3.0 (Jandelcientific, 1991). The R2 values ranged from 0.65 to 0.99.

Kernel moisture concentration values at physiological maturityere determined using a bi-linear model relating kernel dry weight

nd kernel moisture concentration data (Eqs. (4) and (5)) (Jandelcientific, 1991) following Borrás et al. (2009):

W = h − iMC for MC ≥ j (4)

W = h − ij for MC < j (5)

here KW is kernel weight, MC is moisture concentration (g kg−1),is the Y-intercept (mg), i is the rate of kernel MC decline during

rain filling (mg[g kg−1]−1) and j is the critical MC at physiolog-cal maturity (g kg−1). This model was fitted independently forach genotype × treatment × replicate combination. The R2 valuesanged from 0.68 to 0.98.

Potential KW for each genotype × treatment × replicate combi-ation was calculated according to Borrás and Westgate (2006),

sing WC values when MC ranged between 650 and 800 g kg−1.

Plant biomass was estimated at the pre- and post-floweringtages using non-destructive allometric models developed earlierBorrás and Otegui, 2001; Echarte et al., 2004). This approach madet possible to quantify plant biomass at the individual plant level,

search 120 (2011) 360–369

and ensured the closest representation of plant biomass corre-sponding to tagged plants that remained in the field until finalharvest. The pre-flowering measurement occurred when plants had5 remaining leaves to expand (ca. 15 days before anthesis), and atthe same time the thinning treatment was performed in Exp. III. Thepost-flowering biomass sample was taken 14 DAA. The allometricmodels were constructed from 18 additional tagged plants for eachgenotype × treatment combination (6 plants per replicate in Exps.I and III) and 9 plants per genotype (3 plants per replicate in Exp.II). These plants were taken from the borders of the plot and werespecifically selected, choosing a small, an average and a large plantfrom each plot. This was done to ensure that each model covered abroad range of plant sizes. The pre-flowering model was based onthe linear regression between shoot biomass per plant and the stemvolume of each plant. The stem volume was calculated using plantheight (from the soil surface level up to the uppermost collar) andstem diameter at the base of the stalk. The R2 values of these mod-els ranged between 0.51 and 0.97 across genotypes and treatments.The post-flowering model involved stem volume and maximumdiameter of every ear showing silks, and was fitted using a multi-ple linear regression analysis. The R2 values ranged between 0.67and 0.99. These pre- and post-flowering models were used to esti-mate the plant shoot biomass of 15 (Exp. I), 10 (Exp. II) or 20 plantsper replicate (Exp. III) that remained in the field until physiologicalmaturity. All tagged plants that remained in the field were consec-utive plants within the row. At physiological maturity (defined as75% milk line; Hunter et al., 1991), these plants were harvested andused to measure individual kernel number per plant and averagekernel weight of 15 kernels from floret positions 10–15 from thebase of the rachis from the apical ear. The above-ground weight ofeach plant was measured after drying in a forced-air oven at 65 ◦Cfor at least 7 days.

Plant growth rate around flowering (mg plant−1 ◦Cd−1) wascalculated as the quotient between plant biomass increase fromthe pre- to the post-flowering biomass sample (both estimatedthrough the non-destructive allometric models), and the thermaltime interval between sample dates. Daily thermal time valueswere calculated using a base temperature of 8 ◦C (Ritchie andNeSmith, 1991). Plant growth rate per kernel around flowering(mg plant−1 ◦Cd−1 kernel−1) was obtained as the ratio between PGRduring this period and the kernel number per plant counted atharvest. Plant growth rate during the linear phase of grain fill-ing (mg plant−1 ◦Cd−1) was calculated as the quotient betweenplant biomass increase from the post-flowering sample to phys-iological maturity and the thermal time interval between thesestages, using 0 ◦C as base temperature (Muchow, 1990). Plantgrowth rate per kernel during the linear phase of grain filling(mg plant−1 ◦Cd−1 kernel−1) was calculated as the ratio betweenPGR during this period and final kernel number per plant.

For analyzing kernel type, genotype, plant densities and sourcemanipulation treatment effects and its interactions in each exper-iment, we did an ANOVA by fitting a general linear modelwith nested effects (‘genotype’ nested inside ‘kernel type’) (RDevelopment Core Team, 2010).

3. Results

3.1. Plant growth around flowering, ear growth and kernelnumber per plant

There was a fairly large and significant (p < 0.001) variation inPGR around flowering among the genotypes examined in this study(Table 1). Dent genotypes grew faster than the popcorn genotypesin the three experiments (p < 0.001). As expected, reducing plantdensity (Exp. I) and thinning (Exp. III) before flowering increased

A.D. Severini et al. / Field Crops Research 120 (2011) 360–369 363

Table 1Kernel number per plant (KNP), plant growth rate (PGR) around flowering, plant growth rate per kernel around flowering and during the effective grain-filling period,individual kernel weight (KW), ears biomass per plant at 14 days after anthesis (EB14DAA), kernel number per g ear biomass and yield per plant. KT, kernel types; G,genotypes; D, plant density; T, source treatments. Def − 15, defoliation of 75% leaf area at −15 days from anthesis; Def + 17, defoliation of 75% leaf area at +17 days fromanthesis; Thin − 15, thinning of 50% plant population at −15 days from anthesis; Thin + 17, thinning of 50% plant population at +17 days from anthesis.

Exp. KT G D (pl m−2) T KNP(kernel pl−1)

PGR flowe-ring (mg ◦-Cd−1 pl−1)

PGR per ker-nel flowering(mg ◦Cd−1 pl−1

kernel−1)

KW(mg ker-nel−1)

PGR per ker-nel grain filling(mg ◦Cd−1 pl−1

kernel−1)

EB14DAA(g ear−1)

KNP(EB14DAA)−1

(kernel g−1)

Yield perplant(g pl−1)

I Dent B73 3 Control 741 394 0.540 233 0.365 37.8 20.1 149.0Def + 17 616 399 0.671 177 0.013 36.2 17.3 89.3

9 Control 408 191 0.488 228 0.388 12.5 32.3 78.4Def + 17 420 195 0.470 151 0.066 13.4 31.9 52.3

Mo17 3 Control 435 312 0.741 328 0.455 16.9 25.7 117.1Def + 17 396 325 0.843 244 0.052 17.3 24.1 74.3

9 Control 328 183 0.583 271 0.347 6.7 52.9 76.2Def + 17 296 177 0.612 181 0.004 6.0 56.0 41.6

N209 3 Control 796 354 0.464 229 0.369 38.6 21.9 154.0Def + 17 698 332 0.505 143 0.070 37.2 18.2 78.7

9 Control 419 161 0.409 193 0.294 16.1 27.6 69.7Def + 17 347 158 0.570 123 −0.157 14.4 25.2 36.6

Popcorn IDS69 3 Control 367 174 0.487 115 0.351 15.0 25.5 35.9Def + 17 357 174 0.498 101 −0.004 14.1 26.9 30.6

9 Control 311 144 0.485 116 0.238 9.6 35.3 30.3Def + 17 287 150 0.583 97 −0.228 8.9 34.9 23.2

IDS91 3 Control 394 230 0.603 125 0.313 15.0 27.8 45.3Def + 17 374 223 0.609 114 −0.112 15.0 26.7 36.5

9 Control 327 163 0.505 135 0.203 7.8 47.0 36.8Def + 17 235 132 0.672 107 −0.272 4.8 54.4 20.8

R18 3 Control 845 194 0.275 104 0.198 14.8 53.2 70.4Def + 17 772 183 0.258 93 0.047 16.0 49.3 53.5

9 Control 567 142 0.319 98 0.094 8.2 62.7 43.4Def + 17 445 110 0.298 85 −0.044 5.7 85.9 28.7

G *** *** (0.111)*** *** ns *** *** ***

KT ** *** *** *** *** *** *** ***

D *** *** ns *** *** *** *** ***

T ** ns ** *** *** ns ns ***

G × D (154)a,*** (43)*** ns (20)*** ns (6.1)*** (11.9)*** (18.8)***

KT × D (68)* (19)*** (0.081)** (9)*** ns (2.7)*** (5.3)*** ***

G × T ns ns ns ns (0.177)*** ns ns nsKT × T ns ns ns (9)*** ns ns (5.3)* ***

D × T ns ns ns ns ns ns (5.3)* **

G × D × T ns ns ns ns ns ns ns nsKT × D × T ns ns ns ns ns ns ns (14.1)***

II Dent B73 9 450 242 0.583 234 0.546 24.4 19.3 97.1N209 371 208 0.701 247 0.737 20.3 18.9 77.8

Popcorn 95:2 295 198 0.705 127 0.640 13.1 22.9 33.1IDS69 291 146 0.545 122 0.290 7.8 40.8 29.8IDS91 327 183 0.575 133 0.308 13.6 25.5 38.5R18 506 139 0.298 96 0.220 10.7 49.4 42.1R-28-2 391 129 0.544 112 0.479 16.6 31.7 42.5R-53-1 453 130 0.300 115 0.643 16.2 29.2 44.7G (180)** (47)*** ns ns ns (6.3)*** (12.7)*** nsKT ns *** ns *** ns *** *** ***

III Dent AW190 9 Control 560 269 0.483 269 0.202 21.2 29.3 132.4Def − 15 391 190 0.491 226 0.102 12.3 41.5 75.8Thin − 15 917 409 0.453 293 0.256 43.8 21.9 231.2Def + 17 486 247 0.504 198 0.155 19.1 34.0 84.3Thin + 17 600 267 0.447 307 0.346 22.4 28.5 160.2

Popcorn P625 Control 465 174 0.378 144 0.173 19.4 24.9 60.8Def − 15 366 138 0.399 129 0.062 12.8 30.4 43.3Thin − 15 503 272 0.589 157 0.165 36.1 14.2 71.3Def + 17 415 166 0.411 128 0.127 18.3 24.5 47.7Thin + 17 472 165 0.362 142 0.203 19.7 28.5 61.2

P802 Control 449 183 0.412 162 0.152 18.0 27.7 66.8Def − 15 392 140 0.376 143 0.090 12.1 39.5 51.1Thin − 15 555 253 0.466 177 0.195 26.0 23.2 88.9Def + 17 409 166 0.408 123 0.109 17.2 27.6 46.1Thin + 17 463 183 0.395 157 0.139 18.5 26.7 66.9

G ns ns ns (9)** ns ** (5.5)* ***

KT *** *** *** *** *** *** * ***

T *** *** *** *** ** *** (7.0)*** ***

G × T ns ns (0.124)** ns ns (8.8)* ns (14.4)*

KT × T (79)*** (35)*** (0.098)*** (26)*** (0.126)* (7.0)*** ns (11.4)***

ns = not significant.a LSD value for p ≤ 0.05.* Significant at p ≤ 0.05.

** Significant at p ≤ 0.01.*** Significant at p ≤ 0.001.

3 ps Re

Pflnagd

((stotlTw

ttTEtdlIpttsh

PdTiirikK1rwtwadRsksa

3

mgtiteiaAa

64 A.D. Severini et al. / Field Cro

GR around this period. Likewise, the defoliation treatment beforeowering (Exp. III) significantly decreased PGR. A significant ker-el type × source treatment interaction (p < 0.001) was observed inll experiments, however, as dent genotypes showed greater plantrowth responses to defoliation, thinning or plant densities thanid popcorn genotypes (Table 1).

Genotypes varied in the accumulated ear weight at 14 DAATable 1) and dent genotypes showed larger values than popcornsp < 0.001). At 9 pl m−2, for example, dent inbred lines B73 and N209howed the greatest ear weight values in Exps. I and II. In Exp. IIIhe largest ear biomass was observed in dent hybrid AW190. Asbserved with PGR, the reduced plant density (Exp. I) and thinningreatments (Exp. III) increased ear weight at 14 DAA. Partially defo-iating the canopy (Exp. III) before flowering decreased ear weight.hese differences in ear weight among genotypes and treatmentsere highly significant (p < 0.001) in most cases.

Kernel number per plant differed significantly among kernelypes in Exps. I and III, where the general trend was dent geno-ypes setting more kernels than popcorns (p < 0.01 and p < 0.001;able 1). There was also variation in KNP between genotypes inxps. I and II (p < 0.01 and p < 0.001), where inbred line R18 showedhe largest KNP across all trials when compared at a similar plantensity. The general trend was an increase in KNP in response to

ower plant density (Exp. I) and thinning prior to flowering (Exp.II), and fewer kernels per plant in response to canopy defoliationrior to flowering (Exp. III). A kernel type × source manipulationreatment significant interaction in Exp. III (p < 0.001) and a kernelype per × plant density significant interaction in Exp. I (p < 0.05)howed dents to increase KNP more than popcorns in response toigher plant growth rates around flowering.

There was a positive association between KNP at maturity andGR around flowering (Table 1). Genotypic differences were evi-ent in the number of kernels set per unit of PGR around flowering.he greatest efficiency in kernel set per unit of PGR was observedn the inbred line R18 (Table 1). In an attempt to explain differencesn kernel number per plant between genotypes, we examined theelationship between ear biomass 14 DAA and PGR around flower-ng as an indicator of reproductive partitioning needed to supporternel set. In general, genotypes and treatments with the highestNP and more vigorous PGR also had the highest ear biomass at4 DAA (Table 1). Growth conditions that altered PGR, however,evealed the efficiency of kernel set per unit ear biomass 14 DAAas not a constant for each genotype, but varied by plant densi-

ies and source manipulation treatments (Table 1). The efficiencyas greater at slower PGRs both in Exps. I and III. Also, there wassignificant kernel type effect in Exps. I, II and III (p < 0.05) whereents showed lower efficiencies than popcorns. Popcorn genotype18 was most efficient in setting kernels per unit ear biomass, con-istent in both Exps. I and II (Table 1). In this genotype, the greaterernel set per unit PGR was associated with greater efficiency inetting kernels per g ear biomass rather than greater ear biomassccumulated at flowering.

.2. Kernel weight and kernel growth patterns

Kernel types significantly differed in their individual KW ataturity (p < 0.001; Table 1), where dent genotypes always showed

reater KW than popcorns. Kernel type × plant density, geno-ype × plant density and kernel type × source treatment significantnteractions were detected for KW in Exps. I and III. When dissectinghese interactions it was observed that, when PGR around flow-

ring was increased, KW only augmented in dent genotypes. Fornstance, reducing plant density in Exp. I increased KW in Mo17nd N209, and thinning at pre-flowering in Exp. III increased KW inW190. Decreasing PGR around flowering by pre-flowering defoli-tion reduced KW in these same genotypes, and only in the popcorn

search 120 (2011) 360–369

genotype P802 in Exp. III. In general, KW in popcorn genotypes wasless affected by changes in PGR around flowering than it was in thedent genotypes.

Differences in KW between popcorn and dent genotypes werea consequence of differences in kernel growth rate and durationof grain filling (Table 2). Popcorn kernels were smaller than dentkernels because of a slower growth rate and a shorter grain fill dura-tion. Treatments imposed to modify PGR around flowering (plantdensity in Exp. I, and defoliation and thinning at pre-floweringin Exp. III) had no significant effect on the rate of kernel growth(Table 2).

There were significant differences (p < 0.001) in MWC amongkernel types (Table 2), where dent kernels showed greater val-ues than popcorn kernels. Also, we observed differences in MWCbetween plant densities and treatments in both experiments.The low plant density (Exp. I) and the thinning at pre-flowering(Exp. III) increased MWC, and the pre-flowering defoliation treat-ment in Exp. III reduced MWC. Significant kernel type × plantdensity and kernel type × source modification treatment interac-tions were detected (Table 2). Although the thinning treatment atpre-flowering in Exp. III had a higher effect on MWC in popcornP625, dent kernels MWC always responded more than popcorns inthe rest of the treatments. Whenever PGR around flowering wasdecreased (by a higher plant density in Exp. I or the defoliationtreatment in Exp. III) MWC was more markedly reduced in dentsthan in the popcorn genotypes (Table 2).

Genotypic differences in potential KW, calculated from kernelwater content at ca. 70–80% of MC, were evident in Exps. I andII (p < 0.001), but not in Exp. III, for potential KW differences inthat experiment were mainly a cause of kernel type differences(p < 0.001; Table 2). In addition, potential KW was greater whenplants were grown at the low density of Exp. I (Table 2). Plantssubjected to thinning at 15 days before anthesis in Exp. III hadgreater potential KW than those defoliated at that time. A kerneltype × plant density interaction was observed in Exp. I since thehigher plant density reduced potential KW more markedly in thedent inbred Mo17.

Kernel moisture concentration (MC, on a fresh weight basis)at physiological maturity showed significant differences betweenkernel types in the three experiments (p < 0.01; Table 2). Across thethree experiments, popcorn genotypes always showed lower MCvalues at maturity than did dent genotypes. MC was not affectedby treatments aimed to modify plant growth around the flower-ing period (plant density in Exp. I, and pre-flowering thinning anddefoliation treatments in Exp. III).

3.3. Kernel weight and plant growth rate per kernel aroundflowering

We analyzed the relationship between KW and PGR per kernelaround flowering (Fig. 1). We expected PGR per kernel around flow-ering to explain differences in KW at maturity for all genotypes,including dents and popcorns. However, KW was only partiallyrelated to PGR per kernel around flowering (Fig. 1A). Althoughsignificant (p < 0.05), a single relationship did not describe all geno-types and growth conditions because dent genotypes achievedgreater KWs for the same PGR per kernel around flowering com-pared to most popcorn genotypes. As such, PGR around floweringdid not account for the observed variation in KW across all geno-types and treatments.

Kernel growth rate, MWC and potential KW were posi-

tively associated with PGR per kernel around flowering (p < 0.05;Tables 1 and 2). For any given PGR per kernel, however, most pop-corn genotypes exhibited a smaller kernel growth rate, MWC orpotential KW than did dent genotypes. This result implies that dentgenotypes, which achieved a greater final KW for the same plant

A.D. Severini et al. / Field Crops Research 120 (2011) 360–369 365

Table 2Kernel growth rate, duration of the grain-filling period, maximum water content, potential kernel weight and kernel moisture concentration at physiological maturity. KT,kernel types; G, genotypes; D, plant density; T, source treatments; Def − 15, defoliation of 75% leaf area at −15 days from anthesis; Def + 17, defoliation of 75% leaf area at+17 days from anthesis; Thin − 15, thinning of 50% plant population at −15 days from anthesis; Thin + 17, thinning of 50% plant population at +17 days from anthesis.

Exp. KT G D (pl m−2) T Kernel growth rate(mg ◦Cd−1 kernel−1)

Grain-fillingduration(◦Cd)

Maximumwater content(mg kernel−1)

Potentialkernel weight(mg kernel−1)

Moistureconcentration atphysiologicalmaturity (g kg−1)

I Dent B73 3 Control 0.268 1123 165 237 300Def + 17 0.321 844 155 245 424

9 Control 0.287 1060 163 251 325Def + 17 0.261 855 142 237 444

Mo17 3 Control 0.435 1017 267 370 344Def + 17 0.361 925 254 353 457

9 Control 0.390 993 204 291 360Def + 17 0.388 760 188 277 473

N209 3 Control 0.320 992 238 308 463Def + 17 0.284 803 210 306 573

9 Control 0.256 976 196 284 440Def + 17 0.239 765 192 282 599

Popcorn IDS69 3 Control 0.214 783 68 155 286Def + 17 0.167 800 66 157 308

9 Control 0.176 843 65 151 294Def + 17 0.158 816 59 150 305

IDS91 3 Control 0.187 879 76 161 295Def + 17 0.201 816 73 164 302

9 Control 0.182 993 75 164 266Def + 17 0.202 810 70 165 317

R18 3 Control 0.140 927 59 152 273Def + 17 0.125 928 57 152 238

9 Control 0.162 846 58 151 292Def + 17 0.132 828 51 148 322

G (0.061)a,*** (104)* *** *** (48)***

KT *** *** *** *** ***

D ns ns *** *** nsT ns *** *** ns ***

G × D ns ns *** (29)*** nsKT × D ns ns (6)*** (13)*** nsG × T ns ns ns ns nsKT × T ns (76)*** (6)** ns (35)***

D × T ns ns ns ns nsG × D × T ns ns (23)* ns nsKT × D × T ns ns ns ns ns

II Dent B73 9 0.298 1071 168 227 412N209 0.287 1084 219 312 395

Popcorn 95:2 0.169 938 72 163 304IDS69 0.195 880 68 153 317IDS91 0.196 911 78 168 302R18 0.170 802 64 156 361R-28-2 0.179 854 77 162 356R-53-1 0.143 938 72 153 265G ns ns (32)** (36)*** nsKT *** *** *** *** **

III Dent AW190 9 Control 0.347 1062 193 278 335Def − 15 0.325 970 170 261 362Thin − 15 0.323 1146 210 324 308Def + 17 0.319 887 181 284 404Thin + 17 0.350 1147 207 283 332

Popcorn P625 Control 0.205 921 85 177 309Def − 15 0.219 844 80 166 328Thin − 15 0.223 911 101 192 314Def + 17 0.220 810 84 175 335Thin + 17 0.211 887 87 175 317

P802 Control 0.236 944 99 180 299Def − 15 0.208 947 84 163 299Thin − 15 0.228 1007 107 200 290Def + 17 0.231 808 89 178 354Thin + 17 0.217 973 93 172 295

G ns (50)** (5)** ns nsKT *** *** *** *** ***

T ns *** *** (33)* (49)*

G × T ns ns ns ns nsKT × T ns (142)* (13)*** ns ns

ns = not significant.a LSD value for p ≤ 0.05.* Significant at p ≤ 0.05.

** Significant at p ≤ 0.01.*** Significant at p ≤ 0.001.

366 A.D. Severini et al. / Field Crops Research 120 (2011) 360–369

200

300

400

ht (m

g k

ern

el-1

)

R2 = 0.21

p < 0.05

A B AW190

B73

Mo17

N209

95:2

IDS69

IDS91

P625

0.0 0.2 0.4 0.6 0.8

0

100

Kern

el w

eig

h

0.0 0.2 0.4 0.6 0.8

P625

P802

R-28-2

R-53-1

R18

0.0 0.2 0.4 0.6 0.8

Plant growth rate per kernel around

flowering (mg plant-1 °Cd-1 kernel-1)

Plant growth rate per kernel during

grain filling (mg plant-1 °Cd-1 kernel-1)

0.0 0.2 0.4 0.6 0.8

F or dut . The rs pcorn

gla

3g

fift

rwngDDtnrTKl

bgdaItigmurip

3

s

dents.There was a linear relationship between measured KW at matu-

rity and potential KW estimated from kernel water content earlyin grain filling for all genotypes exposed to treatments that altered

0

20Popcorns

y = (0.09 ± 0.02) x

R2 = 0.36

p < 0.01

-20 Dents

y = (0.28 ± 0.04) x

R2 = 0.79

p < 0.001

3002001000-100-200-300

Re

lative

ch

an

ge

in

ke

rne

l w

eig

ht (%

)

-40

Relative change in plant growth rate

per kernel during grain filling (%)

Fig. 2. Relationship between the relative change in final kernel weight and the rel-

ig. 1. Kernel weight response to plant growth rate per kernel around flowering (A)reatments starting at the end of the flowering period (17 DAA from Exps. I and III)ymbols represent dent genotypes and open symbols (either clear or dotted) are po

rowth rate per kernel around flowering (Fig. 1A), must have estab-ished a greater potential kernel weight for a similar PGR per kernelround flowering.

.4. KW and plant growth during the linear phase of therain-filling period

Plant growth rate per kernel during the linear phase of the grain-lling period showed no obvious relationship with KW at maturity

or popcorn or dent genotypes (Fig. 1B) when the treatments aimedo modify PGR starting at pre-flowering are considered.

We also evaluated genotypic differences in KW and yieldesponse to a severe defoliation imposed after the lag phasehen potential sink capacity (kernel number and potential ker-el weight) was already established. This treatment slowed plantrowth rate per kernel dramatically during grain-filling (Table 1).efoliations at this stage reduced KW at maturity for all genotypes.ifferences between popcorn and dent genotypes were evident in

he percent decrease in KW relative to the reduction in PGR per ker-el (Fig. 2). The relative reduction in KW at any given PGR per kerneleduction was less for popcorn genotypes than for dent genotypes.his result suggests that popcorns would be less prone to reduceW than dents when a severe source limitation occurs during the

inear phase of the grain-filling period.Moisture concentration at physiological maturity was affected

y plant growth manipulation treatments aimed to reduce the plantrowth during grain filling (Table 2, Exps. I and III). Plants in theefoliation treatments starting at 17 DAA showed higher MC valuest maturity. However, this effect was not evident for all genotypes.n Exp. I, there was a significant kernel type × source modificationreatment interaction (p < 0.001), since defoliating the canopy dur-ng grain filling increased MC at physiological maturity more in dentenotypes than in popcorn genotypes (Table 2). MC at physiologicalaturity remained more stable across plant densities and manip-

lative source treatments in popcorn kernels, and MC at maturityemained unchanged when PGR per kernel during grain filling wasncreased by thinning at the end of the lag phase for dent andopcorn genotypes.

.5. Yield per plant

Yield per plant varied among kernel types, plant densities andource treatments in all experiments (always p < 0.001), where

ring the effective grain-filling period (B). The figure does not include the defoliationegression line in (A) represents the equation KW = 250 × PGR kernel−1 + 52. Closedgenotypes.

dent genotypes yielded more than did popcorn genotypes. Therewere also significant genotypic effects within each kernel typein Exp. I (p < 0.05; Table 1). In most cases, the reduced yieldof popcorns was explained both by a decrease in kernel num-ber per plant and in kernel weight (Table 1). Yield per plantalso changed in response to treatments imposed to alter PGRaround flowering or during the linear phase of grain filling. Therewere significant kernel type × treatment interactions (p < 0.05),because yield in popcorn genotypes remained more stable than

ative change in plant growth rate per kernel during the linear phase of grain filling,both calculated as the relative difference between the plant growth source modi-fication treatment and the control treatment. Source treatments were thinning ordefoliation, both imposed 17 days after 50% anthesis. Fitted models were forcedto have zero-intercept, and values within parenthesis are slope ± standard error.Symbols are the same as Fig. 1.

A.D. Severini et al. / Field Crops Research 120 (2011) 360–369 367

100 200 300 400

100

200

300

400

0 100 200 300

0

100

200

300

100 200 300 400

0 100 200 300

Kern

el w

eig

ht (m

g k

ern

el-1

)

Potential kernel weight (mg kernel-1)

A B

Controls Defoliated

Yie

ld (

g p

lant-1

)

Potential sink capacity (g plant-1)

C D

Controls Defoliated

R2 = 0.91

p < 0.001

R2 = 0.74

p < 0.001

R2 = 0.76

p < 0.001

R2 = 0.93

p < 0.001

Fig. 3. Relationship between final and potential kernel weight for the controls (A) and for the source treatments that had a severe defoliation imposed 17 days after 50%anthesis (DAA) (B), and between final yield and potential sink capacity established at 14 DAA for the controls (C) and source treatments that had a severe defoliation imposed17 DAA (D). Fitted models are (A) kernel weight = 0.96 × potential kernel weight − 25.70, (B) kernel weight = 0.56 × potential kernel weight + 13.80, (C) yield = 0.75 × potentials re the

ptco1petgsrKnrRtpie

airtp

ink capacity − 3.05, and (D) yield = 0.35 × potential sink capacity + 10.70. Symbols a

lant growth rate during flowering (Fig. 3A). This result implieshat variability in final KW among the popcorn and dent genotypesan be related to their potential KWs. The same relationship wasbserved between potential sink capacity per plant established at4 DAA (kernel number per plant × potential kernel weight) andlant grain yield at maturity. This result suggests that genotypic andnvironmental differences in yield were a consequence of poten-ial yield established at the end of the lag phase (Fig. 3C). Dentenotypes produced the greatest yields due to a greater potentialink capacity established prior to grain filling. This potential waselated not only to a greater KNP but also to a greater potentialW (Table 1), showing the importance of potential KW determi-ation on yield formation. An extreme example of the impact of aeduced potential KW on yield was observed for popcorn inbred18. Although this genotype showed the highest efficiency for set-ing kernels per unit of ear biomass (Table 1), the significantly lowerotential kernel weight determined a smaller potential sink capac-

ty (and yield) for this genotype than for any of the other dentntries.

Defoliation at 17 DAA reduced the capacity of the kernels to

chieve their potential KW (Fig. 3B) and the total plant sink capac-ty (Fig. 3D) established at the end of the lag phase. This yieldesponse to defoliation was greater in dent than in popcorn geno-ypes (p < 0.001), in agreement with the lower susceptibility ofopcorns to decrease KW (Fig. 2).

same as Fig. 1.

4. Discussion

Dent and popcorn inbreds and hybrids used in this study showedan ample range of KNP and KW that exposed different devel-opmental patterns determining the final values for these yieldcomponents. Separating the relationship between PGR aroundflowering and KNP into ear biomass partitioning and kernel setper unit of ear biomass was particularly useful for distinguishinghow each genotype established final KNP. However, there seemsto be a need to determine how each specific genotype deter-mines yield independently of the kernel type. Although popcornsshowed a higher kernel set per unit of ear biomass accumulationaround flowering, popcorn inbred line R18, for example, showed aunique higher efficiency for setting kernels per unit of PGR aroundflowering by setting a large number of kernels per unit of earbiomass (Table 1). The efficiency for setting kernels per unit of earbiomass, however, was not constant for any given genotype, butwas affected by treatments intended to modify the PGR aroundflowering (Table 1). Previous studies have shown genotypic andenvironmental effects on biomass partitioning to the growing ear

around flowering (Echarte et al., 2004; Borrás et al., 2007). Ourresults emphasize the need to address changes in the efficiencyfor setting kernels per unit of ear biomass, which is supported byrecent observations published by Echarte and Tollenaar (2006) andD’Andrea et al. (2009).

3 ps Re

fltc(rctwePKdpaaK

ed(cd1tfikagoipisnuk

taapsdasgts(

ooyegamwiogBcKb

to contrasting nitrogen supply levels. Field Crops Res. 114, 147–158.Echarte, L., Tollenaar, M., 2006. Kernel set in maize hybrids and their inbred lines

68 A.D. Severini et al. / Field Cro

Gambín et al. (2006, 2008) proposed that PGR per kernel aroundowering could be used to predict KW differences among geno-ypes, as KW usually depends upon the potential kernel sinkapacity established during the early stages of kernel developmentBorrás and Westgate, 2006). We tested this concept for a widerange of kernel sizes and two endosperm types, dent and pop-orn. A single model for all genotypes could not be establishedo relate KW at maturity or potential KW determined 15 DAAith PGR per kernel around flowering. Most popcorn genotypes

stablished a smaller KW than did dent genotypes for the sameGR per kernel around flowering. This result is in agreement withiniry et al. (1990) who reported that KW of popcorn genotypesid not increase in response to decreased KNP through restrictedollination. Evidently there are clear differences between dentnd popcorn endosperm-types in their capacity to utilize availablessimilates per kernel around flowering for establishing potentialW.

Popcorns are selected not only for yield but for their endospermxpansion capacity when heated. Kernel expansion capacityepends on KW, and there is an optimum KW for kernel expansionAllred-Coyle et al., 2000). Popcorn pericarps are thicker, have moreell layer numbers and a higher content of extensin proteins than doent and sweetcorn genotypes (Tracy and Galinat, 1987; Hood et al.,991). Our study shows popcorn genotypes have a smaller poten-ial KW per unit of plant growth per kernel during the early grainlling stages than dents. This smaller potential is reflected in slowerernel growth rates, lower MWC, and lower potential KW and KWt physiological maturity when compared to dents at similar plantrowth rates per kernel around flowering. Selection to maintain anptimum KW to maximize yield and endosperm expansion capac-ty has apparently altered the flexibility of kernels to modify theirotential weight in regard to changes in assimilate availability dur-

ng early kernel development. A logical future research would be totudy pericarp development in popcorns. This structure may domi-ate the relationships between kernel water uptake and assimilatetilization because it might be controlling the rate and extent ofernel expansion.

When plant growth per kernel was severely reduced duringhe linear phase of the grain-filling period, popcorn KW was lessffected than dent KW. Hence, popcorns would be expected to showmore stable KW compared to dents for a similar reduction in

lant growth per kernel during grain filling. Echarte et al. (2006)howed genotypic differences in KW susceptibility under severeefoliations during the linear phase of the grain-filling period,lthough they did not show if this differential response was a con-equence of a smaller source–sink ratio in the more susceptibleenotypes. We observed that the source–sink ratio during the effec-ive seed filling (measured as plant growth per kernel during thistage) was similar for genotypes with different KW susceptibilityFig. 2).

Although severely reducing crop growth during the linear phasef the grain-filling period can affect KW and limit the achievementf a higher yield, we observed a close relationship between finalield and potential sink capacity established at the end of the flow-ring period for dent and popcorn genotypes. These results are ineneral agreement with the concept that focusing on the periodround flowering in this species is important for yield improve-ent through both yield components, kernel number and kerneleight. At present it seems that maize potential yield can only be

mproved by modifying the sink capacity established at the endf the flowering period and increasing the source strength duringrain filling so as to fulfill this potential (Lee and Tollenaar, 2007;orrás and Gambín, 2010). However different specific responses

ould be observed between dent and popcorn germplasm inNP and KW determination, these concepts seem valid foroth.

search 120 (2011) 360–369

5. Conclusions

Key issues emerging from this study are:

• KNP in dent and popcorn genotypes was very dependent uponPGR and biomass partitioning to the ear around the floweringperiod. However, consistent differences among genotypes in theefficient use of biomass allocated to the ear for kernel numberdetermination existed. Therefore, using a single efficiency valuefor all kernel types, genotypes and growth conditions would notseem appropriate.

• Unlike the situation when only dent hybrids were evaluated(Gambín et al., 2006, 2008), it was not possible to predict poten-tial KW and kernel growth rate from plant growth rates perkernel around flowering across the more diverse germplasmbase including popcorn genotypes. Pericarp restrictions to earlyexpansion in popcorns may control this response.

• Popcorn kernels were less prone than dent kernels to decreaseKW under stressful conditions causing severe reductions in plantgrowth during grain-filling. Compared to dents, popcorn kernelsevidently develop with a surplus of assimilates during the lagphase and also the linear phase of the grain-filling period. Thismay alleviate source limitations occurring during grain filling.

• Despite different patterns of KNP and KW determination, yieldvariation across genotypes and source manipulation treatmentscorresponded closely to the potential sink capacity established bythe end of the lag phase. This result emphasizes the importanceof favorable growth during the flowering period to establish KNPand KW in maize.

Acknowledgements

Authors wish to thank M.E. Otegui, and J.L. Rotundo and J. Kimfor field assistance at Iowa, L.B. Blanco, E.M. Chintio, W.R. Miranda,M. Rossini, L. Ferrand, J. Lobo and S. Patrouilleau at Pergamino. A.D.Severini held a grant from INTA, the Argentinean National Instituteof Agronomical Technology. L. Borrás is a member of CONICET, theNational Research Council of Argentina.

References

Agricultural Alumni Seed Improvement Association, Inc. Romney, Indiana, USA.Available from: http://www.agalumniseed.com.

Allred-Coyle, T.A., Toma, R.B., Reiboldt, W., Thakur, M., 2000. Effects of mois-ture content, hybrid variety, kernel size, and microwave wattage on theexpansion volume of microwave popcorn. Int. J. Food Sci. Nutr. 51, 389–394.

Andrade, F.H., Vega, C.R.C., Uhart, S., Cirilo, A., Cantarero, M., Valentinuz, O., 1999.Kernel number determination in maize. Crop Sci. 39, 453–459.

Borrás, L., Otegui, M.E., 2001. Maize kernel weight response to postflowering source– sink ratio. Crop Science 41, 1816–1822.

Borrás, L., Gambín, B.L., 2010. Trait dissection of maize kernel weight: towards inte-grating hierarchical scales using a plant biomass framework. Field Crops Res.118, 1–12.

Borrás, L., Westgate, M.E., 2006. Predicting maize kernel sink capacity early in devel-opment. Field Crops Res. 95, 223–233.

Borrás, L., Westgate, M.E., Otegui, M.E., 2003. Control of kernel weight and ker-nel water relations by post-flowering source–sink ratio in maize. Ann. Bot. 91,857–867.

Borrás, L., Westgate, M.E., Astini, J.P., Echarte, L., 2007. Coupling time to silking withplant growth rate in maize. Field Crops Res. 102, 73–85.

Borrás, L., Zinselmeier, C., Senior, M.L., Westgate, M.E., Muszynski, M.G., 2009. Char-acterization of grain-filling patterns in diverse maize germplasm. Crop Sci. 49,999–1009.

D’Andrea, K.E., Otegui, M.E., Cirilo, A.G., Eyhérabide, G.H., 2009. Ecophysiologicaltraits in maize hybrids and their parental inbred lines: phenotyping of responses

exposed to stress. Crop Sci. 46, 870–878.Echarte, L., Andrade, F.H., Vega, C.R.C., Tollenaar, M., 2004. Kernel number determi-

nation in Argentinean maize hybrids released between 1965 and 1993. Crop Sci.44, 1654–1661.

ps Re

E

E

G

G

G

G

H

H

J

J

K

A.D. Severini et al. / Field Cro

charte, L., Andrade, F.H., Sadras, V.O., Abbate, P., 2006. Kernel weight andits response to source manipulations during grain filling in Argentineanmaize hybrids released in different decades. Field Crops Res. 96, 307–312.

dmeades, G.O., Bolanos, J., Hernández, M., Bello, S., 1993. Causes of silk delay in alowland tropical maize population. Crop Sci. 33, 1029–1035.

ambín, B.L., Borrás, L., 2010. Resource distribution and the trade-off between seednumber and weight: a comparison across crop species. Ann. Appl. Biol. 156,91–102.

ambín, B.L., Borrás, L., Otegui, M.E., 2006. Source–sink relations and kernel weightdifferences in maize temperate hybrids. Field Crops Res. 95, 316–326.

ambín, B.L., Borrás, L., Otegui, M.E., 2007. Kernel water relations and duration ofgrain filling in maize temperate hybrids. Field Crops Res. 101, 1–9.

ambín, B.L., Borrás, L., Otegui, M.E., 2008. Kernel weight dependence upon plantgrowth at different grain-filling stages in maize and sorghum. Aust. J. Agric. Res.59, 280–290.

ood, E.E., Hood, K.R., Fritz, S.E., 1991. Hydroxyproline-rich glycoproteins in cell-walls of pericarp from maize. Plant Sci. 79, 13–22.

unter, J.L., Tekrony, D.M., Miles, D.F., Egli, D.B., 1991. Corn seed maturity indicatorsand their relationship to uptake of C-14 assimilate. Crop Sci. 31, 1309–1313.

andel Scientific, 1991. Table Curve V 3.0. User’s Manual Version 3.0. AISN Software.Jandel Scientific, Corte Madera, CA.

ones, R.J., Schreiber, B.M.N., Roessler, J.A., 1996. Kernel sink capacity in maize:genotypic and maternal regulation. Crop Sci. 36 (2), 301–306.

iniry, J.R., Wood, C.A., Spanel, D.A., Bockholt, A.J., 1990. Seed weight response todecreased seed number in maize. Agron. J. 82, 98–102.

search 120 (2011) 360–369 369

Lee, E.A., Tollenaar, M., 2007. Physiological basis of successful breeding strategiesfor maize grain yield. Crop Sci. 47 (S3), S202–S213.

Monsanto Argentina, 2008. Maíz: Híbridos Dekalb. Catálogo de Productos.Muchow, R.C., 1990. Effect of high-temperature on grain-growth in field-grown

maize. Field Crops Res. 23, 145–158.R Development Core Team, 2010. R: A Language and Environment for Statistical

Computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN: 3-900051-07-0. URL http://www.R-project.org.

Reddy, V.M., Daynard, T.B., 1983. Endosperm characteristics associated with rate ofgrain filling and kernel size in corn. Maydica 28, 339–355.

Ritchie, S.W., Hanway, J.J., Benson, G.O., 1993. How A Corn Plant Develops. Spe-cial Report Number 48. Iowa State University of Science and Technology, Ames,IA.

Ritchie, J.T., NeSmith, D.S., 1991. Temperature and crop development. In: Hanks, J.,Ritchie, J.T. (Eds.), Modelling Plant and Soil Systems. American Society of Agron-omy, Crop Science Society of America, Soil Science Society of America, AgronomySeries 31, Madison, WI, pp. 5–29.

Sadras, V.O., 2007. Evolutionary aspects of the trade-off between seed size andnumber in crops. Field Crops Res. 100, 125–138.

Tollenaar, M., Dwyer, L.M., Stewart, D.W., 1992. Ear and kernel formation in maize

hybrids representing three decades of grain yield improvement in Ontario. CropSci. 32, 432–438.

Tracy, W.F., Galinat, W.C., 1987. Thickness and cell layer number of the pericarp ofsweet corn and some of its relatives. Hortscience 22, 645–647.

Ziegler, K.E., 2001. Popcorn. In: Hallauer, A.R. (Ed.), Specialty Corns, 2nd ed. CRCPress, Boca Raton, FL, pp. 199–234.