Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/305317391

Agronomicandqualitycharacteristicsofold,modernandmixturewheat(Triticumaestivumspp.)varietiesandlandracefor...

ArticleinEuropeanJournalofAgronomy·July2016

CITATIONS

0

READS

289

1author:

Someoftheauthorsofthispublicationarealsoworkingontheserelatedprojects:

DIVERSIFOOD:EmbeddingcropdiversityandnetworkingforlocalhighqualityfoodsystemsView

project

Strengtheningcultivardiversityofbarleyanddurumwheattomanageclimaterelatedrisksand

fosterproductivityinmarginalareasofEthiopiaViewproject

SalvatoreCeccarelli

SelfEmployed

193PUBLICATIONS5,163CITATIONS

SEEPROFILE

AllcontentfollowingthispagewasuploadedbySalvatoreCeccarellion14July2016.

Theuserhasrequestedenhancementofthedownloadedfile.Allin-textreferencesunderlinedinblueareaddedtotheoriginaldocumentandarelinkedtopublicationsonResearchGate,lettingyouaccessandreadthemimmediately.

Awe

PSa

b

c

d

a

ARRA

KOMAWOPEBC

1

sp

h1

Europ. J. Agronomy 79 (2016) 131–141

Contents lists available at ScienceDirect

European Journal of Agronomy

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

gronomic and quality characteristics of old, modern and mixtureheat varieties and landraces for organic bread chain in diverse

nvironments of northern Italy

aola Migliorini a,∗, Sandra Spagnolo b, Luisa Torri a, Marco Arnoulet b, Giulio Lazzerini c,alvatore Ceccarelli d

University of Gastronomic Sciences, Piazza Vittorio Emanuele 9, 12060 Bra, CN, ItalyAIAB in Piemonte, Via Stura, 24-100098 Rivoli, TO, ItalyUniversity of Florence, Department of Agrifood Production and Environmental Sciences (DISPAA), Piazzale delle Cascine 18, 50144 Firenze, ItalyConsultant, Rete Semi Rurali, Via di Casignano 25, Scandicci, Firenze, Italy

r t i c l e i n f o

rticle history:eceived 6 January 2016eceived in revised form 19 May 2016ccepted 23 May 2016

eyword:ld wheat varietiesixtures

gronomic characteristic of wheatheat flour quality

rganic agriculturearticipatory plant breedingvolutionary plant breedingread sensory analysisonsumer preference

a b s t r a c t

Wheat landraces and old varieties could have an important role for food security not only as source ofgene readily available for breeders, but also because they perform well in marginal environments andare more resilient as compared to the modern cultivars. The Italian cereal sector suffers from lack ofseed companies that breed specifically adapted varieties for organic and biodynamic farms. Participa-tory and evolutionary plant breeding (PPB and EPB) have been used in this research to (i) evaluate theagronomic characteristics of old, modern and mixture of varieties and landrace of bread wheat (Triticumaestivum spp.) and their adaptability to organic farming in hilly and mountainous areas; (ii) assess thetechnological, nutritional and functional properties of grains (rheological characteristic, macro and microelements contents and antioxidants); (iii) explore the consumers’ preferences for breads obtained by oldand modern varieties. Between five and seven old (Sieve, Verna, Gentil Rosso, Andriolo, Gambo di ferro,Frassineto and Abbondanza), two mixtures and four modern (Bolero, Blasco, Arabia and Bologna) vari-eties were tested for two years in between two and three organic farms (FARM1, FARM2 and FARM3) inhilly areas of Piedmont. Agronomic characteristic were strongly affected by locations and years. On aver-age, Bologna, Abbondanza and Arabia, and the two mixtures were the highest yielding varieties. Flourstrength (W) varied greatly ranging from 230 in 2011 for Andriolo to 38 in 2012 for Gambo di ferro. Glutenquality, expressed by GI, was found to be almost within the optimal range but was affected by the year.All six bread samples were acceptable to the 233 consumers who scored them, but the bread producedwith old wheat varieties, particularly with Andriolo and Gambo di Ferro, was the preferred one. The oldvarieties and their mixtures yielded less than the modern varieties but with higher stability as shown

by the inability of the modern varieties in FARM1 to survive the winter (they were not harvested) whilethe old varieties reached maturity showing higher robustness, Therefore, the use of old bread wheatvarieties and their mixture, assessed with participatory and evolutionary plant breeding, could representa strategy for local communities to cope with climate change while improving food security and foodquality.

© 2016 Elsevier B.V. All rights reserved.

. Introduction

Bread wheat (Triticum aestivum L.) is the second most importanttaple crop in the world (FAOSTAT, 2014). Given its predominantresence in human diet, cultivated wheat has to meet quality cri-

∗ Corresponding author.E-mail address: p.migliorini@unisg.it (P. Migliorini).

ttp://dx.doi.org/10.1016/j.eja.2016.05.011161-0301/© 2016 Elsevier B.V. All rights reserved.

teria. Bread wheat production in Italy is insufficient to satisfy thedemand, and 4,7 Mt are imported annually, of which 77% fromEU countries (France, Austria, Germany), 12% from OCSE (Canada,Australia and USA) and 11% from other countries (Ukraine, Mexicoand Turkey) to fulfil the market needs for flour (Zuppiroli, 2013).

The domestic wheat production (2.9 Mt) meets about 38% of thenational requirement of the volumes processed by the millingindustry (80%) and confectionery (20%). The massive import doesnot stimulate the selection of grains of quality, in particular for the

1 J. Agro

ovrblifhasdbteahtf2vnwTb((utbr

asITm12a2R

aqhtIf(hairotgtia((iasti

32 P. Migliorini et al. / Europ.

rganic sector, which has a limited choice of specifically adaptedarieties. A range of diverse varieties, adapted to different envi-onments, is needed to respond to the needs of the organic andiodynamic farms that need locally adapted varieties. During the

ast centuries, wheat breeding efforts have concentrated on yieldncreases, and modern varieties are characterized by genetic uni-ormity and adaptation to conventional agriculture typically usingigh-energy inputs in terms of fertilizers, herbicides, insecticidesnd fungicides (Guarda et al., 2004). Modern varieties are oftenelected in favourable environments that do not represent theiversity of local conditions. Modern agriculture and conventionalreeding have resulted in a reduction of genetic diversity andhe stagnation of cereal yields in less favourable areas (Newtont al., 2010). A local variety of plant (and animal), or landrace is

domesticated, traditional, regional ecotype, locally adapted thatas developed over time becoming adapted to its natural and cul-ural agricultural environment (and pastoralism) due to isolationrom other populations of the same species (Jones et al., 2008; FAO,013; Camacho et al., 2005). Local wheat varieties, or landraces, areery tall, less productive than modern varieties, not homogeneousor regular, with variable grain protein content, and with glutenith less gliadin and less glutenins (Sanchez-Garcia et al., 2015).

heir flours are not suitable for the industrial processing and theest results are obtained when milled with natural stone groundflour type 2), with attractive results from the quality point of viewsensorial and functional); they produce excellent bread with these of sourdough (Torri et al., 2013). As they are dynamic (evolu-ionary) populations, they represent a way to protect and enhanceiodiversity, to create local supply chains for small producers whileepresenting an historical/cultural/social heritage.

The lack of suitable genotypes as well as the difficulty of using high-input based agriculture, are the main causes of the con-tant and progressive abandonment of cereal cultivation in mosttalian marginal areas (hilly areas), once extensively cultivated.he area planted with autumn-winter cereals (durum wheat, com-on wheat, barley, oats and other cereals) in Italy decreased by

0% in 2013 compared to 2012, to 2.1 million hectares (MIPAAF,013). This was due to both poor autumn weather, which in manyreas delayed planting, and to reduced market prices. Piedmont in013 was the third highest producing region in Italy, after Emiliaomagna and Veneto, with 5.1 million tons (MIPAAF, 2013).

The production of certified organic wheat allowed to maintain premium price compared with the conventional market. Grainuality has been a secondary objective of breeding programme andas been limited to few attributes such as test weight, protein con-ent and gluten quantity and quality (Canerava et al., 1994). In thetalian context, bread wheat flour is commercially classified in dif-erent quality categories on the basis of a Synthetic Index of QualityIndice Sintetico di Qualità, ISQ). ISQ is based on parameters such asectolitre weight (HeW), falling number (determining the amylasectivity of some enzymes in dough, that influence dough viscos-ty), protein content, alveographic indexes (indicating the degree ofesistance opposed by the dough to a deformation stress, in termsf dough ‘strength’ – alveograph W – and the ratio between doughenacity and dough extensibility – alveograph P/L ratio), and farino-raph stability (expressing the time for which the dough maintainshe maximum consistency during the dough mixing process). Rang-ng from the stronger to the weaker type, the four quality categoriesre defined as improved wheat (FF), superior bread making wheatFPS), ordinary bread making wheat (FP) and wheat for biscuitsFB). This classification was proposed in 1997 (Borasio, 1997) andt is suitable for industrial processing, which require grains with

high gluten content and high dough strength, in order to obtaintandardized loaves with the use of selected yeasts and short risingimes. Another very important parameter for the milling industrys the Hardness Index as it affects the yield of flour.

nomy 79 (2016) 131–141

In the seed market, wheat varieties are classified according totheir potential technological features following the ISQ, based onwhich prices for the various commodity exchanges are determined.Therefore, it becomes essential the varietal selection for yield andmilling characteristics, which in turn depends on soil characteris-tics, climate (rainfall, temperature), as well as biotic (weeds, insects,cryptogams) and abiotic (lodging) stresses.

In Italy, bread production was liberalized as an activity relatedto agriculture in 2006 (Decreto Legge n.233 del 4 Luglio 2006) andin 2010 (Decreto del Ministero dell’Economia e delle Finanze del 5Agosto 2010). This authorization leads to some technical consider-ations about whether the farm is able to produce the right qualityfor commercial bread, or rather a type of bread that can go beyondthat offered by the current market. Making bread in a farm meansmaking technical choices that affect the cultivation, processing andmarketing.

The main objective of this study was to recover and evaluateold varieties on-farm and to practice evolutionary plant breed-ing (Ceccarelli, 2009). In situ and on farm conservation offer stablesolutions to biotic and abiotic problems, provide opportunities forscientific research and for genetic improvement, meet the needs ofboth farmers and the market for local products, and respect histor-ical and cultural traditions. This also allows the development of alocal supply chain for the production of quality bread that involvesresearchers, producers, processors and consumers, using a partic-ipatory plant breeding approach (PPB). PPB is an important toolof agricultural research to find varieties for low-input agricultureacceptable to farmers (Ceccarelli et al., 2000). Cereal improve-ment programs conducted to date addressed mainly higher yieldsand improved technological characteristics for industrial transfor-mation, while neglecting nutritional, functional, digestibility andpotential allergenicity characteristics (Van den Broeck et al., 2010).

A further development of PPB is evolutionary plant breeding(EPB) which is a simple but very innovative and revolutionary ideainitially proposed by Suneson (1956) but never applied. By lettingmixtures of varieties, or of breeding lines, or of segregating popu-lations, evolve under the selection pressures of soil type, climate,agronomic practices and pests, every year the genetic composi-tion of the seed, which is harvested, will be different from thatof the seed which was planted. The farmers will then harvest avintage wheat, which is the fruit of a blend of those local varietiescultivated in that place, that have been more productive, more com-petitive, more successful in that year. Just like wine, we will havea vintage bread and typical of the area, without having to resort todisciplinary and bureaucratic super-controls. In addition, this pop-ulation will evolve over time, adapting gradually and continuouslyto climate change (Ceccarelli, 2014).

The specific objectives of this work were to: (i) evaluate theagronomic characteristics of old, modern and mixture of varietiesand landrace of wheat (Triticum aestivum L.) and their adaptabil-ity to organic farms in hilly and mountainous areas; (ii) assessthe technological, nutritional and functional properties of grains(rheological characteristic, macro and micro elements contents andantioxidants); (iii) explore the consumers’ preferences for breadsobtained by old and modern varieties of wheat.

2. Materials and methods

2.1. Farming techniques

The research was conducted for three years (2010/11, 2011/12

and 2012/13) in three organic farms (FARM1, FARM2 and FARM3).In 2010/2011 and 2011/2012, in the context of a project of vari-etal comparison of wheat for bread-making quality, five old (Sieve,Verna, Gentil Rosso, Andriolo and Gambo di ferro) and three mod-

P. Migliorini et al. / Europ. J. Agronomy 79 (2016) 131–141 133

Table 1Characteristic of the variety utilised in the research.

ID Variety Constitutor Year Origin zone/Adaptability

1 Andriolo Landrace – Pistoia, mountain2 Gambo di ferro – – –3 Gentil rosso Landrace – Tuscany and South Italy4 Sieve Mario Gasparini 1960 mountain5 Verna Mario Gasparini 1953 mountain10 Frassineto Marco Michaellis 1932 Mountain and hilly12 Abbondanza Marco Michaellis 1950 Hilly and plain11 Mixture1 UNISG/AIAB 2008 Mountain and hilly13 Mixture2 Salvatore Ceccarelli 2000 Dry areas6 Bolero* (FPS) 356-Ets Claude Camille Benoist, 159-Venturoli Sementi S.R.L. 19877 Blasco* (FPS) 91-Co.Na.Se. Consorzio Nazionale Sementi S.R.L. 20028 Arabia* (FP) 101-Apsovsementi S.P.A., 91-Co.Na.Se. Consorzio Nazionale Sementi S.R.L. 20099 Bologna* (FF) 356-Ets Claude Camille Benoist, 159-Venturoli Sementi S.R.L. 1999

F bread(

etaUw1ro“vvwo

atMittIaar

T

2

q

2

hheiF

2

g(qs

m

PS: superior bread making wheat (frumento panificabile superiore); FP: ordinarySource: http://www.sian.it/mivmPubb/listeVarieta.do).

rn (Bolero FP, Blasco FPS e Arabia FB) varieties were tested forwo years in two hilly areas of Piedmont: FARM1 in Costa Vescov-to (Alessandria province) and FARM2 in Barge (Cuneo province).nfortunately, the modern varieties in FARM1 didn’t survive theinter and were not harvested. The sowing dates were 16/12/2010,

1/11/2010 and 11/10/2011, 17/11/2011 in FARM1 and FARM2espectively, at a seed rate of 450 seeds m−2. In addition, as partf the trial “National Network of organic wheat”, coordinated byUnità di ricerca per la selezione dei cereali e la valorizzazione dellearietà vegetali (CREA-SVC) in S.Angelo Lodigiano (LO), 16 modernarieties were cultivated and compared in the same two farms. Weill refer to the performance of modern varieties for comparison

nly.In 2012/2013, two old varieties (Frassineto and Abbondanza)

nd a modern reference variety (Bologna, FF) were added to therial together with two variety mixtures (mixture 1 and mixture 2).

ixtures 1 is a blend of the old 7 varieties listed above (the 5 testedn 2010 and 2011 and the 2 added in 2012). Mixture 2 is a populationhat contains several hundred breeding lines obtained by one ofhe authors (SC) from the bread wheat breeding program of thenternational Centre for Agriculture Research in Dry Areas (ICARDA)nd that was assembled in 2009. The sowing dates were 24/10/2012nd 08/11/2012 in FARM1 and FARM3 in Valgrana (Cuneo province)espectively.

The list of varieties and their characteristics are summarised inable 1.

.2. Data collection

Data were collected on (i) agronomic traits, (ii) grain and flourualitative traits and (iii) consumers’ preference on bread.

.2.1. Agronomic traitsThe agronomic traits included humidity of harvested grain (%),

ectolitre weight (kg h L−1), grain yield (t/ha), tillering (1–10), planteight (cm), frost damage (1–10), lodging (1–10), and date of earmergence (DAS – days after sowing). These traits were measuredn 2010/11 and in 2011/12 in FARM1 and FARM2, in 2012/13 inARM1 and FARM3.

.2.2. Quality traitsQuality traits included hardness (Ha), protein content (PC), total

luten content (Gtot), gluten index (GI), Hagberg falling numberHFN), total polyphenols and flavonoids, and bread production

ualities of the flour such as P (resistance to stretching), L (exten-ibility) and W (strength).

Grain samples of 50 g were ground with a 1-mm sieve Cyclotecill (Tecator, Sweden). Protein content (N × 5.7, dry weight, AACC

making wheat (frumento panificabile); FF: improved wheat (frumento di forza).

39-10 (AACC International, 2000)), and hardness (AACC 39–70(AACC International, 2000)) were determined on the wholemealby a NIR System Model 6500 (FOSS NIRSystems, Laurel, MD). TheHagberg falling number, an indicator of sprouting resistance, wasdetermined according to UNI ISO 3093:2009 in order to estimate�-amylase activity in cereal grains. Wet gluten content (Gtot)and gluten index (GI) were analysed according to AACC 38-12.Gluten was separated from whole-wheat flour by centrifugation.The GI determines the gluten characteristics, indicating whetherthe gluten is weak or strong.

In order to evaluate the functional content, the extraction ofsoluble (free) and insoluble (bound) phenolic compounds was per-formed according to the method of Dinelli et al. (2011). Polyphenolcontent in both the free and bound fractions was measured usingthe spectrophotometric Folin-Ciocalteu method (Lamba 25 Spec-trophotometer, Perkin-Elmer Corporation, USA) with gallic acidas the reference standard (Singleton et al., 1999). Similarly, theflavonoid content was determined using a colorimetric methodwith catechin as the reference standard (Adom et al., 2003). The freeand bound polyphenol were added to the flavonoid contents to pro-vide the total content. C and N were measured in flour samples (2 g)using a Flash Elemental Analyser 1112 NC (Thermo Fisher Scientific,Waltham, MA, USA) according to the manufacturer’s instructions.

Seventeen mineral elements (Al, B, Ba, Ca, Cd, Co, Cr, Cu, Fe, K,Mg, Mn, Na, Ni, P, Sr and Zn) were extracted by nitric/perchloric aciddigestion and measured by inductively coupled argon plasma (ICP)emission spectroscopy (IRIS Intrepid II XSP Radial, Thermo FisherScientific), using same techniques of other Italian studies (Ghiselliet al., 2008; Ghiselli et al., 2013).

The Chopin Alveograph was used to measure WPL by ChemicalLaboratory of the Chamber of Commerce of Turin according to UNIEN ISO 27971:2008.

W, P/L, hardness (Ha), protein content (PC), total gluten con-tent (Gtot), gluten index (GI), Hagberg falling number (HFN) weremeasured in 2010/11 and 2011/12 from FARM1 wheat samples;antioxidants in 2010/11 from FARM1 and in 2011/12 from FARM1and FARM2; mineral elements were measured in 2011/12 fromFARM1 and FARM2.

2.2.3. Bread samples and consumer testConsumers’ preference of bread was measured in 2010/11, from

the samples harvested in FARM1. Five wheat flour samples (type 2)were obtained from the stone milling of the following old wheat

varieties grown in an experimental field under organic agricul-ture management practice: Gentil Rosso, Sieve, Andriolo, Gambodi Ferro, Verna. A flour (type 2) obtained by the stone milling of amodern wheat variety (Bologna) was used as control.

134 P. Migliorini et al. / Europ. J. Agronomy 79 (2016) 131–141

Table 2Accumulated analysis of variance: one-way and interactions are shown for agronomic data. (*p < 0.05; **p < 0.01; ***p < 0.001; ns: not significant).

Source of variation Degree of freedom Humidity HeW Yield Denseness Height Frost damage Lodging Heading date

sig sig sig sig Sig sig sig sigYear 2 ** ns *** *** *** *** *** ***Locality 2 *** ns *** *** *** *** *** ***Variety 12 *** * *** ** *** *** *** ***YxL 1 ns ns ns ns * *** ns ***

wbwlwaos

e

poase

pwwp

ah(fsQ2(c

2

acaD(CtoaC

dsrotrd

YxV 11 ns ns **

LxV 15 ns ns **

YxLxV 4 ns ns ns

We used the same bread dough formula(5 kg of flour, 5 kg ofater, 1.7 kg of sourdough, 150 g of salt) and we followed the same

read making process: first sea salt is put into the mixer and slightlyarm water is added to melt it, then the flour is mixed for 4–5 min;

eft to rest for about 15–20 min; sourdough is added with a bit ofater and kneaded for a few minutes; left to rests at 27 ◦C for 80 min

nd then loaves are formed (600 g size); rising 29 ◦C for 4 h; etchingf the surface of the bread; baking in two consecutive steps (firsttep: 5 min at 240 ◦C; second step: 45 min at 220 ◦C).

For the sensory analysis, the samples were made by a local bak-ry and delivered to the evaluation site within 2 h from baking.

In order to assess the sensory acceptability of the bread sam-les, a consumer test was conducted with 233 regular consumersf bread (39% males, 61% females, ranging from 7 to 82 years ofge, with a mean age of 27), who voluntarily participated to theensory evaluation. Written informed consent was obtained fromach subject after the experiment has been described to them.

The test consisted in a sensory evaluation of the six bread sam-les. 1-cm bread slices (∼50 g), including the crust and the crumbs,ere served in blind conditions, in white paper plates and codedith a random three-digit number. Samples were served in com-

letely randomized and balanced order.Consumers were asked to observe, smell and taste the samples,

nd rate their overall liking. Liking was expressed on a 7-pointedonic scale ranging from ‘dislike extremely’ (1) to ‘like extremely’7). Participants were required to rinse their mouth with still wateror about one minute between samples. After the liking test, con-umers were asked to fill in a reduced version of the Food Choiceuestionnaire (13 items, 4-point scale; 1 = not at all important;

= a little important; 3 = moderately important; 4 = very important)Steptoe et al., 1995). Consumers took between 15 and 20 min toomplete their tasks.

.3. Experimental design and statistical analysis

The experimental design used for the on farm trials in 2010/2011nd 2011/2012 was a randomized block design with three repli-ations for a total of 15 plots (5 varieties and 3 replications)rranged in three rows and five columns in an area of about 30 m2.ata collected were subjected to univariate analysis of variances

ANOVA) utilising the GenStat statistical software package (GenStatommittee, 2011) for the following source of variation: year, loca-ions, variety and their interactions (Table 2). Univariate analysisf the data was carried out by a comparison of the mean values,ccording to Fisher’s least significant difference test (Snedecor andochran, 1980) at P ≤ 0.05 and P ≤ 0.01 probability levels.

In 2012/2013, the experimental design was a partially replicatedesign with ten varieties of which four replicated three times andix replicated two times, for a total of 24 plots arranged in threeows and eight columns. The best linear unbiased estimates (BLUEs)

f the genotype effects were calculated using the most suitable spa-ial model determined for the associated individual field layout,ows and columns. The procedure of Singh et al. (2003) was used toetect the best of the nine applicable models. These models com-

ns Ns ns *** ***** Ns * *** ***** * ns *** ***

prised factorial combinations of: (1) three ways of accounting forlinear trends in column direction, with or without a linear trendand random cubic smoothing spline (CS) in column number; and(2) three structures for plot errors, first-order autoregressive (AR)errors along rows, first-order autoregressive (AR) errors along rowsas well as AR along columns, and independent errors. Each of theabove models was fitted for the data of each trial by expressingthe plot position in row and column number. The REML (restrictedmaximum likelihood) method of the GenStat software (GenStatCommittee, 2011) was used to fit the model by setting the asso-ciated directives. For further details, see also Rollins et al. (2013).

In the case of the agronomic data and because of the additionalvarieties used in 2012, we conducted an unbalanced ANOVA inGenStat (GenStat Committee, 2011).

We used the BLUEs or the means from the unbalanced ANOVA toproduce a GGE biplot, using the GGE biplot software (Yan, 2001) toindicate possible specific adaptations of lines to the environments.The biplot was environment scaled/focused (Yan et al., 2000). Inthe GGE biplot, the genotypic PC1 scores are proportional to theexpression of the traits, while the PC2 scores are proportional todeviations associated with G × E interaction. Positive and negativeenvironmental PC1 scores indicate G × E interaction of cross-overtype, while positive and negative environmental PC2 scores indi-cate G × E interaction of non-cross-over type.

When the data is sufficiently approximated by the biplot, thecosine of the angle between the vectors of two testers approximatesto the correlation coefficients between them. In particular: (a) theorigin is the point with coordinates PC1 = 0 and PC2 = 0; (b) twotesters are positively correlated if the angle between their vectors is<90◦; (c) two testers are negatively correlated if the angle betweentheir vectors is >90◦; (d) two testers are independent if the anglebetween them is near 90◦; (e) 0◦ means r = 1; (f) 90◦ means r = 0;(g) 180◦ means r = −1.

Eventually similar genotypes are positioned closely; genotypesthat are similar in GGE value directions have a small angle acuteangles or <90◦, while dissimilar genotypes have a large angle (theangle formed between the first genotype, the origin and the secondgenotype) obtuse angle or between 90◦ and 270◦.

Genotypes far from the origin have a large genotype plus inter-action effect. If a given genotype and a given location vector are onthe same side of the origin that genotype performs above average inthat location. By contrast, a genotype, which is at the opposite sideof a location vector origin, performs below average in that environ-ment. Genotypes close to the origin have average performance in allenvironments. Locations with longer vectors are more discriminat-ing of the entries; those with short vectors are less discriminating;those located at the biplot origin are not discriminating.

One interesting feature of the GGEbiplot software is the ‘whichwon where’: in this feature the biplot is divided in sectors obtainedby first connecting all the most extreme genotypes in a polygon,

and secondly drawing perpendiculars from the origin to each sideof the polygon. The genotypes included in the same sector as one ormore testers (locations, years or traits) are those performing betterin those locations or years or for those trait. Genotypes included in

J. Agronomy 79 (2016) 131–141 135

sivo

AAofdccttcct

3

3

ra(

eaete(u(dyr

e(dFea

dD(Fevm

dtFo(

fwpai

ata

on

wh

eat v

arie

ty: t

he

hu

mid

ity

of

har

vest

ed

grai

n

(%);

the

hec

toli

tre

wei

ght (

kg

hL−1

);

the

yiel

d

(t/h

a); t

he

den

sen

ess

of

pla

nts

(1–1

0); p

lan

t hei

ght (

cm);

fros

t dam

age

(1–1

0); l

odgi

ng

(1–1

0); d

ate

of

ear

emer

gen

ceow

ing)

. Res

ult

s

of

one-

way

AN

OV

A

and

inte

ract

ion

s

are

show

n. L

DS

and

stan

dar

d

erro

r

valu

e

are

show

n

only

in

case

of

sign

ifica

nt

AN

OV

A

(val

ues

in

row

foll

owed

by

dif

fere

nt

lett

ers

are

sign

ifica

ntl

y

dif

fere

nt)

.

ety

Hu

mid

ity

(%)

HeW

(kg

hL−1

)

Yie

ld

(t/h

a)

Den

sen

ess

(1–1

0)

Hei

ght

(cm

)

Fros

t dam

age

(1–1

0)

Lod

gin

g

(1–1

0)

Hea

din

g

dat

e

(DA

S)

riol

o

12.9

9

b

0.39

77.6

0

a

1.97

0

2.07

e

0.19

2

8.06

bcef

0.25

5

106.

89

b

2.31

4

0.40

a

0.09

2

0.66

d

0.96

3

55.6

7

b

0.60

1bo

di f

erro

12.5

8

c

0.42

75.3

2

a

2.13

3

1.73

e

0.20

7

8.40

abce

0.25

9

126.

72

a

2.34

6

0.35

ab

0.09

3

1.51

b

0.98

0

59.2

1

ab

0.61

2il

ross

o

12.5

7

c

0.39

74.6

1

a

2.06

1

1.57

e

0.20

1

8.24

bcef

0.25

9

122.

45

ab

2.34

6

0.35

ab

0.09

3

1.30

cd

0.98

0

58.9

8

ab

0.61

2e

12.5

3

c

0.40

73.5

6

a

2.18

1

1.51

e

0.19

9

8.35

bce

0.26

4

107.

80

b

2.39

4

0.42

a

0.09

5

0.84

d

1.00

7

56.6

4

ab

0.62

9a

12.5

5

c

0.38

74.2

1

a

1.98

8

2.08

e

0.18

9

8.47

abce

0.25

1

116.

42

b

2.27

0

0.33

ab

0.09

0

1.07

d

0.93

9

59.7

8

ab

0.58

6si

net

o

12.1

5

c

0.97

78.1

6

a

4.84

9

2.09

e

0.47

9 9.

50

a

0.56

0

130.

98

a

5.07

1

0.00

b

0.20

1

2.93

a

1.87

6

59.3

4

ab

1.17

1n

dan

za

13.1

5

ab

0.87

76.9

9

a

4.38

4

2.44

de

0.43

2 9.

07

abce

0.56

0

69.6

2

c

5.07

1

0.00

b

0.20

1

0.00

d

1.87

6

61.8

3

a

1.17

1u

re1

12.4

2

c

0.87

79.4

6

a

4.38

4

2.98

c

0.43

2

9.15

abc

0.56

0

131.

59

a

5.07

1

0.00

b

0.20

1

4.78

a

1.87

6

60.7

2

a

1.17

1u

re2

8.98

d

0.77

57.7

3

b

3.84

8

2.02

e 0.

432

9.24

ab

0.45

7

106.

63

b

4.14

2

0.00

b

0.16

4

0.00

d

1.53

3

60.9

3

a

0.95

7ro

14.6

5

ab

0.52

71.2

0

a

2.68

6

2.64

d

0.26

0

7.21

fg

0.34

7

58.0

7

d

3.14

7

0.64

a

0.12

5

0.00

d

1.18

9

39.9

1

cd

0.74

2co

16.4

3

a

0.52

72.7

0

a

2.68

6

2.14

e 0.

260

6.25

g

0.34

7

60.4

5

cd

3.14

7

0.32

ab

0.12

5

0.00

d

1.18

9

42.4

0

c

0.74

2ia

14.8

9

ab

0.52

72.8

6

a

2.84

3

3.57

b

0.26

0

7.47

ef

0.34

7

66.4

6

cd

3.14

7

0.21

ab

0.12

5

0.00

d

1.18

9

39.1

3

d

0.74

2gn

a

12.8

6

c

0.76

79.6

9

a

3.78

8

3.87

a

0.37

3

8.84

abce

0.48

8

68.7

9

c

4.42

0

0.00

b

0.17

5

0.00

d

1.63

4

59.7

3

ab

1.02

01.

74

8.85

0.87

1.10

9.96

0.40

3.80

2.38

P. Migliorini et al. / Europ.

ectors void of testers (locations, years or traits) are those perform-ng poorly everywhere and/or for each year or trait, while sectorsoid of genotypes (locations, years or traits) are those in which nonef the genotypes performed well.

Preference data from consumers were submitted to a two-wayNOVA mixed model (fixed factor: sample; random factor: subject).

Fisher LSD post hoc test was used to test the significance (P ≤ 0.05)f relative mean differences. A k-means cluster analysis was per-ormed on preference data from consumers and three groups wereefined: cluster 1 (Cl1) consisting of 106 consumers, cluster 2 (Cl2)onsisting of 57 consumers and cluster 3 (Cl3) consisting of 70onsumers. Preference data from each cluster were submitted to awo-way ANOVA mixed model (fixed factor: sample; random fac-or: subject). Ratings of Food Choice Questionnaire items from eachluster were submitted to a one-way ANOVA model (fixed factor:luster). After ANOVA, a Fishers’ LSD test was performed to comparehe mean values (P < 0.05).

. Results

.1. Agronomic characteristic of wheat varieties

Accumulated analysis of variance for agronomic data (Table 2)evealed highly significant difference for most of the sources of vari-tion/variables. Results of ANOVA are shown for the 12 varietiesTable 3).

Average yields of all varieties was 2.36 t/ha. The old vari-ties, landrace and their mixture (non-modern) yielded 2.05 t/hand the modern 3.05 t/ha. Bologna (modern) resulted the high-st yielding variety followed by Arabia (modern) and Mixture1,he blend of local old varieties. Blasco was not significantly differ-nt from all non-modern varieties. A highly significant differenceP < 0.001) was found between the yields of the locations. In partic-lar, FARM2 showed a higher average yield than FARM3 and FARM12.15 ± 0.1039, 2.02 ± 0.1164 and 1.92 ± 0.2027 t ha−1 respectively,ata not shown). Yield was also significantly different betweenears with 1.55 ± 0.1294, 2.05 ± 0.1154 and 2.55 ± 0.1525 t ha−1

espectively in 2011, 2012 and 2013.Tillering was highest in Frassineto and lowest in Blasco. As

xpected, the old varieties were much taller than modern varieties113.2 vs 63.4 cm), data not shown. There were highly significantifferences in plant height among the varieties, with Mixture1,rassineto, Gambo di ferro and Gentil Rosso being the tallest vari-ties. This affected lodging that was higher in Mixture1, Frassinetond Gambo di ferro compared to the other varieties.

There were highly significant differences (P < 0.001) for headingate between the years (57.0 ± 0.37; 46.9 ± 0.40; and 61.0 ± 0.37AS in 2011, 2012 and 2013 respectively), the locations

55.0 ± 0.28; 43.6 ± 0.38; and 81.9 ± 0.48 DAS in FARM1, FARM2 andARM3 respectively) and among the varieties, with Arabia being thearliest (39.1 DAS) and Abbondanza the latest (61.8 DAS). No rele-ant frost damage was observed in any varieties that resulted in aean value of 0.23 (Table 3).

Average hectolitre weight (HeW) was 74.2 kg hl−1; significantifferences were found among the varieties with Mixture2 showinghe lowest value of 57.7 kg hl−1. In particular, Bologna, Mixture1,rassineto, Andriolo, Abbondanza and Gambo di ferro had a valuef HeW above the reference used in the Synthetic Index of Quality75.00 kg hl−1).

A strong interaction was observed for HeW, as shown by theact that the vectors for the two years are in opposite direction,

hile no correlation between the two years for observed for theercent humidity of harvested grain (U%) (Fig. 1). Gambo di ferrond Frassineto (nr. 2 and 10) were the most susceptible to lodg-ng together with Mixture 1, which was also the tallest. Bologna Ta

ble

3A

gron

omic

d(D

ays

Aft

er

S

ID

Var

i

1

An

d2

Gam

3

Gen

t4

Siev

5

Ver

n6

Fras

12

Abb

o11

Mix

t13

Mix

t6

Bol

e7

Bla

s8

Ara

b9

Bol

oLS

D

136 P. Migliorini et al. / Europ. J. Agronomy 79 (2016) 131–141

Fig. 1. The “which won where” feature of the GGE biplot for the agronomic dataof 10 bread wheat varieties in two years U% = percent humidity of harvested grain.HeW = hectolitre weight. GY = grain yield. FIT = tillering. PH = plant height. CD = frostdamage. LDG = lodging. DH = Date of ear emergence. Underscore is followed by theyear code.

(hGh2

BtdBah

Fig. 2. The “which won where” feature of the GGE biplot for the agronomic data of13 bread wheat varieties in three years U% = percent humidity of harvested grain.

W indices, i.e., an increase of protein leads to an increase of W. Thehardness score, expression of the endosperm texture and gener-ally formulated as a resistance to plastic strain and cracking at a

TQgRd

9) was the highest yielding while no consistency was found forectolitre weight (because of the large interaction) with varietiesentil Rosso, Mixture2 and Mixture1 (3, 11 and 13) having theighest HeW in 2011 and Mixture 1 having the highest HeW in013.

Overall, using the unbalanced analysis over the three years,ologna, Abbondanza and Arabia (9, 12 and 8), and the two mix-ures were the highest yielding varieties (Fig. 2), Mixture2, Gamboi ferro, Gentil Rosso, Verna (11, 2, 3 and 5) the tallest, varietiesolero and Blasco (6 and 7) those with more humidity at harvest,nd Frassineto (10) the most susceptible to lodging, the latest ineading and the highest in tillering.

able 4uality parameters of wheat grain and bread production qualities of the flour: hardnessluten content (Gtot) and gluten index (GI); Hagberg falling number (HFN); total polypheesults of one-way ANOVA and interactions are shown. LDS is provided in case of signiifferent letters are significantly different.

W P/L Hardness

Source of variation (J 10−4) sig (index) sig (score) sig

Year *** *** ***

2011 139.70 a 0.83 a 46.87 b

2012 59.30 b 0.40 b 55.80 a

Variety *** ns ***

Andriolo 160.80 a 0.61 82.50 a

Gambo di ferro 57.50 e 0.64 34.00 c

Gentil rosso 78.30 d 0.55 37.83 c

Sieve 113.30 b 0.56 64.67 b

Verna 87.50 c 0.72 37.67 c

YxV ** ns ns

2011 Andriolo 230.00 a 0.82 75.33

2011 Gambo di ferro 76.70 def 0.68 31.33

2011 Gentil rosso 110.00 cd 0.82 32.33

2011 Sieve 156.70 b 0.76 62.33

2011 Verna 125.00 bc 1.047 33.00

2012 Andriolo 91.70 cde 0.40 89.67

2012 Gambo di ferro 38.30 g 0.59 36.67

2012 Gentil rosso 46.70 fg 0.28 43.33

2012 Sieve 70.00 efg 0.35 67.00

2012 Verna 50.00 fg 0.40 42.33

HeW = hectolitre weight. GY = grain yield. FIT = tillering. PH = plant height. CD = frostdamage. LDG = lodging. DH = Date of ear emergence.

3.2. Bread wheat flour quality

As expected, bread quality was strongly affected by year andvariety (Table 4). Flour strength (W) varied between years and vari-eties ranging from 230 in 2011 for Andriolo to 38 in 2012 for Gambodi ferro. Andriolo had the highest strength, which could lead to goodincrease in bread volume (classified as FP: ordinary bread makingwheat, according to ISQ). In addition, the P/L showed a mean value(0.61), which classified the wheat as FP but without any signifi-cant difference among years and varieties. The mean value of totalprotein content was very high (14.4% dm) and classified the wheatas improved wheat (FF). However, there was great variability asin the 2012 protein content was about 30% lower than in 2011.There was a strong linear association between protein content and

(score) and hardness index (hard, medium, soft); protein content (N x 5.7); totalnols and flavonoids; P (resistance to stretching), L (extensibility) and W (strength).ficant ANOVA (*p < 0.05; **p < 0.01; ns: not significant). Values in row followed by

HI Prot Gtot GI HFN

(% dm) sig (% dm) sig (%) sig (sec) sig*** ns ** ns

M 16.14 a 10.12 59.30 b 394.30M 12.63 b 9.67 70.90 a 385.90

ns * *** ***H 14.25 9.82 b 75.70 a 430.50 aS 14.15 9.69 b 63.20 c 346.50 bS 14.57 10.04 ab 68.80 b 327.70 bM 13.80 8.76 b 56.50 d 425.30 aS 15.15 11.19 a 61.50 c 420.50 a

ns ns ns nsM 15.57 10.07 71.00 450.30S 16.07 9.77 57.30 317.70S 16.43 10.44 67.30 328.30M 15.30 8.45 44.00 442.70S 17.33 11.88 57.00 432.70H 12.93 9.57 80.30 410.70S 12.23 9.60 69.00 375.30M 12.70 9.63 70.30 327.00M 12.30 9.07 69.00 408.00M 12.97 10.50 66.00 408.30

P. Migliorini et al. / Europ. J. Agronomy 79 (2016) 131–141 137

Table 5Mineral element. Results of one-way ANOVA and interactions are shown. The concentration is expressed as mg/kg of s.s. LDS is provided in case of significant ANOVA (*:p < 0.05; **: p < 0.01; ns: not significant). Aluminium (Al). Barium (Ba). Calcium (Ca). Cadmium (Cd). Cobalt (Co). Chromium (Cr). Copper (Cu). Iron (Fe). Potassium (K). Lithium(Li); Magnesium (Mg); Manganese (Mn); Molybdenum (Mo); Sodium (Na); Nickel (Ni); Phosphorus (P); Lead (Pb); Selenium (Se); Strontium (Sz); Vanadium (V); Zinc (Zn).

Al Ba Ca Cd Co Cr Cu Fe K Li Mg

Source of variation sig sig sig sig sig sig sig sig sig sig sigLocality ns *** * *** ns ns * ns ns ns ***FARM1 2.97 1.81 b 345.60 a 0.08 b 0.01 0.26 4.94 a 43.00 2975.00 0.01 1223.00 aFARM2 9.95 4.15 a 307.50 b 0.12 a 0.04 0.32 4.22 b 52.30 2997.00 0.09 1072.00 bVariety ns *** ns ns ns ns ns ns *** ns nsAndriolo 3.50 3.0 b 354.20 0.09 0.04 0.24 4.07 38.00 2596.00 d 0.02 1167.00Gambo di ferro 3.95 3.0 b 306.40 0.09 0.02 0.15 4.02 39.10 2922.00 c 0.03 1128.00Gentil rosso 8.27 2.5 bc 343.30 0.08 0.00 0.32 4.75 72.00 3142.00 b 0.04 1202.00Sieve 12.13 4.2 a 336.20 0.11 0.03 0.44 4.86 47.30 3269.00 a 0.04 1146.00Verna 3.96 2.1 c 292.80 0.12 0.03 0.32 5.14 41.90 3002.00 c 0.13 1094.00LxV ns ns ns ns ns ns ns ns ns ns *FARM1 Andriolo 1.68 1.47 397.20 0.08 0.01 0.43 4.84 40.10 2626.00 0.01 1219.00FARM2 Andriolo 5.44 4.61 311.30 0.10 0.08 0.05 3.34 35.90 2566.00 0.04 1115.00FARM1 Gambo di ferro 0.74 1.73 338.80 0.07 0.00 0.19 4.42 36.80 3014.00 0.01 1177.00FARM2 Gambo di ferro 7.38 4.32 273.90 0.10 0.03 0.10 3.65 41.40 2829.00 0.06 1078.00FARM1 Gentil rosso 12.33 1.61 329.80 0.07 0.00 0.22 5.27 59.00 3098.00 0.01 1335.00FARM2 Gentil rosso 1.97 3.45 356.80 0.09 0.00 0.41 4.41 84.90 3185.00 0.06 1069.00FARM1 Sieve 1.23 2.79 338.10 0.08 0.01 0.25 5.05 36.80 3049.00 0.02 1131.00FARM2 Sieve 23.82 5.58 334.30 0.15 0.05 0.63 4.69 57.90 3490.00 0.07 1161.00FARM1 Verna 0.44 1.45 324.10 0.09 0.03 0.22 5.23 42.40 3087.00 0.01 1253.00FARM2 Verna 7.73 2.80 261.50 0.15 0.03 0.42 5.05 41.50 2916.00 0.24 935.00

Mn Mo Na Ni P Pb Se Sr V Zn

Source of variation sig Sig sig sig sig sig sig sig sig sigLocality *** *** ns ns *** *** * *** *** nsFARM1 19.25 b 2.06 a 12.11 0.53 2450.00 b 0.69 a 0.01 b 3.53 a 0.30 b 52.55FARM2 67.72 a 0.95 b 12.35 0.65 3382.00 a 0.00 b 1.68 a 1.21 b 1.60 a 49.37Variety * ns ns ns ns ns ns ns ns nsAndriolo 50.29 a 1.25 11.93 0.53 2649.00 0.34 0.16 2.38 1.02 52.03Gambo di ferro 44.05 ab 1.49 9.16 0.53 2982.00 0.34 0.37 2.37 0.99 49.15Gentil rosso 41.71 b 1.42 12.48 0.56 3042.00 0.35 0.49 2.34 0.91 50.29Sieve 43.24 b 1.66 16.20 0.64 3081.00 0.32 1.40 2.76 0.98 52.99Verna 38.12 b 1.71 10.99 0.68 2829.00 0.37 1.81 1.99 0.84 49.67LxV * ns ns ns ns ns ns ns ns nsFARM1 Andriolo 19.50 d 1.81 13.86 0.61 2160.00 0.68 0.00 3.36 0.28 55.02FARM2 Andriolo 81.08 a 0.68 9.86 0.45 3138.00 0.00 0.31 1.39 1.76 49.25FARM1 Gambo di ferro 17.53 d 2.25 11.95 0.56 2437.00 0.67 0.00 3.61 0.31 44.76FARM2 Gambo di ferro 70.57 b 0.73 6.16 0.51 3526.00 0.00 0.73 1.14 1.68 53.25FARM1 Gentil rosso 21.88 d 2.04 11.81 0.50 2706.00 0.71 0.00 3.32 0.35 54.81FARM2 Gentil rosso 61.54 bc 0.80 13.2 0.62 3379.00 0.00 0.99 1.37 1.47 47.39FARM1 Sieve 20.27 d 2.08 10.83 0.41 2352.00 0.64 0.07 4.19 0.27 52.08

381259306

fweaV

e(

gebVa

ttfo

aS

FARM2 Sieve 66.22 bc 1.25 21.96 0.86

FARM1 Verna 17.05 d 2.13 12.13 0.55

FARM2 Verna 59.19 c 1.28 9.23 0.81

orce concentrated on the surface of a given body, classified theheat as “medium” Hardness Index. There were significant differ-

nces among varieties: Andriolo had the hardest texture classifieds “hard”, Sieve as “medium” and Gambo di ferro, Gentil Rosso anderna as “soft” grain.

The HFN mean value was 390 s with highly significant differ-nces among varieties, with Andriolo showing the highest value430.5 s).

Gluten index (GI) is correlated with the strength and elasticity ofluten, and does not strictly depend on protein content. GI param-ter can take values between 0 and 100 and the optimum value isetween 65 and 80. Values greater than 80 indicate a strong gluten.alues lower than 65 describe a gluten having an initial proteolyticctivity from moderate to strong.

Gluten quality, expressed by GI, was found to be almost withinhe optimal range (between 57 and 80.3) but with an influence ofhe year (2012 was higher than 2011); Andriolo had the highest GIollowed by Gentil Rosso: they were statistically different from thethers.

The mineral element (Table 5) shows some statistical differencemong locations − but without any clear trend. Among varieties,ieve had a higher content of Ba and K (4.2 and 3269 �g−1, respec-

0.00 0.00 2.73 1.33 1.70 53.848.00 0.73 0.00 3.16 0.29 57.050.00 0.00 3.61 0.83 1.38 42.79

tively) and Andriolo of Mn (50 �g−1). Significant genotype byenvironment interaction effects were observed only on Mg and Mnconcentrations.

There were highly significant differences between the totalpolyphenols and flavonoids content in the two years (Table 6) witha content of all antioxidants expect for bound flavonoids higherin 2012 than in 2011. Significant differences were also observedbetween varieties, showing that Gentil Rosso had a much higheramount of total, free and bound polyphenols.

Those results confirm the dependency of biosynthesises andaccumulation of phenolic compound on both variety and environ-mental conditions (abiotic and biotic stress).

3.3. Consumers’ preferences

The average overall liking ratings expressed by all the 233 con-sumers for the six bread samples are reported in Table 7. Averageratings of all samples were higher than the central value of the

scale (4 = neither like nor dislike) to indicate that all tested breadswere acceptable for consumers. The control bread was significantlythe least liked sample (P < 0.0001). Consumers clearly preferred thebread samples produced with old wheat varieties. In particular,

138 P. Migliorini et al. / Europ. J. Agronomy 79 (2016) 131–141

Table 6Antioxidant. Results of one-way ANOVA and interactions are shown. LDS is provided in case of significant ANOVA (*: p < 0.05; **: p < 0.01; ***: p < 0.001; ns: non-significant).

Free polyphenol Bound polyphenol Tot. Polyphenol Free flavonoids Bound flavonoids Total flavonoidsSource of variation sig sig sig sig sig sig

Year *** *** *** *** ns ***2011 0.45 1.67 2.12 0.15 0.60 0.752012 0.94 2.26 3.20 0.22 0.65 0.87

Locality ns ns ns *** ns nsFARM1 0.784 2.09 2.88 0.18 0.64 0.823FARM2 0.763 2.01 2.77 0.22 0.63 0.847

Variety * ** *** ns * nsAndriolo 0.776 b 2.06 b 2.83 c 0.21 0.66 a 0.86

Gambo di ferro 0.752 bc 2.06 b 2.81 c 0.18 0.62 a 0.80Gentil rosso 0.846 a 2.21 a 3.06 a 0.19 0.68 a 0.87

Sieve 0.731 c 1.83 c 2.56 d 0.20 0.64 a 0.84Verna 0.782 b 2.16 ab 2.94 b 0.19 0.59 b 0.78

YxV ns ns ns *** * ***2011 Andriolo 0.388 1.66 2.05 0.12 d 0.63 abc 0.74 c

2011 Gambo di ferro 0.971 2.25 3.22 0.25 a 0.67 ab 0.93 a2011 Gentil rosso 0.438 1.67 2.10 0.15 c 0.59 abc 0.74 c

2011 Sieve 0.909 2.25 3.16 0.20 b 0.64 abc 0.84 b2011 Verna 0.563 1.65 2.21 0.19 b 0.63 abc 0.82 b

2012 Andriolo 0.987 2.49 3.48 0.20 b 0.71 a 0.90 a2012 Gambo di ferro 0.444 1.49 1.93 0.12 d 0.51 c 0.63 d

2012 Gentil rosso 0.875 2.00 2.88 0.24 a 0.70 a 0.94 a2012 Sieve 0.417 1.88 2.30 0.16 c 0.67 ab 0.83 b

2012 Verna 0.965 2.30 3.26 0.21 b 0.54 b 0.75 cLxV * ns ns ** ns ns

FARM1 Andriolo 0.795 bc 2.10 2.89 0.1825 c 0.68 0.87FARM1 Gambo di ferro 0.738 d 1.97 2.71 0.2598 a 0.60 0.86

FARM1 Gentilrosso 0.794 b 2.01 2.81 0.1684 c 0.63 0.80FARM1 Sieve 0.666 e 2.14 2.81 0.2099 b 0.60 0.81

FARM1 Verna 0.797 b 2.26 3.05 0.2011 b 0.69 0.89FARM2 Andriolo 0.943 a 2.12 3.07 0.1761 c 0.67 0.85

FARM2 Gambo di ferro 0.738 d 1.85 2.59 0.1937 c 0.61 0.81FARM2 Gentilrosso 0.717 d 1.79 2.51 0.2107 bc 0.69 0.90

FARM2 Sieve 0.798 b 2.23 3.03 0.1679 c 0.59 0.76FARM2 Verna 0.752 cd 2.01 2.76 0.2422 ab 0.58 0.82

Table 7Overall liking for six bread samples expressed by all consumers (n = 233). Cluster 1 (n = 46). Cluster 2 (n = 57) and Cluster 3 (n = 70). Data are average ± error standard. Valuesin the same column with differing letters are significantly different (Fisher’s LSD. p < 0.05).

Bread samples All Cluster 1 Cluster 2 Cluster 3

Control 4.3 ± 0.1c 3.8 ± 0.1c 3.2 ± 0.2 6.0 ± 0.1abSieve 4.7 ± 0.1b 4.8 ± 0.1b 3.5 ± 0.2 5.4 ± 0.2cAndriolo 5.0 ± 0.1a 5.2 ± 0.1a 3.6 ± 0.2 5.7 ± 0.1bcGambo di Ferro 5.0 ± 0.1a 4.9 ± 0.1ab 3.9 ± 0.2 6.1 ± 0.1aVerna 4.8 ± 0.1ab 4.9 ± 0.1ab 3.5 ± 0.1 5.6 ± 0.1c

5.0 ±17.32<0.00

tAgowf

biac4molPl3a

Gentilrosso 4.8 ± 0.1ab

F 9.867

P <0.0001

he highest average ratings were given to the bread obtained fromndriolo and Gambo di Ferro; these were comparable to the ratingiven to Verna and Gentil Rosso and significantly higher than theverall rating given to Sieve (P < 0.0001). No significant differencesere found among the ratings given to the bread samples obtained

rom Sieve, Verna, and Gentil Rosso.Three subject groups with similar preferences were identified

y the k-means cluster analysis performed on consumer liking rat-ngs: Cluster 1 (Cl1: n = 106; 45.5%), Cluster 2 (Cl2: n = 57; 24.5%),nd Cluster 3 (Cl3: n = 70; 30.0%). No significant differences amonglusters were found for demographic variables (Cl1: average age 27,0 males; Cl2: average age 26, 39 males; Cl 3: average age 27, 39ales). On the other hand, the clusters differed significantly for five

ut of the 13 items from the Food Choice Questionnaire (contains aot of vitamins and minerals, P = 0.005; high in fibre and roughage,

= 0.016; contains natural ingredients, P = 0.005; helps me control-

ing my body weight, P = 0.094; keeps me healthy, P = 0.089). Cluster

was characterized by higher ratings for the items related to healthnd natural content than Cluster 2 (Fig. 3). Thus, it appears that

0.1ab 3.4 ± 0.1 5.7 ± 0.1bc4 1.911 4.43001 0.093 0.001

subjects in Cluster 3 pay more attention than consumers in Clus-ter 2 to the relation between the quality of food they consume andtheir health condition. Cluster 1 was characterized by intermediateaverage ratings. No significant differences were observed amongclusters in terms of importance given to the sensory appeal (looksnice; smells nice; tastes good; has a pleasant texture).

The mean liking ratings given by the three clusters of consumersfor bread samples produced with different organic wheat varietiesare shown in Table 5. Cluster 1 significantly (P < 0.0001) preferredthe five bread samples from the old wheat varieties (average ratings∼5) than the control bread, which did not reach the acceptabilitylevel (average value <4). Cluster 2 did not like the tasted samples,independently by the wheat variety used: this group of subjectsrated all samples under the acceptability level and did not dis-criminate among them. Cluster 3 rated all samples highly (averageratings higher than 5) and discriminated among samples better

than the others two clusters. In particular, Cluster 3 significantlypreferred the control and Gambo di Ferro better than the othersbreads (P = 0.001).

P. Migliorini et al. / Europ. J. Agronomy 79 (2016) 131–141 139

F 1 (n =f .05).

4

aonso

melmmayamoipclwie

sto

ig. 3. Average ratings for the items of the Food Choice Questionnaire from Clusteror the same item with differing letters are significantly different (Fisher’s LSD. p < 0

. Discussion

The objectives of this work were to evaluate the agronomic char-cteristics of old, modern and mixture of varieties and landracesf wheat (Triticum aestivum L.) and to assess the technological,utritional and functional properties of grains including the con-umers’ preferences for bread obtained by old and modern varietiesf wheat.

The results of the analysis of agronomic characteristics of old,odern and mixture varieties and landrace of wheat were as

xpected. The old varieties, landrace and their mixture yieldedess than the modern varieties but with higher stability. The non-

odern varieties of bread wheat were more adaptable than theodern one to hilly and mountainous areas of organic farms,

lthough there was a large variability in yield among locations andears for all varieties. This shows that there is sufficient variabilityvailable for rare alleles, which have been eliminated in breedingodern varieties. Bordes et al. (2008) found that the largest ranges

f variation were found in landraces and old cultivars rather thann varieties that are more recent. The lower yield of FARM1 com-ared to FARM2 can be due to the later sowing time and adverselimatic conditions (raining autumn and sever winter). In particu-ar, the modern varieties in FARM1 did not survive the winter and

ere not harvested, while the old varieties reached maturity. Thiss an important result in view of climate change, food security andconomic sustainability.

Average yield of modern varieties in FARM2 was 2.15 t/ha, noto different from the average yield of the 16 modern varieties cul-ivated in FARM 2 (2.27 t/ha) (Perenzin et al., 2011). Moreover,n one hand the tillering of plants is comparable with that of 16

46). Cluster 2 (n = 57) and Cluster 3 (n = 70). Bars represent standard errors. Values

modern varieties (6.06 vs 7.0 respectively), cultivated in the sameFARM2 (Perenzin et al., 2011). On the other hand, the old varietieswere much taller than the 16 modern varieties (113.23 vs 59.0 cm)cultivated in the FARM2 (Perenzin et al., 2011).

Nutritional and functional properties of grains are importantcomponents of grain quality characteristics although mineral den-sities in wheat grains were not a priority of genetic improvements.As a result, grains of modern cultivated wheat cultivars mighthave relatively low mineral densities (Fan et al., 2008; Hussainet al., 2012). Minerals are not only essential for plants but also forhumans and animals. Micronutrient malnutrition affects 1⁄3 − ½of the world population (Miller and Welch, 2013). A major causeof micronutrient malnutrition in low-income populations is lackof access to a variety of foods. When incomes are low, people relyon inexpensive sources of calories, such as cereals and tubers, tomeet energy needs (Miller and Welch, 2013). These foods tend to bepoor sources of many micronutrients. Recommended daily dietaryintake of some minerals is generally not achieved by the humanpopulation in developing countries (Brown et al., 2001; Gibson2006). Therefore, an important objective of the paper was to clar-ify the quality of grains of, and of products from old and modernvarieties.

Piergiovanni (2013) evaluated three Italian bread wheat lan-drace, a mixture of durum and bread wheat landrace and threemodern varieties in north Italy. The high variation of gliadinprofile detected within all the landraces indicates that they

have retained the genetic heterogeneity typical of the old wheatlandraces not released by breeders. These findings agree withFigliuolo et al. (2007) who, using molecular and biochemical mark-ers observed a greater genetic diversity in indigenous landraces

1 J. Agro

pg

td

afHa

6ce2

FrgaarW

tp

1

2

bi

paisaac2iw

5

itsnlnfp

tv

40 P. Migliorini et al. / Europ.

resent in Italy before 1915 than in the material developed andrown by farmers in subsequent periods.

Average hectolitre weigh of old varieties was 74.16 kg/hl, higherhan in the modern varieties cultivated in the same FARM2 (70.0)emonstrating better quality.

The HFN mean value above 350 s suggest an enzyme deficiencynd the amylase activity is considered very low (Suas, 2008). Theermentation will be slow and the bread may have less volume. TheFN can become excessive due to the presence of sprouted grains,ffecting bread making (soft and sticky dough).

In particular grain Zn concentration ranged from 29.9 to4.7 �g−1, with an average of 51 �g−1 showing much higher valueompared to the Pakistan cultivars after green revolution (Hussaint al., 2012) and to 265 Chinese modern genotypes (Yong et al.,010) that resulted respectively in an average of 29 and 32 �g−1.

Grain concentrations of Ca ranged from 220 to 455 �g−1 ande from 31 to 91 �g−1 with an average of 326.5 and 42.83 �g−1

espectively and are in-line with Pakistani cultivars and Chineseenotypes (Hussain et al., 2012; Yong et al., 2010). The level of Fend Zn of this old Italian varieties are near to the desired one (57nd 41 mg kg−1 for Fe and Zn, respectively) based on personal dailyequirement and the information of bioavailability provided by the

HO (2006).A relationship among different grain yield mineral concentra-

ions and uptake capacity of varieties needs precise description. Inarticular, it could depends:

. On genotype characteristics (Verna and Sieve derived both fromEst Mottin but have different “maternal parent” – Mont Calme245 from Switzerland and Bellevie II from France, respectively,that provide high variability in F2 generation);

. On location, soil type and soil fertility (different mineral avail-ability could depend on different local pH, mycorrhiza presenceand different varietals symbiosis interaction); climatic differ-ence among years, that interfere with mineral absorption.

Tracing back the grain yield mineral concentrations is not possi-le sas of now for local wheat cultivars of Italy and further research

s needed.Polyphenols are the most representative antioxidant com-

ounds of the wheat grain. They are mainly made up of phenoliccids and flavonoids. In wheat kernel, polyphenols can be foundn the free soluble and the bound insoluble forms, both possessingtrong antiradical power as well as anticancer, anti-inflammatorynd nutraceutical properties (Dinelli et al., 2013). Both polyphenolsnd flavonoids resulted higher in all fractions (free, total and bound)ompared with durum wheat (Bellato et al., 2013; Dinelli et al.,013) and others Triticum species (Ciccoritti et al., 2013) show-

ng a great nutraceutical potential of this old varieties of commonheat.

. Conclusion

The old wheat varieties and their mixture are agronomicallynteresting especially in adverse climatic years. In FARM1, wherehe adverse weather conditions have affected strongly the autumnowing, old varieties manage in late winter to compete with theatural flora due to their tall plants while modern varieties were

ess resilient as in bad years didn’t developed properly and it wasot possible to harvest them. Andriolo is a very interesting variety

or all aspects (agronomic, quality and sensory) and mixture are

romising for their evolutionary ability.

The results of the consumer test demonstrated an apprecia-ion cheering by consumers of the bread obtained by old wheatarieties; this could open new market opportunities not only in

nomy 79 (2016) 131–141

the production of traditional niche food products but also in thepreparation of novel products naturally enriched with health-beneficial compounds and new chance of on farm survival for theseimportant genetic resources Both agronomic characteristics andconsumer rating, which can be visually scored or easily measured,are amenable of being evaluated by a representative sample ofpotential beneficiaries, with gender equally represented, in par-ticipatory and evolutionary breeding programs, in which trials likethose described in this paper, are planted and evaluated in a numberof farmers’ fields.

The results of this study suggest that the old bread wheat vari-eties and their mixture, assed with participatory and evolutionaryplant breeding, could represent a solution for local communities tocope with climate change mitigation, food security and food quality.

Acknowledgment

This research was partial financed by the Piedmont Regioninside the FRULOGICO Project. We would like to thank Anne Wit-tacher, Luisa Andrenelli and Stefano Benedettelli from Universityof Florence; Patrizia Vaccino and Maurizio Perenzin from Unità diricerca per la selezione dei cereali e la valorizzazione delle varietàvegetali (CREA-SVC) in S.Angelo Lodigiano (LO).

References

AACC International, 2000. Approved Methods of the American Association ofCereal Chemists, tenth ed. American Association of Cereal Chemistry, St. Paul.

Adom, K.K., Sorrells, M.E., Liu, R.H., 2003. Phytochemical profiles and antioxidantactivity of wheat varieties. J. Agric. Food Chem. 51, 7825–7834.

Bellato, S., Ciccoritti, R., Del Frate, V., Sgrulletta, D., Carbone, K., 2013. Influence ofgenotype and environment on the content of 5-n alkylresorcinols: totalphenols and on the antiradical activity of whole durum wheat grains. J. CerealSci. 57, 162–169.

Borasio, E., 1997. Classificazione merceologica del frumento con indici di qualità.Agricoltura 9, 59–61.

Bordes, J., Branlard, G., Oury, F.X., Charmet, G., Balfourier, F., 2008. Agronomiccharacteristics, grain quality and flour rheology of 372 bread wheats in aworldwide core collection. J. Cereal Sci. 48, 569–579.

Brown, K.H., Wuehler, S.E., Peerson, J.M., 2001. The importance of zinc in humannutrition and estimation of the global prevalence of zinc deficiency. Food Nutr.Bull. 22, 113–125.

Camacho, V., Taina, C., Maxted, N., Scholten, M., Ford-Lloyd, B., 2005. Defining andidentifying crop landraces. Plant Genet. Resour. 3, 373–384.

Canerava, M.G., Romani, M., Corbellini, M., Perenzin, M., Borghi, B., 1994.Evolutionary trends in morphological physiological, agronomical andqualitative traits of Triticum aestivum L. cultivars bred in Italy since1900. Eur. J.Agron. 3, 175–185.

Ceccarelli, S., 2009. Evolution, plant breeding and biodiversity. J. Agric. Environ. Int.Dev. 103, 131–145.

Ceccarelli, S., 2014. Drought. In: Jackson, M., Ford-Lloyd, B.V., Parry, M.L. (Eds.),Plant Genetic Resources and Climate Change. CAB International, Boston, pp.221–235.

Ceccarelli, S., Grando, S., Tutwiler, R., Baha, J., Martini, A.M., Salahieh, H., Goodchild,A., Michael, M., 2000. A methodological study on participatory barley breeding.I. Selection phase. Euphytica 111, 91–104.

Ciccoritti, R., Carbone, K., Bellato, S., Pogna, N., Sgrulletta, D., 2013. Content andrelative composition of some phytochemicals in diploid: tetraploid andhexaploid Triticum species with potential nutraceutical properties. J. CerealSci. 57, 200–206.

Dinelli, G., Segura-Carretero, A., Di Silvestro, R., Marotti, I., Arráez-Román, D.,Benedettelli, S., Ghiselli, L., Fernandez-Gutierrez, A., 2011. Profiles of phenoliccompounds in modern and old common wheat varieties determined by liquidchromotography coupled with time-of-flight mass spectrometry. J.Chromatogr. A 1218, 7670–7681.

Dinelli, G., Marotti, I., Di Silvestro, R., Bosi, S., Bregola, V., Accorsi, M., Di Loreto, S.,Benedettelli, S., Ghiselli, L., Catizone, P., 2013. Agronomic: nutritional andnutraceutical aspects of durum wheat (Triticum durum Desf.) cultivars underlow input agricultural management. Ital. J. Agron. 8, 85–93.

FAO, 2013. In Vivo Conservation of Animal Genetic Resources. FAO, AnimalProduction and Health Guidelines, 14, Rome.

FAOSTAT, 2014. http://faostat.fao.org (accessed 05.02.15.).

Fan, M.S., Zhao, F.J., Fairweather-Tait, S.J., Poulton, P.R., Dunham, S.J., McGrath, S.P.,

2008. Evidence of decreasing mineral density in wheat grain over the last 160years. J. Trace Elem. Med. Biol. 22, 315–324.

Figliuolo, G., Mazzeo, M., Greco, I., 2007. Temporal variation of diversity in Italiandurum wheat germplasm. Genet. Resour. Crop. Evol. 54, 615–626.

J. Agro

G

G

G

G

G

H

J

M

M

N

P

P

R

Yong, Z., Qichao, S., Jun, Y., Jianwei, T., Rongrong, Z., Yueqiang, Z., Ortiz-Monasterio,

VV

P. Migliorini et al. / Europ.

enStat Committee, 2011. In: Payne, R.W. (Ed.), The Guide to GenStat® Release 16.Part 2: Statistics. VSN International, Hemel Hempstead.

hiselli, L., Dinelli, G., Marotti, L., Sofi, F., Romagnoli, S., Baglio, A.P., Benedettelli, S.,2008. In: De Bellis, L., Marchiori, S., Miceli, A. (Eds.), Studio Delle QualitàProduttive E Funzionali Di Germoplasma Di Vecchie Varietà Di FrumentoTenero. Atti VIII Convegno Nazionale sulla Biodiversità, Lecce, pp. 211–213.

hiselli, L., Sofi, F., Whittaker, A., Gori, A.M., Casini, A., Abbate, R., Gensini, G.F.,Dinelli, G., Marotti, I., Benedettelli, S., 2013. Effect of pasta consumptionobtained by an old Italian durum wheat variety on cardiovascular parameters:an intervention study. Prog. Nutr. 15, 265–273.

ibson, R.S., 2006. Zinc: the missing link in combating micronutrient malnutritionin developing countries. Proc. Nutr. Soc. 65, 51–60.

uarda, G., Padovan, S., Delogu, G., 2004. Grain yield, nitrogen-use efficiency andbaking quality of old and modern Italian bread-wheat cultivars grown atdifferent nitrogen levels. Eur. J. Agron. 21, 181–192.

ussain, S., Maqsood, M.A., Rengel, Z., Khan, M.K., 2012. Mineral bioavailability ingrains of Pakistani bread wheat declines from old to current cultivars.Euphytica 186 (1), 153–163.

ones, H., Lister, D.L., Bower, M.A., Leigh, F.J., Smith, L.M., Jones, M.K., 2008.Approaches and constraints of using existing landrace material to understandagricultural spread in prehistory. Plant Genet. Resour. 6, 98–112.

inistero delle Politiche Agricole Alimentari e Forestali (MIPAAF), 2013. BollettinoAGRIT. https://www.politicheagricole.it/flex/cm/pages/ServeBLOB.php/L/IT/IDPagina/6557.

iller, B.D., Welch, R.M., 2013. Food system strategies for preventingmicronutrient malnutrition. Food Policy 42, 115–128.

ewton, A.C., Akar, T., Baresel, J.P., Bebeli, P.J., Bettencourt, E., Bladenopoulos, K.V.,Czembor, J.H., Fasoula, D.A., Katsiotis, A., Koutis, K., Koutsika-Sotiriou, M.,Kovacs, G., Larsson, H., De Carvalho, M.A.A.P., Rubiales, D., Russell, J., DosSantos, T.M.M., Vaz Patto, M.C., 2010. Cereal landraces for sustainableagriculture. Rev. Agron. Sustain. Dev. 30, 237–269.

erenzin, M., Gallina, K., Gasparini, A., Notario, T., Sanzone, E., 2011. Le varietà digrano tenero più adatte per il biologico. L’Informatore Agrario 39, 52–57.

iergiovanni, A.R., 2013. Evaluation of genetic variation and grain quality of oldbread wheat varieties introduced in north-western Italian environments.

Genet. Resour. Crop. Evol. 60, 325–333.

ollins, J.A., Drosse, B., Mulki, M.A., Grando, S., Baum, M., Singh, M., Ceccarelli, S.,von Korff, M., 2013. Variation at the vernalisation genes Vrn-H1 and Vrn-H2determines growth and yield stability in barley (Hordeum vulgare) grownunder dryland conditions in Syria. Theor. Appl. Genet. 126, 2803–2824.

iew publication statsiew publication stats

nomy 79 (2016) 131–141 141

Sanchez-Garcia, M., Álvaro, F., Peremarti, A., Martín-Sánchez, J.A., Royo, C., 2015.Changes in bread-making quality attributes of bread wheat varieties cultivatedin Spain during the 20th century. Eur. J. Agron. 63, 79–88.

Singh, M., Malhotra, R.S., Ceccarelli, S., Sarker, A., Grando, S., Erskine, W., 2003.Spatial variability models to improve dryland field trials. Exp. Agric. 39,151–160.

Singleton, V.L., Orthofer, R., Lamuela-Raventos, R.M., 1999. Analysis of totalphenols and other oxidation substrates and antioxidants by means ofFolin–Ciocalteau reagent. Methods Enzymol. 299, 152–178.

Snedecor, G.W., Cochran, W.G., 1980. Statistical Methods, Seventh ed. Iowa StateUniversity Press, Ames.

Steptoe, A., Pollard, T.M., Wardle, J., 1995. Development of a measure of themotives underlying the selection of food: the Food Choice Questionnaire.Appetite 25, 267–284.

Suas, M., 2008. Advanced Bread and Pastry. Delmar Cengage Learning, Clifton Park.Suneson, C.A., 1956. An evolutionary plant breeding method. Agron. J. 48, 188–191.Torri, L., Migliorini, P., Masoero, G., 2013. Sensory test vs: electronic nose and/or

image analysis of whole bread produced with old and modern wheat varietiesadjuvanted by means of the mycorrhizal factor. Food Res. Int. 54, 1400–1408.

UNI ISO 3093:2009 Wheat, rye and their flours, durum wheat and durum wheatsemolina −Determination of the falling number according to Hagberg-Perten.

Van den Broeck, H.V., de Jong, H.C., Salentijn, E.M.J., Dekking, L., Bosh, D., Hamer,R.J., Gilissen, L.J.W.J., van del Meer, I.M., Smulders, M.J.M., 2010. Presence ofceliac disease epitopes in modern and old hexaploid wheat varieties: wheatbreeding may have contributed to increased prevalence of celiac disease.Theor. Appl. Genet. 121, 1527–1539.

World Health Organization, 2006. Guidelines on Food Fortification withMicronutrients. World Health Organization, Geneva.

Yan, W., 2001. GGEbiplot—a windows application for graphical analysis ofmultienvironment trial data and other types of two-way data. Agron. J. 93,1111–1118.

Yan, W., Hunt, L.A., Sheng, Q., Szlavnics, Z., 2000. Cultivar evaluation andmega-environment investigation based on the GGE biplot. Crop Sci. 40,597–605.

I., 2010. Mineral element concentrations in grains of Chinese wheat cultivars.Euphytica 174, 303–313.

Zuppiroli, M., 2013. Prezzi e filiere agroalimentari nello scenario internazionale, in:Origini e garanzie materie prime agricole. Coop Italia, Bologna, 10–18.