Evaluation of softening characteristics of fruit from 14 species of Actinidia

Postharvest Biology and Technology 35 (2005) 143–151

Evaluation of softening characteristics of fruitfrom 14 species ofActinidia

Anne White∗, H. Nihal de Silva, Cecilia Requejo-Tapia, F. Roger Harker

The Horticulture and Food Research Institute of New Zealand Ltd., Mt. Albert Research Centre,Private Bag 92169, Auckland, New Zealand

Received 9 March 2004; accepted 21 August 2004

Abstract

Softening of 25 genotypes of kiwifruit, representing 14 species and three families of the genusActinidia, was characterizedduring ripening at 20◦C. Small-diameter flat-tipped probes (2.5 mm and 2.0 mm diameter) were used to measure firmness ofwhole fruit and individual tissue zones in order to ensure that even small-fruited species would be represented in the study.Softening was modelled according to Boltzman function or a simple exponential decay model as appropriate. For three of thegenotypes,Actinidia chinensis‘Hort16A’, A. arguta ‘Hortgem Tahi’ andA. chrysanthafruit that were immature, mature, orover-mature were ripened. While maturity had an impact on the lag before initiation of softening, once fruit had started to softenthe curves were relatively consistent. Softening of tissue zones including the outer pericarp, inner pericarp and core, generallyfollowed the same pattern as that found during whole-fruit firmness measurements. The exceptions were genotypes in the speciesA zones.T particular,f .©

K

1

tit

darddec-areur.

cies.d onnce

0

. chinensis, A. glaucophyllaandA. rufa in which the core failed to soften or softened at a slower rate than other tissuehe study found that the spectrum of softening behaviour was broader than occurs in current commercial cultivars. In

ruit from some small-fruited genotypes tended to remain relatively firm even towards the end of the ripening process2004 Elsevier B.V. All rights reserved.

eywords:Kiwifruit; Firmness; Fruit; Ripening

. Introduction

The commercial success of current kiwifruit, par-icularly ‘Hayward’, was contributed to by the abil-ty of fruit to maintain firmness during storage andransport. Even when coolstored in air, fruit will main-

∗ Corresponding author. Tel.: +64 9 8154200; fax: +64 9 8154201.E-mail address:[email protected] (A. White).

tain their firmness above the minimum export stanfor up to 20 weeks (McDonald, 1990). Breeding andevelopment of new kiwifruit cultivars needs to rognize the importance of creating new fruit thatequally robust in terms of their softening behavioKiwifruit are from the genusActinidia, which is ex-tremely diverse and comprised of about 65 speThe genus is divided into four main sections basemorphological characteristics, primarily the prese

925-5214/$ – see front matter © 2004 Elsevier B.V. All rights reserved.doi:10.1016/j.postharvbio.2004.08.004

144 A. White et al. / Postharvest Biology and Technology 35 (2005) 143–151

or absence of hairs on the stems, leaves and fruit, thetype of hair, if present, and the presence or absenceof spots on the fruit surface (Ferguson, 1990). Overmany years, fruit breeders at HortResearch have builtup an extensive collection ofActinidiagenotypes. Verylittle is known about the softening characteristics ofthese fruit and measurement of firmness is technicallydifficult.

The conventional method of measuring kiwifruit(Actinidia deliciosa) firmness involves pushing an Ef-fegi penetrometer with 7.9 mm diameter probe, 8 mm,into the fruit (Harker et al., 1996). Measurements us-ing such a large diameter probe are not possible forgenotypes that produce small fruit. The process ofdriving flat-tipped cylindrical probes into fruit tissueproduces a combination of compression and shearstresses (Bourne, 1975). The compression force is as-sociated with the cross-sectional area of the probeand the shear force is associated with the perimeterof the probe, which are generally characterised bythe area-dependent coefficient (Ka) and the perimeter-dependent coefficient (Kp), respectively (Bourne, 1966,1975). Thus, firmness values collected using differentdiameter probes are related to each other but the re-lationship is complicated (Volz et al., 2003). In thepresent study, a 2.5 mm diameter probe was used tomeasure firmness of all genotypes so that small-fruitedgenotypes could be readily compared to large-fruitedgenotypes.Actinidia fruit are composed of four tissue zones

( itha nalw pha-sa dif-fc it( s-i Theot amet rm-n houtr -s e sot ingo riort s.

The objective of this study was to determine the vari-ability in softening behaviour for 25Actinidia geno-types and determine if differences are associated withparticular families and species ofActinidia. Soften-ing was determined for whole fruit and for individ-ual tissue zones using cylindrical probes of a size thatcould be utilised for both very small and very largefruit.

2. Materials and methods

2.1. Experimental design

Depending on fruit availability, 80–100 fruit from 25Actinidia Lindl. genotypes were harvested when ma-ture from vines growing at the HortResearch ResearchOrchard in Kerikeri and Te Puke, New Zealand in 2001and 2002. Within two days of harvest a sample of eightfruit was assessed for quality and firmness character-istics at Mt Albert Research Centre, Auckland and theremaining fruit, depending on size, packed into eitherplastic punnets without ventilation holes (small fruit)or commercial cardboard trays with plastic pocketpackand polyliner (large fruit) and held at 20◦C to ripen.Firmness was measured on a number of occasions dur-ing ripening until fruit were deemed over-ripe. In 2002,fruit from threeActinidiagenotypes was harvested onthree occasions to determine the influence of maturityon softening behaviour. Fruit were harvested when im-m 00%d ver-m n ont

2

2ined

a afterc e-t g thefl nerp digi-t ) arep ch tot g ane

skin, outer pericarp, inner pericarp and core) wll but the skin usually being consumed. Conventiohole-fruit measurements of firmness tend to emise changes in texture of the outer pericarp (Jacksonnd Harker, 1997). Firmness measurements of the

erent anatomical parts of tomato (Holt, 1969), cu-umber (Thompson et al., 1982) and strawberry fruOurecky and Bourne, 1968) have been undertaken ung probes ranging in size from 1.0 mm to 9.5 mm.uter and inner pericarp and core of fiveActinidiageno-

ypes have been shown to begin softening at the sime and at the same rate so that differences in fiess between the tissues were maintained througipening (Jackson and Harker, 1997). Firmness meaurements should accurately describe fruit texturhat any potential problems with whole fruit softenr non-uniform tissue softening can be identified p

o commercial development of promising genotype

ature (approximately 80% dark seeds), mature (1ark seeds, fruit not softening on the vine) and oature (100% dark seeds, fruit beginning to softe

he vine).

.2. Fruit quality measurements

.2.1. Fruit sizeFruit weight and tissue dimensions were determ

t harvest. Tissue dimensions were determinedutting fruit in half through the equator. The diamer (two perpendicular measurements representinat and rounded sides of each fruit) of the core, inericarp and outer pericarp was determined using

al callipers. Tissue ratios (core and inner pericarpresented as estimated cumulative % area of ea

he whole fruit in equatorial cross-section assuminllipse for the shape of each sector.

A. White et al. / Postharvest Biology and Technology 35 (2005) 143–151 145

2.2.2. Soluble solids concentrationSoluble solids concentration was determined at har-

vest by squeezing a drop of juice from the fruit equa-tor onto a digital refractometer (model PR-100, Atago,Tokyo).

2.2.3. Dry matterA transverse slice was cut from near the fruit equator

(skin on) from each fruit during the at-harvest assess-ment and dried at 65◦C for 48 h to a constant weight.

2.2.4. Whole fruit puncture measurementsOn each assessment day, eight fruit were randomly

selected and a slice of skin was removed from thefruit prior to conducting the puncture test, which in-volved driving a 2.5 mm diameter flat-tipped probeinto the flesh at a speed of 4 mm s−1 using a mate-rials testing machine (model 4301, Instron, Canton,MA). Force–deformation curves were evaluated foreach measurement and maximum force was recorded.

2.2.5. Tissue zone puncture measurementsA transverse slice (∼5 mm thick) was cut from the

fruit equator and a 2 mm flat-tipped probe driven per-pendicularly into the core, inner pericarp and outer peri-carp until maximum force at a speed of 4 mm s−1 usinga materials testing machine (model 4301, Instron, Can-ton, MA).

2.3. Statistical analysis

e-v yn ware(

3

s arep an-u itht t inA str rom1 so-c thec

3.1. Softening behaviour of Actinidia genotypes

A range of different types of softening behaviourwas observed for the different genotypes (Fig. 1). Thiswas summarised according to properties of the soft-ening curve such as the upper and lower asymptoticvalue and the rate of change of flesh firmness with time.Where observed data clearly indicated an upper asymp-tote, the Boltzman function, which is a reparameterisa-tion of the commonly used logistic equation, was fitted:

FF = FF min + �FF1+e(t−tk )/λ

dFF

dt= −�FF

λ

e(t−tk)/λ

[1 + e(t−tk)/λ

]2

(1)

where FFmin is the lower asymptotic value,�FFthe difference in FF between the upper and the lowerasymptote (this may be large and unreliable due toinsufficient data),t the time after harvest (days),tk thetime in days for firmness to drop by�FF/2 from theupper asymptote, andλ a parameter that describes theshape of the curve between the upper and lower asymp-totes. Note: the latter expression is the rate function.

For remaining genotypes a simple exponential de-cay model was fitted:

FF = FF min + �FFe−kt

dFF

dt= −k �FFe−kt

(2)

where FFmin is the lower asymptotic value,�FF thed mp-t tialcl anyg 1–2d ys af-t um-m ayb

3

be-h rac-t oft olidsc ctedt data

The models utilised for curve-fitting build from prious studies (Benge et al., 2000), and were fitted bon-linear least squares using SAS statistical softSAS, 2000).

. Results and discussion

The harvest characteristics of the 25 genotyperesented inTable 1. Harvest date ranged from late Jary (mid-summer) to mid-June (early winter), w

he majority of genotypes being ready for harvespril/May (autumn). Average fruit weight at harve

anged from 2 g to 134 g, and dry matter ranged f1% to 22%. The relative proportion of the fruit asiated with the core ranged from 1% to 21%, andore plus inner pericarp from 31% to 82%.

ifference in FF between harvest and the lower asyote, andk the relative rate constant. For the exponenurve maximum rate of decay occurs att = 0. Note: theatter expression is the rate function. Since for menotypes the actual FF data were only availableays after harvest the rate estimates are at 2 da

er harvest. In the case of Boltzman the same is sarised by the maximum rate, which occurs halfwetween the upper and lower asymptotes.

.2. Cluster analysis

The parameters used to describe softeningaviour (see above) along with inherent fruit cha

eristics including fresh weight, cumulative % areahe core and inner pericarp, as well as soluble soncentration and dry matter at harvest, were subjeo cluster analysis. Examination of the analysis (


Fig. 1. Softening curves of 25Actinidiagenotypes. Whole fruit firmness was measured using a 2.5 mm diameter flat-tipped probe driven intothe fruit after removal of skin. The Boltzman function or exponential decay was used to model curves as appropriate. Each point is the meanvalue of an eight-fruit sample.

not shown) suggested that clustering associated withrelated genotypes was confounded by genotype size(average fruit weight). The relationship between geno-type, fruit size and softening behaviour was examinedin scatter plots (Fig. 2).

3.3. Softening within different tissue zones

Softening in the three tissue zones tended to fol-low the same pattern as whole fruit softening and theBoltzman function was utilised to fit softening curvesfor each tissue zone where appropriate. The fruit corewas consistently firmer and inner pericarp consistentlysofter until fruit were ripe to over-ripe by which timefirmness was very similar in all tissue zones. This pat-tern of softening is similar to that observed previouslyfor Actinidia genotypes where tissue zones softenedat the same time and rate (Jackson and Harker, 1997).

For two of the genotypes (A. glaucophyllaGD01 andA. rufa RE0204) the core did not appear to softenafter harvest while the inner and outer pericarp didsoften. For genotypesA. chinensis‘Wuzhi 2’ andA.rufa RE0101 the core did soften after harvest butremained significantly firmer than the pericarp tissueeven when fruit were very ripe. An example of the threetypes of core softening is presented inFig. 3.

3.4. Modelling softening of fruit harvested atdifferent maturities

The influence of harvest maturity on the soften-ing pattern of kiwifruit genotypes was investigated forgenotypesA. chinensis‘Hort16A’, A. arguta‘HortgemTahi’ andA. chrysantha(CN01 02) (Fig. 4). Flesh firm-ness tended to follow a sigmoid shape curve, with aslow initial phase followed by a rapid decline and then

A.W

hiteetal./P

ostharve

stBiologyandTechnology35(2005)143–151

Table 1Summary of fruit attributes at harvest for 25Actinidiagenotypes evaluated in 2001 and 2002 ordered alphabetically within section where L =Leiocarpae(Dunn) Li, M =MaculataeDunn and S =StellataeLi.

Section Species Genotype I.D. Harvest date FWta (g) SSCb (%) FFc (N) DMd (%) Cumulative % area

Core Core + inner

L arguta ‘Hortgem Tahi’ 15/3/2001 16.3± 0.70 7.1± 0.55 9.8± 1.46 18.4± 0.67 12.1± 1.24 77.4± 0.65L arguta ‘Hortgem Wha’ 22/3/2001 10.1± 0.32 10.2± 0.81 7.9± 0.88 20.3± 0.75 6.9± 0.98 66.8± 0.81L arguta ‘Hortgem Toru’ 30/1/2002 10.7± 0.58 7.8± 0.50 7.5± 0.91 18.1± 0.20 11.7± 0.67 69.6± 0.75L macrosperma MA01 02 8/3/2001 11.0± 0.30 5.7± 0.09 12.8± 0.99 11.2± 0.31 1.3± 0.06 70.5± 1.85L polygama PC0202 8/3/2001 8.4± 0.55 5.5± 0.10 6.5± 0.48 17.7± 0.28 13.2± 0.74 81.5± 1.29M chrysantha CN01 02 23/4/2001 14.5± 0.65 5.7± 0.13 10.0± 1.54 15.8± 0.22 8.1± 0.32 38.1± 1.28M chrysantha CN03 01 23/4/2001 12.3± 0.78 6.5± 0.18 9.9± 0.34 11.7± 0.16 4.3± 0.3 38.7± 0.94M glaucophylla GD01 7/5/2002 2.6± 0.08 16.3± 0.45 6.8± 0.66 22.4± 1.75 9.8± 0.16 79.4± 2.19M indochinensis IA01 01 15/5/2001 14.9± 0.39 6.3± 0.28 9.0± 0.29 11.4± 0.46 10.2± 0.51 51.6± 0.84M indochinensis IA01 15 5/6/2001 8.6± 0.45 10.2± 0.57 2.3± 0.41 15.6± 0.51 7.6± 0.55 54.8± 2.24M rufa RE0101 15/5/2001 23.7± 1.33 6.7± 0.12 8.7± 0.30 12.7± 0.22 5.5± 0.43 47.5± 2.01M rufa RE0204 24/5/2001 16.8± 0.78 6.1± 0.12 8.7± 0.42 11.9± 0.40 5.1± 0.34 59.0± 0.80S chinensis CK01 03 23/4/2001 48.8± 1.7 9.2± 0.40 8.8± 0.37 16.4± 0.35 4.4± 0.14 47.7± 1.19S chinensis ‘Lushanxiang’ 23/4/2001 101.8± 6.21 5.5± 0.37 9.7± 0.26 13.0± 0.29 4.5± 0.17 42.4± 1.11S chinensis ‘Wuzhi 2’ 2/5/2001 70.4± 7.93 6.6± 0.11 5.8± 0.17 13.2± 0.32 3.1± 0.21 40.8± 0.92S chinensis ‘Jinfeng’ 23/4/2001 133.6± 3.62 5.2± 0.07 9.9± 0.38 14.1± 0.23 4.4± 0.35 31.0± 0.50S chinensis ‘Hort16A’ 13/5/2002 91.1± 3.43 12.5± 1.38 18.3± 0.63 16.7± 0.44 2.7± 0.55 51.7± 12.8S deliciosa DA36 01 5/6/2001 30.1± 1.90 15.5± 0.37 7.9± 0.71 19.9± 0.26 5.7± 0.25 52.7± 0.53S deliciosa ‘Hayward’ 14/6/2001 114.8± 5.74 11.4± 0.35 6.9± 0.92 15.7± 0.37 7.2± 0.39 58.3± 1.04S deliciosa ‘Qinmei’ 5/6/2001 85.1± 4.14 10.9± 0.20 9.3± 0.23 14.8± 0.28 7.9± 0.56 49.5± 2.33S eriantha EA01 01 24/5/2001 36.2± 3.43 8.2± 0.29 4.3± 0.41 15.3± 0.38 10.3± 0.22 82.0± 1.15S fulvicoma FI01 01 18/5/2001 2.8± 0.20 10.2± 0.35 11.5± 0.83 14.4± 0.55 21.1± 1.25 72.0± 1.22S lanceolate LB01 01 18/5/2001 2.0± 0.19 10.6± 0.25 12.2± 1.75 17.6± 0.70 7.7± 0.64 64.8± 1.28S latifolia LC01 13 18/5/2001 2.2± 0.26 8.9± 0.58 5.5± 1.32 17.1± 0.62 8.8± 0.61 71.1± 1.34S setosa SB02-01 20/2/2002 32.6± 1.44 5.1± 0.14 6.2± 0.54 11.1± 0.28 5.3± 0.48 68.8± 2.88

Values are average of an eight-fruit sample± standard errora Fruit weight.b Soluble solids concentration.c Whole fruit firmness.d
Dry matter content.
147


Fig. 2. Estimates for parameters used to characterise softening curves of 25 genotypes ofActinidia. (Refer to Eqs.(1) and(2)).

a slower phase towards a lower asymptote (Fig. 4). Thedata did not always show the full range of the softeningcurves so that initial plots provided little evidence of thepresence of a lower asymptote, i.e. fruit ofA. chinen-sis ‘Hort16A’ andA. chrysantha(CN01 02) harvestedwhen immature. Hence, in formulating a model, flesh

Table 2Parameter estimates and their S.E. for the Boltzman model whereα is the upper asymptotic value of FF (N),tk is the time in days when FF =α/2andλ is a parameter that describes the width of the curve along the time axis

Parameter GenotypeA. chinensis‘Hort16A’ A. arguta‘Hortgem Tahi’ A. chrysantha(CN01 02)

α (N) 21 ± 0.8 10± 0.5 19± 0.9t1 (days)—immature 29± 1.4 13± 0.4 15± 0.6t2 (days)—mature 22± 1.4 2± 0.4 7± 0.6t3 (days)—overmature 9± 1.1 3± 0.3 4± 0.6λ (days) 4± 0.7 1± 0.2 2± 0.4

firmness was assumed to finally drop down to zero overtime. The objective here was to use an empirical modelto describe and predict changes in firmness at differ-ent maturity levels. As with previous data, the Boltz-man function was found to best describe the changesin fruit firmness (FF). The Boltzman function is now


Fig. 3. Firmness of core (�), inner (+) and outer pericarp (�) andwhole fruit (©) during ripening at 20◦C for fruit of A. chinensis(CK01 03) (A), A. chinensis‘Wuzhi 2’ (B) andA. rufa (RE0204)(C). Tissue zone firmness was measured using a 2 mm diameter flat-tipped probe driven perpendicularly into a transverse fruit slice andwhole fruit firmness measured using a 2.5 mm diameter flat tippedprobe driven into the fruit after removal of skin. The Boltzman func-tion was used to model curves where appropriate.

of the form:

FF = �FF

1 + e(t−tk)/λ, (3)

as described in Eq.(1). The parameter,λ, is assumedto be invariant of harvest maturity.

The model fitted reasonably well and accounts forthe maturity effect by an offset parameter on the timeaxis (Table 2). This means a family of curves couldbe generated for different maturity levels. The modelimplies that the difference in offset parameter (tk) be-tween two harvests measures the equivalent number ofdays it takes for FF to reach the value at the time ofthe second harvest if the fruit had been held at 20◦C

Fig. 4. Influence of maturity on fruit softening at 20◦C ofA. chrysan-tha (CN01 02) (A), A. arguta‘Hortgem Tahi’ (B) andA. chinensis‘Hort16A’ (C). Whole fruit firmness was measured using a 2.5 mmdiameter flat-tipped probe driven into the fruit after removal of skin.The Boltzman function was used to model curves.

from the first harvest. This is potentially a useful way tointerpret the differences in parameter estimates usingknown harvest dates.

Examination ofFig. 4 and Table 2suggests thatwhile harvest maturity has a considerable impact onsoftening behaviour, there are characteristics that seemgenotype specific. For example, the onset of softeningis delayed in immature fruit from all three genotypes,but the maximum rate of softening andλ seem rela-tively well conserved at all maturities.

3.5. Interpretation of softening behaviour amonggenotypes

Various characteristics of the softening curves seemto be differentially affected by genotype× harvest ma-


turity. The data presented inFig. 4suggests that the lagbefore softening is initiated and correspondingly thetime taken to become fully ripe are strongly influencedby harvest maturity. However, the maximum rate ofsoftening andλ seem to be genotype specific and lessinfluenced by harvest maturity. Furthermore, it is an-ticipated that minimum firmness, when it is identifiedas a lower asymptote, should be genotype specific (i.e.Fig. 2b). These observations provide some reassurancethat generalised interpretation of differences in soften-ing behaviour among clusters of genotypes has somevalidity. Moreover, they suggest that measurements oftime to reach eating ripe are not appropriate measuresfor comparing genotypes.

There was a broad scatter of initial harvest firm-ness across all genotypes (Fig. 2a). However, it ispossible that firmness will change during fruit devel-opment as occurs for apple (Volz et al., 2003), andtherefore firmness may be greatly influenced by har-vest date. Similarly, there were few identifiable trendsamong genotypes for the change in firmness (Fig. 2c)or maximum rate of softening (Fig. 2d). The excep-tion was genotypes from the family Maculatae (in-cludes speciesA. chrysantha, A. indochinensis, A.rufa), which seemed to lack individual genotypes withhigh rates and large magnitudes of softening as exhib-ited among genotypes from other families (seeFig. 2cand d).

A more obvious trend was found in the relation-ships between fruit size, genotype and final firmness( pes(a m-n val-up e int ruitf N,o thecn ands ciesc dary.WA -t ingA -f e

more important than family in determining likely soft-ening behaviours.

Information on the softening behaviour ofActini-dia species is important for fruit breeding/product de-velopment. Softening characteristics and the way fruithandles through the distribution chain are importantfactors that influence adoption of new products (e.g.Reid and Buisson, 2001). Furthermore, softening be-haviour of fruit impacts on three of the five aspects ofconvenience that have been identified as being relevantto consumption of fruit and vegetables (Jaeger, 2003):how messy the fruit is to eat, how the fruit handles inthe home, and the length of time that it is available inthe shops.

The present study suggests that the softening be-haviours ofActinidia fruit extends beyond those thatmight be expected according to commercial experiencewith cultivars ‘Hayward’ and ‘Hort16A’. Maturity atharvest appears to influence the lag before initiationof softening more than any other softening parameter.While the outer and inner pericarp tissues softened ina similar temporal pattern, in some genotypes the corefailed to soften or softened at a slower rate. The dif-ferences in softening of fruit from different genotypeswere largely associated with the final firmness when‘fully ripe’. Fruit from large-fruited species tended tobe softer than fruit from small-fruited species in thisrespect.

A

atac ilipM har-v da-t tractC

R

B piri-ruit.

B nents

B heartud.

Fig. 2b). It was notable that large-fruited genoty>25 g), including speciesdeliciosa, chinensis, setosanderiantha, tended to be less variable in final firess. For these fruit, firmness tended to decline toes below 0.4 N during ripening (Fig. 2b). By com-arison, small-fruited genotypes were very variabl

erms of the final firmness. While the firmness of from all A. argutagenotypes declined below 0.5ther small-fruited genotypes were much firmer atompletion of the ripening process (Fig. 2b). Unfortu-ately, the split between the group of large-fruitedmall-fruited genotypes occurs abruptly, and no speontributed genotypes that crossed the size bounhile the large-fruited genotypes (fromA. chinensis,. deliciosa, A. setosa, andA. eriantha) belong to sec

ion Stellatae, other members of this section (includ. latifolia,A. lanceolata, andA. fulvicoma) are small

ruited. This suggests that the fruit size× genotype ar

cknowledgements

Linda Boyd and Mary Petley are thanked for dollection and Ross Ferguson, Mark McNeilage, Phartin and Thomas Patterson for sourcing and

esting fruit. This work was funded by the Founion for Research Science and Technology (Con06X0213).

eferences

enge, J.R., De Silva, H.N., Banks, N.H., Jeffery, P.B., 2000. Emcal modelling of postharvest changes in the firmness of kiwifPostharvest Biol. Technol. 19 (3), 211–220.

ourne, M.C., 1966. Measure of shear and compression compoof puncture tests. J. Food Sci. 31, 282–291.

ourne, M.C., 1975. Method of obtaining compression and scoefficients of foods using cylindrical punches. J. Texture S5, 459–469.


Ferguson, A.R., 1990. Kiwifruit (Actinidia). In: Moore, J.N., Balling-ton Jr., J.R. (Eds.), Genetic Resources of Temperate Fruit and NutCrops (Acta Horticulturae 290). Int. Soc. Hortic. Sci., Wagenin-gen, pp. 603–653.

Harker, F.R., Maindonald, J.H., Jackson, P.J., 1996. Penetrometermeasurement of apple and kiwifruit firmness: operator and in-strument differences. J. Am. Soc. Hortic. Sci. 121, 927–936.

Holt, C.B., 1969. Measurement of tomato firmness with a universaltesting machine. J. Texture Stud. 1, 491–501.

Jackson, P.J., Harker, F.R., 1997. Changes in firmness of the outerpericarp, inner pericarp, and core ofActinidia species duringripening. NZ J. Crop Hortic. Sci. 25, 185–189.

Jaeger, S.R., 2003. Innovation in the fruit industry: need for conve-nience. Food Aus. 55 (4), 129–132.

McDonald, B., 1990. Precooling, storage, and transport of kiwifruit.In: Warrington, I.J., Weston, G.C. (Eds.), Kiwifruit Science and

Management. NZ Soc. Hortic. Sci. Inc., Wellington, pp. 429–459.

Ourecky, D.K., Bourne, M.C., 1968. Measurement of strawberrytexture with an Instron machine. J. Am. Soc. Hortic. Sci. 93,317–325.

Reid, M., Buisson, D., 2001. Factors influencing adoption of newapple and pear varieties in Europe and the UK. Int. J. RetailDistrib. 29 (6), 315–327.

SAS Institute Inc., SAS OnlineDoc®, Version 8. SAS Institute Inc.,Cary, NC, 2000.

Thompson, R.L., Fleming, H.P., Hamann, D.D., Monroe, R.J., 1982.Method for determination of firmness in cucumber slices. J. Tex-ture Stud. 13, 311–324.

Volz, R.K., Harker, F.R., Lang, S., 2003. Firmness decline in ‘Gala’apple during fruit development. J. Am. Soc. Hortic. Sci. 128,797–802.

Documents

Evaluation of softening characteristics of fruit from 14 species of Actinidia