This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Firmness is a primary edible quality factor that directly contributes to apple fruit texture. Despite the well-developed understanding of the process of firmness loss in storage, there is very limited information about potential pre-harvest and at-harvest causes of the variation in fruit quality in the marketplace. The objective of the present study was to address the respective roles that the factors of time and intensity of crop thinning, fruit size and fruit dry matter concentration (DMC) each may have in determining fruit firmness of 'Royal Gala' apple at harvest and during storage. Loss of firmness during storage of all thinning treatments and of fruit size and DMC categories was described by a bilinear equation. Time of thinning did not influence any aspect of fruit softening curves during cold air storage at 0.5 Â°C. Comparing the crop loads, a lower crop load (100 fruit per tree) resulted in firmer fruit at harvest. Unlike thinning time, crop load appeared to influence the loss of firmness through storage where the lower crop load softened rapidly during the second slow softening phase. Fruit firmness was positively correlated with fruit size where larger fruit were slightly firmer than smaller fruit at harvest but not after storage. The softening profiles of different sized fruit were similar except for a class of extremely small fruit which appeared to continue softening during the second slow softening phase of storage. Both at-harvest and post-harvest fruit firmness were influenced by fruit DMC in which firmness was increased significantly as DMC increased from 13% to above 16%. Despite having significantly different initial firmness, all fruit with DMC higher than 13% softened at a similar rate during both the initial rapid and second slow softening phases and the transition between the two phases occurred after the same time in storage. In contrast, fruit with very low DMC, less than 13%, had a greater rate of softening in the second phase. These results indicate that variation in fruit firmness at harvest and after storage is influenced by processes that affect and alter fruit DMC during fruit development. In this respect crop load control which is used to improve fruit size, appears as important in altering fruit DMC thereby affecting firmness at harvest and after storage. Furthermore, the effects of DMC on fruit firmness were independent of fruit size.
Keywords: Apples; Fruit size; Dry matter content; Softening, Model fitting
Firmness is a primary edible quality factor that directly contributes to apple fruit texture, where firmer fruit are generally considered to have better quality characteristics than softer fruit (Harker et al., 1997). Many markets are increasingly requiring minimum levels of fruit firmness and soluble solids concentration (SSC) as quality standards for apples, to ensure that consistently high quality fruit can be supplied to the consumers over a long period of time (Johnston et al., 2002a). New Zealand 'Royal Gala' apple fruit must have a minimum of 65N firmness and 12% SSC after storage and on arrival into the market (Anon, 2005). Excessive post-harvest loss of fruit firmness therefore is one of the major problems affecting apple quality (Johnston et al., 2002a) and can contribute to lowering market demand by reducing consumers' acceptance of fruit and their willingness to buy (Harker et al., 2008). Harker et al. (2008) showed that after long term controlled atmosphere storage of 'Royal Gala' apples, consumer acceptability increased as firmness increased from about 30 N to 75 N. The same trends but with slightly different response ranges were evident for the cultivars 'Red Delicious', 'Golden Delicious', 'Fuji', 'Braeburn' and short-term stored 'Royal Gala'.
Some pre-harvest factors are thought to influence fruit quality and texture at harvest and after storage including cultural practices that involve timing and extent of thinning affect fruit quality at harvest (Harker et al., 1997; Johnston et al., 2002a). Fruit of 'Royal Gala' apple from light cropping trees were found to be 10% firmer than those from heavier cropping trees at harvest (Volz et al., 2004). In contrast, Cmelik et al. (2006) did not observe any significant effect of different crop loads on fruit firmness at harvest of 'Fuji' apple. Delong et al. (2006) showed with 'Honeycrisp' apple that as crop load was increased (comparing three, six and nine fruit per cm2 trunk cross-sectional area), fruit firmness after storage decreased, regardless of pre-storage conditioning treatments and storage environment. Interestingly despite many studies having been carried out determining the effects of pre-harvest factors on fruit quality at harvest, Johnston et al. (2002a) highlighted the lack of information on the effect of these factors on fruit softening behaviour during storage.
Fruit size is a trait considered to influence the firmness and post-harvest softening of apples (Johnston, 2002b). Differences in fruit size at harvest typically reflect changes in fruit development induced by timing and intensity of thinning to regulate crop load. It is frequently reported for apple that small fruit are firmer than large fruit both at harvest and after storage ( Marmo et al., 1985; Siddiqui & Bangerth, 1995; Harker et al., 1997). Some studies have attempted to relate fruit firmness with the biophysical properties of different sized fruit. On 'Empire' apple Goffinet et al. (1995) only found a positive correlation of fruit size with cortical cell number but no influence of cell packing or intercellular air space. With 'Royal Gala' apple, Volz et al. (2004) found larger fruit had lower firmness, larger cells, less cell packing and more intercellular airspace than smaller fruit. Despite these differences, cell volume and cell packing showed poor association with fruit firmness for individual fruit from within a crop load and region. Koorey & Brookfield (1999) suggested that the effect of fruit size on fruit firmness arose from differences in physiological maturity of fruit of different sizes, picked on the same date. In another study, fruit size did not affect any aspect of the softening curve of apples stored at 0.5-3 Â°C when fruit were harvested at an early stage of maturity, but when fruits were harvested at a later stage of maturity, smaller fruit tended to soften more slowly than larger fruit (Johnston et al., 2002b).
The use of fruit dry matter concentration (DMC) has recently been suggested as a quality prediction tool in fruit species as diverse as kiwifruit, mango and avocado (Harker et al., 2009) and references therein). With kiwifruit, a meta-analysis of five separate consumer studies on fruit quality demonstrated that fruit with DMC >18% were significantly more liked and that the majority of consumers preferred the flavour of kiwifruit with high SSC when eating ripe, i.e. fruit with high DMC (Harker et al., 2009). The utility of DMC as a quality predictor of taste and flavour in kiwifruit is because measurements can be made before or at harvest and because DMC is principally composed of carbohydrates accumulated by the fruit, with starch being a major constituent which breaks down to sugars during storage and ripening. With 'Royal Gala' apple fruit, predictive measurement of fruit DMC at harvest using near-infrared spectroscopy provided a very strong positive relationship with fruit SSC after storage (McGlone et al., 2003). A similar pattern to kiwifruit was evident in which SSC after storage increased as fruit DMC increased. The high correlation between ripe SSC and fruit DMC for both kiwifruit and apple is not surprising given that fruit of both have a DMC of around 15% with 60-70% of that dry matter in the form of sugars when fully ripe (Palmer, 2007). Consumer acceptance of apple is responsive to quality traits like SSC and titratable acidity but usually only if the firmness of fruit is acceptable (Harker et al., 2008). So, is there evidence for an influence of fruit DMC on textural as well as flavour properties of apple fruit?
There is only sparse and inconclusive literature to implicate the influence of fruit DMC on fruit firmness at harvest or after storage. In a study of the storage of 'Cox Orange Pippin' apples in relation to fruit mineral composition, Johnson et al. (1987) reported that in two out of three years, firmness after storage was positively correlated with fruit DMC at harvest (r = 0.49-0.59), commenting that the accumulation of photosynthates may be important for the formation of firm fruit tissue and merits further investigation. From a survey on post-harvest loss of quality using up to 70 orchards each of four commercial apple cultivars in the Netherlands industry, firmness at harvest was correlated with dry matter concentration of 'Elstar' in one of the two seasons surveyed, whilst no correlations were found with the cultivars 'Jonagold', 'Boskoop' and 'Cox Orange Pippin' (De Jager & De Putter, 1999).
Despite the well-developed understanding of the processes of apple softening in storage, there remains inconclusive knowledge of pre-harvest causes of the variation in firmness in the marketplace (Johnston et al., 2002a). The aim of the present study was to address the question of the respective roles that the pre-harvest factors of time and intensity of crop thinning and the response variables, fruit size and fruit DMC, have in determining fruit firmness and textural properties of apple fruit during storage. A series of crop load treatments were devised that would provide a matrix of fruit development conditions that were believed to influence fruit size distribution and fruit DMC at harvest. These treatments would provide a large number of individual fruit from precisely-known pre-harvest origins to use in extended post-harvest time-course firmness loss studies under low temperature air storage.
2. Material and methods
2.1. Crop load treatments to induce differences in seasonal fruit growth and fruit size at harvest
The study was conducted using fruit in a block of 16-year old 'Imperial Gala' / M.9 apple trees grown at the HortResearch Crosses Road Research Orchard located in Havelock North, Hawke's Bay, New Zealand.
Three crop load (low = 100, medium = 200 and high = 300 fruit per tree) and two thinning time treatments (30 and 60 days after full bloom (dafb)) were applied in a factorial combination together with an additional treatment of completely unthinned trees included as an 'unmodified crop' control. Treatments were set up in a Randomized Complete Block Design of 28 representative and healthy trees selected for uniform flowering. The treatments were set out as single tree plots in four statistical blocks in which the blocks were defined by tree canopy size. An initial chemical-thinning was done using 7.5 ppm Naphthalene Acetic Acid + 0.05% Regulaid at full bloom which was followed by hand-thinning at the various times during the growing season to produce the matrix of crop load Ã- thinning time treatments.
2.2. Harvest and post-harvest quality assessments
Harvest date was decided by assessing starch pattern index (SPI) and fruit background colour (BGC) of the most mature fruit in the block and commenced when such fruit reaching the commercial harvest indicator (SPI 1.0-1.2). All visibly-mature fruit were harvested at several sequential selective picks approximately 6 days apart. From each harvest date, 10 fruit were randomly selected from the harvested fruit of every replicate tree, weighed and assessed for fruit firmness, soluble solid concentration (SSC), dry matter concentration, SPI and titrable acidity (TA). Apart from the 10 fruit for harvest quality assessment, another sample of 100 fruit per tree was selected from the 2nd pick for quality assessment after different durations of storage at 0.5 Â°C in air. Fresh weight of individual fruit prior to storage was recorded for calculation of weight loss during the storage period for recalibration of stored fruit to harvest DMC. Ten fruit per tree were removed from storage initially at two week intervals until 84 days after harvest and thereafter at intervals of 20 and 50 days until reaching 155 days in cold storage. At each storage interval, fruit were equilibrated at 20 ËšC for one day and then assessed for fresh weight, fruit firmness, SSC, dry matter concentration, starch pattern index and titrable acidity.
2.3. Fruit quality measurements
2.3.1 Fruit size
Fruit fresh weight was determined using an electronic balance at harvest and during storage assessments for calculation of fruit fresh weight and weight loss during the storage period for subsequent adjustment of measured storage sample dry matter concentration to 'at-harvest' values.
2.3.2 Fruit Firmness
Fruit firmness was measured using a Fruit Texture Analyser (Güss, model GS14, South Africa) fitted with an 11.1 mm-diameter probe. A small skin area was removed from two opposite sides of each fruit around the equator corresponding to the blushed and shaded sides. The probe penetrated 9 mm into skinless apple flesh and the maximum force (kgf) achieved during tissue penetration was used as the firmness value. The two readings taken on opposing sides of each fruit were averaged to obtain a mean firmness value for each fruit. Data were converted to Newton's force.
2.3.3 Soluble solids concentration
Fruit SSC was determined using a digital refractometer (Atago, model PAL-1, Japan) using juice from the tip of the penetrometer probe released by both firmness measurements of each fruit.
2.3.4 Dry matter concentration
Dry matter concentration was determined using a standardised sampling method of two longitudinal wedges of fruit tissue (not including core material) from opposite sides of each fruit. Sample fresh weight (~25g) was recorded then dried in a dehydrator at 65 ËšC until constant weight and then re-weighed to enable calculation of the percent dry matter of the sample.
2.3.5 Starch pattern index
The SPI was assessed by spraying the cut surface of one half of equatorially-sliced apples (after taking longitudinal wedge samples for dry matter and TA determinations) with an iodine solution (2.5 g l-1 iodine and 10 g l-1 potassium iodide). SPI was scored on a scale of 0 to 6 according to a commercial (ENZAFRUIT New Zealand International) starch pattern index for apple, where 0 (cut surface completely stained) indicates the least and 6 (cut surface with no staining) represented complete conversion of starch to sugars.
2.3.6 Titrable acidity
A pooled sample of two longitudinal wedges from all 10 fruit per tree was used for the determination of fruit titrable acidity. Juice was extracted by placing all longitudinal apple slices into a juicer and the juice sample was then frozen at -20 Â°C until analysis. TA was determined by titrating 5 ml of the juice to an endpoint of pH 8.1 with 0.1 N NaOH using a Meterohm Autotitrator (model 716 DMS Titrino). Titration results were expressed as malic acid equivalents per 100 ml of sample juice using a malic acid standard curve.
2.4. Data analysis
2.4.1 Analysis of variance
Quality assessment data were taken at time intervals from harvest date until the end of storage. Measurements taken over time on the same experimental plot are often correlated and called repeated measures. Therefore a repeated measures analysis of variance was conducted to test the factorial effect of crop load and thinning time treatment plus an unthinned control on measured traits using the Mixed procedure of SAS system (SAS Institute, Inc., Cary, N.C.). Measurements were not taken at regular time intervals; hence a spatial covariance structure was used. All means were separated by least significant difference test determined by Proc mixed models procedure of SAS (Pâ‰¤0.05).
2.4.2 Model fitting of fruit softening patterns
Because of the focus of this study to investigate how factors regulating fruit growth may affect the retention of fruit texture during storage, the following bilinear model (LinBiExp model) was fitted to the time-course firmness loss data by the method of nonlinear least squares using the NLIN procedure of the SAS statistical software (SAS, 9.1) with the estimation of parameter statistics as follows:
A total of five unrestricted parameters are needed to have a completely general bilinear model (LinBiExp): two slopes (Î± and Î²) for the rapid and slow softening phases, the location of the turning point between the two softening phases using a constant (f) for positioning along the FF axis (fruit firmness) and a constant (t) for positioning along the time axis (time in storage), plus a parameter (h) for adjusting the smoothness/abruptness of the transition between the two softening phases (Buchwald, 2006).
Initially to investigate the effect of experimental factors (time of thinning, crop load, the response of fruit from completely unthinned trees) on the change in fruit firmness during storage the model was fitted separately on firmness data from each replicate of all treatments. Since the model (1) did not address the fruit firmness at harvest (t=0), a recalibrated form of the model was written:
where the model parameters are the same as in model (1) except the constant (f) for positioning along the fruit firmness axis which was positioned to represent initial firmness at harvest (f0). The estimated model parameters were statistically compared using SAS proc Mixed procedures.
The study then investigated the effect of the two fruit quality response variables, fruit weight (size) and fruit DMC at harvest, on the softening behaviour of fruit during storage. The combined population of fruit from all treatments on each assessment date was segregated into five categories of fruit size (FS): FS<120 g, 120<FS<150 g, 150<FS<170 g , 170<FS<200 g and fruit greater than 200 grams. To perform comparisons among the five fruit size categories, two versions of the model were required. The full version of the model was compared with reduced versions which were derived from constraining the parameters of the full model (Schabenberger & Pierce, 2002). Sum of square reduction tests were used to compare the full model and reduced models. The analysis commenced with a fit of the full model (3):
where the subscript i = 1, ..., 5 denotes the fruit size category and tij is the jth time in storage at which the firmness of fruit with different fruit size i was observed.
To test whether there were any differences in fruit softening in storage among the five size categories the full model (3) was tested against the most reduced model (2) which represents all data using a single set of parameters. If a difference was found between the full and the most reduced model, the following hypotheses tested each parameter in turn to identify where the differences in responses caused by fruit size occurred as follows:
1) H0: Î±1 = Î±2 = Î±3 = Î±4 = Î±5. If this hypothesis was rejected, the procedure was to test the parameters in pairs.
2) H0: Î²1 = Î²2 = Î²3 = Î²4 = Î²5, the treatments do not differ in the parameter Î² and the value is equal to zero. If the hypothesis was rejected then we tested separately H0: Î²i=0 to find out which fruit size categories differed in the Î² parameter and were not equal to zero.
3) H0: f01 = f02 = f03 = f04 = f05. If this hypothesis was rejected, the procedure was to test the parameters in pairs.
4) H0: Ï„1 = Ï„2 = Ï„3 = Ï„4 = Ï„5. If this hypothesis was rejected the procedure was to test the parameters in pairs.
The main form of the model, (1), was used to compare both firmness at turning point and post-storage firmness (firmness at day 155) between different fruit size categories.
The same combined population of fruit from all treatments on each assessment date were re-segregated, this time into five categories of fruit dry matter concentration (%): DMC <13, 13<DMC<14, 14<DMC<15, 15<DMC<16 and DMC>16. To compare the effect of different dry matter bands on fruit firmness during storage we proceeded similarly with model fitting and analysis as described above for the analysis of fruit size class influences.
3. RESULTS AND DISCUSSION
3.1 Effects of crop load and time of thinning on post-harvest fruit quality traits
The quality characteristics of 'Royal Gala' apple, presented as the trait means from nine storage interval assessments, were significantly influenced by both the timing and intensity of early-season crop thinning, typically used to set tree crop loads for determining fruit size (Tables 1 - 3). Unthinned control trees had fruit numbers 2.5 times greater than the highest crop levels set by crop load treatments which contrast analysis showed was associated with significantly lower fruit firmness, SSC, titrable acidity and DMC as well as smaller fruit size. With Cox's Orange Pippin apple, Johnson (1992) similarly concluded that all thinning treatments significantly increased fruit size, post-storage firmness and DMC when comparing with no thinning. Two aspects of fruit metabolism during storage showed contrasting trends, in which weight loss by fruit from unthinned controls was not different whereas starch hydrolysis was significantly more rapid than in all other crop load treatments. The fruit characteristics from these unthinned trees with a naturally-set very high crop load quantify the genotypic expression of fruit quality traits rarely defined or available to use in interpretation of effects of tree management on storage fruit quality.
The main effect of 'time of thinning' had no influence on post-storage fruit firmness, soluble solid concentration, starch pattern index and fruit DMC but did influence mean fruit weight and titrable acidity. Fruit from trees thinned at 30 dafb had slightly higher titrable acidity than those thinned at 60 dafb (Tables 1 - 3). As expected, fruit fresh weight from treatments thinned at 30dafb were significantly larger than from treatments thinned at 60 dafb typical of reported thinning time responses ( McArtney et al., 1996). In contrast, the main effect of 'crop load' significantly influenced all fruit quality traits with a general response for each trait to quantitatively decline as crop load was increased.
Significant interactions between crop load and time of thinning were found for fruit firmness, SSC and DMC (Tables 1 and 2). Specifically, both fruit firmness and soluble solid concentration were significantly higher for fruit from treatments thinned at 30 dafb to the low crop level (100 fruit per tree) (Table 3). The mean fruit weight and DMC were affected by differences in crop load (Table 2). As number of the fruit per tree increased both fruit weight and DMC decreased. For DMC, the interaction between crop load and time of thinning was significant where fruit from trees thinned at 30 dafb to the low crop level had the highest DMC. The findings are similar to data presented by Delong et al (2006) from 'Honeycrisp' apple in which they found that fruit mass, firmness, SSC and TA increased as the number of fruit per cm2 TSCA decreased regardless of storage environment, or pre-storage treatment. McArtney et al. (1996) also reported a positive relationship between reducing the number of fruit per tree (expressed as fruit number/cm2 TCSA) and mean fruit weight at harvest of 'Royal Gala' apple. In contrast, in a study by Cmelik et al (2006), crop loads did not influence any of fruit quality characteristics such as titratable acidity, soluble solid concentration, firmness and starch index. In the present study, fruit firmness, DMC and SSC all responded to the combination of treatments in the same manner, suggesting close relationships exist between these internal fruit quality traits.
Although treatments induced quite variable quality characteristics, fruit weight loss during storage was completely unaffected by any crop development factors and was extremely stable across all treatment combinations, suggesting that weight loss is primarily a function of storage conditions with little influence of the pre-harvest factors that altered other fruit quality traits.
3.2 Modelling treatment effects on fruit firmness changes during storage.
Initial investigation of the changes in firmness during storage suggested the use of a generalized bilinear model fitted to the time-course storage firmness data. Estimates for all model parameters and predicted final firmness were obtained for every replicate of each treatment and then subjected to ANOVA analysis to compare effects of crop load and time of thinning on fruit firmness changes during storage and final firmness at day 155. Figures 1 and 2 show the fitted curves of the LinBiExp model and observed fruit firmness of 'Royal Gala' apple in storage at given thinning times and crop load levels, respectively.
The initial firmness at harvest (Â¦0, N) was significantly different when comparing unthinned control versus all other treatments where fruit from unthinned control trees were less firm (Tables 4 and 5). Time of thinning had no influence on initial firmness at harvest (Table 4). The higher firmness of fruit at harvest from applying thinning treatments supports Johnson (1992) who showed positive effects of thinning on fruit texture and quality at harvest compared with fruit from unthinned trees. Richardson (1986), also, remarked the adverse effect of productivity on fruit texture. Comparing the crop loads, the lowest crop load (100 fruit per tree) resulted in firmer fruit at harvest (Table 5). The increase in firmness with decreasing crop load agreed with another study of 'Royal Gala' apple (Volz et al., 2004) but not with 'Fuji' (Cmelik et al., 2005). Time of thinning had no significant influence on any aspect of fruit softening behaviour during storage (Table 4, Fig 1). Crop load appeared to influence the firmness at which the turning point between softening phases occurred, but no other aspects of softening behaviour (Table 4 and Fig 2).
The results of contrast analysis showed that fruit from unthinned control trees softened more slowly during the rapid softening phase, although from a lower initial firmness, compared with those from the other treatments. Fruit from trees thinned either at 30 or 60 days after full bloom to the given crop loads softened at similar rates during the rapid phase of softening (Tables 4 and 5).
The time in storage at which the transition between the two phases of fruit softening occurred was similar for all fruit, irrespective of crop load and thinning time treatments (Tables 4 and 5). Despite this, fruit of some treatments differed in firmness at this point of transition which was related to crop load. Fruit from the lowest crop load had significantly higher firmness at the time of transition, which corresponded with having a higher initial firmness (Table 5 and Fig 2).
Fruit from unthinned control trees softened more rapidly in the second phase of firmness decline (Î²) compared with fruit from all other treatments resulting in significantly lower firmness after 155 days of storage (ff) (Table 5 and Fig 2). Among all other thinning time and crop load treatments there were no significant effects on the second phase softening parameter Î² (Tables 4 and 5),so that firmness among crop load treatments at the end of long term storage were very similar(Tables 4 and 5). Studies on 'Braeburn' apple by Mpelasoka et al., (2001) comparing effects of a whole-season deficit irrigation regime with a fully-irrigated control also showed that despite significant difference in fruit firmness between treatments in the first 10 weeks in storage 0 ËšC, fruit of the different treatments reached similar firmness later in the storage period. Together these studies indicate that various treatments imposed that affect fruit growth and development do not closely correlate with all major aspects of softening behaviour of fruit during storage.
3.3 Modelling effects of fruit size on fruit firmness changes during storage.
Estimates for the parameters of the recalibrated LinBiExp model (2) of the firmness loss of 'Royal Gala' apple in storage for the five fruit size classes are presented in Table 6. Figure 3 presents the fitted curves of the LinBiExp model together with measured data points of the changes in fruit firmness of 'Royal Gala' apple during storage for fruit of different sizes.
Hypothesis testing of the parameter estimates revealed that initial firmness at harvest (f0) was the parameter most significantly affected by fruit size (Table 7). For very small sized fruit (class <120 g), the initial firmness was significantly lower than all other size classes. Among the remaining size classes, there was a trend for increasing initial firmness as fruit size became larger, but only within a narrow measured range in initial firmness of between 85 to 90 N. All other softening parameters were not different among the five size classes of fruit (Table 7, Figure 3).
It is commonly thought that firmness is negatively correlated to fruit size (Harker et al. 1997) with larger fruit being softer than smaller fruit both at harvest and after storage (Marmo et al. 1985; Siddiqui & Bangerth 1995). This difference in firmness between different sized fruit has been linked to supposed variation in tissue strength caused by differences in cell size, cell number, intercellular airspace and cell wall material per unit volume of fruit (Volz et al., 2004; Johnston et al., 2002a; Johnston et al., 2002b). In contrast, Koorey & Brookfield (1999) reported that the effect of fruit size on firmness was most likely caused by maturity differences where large fruit picked on the same date as small fruit, had more advanced harvest maturity indices. In the study by Johnston et al (2002b), despite the harvest indices being similar for different sized fruit, large fruit of 'Royal Gala' were slightly softer than both medium and small fruit on the second harvest date, but not so for the cultivar 'Cox's Orange'. In contrast with these studies, the data of present study showed positive correlation of firmness to fruit size and fruit of different sizes had similar SPI at harvest (the selective harvest of fruit was decided by background colour and SPI (Figure 6). Therefore the variation in initial fruit firmness of different size classes could result from physical differences in tissue strength rather than differences in harvest maturity although firmness differences in this study were quite small. These findings add to the inconsistencies reported on fruit size effects on firmness suggesting that the physical differences in cell size and cell number caused by different fruit sizes may not provide an adequate explanation of the variation in firmness at harvest.
Fruit in the different size classes softened at a similar rate in the first softening phase and little or no firmness loss occurred during the final phase (Î²=0) apart from the smallest fruit class (mostly from unthinned control trees) which appeared to continue softening during this phase (Table 7, Figure 3). For all fruit sizes, transition between the two softening phases (Ï„) occurred at a similar time during storage (Table 7, Figure 3).
The main form of LinBiExp model (1) was used to estimate fruit firmness at the transition point between the two phases of softening (parameter f), to determine whether there was any significant difference in transition phase fruit firmness among the five fruit size classes. Firmness was similar for all size classes when transition between the rapid and slow softening phases occurred (Figure 5). To enable parameter f to represent post-storage firmness at day 155 (ff), the Î²1 = Î²2 = Î²3 = Î²4 = Î²5 = 0 constraint was imposed in the model (1) and was used as a full model to test the hypothesis H0: ff1 = ff2 = ff3 = ff4 = ff5. This hypothesis was rejected (P < 0.01). Further analysis showed that fruit smaller than 150 g (<120 and 120-150 g classes) had similar and lower post-storage firmness after 155 days of storage and fruit larger than 150 g (150-170, 170 to 200 and fruit greater than 200 g classes) had similar and higher firmness (Figure 5). Fruit larger than 200 g appeared to soften very slightly faster in the second phase than fruit of the other larger sizes and as a consequence, post- storage firmness at day 155 was not significantly higher for these fruit.
There is very limited information available on the influence of fruit size on softening behaviour of apple fruit through storage. In the study using 'Royal Gala' and Cox's Orange Pippin' apples by Johnston et al (2002b), smaller fruit tended to soften more slowly than larger fruit only when fruit were harvested at a later stage of maturity. They commented that these differences in softening rate were because of the interaction between maturity stage and fruit size and the reason for that is not known. In the present study very small fruit were significantly softer than medium and larger fruit at harvest and despite non significant differences, small sized fruit tended to soften more slowly in the rapid softening phase and more rapidly during the final phase of softening (Figure 3). These differences between two studies could be related to differences in maturity stage because in Johnston's study, differences between different sized fruit were only observed in fruit that were picked later than commercial harvest maturity.
3.4 Modelling effects of fruit dry matter content on fruit firmness changes during storage.
Estimates were obtained for the parameters of the recalibrated LinBiExp model for the firmness loss in storage of 'Royal Gala' fruit sorted into five DMC classes (Table 8). For fruit with DMC less than 13%, all the parameters were significant except for the parameter Î·. For fruit of the other DMC classes, in addition to the parameter Î·, parameter Î² (second stage softening) was also non-significant. The fitted curves and their associated measured data for the five dry matter classes are presented in Figure 6.
To compare the regression curves of different DMC classes, the same model analysis procedures used to compare fruit size classes were followed. The full model (3) in which all parameters varied by DMC, was tested against the most reduced model (2). There was clear evidence that the full model with separate curves fitted these data significantly better i.e. there was significant difference in firmness responses among the five DMC classes (Table 9). The subsequent hypothesis tests showed that the softening curves of fruit of different DMC classes shared common values for the parameters Î±, Ï„ and Î· but the parameters Î² and f0 differed for fruit of different DMC classes. This demonstrated that despite having different initial firmness at harvest, fruit of all DMC classes softened at the same rate during the first rapid phase of softening and that transition to the second slow phase of softening occurred at the same time in storage.
Paired comparisons between DMC classes of the parameters f0 and Î², showed that fruit with DMC less than 14% (classes DMC < 13% and 13< DMC <14%) had similar and significantly lower initial firmness at harvest (f0) than fruit in the higher DMC classes (Table 9). Initial firmness increased significantly with each DMC class as DMC increased from 14% to above 16%. Jager & Putter (1999) have previously reported firmness at harvest to be significantly related to fruit dry matter of 'Elstar' in only one of the two seasons evaluated but not with any of the other cultivars in their study.
Figure 7 shows the starch clearance pattern for fruit of different DMC classes through storage. Fruit from all DMC classes had virtually identical starch clearance pattern at harvest although in storage starch hydrolysis in fruit with higher DMC occurred more slowly thereafter. Therefore, the variation in initial firmness of fruit in different DMC classes cannot be explained by differences in harvest maturity indices.
During the second phase of softening (Î²), fruit with DMC higher than 13% showed almost no further firmness loss whereas fruit with DMC less than 13% continued to soften at a higher rate so that fruit were significantly less firm after 155 days of storage (Table 9, Figure 6).
Following the same procedures used to analyse the effects of fruit size on fruit softening behaviour, the model (1) was used to investigate the effect of DMC on firmness at the transition point between the two softening phases (f) and after 155 days of storage (ff). Differences in fruit DMC resulted in variation in firmness at the transition point and at day 155 (Figure 8). Fruit with DMC less than 15% had similar firmness at transition between the two phases of softening which was significantly lower than fruit in higher DMC classes. As fruit DMC increased from 15% to above 16%, fruit firmness at the transition point increased significantly with each 1% increase in DMC (Figure 8). This showed that fruit with higher firmness at harvest tended to still be firmer at the end of the rapid first phase of softening during storage, supporting the idea that firmer fruit at harvest tend to be firmer during storage.
The model (1) restricted under Î² =Î²2=Î²3= Î²4=Î²5=0 was fitted to derive final firmness at day 155. The resulting model was then used as the full model to investigate the effect of dry matter concentration on firmness at day 155. Fruit with DMC less than 13% were significantly less firm than all other higher DMC classes at day 155 and final firmness was higher with each increase in DMC from 13% to above 16% (Figure 8). These modelled responses show how fruit of low DMC become the least firm during storage through a combination of lower initial firmness, a lower firmness at the transition between phase one and phase two softening and a slightly greater rate of firmness loss during phase two softening compared with fruit from higher DMC classes. The converse shows how fruit of high DMC retained significantly higher firmness during storage, where firmness differences of >15N after 155 days storage correlated with differences of ~3% in fruit DMC.
The effect of fruit dry matter content on the change of firmness in storage has received some attention. Velemis et al. (1997) comparing kiwifruit from 35 orchards grouped in three different DMC classes (14%-15%, 16%-17% and 18%-19%) reported that despite being similar in firmness both at harvest and after seven months of storage, fruit with DMC between 17 and 19% appeared to keep slightly firmer during the first six month of storage than those with lower DMC although the difference was not significant. However no information has been found in the literature regarding the effect of dry matter content on the softening behaviour of apple fruit. Only in the study by Johnson et al. (1992), fruit DMC at harvest provided a positive relationship with post-storage firmness of CA-stored 'Cox's Orange Pippin' in which they suggested that the accumulation of dry matter is critical for the formation of firm apple tissue. Apple fruit dry matter at harvest is largely made up of starch and sugars comprising between 64-80% of the total dry matter whilst on average 14% of the dry matter is associated with the cell walls (hemicelluloses, cellulose, lignin) (Withy et al., 1978; Salo & Korhonen, 1972). Therefore, fruit with high dry matter content could have more cell wall material per unit volume resulting in more dense fruit. Richardson et al. (1997) demonstrated the consistency of the relationship between % DM and fruit density with kiwifruit.
In conclusion, this study demonstrated that crop load, across a commercial range of fruit numbers per tree, and the time at which that crop load was set, had minimal effect on apple fruit firmness at harvest, or on fruit softening patterns in storage. Contrary to many reports, apple fruit size had little influence on fruit firmness at harvest or on softening behaviour in storage, when fruit were harvested at the same maturity. A minor exception was a class of extremely small fruit <120g, largely arising from completely unthinned trees, which was less firm at harvest. In contrast, the DMC of apple fruit was found to greatly influence fruit firmness at harvest as well as the rate of starch hydrolysis in storage. Despite these effects, most of the softening properties of fruit during storage were shown to be the same among fruit with different DMC. Consequently fruit with higher DMC had higher firmness at harvest and remained relatively firmer during storage than fruit of lower DMC. A slight exception occurred with fruit with very low DMC, less than 13%, which lost firmness more rapidly during second phase softening. These results provide strong evidence that the DMC component of fruit is important to the enhancement and regulation of fruit texture and its retention during storage. Further, this study indicates the possibility for fruit DMC to be an important tool to enable apple growers and marketers to more reliably supply fruit with improved texture characteristics than is currently possible.