Harvest of Southern Highbush Blueberry with a Modiﬁed, Over-the-Row Mechanical Harvester: Use of Soft-Catch Surfaces to Minimize Impact Bruising

: Harvest of fresh market, southern highbush blueberries (SHB) is labor intensive and costly, leading to a demand for alternative harvest methods. Recent research has shown potential for mechanically harvesting blueberries with minimal bruising by using a modiﬁed over-the-row (OTR) harvester. For two harvests, SHB cultivars Optimus and Vireo were either hand-harvested (HH) or mechanically harvested (MH) by two commercial harvesters, one unmodiﬁed with standard hard-catch surfaces (HCS) or by the other modiﬁed with soft-catch surfaces (SCS). For Harvest 1, fruit from all harvest methods were hand-sorted into the following categories: blue fruit (marketable), red fruit, green fruit and culls. Samples from each cultivar and treatment were then held at 24 ◦ C overnight, and the following day ﬁrmness and impact bruise severity were determined. Harvest 2 had identical harvest treatments; however, the fruit were sorted on a commercial packing line prior to packing in clamshells (n = 16). A subsample was held overnight as in Harvest 1 for bruise severity rating, while the remaining fruit were stored at 1 ◦ C for 14 days (d). Although percent marketable fruit was more affected by cultivar than harvest method, determination of bruise severity revealed the beneﬁt of harvest with SCS over HCS for both cultivars. The former had 16–26% severe bruising, compared to 27–40% for the latter; HH had 1–4%. During storage, HH fruit remained ﬁrmer than HCS and SCS, which were similar. After 14 d, ﬁrmness for “Optimus” was 239 N (HH), 157 N (HCS and SCS) and for “Vireo” it was 189 (HH), 155 N (HCS and SCS). Fruit weight loss increased during storage (1.1–4.4%), but there was no difference due to catch plate surface. SSC, TTA and Ratio were not affected by catch surface type or storage period within cultivar. Using modiﬁcations such as “soft” catch surfaces on currently available OTR harvesters reduces impact bruise damage; however, impacts incurred during commercial packing operations can negate this effect. MH for fresh market blueberries may provide an economical alternative for blueberry growers; however, with current technology, the fruit should be utilized within a week of harvest.


Introduction
The United States (U.S.) is the world's largest producer of blueberries, with estimated production in 2019 at 309 million kg, an increase of 21% from the previous year. Production for fresh and processing markets was 169 and 136 million kg, respectively, for a total value of $909 million. U.S. production begins in Florida in the late March market window when prices are high. In 2019, Florida growers harvested 2064 ha and produced 24 million pounds of primarily southern highbush blueberries (SHB) with a value over $62 million [1]. However, increased imports of blueberries during the early spring season has reduced prices for Florida growers and emphasized the importance of increased production efficiency to remain competitive internationally.
Blueberries destined for fresh market require intensive labor during the harvest season, and this cost is a major factor leading to a demand for alternatives to reduce manual labor. For example, labor costs comprise approximately 60% of the annual Florida blueberry production budget and <10% is harvested mechanically for the processing market [2]. According to a survey of U.S. blueberry growers conducted by Gallardo et al. [3], growers used the following factors to determine harvest method: market price (77%), labor availability (63%), labor cost (56%), fruit quality (39%), cultivar characteristics (18%), mechanical harvest cost (17%), infrastructure for fresh fruit (7%), and knowledge of machine harvesting (5%). Most fresh blueberry growers indicated that they would be receptive to adopting new machine harvesting technologies, especially if they were concerned about labor uncertainty or price [4]. Mechanical harvesting (MH) of northern highbush blueberries specifically for fresh market has been investigated since the mid-1990s, beginning with the introduction of the V45 blueberry harvester, which used slanted shaker drums and padded fruit catching surfaces [5,6]. Later, Takeda et al. [7] determined that harvesting southern highbush and rabbiteye blueberries with the V45 harvester improved packout and reduced impact damage compared to commercial, over-the-row (OTR) harvesters available at that time.
Harvesting blueberries with the V45 also reduced ground loss and increased shelf life. Bruising resulted from impacts with hard surfaces of the harvester such as beater rods, catch plates, and conveyer belts.
To characterize the impacts created by mechanical harvesters, Yu et al. [8,9] developed a blueberry impact recording device (BIRD) and found that rotary-type harvesters generated fewer and lower-magnitude impacts than either sway-or slapper-type harvesters. These studies revealed that detached blueberry fruit were more likely to strike the catch plates than the tunnel wall and that maximum impact occurred when the fruit hit the catch plates [8,9]. Takeda et al. [10] achieved a significant reduction in bruise incidence by incorporating soft-catch surfaces onto a semi-mechanical harvesting system. Additional research has shown potential for harvesting berries with minimal bruising by modifying a commercial OTR harvester. In 2017, studies were conducted in Florida [11] and Washington [12] in which a small harvest crew rode on an OTR harvester with various soft-catch surfaces and used handheld shakers to remove fruit as a means to make harvest labor more efficient. These studies indicated that the semi-mechanical harvesting method and capturing detached berries on soft surfaces had potential for small-sized farms.
In addition to mechanical harvester improvements, there has been continued effort to develop SHB varieties that are suitable for machine harvest for fresh market. The U.S. and Canada blueberry industry indicated that the most important breeding traits were fruit firmness, flavor, and shelf life [13]. However, to develop cultivars to withstand machine harvest, the following factors should be considered: bush architecture, harvest timing, loose fruit clusters, easy detachment of ripe fruit, minimal stem retention or scarring, firmness [14][15][16][17], and most importantly, fruit attributes that contribute to minimizing internal bruising from impact force. Several currently grown SHB cultivars, such as Farthing, Meadowlark, Indigocrisp, Keecrisp, and Optimus, have been released through the University of Florida breeding program and are considered suitable for fresh market machine harvest [17,18]. Other studies have focused on improving harvest efficiency. Casamali et al. [19] investigated grafting blueberry cultivars onto Vaccinium arboretum, a species with single-trunk growth habit, with the goal of improving mechanical harvest efficiency by reducing ground loss. Other studies examined fruit detachment force for several SHB cultivars, and green "Farthing" fruit were found to have higher detachment force than blue fruit, which should reduce the number of unripe fruit harvested, thereby increasing packout [20,21]. It was also determined that extending the harvest interval from each 3 d (with manual harvest) to 7 d nearly doubled the amount of soft fruit harvested. Extending the harvest interval also negatively affected fruit quality and shortened shelf life due to the advanced maturity of the fruit [19]. Although there are many challenges to overcome, current efforts are underway by a number of research groups including USDA, University of Florida, University of Georgia, Michigan State University, Oregon State University, and Washington State University. Several commercial manufacturers are developing innovative mechanical harvest systems for fresh-market blueberries. For example, commercial engineering approaches to reduce the interaction between blueberry fruit and the shaking rods is pursued by BIOSK in Serbia with its air jet harvesting concept in which horizontally directed air is used to displace mature blueberries from their stems and detached blueberries land on pneumatic pillows to reduce impact damage. Another concept pursued by A&B Packaging (Lawrence, MI, USA) is reducing the drop height by engaging the harvesting apparatus from above the blueberry plants instead of from the side as previously described [6,7]. Another manufacturer has focused on collecting machinedetached blueberries on soft catching surfaces (Oxbo International, Lynden, WA, USA) in which hard, polycarbonate catch plates are replaced with plates consisting mostly of soft elastomer neoprene sheet and additional soft surfaces cover the horizontal conveyance system that transport fruit to containers. Another leading manufacturer of a mechanical harvest system for fresh market blueberries is FineField, Melderslo, Netherlands.
The objective of this research was to compare fruit quality and bruise severity of SHB cultivars harvested by (a) a commercial OTR harvester with standard, hard-catch surfaces, and (b) a commercial OTR harvester modified with soft-catch surfaces with handharvested fruit.

Plant Material and Field Conditions
In spring of 2019, commercially grown, SHB cultivars Optimus (released in 2017) and Vireo (released in 2009), and selection FL09-311 (unreleased) were harvested in Waldo, Florida. These genotypes were selected because of high fruit firmness and uniform ripening, characteristics which may aid in machine harvesting for fresh fruit. Three-year-old plants were used in this study and were mechanically top-hedged the previous years. The plants were covered with bird netting approximately one week before the first mechanical harvest and left covered between harvests to prevent bird damage and accidental harvest from farm workers. An entire field row of each cultivar was MH using a conventional or modified OTR harvester, while an adjacent row was hand-harvested (HH). Four field lugs of harvested blueberries were randomly selected per treatment for the experiment.

Blueberry Harvesting with OTR Harvesters
Fruit were mechanically harvested with a conventional OTR (Model Korvan 8000; Oxbo International Corporation, Lynden, WA, USA) with standard hard-catch surfaces (HCS) (Figure 1) or with a Model 8040, Oxbo International Corporation, Lynden, WA, USA, to which soft-catch surfaces (SCS) had been installed ( Figure 2). Modifications to the Model 8040 harvester included installation of a prototype soft-catch surface, which consisted of an elastomeric polymer installed into a frame (e.g., hollowed out polycarbonate catch plates) and a soft material mounted on a rectangular catch frame suspended over the conveyor belt and partially overlapping the catch plate surfaces. A video link showing these modifications is available in Supplemental Materials. Previous impact measurement analysis using the BIRD II impact recording device [10] showed that neoprene elastomer sheets generated significantly less impact force during laboratory tests and blueberries dropped on the elastomer sheet had less bruise damage than berries dropped on the polycarbonate catch plate found on conventional OTR harvesters. At each harvest, fruit from the same cultivars were HH from an adjacent row for later comparison with mechanically harvested fruit. tion). Machine settings were selected by the operator to optimize mature fruit removal and minimize removal of green berries in adjacent rows. The chosen OTR settings were used for each cultivar and both harvests. The selected machine settings for this experiment were 0.75 km/h ground speed and head speed of 325 rpm, with rotary shakers positioned for shaker rods of one drum overlapping with rod tips of the other drum by about 2.5 cm. The horizontal displacement of the shaker rods at their tips was ~8 cm. At both harvest dates, the berries were harvested between 9:00 and 11:00 a.m.  Both harvesters had two rotary shakers and harvested blueberries were transferred to the rear of the harvester on horizontal conveyor belts located on either side of the tunnel, where a fan separated leaves and debris and only blueberries dropped into collection lugs. Prior to harvesting test plots, both harvesters were adjusted in adjacent rows by an experienced harvester operator from Oxbo International who also drove the harvester over the rows used for the experiment (B. Foote, Oxbo Intl. Corp., personal communication). Machine settings were selected by the operator to optimize mature fruit removal and minimize removal of green berries in adjacent rows. The chosen OTR settings were used for each cultivar and both harvests. The selected machine settings for this experiment were 0.75 km/h ground speed and head speed of 325 rpm, with rotary shakers positioned for shaker rods of one drum overlapping with rod tips of the other drum by about 2.5 cm. The horizontal displacement of the shaker rods at their tips was~8 cm. At both harvest dates, the berries were harvested between 9:00 and 11:00 a.m.

Sorting, Packing and Storage
Following Harvest 1 (16 April), "Optimus", "Vireo", and "09-311" fruit in field lugs were transported to the Postharvest Horticulture Laboratory at the University of Florida in Gainesville (approximately 27 km) to determine packout data as follows. A 500-g subsample was removed from each lug (n = 4) then sorted into three commercial categories: marketable (blue), unmarketable due to off-color (red or green), and culls (soft and chaff). No storage test was conducted for Harvest 1. Based on poor harvest performance in this harvest, breeding line 09-311 was eliminated from further testing.

Sorting, Packing and Storage
Following Harvest 1 (16 April), "Optimus", "Vireo", and "09-311" fruit in field lugs were transported to the Postharvest Horticulture Laboratory at the University of Florida in Gainesville (approximately 27 km) to determine packout data as follows. A 500-g subsample was removed from each lug (n = 4) then sorted into three commercial categories: marketable (blue), unmarketable due to off-color (red or green), and culls (soft and chaff). No storage test was conducted for Harvest 1. Based on poor harvest performance in this harvest, breeding line 09-311 was eliminated from further testing.
For Harvest 2 (24 April), blueberry cultivars Optimus and Vireo were harvested by hand and mechanically, as in Harvest 1. However, these fruit were handled under commercial conditions, where they were transported to the packing facility at the farm for partial forced-air cooling to 10 °C and held overnight at that temperature in the facility. The following morning, lugs from each treatment (n = 4 lugs) were manually transferred onto a commercial packing line and samples were collected after the fruit were removed at the following points: (1) trash removal via air classifier (small berries, leaves, twigs), (2) inline color sorter (removed off-color green and red fruit) (WECO, Woodside Electronics Corp., Woodland, CA, USA), (3) soft fruit sorter (WECO, Woodside Electronics Corp., Woodland, CA, USA), (4) cull chute (hand-sorted to remove soft, shriveled fruit), and (5) after volume packing into 170 g clamshell containers. Packing line parameters for the color and soft fruit sorters were set by the packinghouse manager, based on his experience with these cultivars.
For both harvests, impact bruise severity was determined as follows, based on Brown et al. [22] and Yu et al. [9]. On the day of harvest, a subsample of marketable blueberries For Harvest 2 (24 April), blueberry cultivars Optimus and Vireo were harvested by hand and mechanically, as in Harvest 1. However, these fruit were handled under commercial conditions, where they were transported to the packing facility at the farm for partial forced-air cooling to 10 • C and held overnight at that temperature in the facility. The following morning, lugs from each treatment (n = 4 lugs) were manually transferred onto a commercial packing line and samples were collected after the fruit were removed at the following points: (1) trash removal via air classifier (small berries, leaves, twigs), (2) inline color sorter (removed off-color green and red fruit) (WECO, Woodside Electronics Corp., Woodland, CA, USA), (3) soft fruit sorter (WECO, Woodside Electronics Corp., Woodland, CA, USA), (4) cull chute (hand-sorted to remove soft, shriveled fruit), and (5) after volume packing into 170 g clamshell containers. Packing line parameters for the color and soft fruit sorters were set by the packinghouse manager, based on his experience with these cultivars.
For both harvests, impact bruise severity was determined as follows, based on Brown et al. [22] and Yu et al. [9]. On the day of harvest, a subsample of marketable blueberries was selected for each cultivar and harvest method, placed into individual clamshell replicates (n = 4 clamshells) (6 oz, H160, Highland Packaging Solutions, Plant City, FL, USA) and weighed, then held overnight at room temperature (24 • C) to allow development of impact bruises. The following day, each clamshell was reweighed to determine weight loss, then 25 fruit were removed from each clamshell replicate for firmness and bruising evaluations. For Harvest 2, fruit samples were taken at two time points after harvest: day of harvest and the following day after fruit were run over the packing line.
Additionally, for Harvest 2 only, a storage experiment was set up to represent commercial storage conditions. Subsamples from each cultivar and harvest method were taken immediately after commercial sorting and packed into 170 g (6 oz) clamshells (H160, Highland Packaging Solutions, Plant City, FL, USA) forced-air cooled to 1 • C, and held at that temperature and 95% relative humidity (RH) under normal atmosphere. Clamshell subsamples were evaluated at 0, 7, and 14 d for weight loss, firmness and impact bruise severity (as above); another subset of these samples was frozen (−30 • C) to minimize enzymatic activity until compositional analyses could be performed at a later date [11].

Firmness and Bruise Severity
Firmness of individual blueberries was measured at the equator using a FirmTech 2 (BioWorks, Inc., Wamego, KS, USA). After firmness measurement, each fruit was sliced through the equatorial plane and images of the cut surfaces were taken for later visual rating of internal bruise severity. Bruise severity was determined by subjectively estimating the area of internal discoloration on the cut surface of 25 fruit halves using a scale from 0% to 100% [9,22]. Fruit in this study were considered marketable when the bruise severity rating was <20%.

Soluble Solids Content (SSC) and Total Titratable Acidity (TTA)
To determine blueberry composition, the frozen samples were thawed, homogenized, and centrifuged at 17,600× g for 20 min. The juice supernatant was filtered through cheesecloth and was used to determine soluble solids content (SSC), total titratable acidity (TTA) and pH. SSC was determined with a digital benchtop refractometer (Model r2i300, Reichert Technologies, Depew, NY, USA) and expressed as • Brix. TTA and pH were determined using an automatic titrimeter (Model 905 Titrando; Metrohm Ion Analysis, Ionenstrasse, Switzerland). Blueberry juice (3 mL) was diluted with 50 mL deionized water and TTA was determined by titration with 0.1 N sodium hydroxide (NaOH) to an end point of pH 8.2. TTA was expressed as percent citric acid basis.

Statistical Analysis
The tests were set up using a Completely Randomized Design and analyzed by cultivar and harvest method (SAS 9.4; SAS Institute Inc., Cary, NC, USA). Means were separated using Tukey's honestly significant difference test at p ≤ 0.05.

Marketable Fruit
There was a significant effect of harvest time on marketable fruit (packout); therefore, results are presented separately for Harvests 1 and 2. For Harvest 1 (sorted in the laboratory), there was no effect of the mechanical harvest method on packout, which ranged from 76% to 92% (Table 1). Breeding line "09-311" had the most unmarketable fruit (23.8%) and, combined with poor field production traits, was eliminated from Harvest 2. For Harvest 2 (sorted on a commercial packing line), "Vireo" had approximately 15% to 20% more marketable blueberries than "Optimus" ( Table 2). "Optimus" had lower packout because there were 3 to 4 times the amount of off-color (red and green) fruit and trash removed during harvest. For both cultivars, soft fruit removed by the automated sorter was similar to that for culls later removed manually by graders (about 5% for each). Packout data were not reported for HH fruit because there was no need for sorting (i.e., 100% considered marketable). Table 1. Packout distribution on day of harvest for "Optimus", "Vireo", and "09-311" blueberries for combined mechanical harvest methods. All fruit from Harvest 1 were manually sorted in the laboratory.

Off-Color Cull
Optimus 89.4 a

Firmness and Internal Bruise Severity at Harvest and after Packing
Initial fruit firmness (after holding for 24 h at 24 • C), was not statistically different due to harvest time, so firmness data for Harvests 1 and 2 were combined for each cultivar. HH blueberries were firmer (overall mean = 231 g/mm) than mechanically harvested fruit (overall mean = 207 g/mm); however, there was no effect of fruit catch surface (Table 3). Table 3. Initial fruit firmness and severe internal bruising for "Optimus" and "Vireo" (Harvests 1 and 2 combined), and for "09-311" (Harvest 1 only), after hand harvest or mechanical harvest onto hard-or soft-catch surfaces. Fruits were sampled prior to commercial packing and held for 24 h at 24 • C. Severe bruise: ≥20% sliced surface discoloration. There was no difference in severe bruising (≥20% of the sliced fruit surface) due to harvest time, so the bruise data for both harvests of "Optimus" and "Vireo" were combined. HH fruit exhibited only 1% to 4% severe bruising whereas in MH fruit it ranged from 16% to 40% (Table 3). Modification of the harvester with SCS significantly reduced severe bruising in both "Optimus" and "Vireo" fruit compared to those harvested with the standard HCS. The beneficial effect of the SCS over HCS was most apparent for "Optimus" (16% vs. 40% severe bruising, respectively); "Vireo" benefitted less but SCS was still effective (26% vs. 40% severe bruising, respectively). Compared to the two cultivars, severe bruising was lower for "FL09-311" (mean of 23%), but there was no difference due to fruit catch surface; laboratory grading revealed that more unripe fruit were harvested in this breeding line than either "Optimus" or "Vireo".

Cultivar
For Harvest 2 (commercially graded), firmness values of cultivars Optimus and Vireo were similar (average 208 g/mm) whether they were sampled before or after going over the commercial packing line and were considered acceptable for fresh market. However, in packed fruit, the percentage of fruit with severe bruising had increased to more than 50% (Table 4), whereas prior to packaging on a commercial packing line, the combined mean bruise severity for machine harvested was about 31% (Table 3). Table 4. Firmness and unmarketable blueberries due to severe internal bruising after commercial packing for "Optimus" and "Vireo" (Harvest 2 only). Fruit held overnight at 10 • C, commercially packed, held 24 h at 24 • C, then measured for firmness and bruise severity (severe bruise: ≥20% sliced surface discoloration).

Weight Loss
From day 7 to day 14, weight loss for fruit for both cultivars roughly doubled, regardless of harvest method (Table 5). There were only two time points where the difference in weight loss for HH was lower than MH: after 7 d, HH "Optimus" lost 1.1% compared to HCS and SCS (mean = 1.55%); after 14 d, HH and SCS "Vireo" were 2.6% and 4.2%, respectively.

Firmness
Immediately after harvest, HH fruit were approximately 25% firmer than MH fruit, regardless of catch surface (Table 6). Although initially firmer, MH "Optimus" softened at a faster rate than "Vireo" during storage.

SSC and TTA
There were minor differences in SSC and TTA due to catch surface or storage period. Overall means for "Optimus" were SSC = 11.7 • Brix, TTA = 0.4% and SSC/TTA = 29.3 SSC/TTA, whereas these respective values for "Vireo" were 11.5 • Brix, 0.8% and 14.4 (data not shown). "Vireo" fruit were more acidic than "Optimus" and therefore had a much lower SSC/TTA ratio.

Discussion
Some of the first mechanical harvesters tested for fresh market blueberries were reported to significantly reduce yield compared to hand-harvested fruit; chief causes for lower yields were excessive green fruit removal, failure to remove ripe fruit, and ground drop losses [22,23]. However, with recent advances in harvester technology and availability of firmer breeding lines, efficiency of mechanically harvested fruit has increased. Current OTR harvesters can achieve acceptable packout when used in conjunction with blueberry cultivars selected for mechanical harvest. Takeda et al. [7] found that some cultivars with firm-textured blueberries picked with a conventional OTR harvester (Korvan 8000) had similar packout (90%) to those HH (92%), especially when the OTR harvester was used on young, short-stature blueberry bushes. These authors also noted that mechanically harvested fruit were softer and had more attached stems than hand-harvested.
In the current study, the packout determined for Harvest 1 was higher than that for Harvest 2 due to the former fruit being sorted in the lab. However, for both harvests, cultivar influenced packout (71% to 81%) more than catch surface. Studying NHB cultivars, DeVetter et al. [12] found differences in packout between "Duke" (84%) and Draper (74%). In contrast, marketable packout was similar for SHB cultivars Meadowlark (88%) and Farthing (91%) [11], in which the decrease in packout was due to a higher percentage of unripe fruit being picked. Harvest of unripe fruit could be mediated by limiting mechanical harvest to only those cultivars in which unripe fruit have higher fruit detachment force than that of blue, fully ripe fruit. Green fruit often require higher detachment force, but this is not the case with every cultivar [20]. When optical sorting technology for internal bruise damage becomes available for commercial packing lines, then machine-harvested blueberries can be more effectively sorted for surface defects as well as for internal damage.
Other researchers, evaluating a handheld shaker device to determine differences in fruit detachment of rabbiteye and SHB types, found that a significant portion of the harvested fruit from both fruit types had stems attached (stemminess), and that up to 23% of SHB fruit were unripe [24]. During mechanical harvest of fruit, detachment does not occur at the fruit-pedicel junction. The resultant stemminess is scored as a quality defect that limits fruit from being graded as U.S. No. 1 [25,26]. Vashisth et al. [27] reported that genes associated with phytohormone metabolism regulated differences in fruit detachment of the genotypes studied. One line of breeding research would be to focus on fruit detachment force such that the force to remove ripe blueberries would be lower than that for less ripe or immature fruit; this could also reduce stemminess.
In the 1990s, the first firm or crisp-textured SHB cultivars (e.g., "Reveille") were released by the North Carolina blueberry breeding program and subsequently by the UF breeding program (e.g., "Bluecrisp"). This trait was described by Ehlenfeldt [28] as fruit having the ability to ripen on the bush while retaining firmness. SHB cultivars with crisp fruit texture described by Padley [29] and Blaker et al. [30] have been shown to facilitate mechanical harvest, reduce postharvest losses, and reduce decay development during storage [7,31].
In the present study, HH blueberries were firmer than MH fruit 24 h after harvest, but with no difference due to fruit catch surface. "Optimus", released by the UF breeding program for MH, was firmer than the other cultivars. DeVetter et al. [12] found that firmness of NHB "Draper" was affected by catch surface when using hand-held shakers; however, firmness was not affected with "Duke". In a subsequent test using a modified OTR harvester, there was no significant difference in fruit firmness due to catch surface (hand, HCS or SCS) although there was a difference between cultivars 'Elliot' and 'Aurora'. In addition, Sargent et al. [11] found that SHB blueberries 'Meadowlark' and 'Farthing' also were firmer after HH compared to those MH but with no significant effect of catch surface.
Previous research determined that conventional OTR harvesters induced excessive damage to the berry skin as well as to internal tissues that reduced blueberry quality and, consequently, research shifted to modifying harvesters to decrease impact force [32,33]. Brown, et al. [22] showed that an experimental blackberry harvester reduced blueberry bruising 10% (from 77% to 68%) compared to commercial harvesters available at that time. Various researchers have reported that blueberry impact damage can be mediated by cushioning hard impact surfaces [5,6,[10][11][12]22]. The present study showed that, although bruise incidence was significantly higher in MH versus HH blueberries, retrofitting the OTR harvester with SCS reduced bruising severity by 30% to 60%, depending on the cultivar. 'Optimus' (recommended for mechanical harvest) developed less bruising than the other cultivars tested, comparable to results from DeVetter et al. [12] and Yang et al. (unpublished data) who reported that bruise severity of NHB 'Aurora' was similar for HH and MH with SCS 24 h after harvest; however, 'Elliott' had more bruising when mechanically harvested with either HCS or SCS that HH. In comparison, there was no effect of catch surface on mechanically harvested SHB cultivars, although bruising was lower for 'Meadowlark' than 'Farthing' [11]. These data confirm that modified catch surfaces and cultivar selection can reduce bruising incidence and severity.
Blueberries are easily bruised during harvest and handling, leading to softening and subsequent decay during storage and shipping; this topic has been extensively examined over the years [34][35][36][37][38]. The present study found that fruit softened during storage regardless of cultivar or harvest method; however, HH fruit remained firmer than MH, regardless of catch surface. Internal bruising was not evaluated during storage since previous studies documented little change in bruise volume over time [10]. Weight loss increased during storage for all harvest methods and there was no effect of catch surface. Blueberry composition was unaffected by harvest method and was considered acceptable. The results of this storage study are similar to those previously reported by Sargent et al. [11], where blueberry quality during storage was more dependent on cultivar than harvest method. Another factor that affects fruit quality is time of harvest; fruit harvested early in the season tended to be more resistant to impact damage and typically had higher acidity resulting in less sweet tasting fruit [21]. More recently, Moggia [33] showed that blueberry firmness at harvest is not a good indicator of fruit quality. Due to lengthened harvest intervals for MH, the presence of softer fruit could be a result of more advanced maturity at harvest favoring increased development of decay during storage.

Conclusions
Blueberry firmness is a common parameter employed to evaluate handling operations and to predict shelf life. However, internal impact bruising is a hidden defect not readily determined on commercial packing lines and does not correlate well with whole fruit firmness. In these tests, 'Optimus' and 'Vireo' fruit harvested with an OTR harvester retrofitted with SCS had minimally acceptable packout and firmness according to commercial grade standards. Both cultivars also had significantly lower severe bruises (21%) compared to fruit picked with an unmodified harvester (40%). It is noteworthy that commercial sorting negated the benefit of SCS as severe impact bruising increased to approximately 50%, irrespective of catch surface. After 7 d at 1 • C MH fruit firmness decreased by 10%, whereas HH fruit did not soften and only by 13% after 14 d. Therefore, it is recommended that mechanically harvested SHB be marketed within one week of harvest to avoid significant losses in quality.
Finally, mechanical harvesting systems for blueberries for the last 60 years had focused on improving harvest efficiency and selectivity. Until recently, harvesting of blueberries for fresh market in the US and elsewhere around the world had been limited solely to manual hand picking. With surging global blueberry production, high demand for fresh blueberries and increasing labor cost, the number of blueberry growers who use machines to harvest fresh market blueberries has risen sharply in the last ten years. Also, the focus of machineharvesting research and engineering advancements has shifted to reducing impact force during harvest, with the result that fruit quality now approaches that of hand-harvested blueberries. Once packed, blueberries are often shipped over long distances to market or held for extended periods in cold storage. Advances in optical sorting systems will permit detection of internally bruised fruit prior to packing by incorporating specific wavelength of light in the near-infrared region and applying algorithms for decision making such that only the fruit that are firm, free from internal bruising and without other quality defects will be packed for shipment to distant markets [39].