Effects of year and plant spacing (0.75 or 1.5 m apart within row) (A) or training method (August or new-over-old training of primocanes) (B) on yield per plant in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. Error bars represent standard error.
Fig. 2.
Interactive effects of year, within-row plant spacing, and primocane training method on yield per hectare in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. Error bars represent standard error.
Fig. 3.
Interactive effects of year and primocane training method on the number of fruit per lateral on plants spaced 0.75 m (A) or 1.5 m (B) apart within the rows in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA.
Fig. 4.
Effect of training method (August or new-over-old training of primocanes) in 2021 (A), 2022 (B), 2023 (C), and 2024 (D) on the percentage of total yield on each harvest date in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. *, **, *** Significant at P ≤ 0.05, 0.01, and 0.001, respectively.
Plant Growth, Yield, Machine-harvest Efficiency, Fruit Quality, and Labor Costs in ‘Columbia Star’ Trailing Blackberry: Risks and Advantages of Closer In-row Spacing, Primocane Suppression, and “New-over-old” Primocane Training
Authors:
,
,
,
, and
Click on author name to view affiliation information
‘Columbia Star’ is a high-quality, thornless, trailing blackberry (Rubus L. subgenus Rubus Watson) grown for the machine-harvested processed market in Oregon and Washington, USA. Because it is a relatively new cultivar, little is known about how to manage it best. Therefore, a new planting was established in Fall 2019 to study the impact of plant spacing (0.75 vs. 1.5 m), chemical primocane suppression (with or without), and primocane training [August training vs. “new-over-old” (unpruned)] on its growth, yield, fruit quality, machine-harvest efficiency, and labor requirements. In each treatment, primocanes were trained to a vertical two-wire trellis system the first year after planting. By the second year and each year afterward, a chemical suppressant was sprayed on the primocanes when they reached ∼0.3 m in length in late April or early May, and the planting was machine-harvested on four to six dates from late June to late July. When averaged across plant spacings, the combination of August training and primocane suppression resulted in more primocanes (15 canes/plant) than any combination of new-over-old training and/or no primocane suppression (averaged 12 canes/plant). New-over-old training, on the other hand, increased the number of fruit per lateral, depending on the year and plant spacing, and resulted in higher yields than August training in the final 2 years of the study (2023 and 2024). The plants also produced a higher yield when they were spaced at 0.75 m than at 1.5 m, but only during the first 2 years of fruit production. Primocane suppression had no effect on yield in any year, likely because it resulted in more canes per plant, which compensated for shorter cane lengths. Any treatment differences in fruit weight were small and were <0.2 g/fruit. However, total soluble solids in the fruit were nearly 1% less on average with new-over-old training, perhaps because a portion of the fruit were knocked off earlier during harvest than with August training. In terms of labor, new-over-old training required 87% less time than August training when the plants were spaced at 0.75 m, and 79% less time when they were spaced at 1.5 m. Primocane suppression reduced the time by an average of 7% and 13% with new-over-old and August training, respectively. After the first 3 years of fruit production, new-over-old training resulted in a cumulative net return of $52,875 and $49,006/ha when the plants were spaced at 0.75 and 1.5 m, respectively, when averaged across treatments with and without primocane suppression. August training, in comparison, resulted in a loss of $11,606/ha at 0.75 m and a return of only $7218/ha at 1.5 m. Across plant spacings, primocane suppression increased cumulative net returns by $7217/ha when the primocanes were trained in August, but only by $1067/ha when the primocanes were trained using the new-over-old method. Given economic and labor constraints and positive impact on yields, the new-over-old method is expected to become the industry standard very quickly.
The northwestern United States is a leading production region for trailing blackberry (Rubus L. subgenus Rubus Watson), which is machine-harvested for the processed market (Finn and Strik 2016). Many cultivars are widely grown for this purpose, including ‘Columbia Star’, which was released in 2013 (Finn et al. 2014). It accounted for 38% of the blackberry plants sold in Oregon (26% more than the next highest cultivar) and 32% of total blackberry sales in the Pacific Northwest (Oregon and Washington, USA; and British Columbia) in 2024 (Northwest Berry Foundation 2025). ‘Columbia Star’ also accounted for 20% to 25% of total blackberry production in Oregon in 2022–23 (Oregon Raspberry and Blackberry Commission 2024). This newer cultivar has a very different growth habit than two other predominant trailing blackberry cultivars grown in Oregon, USA: ‘Marion’ and ‘Black Diamond’. Specifically, ‘Columbia Star’ is thornless and produces an abundant number of canes that are intermediate in length compared with ‘Marion’ (longer) and ‘Black Diamond’ (shorter), and it is also more cold hardy (Finn et al. 2014).
Trailing blackberries are characterized as perennials with a biennial growth cycle of cane growth (Strik 2017). Canes growing in the first year of the cycle are called primocanes. They are solely vegetative and are not self-supporting—meaning, they grow along the ground when they reach ∼1 m in length. After winter dormancy, the canes become what are referred to as floricanes, as they produce fruiting laterals that flower in the spring and fruit in the summer. After fruiting, the floricanes senesce, thus completing the life cycle. Although plantings can be managed to have only primocanes or floricanes in a given season (called alternate-year production), typically both types of canes are present simultaneously to maintain stable production every year. In this case, primocanes must be trained before harvest so they are not damaged by equipment, including the over-the-row harvesters used to pick fruit.
The annual process of pruning out dead floricanes (“caning out”) and training primocanes onto a trellis accounts for the largest share of production costs in machine-harvested trailing blackberry fields (Delbridge et al. 2024). Therefore, leaving dead floricanes on the trellis each year (no caning out) reduces these costs, as the new primocanes may then be thrown over the old, dead canes, and may also reduce cane breakage during training. This process is repeated for multiple years, although it is not known whether there is an optimal number of years to train canes this way before a decline in yield, fruit quality, or plant health is observed. This approach, referred to as a new-over-old system, produced higher yields but more harvest contaminants, especially thorns, than conventional pruning and training in ‘Marion’ blackberry. Therefore, it was not recommended for that cultivar (Strik 2018). ‘Columbia Star’ is thornless, but pieces of dead cane and any insects or other debris might still be a contaminant when machine harvesting. Growers have begun trying the new-over-old method in trailing blackberry to save production costs, but no studies have been performed on this technique in cultivars other than ‘Marion’ to evaluate yield, plant growth, or longer term effects on productivity.
Another common cane management technique is primocane suppression, or primocane burning, during which primocanes are removed chemically or mechanically early in the growing season (Becerra-Alvarez 2025; Norton 1980). This practice is used to increase yield and machine-harvest efficiency (MHE), and manage weeds, and may increase cold hardiness of the subsequent canes (Bell et al. 1995a, 1995b; Cortell and Strik 1997a, 1997b). Plants produce another flush of primocanes after treatment (Strik 2017). In ‘Marion’ blackberry, primocane suppression weakened plants over time, especially those that had lower plant vigor, and reduced the number of primocanes per plant and thus subsequent yield. However, in ‘Boysen’ (Rubus ursinus hybrid), primocane suppression early during the growing season increased the number of primocanes per plant and did not affect yield or fruit size (Stanley et al. 1999). In red raspberry (Rubus idaeus L.), suppression either increased yield as a result of competition for resources between floricanes and primocanes (Crandall et al. 1980; Wright and Waister 1982) or had no effect on productivity (Miller et al. 2008). Because ‘Columbia Star’ is a relatively new cultivar, there have been no studies on these various primocane management techniques and their short- or long-term impact on fruitfulness or growth.
Trailing blackberries are typically grown at in-row spacings between 0.9 and 1.8 m (Strik et al. 2007). Planting density is known to affect growth and yield in many fruit crops (Menzel and Le Lagadec 2014; Sansavini et al. 2008, Wertheim et al. 2001), and in blackberry has been shown to impact cane number and length, as well as percent budbreak, yield, and berry size (Strik 2018). When ‘Marion’ blackberry was planted at different in-row spacings (0.6, 0.9, and 1.5 m), plant density had inconsistent results on yield per acre across three seasons (Strik 2018). Because ‘Columbia Star’ has a different growth habit than ‘Marion’ and other commonly planted trailing blackberries in Oregon, USA, growers are interested to know how increasing density from 1.5 to 0.75 m will affect yield and management costs, considering that establishment costs would be higher.
The objective of our study was to determine how plant spacing and new cane management techniques affect plant growth, yield, MHE, fruit quality, and labor costs in ‘Columbia Star’ trailing blackberry. The results will be used to optimize the best management practices for this cultivar and to provide growers with information on the advantages and any risks associated with the “new-over-old” alternative system, primocane suppression, and closer in-row spacing.
Materials and Methods
A 0.16-ha field of ‘Columbia Star’ trailing blackberry was established Oct 2019 at the North Willamette Research and Extension Center (lat. 45°16′47″N, long. 122°45′23″W), Aurora, OR, USA. One year before planting, soil was tested for pH and nutrient concentrations. To increase soil pH, lime was applied and incorporated at a rate of 4483 kg·ha–1. No other nutrients were found to be deficient based on recommended ranges for blackberry (Davis et al. 2024). In 2019, an annual rye (Lolium multiflorum) cover crop was planted in the spring and mowed in early summer; ground preparation took place in late summer. Planting of standard tissue-cultured plug plants occurred 27 Sep 2019. Approximately one third of the transplants died from cold temperatures during the winter and were replaced with new plugs in Apr 2020. A trellis was constructed in Summer 2020 that consisted of three wires, one at 0.3 m above the ground, to which the drip irrigation line was clipped; a center wire at 1.3 m; and a top wire at 1.8 m above the ground. Plants were irrigated and fertilized using a single line of drip tubing with 1.9 L·h–1 inline emitters every 0.6 m. Fertigation occurred weekly each year, starting in late March and continuing through early July, at a rate of 5.6 to 11.2 kg·ha–1 nitrogen per week of calcium ammonium nitrate in 2020 (17N–0P–0K, 8.8% calcium), ammonium nitrate in 2021 (20N–0P–0K), or urea ammonium nitrate in 2022 through 2024 (32N–0P–0K). Composite soil and leaf tissue samples were collected and submitted to a testing laboratory (Brookside Laboratories, New Bremen, OH, USA) each year according to recommended sampling methods (Davis et al. 2024), and the results were used to adjust nutrient applications to the planting. A total of 1.1 kg·ha–1 boron (Solubor; U.S. Borax, Chicago, IL, USA) was applied to canes and foliage 5 Apr 2021, 31 Mar 2022, and 1 Apr 2024; and 90 kg·ha–1 potassium oxide was applied 17 Mar 2022 as a granular application of potassium chloride (0N–0P–60K).
Eight treatments were arranged in a split–split plot design with two in-row plant spacings (0.75 and 1.5 m) as main plots (whole rows), two training methods (August and new-over-old) as subplots (half rows), and two levels of primocane suppression (with or without) as sub-subplots, in four replications. There were four replicates of each treatment, for a total of 32 plots. Each sub-subplot was 6.1 m long and was separated by 3 m of unplanted space to allow for clearing of fruit from the machine harvester sorting belt between adjacent treatment plots. Rows were spaced 3 m apart and aisles were maintained using tillage (planting year only), typical herbicide spray practices (Becerra-Alvarez 2025), and hand-hoeing as needed. The planting was flanked by guard rows and included five rows in the center of the field used for a separate research trial (Carroll et al. 2024).
Treatments included in-row spacing (0.75 or 1.5 m between plants), training method (August training, in which floricanes were removed in mid- to late-August after senescence and new primocanes were trained onto the trellis wires; or new-over-old training, in which floricanes were not removed and new primocanes grew through or were thrown up and over the older canopy), and the presence or absence of primocane suppression. For primocane suppression, one application of carfentrazone (Aim EC; FMC Corp, Philadelphia, PA, USA) at a rate of 0.47 L·ha–1 was applied with a backpack sprayer and was directed to the bottom 0.5 m of the plant canopy when the first seasonal canes reached ∼0.3 m long. During the first 2 years of the study, primocanes were measured in the spring before suppression, and there was no significant difference in primocane length at that time (data not shown), so all plots receiving primocane suppression were sprayed on the same day.
Data collection.
Primocanes were counted on two plants per plot, and the length of two primocanes per plot was measured in August each year. Five fruiting laterals per plot were selected from senesced floricanes each year and were measured for length and number of fruiting sites per lateral.
Fruit were picked on four to six harvest dates between 24 Jun and 25 Jul using an over-the-row Littau Harvester (Stayton, OR, USA) in 2021 and 2022, and an Oxbo 7450 harvester (Lynden, WA, USA) in 2023 and 2024. During harvest, unmarketable fruit (underripe, overripe, moldy, or damaged) were sorted out and weighed separately to determine the amount of culled fruit per plot. The remaining fruit were loaded automatically into plastic harvest lugs and weighed to determine the marketable yield of each plot. Any fruit that fell to the ground during harvest was also collected and weighed from a 1 m section of each plot (on both the east and west sides of the canopy). Total fruit production was calculated as the sum of marketable, culled, and dropped fruit. The percentage of dropped fruit [(Drop weight/Total fruit production) × 100], percentage of culled fruit [(Cull weight/Total fruit production) × 100], and the MHE (Marketable yield/Total fruit production) × 100] were calculated.
A random subsample of 25 fruit was taken from each plot to determine the average weight on each harvest date. After the subsamples were weighed in 2023 and 2024, the fruit were crushed and homogenized by hand in a zippered plastic bag and measured for total soluble solid (TSS) content using a temperature-compensating digital refractometer (Atago, Bellevue, WA, USA).
Labor and costs.
From 2021–23, the amount of time required to perform training and primocane suppression tasks was recorded for at least one plot per treatment. These data could not always be replicated because of the concurrent collection of plant growth data. Plant costs were estimated based on commercial nursery pricing, including royalties, for a new 8-ha field, with reduced pricing for a higher volume of plants needed for 0.75 m [US dollars (USD)1.40/plant] compared with 1.5 m (USD1.47/plant) spacing. Fruit sales revenue was calculated using a price of USD1.92/kg paid to growers based on assessment reports from 2018–20 and 2023 collected by the Oregon Raspberry and Blackberry Commission (2024). Prices from 2021 and 2022 were omitted from the calculation because of the devastating crop loss impact of the 2021 heat dome (Bell 2021), which inflated prices beyond typical levels (Oregon Raspberry and Blackberry Commission 2024). Labor costs were estimated at USD22/h for general labor and USD30/h for tractor operators, which included workers’ compensation, unemployment, insurance, and overhead expenses.
Data analysis.
Statistical analyses were performed using PROC MIXED in SAS v. 9.4 (SAS Institute, Cary, NC, USA), and means were separated at the 5% level using Tukey’s honestly significant difference test. The data were analyzed as a 4 × 2 × 2 × 2 factorial experiment, with year treated as the main plot, and plant spacing, training method, and primocane suppression treated as subplots, sub-subplots, and sub-sub-subplots, respectively.
Results and Discussion
Cane growth and fruiting components.
Plant production of new primocanes was affected significantly by each treatment, including plant spacing, training method, primocane suppression, as well as interactions between year and training method, year and primocane suppression, and training method and primocane suppression (Table 1). On average, the plants produced four more primocanes when they were spaced at 1.5 m than at 0.75 m. They also produced an average of four more primocanes with than without primocane suppression in 2021 and 2023, which were years 2 and 4 after planting, and an average of two to four more primocanes with August training than with new-over-old training in 2022 and 2023. In contrast, the primocanes were 0.4 m shorter on average with primocane suppression (Table 1). In other studies, primocane suppression likewise reduced primocane length while increasing the number in ‘Boysen’ and ‘Marion’ blackberry (Bell et al. 1995a; Stanley et al. 1999). When averaged across plant spacings in our study, the combination of August training and primocane suppression resulted in more primocanes (15 canes/plant) than any combination of new-over-old training and/or no primocane suppression (average, 12 canes/plant). Primocanes were also longer on average in years with less fruit production, which included 2021 and 2023 (Tables 1 and 2). In other studies, Cortell and Strik (1997a, 1997b) found a similar relationship between primocane growth and fruit production in ‘Marion’ blackberry.
Table 1.The effects of year, plant spacing, primocane training method, and primocane suppression on development of the primocanes and floricanes during the first 4 years of fruit production in ‘Columbia Star’ trailing blackberry in western Oregon, USA.
Table 2.The effects of year, plant spacing, primocane training method, and primocane suppression on yield and fruit quality during the first 4 years of fruit production in ‘Columbia Star’ trailing blackberry in western Oregon, USA.
Like primocanes, fruiting laterals on the floricanes were longer in 2023 than in 2022 or 2024 (Table 1). The length of the laterals was also affected by a three-way interaction among plant spacing, training method, and primocane suppression, although any difference among the treatments was small (≤ 0.04 m) and likely insignificant biologically. However, an interaction among year, spacing, and training method indicated the plants produced more fruit per lateral with new-over-old than with August training when spaced at 0.75 m in 2023 or 1.5 m in 2022 and 2024 (Fig. 1).
Fig. 1.Effects of year and plant spacing (0.75 or 1.5 m apart within row) (A) or training method (August or new-over-old training of primocanes) (B) on yield per plant in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. Error bars represent standard error.
Increased light exposure improves initiation and development of the flower buds in many perennial fruit crops (Morgan et al. 1985; Peavey et al. 2020; Wang et al. 2020; Yanez et al. 2009), including blackberry, where the development occurs during the fall and winter months (Strik 2012). During this time, primocanes trained using the new-over-old system are exposed to more light than those that are tightly bundled during August training. Many leaves are also stripped unintentionally during August training, reducing the photosynthetic capacity of the plants and potentially the number of berries per lateral.
Yield and fruit quality.
The plants produced less fruit in 2021 than in the following years (Table 2). This was expected, given that it was the first year the field was cropped (often called a “baby crop” year). There was also an early heat dome event in 2021 that occurred just before the second harvest in late June (White et al. 2023). During this event, ambient air temperature climbed to a record-setting high of 45.1 °C, whereas relative humidity remained very low, ranging between 14% and 28% throughout the day and reaching no higher than 42% at night. Many ripe and immature berries were heavily damaged by the heat. There was also a considerable amount of leaf scorching on the fruiting laterals, particularly on the west side of the canopy. On average, nearly 40% of the fruit was lost to heat damage in 2021 (Carroll et al. 2024). In the following years, greater yields corresponded to years with shorter but more fruitful laterals (Tables 1 and 2). Past studies have shown a compensatory ability for increased budbreak, fruitfulness, or fruit size in both blackberry and raspberry whenever there are fewer or shorter floricane laterals on the plants (Bell et al. 1995a; Cortell and Strik 1997a, 1997b; Freeman et al. 1989; Oliveira et al. 2004).
Predictably, the plants produced a greater yield each year when they were spaced at 1.5 m than at 0.75 m (Fig. 1). Yield per hectare, on the other hand, was greater when plants were spaced at 0.75 m, but only during the first 2 years of fruit production in 2021 and 2022 (Fig. 2). Consequently, the benefit of producing more fruit early on was outweighed by additional costs associated with planting at a greater density. Strik (2018) observed a similar response when ‘Marion’ blackberries were planted 0.6, 0.9, or 1.5 m apart. In that case, planting at a greater density only increased yield per hectare during the first year of fruit production. Cumulative yield per plant was 45% less at the 0.75 m spacing compared with 1.5 m, whereas cumulative yield per hectare was 9% greater with the 0.75 m spacing (Table 3).
Fig. 2.Interactive effects of year, within-row plant spacing, and primocane training method on yield per hectare in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. Error bars represent standard error.
Table 3. The effects of year, plant spacing, primocane suppression, and primocane training method on cumulative yield per plant and per hectare from 2021–24 for ‘Columbia Star’ trailing blackberry grown at Oregon State University’s North Willamette Research and Extension Center, Aurora, OR, USA.
Unlike different plant spacings, new-over-old training resulted in greater yields on both a per-plant and per-acre basis than August training in 2023 and 2024 (Figs. 1 and 2). In ‘Marion’ blackberry, Strik (2018) hypothesized that yield was greater with new-over-old training as a result of less cane damage during training and possibly increased carbohydrates and nutrients, because the floricanes remained on the plants through senescence. Cane damage that occurs during training is not always immediately apparent, and although damage was not quantified during cane counting in August in our study, we observed increased cane damage with August training compared with new-over-old in the weeks following training. In our case, greater yields with new-over-old was the result, at least in part, of more fruit per lateral (Fig. 3), which was not observed in ‘Marion’ blackberry (Strik 2018). New-over-old training resulted in a 15% greater cumulative yield per hectare compared with August training during the 4-year study (Table 3).
Fig. 3.Interactive effects of year and primocane training method on the number of fruit per lateral on plants spaced 0.75 m (A) or 1.5 m (B) apart within the rows in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA.
Primocane suppression had no effect on yield per plant or yield per acre during any year of the study, likely because the practice resulted in more canes per plant, which compensated for shorter cane lengths. Stanley et al. (1999) observed a similar response to primocane suppression in ‘Boysen’ blackberry. However, Bell et al. (1995a) found that primocane suppression increased the yield per plant in ‘Marion’ blackberry when it was done before July. Primocane suppression also increased yield in several cultivars of red raspberry (Freeman et al. 1989), although this may vary depending on the formulation of the chemical suppressant (Miller et al. 2008).
On average, the individual fruit on the plants were heavier in years with a greater yield (2022 and 2024) than in those with lower yield (2021 and 2023) (Table 2). In most cases, the effects of the treatments on fruit weight were small and <0.2 g/fruit. For example, the weight averaged 7.2 and 7.4 g/fruit with and without primocane suppression, respectively, when the plants were spaced 1.5 m apart or the primocanes were trained in August. Fruit weight was unaffected by primocane suppression when the plants were spaced 0.75 m apart or the primocanes were trained using the new-over-old method.
TSS contents were greater in 2023 than in the following year, which was likely the result of differences in yield and fruit weight (Table 2). Sugars and other soluble solids are often more concentrated when yields are low and fruit are smaller (Dixon et al. 2015; Strik and Davis 2021). August training also resulted in a greater concentration of soluble solids in the fruit than new-over-old training, which again could have been the result of differences in yield (but not fruit weight). However, because of the larger canopy, it is also possible that a portion of fruit was knocked off earlier by the machine harvester in the new-over-old plots. In fact, a greater percentage of fruit was picked during the first harvest from new-over-old- than from August-trained plots in 2023 and 2024 (Fig. 4). Nonetheless, in each treatment, TSS contents were comparable to those reported in the cultivar release for ‘Columbia Star’ (Finn et al. 2014), suggesting that the fruit were marketable and fully ripe when harvested from plots with either training method.
Fig. 4.Effect of training method (August or new-over-old training of primocanes) in 2021 (A), 2022 (B), 2023 (C), and 2024 (D) on the percentage of total yield on each harvest date in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. *, **, *** Significant at P ≤ 0.05, 0.01, and 0.001, respectively.
During harvest, MHE increased from 2021–24, whereas the percentage of dropped fruit declined (Table 2). For the first 2 years, as a full canopy was being established, MHE was greater at the 0.75 m spacing, but in the subsequent 2 years, there was no difference in MHE resulting from plant spacing. Strik and Buller (2002) also found that planting density in 4- to 7-year-old highbush blueberry (Vaccinium corymbosum L.) did not affect MHE. Simultaneously, new-over-old training resulted in greater MHE than August training at the 0.75 m spacing in 2021, and greater MHE at the 0.75 m spacing compared with 1.5 m spacing. New-over-old also had greater MHE than August training in 2023, but only at the 1.5 m spacing. In contrast, new-over-old had a lower MHE at both plant spacings in 2022, and training method did not affect MHE in 2024. Related to this, percent drop was less for new-over-old training in 2021 at the 0.75 m spacing and in 2023 at the 1.5 m spacing, but was greater for this training method at 1.5 m in 2021 and for both plant spacings in 2022.
Cull fruit sorted out on the harvester ranged from an average of 2% to 9% of total plant production each year (Table 2). In most cases, percent cull was similar between the training methods; however, it was slightly greater with new-over-old than with August training in 2022 (7.6% vs. 6.2%, respectively). This result is positive given the concern that a denser canopy with new-over-old could result in more problems with fungal disease or insect pests (Pscheidt and Ocamb 2025). Further study is warranted to determine whether and when these fields will need to be pruned and retrained. Percent cull was also affected by an interaction between plant spacing and primocane suppression (Table 2). In this case, primocane suppression reduced the percentage by 0.5% when the plants were spaced at 1.5 m, but had no effect on cull fruit when they were spaced at 0.75 m.
Amount of labor time needed for primocane training.
On average, August training required 19% less time per plant but 60% more time per linear meter of row when the plants were spaced at 0.75 m (10.9 min/plant and 14.4 min·m–1, respectively) than at 1.5 m (13.4 min/plant and 8.9 min·m–1, respectively). Nelson and Martin (1984) observed a similar increase in labor time when ‘Marion’ blackberry was August-trained and spaced at 0.5 or 1.0 m instead of 1.5 m. New-over-old training also required less time per plant and more time per linear meter when the plants were spaced at 0.75 m (1.2 min/plant and 1.5 min·m–1, respectively) than at 1.5 m (1.9 min/plant and 1.2 min·m–1, respectively), but it required substantially less time than August training. Relative to training the primocanes in August, new-over-old training reduced the total number of labor hours required from 2021–23 by 87% when the plants were spaced at 0.75 m and by 79% when they were spaced at 1.5 m. New-over-old training was also faster in 2022 than in 2021 because there was a larger existing canopy for primocanes to grow into, thus requiring less labor later in the season. Training time increased in 2023 for new-over-old relative to prior years because more time was needed to tuck the primocanes into the old canopy and prevent it from getting too wide to allow tractor access through the rows, but further increases in training time for new-over old are not anticipated because the canopy is established and relatively stable, with old floricanes decaying and new primocanes replacing them. Across seasons, primocane suppression reduced training time by an average of 13% with August training and 7% with new-over-old training. Although these times were based on student labor and, therefore, slower than the time needed by trained professional field workers, the relative differences remain applicable to a commercial setting.
Economics.
Based on the fruit yield and labor requirements for each treatment, a baseline economic analysis was completed (Table 4). Primocane suppression averaged USD143/ha per year for labor and herbicide, which was relatively low compared with the cost of primocane training. Training costs averaged USD2677/ha per year when the primocanes were trained using the new-over-old method and USD15,548 to USD26,556/ha, depending on the year and plant spacing, when the primocanes were trained traditionally in August. When yield was low in 2023, August training cost more than the potential income from fruit sales. Plants spaced at 0.75 m and trained using the new-over-old method produced the highest cumulative net return among the treatments, totaling USD3869/ha more than those spaced at 1.5 m and trained using the same method. August-trained plants had a net loss when spaced at 0.75 m, and only 15% of the net return calculated for new-over-old plants spaced at 1.5 m. When averaged across plant spacings and training methods, cumulative net returns were 24% higher with (USD30,317/ha) than without (USD23,100/ha) primocane suppression, which was primarily the result of less labor costs for primocane training; however, the difference was only USD1067/ha when the primocanes were trained using the new-over-old method (data not shown).
Table 4.Cash costs and returns for production of ‘Columbia Star’ trailing blackberries in western Oregon, USA.i
Conclusion
Planting ‘Columbia Star’ at a greater than normal density increased yield per hectare during the first 2 years of fruit production but had no effect on yield or fruit quality as plants matured. In contrast, using the reduced-labor, new-over-old cane management technique increased yield as plants matured compared with traditional August training. Primocane suppression generally increased cane number and reduced cane length, reduced training time in some cases, and had little to no effect on yield or fruit quality. The highest cumulative economic returns came from plants spaced at 0.75 m that were trained using the new-over-old method, followed closely by those with a 1.5 m spacing and likewise trained with the same method. In contrast, August training resulted in an economic loss at a greater density and small gains at the lower density spacing. Primocane suppression improved returns when combined with August training, but was less beneficial with new-over-old training. Additional research on this method is warranted to address any ongoing concerns with pests and diseases. In addition, the longevity of new-over-old production before renovation, and subsequent noncropping year, are needed.
Received: 07 Aug 2025
Accepted: 03 Nov 2025
Published Online: 26 Nov 2025
Published Print: 01 Dec 2025
Fig. 1.
Effects of year and plant spacing (0.75 or 1.5 m apart within row) (A) or training method (August or new-over-old training of primocanes) (B) on yield per plant in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. Error bars represent standard error.
Fig. 2.
Interactive effects of year, within-row plant spacing, and primocane training method on yield per hectare in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. Error bars represent standard error.
Fig. 3.
Interactive effects of year and primocane training method on the number of fruit per lateral on plants spaced 0.75 m (A) or 1.5 m (B) apart within the rows in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA.
Fig. 4.
Effect of training method (August or new-over-old training of primocanes) in 2021 (A), 2022 (B), 2023 (C), and 2024 (D) on the percentage of total yield on each harvest date in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. *, **, *** Significant at P ≤ 0.05, 0.01, and 0.001, respectively.
. 2015. Weed management, training, and irrigation practices for organic production of trailing blackberry: I. Mature plant growth and fruit production. HortScience. 50(
. 2020. The effects of canopy height and bud light exposure on the early stages of flower development in Prunus persica (L.) Batsch. Plants (Basel). 9:1073. https://doi.org/10.3390/plants9091073.
. 2002. Improving yield and machine harvest efficiency of ‘Bluecrop’ through high density planting and trellising. Acta Hortic. 574:227–231. https://doi.org/10.17660/ActaHortic.2002.574.34.
. 2021. Individual and combined use of sawdust and weed mat mulch in a new planting of northern highbush blueberry: III. Yield, fruit quality, and costs. HortScience. 56(
. 2020. Evaluating the effect of light intensity on flower development uniformity in strawberry (Fragaria ×ananassa) under early induction conditions in forcing culture. HortScience. 55(
Funding support was provided by the Oregon Raspberry and Blackberry Commission and the Northwest Center for Small Fruits Research. We appreciate the support.
Effects of year and plant spacing (0.75 or 1.5 m apart within row) (A) or training method (August or new-over-old training of primocanes) (B) on yield per plant in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. Error bars represent standard error.
Fig. 2.
Interactive effects of year, within-row plant spacing, and primocane training method on yield per hectare in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. Error bars represent standard error.
Fig. 3.
Interactive effects of year and primocane training method on the number of fruit per lateral on plants spaced 0.75 m (A) or 1.5 m (B) apart within the rows in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA.
Fig. 4.
Effect of training method (August or new-over-old training of primocanes) in 2021 (A), 2022 (B), 2023 (C), and 2024 (D) on the percentage of total yield on each harvest date in a mature planting of ‘Columbia Star’ trailing blackberry in western Oregon, USA. *, **, *** Significant at P ≤ 0.05, 0.01, and 0.001, respectively.
