Materials and methods

The study was integrated into a larger research vineyard (2 acres) that included additional studies on grafted ‘Chambourcin’ vines established in 2008 and 2009 at the University of Missouri’s Southwest Research Center near Mt. Vernon (lat. 37.074297°N, long. 93.879708°W, USDA Plant Hardiness Zone 6a) (Migicovsky et al., 2019; Thomas et al., 2017). The soil was a Hoberg silt loam (fine loamy, siliceous, mesic Mollic Fragiudalfs) that is upland, deep, gently sloping, and moderately well drained to a fragipan at 40 to 90 cm (Hughes, 1982). Soil was pushed into berms (≈25 cm high) to form vineyard rows, and vines planted thereon to improve soil drainage and root penetrable soil depth over the fragipan. Drip irrigation [emitters rated 0.42 gal/h spaced 36 inches (Netafim, Fresno, CA)] was used to supplement rainfall as needed to supply 1 acre-inch of water per week during the growing seasons. Vines in this study were established with paired trunks and trained to high bilateral cordons (single curtain), with average height of 5.8 ft above soil level. More detail on climate, trellis system, and site establishment is provided in Thomas et al. (2017), along with protocols on soil fertilization, pest management, weed control, winter graft union protection, and pruning.

‘Chambourcin’ scions were bench-grafted to ‘UCD GRN-2’, ‘UCD GRN-3’, ‘UCD GRN-4’, ‘UCD GRN-5’, and 3309C rootstocks, and own-rooted ‘Chambourcin’ grapevines were produced and donated by Wonderful Nurseries (Wasco, CA). All were transplanted 16 July 2010 as green-growing potted vines, with graft unions positioned ≈5 inches above soil level to prevent scion rooting. Spacing was 7 ft between vines in-row, and 9.7 ft between rows. Graft unions were afforded winter protection with mounds of mushroom and/or municipal compost (J-M Farms, Miami, OK; City of Monett, Monett, MO, respectively) during the first five winters; the compost was thereafter spread into the vineyard rows each spring. The study was established in a completely randomized block design, with six rootstock treatments (five rootstocks plus own-rooted vines) and four single-row blocks (replications) of each rootstock. Each plot contained two side-by-side vines of the same rootstock, for a total of 48 vines in the study. Additional ‘Chambourcin’ guard vines were deployed at the row ends and between blocks such that all vines were grown in near-identical environments.

During the establishment seasons of 2010–12, all inflorescences were removed so that no fruit developed to encourage robust vine establishment. All fruit from the experiment was harvested 1 Oct. 2013, 30 Sept. 2014, and 3 Oct. 2015. Harvest data included total fruit yield and cluster number per plot. Samples of 100 random ripe berries per plot were collected, weighed, immediately refrigerated, and then analyzed within 2 d for soluble solids concentration, pH, and titratable acidity in terms of tartaric acid [laboratory methods and instruments used are described in Thomas et al. (2017)]. Mean berry size was calculated, and mean number of berries per cluster calculated by dividing mean cluster weight by mean berry weight. Pruning weight data were collected in Feb./Mar. 2013–16 to quantify vegetative growth during the previous year (2012–15), and Ravaz index (yield/pruning weight) calculated for the 2013–15 growing seasons. Vines were balance-pruned to 20 + 20 formula (20 nodes retained for every 1 lb of pruning weight).

To quantify rootstock influence on scion response to the regional climate, dates of budbreak and shootless nodes data were collected. Spring budbreak date, defined as the date on which 50% or more of buds among both vines in a plot exhibited emergence of green tissue beneath the bud scales (Dokoozlian, 1999), was recorded in seasons 2016–18. Shootless nodes, defined as count nodes retained during dormant pruning that failed to produce viable shoots, were assayed in July 2017 when eight randomly selected nodes per vine were examined for the presence of live shoot growth. This metric, which measures bud survival and the sustainability of a vine’s bearing surface, was used to quantify observed differences between rootstock treatments several years after establishment.

All data were subjected to analyses of variance in the form of a randomized complete block design using a general linear statistical model (SAS version 9.4; SAS Institute, Cary, NC) to elucidate differences in fruit production, fruit characteristics, phenology, and vine size among the rootstocks. Means were separated by Fisher’s least significant difference test (P ≤ 0.05). The experimental unit was the two-vine plot, thus two-vine plot data means were statistically analyzed; however, all appropriate data are reported on a single-vine basis.

Results and discussion

‘Chambourcin’ initially grew satisfactorily on all rootstocks, as well as on its own roots, as demonstrated by acceptable pruning weights after 2.5 years of growth in 2012 (Table 1). For reference, Jordan et al. (1981) recommend optimum pruning weights of 2 to 3 lb/vine (for ‘Concord’ with 8-ft cordons) or 0.3 to 0.4 lb/ft of cordon, and Kliewer and Casteel (2003) suggest optimal values of 0.2 to 0.4 lb/ft. After the establishment period, the rootstock ‘UCD GRN-3’ consistently produced pruning weights among the highest recorded, although statistical differences from other rootstocks varied among years. The exception was ‘UCD GRN-5’, for which pruning weights early in the study (2012, 2013) were lower than most other rootstock treatments. Although ‘UCD GRN-5’ produced a fruit crop comparable to the other rootstocks in 2013 (detailed yield data not shown), vegetative growth was greatly reduced (likely due to fruiting stress) as indicated by low pruning weights in 2013, resulting in the decision to de-fruit the vines in 2014. After 1 year of rest, ‘UCD GRN-5’ produced a fruit crop comparable to the other treatments in 2015, but, once again, at the expense of vegetative growth; 2015 pruning weights for ‘UCD GRN-5’ were significantly lower than all other rootstocks, resulting in very high Ravaz indices.

Table 1.

Pruning weight and Ravaz index (yield/pruning weight) for ‘Chambourcin’ winegrape grown on five rootstocks and own-rooted vines at Mt. Vernon, MO in 2012–15.

View Table

Fruit yields, as well as cluster, berry, and juice characteristics across the 3 harvest years and five rootstocks/own-rooted are shown in Table 2. Fruit yields generally decreased across the 3 years, from 13.3 kg/vine in 2013 to 9.6 kg/vine in 2015, which roughly extrapolates to 9.4 and 6.8 tons/acre in 2013 and 2015, respectively. For comparison, Bordelon (2009) indicates 5.1 tons/acre is an expected yield for ‘Chambourcin’ in Indiana, and Dami et al. (2005) produced ‘Chambourcin’ yields ranging from 4.5 to 7.1 tons/acre in Ohio and 4.8 to 6.7 tons/acre in Illinois in a crop load study. Thus, ‘Chambourcin’ was highly productive at the study site, which created challenges in achieving and maintaining vine balance. Although seasonal influence had a statistically significant impact on number of clusters per vine, cluster weights, and number of berries per cluster, rootstock did not influence such factors during the 3 harvest years (remembering that ‘UCD GRN-5’ was de-fruited in 2014). This indicates that the GRN rootstocks 2, 3, and 4 performed as well as both the benchmark 3309C rootstock and own-rooted ‘Chambourcin’ vines for these economically important parameters. Berries were smaller (2.17 g) in 2013, likely a reflection of the high yields generated during that first year of production. Individual berry weights also varied by rootstock, but such differences were generally modest and of limited commercial significance.

Table 2.

‘Chambourcin’ winegrape fruit yields, and cluster, berry, and juice characteristics grown on five rootstocks and own-rooted vines at Mt. Vernon, MO in 2013–15.

View Table

Juice characteristics were variable across the 3 years but all were within commercially acceptable ranges (Boulton et al., 1999). The GRN rootstocks all produced berries with higher soluble solids concentration (collective mean 22.2%) compared with ungrafted vines (21.2%), but only ‘UCD GRN-3’ was higher than 3309C for this parameter. Although some effect of rootstock was observed for pH and titratable acidity, values for fruit grown on the GRN rootstocks generally were near or between those for 3309C and ungrafted vines. When evaluating growing year × rootstock statistical interactions, the only factors that were affected were single berry weight and titratable acidity.

Observed weather during this experiment was generally representative of a midwestern U.S. continental climate, with routine seasonal fluctuations in temperatures and precipitation, and occasional temperature extremes (Table 3). Although extreme temperatures are, by definition, rare, their “frequent” occurrence in the midwestern United States presents a perpetual challenge to economical winegrape production in the region. Maximum summer temperature and seasonal heat summation data (growing degree days, base 50 °F) confirm exceptionally hot summers during the establishment years of 2011 and 2012. More normal temperatures occurred during the data years of 2013–15, except for extreme cold during the winter of 2013–14. Despite these seasonal weather challenges, summer irrigation should have reduced moisture stress during dry periods/seasons, and winter graft union protection should have mitigated freezing stresses somewhat during this experiment.

Table 3.

Maximum (Max) and minimum (Min) ambient temperatures, total annual precipitation, and growing degree days [base 50 °F (10.0 °C)] recorded at or near the study of ‘Chambourcin’ winegrape grown on five rootstocks and own-rooted vines at Mt. Vernon, MO in 2010–17.

View Table

The challenges observed in this study with ‘Chambourcin’ scions grafted to ‘UCD GRN-5’ are difficult to explain given a limited dataset. One plausible explanation is that ‘UCD GRN-5’ may simply not be adapted to the environmental conditions under which this experiment was conducted, and may have suffered freeze damage or other abiotic stress. A second is that ‘UCD GRN-5’ may be a low capacity or late-establishing rootstock; early cropping stresses may have permanently weakened the vines, thereby directly and/or indirectly increasing susceptibility to additional stresses. When ‘UCD GRN-5’ vines suffered damage during the first years of the study, both the rootstock and ‘Chambourcin’ scion survived, and cordons were reestablished from epicormic buds repeatedly as necessary. By 2017, however, nearly all the ‘UCD GRN-5’ vines had suffered significant long-term cordon or whole vine damage as they continued to weaken. Of eight ‘UCD GRN-5’ nodes evaluated per vine in July 2017, 7.9 nodes were shootless, whereas the other rootstock treatments continued to thrive (Table 4). Although spring frost/freeze injury was not visibly evident during this experiment, budbreak data show that ‘UCD GRN-5’ apparently hastened spring budbreak by 2 to 3 d in 2016 and 3 to 5 d in 2018, possibly increasing the potential risk of late spring freeze damage (Johnson and Howell 1981; Kovacs et al., 2003). Grapevine rootstocks have been shown to influence scion phenology and budbreak in numerous cultivars (e.g., Ferroni and Scalabrelli, 1995; Menora et al., 2015; Neal et al., 2016; Reddy, 1990).

Table 4.

Budbreak and shootless node data for ‘Chambourcin’ winegrape grown on five rootstocks and own-rooted vines at Mt. Vernon, MO in 2016–18.

View Table

Kamas et al. (2020) and Scheiner et al. (2020) evaluated these same nematode-resistant rootstocks (‘UCD GRN-2’, ‘UCD GRN-3’, ‘UCD GRN-4’, and ‘UCD GRN-5’), along with additional rootstocks, grafted to ‘Sangiovese’ and ‘Blanc du Bois’, respectively, in Texas. ‘UCD GRN-5’ performed poorly on a pH 8.5 soil with ‘Sangiovese’ (attributed to pH-induced nutritional deficiency), but all four performed satisfactorily on a pH 8.1 soil with ‘Blanc du Bois’ and a pH 6.6 soil with ‘Sangiovese’. In the ‘Sangiovese’ study on pH 6.6 soil, the UCD rootstocks did not statistically influence fruit yields compared with own-rooted vines; however, ‘UCD GRN-5’ produced among the lowest pruning weights and among the lowest number of berries per cluster. Similarly, the UCD rootstocks grafted to ‘Blanc du Bois’ did not statistically influence pruning weight, and only ‘UCD GRN-4’ increased fruit yields compared with own-rooted vines in 1 production year.

Because the rootstocks ‘UCD GRN-2’, ‘UCD GRN-3’, and ‘UCD GRN-4’ generally performed similarly to the standard 3309C and own-rooted vines under identical ambient conditions in our study, we can conclude that they show promise for use in midwestern U.S. winegrape production. ‘UCD GRN-3’ was statistically more vigorous and produced fruit with a higher soluble solids concentration compared with most other rootstocks or own-rooted vines. Concurrently, our study indicates that ‘UCD GRN-5’ may not be suitable for southern Missouri growing conditions, and presumably USDA Hardiness Zone 6a and colder zones. Although these rootstocks are resistant to the enumerated nematodes in California (Ferris et al., 2012), they have not been specifically tested for resistance to nematode species and races that are prevalent in the midwestern United States. In particular, although the rootstocks are resistant to dagger nematode, their resistance to american dagger nematode, which is of greater concern in the midwestern United States, is presently unknown. This was the first significant trial of these new nematode-resistant rootstocks in the midwestern United States; more testing is needed, including an evaluation of the rootstocks’ resistance to problematic nematodes in the region, before they can be recommended with confidence.

Units

View Table