Materials and Methods

Plant material

The rootstocks were developed in Florida (University of Florida), California (University of California, Riverside), and Spain (Valencian Institute of Agricultural Research). They contain sexual and somatic hybrids from a wide range of different germplasms (Table 1). The same rootstocks were used in both trials, except for sour orange, which was only included in one trial. Rootstock liners were grown from seeds and grafted with ‘Hamlin’ orange (C. sinensis) scion using standard procedures (Albrecht et al., 2017; Lewis and Alexander, 2008). Trees were produced in a commercial citrus nursery (Lykes Citrus, Basinger, FL) and planted in 2015 in commercial citrus orchards at the locations described.

Table 1.Rootstock names and their parentage.

View Fullscreen

Results

Canopy health, tree survival, and leaning

The rootstock effect was significant for foliar disease symptoms and canopy thickness in both locations (Supplemental Table 1). In general, most trees looked healthy, with average disease indices not exceeding 1.6. At the Basinger location, trees on ‘Amb+Benton’ had a significant lower HLB disease index (1.2) than trees on ‘ES-3’ (1.6). At the Lake Wales location, trees on ‘UFR-5’ had a significant lower disease index (1.1) than trees on ‘ES-7’ (1.5). Most rootstocks induced thick canopies at either location. At the Basinger location, ‘C-54’ induced a significantly thicker canopy (4.8) than ‘Changsha+Benton’ (4.3). At the Lake Wales location, most rootstocks induced significantly thicker canopies (4.6–4.8) than ‘Green-3’ (3.8). Canopy color ratings ranged from 4.7 to 4.9, but there were no significant differences among rootstocks.

The rootstock significantly affected tree survival and leaning (Table 2). The survival rate was 98%, on average, in both locations, and significant rootstock effects were only found for the Basinger location, where trees on ‘Green-3’ had a significantly lower survival rate (85.4%) compared with most of the other rootstocks. At the Basinger location, the percentage of leaning trees was highest for trees on ‘Amb+Benton’ (41.7%), followed by ‘C-22’ (33.3%) and ‘Changsha+Benton’ (31.3%), and no leaning was observed for trees on ‘Green-7’, ‘UFR-2’, ‘Wgft+50-7’, and ‘UFR-5’. At the Lake Wales location, trees on ‘Amb+Benton’, ‘Changsha+Benton’, and ‘C-22’ had the highest percentage of leaning (52.1% to 66.7%), whereas less than 10% of trees leaned on most other rootstocks.

Table 2.Tree survival and leaning of ‘Hamlin’ orange trees on different rootstocks.

View Fullscreen

Across both locations, the lowest percentage of survival was found for trees on ‘Green-3’ (90.6%). The percentage of leaning trees was highest for trees on ‘Amb+Benton’ (54.2%), followed by trees on ‘Changsha+Benton’ and ‘C-22’ (42.7%). The leaning of trees at the Basinger location and the Lake Wales location was significantly and inversely correlated with tree height (R = −0.24 and R = −0.35) and canopy volume (R = −0.27 and R = −0.30). A significant but weak correlation was also found between tree survival and canopy volume (R = 0.15 and R = 0.24).

Tree size

Tree height, canopy volume, scion trunk diameter, and scion-to-rootstock trunk diameter ratio (SRR) were significantly influenced by rootstock in both locations (Table 3). ‘X-639’, ‘C-54’, ‘C-57’, and ‘C-146’ induced the tallest trees (2.3 m) at the Basinger location and were among the rootstocks producing the tallest trees (2.2–2.3 m) at the Lake Wales location. ‘Green-3’, ‘Changsha+Benton’, and ‘Amb+Benton’ produced the smallest trees (1.5–1.6 m) at the Basinger location and were among the rootstocks producing the smallest trees (1.4–1.6 m) at the Lake Wales location.

Table 3.Tree size of ‘Hamlin’ trees on different rootstocks.

View Fullscreen

Similar results were found for the canopy volume, which was largest for trees on ‘X-639’ and ‘C-54’ (2.8–3.1 m³) at the Basinger location and for trees on ‘X-639’, ‘C-54’, and ‘C-57’ (3.0–3.3 m³) at the Lake Wales location. Trees on ‘UFR-3’, ‘Green-3’, and ‘Changsha+Benton’ produced the smallest canopy volume (1.0–1.1 m³) at the Basinger location, and together with ‘Sorp+Sh-991’ at the Lake Wales location (0.8–1.0 m³).

Scion trunk diameters were largest in trees on ‘C-54’ and ‘C-57’ (8.5 cm) at the Basinger location and on ‘C-57’ (8.5 cm) at the Lake Wales location. The smallest scion trunk diameters were found in trees on ‘Green-3’ (5.3 cm) at the Basinger location and in trees on ‘UFR-3’ (5.1 cm) at the Lake Wales location. The SRR was highest (0.86) for trees on rootstocks ‘ES-5’ and sour orange at the Basinger location and for trees on ‘Changsha+Benton’ (0.88) and ‘Amb+Czo’ (0.87) at the Lake Wales location. The lowest SRR (0.63) was found for the trees on ‘Swingle’ at both locations along with trees on ‘Wgft+50-7’ (0.63) at the Basinger location and ‘UFR-3’ (0.65) at the Lake Wales location.

Tree height was significantly correlated with both canopy volume and scion trunk diameter at the Basinger location (R = 0.92 and R = 0.88, respectively) and the Lake Wales location (R = 0.89 and R = 0.85, respectively).

Fruit yield

Fruit yield was measured in production years 2018–19 and 2019–20, and the cumulative yield was calculated. Rootstock effects on yield were significant for both years and at both locations (Table 4). At the Basinger location, the average yield was nearly three-fold lower in 2018–19 (5.6 kg/tree) than in 2019–20 (15.4 kg/tree). Yield per tree ranged from 2.8 kg/tree for trees on ‘ES-5’ to 8.5 kg/tree for trees on ‘US-897’ in 2018–19, and from 9.6 kg/tree for trees on ‘UFR-3’ to 20.2 kg/tree for trees on ‘ES-1’ in year 2019–20; however, separation of means was not significant in either production year.

Table 4.Fruit yield of ‘Hamlin’ orange trees on different rootstocks.

View Fullscreen

The average yield at the Lake Wales location was similar in both production years (Table 4). ‘UFR-5’ and ‘ES-1’ induced the highest yields per tree in 2018–19 (20.2 and 18.0 kg), and ‘X-639’ and ‘C-54’ induced the highest yields per tree (19.9 and 18.3 kg) in 2019–20. The lowest yields per tree were induced by ‘Sorp+Sh-991’, ‘Wgft+50-7’, ‘UFR-3’, and ‘Amb+Benton’ (4.6–6.8 kg) in year 2018–19, and by ‘Changsha+Benton’ and ‘Sorp+Sh-991’ (4.6 and 4.9 kg) in 2019–20.

Across both locations, the most productive rootstocks in terms of average cumulative yield were ‘ES-1’ (29.5 kg/tree), followed by ‘UFR-5’, ‘C-54’, and ‘X-639’ (27.8–28.0 kg/tree). The least productive rootstocks were ‘UFR-3’ (13.1 kg), followed by ‘Sorp+Sh-991’ (14.1 kg), ‘Changsha+Benton’ (15.3 kg), and ‘Amb+Benton’ (15.6 kg/tree). Fruit yield per tree was significantly correlated with canopy volume at the Basinger location (R = 0.39) and at the Lake Wales location (R = 0.66).

Yield efficiency

The rootstock effect was significant for yield efficiency in both locations and both production years (Table 5). In 2018–19, yield efficiency at the Basinger location ranged from 2.1 kg/m³ for trees on ‘X-639’ to 9.9 kg/m³ for trees on ‘Amb+Benton’, but the mean separation was not significant. In 2019–20, the average yield efficiency was highest for trees on ‘Green-3’ (14.3 kg/m³) and lowest for trees on ‘C-146’, ‘C-54’, and ‘C-22’ (4.3–5.4 kg/m³).

Table 5.Yield efficiency of ‘Hamlin’ orange trees on different rootstocks.

View Fullscreen

At the Lake Wales location, ‘UFR-5’, ‘Green-3’, and ‘Changsha+Benton’ induced the highest yield efficiency (15.7–17.3 kg/m³) and ‘ES-2’, ‘ES-3’, ‘C-57’, and ‘UFR-2’ induced the lowest (6.1–6.2 kg/m³) in 2018–19. In 2019–20, ‘Green-3’ induced the highest yield efficiency (10.5 kg/m³) and ‘UFR-6’, ‘UFR-17’, ‘ES-6’, ‘UFR-4’, ‘UFR-2’, ‘ES-3’, and ‘ES-2’ induced the lowest yield efficiency (4.2–5.1 kg/m³).

Across both locations and years, the highest average yield efficiency was induced by ‘Green-3’ (11.9 kg/m³), followed by ‘Amb+Benton’ (10.5 kg/m³), ‘Changsha+Benton’ (10.2 kg/m³), and ‘ES-4’ (10.0 kg/m³), whereas the lowest yield efficiency was induced by ‘ES-2’ (4.8 kg/m³), followed by ‘C-146’, ‘ES-3’, and ‘X-639’ (5.2–5.4 kg/m³). Yield efficiency was significantly and inversely correlated with canopy volume at the Basinger location (R = −0.59) and the Lake Wales location (R = −0.30).

Fruit quality

Fruit quality variables were measured in the production year 2019–20. The rootstock effect was significant for most variables in both locations. In the Basinger trial, ‘Green-3’ induced the highest weight fruit (174 g) and ‘ES-2’ induced the lowest fruit weight (140 g) (Table 6). The juice percentage was highest for trees on ‘ES-4’ (59.1%) and lowest for trees on ‘ES-7’ (48.6%). ‘Green-3’, ‘White-1’, and sour orange induced the largest amount of TSS (9.40% to 9.80%), and ‘C-22’, ‘C-146’, and ‘ES-2’ induced the smallest amount of TSS (7.95% to 8.10%). The acid percentage was highest for trees on ‘Green-3’ (0.565%) and lowest for trees on ‘ES-2’ (0.443%). The TSS-to-acid ratio ranged from 16.2 to 18.4, but the mean separation was not significant.

Table 6.Fruit quality of ‘Hamlin’ orange trees on different rootstocks in trial 1 (Basinger).

View Fullscreen

At the Lake Wales location, the average fruit weight was 147 to 168 g, but there was no significant rootstock effect (Table 7). Juice percentage was highest for trees on ‘Swingle’ (58.3%) and lowest for trees on ‘ES-7’ (49.9%). ‘UFR-5’, ‘Amb+Czo’, and ‘Changsha+Benton’ induced the largest TSS% in the fruit (9.63% to 9.84%), whereas ‘ES-2’ and ‘ES-3’ induced the smallest TSS% (8.38% and 8.42%, respectively). ‘UFR-5’ induced the highest percentage of acid in the fruit (0.557%), and ‘ES-4’ and ‘ES-6’ induced the lowest (0.322% and 0.348%, respectively). The highest TSS-to-acid ratio was induced by ‘ES-4’ (28.4), and the lowest was induced by ‘Swingle’ (17.4). TSS was significantly and inversely correlated with canopy volume at the Basinger location (R = −0.68) and at the Lake Wales location (R = −0.53).

Table 7.Fruit quality of ‘Hamlin’ orange trees on different rootstocks in trial 2 (Lake Wales).

View Fullscreen

Discussion

Considering the devastation caused by HLB, precocity may be one important criterion for rootstock selection to cope with economic losses due to tree decline when the disease reaches an advanced stage of progression. Our results showed significant variations among rootstocks in their effects on most of the horticultural traits measured. The identification of rootstock traits at an early stage of production is valuable for determining rootstock impacts on the economic viability of the mature orchard in the long term.

Although significant rootstock effects were measured, most trees had a healthy canopy with moderate foliar disease expression. The total percentage of dead trees was less than 2% in both locations, which is normal for young commercial citrus plantings. Trees on the small tree size-inducing rootstock ‘Green-3’ had the highest mortality rate among all rootstocks at the Basinger location. The low foliar disease expression of trees on this rootstock suggests that factors other than HLB may have been responsible.

The average percentages of wind-induced leaning of trees at both locations were 15.8% and 10.6% at Lake Wales and Basinger, respectively. The higher percentage of leaning trees at the Lake Wales location may be attributable to this location being part of the Central Florida Ridge, an ancient sand dune representing the highest elevation in Florida, and its characteristic sandy soils. Trees on some of the dwarfing rootstocks leaned more than trees on some of the larger size-inducing rootstocks, but the (inverse) correlation between tree size and leaning was weak. One of the exceptions was ‘C-22’, which induced a high percentage of leaning while producing larger than average trees. The leaning of trees did not appear to have a negative influence on tree health and productivity at the time of evaluation; however, it is expected that growers will prefer wind-tolerant rather than wind-susceptible cultivars. To our knowledge, this is the first report of citrus that compared rootstock cultivar effects on the ability of citrus trees to withstand tropical force winds. A study of apple rootstocks also found that trees on dwarfing rootstocks leaned more than trees on vigorous rootstocks (Schupp, 1992); this was attributed to the heavy crop load of dwarfing trees and weak anchorage. Differences in wind resistance among rootstocks observed in our study are likely associated with different root architectures because trees were bearing few fruit at the time of the hurricane. A study following the aftermath of hurricane Andrew, which destroyed much of the tropical fruit acreage in South Florida in 1992, found that the percentage of toppled and surviving trees was correlated with tree height and age and depended on the crop species and other factors (Crane et al., 1993).

The rootstocks induced a wide range of vigor in the ‘Hamlin’ scion. The diploid rootstocks produced mostly medium to large trees, with ‘X-639’, ‘C-54’, ‘C-57’, ‘C-146’, and ‘ES-1’ producing the largest trees in the trials. Surprisingly, ‘US-897’ and ‘C-22’, which are generally regarded as small tree size-inducing rootstocks (Bowman et al., 2008; Siebert et al., 2010), produced medium trees in both locations when compared with the rootstock standards ‘Swingle’ and sour orange which are known to produce medium trees. Other studies that included the rootstocks ‘C-22’, ‘C-54’, ‘C-57’, ‘C-146’, and ‘X-639’ showed similar results (Castle et al., 2011; Louzada et al., 2008; Roose, 2008). It is anticipated that the size-limiting effect of ‘US-897’ and ‘C-22’ will manifest during later production years. The delayed effect of ‘US-897’ on tree size has also been observed by commercial citrus growers (personal communications). It must be noted that some hybrids of C. reticulata and P. trifoliata (citrandarins) are regarded as HLB-tolerant (Albrecht and Bowman, 2012; Boava et al., 2014). A higher level of tolerance may have contributed to the vigorous growth of trees on the citrandarin rootstocks ‘X-639’, ‘C-54’, ‘C-57’, ‘C-146’, ‘ES-1’, ‘US-897’, and ‘C-22’ observed in our study.

Among the tetraploid rootstocks, ‘Orange-14’, ‘UFR-4’, and ‘UFR-5’ induced the largest tree size, which was average when compared across all rootstocks. Most of the other tetraploid rootstocks induced below-average size trees, which is in accordance with the results from other field trials conducted on both flatwood and Central Ridge sites in Florida (Grosser et al., 2011). During those trials, many of the tetraploid rootstocks induced a tree size smaller than 9 ft (2.7 m) after 10 years of growth. The dwarfing capacity of tetraploid rootstocks was suggested to be a graft union response because diploid and tetraploid cells differ morphologically, with the latter being larger and containing thicker cell walls (Grosser et al., 2011). However, the reduced growth of tetraploid rootstocks in comparison with diploid rootstocks was also observed at the seedling stage (Allario et al., 2011; Guerra et al., 2014). It is generally recognized that the reduced capacity of the dwarfing rootstocks to transport water from the soil to the aboveground part of the plant contributes to the vigor-inducing potential of a rootstock. Lower hydraulic conductivity was reported for apple trees on the dwarfing rootstock M.27 compared to those on the invigorating rootstock MM.106 (Atkinson et al., 2003). Forner-Giner et al. (2014) reported a reduced hydraulic conductance of ‘Navelina’ orange trees on dwarfing rootstocks ‘FA-517’ and ‘FA-418’ compared with trees on Carrizo rootstock. Factors such as soil and environmental conditions, management practices, pest and disease pressure, and rootstock compatibility with the scion may further affect the rootstock influence on tree size (Bowman and Joubert, 2020). The observed variations in the tree size induced at different locations for some of the rootstocks in our study were likely due to the differences in the soil and other environmental conditions because both sites had similar management practices, the same scion, and similar HLB disease pressure.

The SRR represents the smoothness of the graft union, and it has often been considered an indicator of the compatibility of scion and rootstocks (Kallsen and Parfitt, 2011). However, different vigor of the grafting partners can also result in overgrowth of one of the partners without posing any hazard to trunk health and tree physiology (Bowman and Joubert, 2020), as demonstrated by the long-time dominance of ‘Swingle’. Together with ‘Wgft+50-7’, ‘Swingle’ induced the lowest SRR value in both trials.

The average yield at the Basinger location in 2018–19 was considerably lower than that at the Lake Wales location. One reason for this was an unusually high incidence of Xanthomonas citri, the causal organism of citrus canker, at the Basinger location, possibly as aftermath of hurricane Irma. Most of the high vigor-inducing rootstocks, such as ‘X-639’, ‘ES-1’, ‘C-54’, and ‘C-57’, were among the rootstocks that induced the highest cumulative yields across both production years and both locations, amounting to 149 to 158 boxes of fruit per acre (1 box = 90 lb or 40.8 kg of fruit). Although this amount is not yet considered commercially significant, many of the low vigor-inducing tetraploid rootstocks induced less than half this amount (70–83 boxes/acre), which, if this trend continues, may prevent their commercial acceptance. Similar to our findings, in a rootstock trial with ‘Marsh’ grapefruit, ‘X-639’ and ‘C-54’ induced high cumulative yields and large tree sizes (Castle et al., 2011). In that trial, ‘C-146’ also produced high cumulative yields; however, our yield on this rootstock was average to low.

It is generally recognized that the size of the canopy is positively related to fruit yield (Anderson, 1987). An association between tree size and yield was also observed during rootstock trials with ‘Valencia’ and ‘Hamlin’ scion in Florida and with ‘Pera’ scion in Brazil (Albrecht et al., 2012; Bowman et al., 2016a; Quaggio et al., 2004). Despite the general relationship between canopy size and yield, in the present study, the medium vigor-inducing rootstock ‘UFR-5’ and the lower vigor-inducing ‘ES-4’ were among the rootstocks that induced higher than average cumulative yields. In contrast, the high vigor-inducing ‘ES-2’ rootstock induced a lower-than-average cumulative yield. The correlation between canopy size and yield in our trials was moderate (0.39) to high (0.66), depending on the location, suggesting that the canopy size of a citrus tree may not always be a good predictor of yield.

High yield-inducing rootstocks do not necessarily increase the productivity of a citrus orchard if the planting density is not optimized for the size of the mature productive tree (Bowman and Joubert, 2020). Hence, yield efficiency (yield per unit canopy volume) is a more suitable variable for determining productivity of a mature orchard. In two recent studies of ‘Ray Ruby’ grapefruit, high-density planting resulted in increased yield per hectare, increased fruit TSS contents, and decreased canopy volume (Phuyal et al., 2020). This demonstrates that the use of high-yield efficient rootstocks and optimized planting density can increase productivity and minimize cost per unit production.

Most of the high vigor-inducing rootstocks, such as ‘X-639’, ‘C-54’, ‘C-57’, and ‘C-146’, and some of the medium vigor-inducing rootstocks, such as ‘ES-2’ and ‘ES-3’, were not as yield-efficient in our study as the lower vigor-inducing rootstocks. Among the rootstocks that induced the highest yield efficiency were the tetraploid rootstocks ‘Green-3’, ‘Amb+Benton’, ‘Changsha+Benton’, ‘Amb+Czo’, and ‘6058+2071-02-2’, which produced some of the smallest trees in the trials. However, despite having high yield efficiency, the overall yield potential of these rootstocks may not be sufficient, in the longer-term, to compete with the more vigorous rootstocks. Moreover, some of these small size-inducing rootstocks were leaning because of the strong winds imposed by the hurricane in 2017; the long-term impact remains to be investigated. ‘ES-4’, ‘UFR-5’, and ‘Swingle’ were among the rootstocks inducing higher-than-average yield efficiency across both locations and production years while producing medium-sized trees.

The priority for different fruit quality traits varies in different production areas based on the scion used and purpose (fresh fruit or juice production). For sweet oranges like ‘Hamlin’ that are grown for juice processing, TSS (Brix) is the most important fruit quality variable, followed by the acid content and the ratio of the two. Similar to other citrus studies, rootstocks in this study induced significant differences in all measured fruit quality variables in both trials (Bowman et al., 2016b; Grosser et al., 2011; McCollum and Bowman, 2017).

Most rootstocks induced more than 50% juice, except for ‘ES-7’, which is a lemon-type hybrid. Despite differences among locations for some of the rootstocks, there was a tendency for dwarfing and semi-dwarfing rootstocks to induce larger amounts of TSS than vigorous rootstocks. The inverse correlation between tree size and TSS was more evident at the Basinger location (R = −0.78) than at the Lake Wales location (R = −0.53). Among the highest TSS-inducing rootstocks in both locations were the smallest size-inducing tetraploid rootstocks ‘Green-3’, ‘White-1’, ‘Changsha+Benton’, ‘Amb+Benton’, ‘Amb+Czo’, ‘Wgft+50-7’, ‘UFR-1’, and ‘UFR-6’. These rootstocks were also found to induce excellent fruit quality during a previous field trial with ‘Valquarius’ scion (Grosser et al., 2011). The low vigor-inducing rootstock ‘US-897’ and the high vigor-inducing rootstock ‘X-639’ induced the highest and lowest TSS contents, respectively, during a study of ‘Ray Ruby’ grapefruit trees (McCollum and Bowman, 2017).

The influence of rootstocks on fruit quality is assumed to be due to the different roles of rootstocks in the translocation of photoassimilates to the fruits (Gardner, 1969) and the movement of water and sucrose to the juice sacs (Castle, 1995). Because these factors also affect the vegetative growth of a tree, the inverse correlation found between TSS and the vigor-inducing capacity of rootstock is expected. In addition to rootstock cultivar, climatic factors (Davies, 1997) and diseases such as HLB can affect the citrus fruit quality (Baldwin et al., 2018; Liao and Burns, 2012) and interact with the rootstock (Dala-Paula et al., 2019). Two of the rootstocks (‘ES-4’ and ‘ES-6’) at the Lake Wales location induced a very low acid content, thereby increasing the TSS-to-acid ratio to more than 24. In general, acid ratios were very low in the 2019–20 production season. Unusually high temperatures and excessive rainfall throughout the year, particularly during the preharvest period, may have been responsible for this. The negative influence of warm temperatures on TSS and acid content was previously highlighted (Reuther, 1980; Zekri, 2011).

It is accepted that rootstocks differ in their capacity to uptake nutrients from the soil and distribute them through the scion (Brown et al., 1994; Toplu et al., 2012; Wutscher, 1973). In our study, the rootstock effect was not significant for most of the leaf nutrients and not consistent across both trials. Overall, most of the nutrients were in the range deemed optimal or high for citrus (Kadyampakeni and Morgan, 2020). The high Cu concentration measured in trees at both locations are the result of frequent Cu applications to control citrus canker, which is widespread in Florida (Dewdney et al., 2001; Zambon et al., 2019).

Conclusions

The diploid and tetraploid rootstocks investigated during this study differed in their effects on the ‘Hamlin’ scion. Among the most prominent effects were those on tree size and productivity. The early production of high-quality fruits and high productivity comprise one strategy that can result in growing citrus profitably under HLB-endemic conditions. Many of the tetraploid rootstocks reduced tree size significantly and consequently reduced yield; however, they increased yield efficiency. Therefore, these rootstocks may be suitable for high-density plantings to increase orchard productivity during the early production years. However, many of the small size-inducing rootstocks rendered the trees more vulnerable to tropical force winds than the large size-inducing rootstocks. Although this study was conducted during the early production years, the results are valuable for predicting rootstock resilience and their economic potential in the longer term.