Investigation of stem anatomy in relation to hydraulic conductance, vegetative growth and yielding potential of ‘Summit’ cherry trees grafted on different rootstock candidates

The trial was carried out in Rimski Šančevi, at the experimental farm of the Faculty of Agriculture, University of Novi Sad, in northern Serbia (45°20′ N and 19°50′ E, altitude 80 m above sea level). The climate of this area is temperate continental, with extremely warm summers and cold winters. During the experimental years 2017–2020, mean annual temperatures were 12.6 °C, 13.2 °C, 13.4 °C and 12.8 °C, with annual precipitation sums of 513 mm, 718 mm, 632 mm and 733 mm, respectively. Seven ungrafted rootstocks were investigated during the 4 years, including five cheery rootstocks belonging to P. fruticosa (Pall.) and P. cerasus (L.) ecovar. ‘Oblačinska’, one ‘Mahaleb’ (Prunus mahaleb) genotype PM_09_01 and ‘Gisela 5’ were used as a control rootstock (Table 1). Concomitantly, the investigation included mid- to late-maturing sweet cherry cultivar Summit grafted on the same rootstocks. P. cerasus genotypes were PC_05_04 and PC_02_01/4, while European ground cherry (P. fruticosa) genotypes were PF_02_16, PF_01_01 and PF_04_09. Trees were planted in autumn 2015, with a planting distance 4 m × 2 m. All investigated plants were equally managed, in conditions without irrigation and with minimal herbicide usage. The trial was located on flat terrain exposed to the cold winds, without a frost protection strategy implemented. The trees were not pruned during the experiment.

One year after orchard establishment, in 2016, an initial anatomical characterisation of rootstocks and scion stems was conducted. Both ungrafted rootstock mother trees, as well as grafted ‘Summit’ trees, were cut back to 10 cm to obtain appropriate cross-sections. Anatomical measurements were performed on seven ungrafted rootstocks and the same seven rootstocks – ‘Summit’ scion combinations. During the winter dormancy period (November 2016), 1-year-old main stems were collected from five replicate plants per rootstock/rootstock-scion combination (five samples per variety). To preserve the structure of plant tissues, samples were fixed in 60% ethanol, with the addition of 10% of glycerin. Before cross-section preparation, plant material was immersed in 50% glacial acetic acid for 1–2 days, to soften stems and facilitate sectioning. Cross-sections were obtained using a hand microtome and examined under a Motic Digital BA310 biological light microscope with a builtin digital camera. Images of cross-sections were taken at 40× and 400× magnifications and measurements were performed using the image analysing system Motic Images Plus 2.0; Motic China Group Co., Ltd. Cross-sections were obtained from the third to fifth internode region (approximately the middle) of 1-year-old stems, 1.5–6.5 mm in diameter, as described by Ljubojević et al. (2021). On each stem cross-section, anatomical characteristics were investigated on four radial segments, 90° apart. Anatomical analyses included both total cross-section and secondary wood measurements. Pith radius, secondary wood outer radius, secondary cortex outer radius and periderm outer radius, which is equal to the stem radius, were measured to calculate stem diameter (SD, mm), cross-section area (CSA, mm²), areas and the percentage of the pith (% P), secondary wood (% SW), secondary cortex (% SC) and periderm (% PD), relative to the total CSA. Stem cross-section was considered a regular circle as indicated by Hajagos and Végvári (2013). Secondary wood/secondary cortex ratio (SW/SC ratio) was calculated for all investigated stems, by dividing the secondary wood area (SWA) with the area of the secondary cortex. At higher magnification, 1–2 visual fields were selected for measurement on each radial segment, depending on the stem thickness. The number and lumen areas of all vessels captured in the field of vision were measured, to calculate average VLA (μm²) and VF (number of vessels per mm²). Total vessel and ray areas, as well as xylem area per each visual field, were measured and calculated and their percentages relative to CSA and SWA were determined. For further hydraulic conductance (k_h) calculation, vessel diameters (VD, μm) were determined from the VLA values [VLA = (VD/2)² π], by assuming the vessels were circular. Based on their lumen area, vessels were classified into three classes: I – VLA <300 μm²; II – VLA in the range 300–700 μm²; and III – VLA >700 μm². The number of vessels per class was expressed as a percentage of the total number of vessels (100%).

During the 4 consecutive years, trunk cross-section measurements of grafted sweet cherry trees were conducted. Small segments of the trunk were cut out carefully using a wood chisel, 10 cm above the graft union. Each year, samples were taken out from the different radial segments of the trunk, so as not to disturb the tree growth in the following years as a result of harming the plant tissues. Sampling was carried out in the winter dormancy period (December), on five replicate plants per scion-rootstock combination. Obtained trunk segments were preliminarily screened for VLAs and frequency, to ensure the accuracy of the determined vessel size ranges and calculations based on entire stem cross-sections sampled in 2016. Trunk diameter 10 cm above grafting point (mm) and trunk CSA (mm²) were measured and calculated. Knowing those values, trunk SWA (mm²) was calculated by measuring the periderm and secondary wood ring thickness and calculating the secondary wood inner and outer radius.

Based on the initial anatomical characterisation in the first year after tree planting, the theoretical axial hydraulic conductance (k_h) per mm² of the stem was determined. The theoretical axial hydraulic conductance of stems was calculated according to the expression given by Tyree and Ewers (1991), based on Hagen-Poisseuille's law: kh=π⋅ρ128⋅η∑i=1ndi4 {k_h} = {{\pi \cdot \rho } \over {128 \cdot \eta }}\sum\limits_{i = 1}^n {d_i^4} where d was the diameter of the vessels in meters, ρ was the fluid density (assumed to be 10³ kg · m⁻³ for water at 20 °C) and η was the viscosity (assumed to be 1.002 × 10⁻⁹ · s for water at 20°C).

Total trunk theoretical axial hydraulic conductance (k_h(T)) was calculated from the stem k_h per mm² and trunk SWA values. Because trunk tissues represent the last barrier to the plant water transport from the roots to the plant's upper parts, trunk k_h was obtained rather than rootstock k_h.

Between 2017 and 2020, the investigated plants of all seven scion-rootstock combinations were measured to estimate the effective tree crown volume (V_e) for each tree. Crown height (cm) and crown diameter (cm) measurements were used in the tree crown volume calculation, according to the formula given by Changok (2007): CD2×CH×crown shape index=tree crown volume {\rm{C}}{{\rm{D}}^2} \times {\rm{CH}} \times {\rm{crown}}\;{\rm{shape}}\;{\rm{index}} = {\rm{tree}}\;{\rm{crown}}\;{\rm{volume}} where CD was crown diameter and CH was crown height. A crown shape index was assigned to every plant as described by the same author.

In the period 2019–2021, during the flowering time, the total number of flowers per tree was determined. The manual counting was carried out on three plants per scion-rootstock combination and the average number per tree was calculated. The year 2018 was excluded from the analysis, due to the very scarce number of flowers per tree (the very first year of fruit occurrence).

The obtained data were statistically processed by analysis of variance and correlation analysis, using STATISTICA 14 software (Tibco, USA). The significance of differences between mean values was determined using Duncan's multiple range tests with the confidence of p ≤ 0.05.

Among the investigated ungrafted rootstocks, significant differences in rootstock stem anatomical properties were observed (Table 2). Regarding SD and CSA, significant variation was found between and within species. Among species, PM_09_01 and the control rootstock ‘Gisela 5’ had the greatest SD, while P. cerasus genotypes had significantly thicker stems in comparison with P. fruticosa genotypes. SD in P. cerasus genotypes was within the range of 2.65–2.88 mm, while rootstock stems of P. fruticosa genotypes varied from 1.61 mm to 2.12 mm. The percentage of secondary wood was lower than % SC in all P. cerasus and P. fruticosa genotypes and did not differ significantly among the two species. On contrary, in both PM_09_01 and ‘Gisela 5’,% SW was higher than % SC, reaching 44.94 and 43.77%, respectively. Such tissue distribution was reflected in a significantly lower SW/SC ratio in analysed sour and ground cherry genotypes, in comparison to ‘Gisela 5’ and PM_09_01 that reached values of 1.08 and 1.19, respectively. P. fruticosa genotypes showed greater % PD than P. cerasus genotypes, with the highest value of 12.37% in PF_02_16. Total vessel area percentages (% V) were significantly higher in rootstock stems of PM_09_01 and the control than in P. cerasus and P. fruticosa genotypes, whose % V varied from 3.52% to 4.68%. The CSA was significantly positively correlated with % V (r = 0.85, as shown in Table S1 in Supplementary Materials). Ray area percentages (% R) were also higher in PM_09_01 and ‘Gisela 5’ compared to other genotypes. Xylem portions (% X) varied between and within species, ranging from 17.53% to 29.49% in P. cerasus, and 20.79–30.86% in P. fruticosa genotypes.

Scion stems of a variety ‘Summit’ grafted on different rootstock candidates showed significant differences regarding examined anatomical features (Table 3). SD varied between trees grafted on rootstock candidates belonging to different species, with higher average values when grafted on P. cerasus (reaching up to 6.28 mm). Concerning % P, significant variation was found between scion stems on P. cerasus and P. fruticosa rootstock candidates, with higher pith portions of 14.73–16.92% in scion stems on P. cerasus. In general, % SW and % SC were higher on P. fruticosa than on P. cerasus rootstock candidates (reaching average values of 34.3 and 47.79%, respectively), due to notably higher percentage of pith area in P. cerasus. For all scion-rootstock combinations, SW/SC ratio has shown that secondary cortex portions were higher than portions of secondary wood in scion stems, with 13% difference on average. CSA was significantly negatively correlated with % SC in scion stems (r = −0.93; Table S2 in Supplementary Materials). No rootstock species-related pattern was observed concerning % V, while regarding % R the highest values were found in scion stems of trees grafted on P. fruticosa (7.54–8.93%). A significant positive correlation of CSA was found with the % V (r = 0.92; Table S2 in Supplementary Materials). Both % R and % X was significantly higher in scion stems in comparison with rootstock stems. Scion stems on all investigated rootstock candidates had a lower % X, when compared to the control. In general, scion stems were approximately 2-fold thicker than rootstock stems. According to the mean values, portions of the secondary wood, as well as the secondary cortex, were similar in both rootstock and scion stems, accompanied by higher % P and lower % PD values in scions.

Within the secondary wood ring, differences between genotypes were determined, both for the rootstock plants and grafted ‘Summit’ trees (Tables 4 and 5, Figure 1). SWA varied from 0.78 mm² to 5.42 mm² in rootstock stems; in scion stems values fell within a range of 6.48–10.66 mm². Regarding rootstock stem analysis, SWA followed CSA values, resulting in the highest SWA observed in PM_09_01. Rootstock stems of P. cerasus genotypes had higher SWA than P. fruticosa genotypes, with average values of 1.93 and 0.97 mm², respectively. The percentage of vessels in secondary wood (% V_SW) differ significantly between P. cerasus and P. fruticosa genotypes as well as within species, while the highest % V_SW was found in P. mahaleb genotype and the control (up to 18.99% in PM_09_01). Genotypes belonging to P. cerasus formed one homogenous group with ray portions of >10%, while the lowest values were found in three P. fruticosa genotypes (ranging from 5.89% to 8.54%). Consequently, the average percentage of xylem in secondary wood (% X_SW) was higher in P. fruticosa genotypes compared to P. cerasus and P. mahaleb rootstock candidates, as well as when compared to control.

Figure 1

Concerning investigated scion-rootstock combinations, maximal % V_SW was achieved in trees grafted on P. cerasus rootstock candidates (15.95%, on average while the average % V_SW in scion stems on P. fruticosa was lowest (13.71%). The percentage of rays in secondary wood was approximately 2-fold higher in scion stems compared to rootstock stems, reaching up to the average value of 22.03%. Such partitioning resulted in the lower % X_SW in scion stems that accounted for about 63.35% of total SWA, on average.

Average VLAs in rootstock and scion stems significantly varied between and within species. In rootstocks, VLA was significantly positively correlated with stem CSA (r = 0.98) and %V (r = 0.86; Table S1 in Supplementary Materials). Generally, in P. cerasus and P. fruticosa genotypes, lower VLA values were found compared to P. mahaleb genotype and the control. Between species, VLA in P. cerasus was higher than in P. fruticosa genotypes, with variation from 264.65 μm² to 332.20 μm² in the former and from 211.81 μm² to 241.58 μm² in the latter. The size of the CSA was not significantly correlated with VF, both in rootstock and scion stems (Tables S1 and S2 in Supplementary Materials). Regarding the average VF in rootstock stems relative to rootstock species, a greater number of vessels was found in those genotypes that had lower VLA values. Thus, VF in P. fruticosa genotypes was 16% higher than average VF in P. cerasus. Consequently, P. mahaleb genotype, which was characterised with the highest VLA, had the lowest VF – about 33% lower than in P. fruticosa, which had 580.17 vessels per mm² on average. In scion stems a significant negative correlation between VLA and VF was found (r = −0.95; Table S2 in Supplementary Materials). Generally, scion stems were characterised by larger average VLA and lower VF values in comparison to the rootstock stems. On average, scion stems on P. cerasus and P. fruticosa recorded the VLA increase and VF decline when compared to rootstock stems of the same genotypes, with the lower average VLA of 257.83 μm² and the greater VF in scion stems on P. fruticosa rootstock candidates, counting 541.28 vessels per mm² in average. In scions, % V_SW was significantly positively correlated with VLA (r = 0.91) but negatively correlated with VF (r = −0.77; Table S2 in Supplementary Materials). In comparison to the control, scion stems on P. cerasus and P. mahaleb rootstock candidates had significantly larger vessels, ranging from 368.44 μm² to 390.55 μm², while VF was similar according to Duncan's multiple range tests.

Portions of vessels per size category are shown in Figure 2. Between rootstock candidates, two patterns of vessel distribution considering categories I and II were observed (Figure 2A). All cherry genotypes had a higher percentage of vessels <300 μm² in comparison with the percentage of vessels belonging to group II. In P. cerasus genotypes, portions of smallest vessels were 52.37% in PC_05_04 and 63.30% in PC_02_01/4, while in P. fruticosa genotypes group I percentages varied from 71.57% in PF_01_01 to particularly high 79.98% in PF_02_16. Consequently, the lowest percentages of vessels from group II were observed in P. fruticosa genotypes, ranging from 19.87% to 27.91%. The second pattern regarding vessel distribution was observed in PM_09_01 rootstock stems, where the highest number of vessels belonged to class II (58.07%) while small and large vessels were similarly present. ‘Gisela 5’ was characterised by the higher number of vessels sized within the range of 300–700 μm² (50.23%), but vessels from class I were also highly present in the rootstock stem secondary wood (46.27%). The share of vessels >700 μm² in the total number of vessels was lowest in all investigated rootstocks. Percentage of largest vessels varied from 0.14% in PF_02_16 to 18.67% in PM_09_01.

Figure 2

Among scion stems, portions of vessels with different VLA did not match strictly patterns observed in rootstock candidates (Figure 2B). Concerning vessel affiliation to different size categories in scion stems of trees grafted on P. fruticosa genotypes where lowest V_e was observed, the highest percentage of vessels <300 μm² (60.08–73.56%), lowest portions of medium vessels (25.22–36.55%) and extremely low percentage of vessels >700 μm² (1.16–3.37%) were found. On the contrary, ‘Summit’ trees of highest V_e, grafted on P. cerasus genotypes, had a similar percentage of small (<300 μm²) and medium vessels (300–700 μm²) in their scion stems, as well as a higher percentage of large vessels (>700 μm²) compared to the previous group of trees, reaching 8.64% for PC_02_01/4 and 12.61% for PC_05_04. In contrast with the vessel categories’ percentages in rootstock stems, PM_09_01 and ‘Gisela 5’ values in scion stems showed larger portions of vessels <300 μm² (47.68 and 58.91%, respectively), while vessels within a group III had the smallest share for both rootstocks (10.55% and 4.23%, respectively).

The CSA was significantly negatively correlated with the percentage of vessels under 300 μm², both in rootstock (r = −0.99) and scion (r = −0.79) stems (correlation coefficients are shown in Tables S1 and S2 in Supplementary Materials). In rootstocks, portions of vessels belonging to the groups II and III were in a significant positive correlation with stem size (correlation coefficients were 0.96 and 0.91, respectively), while in scions significant positive correlation was found with the percentage of medium-sized vessels (r = 0.89). Significant correlations of VF were found in scion stems with vessels percentage from I (r = 0.93), II (r = −0.87) and III (r = −0.93) size class.

Results showed a statistically significant correlation between V_e, k_h(T) and yielding potential of grafted trees, with the highest alteration from the correlation curve recorded for trees grafted on PM_09_01 and ‘Gisela 5’ (Figure 3). Effective tree crown volume of sweet cherry ‘Summit’ trees was strongly influenced by the rootstock selected for grafting (Table 6). Relative to the V_e of trees grafted on control (100%), the estimated V_e in the first experimental year was lower for trees grafted on the following rootstock candidates: PF_02_16 (22%), PF_04_09 (26%), PF_01_01 (48%) and PM_09_01 (35%). ‘Summit’ trees on P. cerasus rootstock candidates had higher V_e in comparison with the control trees, reaching 104% on PC_05_04 and 204% on PC_02_01/4, of V_e value achieved on control. During the 4 years, trees grafted on PC_02_01/4 had the highest crown volume values, from 0.47 m³ in the first year of measuring to 2.97 m³ in 2020. Trees grafted on P. fruticosa genotypes showed lower V_e than on P. cerasus genotypes, with crown volumes within the range of 0.86–1.37 m³ in the fifth year after planting. Although trees on ‘Gisela 5’ achieved similar values as trees on PC_05_04 in 2017 (0.23 m³ and 0.24 m³, respectively) and had higher V_e in the following year, during 2019 and 2020 this genotype influenced higher crown volumes in comparison with the control. Trees with lower V_e determined in previous years obtained higher percentages in 2020 in comparison with trees grafted on ‘Gisela 5’, with an estimated 44% (PF_02_16), 56% (PF_04_09), 61% (PF_01_01) and 65% increase (PM_09_01), while P. cerasus rootstock candidates induced values of 122% and 151% of V_e achieved on ‘Gisela 5’ (PC_05_04 and PC_02_01/4, respectively). Sweet cherry trees on PM_09_01 had lower V_e than ‘Gisela 5’ during all experimental years, reaching 0.08 m³ in 2017 and 1.29 m³ in 2020.

Figure 3

Trunk calculated hydraulic conductivity varied among grafted trees, with maximal calculated values on P. cerasus genotypes and minimal values on P. fruticosa genotypes during the 4 consecutive years. In 2017 trunk k_h values on PC_05_04 and PC_02_01/4 were approximately 3-fold higher than k_h(T) achieved on P. fruticosa genotypes, and about 2-fold higher compared to PM_09_01 and the control. During 2017–2020 maximal values were found on PC_02_01/4, reaching maximal 51.07 × 10⁻⁴ kg · m · MPa⁻¹ · s⁻¹ in 2020. Results obtained in 2020 varied from 8.59 × 10⁻⁴ kg · m · MPa⁻¹ · s⁻¹ in trees on PF_02_16 to previously listed highest value on PC_02_01/4 genotype, where control rootstock influenced 2.5-fold lower trunk conductance than P. cerasus genotypes and 1.5-fold higher k_h(T) when compared to P. fruticosa genotypes. The third highest value in the last year of measuring was found in trees grafted on PM_09_01 (27.97 × 10⁻⁴ kg · m · MPa⁻¹ · s⁻¹), pointing to the approximately 1.5-fold greater k_h(T) in comparison to control, during the 4 consecutive years. Trees grafted on P. cerasus genotypes had significantly higher k_h(T) compared to trees on other rootstock candidates, with about 4-fold and 1.5-fold higher values than trees on P. fruticosa and P. mahaleb in 2020, respectively.

The effective tree crown volumes in the years 2017, 2019 and 2020 were significantly positively correlated with k_h(T) for the same years (coefficients of correlation were 0.86, 0.86 and 0.89, respectively, as shown in Table S3 in Supplementary Materials).

There were significant differences in the number of flowers per tree determined in trees grafted on differing rootstocks (Figure 4). In 2019, P. cerasus genotypes PC_05_04 and PC_02_01/4 induced the highest flowering of sweet cherry trees, forming a homogenous group with 723.76 and 741.59 flowers per tree, respectively. Plants grafted on control rootstock ‘Gisela 5’ counted 614 flowers per tree, while those grafted on PM_09_01 counted 282.89 flowers. P. fruticosa genotypes were characterised with only 25.68 to 43.60 flowers per tree. In comparison with 2019, the number of flowers in 2020 increased by 23% on PC_05_04 and 38% on PC_02_01/4, 6–15 times on P. fruticosa rootstock candidates, 50% on PM_09_01 and only 12% on control. Values from 2020 and 2021 showed that P. cerasus genotypes PC_05_04 and PC_02_01/4 achieved the highest numbers of flowers per tree, reaching up to 889.60 and 1023.40 flowers per tree in 2020 and 1,112.18 and 1,083.69 flowers per tree in 2021, respectively. In 2021, the largest increase in the number of flowers per tree compared to the previous year was noted on trees on P. fruticosa rootstock candidates (1.7-to 3.3-fold higher) and on PM_09_01 (2.6-fold higher). Yielding potential on P. cerasus rootstock candidates and control was raised by 25%, 8% and 44% (PC_05_04, PC_02_01/4 and ‘Gisela 5’, respectively), which led to a similar number of flowers on P. fruticosa and control rootstock, as well as on PC_02_01/4 and PM_09_01. Trees grafted on PF_02_16, PF_01_01 and PF_04_09 achieved the lowest yielding potential, counting 791.72, 919.05 and 918.90 flowers per plant, respectively. A significant positive correlation of yielding potential was found with the V_e and k_h(T) values in 2019 and 2020 (correlation coefficients varied between 0.84 and 0.96; Table S3 in Supplementary Materials).

Figure 4