Initial Compatibility Indicators of Four Coffea arabica Cultivars on Coffea canephora Rootstock

Abstract

Grafting is a strategy to mitigate biotic and abiotic stresses in coffee systems. However, initial compatibility indicators between Coffea arabica scions and C. canephora rootstocks under controlled conditions remain insufficiently documented. We evaluated the physiological and morphological compatibility of four C. arabica cultivars (Bourbon, Geisha, Catuai, and Villa Sarchí) grafted onto C. canephora (Robusta) rootstock in a tropical highland nursery in the Peruvian Amazon. Seven physiological and six morphological variables were measured. Statistical analyses included one-way ANOVA and Kruskal–Wallis tests with Bonferroni post hoc comparisons. Two physiological parameters were significantly higher in Villa Sarchí grafts than in Robusta: PSII quantum yield (+0.044 units; p < 0.05) and electron transport rate (+14.702 µmol e⁻ m⁻² s⁻¹; p < 0.05); by contrast, net photosynthesis, stomatal conductance, intercellular CO₂ concentration, and transpiration did not differ, and maximum PSII efficiency was similar among treatments (p = 0.509). Conversely, no morphological trait showed significant differences, and graft-take was high across all combinations. The results support the use of Coffea robusta as a rootstock for these four cultivated varieties, thereby offering the possibility of improving their resilience in tropical highland regions.

1. Introduction

Coffee is one of the most important tropical crops, underpinning rural economies in developing countries across the humid tropics [1]. It holds major cultural and social significance worldwide, being cultivated in more than 60 countries by approximately 25 million farmers—most of them smallholders dependent on this crop for their livelihoods [2]. In Peru, coffee is the leading agricultural export, supporting over 223,000 producing families, especially in Andean Amazon regions [3,4]. However, production faces critical challenges: soilborne pests and diseases, soil degradation from long-term cultivation, and climate change effects. Plant-parasitic nematodes, particularly root-knot nematodes of the genus Meloidogyne, cause severe yield losses and plant death in Coffea arabica L. across multiple regions [5]. Conventional chemical management in intensified monocultures exacerbates these issues, creating impoverished, pathogen-laden soils that hinder new plantations [6]. Additionally, coffee leaf rust and climatic extremes—such as recurrent drought and high temperatures—are reducing productivity and the suitable area for C. arabica in Latin America [7,8]. These challenges underscore the need for agronomic innovations that balance productivity and sustainability in tropical coffee systems.

Within this context, grafting constitutes a potential strategy to mitigate several of the above problems. The practice has been widely used in fruit trees and other perennial crops, such as Citrus spp. [9], Persea americana Mill. [10], and Theobroma cacao L. [11]. Grafting allows combining a target cultivar (scion) with a rootstock possessing vigorous roots and resistance to adverse factors, thereby optimizing crop performance and adaptability. In coffee, grafting C. arabica cultivars onto Coffea canephora Pierre (Robusta) rootstocks has emerged as a solution to manage soil pests and increase plantation resilience [12]. Specifically, some C. canephora provide valuable resistance; for example, the Robusta cultivar “Nemaya” shows high resistance to root-knot nematodes and is used as a rootstock in infested soils in Central America [5]. Field research has demonstrated notable benefits of grafting in coffee. In Hawai‘i, ref. [5] reported that ungrafted C. arabica “Kona Typica” suffered up to 81% mortality under Meloidogyne infestation, whereas plants grafted onto resistant rootstocks—including Robusta “Nemaya” and Coffea liberica Hiern. accessions—showed greater vigor, survival, and significantly higher yields. Similarly, in Central America, productivity gains have been obtained in C. arabica grafted onto Robusta, attributed to effective control of “coffee decline” caused by nematodes [5]. In addition, integrated nursery approaches combining grafting with bioinoculants have yielded positive results. In Vietnam, for instance, ref. [6] found that grafting C. canephora onto C. liberica, combined with mycorrhizal symbiosis, significantly reduced the density of Pratylenchus coffeae and Meloidogyne incognita in soil, thereby improving seedling survival and growth. This body of evidence supports the use of coffee grafting as an effective tool for managing soil nematodes and other adverse biotic factors.

Beyond resistance to root pests, coffee grafting is also being explored for its effects on tolerance to abiotic stresses and foliar diseases. Under drought, C. arabica grafted onto Robusta rootstocks has shown better physiological responses than ungrafted plants. For example, ref. [7] demonstrated that in C. canephora, reciprocal grafting of a drought-sensitive clone onto a tolerant rootstock increased abscisic acid synthesis in leaves under water stress, inducing timely stomatal closure that attenuates water loss. Consistently, other studies have observed that grafted plants have higher relative leaf water content, lower cellular electrolyte leakage during drought, and faster recovery after rehydration compared to ungrafted plants [13]. Regarding foliar diseases, there are indications that the rootstock influences the scion response. A study in Mexico evaluated the interaction with coffee leaf rust (Hemileia vastatrix), in which grafting the susceptible variety “Garnica” onto the resistant variety “Oro Azteca” significantly reduced fungal colonization of scion leaves and increased expression of defense genes in foliar tissue [8]. These findings suggest that a resistant rootstock can transmit systemic resistance signals to the scion, thereby reinforcing the plant’s capacity to defend against pathogens. Scientific advances in the last decade have demonstrated the potential benefits of grafting in coffee, including nematode control and improved root health, increased drought tolerance, and mitigation of the impact of devastating diseases such as leaf rust, among others.

Despite the above, knowledge gaps remain regarding coffee grafting. Not all interspecific or inter-cultivar combinations are compatible. Grafting success can depend strongly on the specific genetics involved and local environmental conditions. There is documentation of incompatibility in interspecific grafts; for example, Robusta grafted onto the C. liberica variety “Excelsa” has shown deficient graft unions and high failure rates in trials in Indonesia and Vietnam [6]. Even in the most common combination—C. arabica onto C. canephora—results have not always been consistent. A recent study in Costa Rica found that although the arabica–Robusta graft did not present physical vascular blockages at the union, grafted plants exhibited reduced vegetative growth and productivity under shade compared with ungrafted plants, especially when a vigorous, high-yielding arabica clone was used [12]. Such discrepancies underscore that the compatibility and effectiveness of grafts can vary according to the cultivar combination and the production environment. To date, no studies have been published on the compatibility of popular C. arabica cultivars grafted onto C. canephora in high-altitude tropical regions such as the Peruvian Amazon. Limited knowledge of how specific genotypes interact at the nursery stage remains a barrier to informed adoption by nurserymen and farmers. In particular, it is necessary to determine whether arabica scions grafted onto Robusta maintain normal growth and form strong unions in the short term, because poor compatibility could lead to low graft-take, delayed seedling development, or medium-term weakening of the tree.

Investigating graft compatibility contributes to understanding tissue-level and physiological interactions in interspecific grafts [14], an area that remains underexplored in tropical species. Specifically, there is a need to develop better-adapted rootstock cultivars derived from C. canephora var. Robusta [15]. This can generate knowledge for breeding programs focused on coffee rootstocks and other Rubiaceae. It is also practically relevant, since identifying compatible combinations can improve coffee renovation practices in areas affected by infested soils or stressful climates [16]. Thus, if certain elite C. arabica cultivars can be successfully established on resistant Robusta roots, farmers could replant coffee in fields with a history of nematodes or soil fatigue, reducing losses and extending the productive life of plantations. This is especially important in regions such as Amazonas, in the Peruvian Amazon, where coffee is often grown on degraded, suboptimal soils. Only about 11% of the land is classified as suitable for coffee cultivation [17], and intensive agronomic practices have led to soil depletion [18]. The use of Robusta rootstocks could confer greater root vigor, tolerance to high soil aluminum saturation, tolerance to limiting conditions such as short droughts, and improved nutrient uptake, contributing to more resilient production systems. Therefore, optimizing the grafting technique in coffee could consolidate a sustainable alternative to increase productivity and address phytosanitary and climatic challenges, complementing other agronomic strategies.

Accordingly, the objective of this research was to evaluate the initial compatibility indicators of four C. arabica cultivars grafted onto Robusta coffee rootstock under nursery conditions in the Peruvian Amazon. The study focused on determining graft-take, the formation and appearance of the graft union, and the initial growth of grafted plants compared to non-grafted plants for each scion/rootstock combination. In doing so, we identified potential early incompatibilities or differences in vigor among combinations. The findings of this work provide crucial information for the successful use of grafting in the coffee sector of the Peruvian Amazon and other tropical areas. They offer technical guidance for grafted coffee propagation aimed at improving plantation health and productivity in critical environments.

2. Materials and Methods

2.1. Study Site

The experiment was conducted from October 2024 to March 2025 under nursery conditions in the district of Ocúmal, Luya Province, Amazonas Region, Peru, specifically at La Hilda Farm (10°05′33.09″ N, 84°13′55.03″ W). Elevation was 1787 m.a.s.l. Minimum temperatures ranged from 6.17 to 13.65 °C; annual precipitation from 0 to 41.7 mm; relative humidity from 33.88% to 91.81%; and maximum temperatures from 16.12 to 28.72 °C.

2.2. Nursery Conditions

The nursery was built with bamboo, wooden slats, galvanized wire, nails, and Raschel shade cloth and was equipped with a micro-sprinkler irrigation system. Polybags were filled with a previously prepared substrate to favor root development.

The substrate consisted of vermicompost and sand (4:1), with pH 8.00, electrical conductivity 6.15 dS m⁻¹, 29.33% organic matter, cation exchange capacity (CEC) 79.13 meq 100 g⁻¹, total organic C 17.01%, total P 301.33 ppm, total K 6489.01 ppm, exchangeable K 41.6 meq 100 g⁻¹, exchangeable Ca 28.63 meq 100 g⁻¹, exchangeable Na 5.29 meq 100 g⁻¹, exchangeable Mg 17.14 meq 100 g⁻¹, and a sandy loam textural class.

2.3. Grafting Procedure and Nursery Management

Seedlings were initially propagated by controlled germination of certified seed, and a terminal wedge graft [19] was performed onto C. canephora rootstocks. The rootstocks are dug up, then grafted, and finally replanted. In this method, the rootstock seedling is first decapitated, and a vertical cut of 0.5 to 1.5 cm is made along the center of the stem’s axis and the cutting angle of the stem (approximately 30–35°). The scion wedge was carefully inserted into the slit, ensuring alignment with the vascular cambium on at least one side for proper vascular continuity. C. arabica scions were selected for diameter compatibility and secured with Parafilm^® to prevent desiccation. The stem is then pruned to leave between one and three true leaves. The lower end of the stem is cut diagonally on both sides to create a wedge shape. This wedge-shaped graft is inserted into the slit made in the rootstock. Although it may not be necessary, a film can be placed to secure the rootstock and graft it together. The rootstock holds the graft firmly in place. Budwood from C. arabica cultivars was grafted onto C. canephora rootstocks. C. arabica seedlings were selected at an optimal phenological stage—two to four fully expanded true leaves and a stem diameter of 2–3 mm—to ensure adequate vascular fusion. C. canephora rootstocks were selected at a semi-lignified stage with 3–5 mm stem diameter, appropriate for terminal wedge grafting, to ensure an effective union with the scion.

After grafting, a growth regulator (Fitaminas, Comercial Andina, Lima, Peru) was applied at a dosage of 4 mL per 2 L of water for rooting. This mixture was used for 5 min and has been tested in other research for the vegetative propagation of coffee. Additionally, a fungicide (Mertec, Syngenta Crop Protectión, Lima, Peru) was used as a disinfectant at a dose of 5 mL per 2 L of water, and the transplants were then placed in polybags. The substrate used for bagging (see Section 2.1) was pre-prepared and enriched to favor root development. Micro-sprinkler irrigation was applied at frequencies and volumes adjusted to seedling water requirements. Phytosanitary management included applications of Tifon NF (chlorpyrifos) and Parachupadera 740PM (captan and flutolanil) every 30 days. Fertilization was carried out with Basacote^® Plus 6M 16-8-12(+2+TE), Soltagro, Lima, Peru). Weed control was manual, every 30 days.

2.4. Experimental Design

A completely randomized design (CRD) was used with five treatments, three replicates, and 10 observations per treatment. In total, 150 out of 500 randomly selected plants were evaluated (n = 3 individual plants; experimental unit = plant). Treatments included four C. arabica cultivars grafted onto C. canephora rootstock under controlled nursery conditions: T1 (Bourbon), T2 (Geisha), T3 (Catuai), and T4 (Villa Sarchí). The control treatment (Tc) consisted of self-grafted C. canephora (Robusta) plants.

2.5. Measurements

Bark tissue samples were collected during the leaf-drop period from the stem of each plant. For each combination, samples were taken at the graft union—approximately 5–6 cm above (scion) and below (rootstock) the union. Bark tissue samples were collected once during the leaf-drop period from three standardized positions relative to the graft union. Samples were immediately frozen in liquid nitrogen, lyophilized, ground, and stored at −80 °C until further biochemical and physiological analyses. Three biological replicates per combination represented each treatment. Samples were collected from three positions: above the graft union, directly at the union, and below the union.

Table 1. Description of the physiological and morphological variables assessed in the grafted plants.

Variable Type	Variable	Abbreviation	Unit
Physiological	Transpiration	E	mmol H2O m−2 s−1
Physiological	Net photosynthesis (CO2 assimilation)	A	µmol CO2 m−2 s−1
Physiological	Intercellular CO2 concentration	Ci	µmol mol−1 (ppm)
Physiological	Stomatal conductance	Gs	mol H2O m−2 s−1
Physiological	PSII quantum yield (light-adapted)	ΦPSII (=ΔF/Fm′)	Dimensionless
Physiological	Electron transport rate	ETR	µmol e− m−2 s−1
Physiological	Maximum PSII efficiency (dark-adapted)	Fv/Fm	Dimensionless (0–1)
Morphological	Graft-take success	GT	%
Morphological	Plant vigor (visual score)	VI	Dimensionless (1–5)
Morphological	Plant height	H	cm
Morphological	Stem diameter	SD	mm
Morphological	Root length	RL	cm
Morphological	Leaf number	LN	Count

describes the variables evaluated in two categories and the analyses performed using these tissues.

Physiological measurements were conducted on fully expanded leaves located in the middle third of each plant. For gas exchange parameters (A, gs, Ci, E), three leaves from each plant were measured using the LI-6800 system (LI-COR, Lincoln, Ne, USA), under stable environmental conditions, with the average values per plant serving as the experimental unit. Chlorophyll fluorescence parameters (ΦPSII, ETR, Fv/Fm) were recorded using the fluorometry module of the LI-6800 and a Pocket PEA fluorimeter (Hansatech Instruments, Norfolk, UK) following a 30-min dark adaptation. Consequently, each treatment provided three independent values at the plant level for statistical analysis.

2.5.1. Physiological Measurements

Leaf gas exchange was used to assess plant physiological status [20]. Net photosynthesis (A) was estimated from the difference in CO₂ concentrations between the inlet and outlet air streams of the chamber, following standard principles for gas-exchange measurement [21].

Variables A, transpiration (E), intercellular CO₂ concentration (Ci), and stomatal conductance (gs) were recorded with a portable LI-6800 system (LI-COR Inc., Lincoln, NE, USA) equipped with the 6800-01A integrated fluorometry module. Spot measurements were taken on fully expanded leaves from the middle third of the canopy under stable environmental conditions, averaging three readings separated by 10 s per leaf.

Chlorophyll fluorescence in the light was measured with the LI-6800 fluorometry module to obtain PSII quantum yield (ΦPSII = ΔF/Fm′) and electron transport rate (ETR; µmol e⁻ m⁻² s⁻¹) as computed by the instrument. Maximum PSII efficiency (Fv/Fm) was determined after ~30 min dark adaptation using a pulse-modulated fluorimeter (Pocket PEA, ( Hansatech Instruments, Norfolk, United Kingdom)). Physiological measurements were taken once at the end of the nursery period from three leaves per plant, using the mean at the plant level as the experimental unit. For analysis, the per-plant mean (mean of three leaves) was used as the experimental unit.

2.5.2. Morphological Measurements

Six morphological traits were evaluated: graft-take (GT, %), plant vigor (VI, visual scale 1–5), plant height (cm), stem diameter (mm), root length (cm), and leaf number (count) [22,23]. GT was calculated as (number of successful grafts/number of grafts) × 100 [24]. Height, diameter, and root length were measured with a digital caliper, and leaf number was counted on fully expanded leaves and healthy stems, using consistent criteria across treatments [25]. VI was assigned by standardized visual inspection (1 = very low vigor; 5 = high vigor), considering symptoms of translocated incompatibility on leaves and shoots [26]. Morphological measurements were conducted at the same stage of evaluation as the physiological measurements. These measures were implemented only once during the evaluation stage and were not repeated over time, as the study focused on early compatibility rather than growth dynamics.

To clarify, the graft-take success percentage (GT%) was determined by considering all grafted plants within each treatment group, totaling 30 plants (10 plants per treatment multiplied by 3 replicates). The three plants selected from each treatment for physiological and morphological assessments were not included in the GT% calculation. This approach ensured that the percentage values were based on a sufficiently large sample size.

2.5.3. Evaluation of Graft Compatibility

To evaluate tissue formation at the graft union zone in Coffea spp. grafts, one representative graft per treatment was sampled. For each sample, longitudinal cuts (before and after the graft point) and transverse sections were obtained to analyze tissue continuity and differentiation. Transverse sections were sputter-coated with a thin layer of gold (Quorum, Quorum Technologies Ltd., Laughton, East Sussex, UK) to ensure electrical conductivity and examined with a scanning electron microscope (SEM; Zeiss EVO10, Oberkochen, Germany). The corresponding images are shown in Figure 1.

2.6. Statistical Analysis

Data were analyzed with Stata version 17 (see details in Appendix A—

Table A1. Raw data collected for the study variables—replicates by treatment.

Treatments (Rows)	Physiological Variables							Morphological Variables
	E	A	Ci	gs	ΦPSII	ETR	Fv/Fm	GT	VI	H	SD	RL	LN
Bourbon	1.123752897	3.984274709	290.4580393	0.062898392	0.084520555	28.49958031	0.364424593	90	8	29.5	1	17	10
Bourbon	0.639047161	4.037399842	205.5180999	0.034413877	0.061977736	20.8990093	0.402354943	60	6	30.5	1	17	10
Bourbon	0.869073551	4.958759573	226.3375723	0.047753703	0.073947513	24.93392262	0.560836479	80	6	28	1.5	16	10
Geisha	1.550164989	5.671193392	287.0069804	0.087095697	0.079241016	26.71976574	0.520732124	70	6	30.5	1	17	10
Geisha	2.106878068	5.228091207	321.9299015	0.121016557	0.064375663	21.70680134	0.459169592	80	6	34.5	1.5	18	12
Geisha	1.154428483	4.850286435	269.4868015	0.063430674	0.068264383	23.01818019	0.500553993	80	7	30	1	18	10
Catuai	0.829036925	4.00593231	251.3421375	0.045378028	0.085573852	28.85549999	0.418662675	100	8	33	1	17	10
Catuai	1.290259719	5.155256762	275.5585365	0.071115293	0.076971558	25.95407002	0.418868346	80	5	27.5	1	16	12
Catuai	1.084834272	4.736376986	258.3892841	0.056892448	0.054310095	18.31309286	0.490874538	70	4	29.5	1.5	17	10
Villa Sarchí	1.88282726	6.219163447	295.189858	0.104394982	0.103077989	34.75683433	0.520359559	80	6	31.5	1	18	8
Villa Sarchí	0.898521112	4.096647405	258.490047	0.048920878	0.072367708	24.40167208	0.479735213	80	7	29	1	16	8
Villa Sarchí	0.756809676	4.666786785	208.5954658	0.040554735	0.086510578	29.17299361	0.481120132	50	4	28.5	1	18	10
Robusta	0.890914651	3.838112233	261.2159768	0.046856829	0.031503506	10.62291316	0.489236688	100	4	28	1	17	10
Robusta	0.300200664	3.167561295	78.09329745	0.015960146	0.034711148	11.70457872	0.453610503	100	6	35	1	18	10
Robusta	0.593576237	4.109899238	184.6380714	0.031500367	0.06494328	21.89889904	0.439133942	100	8	28.5	1	16	10

Abbreviations: E = transpiration; A = net photosynthesis; Ci = intercellular CO₂ concentration; gs = stomatal conductance; ΦPSII = PSII quantum yield (light); ETR = electron transport rate; Fv/Fm = maximum PSII efficiency (dark-adapted); GT = graft-take; VI = plant vigor; H = plant height; SD = stem diameter; RL = root length; LN = leaf number.

and

Table A2. Summary of the statistical tests applied in data analysis.

Variable Abbreviation	Variable Type	Statistical Tests						Results Assessment				p-Value of the Recommended Test
		p-Value (Shapiro–Wilk)					p-Value (Levene)
		Bourbon	Geisha	Catuai	Villa Sarchí	Robusta		N	H	I	Recommended Test
E	Physiological	0.94393	0.81423	0.87974	0.22112	0.99260	0.24683812	Yes	Yes	Yes	ANOVA	0.0997
A	Physiological	0.09259	0.91236	0.70350	0.50137	0.54236	0.25325899	Yes	Yes	Yes	ANOVA	0.117
Ci	Physiological	0.45332	0.63848	0.54776	0.83242	0.82010	0.17076344	Yes	Yes	Yes	ANOVA	0.1464
gs	Physiological	0.93017	0.80432	0.88411	0.23080	0.99344	0.23117457	Yes	Yes	Yes	ANOVA	0.0934
ΦPSII	Physiological	0.93171	0.48695	0.51486	0.91301	0.16620	0.5401381	Yes	Yes	Yes	ANOVA	0.0429 *
ETR	Physiological	0.93195	0.48696	0.51499	0.91354	0.16620	0.54024316	Yes	Yes	Yes	ANOVA	0.0429 *
Fv/Fm	Physiological	0.34958	0.62507	0.00467	0.05731	0.54345	0.02156893	No	No	Yes	Kruskal–Wallis	0.5089
GT	Morphological	0.6369	−0.00005	0.63690	−0.00005		0.04918814	No	No	Yes	Kruskal–Wallis	0.1767
VI	Morphological	−0.00005	−0.00005	0.46326	0.63690	1.00000	0.44535999	No	Yes	Yes	Kruskal–Wallis	0.9185
H	Morphological	0.78045	0.19389	0.70173	0.29826	0.12232	0.17052693	Yes	Yes	Yes	ANOVA	0.8225
SD	Morphological	−0.00005	−0.00005	−0.00005			0.00368007	No	No	Yes	Kruskal–Wallis	0.8903
RL	Morphological	−0.00005	−0.00005	−0.00005	−0.00005	1.00000	0.45155505	No	Yes	Yes	Kruskal–Wallis	0.5142
LN	Morphological		−0.00005	−0.00005	−0.00005		0.00368007	No	No	Yes	Kruskal–Wallis	0.3764

Abbreviations: E = transpiration; A = net photosynthesis; Ci = intercellular CO₂ concentration; gs = stomatal conductance; ΦPSII = PSII quantum yield (light); ETR = electron transport rate; Fv/Fm = maximum PSII efficiency (dark-adapted); GT = graft-take; VI = plant vigor; H = plant height; SD = stem diameter; RL = root length; LN = leaf number. Notes: * p < 0.05, N = Normality; H = Homogeneity; I = Independence (satisfied by the CRD).

). One-way analysis of variance (ANOVA) was applied to determine significant differences among treatments for variables meeting assumptions of normality and homogeneity of variance, verified by Shapiro–Wilk and Levene tests, respectively. For variables that did not meet these assumptions, the non-parametric Kruskal–Wallis test was used. When parametric assumptions were met and significant effects were detected (p < 0.05), Bonferroni post hoc comparisons were performed [27].

3. Results

3.1. Physiological Parameters of the Grafts

The physiological parameters of five coffee varieties (C. arabica: Bourbon, Geisha, Catuai, Villa Sarchí; C. canephora: Robusta) grafted onto C. canephora were evaluated to compare their performance under nursery conditions.

Table 2. Summary statistics (means ± SD for ANOVA; medians and interquartile ranges for Kruskal–Wallis) and p-values for physiological parameters of coffee cultivars grafted onto C. canephora under nursery conditions.

Variable	Unit	Treatments [(Mean ± Standard Deviations) or Median (IQR: Q1–Q3)]					p-Value	Test
		T1 Bourbon	T2 Geisha	T3 Catuai	T4 Villa Sarchí	Tc Robusta
E	mmol H2O m−2 s−1	0.88 ± 0.24	1.60 ± 0.48	1.07 ± 0.23	1.18 ± 0.61	0.59 ± 0.30	0.0997	ANOVA
A	µmol CO2 m−2 s−1	4.33 ± 0.55	5.25 ± 0.41	4.63 ± 0.58	4.99 ± 1.10	3.71 ± 0.49	0.1170	ANOVA
Ci	µmol mol−1 (ppm)	240.77 ± 44.27	292.81 ± 26.70	261.76 ± 12.46	254.09 ± 43.46	174.65 ± 91.97	0.1464	ANOVA
gs	mol H2O m−2 s−1	0.05 ± 0.01	0.09 ± 0.03	0.06 ± 0.01	0.06 ± 0.03	0.03 ± 0.02	0.0934	ANOVA
ΦPSII	Dimensionless	0.07 ± 0.01	0.07 ± 0.01	0.07 ± 0.02	0.09 ± 0.02	0.04 ± 0.02	0.0429 *	ANOVA
ETR	µmol e− m−2 s−1	24.78 ± 3.80	23.81 ± 2.60	24.37 ± 5.45	29.44 ± 5.18	14.74 ± 6.22	0.0429 *	ANOVA
Fv/Fm	Dimensionless	0.40 (0.36–0.56)	0.50 (0.46–0.52)	0.42 (0.42–0.49)	0.48 (0.48–0.52)	0.45 (0.44–0.49)	0.5089	Kruskal–Wallis

Abbreviations: E = transpiration; A = net photosynthesis; Ci = intercellular CO₂ concentration; gs = stomatal conductance; ΦPSII = PSII quantum yield (light); ETR = electron transport rate; Fv/Fm = maximum PSII efficiency (dark-adapted). Notes: * p < 0.05, IQR = Interquartile Range. For pairwise comparisons with differences, the Bonferroni correction was applied (Table 3).

presents the means ± standard deviations for E, A, Ci, gs, ΦPSII, and ETR (ANOVA) and the medians with interquartile ranges (IQRs) for Fv/Fm (Kruskal–Wallis). Figure 1 shows the corresponding boxplots.

Physiological measurements also revealed treatment-dependent responses. Differences in ΦPSII and ETR suggest that Villa Sarchí may exhibit a more efficient photochemical functioning during the early phases of graft establishment. The controlled conditions of the nursery—designed to minimize stress variation—may also reduce the magnitude of physiological divergence among treatments at this early stage.

No significant differences were detected among treatments for E (p = 0.0997), A (p = 0.1170), Ci (p = 0.1464), or gs (p = 0.0934). By contrast, differences were detected for ΦPSII (p = 0.0429) and ETR (p = 0.0429) (Table 2). Bonferroni post hoc comparisons indicated that Villa Sarchí exceeded Robusta in both ΦPSII (Δ mean = 0.044; p = 0.039) and ETR (Δ mean = 14.702 µmol e⁻ m⁻² s⁻¹; p = 0.039) (

Table 3. Bonferroni-adjusted pairwise comparisons for variables with significant treatment effects.

Variable	Treatment Pair	Mean Difference	Adjusted p-Value
ΦPSII	Bourbon vs. Geisha	−0.003	1.000
ΦPSII	Bourbon vs. Catuai	−0.001	1.000
ΦPSII	Bourbon vs. Villa Sarchí	0.014	1.000
ΦPSII	Bourbon vs. Robusta	−0.030	0.290
ΦPSII	Geisha vs. Catuai	0.002	1.000
ΦPSII	Geisha vs. Villa Sarchí	0.017	1.000
ΦPSII	Geisha vs. Robusta	−0.027	0.441
ΦPSII	Catuai vs. Villa Sarchí	0.015	1.000
ΦPSII	Catuai vs. Robusta	−0.029	0.346
ΦPSII	Villa Sarchí vs. Robusta	−0.044	0.039 *
ETR	Bourbon vs. Geisha	−0.963	1.000
ETR	Bourbon vs. Catuai	−0.403	1.000
ETR	Bourbon vs. Villa Sarchí	4.666	1.000
ETR	Bourbon vs. Robusta	−10.035	0.290
ETR	Geisha vs. Catuai	0.559	1.000
ETR	Geisha vs. Villa Sarchí	5.629	1.000
ETR	Geisha vs. Robusta	−9.073	0.441
ETR	Catuai vs. Villa Sarchí	5.070	1.000
ETR	Catuai vs. Robusta	−9.632	0.346
ETR	Villa Sarchí vs. Robusta	−14.702	0.039 *

Abbreviations: ΦPSII = PSII quantum yield (light); ETR = electron transport rate. Notes: * p < 0.05, p < 0.01. Mean differences were computed as (Treatment 1–Treatment 2).

). In magnitude, ΦPSII for Villa Sarchí was ~100% higher than Robusta (0.09 ± 0.02 vs. 0.04 ± 0.02), and ETR was ~100% higher (29.44 ± 5.18 vs. 14.74 ± 6.22), whereas Fv/Fm did not differ among treatments (p = 0.5089).

Although A, gs, Ci, and E did not reach statistical significance, there was a (non-significant) tendency toward higher physiological values in the arabica grafts compared with the Robusta control (e.g., A in Geisha = 5.25 ± 0.41 vs. A in Robusta = 3.71 ± 0.49 µmol CO₂ m⁻² s⁻¹). Given the per-treatment sample size (n = 3), these patterns should be considered exploratory.

3.2. Morphological Characteristics of the Grafts

Morphological traits were assessed to identify differences in early growth and development.

Table 4. Summary statistics (means ± SD for ANOVA; medians and interquartile ranges for Kruskal–Wallis) and p-values for morphological traits of coffee cultivars grafted onto Coffea canephora.

Variable	Unit	Treatments [(Mean ± Standard Deviations) or Median (IQR: Q1–Q3)]					p-Value	Test
		T1 Bourbon	T2 Geisha	T3 Catuai	T4 Villa Sarchí	Tc Robusta
GT	%	80 (60–90)	80 (70–80)	80 (70–100)	80 (50–80)	100 (100–100)	0.1767	Kruskal–Wallis
VI	Dimensionless (1–5)	6 (6–8)	6 (6–7)	5 (4–8)	6 (4–7)	6 (4–8)	0.9185	Kruskal–Wallis
H	cm	29.33 ± 1.26	31.67 ± 2.47	30.00 ± 2.78	29.67 ± 1.61	30.50 ± 3.91	0.8225	ANOVA
SD	mm	1.0 (1.0–1.5)	1.0 (1.0–1.5)	1.0 (1.0–1.5)	1.0 (1.0–1.0)	1.0 (1.0–1.0)	0.8903	Kruskal–Wallis
RL	cm	17 (16–17)	18 (17–18)	17 (16–17)	18 (16–18)	17 (16–18)	0.5142	Kruskal–Wallis
LN	Count	10 (10–10)	10 (10–12)	10 (10–12)	8 (8–10)	10 (10–10)	0.3764	Kruskal–Wallis

Abbreviations: GT = graft-take; VI = plant vigor; H = plant height; SD = stem diameter; RL = root length; LN = leaf number. Note: IQR = interquartile range.

shows the means ± standard deviations for H (ANOVA) and medians with IQRs for GT, VI, SD, RL, and LN (Kruskal–Wallis), along with boxplots illustrating the distribution of these variables by treatment.

No significant differences were observed among treatments for GT (p = 0.1767), VI (p = 0.9185), H (p = 0.8225), SD (p = 0.8903), RL (p = 0.5142), or LN (p = 0.3764). GT was high across all combinations (medians of 80% for arabica grafts and 100% for the self-grafted Robusta), with no statistically verifiable differences. Mean plant height was ~30 cm across treatments, consistent with homogeneous early growth under nursery conditions. The overall graft-take percentage (GT%) showed clear differences among treatments, with Villa Sarchí and Caturra exhibiting the highest compatibility. These results must, however, be interpreted considering the controlled nursery conditions and the relatively short evaluation period. Under such conditions, graft failure tends to occur rapidly, while successful unions remain stable with limited external influences. Therefore, the GT% observed reflects early compatibility responses rather than long-term field performance.

In contrast, morphological variables did not show significant differences among grafted plants. This result is consistent with the early evaluation stage of the trial; morphological differentiation often requires longer growth periods to become detectable. The absence of differences should therefore not be interpreted as evidence of long-term equivalence among graft combinations, but rather as an indication that early morphological development remains buffered under uniform nursery conditions.

Together, these results provide an initial evaluation of early compatibility in Coffea arabica grafted onto Coffea canephora, while also acknowledging that the short-term, controlled nature of the study inherently limits the extent of physiological and morphological divergence detectable at this stage.

The interpretation of physiological variables is directly related to the goal of evaluating early compatibility between Coffea arabica scions and C. canephora rootstock. The higher values of ΦPSII and ETR observed in the Villa Sarchí × Robusta combination indicate more efficient photochemical functioning during the initial physiological integration of the graft components, suggesting an early indication of compatibility. Additionally, the lack of significant differences in A, gs, Ci, and E is consistent with the brief evaluation period and the uniform conditions in the nursery, which tend to minimize variation and limit the emergence of physiological differences at this early stage.

3.3. Graft Compatibility: Scion–Rootstock Union

Compatibility was evaluated macroscopically using images of the scion–rootstock union taken with a millimeter scale (Figure 1). All combinations exhibited apparent tissue continuity and no visible signs of rejection (necrosis, severe callus swelling, discoloration, or open fissures). Minor variation in callus thickness and union symmetry was observed among combinations, without visual evidence of early incompatibility. These observations are qualitative and complement—rather than replace—the quantitative evidence presented in Table 2 and Table 4.

Overall, the results indicate generalized initial compatibility indicators of arabica–robusta grafts in the nursery, with Villa Sarchí showing superior photosynthetic performance compared to Robusta in ΦPSII and ETR, and no differences in Fv/Fm or the morphological traits evaluated. The pattern of physiological improvement without early morphological change suggests initial functional coupling of the scion with the rootstock, whose agronomic relevance should be verified at later stages (field, stress conditions).

4. Discussion

In this study, grafting C. arabica varieties onto C. canephora rootstocks showed satisfactory initial compatibility indicators. Villa Sarchí stood out for higher photosynthetic performance, whereas no significant differences were observed in morphological growth among treatments.

4.1. Early Compatibility of Popular C. arabica Cultivars Grafted Onto Canephora

In this study, we proposed recommendations to demonstrate that grafting C. arabica varieties onto C. canephora rootstocks shows satisfactory initial compatibility [15]. This compatibility reflects the physiological viability of interspecific grafts during juvenile stages [12]. The high success rate and the absence of morphological abnormalities suggest an adequate physiological and anatomical union between the rootstock and scion tissues [14]. One critical factor for initial survival is the early formation of a functional cambium [28]. Additionally, the observed morphological uniformity and lack of necrosis in the union zone imply rapid vascular reconnection, which is essential for the exchange of photoassimilates and hormonal signals [29].

4.2. Physiological Performance in Coffea arabica/Coffea canephora Grafts

The results of this research showed significant differences in two physiological variables: the PSII quantum yield (ΦPSII) and the electron transport rate (ETR). Specifically, the cultivar Villa Sarchí exhibited higher photosynthetic efficiency compared with Robusta. This finding suggests a favorable scion–rootstock interaction, possibly related to inherent physiological characteristics of Villa Sarchí. Similar findings were reported by [30], who observed improvements in photosynthesis and stomatal conductance in C. arabica plants grafted onto Robusta, especially under water stress. Likewise, ref. [12] highlighted genotypic variability in photosynthetic responses following interspecific grafting. These authors emphasized that specific traditional cultivars, such as Villa Sarchí, respond positively—or at least maintain solid photosynthetic performance—when grafted onto Robusta. The behavior described here could also be attributed to better adaptation of the arabica scion (Villa Sarchí) to the temperate nursery climate in Ocumal (Amazonas Region), given that C. canephora generally thrives under warmer conditions [15]. Under cooler conditions, quantum efficiency in Robusta may be limited, favoring Villa Sarchí by comparison.

4.3. Morphological Performance and Visual Compatibility

In contrast, morphological variables did not show significant differences among the treatments. This indicates that, in terms of early growth and development, all grafted varieties performed similarly. These findings agree with previous nursery studies where Arabica–Robusta grafts exhibited high graft-take rates [31]. Visual assessments showed well-formed scion–rootstock unions across combinations, suggesting good initial compatibility, with no external signs of rejection or incompatibility. However, long-term research indicates that the absence of early differences does not rule out potential divergence later in the field [32]. Further extended follow-up studies could reveal subtle differences in vigor, yield, or plant architecture after several productive cycles.

4.4. Comparison with the Literature

Grafting C. arabica onto C. canephora has been investigated primarily for its potential to reduce sanitary problems and enhance tolerance to water stress and nematodes. Ref. [30] reported benefits in yield and stability under adverse conditions after grafting arabica onto Robusta. In Mexico, ref. [33] documented significant yield advantages in nematode-infested soils. Nevertheless, under optimal environments, grafting may not provide clear advantages and might even slightly affect growth or yield due to partial incompatibility [12,15]. In the controlled nursery conditions of the present study, while short-term morphological differences were minimal, certain combinations—such as Villa Sarchí—stood out for superior physiological performance. This advantage could translate into field benefits under moderate stress.

Although this study was conducted under controlled nursery conditions at a single altitude, the interpretation of our results can be better understood when considering the contrasting environments where coffee grafting is commonly applied. High-altitude regions typically present lower temperatures [4,34], greater diurnal thermal amplitude, and reduced vapor pressure deficit [35], conditions that generally favor wound healing and callus formation at the graft interface. In contrast, low-altitude environments tend to impose higher heat loads and evaporative demand, which may compromise early physiological stability and increase the risk of desiccation at the graft union. Within this context, the physiological patterns observed in our study (particularly the differences in ΦPSII and ETR) are consistent with conditions that favor efficient photochemical functioning during early graft establishment. While our experiment does not directly compare altitudinal gradients, these environmental considerations help situate our findings within the broader framework of how grafted coffee plants may respond differently across elevations.

4.5. Physiological and Genetic Causes

The observed compatibility can be interpreted through three main factors. First, a physiological factor: Villa Sarchí may have developed a more efficient vascular connection with the Robusta rootstock, optimizing water and nutrient supply to the scion and thereby enhancing photosynthetic performance [30]. Second, a genetic factor: compatibility is linked to the genetic proximity between the rootstock and the scion. Although C. arabica and C. canephora differ in ploidy (tetraploid versus diploid), specific arabica cultivars may have greater genetic affinity with Robusta, supported by their origin or specific lineage [31,32]. Third, a technical factor: in this study, the grafting technique and controlled nursery management favored high graft take and good initial uniformity, thereby minimizing technical differences among treatments.

The physiological basis of graft compatibility in Coffea involves coordinated vascular reconnection, efficient transport of photoassimilates, and genetic affinity between the scion and rootstock. In the present study, the higher quantum yield of photosystem II (ΦPSII) and electron transport rate (ETR) observed in the combination Villa Sarchí × Robusta suggests a more effective vascular linkage, which facilitates the redistribution of assimilates and signals across the graft union. This functional integration is consistent with the anatomical integrity and balanced growth patterns recorded in this combination, reflecting successful cambial alignment and reduced physiological stress [28]. Although no direct molecular markers were analyzed, previous studies have shown that compatible grafts of C. arabica and C. canephora share similar lignin biosynthesis and auxin transport profiles, enhancing vascular differentiation [36]. Therefore, the observed physiological stability may indirectly reflect underlying genetic affinity, which facilitates metabolic synchronization between the grafted partners. Future analyses should include biochemical and molecular markers to validate these hypotheses and better elucidate the mechanisms of long-term compatibility in coffee grafts.

4.6. Practical Implications for Coffee Production

The high physiological and morphological compatibility observed has a direct impact on tropical agronomy, particularly in regions such as the Peruvian Amazon. Grafting can confer significant advantages against nematodes and water stress. Moreover, it allows combining the resistance of Robusta with the sensory quality of arabica without compromising cup quality [30]. From an economic standpoint, grafting tends to be particularly advantageous under unfavorable conditions, as it does not necessarily increase vigor in optimal settings. Further studies are warranted.

4.7. Limitations and Future Research

This research has some limitations. A key limitation is the short duration of the study, which was conducted exclusively in the nursery. Many signs of incompatibility may appear later in the field [15]. Another limitation is the use of a single Robusta rootstock genotype, which limits extrapolation to other commercial clones or hybrids. Based on the present findings, long-term field studies are recommended to evaluate productivity, yield, and cup quality fully. It would also be valuable to test different Robusta rootstock genotypes to identify more efficient combinations. Finally, physiological and molecular studies could elucidate specific mechanisms of compatibility or incompatibility, thereby strengthening the scientific basis for rootstock–scion selection and management.

This study primarily focused on the early morphophysiological characteristics of grafts. However, we acknowledge that successful bonding between rootstock and graft involves complex biochemical processes. These include the accumulation of phenolic compounds, the activity of antioxidant enzymes, and the regulation of oxidative stress [37]. Future research should aim to monitor these biochemical indicators to complement the current findings, providing a more comprehensive understanding of the compatibility and longevity of graft-rootstock combinations. This line of inquiry will help optimize the selection of graft and rootstock combinations, particularly in coffee systems in high-Andean conditions that face increasing environmental stress.

Although this study focused on the initial compatibility indicators phase under nursery conditions, it is acknowledged that assessing the yield and bean quality of grafted plants is essential for determining the agronomic feasibility of the technique. Future investigations will include these evaluations to establish long-term productivity and quality implications of Coffea arabica scions grafted onto C. canephora rootstocks.

5. Conclusions

This study demonstrates the initial compatibility indicators of interspecific grafts of Coffea arabica (Bourbon, Geisha, Catuai, and Villa Sarchí) onto C. canephora (Robusta) under nursery conditions. Physiologically, the Villa Sarchí combination showed higher PSII quantum yield (ΦPSII) and electron transport rate (ETR) than the self-grafted Robusta control, whereas Fv/Fm did not differ among treatments. Morphological traits (graft-take, plant vigor, height, stem diameter, root length, and leaf number) showed no significant differences, supporting an initial functional coupling of scion and rootstock without early growth penalties.

Taken together, these findings indicate that grafting elite C. arabica cultivars onto Robusta rootstock is a viable horticultural option for tropical highland nurseries, with potential to enhance resilience to edaphic and phytosanitary constraints without compromising early performance. Nevertheless, the evidence is short-term and based on a single rootstock genotype; thus, field validation is required to confirm the persistence of compatibility and its translation into yield, longevity, and cup quality. Future research should evaluate additional Robusta genotypes/clones and incorporate histological and/or molecular analyses of the graft union.