Identification and development of drought-tolerant cocoa hybrids: physiological insights for enhanced water use efficiency under water stress conditions

Abstract

Background

Water stress affects the most important determinants of yield-canopy architecture, photosynthesis, and partitioning of assimilates. Being a perennial crop, the water requirement in cocoa is fairly high. Therefore, identifying drought-tolerant plant varieties within cocoa species is critically important, especially in the context of ongoing climate variability attributed to climate change. This research aimed to develop cultivars that efficiently utilize available/minimal water resources and maintain optimal yields despite environmental stressors. Hence, efforts were taken to identify drought-tolerant genotypes, taking physiological characteristics as the main parameters. Drought-tolerant parents (M 13.12, G I 5.9, G II 19.5, and G VI 55) identified in previous studies were crossed in all possible combinations to attain drought-tolerant hybrids. All the crosses were obtained except for GV1 55 x M 13.12, which may be due to an incompatibility reaction between these genotypes, which is a common mechanism in cocoa. The hybrids were then subjected to drought stress (under 40% field capacity) and were analyzed for various physiological parameters such as chlorophyll stability index, membrane stability, relative water content, photosynthetic rate, transpiration rate, leaf temperature and chlorophyll content.

Results

The highly tolerant and tolerant hybrids in almost all the crosses studied had relatively high chlorophyll stability index (highest being 86.73% in highly tolerant hybrid of cross M 13.12 x G I 5.9), membrane stability (highest value of 86.36% observed in tolerant hybrids of cross M 13.12 x G I 5.9), relative water content (highest value being 79.43% observed in the highly tolerant hybrid of cross G II 19.5 x G VI 55 while the lowest value of 23.51% being shown by the susceptible hybrids of cross G VI 55 x G II 19.5), photosynthetic rates (highest being 1.627 µmol CO₂ m^− 2s^− 1 observed in cross M 13.12 x G I 5.9) and chlorophyll content (highest being 41.27 SPAD units observed in the highly tolerant hybrids of G II 19.5 x G VI 55) as compared to susceptible hybrids. Tolerant hybrids had lower transpiration rates (lowest being 0.306 mmol H₂O m^− 2 s^− 1 in cross G VI 55 x G II 19.5) than susceptible hybrids (highest being observed in the susceptible hybrids of cross M 13.12 x G I 5.9 having the value of 2.067 mmol H₂O m^− 2 s^− 1) indicating their efficiency in handling water stress. However, all the tolerant, susceptible and fully irrigated hybrids showed comparable values ranging between 30 and 33⁰ C for the leaf temperature indicating the efficiency of cocoa plants in regulating the water stress even during drought. Correlation and path analysis revealed that relative water content and photosynthetic rate were positively associated with the dependent variable, the number of leaves retained. However, the transpiration rate showed a negative correlation with several parameters such as cell membrane stability (-0.550), relative water content (-0.528) and chlorophyll stability index (-0.319). Binary regression analysis indicated that relative water content and photosynthetic rate will show 51.87 and 66.29% improvement over the base population, respectively, if used as selection criteria for a new population in future drought breeding programmes.

Conclusion

This experiment indicated how the plants were able to regulate various physiological mechanisms under drought stress and how these parameters can be utilized in distinguishing between drought tolerant and susceptible plants. Ranking based on the values of physiological parameters revealed that crosses M 13.12 x G I 5.9, G II 19.5 x G VI 55, and G II 19.5 x G I 5.9 produced hybrids with favourable responses to drought tolerance. It indicated that these crosses could be further utilized in developing drought-tolerant genotypes. Further, the binary regression studies indicated that relative water content and photosynthetic rate can be used as selection parameters to identify drought-tolerant hybrids more efficiently in the future cocoa stress breeding programmes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-025-06448-3.

Background

Drought is a critical challenge to global agricultural production and is expected to become more severe as climate change progresses. It is often associated with elevated temperatures, which in turn increases evapotranspiration and disrupts photosynthetic processes, thereby impairing the various molecular, biochemical, physiological, morphological, and ecological mechanisms in plants and further reducing crop yields. A water-deficit environment negatively affects crops’ yield and quality, contributing to the progression of land degradation, desertification, and salinization—processes that significantly affect the productivity of crops and threaten food security [1]. Tolerance to drought is a complex quantitative trait; therefore, it is essential to understand the underlying physiological and genetic mechanisms to comprehend the nuances associated with the trait. Gaining insights into these physiological processes can enhance our understanding of the complex network of drought tolerance-related traits, thereby improving selection efficiency. The trait dissection represents a crucial step in understanding the genetic mechanism that will ultimately enhance the effectiveness of molecular breeding strategies [2].

Drought stress disrupts the routine of various physiological processes, which include a reduction in germination, seedling development, and overall plant growth. It affects the elongation rate, plant biomass, leaf area and size, plant height, and relative water content. Prolonged exposure to drought stress adversely impacts the flowering and grain-filling stages, which ultimately leads to a substantial reduction in crop yield. One of the main physiological parameters that gets affected first is the photosynthesis rate. Studies indicate that due to water stress, plant organelles involved in various photochemical reactions change, leading to long-lasting effects on photosynthesis. These effects include reduced leaf growth, improper functioning of the photosynthetic apparatus as well as early leaf senescence. Stomatal closure reduces the molar fraction of CO₂ available in the chloroplast, making the plant more susceptible to photo-oxidative damage. Moreover, limited soil moisture conditions negatively impact photosynthetic pigments, damage the photosynthetic apparatus, and decrease the concentration of main enzymes involved in photosynthetic reactions, which in turn reduces crop growth and yield [3].

Another effect of drought stress on plants is a reduction in relative water content (RWC). The rate of RWC decline is influenced by factors such as leaf water potential, transpiration rate, and the closure of stomata. Increased leaf temperature under drought stress disrupts key metabolic processes, including respiration, photosynthesis, ion and nutrient uptake, and the synthesis of amino acids and proteins. Moreover, drought stress often leads to oxidative and osmotic stress, causing imbalances in ion concentrations and inducing significant alterations in cell membrane structure and other cellular functions. As a result, plants under drought stress typically exhibit reduced RWC and weakened cell membrane integrity. Additionally, drought stress compromises lipid membranes, leading to increased permeability of the cell membrane and higher electrolyte leakage. Hence, it is evident that the physiological parameters such as photosynthesis, relative water content, cell membrane integrity, stomatal conductance, transpiration rate, and chlorophyll content are interconnected and change or disruption in any one of the factors can cause changes in others as well [4].

The internal changes can strongly modify the morphological features of the plants. A strong correlation exists between plant growth and water availability, as water deficit primarily impairs cell expansion rather than cell division. Under such conditions, plant growth is inhibited due to a reduction in cell wall extensibility and turgor pressure. In cases of severe drought, respiration rates may decline. Advanced stages of drought stress include increased rates of leaf senescence, drooping, scorching, rolling of leaves and brittleness, closed flowers and flower wilting, etiolation, wilting, loss of turgidity, premature leaf abscission, and yellowing. While less frequent, other manifestations of drought stress include twig cracking, branch dieback, necrosis, stunted growth and thinning of trees [5].

As a shade-loving crop, cocoa relies heavily on water for its physiological and reproductive development. Compared to other tree crops, cocoa is less adept at managing water stress, making it particularly vulnerable to prolonged periods of drought [6, 7]. Water stress significantly impacts various physiological processes, leading to a decline in yield due to reduced net photosynthetic rates. This reduction can stem from decreased stomatal conductance (gs) or impaired metabolic functions [8]. Drought stress inhibits growth and induces changes in the plant’s metabolism and physiology, affecting morphogenesis and overall plant health [9–12]. As cocoa faces increased water scarcity, understanding these physiological responses is critical for developing strategies to enhance drought tolerance and maintain yield.

With the ongoing decline in available water sources for everyday needs, developing cocoa genotypes that can withstand prolonged periods of water scarcity is increasingly crucial. Recent reports suggest that climate change could push cocoa to the brink of extinction within 40 years due to rising temperatures, which the crop is susceptible to [13]. According to recent news in Financial Times, a 13% drop was seen in global cocoa output due to much lesser production in the countries of Ivory Coast and Ghana- which are responsible for half of the world’s production. The main reason for this decrease was attributed to the El Nino warming weather phenomenon, resulting in poor harvests and increased disease incidence [14]. Studies have concluded that El- Nino-related drought stress caused high cocoa tree mortality of about 15% and severely decreased cocoa yield by up to 89% [15]. Due to this short supply of cocoa, the prices are being doubled; as a result, conglomerates like Mondelez and Hershey are implementing price increase and are planning to offer products at varied prices. This has ultimately led to a decline in their European sales volumes [16]. This urgency highlights the need to breed drought-tolerant varieties or identify existing ones so that the cocoa farmers are not affected much. Given that cocoa is now priced higher than ever (Rs.1000 per kg in India) and is regarded as a “golden tree”, farmers are getting encouraged to grow cocoa in non-traditional areas with sufficient irrigation facilities. Hence, the need of the hour is to breed varieties suitable for drought conditions and also to identify them. The ability to identify the genotypes that combine the traits for good growth and high yield with efficient Water Use Efficiency (WUE) is an essential requirement for breeding cocoa for drought-affected areas [17]. This paper discusses several physiological parameters that are instrumental in identifying such genotypes, providing insights into how these traits can guide future breeding programs.

Materials and methods

Experimental site and hybridization program

The experiment was conducted at the Cocoa Research Centre (CRC), Kerala Agricultural University, Thrissur, Kerala, India. Four genotypes previously identified as drought-tolerant (M 13.12, G I 5.9, G II 19.5, and G VI 55) in a prior study [18] were crossed in all possible combinations to generate F₁ progeny. The progeny from each family group (derived from a single cross) exhibited segregation due to the heterozygous nature of the parental lines. Subsequently, the seedlings progressed to the next phase of evaluation (Table 1).

Table 1.

List of parents used for hybridisation

No	Accession No.	Source
1	M 13.12	Progeny of pods from Karnataka
2	G I 5.9	T76/1224/1201 (Amazon)
3	G II 19.5	Progeny of pods from Nileshwar
4	G VI 55	Progeny of pods from Cadbury farm, Chundale

Screening for initial vigour and imposing stress

At the three-month growth stage, the hybrids were evaluated for initial vigour using a method proposed by [19]. His findings indicated that plants exhibiting a higher HD² (Height × Diameter²) value at an early stage tended to yield more significantly during stable production stages upon maturity (Table 1 of supplementary data). Other parameters such as chlorophyll content and number of leaves were also taken at this stage. In the fifth month, drought stress was imposed following the gravimetric method for two weeks [20]. Field capacity was maintained at 40% (The plants died at less than 40% field capacity, hence, 40% was chosen to impose drought stress) and the chosen 120 hybrids were analyzed for various physiological parameters.

Screening for drought tolerance

After two weeks, based on the percentage of leaves retained on the plant, the hybrids were categorized into four groups (Table 2). Various physiological parameters were estimated on the stress imposed plants. A control representing each cross were kept at fully irrigated condition (Table 2 of supplementary data).

Table 2.

Score chart depicting the leaves retained in the hybrids after drought stress imposition

Sl No.	Percentage of leaves retained	Classification
1	0–10	Highly susceptible
2	10.1–40	Susceptible
3	40.1–70	Tolerant
4	More than 70	Highly tolerant

Physiological analysis

Chlorophyll stability index

Fresh leaf samples (two in number) weighing 0.1 g each were placed in test tubes containing 7 mL of dimethyl sulfoxide (DMSO). One sample was treated at 55 °C for 30 min in a hot water bath, while the other was maintained at room temperature as a control. After 30 min, 3 mL of DMSO was added to each test tube to achieve a final volume of 10 mL (V). Absorbance was measured at 652 nm, following the method established by [21].

The Chlorophyll Stability Index was worked out using the following formulae:

Cell membrane stability (%)

Leaf disc weighing 0.1 g were placed in a test tube, and 15 mL of distilled water was added, allowing the setup to incubate for three hours. The initial electrical conductivity of the solution was recorded (C₁). After this measurement, the leaf discs were returned to the original solution and boiled for 10 min. Following boiling, the leaf discs were removed, and the solution was allowed to cool. The electrical conductivity of the solution was then measured again (C₂), as described by [22].

Relative water content (%)

Twenty leaf discs, each measuring 1 cm in diameter, were excised from the youngest mature leaf, and their fresh weight was recorded. The discs were then kept in water in petri dishes and left at room temperature under ambient light for four hours. After this period, the discs were gently blotted with tissue paper, and their turgid weight was recorded. Subsequently, the leaf discs were oven-dried at 80 °C for 24 h, and the dry weight was measured, following the methodology described by [23].

Leaf temperature (⁰C)

The leaf temperature was measured using the infrared gas analyzer (IRGA). IRGAs are widely used in ecological and agricultural research to monitor gas exchange in plants, helping to study processes like photosynthesis and respiration. The reading (measured in °C) was recorded during morning and evening hours [24].

Photosynthetic rate (µmol CO₂ m^− 2 s^− 1)

The photosynthetic rate of leaves was measured using the infrared gas analyzer (IRGA). The reading was recorded during morning and evening hours [24].

Transpiration rate (mmol H₂O m^− 2 s^− 1)

The transpiration rate of leaves was measured using the infrared gas analyzer (IRGA). The reading was recorded during morning and evening hours [24].

Chlorophyll content (SPAD units)

Chlorophyll content was observed using a SPAD meter. It is used to measure chlorophyll content in leaves and is valued for its simplicity, speed, and non-destructive measurement method, making it a practical tool for both research and agricultural management [24].

Statistical analysis

Analysis of variance

The data recorded were statistically analyzed using analysis of variance (ANOVA) as applied to a completely randomized block design (CRD) using SPSS software. Critical difference (CD), Coefficient of variation (CV), and Standard error were also analyzed [25]. The probability value (p-value) was used to validate the hypothesis that all treatments were the same against the observed data. P-values below 0.05 were considered significant.

Correlation and path analysis

The Pearson correlation coefficient was employed to evaluate the relationship between various physiological traits and the percentage of leaves retained [26]. Correlation analysis was performed using R software Version 3.6.1. Additionally, path analysis was conducted using SPSS software to assess both the direct and indirect effects of each physiological parameter on the percentage of leaves retained, as well as their interrelationships [27, 28].

Binary regression analysis

A binary regression analysis was conducted on the hybrids, with the independent variable defined as a binary response (tolerant vs. susceptible). The objective of this analysis was to examine the relationship between the dependent variables (physiological traits) and the independent variable, which represented the number of leaves retained after two weeks of drought. Additionally, the analysis aimed to assess the effectiveness of these physiological traits in classifying the hybrids into tolerant and susceptible categories within a subsequent population compared to the original population. This analysis was performed using SPSS software.

The percent improvement in selection over the base population was calculated by:

Result

Chlorophyll stability index (CSI)

In the cross M 13.12 x G I 5.9 (Table 3), the highest mean value of the chlorophyll stability index (CSI) was observed at 86.73% for highly tolerant hybrids, while susceptible hybrids had the lowest CSI value of 51.47%. The control hybrid exhibited a maximum value of 92.65%.

Table 3.

Mean values of physiological parameters for different crosses with CD and CV values

Crosses	Reaction to drought	Chlorophyll stability Index (%)	Transpiration rate (mmol H2O m− 2 s− 1)	Chlorophyll content (SPAD units)
M 13.12 x G I 5.9	HT	86.73	0.351	40.57
	T	70.66	0.612	38.73
	S	51.47	2.067	28.3
	CNTRL	92.65	2.868	44.8
	CV (%)	1.815	5.736	4.112
	CD (005%)	3.968	0.182	4.665
	(0.01%)	7.288	0.335	8.567
	SE (m)	0.865	0.063	0.900
	SE (d)	1.223	0.089	1.272
	P- value	0.00**	0.037**	0.001**
M 13.12 x G II 19.5	HT	66.03	0.484	32.70
	T	59.64	0.392	36.19
	S	47.43	0.743	25.54
	CNTRL	91.07	1.086	46.30
	CV (%)	1.934	9.537	2.961
	CD (0.05%)	3.507	0.162	2.924
	(0.01%)	6.442	0.297	5.370
	SE (m)	0.682	0.039	0.594
	SE (d)	0.964	0.055	0.839
	P- value	0.00**	0.001**	0.00**
M 13.12 x G VI 55	HT	74.23	0.426	38.3
	T	67.86	0.435	34.45
	S	52.36	0.634	24.44
	HS	53.87	0.635	26.5
	CNTRL	90.33	0.982	41.2
	CV (%)	1.282	4.818	2.733
	CD (0.05%)	2.194	0.071	2.345
	(0.01%)	3.631	0.118	3.881
	SE (m)	0.507	0.017	0.564
	SE (d)	0.717	0.025	0.797
	P- value	0.00**	0.00**	0.00**
G I 5.9 x M 13.12	T	73.21	0.381	37.05
	S	50.95	0.615	27.95
	CNTRL	91.36	0.950	43.70
	CV (%)	1.871	3.679	0.338
	CD (0.05%)	4.930	0.080	0.473
	(0.01%)	11.385	0.186	1.093
	SE (m)	0.764	0.014	0.262
	SE (d)	1.055	0.019	0.370
	P- value	0.00**	0.00**	0.00**
G I 5.9 x G II 19.5	T	71.81	0.362	34.61
	S	48.65	0.560	24.12
	CNTRL	90.37	1.100	45.7
	CV (%)	0.548	8.478	2.766
	CD (0.05%)	1.427	0.178	3.427
	(0.01%)	3.295	0.412	7.914
	SE (m)	0.722	0.024	0.798
	SE (d)	1.021	0.034	1.129
	P- value	0.00**	0.01**	0.001**
G I 5.9 x G VI 55	HT	74.04	0.447	36.32
	T	68.60	0.460	33.83
	S	45.04	0.686	24.19
	CNTRL	91.44	1.05	42.70
	CV (%)	1.368	5.020	2.436
	CD (0.05%)	2.698	0.083	2.472
	(0.01%)	4.954	0.153	4.540
	SE (m)	0.542	0.017	0.477
	SE (d)	0.767	0.024	0.674
	P- value	0.00**	0.00**	0.00**
G II 19.5 x M 13.12	HT	74.31	0.441	36.60
	T	73.32	0.430	37.65
	S	49.93	0.755	28.50
	CNTRL	96.58	1.105	42.60
	CV (%)	1.542	4.487	1.391
	CD (0.05%)	3.221	0.079	1.530
	(0.01%)	5.915	0.146	2.811
	SE (m)	0.662	0.016	0.349
	SE (d)	0.937	0.023	0.494
	P- value	0.00**	0.00**	0.00**
G II 19.5 x G I 5.9	HT	83.52	0.467	40.65
	T	75.38	0.422	38.24
	S	54.01	0.854	25.05
	CNTRL	95.83	1.39	45.10
	CV (%)	1.503	5.832	3.014
	CD (0.05%)	3.388	0.112	3.326
	(0.01%)	6.223	0.206	6.109
	SE (m)	0.653	0.023	0.715
	SE (d)	0.923	0.032	1.011
	P- value	0.00**	0.00**	0.00**
G II 19.5 x G VI 55	HT	77.77	0.451	41.27
	S	48.63	0.704	25.50
	CNTRL	91.49	0.845	43.50
	CV (%)	0.836	5.816	1.429
	CD (0.05%)	2.269	0.150	2.064
	(0.01%)	5.240	0.347	4.767
	SE (m)	0.499	0.025	0.348
	SE (d)	0.706	0.035	0.492
	P- value	0.00**	0.004**	0.00**
G VI 55 x G I 5.9	T	74.46	0.398	38.78
	S	56.75	0.855	23.55
	CNTRL	94.28	1.25	46.60
	CV (%)	0.737	2.757	1.153
	CD (0.05%)	2.076	0.075	1.557
	(0.01%)	4.794	0.174	3.597
	SE (m)	0.329	0.064	0.81
	SE (d)	0.465	0.091	1.145
	P- value	0.00**	0.004**	0.001**
G VI 55 x G II 19.5	HT	76.83	0.306	32.07
	T	71.21	0.464	34.53
	S	57.17	0.698	27.57
	CNTRL	90.14	0.841	40.20
	CV (%)	1.575	8.460	0.978
	CD (0.05%)	3.411	0.135	0.983
	(0.01%)	6.265	0.249	1.805
	SE (m)	0.676	0.031	0.33
	SE (d)	0.956	0.044	0.466
	P- value	0.0**	0.001**	0.00**

Abbreviations used: HT- Highly Tolerant, T- Tolerant, S- Susceptible, HS- Highly Susceptible, CNTRL- Control

**- significant at 0.05% level, NS- Non- significant

In the cross M 13.12 x G II 19.5, the highest mean CSI recorded was 66.03% for highly tolerant hybrids, compared to a low of 47.43% for susceptible hybrids.

For the cross M 13.12 x G VI 55, highly tolerant hybrids had a mean CSI of 74.23%, followed by tolerant hybrids at 67.86%. Susceptible hybrids recorded a lower value of 52.36%, while highly susceptible hybrids showed the lowest value at 53.87%.

In the cross G I 5.9 x M 13.12, tolerant hybrids achieved a higher mean CSI of 73.21%, whereas susceptible hybrids had a mean of 50.95%. The control in this case recorded the highest mean value of 91.36%. In the cross G I 5.9 x G II 19.5, tolerant hybrids exhibited the highest CSI at 71.81%, while susceptible hybrids had the lowest at 48.65%. The cross G I 5.9 x G VI 55 revealed highly tolerant hybrids with a mean CSI of 74.04%, in contrast to susceptible hybrids, which recorded the lowest value of 45.04%. The control in this cross reached 91.44%.

For the cross G II 19.5 x M 13.12, the highest mean CSI value of 74.31% was observed in highly tolerant hybrids, followed by tolerant hybrids at 73.32%. Susceptible hybrids had a mean of 49.93%. In the cross G II 19.5 x G I 5.9, the highest CSI was 83.52% for highly tolerant hybrids, compared to 54.01% for susceptible hybrids. In the cross G II 19.5 x G VI 55, the highly tolerant hybrid had a CSI of 77.77%, while the susceptible hybrid recorded the value of 48.63%. The control value was 91.49%. In the cross G VI 55 x G I 5.9, tolerant hybrids exhibited the highest CSI at 74.46%, while the lowest mean was found in susceptible hybrids at 56.75%. For the cross G VI 55 x G II 19.5, the highest CSI of 76.83% was observed in highly tolerant hybrids, with susceptible hybrids having the mean CSI value of 57.17%. The control recorded the highest CSI of 90.14%.

Overall, nearly all tolerant hybrids across the different crosses demonstrated higher chlorophyll stability than susceptible hybrids, with control plants showing the highest values. Among the crosses, the tolerant hybrids from M 13.12 x G I 5.9 showed the highest CSI at 86.73%, followed by G II 19.5 x G I 5.9 at 83.52%. The lowest CSI among tolerant hybrids was recorded in the cross M 13.12 x G II 19.5 at 66.03%. These findings suggest that these crosses could be effectively utilized to produce hybrids with stable chlorophyll indices even under drought-stress conditions.

Cell membrane stability

In the cross M 13.12 x G I 5.9 (Fig. 1a), highly tolerant hybrids exhibited the highest membrane stability value of 86.36%, followed by tolerant hybrids at 65.95%. Susceptible hybrids showed the lowest membrane stability value of 51.90%, while the control achieved the highest value of 92.53%, indicating that membrane stability can be significantly affected by drought stress.

Fig. 1.

Variation in Cell membrane stability (CMS) across hybrid categories in different crosses (Mean ± Standard error) a Cross M 13.12 x G I 5.9 (Mean ± Standard error). b Cross M 13.12 x G II 19.5 (Mean ± Standard error). c Cross M 13.12 x G VI 55 (Mean ± Standard error). d Cross G I 5.9 x M 13.12 (Mean ± Standard error). e Cross G I 5.9 x G II 19.5 (Mean ± Standard error). f Cross G I 5.9 x G VI 55 (Mean ± Standard error). g Cross G II 19.5 x M 13.12 (Mean ± Standard error). h Cross G II 19.5 x G I 5.9 (Mean ± Standard error). i Cross G II 19.5 x G VI 55 (Mean ± Standard error). j Cross G VI 55 x G I 5.9 (Mean ± Standard error). k Cross G VI 55 x G II 19.5 (Mean ± Standard error). Error bars indicate standard error. CMS: Cell membrane stability

In the cross M 13.12 x G II 19.5 (Fig. 1b), the highest membrane stability values were observed in highly tolerant hybrids at 86.02%, compared to the lowest values at 54.05% in susceptible hybrids. Similarly, in the cross M 13.12 x G VI 55 (Fig. 1c), highly tolerant hybrids recorded a membrane stability value of 85.90%, while highly susceptible hybrids had the lowest value of 48.44%. The control in this cross displayed a membrane stability value of 92.96%. In the cross G I 5.9 x M 13.12 (Fig. 1d), tolerant hybrids achieved a maximum value of 81.22%, whereas susceptible hybrids recorded a lower value of 53.79%. For the cross G I 5.9 x G II 19.5 (Fig. 1e), tolerant hybrids had higher membrane stability values of approximately 74.74%, while susceptible hybrids were lower at 50.88%. In the cross G I 5.9 x G VI 55 (Fig. 1f), the highly tolerant hybrids showed a mean membrane stability value of 82.98%, followed by tolerant hybrids at 72.89%, with susceptible hybrids at 57.13%.

The cross G II 19.5 x M 13.12 (Fig. 1g) demonstrated that highly tolerant hybrids had a mean value of 79.20%, while tolerant hybrids were slightly higher at 79.87%. Susceptible hybrids recorded a lower value of 56.38%.

In the cross G II 19.5 x G I 5.9 (Fig. 1h), highly tolerant hybrids recorded a mean value of 67.25%, and tolerant hybrids had a slightly higher value at 70.18%, while susceptible hybrids had a mean of 45.56%. In G II 19.5 x G VI 55 (Fig. 1i), the highly tolerant hybrid had the highest membrane stability value at 78.31%, while the susceptible hybrid had the lowest at 58.94%.

In the cross G VI 55 x G I 5.9 (Fig. 1j), tolerant hybrids achieved the highest mean value of 77.54%, whereas susceptible hybrids recorded a lower mean value of 50.90%. In the cross G VI 55 x G II 19.5 (Fig. 1k), highly tolerant hybrids had the highest membrane stability value of 75.72%, while susceptible hybrids displayed the lowest value of 41.99%.

The highly tolerant and tolerant hybrids exhibited greater membrane stability compared to susceptible hybrids, with the control showing the highest stability due to the absence of stress conditions. Among the crosses, hybrids from M 13.12 x G I 5.9 demonstrated the highest membrane stability value of 86.36%, followed closely by those in the M 13.12 x G II 19.5 cross at 86.02%. The lowest membrane stability among tolerant hybrids was found in the cross G II 19.5 x G I 5.9, where the highly tolerant hybrid recorded 67.25%. This variability indicates that each plant’s response to stress parameters is specific for each genotype (Fig. 1).

Relative water content

In the cross M 13.12 x G I 5.9 (Fig. 2a), highly tolerant hybrids exhibited the highest mean relative water content (RWC) value of 60.41%, while susceptible hybrids showed the lowest RWC at 25.40%. The control hybrid had a maximum RWC of approximately 75.53%.

Fig. 2.

Variation in Relative Water Content (RWC) across hybrid categories in different crosses (Mean ± Standard error). a Cross M 13.12 x G I 5.9 (Mean ± Standard error). b Cross M 13.12 x G II 19.5 (Mean ± Standard error). c Cross M 13.12 x G VI 55 (Mean ± Standard error). d Cross G I 5.9 x M 13.12 (Mean ± Standard error). e Cross G I 5.9 x G II 19.5 (Mean ± Standard error). f Cross G I 5.9 x G VI 55 (Mean ± Standard error). g Cross G II 19.5 x M 13.12 (Mean ± Standard error). h Cross G II 19.5 x G I 5.9 (Mean ± Standard error). i Cross G II 19.5 x G VI 55 (Mean ± Standard error). j Cross G VI 55 x G I 5.9 (Mean ± Standard error). k Cross G VI 55 x G II 19.5 (Mean ± Standard error). Error bars indicate standard error. RWC: Relative water content

In the cross M 13.12 x G II 19.5 (Fig. 2b), the highest mean RWC was recorded in highly tolerant hybrids at 76.75%, whereas susceptible hybrids had a significantly lower value of 32.83%. For the cross M 13.12 x G VI 55 (Fig. 2c), highly tolerant hybrids achieved a mean RWC of 73.22%, followed by tolerant hybrids at 64.89%. Susceptible hybrids had lower values of 39.57%, with highly susceptible hybrids showing the least mean RWC at 34.49%.

In the cross G I 5.9 x M 13.12 (Fig. 2d), tolerant hybrids recorded a higher mean RWC of 48.49%, compared to 28.02% in susceptible hybrids, while the control had a mean RWC of 75.01%. The cross G I 5.9 x G II 19.5 (Fig. 2e) showed the highest RWC in tolerant hybrids at 75.00%, with susceptible hybrids exhibiting the lowest at 37.19%.

For the cross G I 5.9 x G VI 55 (Fig. 2f), highly tolerant hybrids recorded a mean RWC of 63.80%, while the lowest RWC was seen in susceptible hybrids at 35.41%. The control in this cross had the highest value of 74.12%. In the cross G II 19.5 x M 13.12 (Fig. 2g), highly tolerant hybrids showed the highest mean RWC at 81.40%, followed by tolerant hybrids at 77.44%, while susceptible hybrids had a lower value of 64.40%.

In the cross G II 19.5 x G I 5.9 (Fig. 2h), the highest mean RWC was 71.97% for highly tolerant hybrids, while susceptible hybrids recorded the lowest at 41.51%. For the cross G II 19.5 x G VI 55 (Fig. 2i), the highly tolerant hybrid had a mean RWC of 79.34%, and the susceptible hybrid had a mean value of 43.63%. The control recorded a mean RWC of 89.40%.

In the cross G VI 55 x G I 5.9 (Fig. 2j), tolerant hybrids displayed the highest mean RWC of 66.75%, indicating their tolerance to drought stress, while susceptible hybrids had the lowest mean value at 37%. In the cross, G VI 55 x G II 19.5 (Fig. 2k), highly tolerant hybrids showed a mean RWC of 53.48%, followed by tolerant hybrids having a mean value of 40.74% with susceptible hybrids having the lowest at 23.51%. The control recorded the highest value of 70.73%.

Higher RWC values were observed in the case of tolerant hybrids, while susceptible hybrids exhibited the lowest values. Notably, the highest RWC among all crosses was observed in G II 19.5 x M 13.12 at 81.40%, followed by G II 19.5 x G VI 55 at 79.34%. The susceptible hybrids ranged from 64.40% in G II 19.5 x M 13.12 to 23.51% in G VI 55 x G II 19.5. This indicates that this parameter can be an effective indicator of drought stress studies as the genotypes showed a wide range of values which means it was significantly affecting the genotypes (Fig. 2).

Leaf temperature

When exposed to drought stress, both drought-tolerant and susceptible hybrids in this study maintained a similar leaf temperature range of 30–33 °C. In contrast, the control plants exhibited a moderate leaf temperature that fell between those of the tolerant and susceptible genotypes. The average leaf temperatures for the different tolerant and susceptible hybrids across the selected crosses are presented in Table 4.

Table 4.

Mean values of physiological parameters for different crosses along with CD and CV values

Crosses	Reaction to drought	Photosynthetic rate (µmol CO2 m− 2s− 1)	Leaf temperature (℃)
M 13.12 x G I 5.9	HT	1.627	37.45
	T	1.482	37.26
	S	0.786	37.44
	CNTRL	2.002	36.4
	CV (%)	1.645	2.627
	CD (005%)	0.043	3.001
	(0.01%)	0.065	5.511
	SE (m)	0.014	0.603
	SE (d)	0.02	0.853
	P- value	0.00**	0.1 NS
M 13.12 x G II 19.5	HT	1.014	31.52
	T	0.884	31.51
	S	0.654	31.03
	CNTRL	1.134	31.25
	CV (%)	3.981	1.509
	CD (0.05%)	0.068	1.491
	(0.01%)	0.102	2.739
	SE (m)	0.027	0.406
	SE (d)	0.038	0.575
	P- value	0.001**	0.831 NS
M 13.12 x G VI 55	HT	0.779	32.31
	T	0.755	31.02
	S	0.607	30.54
	HS	0.514	30.56
	CNTRL	0.909	30.75
	CV (%)	8.226	2.074
	CD (0.05%)	0.103	1.771
	(0.01%)	0.150	2.932
	SE (m)	0.043	0.41
	SE (d)	0.061	0.58
	P- value	0.001**	0.111 NS
G I 5.9 x M 13.12	T	1.100	32.29
	S	0.746	30.47
	CNTRL	1.430	30.55
	CV (%)	4.673	2.197
	CD (0.05%)	0.098	2.964
	(0.01%)	0.162	6.846
	SE (m)	0.148	0.58
	SE (d)	0.032	0.82
	P- value	0.005**	0.630 NS
G I 5.9 x G II 19.5	T	0.998	31.54
	S	0.616	31.03
	CNTRL	1.531	31.20
	CV (%)	2.386	1.480
	CD (0.05%)	0.044	2.002
	(0.01%)	0.072	4.624
	SE (m)	0.012	0.312
	SE (d)	0.017	0.442
	P- value	0.00**	0.145 NS
G I 5.9 x G VI 55	HT	1.053	31.70
	T	0.896	31.66
	S	0.705	31.48
	CNTRL	1.34	31.55
	CV (%)	11.981	2.096
	CD (0.05%)	0.212	2.125
	(0.01%)	NS	3.903
	SE (m)	0.071	0.493
	SE (d)	0.1	0.697
	P- value	0.072 NS	0.424 NS
G II 19.5 x M 13.12	HT	0.983	33.79
	T	0.816	33.79
	S	0.583	33.34
	CNTRL	3.050	33.60
	CV (%)	2.933	1.606
	CD (0.05%)	0.047	1.718
	(0.01%)	0.070	3.156
	SE (m)	0.012	0.563
	SE (d)	0.017	0.796
	P- value	0.00**	0.286 NS
G II 19.5 x G I 5.9	HT	1.133	31.73
	T	1.007	32.44
	S	0.634	30.68
	CNTRL	1.570	31.23
	CV (%)	8.277	1.582
	CD (0.05%)	0.153	1.593
	(0.01%)	0.232	2.925
	SE (m)	0.05	0.314
	SE (d)	0.071	0.444
	P- value	0.006**	0.017**
G II 19.5 x G VI 55	HT	0.971	32.79
	S	0.721	31.77
	CNTRL	1.150	32.65
	CV (%)	7.630	2.081
	CD (0.05%)	0.146	2.847
	(0.01%)	0.243	6.575
	SE (m)	0.039	0.544
	SE (d)	0.056	0.769
	P- value	0.041**	0.551 NS
G VI 55 x G I 5.9	T	0.871	31.38
	S	0.605	30.89
	CNTRL	1.084	31.14
	CV (%)	5.848	2.155
	CD (0.05%)	0.098	2.926
	(0.01%)	0.162	6.758
	SE (m)	0.035	0.642
	SE (d)	0.049	0.876
	P- value	0.002**	0.294 NS
G VI 55 x G II 19.5	HT	0.732	31.11
	T	0.814	31.94
	S	0.583	30.54
	CNTRL	1.061	30.60
	CV (%)	7.836	1.606
	CD (0.05%)	0.111	1.611
	(0.01%)	0.168	2.958
	SE (m)	0.036	0.333
	SE (d)	0.051	0.470
	P- value	0.004**	0.093 NS

Abbreviations used: HT- Highly Tolerant, T- Tolerant, S- Susceptible, HS- Highly Susceptible, CNTRL- Control

**- significant at 0.05% level, NS- Non- significant

In case of cross G II 19.5 x G I 5.9, p-value of 0.017 (significance at less than 0.05% level) indicated significant differences between tolerant, susceptible, and control hybrids. Rest all other crosses displayed non-significant values indicating the inconclusive nature of the parameter in categorizing the hybrids. Due to the overlapping temperature ranges observed in this experiment, leaf temperature may not be a reliable indicator for assessing drought tolerance. Previous studies have shown that drought-stressed plants typically have higher canopy temperatures compared to well-watered plants. This suggests that cocoa plants are efficient in regulating water stress by maintaining a stable canopy temperature, even under drought conditions.

Photosynthetic rate

The highly tolerant hybrids from the cross M 13.12 x G I 5.9 exhibited the highest photosynthetic rate at 1.627 µmol CO₂ m^− 2 s^− 1, while the lowest mean photosynthetic rate was recorded in the highly susceptible hybrids from the cross M 13.12 x G VI 55, at approximately 0.514 µmol CO₂ m^− 2 s^− 1. Overall, highly tolerant and tolerant hybrids consistently demonstrated superior photosynthetic rates, whereas susceptible hybrids exhibited significantly lower values within their respective crosses. The control group displayed even higher photosynthetic rates compared to all hybrids.

These findings suggest that tolerant and highly tolerant hybrids maintain effective photosynthesis even under stress conditions, highlighting their resilience, whereas susceptible hybrids exhibit reduced efficiency in photosynthetic performance. Detailed photosynthetic rates for the various hybrids can be found in Table 4.

Transpiration rate

A contrasting trend was observed in transpiration rates compared to photosynthesis. Hybrids classified as highly tolerant and tolerant displayed lower transpiration rates, whereas susceptible hybrids showed significantly higher rates. The control group also exhibited elevated transpiration rates. A detailed analysis of the transpiration rates across various hybrids can be found in Table 3.

In the cross M 13.12 x G I 5.9, highly tolerant hybrids exhibited the lowest transpiration rate of 0.351 mmol H₂O m^− 2 s^− 1, while susceptible hybrids showed the highest transpiration rate of 2.067 mmol H₂O m^− 2 s^− 1. The control hybrid had the value of 2.868 mmol H₂O m^− 2 s^− 1, which was the highest considering water availability.

In the cross M 13.12 x G II 19.5, the highest mean transpiration rate of 0.743 mmol H₂O m^− 2 s^− 1 was recorded in susceptible hybrids. In contrast, tolerant and highly tolerant hybrids had a significantly lower value of 0.392 mmol H₂O m^− 2 s^− 1 and 0.484 mmol H₂O m^− 2 s^− 1. For the cross M 13.12 x G VI 55, highly tolerant hybrids had the lowest value of 0.426 mmol H₂O m^− 2 s^− 1 followed by tolerant hybrids at the rate of 0.435 mmol H₂O m^− 2 s^− 1. Highly susceptible hybrids had the highest mean value of 0.634 mmol H₂O m^− 2 s^− 1, with susceptible hybrids showing the transpiration rate of 0.634 mmol H₂O m^− 2 s^− 1.

In the cross G I 5.9 x M 13.12, tolerant hybrids recorded a lower mean value of 0.381 mmol H₂O m^− 2 s^− 1, compared to 0.615 mmol H₂O m^− 2 s^− 1 in susceptible hybrids, while the control had a mean value of 0.950 mmol H₂O m^− 2 s^− 1. The cross G I 5.9 x G II 19.5 showed the highest value of transpiration rate in susceptible hybrids at 0.560 mmol H₂O m^− 2 s^− 1, with tolerant hybrids exhibiting higher values at 0.362 mmol H₂O m^− 2 s^− 1.

For the cross G I 5.9 x G VI 55, highly tolerant hybrids recorded the mean rate of 0.447 mmol H₂O m^− 2 s^− 1, while the highest value was seen in susceptible hybrids at having the rate of 0.686 mmol H₂O m^− 2 s^− 1. The control in this cross had the highest value of 1.05 mmol H₂O m^− 2 s^− 1. In the cross G II 19.5 x M 13.12, highly tolerant hybrids exhibited lower mean values of 0.441 mmol H₂O m^− 2 s^− 1, followed by tolerant hybrids at 0.430 mmol H₂O m^− 2 s^− 1, while susceptible hybrids had the highest value of 0.755 mmol H₂O m^− 2 s^− 1.

In the cross G II 19.5 x G I 5.9, the highest mean value for transpiration rate having 0.854 mmol H₂O m^− 2 s^− 1 was observed in susceptible hybrids, while highly tolerant and tolerant recorded the lowest rate at 0.467 mmol H₂O m^− 2 s^− 1 and 0.422 mmol H₂O m^− 2 s^− 1, respectively. For the cross G II 19.5 x G VI 55, the highly tolerant hybrid had a mean transpiration rate of 0.451 mmol H₂O m^− 2 s^− 1, and the susceptible hybrid had a mean value of 0.704 mmol H₂O m^− 2 s^− 1. The control recorded the mean transpiration rate of 0.845 mmol H₂O m^− 2 s^− 1.

In the cross G VI 55 x G I 5.9, tolerant hybrids displayed the lowest mean value of 0.398 mmol H₂O m^− 2 s^− 1, indicating their tolerance to drought stress, while susceptible hybrids had a higher mean value at 0.855 mmol H₂O m^− 2 s^− 1. In the cross, G VI 55 x G II 19.5, highly tolerant hybrids showed a mean transpiration rate of 0.306 mmol H₂O m^− 2 s^− 1, with susceptible hybrids having the highest value at 0.464 mmol H₂O m^− 2 s^− 1. The control recorded the highest value of 0.698 mmol H₂O m^− 2 s^− 1.

Higher transpiration rates were observed in the case of susceptible and highly susceptible hybrids, while tolerant hybrids exhibited lower values comparatively. The highest rate among all crosses was observed in the susceptible hybrids of cross M 13.12 x G I 5.9 having 2.067 mmol H₂O m^− 2 s^− 1 while the lowest mean value was observed in highly tolerant hybrids of cross G VI 55 x G II 19.5 having the mean value of 0.306 mmol H₂O m^− 2 s^− 1. The tolerant genotypes were somewhat able to regulate the loss of water through transpiration more effectively than the susceptible hybrids. The control genotypes of each cross had the highest mean transpiration values due to the availability of water.

Chlorophyll content

In this experiment, chlorophyll content was measured using a SPAD meter, with values ranging from 0 to 50, as shown in Table 3. The results indicated that susceptible hybrids had lower chlorophyll content compared to tolerant hybrids and control plants, highlighting the importance of chlorophyll in drought tolerance.

In the cross M 13.12 x G I 5.9 (Table 3), the highest mean value of the chlorophyll content was observed at 40.57 SPAD units for highly tolerant hybrids, while susceptible hybrids had the lowest value of 28.30 SPAD units. The control hybrid exhibited a maximum chlorophyll content of approximately 44.80 SPAD units.

In the cross M 13.12 x G II 19.5, the highest mean chlorophyll content recorded was 36.19 SPAD units for tolerant hybrids, compared to a low value of 25.54 SPAD units in case of susceptible hybrids.

For the cross M 13.12 x G VI 55, highly tolerant hybrids had the chlorophyll content of 38.30 SPAD units followed by tolerant hybrids at 34.45 SPAD units. Susceptible hybrids recorded the lowest value of 24.44 SPAD units, while highly susceptible hybrids recorded a value of 26.50 SPAD units.

In the cross G I 5.9 x M 13.12, tolerant hybrids achieved a higher mean value of 37.05 SPAD units, whereas susceptible hybrids had a mean of 27.95 SPAD units. The control in this case recorded the highest mean value of 43.70 SPAD units. In the cross G I 5.9 x G II 19.5, tolerant hybrids exhibited the highest value for chlorophyll content at 34.61 SPAD units, while susceptible hybrids had the lowest value at 24.12 SPAD units. The cross G I 5.9 x G VI 55 revealed highly tolerant hybrids with a mean value of 36.32 SPAD units, in contrast to susceptible hybrids, which recorded the lowest value of 24.19 SPAD units. The control in this cross had the chlorophyll content of 42.70 SPAD units.

For the cross G II 19.5 x M 13.12, the highest mean value of 37.65 SPAD units was observed for tolerant hybrids, followed by highly tolerant hybrids at 36.60 SPAD units. Susceptible hybrids had a mean value of 28.50 SPAD units. In the cross G II 19.5 x G I 5.9, the highly tolerant hybrids exhibited the highest value of chlorophyll content at 40.65 SPAD units, compared to 25.05 SPAD units for susceptible hybrids. In the cross G II 19.5 x G VI 55, the highly tolerant hybrid had a chlorophyll content of 41.27 SPAD units, while the susceptible hybrid recorded 25.50 SPAD units. The control value was 43.50 SPAD units. In the cross G VI 55 x G I 5.9, tolerant hybrids exhibited the highest value at 38.78 SPAD units, while the lowest mean was found in susceptible hybrids at 23.55 SPAD units. For the cross G VI 55 x G II 19.5, the highest chlorophyll content of 34.53 SPAD units was observed in tolerant hybrids, followed by highly tolerant hybrids at 32.07 SPAD units. The susceptible hybrids recorded a value of 27.57 SPAD units. The control recorded a value of 40.20 SPAD units.

The highest chlorophyll content was observed in the highly tolerant hybrids from the crosses G II 19.5 x G VI 55 (41.27 SPAD units) and G II 19.5 x G I 5.9 (40.65 SPAD units). In contrast, the lowest chlorophyll content was recorded in the cross G VI 55 x G I 5.9, which had a mean value of 23.55 SPAD units for the susceptible hybrids. These findings further reinforce the role of chlorophyll content in enhancing drought tolerance, as tolerant hybrids consistently exhibited higher chlorophyll levels than their susceptible counterparts.

Correlation and path studies

All measured traits were positively correlated, except transpiration rate, which exhibited negative correlations with several key parameters. Notable negative correlations included cell membrane stability (-0.550), relative water content (-0.528), chlorophyll stability index (-0.319), and the dependent variable, the number of leaves retained (-0.463). This suggests that increased transpiration rates may detrimentally impact plant health by compromising chlorophyll stability, water retention, and leaf integrity, which are crucial for regulating these physiological mechanisms (Fig. III). Interestingly, transpiration rate was positively correlated with leaf temperature, indicating that as transpiration increases, leaf temperature may be affected. However, in cocoa, this relationship differs from that observed in many other plants, as higher transpiration rates did not consistently lead to lower leaf temperatures. This suggests that the plants were able to regulate their transpiration rate effectively, minimizing water loss while maintaining internal hydration levels.

In contrast, several traits showed significant positive correlations with the dependent variable, the number of leaves retained, including chlorophyll stability index (0.698), cell membrane stability (0.693), relative water content (0.635), photosynthetic rate (0.505), and chlorophyll content (0.690). Transpiration rate, however, was negatively correlated with the number of leaves retained, and leaf temperature exhibited minimal significant correlation with this variable (Fig. 3).

Fig. 3.

Correlation plot diagram between different physiological traits. (CSI- Chlorophyll stability index, MSI- Membrane stability index, RWC- Relative water content, PHOTOSYN.- Photosynthetic rate, T.R.- Transpiration rate, CHL CNTNT- Chlorophyll content, LEAF TEMP- Leaf temperature)

Path analysis revealed medium positive effects from cell membrane stability (0.284) and lower direct effects from relative water content (0.121) and photosynthetic rate (0.133) on the number of leaves retained. These results highlight the importance of these traits in evaluating drought tolerance in cocoa genotypes (Fig. 4). The residual effect observed for path analysis was 0.148 indicating that the selected traits accounted for 85.2% variance, leaving only 14.8% unexplained variance in the study.

Fig. 4.

Path diagram for the physiological characters having direct effects on percent of leaves retained

Binary regression studies on the hybrids

The phenes or traits influencing drought tolerance and potential improvements over the base population through selection are illustrated in Fig. 5. The positive and comparable values of the odds ratio (Exp (B)), along with positive correlations, indicate that relative water content and photosynthetic rate are significantly associated with drought stress tolerance. Both traits demonstrated significant values below 0.05, indicating a 95% confidence level in the results (Table 5).

Fig. 5.

Phenes and their association with drought stress analysis in cocoa

Table 5.

Logistic estimate of characters influencing drought tolerance in cocoa

Variables	Coefficient	Standard error	Wald	Significance	Exp(B)	Expected per cent improvement over population (%)
Relative water content**	0.075	0.008	87.803	0.002	1.078	51.87
Photosynthesis**	12.819	1.328	84.290	0.001	1.967	66.29

**Significant value less than 0.05

Based on the Exp (B) value from regression model, expressed percentage for drought tolerance over the base population was calculated and it was found that if selection is based on relative water content, it will show about 51.87 per cent improvement over the base population regarding the drought tolerance and in case of photosynthetic rate, 66.29 per cent improvement will be observed over the base population (Table 5).

Thus, incorporating these traits as screening methods for identifying drought-tolerant genotypes could enhance the selection process for hybrids in future breeding programs.

Ranking of crosses based on physiological crosses

Crosses were evaluated based on their physiological parameters, and it was found that the cross M 13.12 x G I 5.9 exhibited the highest values for the chlorophyll stability index, cell membrane stability, photosynthetic rate, and leaf temperature. This was followed by the cross G II 19.5 x G VI 55, which showed elevated values for the chlorophyll stability index, relative water content, leaf temperature, and chlorophyll content (Table 6). Additionally, the cross G II 19.5 x G I 5.9 demonstrated higher values for the chlorophyll stability index, photosynthetic rate, and chlorophyll content among the tolerant hybrids.

Table 6.

Ranking of crosses based on superiority in various physiological parameters

Crosses	Parameters
	CSI	CMS	Transpiration rate	Photosyn. rate	Leaf temp.	RWC	Chl content
M 13.12 x G I 5.9	1	1	2	1	1	9	3
M 13.12 x G II 19.5	11	2	5	5	9	3	9
M 13.12 x G VI 55	7	3	8	11	4	5	5
G I 5.9 x M 13.12	9	5	4	3	5	11	7
G I 5.9 x G II 19.5	10	10	3	6	8	4	10
G I 5.9 x G VI 55	8	4	10	4	7	8	8
G II 19.5 x M 13.12	6	6	9	7	2	1	6
G II 19.5 x G I 5.9	2	11	7	2	6	6	2
G II 19.5 x G VI 55	3	7	11	8	3	2	1
G VI 55 x G I 5.9	5	9	6	9	10	7	4
G VI 55 x G II 19.5	4	8	1	10	11	10	11

Abbreviations used: Photosyn. Rate: Photosynthetic rate, Leaf temp.: Leaf temperature, Chl content: Chlorophyll content, RWC: Relative water content, CSI- Chlorophyll Stability Index, CMS- Cell Membrane Stability

In contrast, the cross G VI 55 x G II 19.5 recorded the lowest values for tolerant hybrids, despite having the lowest transpiration rate. This data suggests that the crosses M 13.12 x G I 5.9, G II 19.5 x G VI 55, and G II 19.5 x G I 5.9 produced hybrids with favourable responses to drought tolerance. Essentially, hybrids derived from these crosses maintained better physiological conditions, facilitating enhanced adaptation to drought stress. Therefore, these crosses are recommended for future breeding programs focused on developing drought-tolerant hybrids (Table 6).

Discussion

Water stress significantly impacts various physiological processes, including radiation capture, leaf temperature (LT), water content, stomatal conductance, transpiration, electron transport, photosynthesis, and respiration, ultimately influencing the source and sink dynamics of plants [29]. The quantity of water utilized by a crop is closely linked to its photosynthetic activity, dry matter production, and overall yield across many species [29, 30]. However, due to water scarcity, the total photosynthetic potential is rarely achieved. One promising solution is to enhance the drought tolerance of crop varieties through targeted breeding programs. Understanding the physiological processes is essential for identifying desirable gene combinations of hybrids that can tackle drought stress. Due to the irregular and unpredictable nature of drought responses, screening for resistant cultivars under open field conditions is not feasible; however, it can be managed in controlled or sheltered environments. Drought studies nowadays, focuses on the genetic analysis of root architecture, relative water content, and osmotic potential to identify potential genotypes [31].

In the conducted study, most highly tolerant and tolerant hybrids exhibited elevated values for chlorophyll stability index, membrane stability, photosynthetic rate, chlorophyll content, and relative water content, while susceptible hybrids showed significantly lower values. Similar findings have been reported in various crops, including spring wheat [32], cocoa [33–35], sugarcane [36], and olive trees [37]. This consistency across different species underscores the importance of these physiological traits as indicators of drought tolerance.

The measurement of transpiration is crucial for the establishment and management of crops, as it is a key factor in determining water demand. The transpiration rate in the conducted study exhibited an inverse trend, with susceptible hybrids demonstrating higher rates when compared to those of the tolerant hybrids. The control group exhibited highest rates maintained under fully irrigated conditions. Highly tolerant and tolerant hybrids displayed lower transpiration rates, indicating their enhanced ability to mitigate water stress. This suggests that these tolerant hybrids effectively conserve water, which may contribute to their resilience under drought conditions. Transpiration in young cocoa trees is highly sensitive to soil moisture levels, and the rate of transpiration decreases when water availability decreases. Under drought stress, both tolerant as well as susceptible hybrids experienced reduced transpiration rates but the tolerant hybrids exhibited mechanisms to maintain water use efficiency and higher yields compared to their susceptible counterparts. Similar results were observed in a drought related study done in corn [38]. Cocoa plants are unable to maintain optimal transpiration when soil water availability fell below a certain critical threshold. This indicates that water availability plays a key role in regulating water uptake and transpiration in cocoa trees, and there may be a point beyond which the plants are unable to compensate for reduced water availability, thereby, affecting the overall plant health. Hence, genotypes with lower transpiration rates or lower stomatal conductance can tackle this problem to a certain extent [31, 39].

Leaf temperature is another critical factor influencing leaf water status under water deficit conditions [40]. While many studies highlight its role in drought assessment, the current study found no significant variation in leaf temperature among tolerant and susceptible hybrids. All categories—highly tolerant, tolerant, susceptible hybrids, and the control—exhibited somewhat similar leaf temperature values across multiple crosses. Similar results were found in another study done in cocoa genotypes where the plants were kept under drought conditions and there was no significant difference observed between tolerant, susceptible or the control genotypes for leaf temperature [41]. Another drought stress study in sugarcane revealed elevated surface temperature on tolerant and susceptible hybrids indicating that leaf temperature cannot be used as a reliable parameter for drought stress studies [42]. However, in most crosses (except for M 13.12 x G I 5.9), the overall mean temperature for tolerant hybrids was relatively higher than that of susceptible hybrids. This suggests that the higher transpiration rates observed in susceptible hybrids may have contributed to a reduction in surface leaf temperature, leading to lower temperatures in these hybrids compared to the tolerant ones. Cocoa plants typically exhibit low stomatal conductance under water stress and low relative humidity [43], in contrast to normal stomatal opening under non-limiting water conditions and high relative humidity [44]. In this study, all genotypes demonstrated an inability to effectively regulate stomatal opening and closing, resulting in no significant differences between tolerant and susceptible genotypes. Consequently, stomatal conductance proved to be an unreliable parameter for assessing drought tolerance in cocoa.

Correlation and path studies showed a significant positive correlation of chlorophyll stability index, cell membrane stability, relative water content, photosynthetic rate, and chlorophyll content with the dependent variable. Transpiration rate, however, was negatively correlated with the number of leaves retained, and leaf temperature exhibited minimal significant correlation with this variable. Similar results were observed in crops such as cocoa [41], maize [45] and tomato [46].

Binary regression analyses identified relative water content and photosynthetic rate as effective selection criteria for distinguishing drought-tolerant genotypes. Numerous studies highlight the significance of these parameters in differentiating between drought-tolerant and susceptible plants. The photosynthetic process comprises several integral components, including photosynthetic pigments, photosystems, the electron transport chain, and carbon dioxide assimilation pathways. Any disruption to these components can compromise a plant’s photosynthetic capacity [47]. Consequently, quantifying the photosynthetic rate serves as a reliable biomarker for assessing plant stress, offering insights into its physiological status and resilience under water-deficit conditions. Under drought stress, photosynthetic rates can be reduced significantly, with studies reporting an average decline of 86% in cocoa clones, which returned to baseline levels within seven days following rewatering [29]. This underscores the pivotal role of water availability in regulating photosynthetic activity. Additionally, drought stress impedes the growth of cocoa seedlings primarily through the reduction of their photosynthetic efficiency. Evidence suggests that drought-tolerant genotypes exhibit enhanced physiological traits relative to susceptible clones, including increased leaf area, elevated chlorophyll a and b content, higher total chlorophyll, improved relative water content (RWC), and superior photosynthetic rates under drought conditions [30].

Relative water content (RWC) serves as a vital indicator of a plant’s water status, reflecting the degree of stress, particularly in response to drought or elevated temperatures. It is a crucial parameter for assessing plant health and resilience, as it helps in determining the extent of water deficit experienced by the plant [23, 48]. High relative water content (RWC) is recognized as a key resistance mechanism that helps plants cope with drought. This resilience is often attributed to either osmotic adjustments in more tolerant genotypes or reduced elasticity in cell wall tissues [49]. Drought stress typically leads to a decrease in RWC, which has been documented in various crops, including mungbean [50], maize [51], barley [52] and wheat [53]. It, thus, acts as one of the indicators of stress conditions in plants. RWC is significantly correlated with critical parameters such as cellular membrane integrity, intercellular CO₂ concentration, and photosynthesis rates. Thus, it serves as a reliable indicator of plant performance under stress conditions, making it a valuable selection criterion in breeding programs aimed at enhancing drought tolerance [54]. In cocoa research, studies have highlighted the effectiveness of RWC alongside other parameters like proline accumulation and leaf trichome density in identifying drought-tolerant genotypes. For instance, genotypes T63/971 x SCA9 and T60 x POUND10 exhibited relatively high RWC and proline levels, suggesting their resilience to drought stress [55]. Similarly, another study found that 13 cocoa seedlings experienced over a 50% reduction in RWC during drought, with increased proline content distinguishing drought-tolerant from susceptible seedlings [56]. Hence, genotypes with relatively high RWC can be utilized in breeding programmes or used as such to develop cultivars having tolerance to water stress.

The study was able to identify drought-tolerant cocoa hybrids which could be utilized to identify the candidate genes responsible for drought tolerance or can be utilized as parents for further drought breeding programmes. These developed genotypes can also be released as drought tolerant hybrids for commercialization.

Conclusion

Physiological parameters are crucial for understanding the cellular mechanisms plants undergo during drought stress. Among the various parameters examined, relative water content (RWC) and photosynthetic rate emerged as the most reliable indicators for differentiating drought-tolerant from drought-susceptible cocoa genotypes. Identifying genotypes that maintain efficient photosynthetic activity under water stress is crucial, as the primary goal for farmers is to achieve high crop yields even under resource-limited conditions. Thus, genotypes that can retain adequate water and produce high yields in water-scarce environments should be prioritized. In the study, these parameters demonstrated a direct correlation with the number of leaves retained, which was confirmed through path analysis. This highlights the role of these parameters as primary indicators of drought stress.

The cross M 13.12 x G I 5.9 demonstrated superior performance across multiple traits, suggesting that hybrids from this cross could serve as promising candidates for future drought breeding programs. This approach would optimize time and resources by focussing future breeding efforts on the hybrid produced from this cross or utilizing them as parental lines in desired combinations. Regression analysis further corroborated the value of RWC and photosynthetic rate as reliable criteria for identifying drought tolerance in cocoa. Therefore, selecting genotypes with higher RWC and photosynthetic rates is likely to enhance the success of breeding programs aimed at developing drought-tolerant cocoa varieties.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

RWC

Relative water content

CD

Critical difference

CMS

Cell Membrane stability

CV

Coefficient of variation

CSI

Chlorophyll stability index

TR

Transpiration rate

SE

Standard error

Author contributions

Juby Baby wrote the main manuscript text along with figures and tables. Minimol JS edited the manuscript along with tables and figures. All authors reviewed the manuscript.

Funding

This study was financially supported by Mondelez International (Earlier Cadbury India Ltd) and Kerala Agricultural University, Vellanikkara. Researchers supporting Project number (RSPD2025R676) King Saud University.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The article does not contain any studies with human participants or animals performed by any of the authors. Plant materials used in the experiments were from Cocoa Research Centre, Vellanikkara, Thrissur. All methods in the experiment were performed in accordance with the relevant guidelines/ regulations/legislation of Kerala Agricultural University, Vellanikkara, Thrissur.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.