Experimental site

The study was conducted at the Research Center El Nus (Agrosavia), San Roque, Antioquia, Colombia (06° 26' 17.2'' N, 74° 49' 32.1'' W). The weather conditions in 2018 were an average minimum temperature of 18 ºC, an average maximum temperature of 36 °C, average relative humidity of 79 % and annual accumulated rainfall of the region of 2229 mm (Figure 1).

Figure 1 Rainfall (mm), minimum (RHmin), mean (RHmean) and maximum (RHmax) relative humidity (%), minimum (Tmin), mean (Tmean) and maximum (Tmax) temperature. Research Center El Nus (Agrosavia), San Roque, Antioquia, Colombia, 2018.  

Plant material and experimental design

Five cocoa clones (TCS01, TCS06, TCS13, TCS19 and, CCN51) were evaluated under a complete randomized blocks experimental design with four replications; the experimental unit was a group of 20 plants per clone.

Plantation management

Cocoa trees were planted in December 2016 (3 years old) at a distance of 3 × 3 m following a triangle arrangement. The study was carried out in an agroforestry system with permanent trees of abarco (Cariniana pyriformis Miers). ^{Agudelo-Castañeda et al. (2018)} recognized abarco for its potential to regulate incident radiation with less variability; it was considered the forest species to generate a permanent shade on the canopy of cocoa trees. The abarco trees were planted in 2015 (4 years old). At the time of evaluation, they had an average height of 5.1 m and a diameter of 6.04 cm at 1.20 m from the soil surface. The trees were established in 16-m double rows at a distance of 4 × 4 m from each other in a triangle arrangement.

Variables measured

Leaf gas exchange measurements were carried out in September 2018 with a portable photosynthesis measuring system that incorporates an infrared gas analyzer (LCi-ADC Bioscience, Hertfordshire, UK). Measurements were performed on a young, fully expanded leaf of six plants from each cocoa clone for five consecutive days. Net photosynthesis (A, μmol CO₂ m^-2 s^-1), transpiration (E, mmol H₂O m² s^-1), stomatal conductance (g_s, mol CO₂ m^-2 s^-1) and photosynthetically active radiation (PAR, µmol photons m^-2 s^-1) were registered every hour from 08:00 to 17:00 hours. In addition, the A/E and A/PAR ratio, respectively, calculated the instantaneous water use efficiency (WUEi, mmol CO₂ mmol^-1 H₂O) and radiation use efficiency.

Vapor pressure deficit (VPD, measured in kPa) was calculated from daily temperature and relative humidity records, according to Equation [1] proposed by ^{Rosenberg et al. (1983)}. The temperature and humidity values were registered with a thermo-hygrometer (Thermo Hygro & Clock, VAPRECISION, Fountain Valley, California, USA) at 30-minute intervals during the day (8:00 to 17:00 hours).

An SS1 SunScan Canopy Analysis System (Delta-T Devices Ltd, Cambridge, UK) was used to measure the percentage of incident light intercepted by shade trees canopy. Incident radiation flux above the canopy, radiation flux transmitted below the shades tree canopy, extinction coefficient (k), and leaf area index (LAI) were also recorded every half hour, from 08:00 to 17:00.

Daily integral photosynthesis

To determinate the accumulated value throughout the day of the physiological variables A, g_s, E and PAR, the area under the curve (AUC) was estimated for each gas exchange variable by fractioning the total into trapezoidal areas, and the individual areas were calculated using the trapezoid (Equation 2) through a macro of the statistical environment SAS^® 9.4 using the analysis referenced by ^{Córdoba et al. (2018)}, which was adapted from routines described by ^{Huang and Xiao (2010)} and ^{Shiang (2004)}.

Where m _i corresponds to the i_th measurement, while t i is the i_th time; therefore, the AUC is the sum of all the individual areas estimated (Equation 3).

Statistical analysis

Analysis of variance of the AUC was performed. The Tukey test separated treatments with statistical differences at P ≤ 0.05 probability. Both analyses were carried out “agricolae” package ^{(Mendiburu, 2020)} of the statistical environment ^{(R Core Team, 2020)}.

Light habitat and vapor pressure deficit

Abarco (Cariniana pyriformis Miers) generated a shade level between 9.93 and 20.3 % over cocoa canopy (Figure 2F). Under that condition, the environment temperature below the tree canopy and over cocoa canopy was between 23 and 31 °C and relative humidity was between 42 and 73 %; PAR was between 200 and 1000 μmol photons m^-2 s^-1 (Figure 2D). Increasing rates of net photosynthesis (A) have been reported as radiation (PAR) reaches photosynthetic photon flux densities (PPFD) between 300 and 750 μmol photons m^-2 s^{-1 (Avila et al., 2016}; ^{Da Matta et al., 2001}; ^{Mielke et al., 2005)}. Likewise, as shown in Figure 3, the highest incident radiation was concentrated between 200 and 500 μmol photons m^-2 s^-1, near light saturation point for cocoa, which is close to 200 y 500 µmol photons m^-2 s^-1, where plants reach 95 % of the maximum photosynthetic activity. Values close to 1000 μmol photons m^-2 s^-1 tend to reduce net photosynthesis. As explained by ^{Baligar et al. (2008)}, the photosynthetic speed decreases if the photosynthetic apparatus is exposed to PPFD higher than 1800 μmol m^-2 s^-1; thus, the exposure affects the photosynthetic apparatus of the cocoa leaf (Raja Harun and Hardwick, 1988; cited by ^{Almeida and Valle 2007}).

Figure 2 Diurnal variation of gas exchange parameters of five cocoa clones. A) net photosynthesis, B) stomatal conductance, C) transpiration, D) photosynthetically active radiation, E) leaf temperature, F) shade level at the top of Theobroma cacao trees. 

Figure 3 Net photosynthetic rate (A) and Photosynthetic Photon Flux Density (PPFD) relation of five Theobroma cacao clones planted in an agroforestry system. 

According to ^{Agudelo-Castañeda et al. (2018)}, abarco, as a forest species of the system, has been recognized for its potential to regulate incident radiation with less variability. It allows cocoa plants to achieve higher rates of photosynthesis (5.39 μmoles CO₂ m^-2 s^-1) because this species has the lowest percentage of natural leaf removal in the dry season. The aforementioned allows maintaining radiation levels suitable for cocoa cultivation throughout the year, compared to other species such as teak (Tectona grandis L. f.- Verbenaceae) and rubber (Hevea brasiliensis (Willd. ex A. Juss.) Müll Arg. - Euphorbiaceae).

The VDP reached the lowest values in the arly hours of the day (0.25 to 0.45 kPa) and in the late afternoon (0.43 to 0.3 kPa), while the highest VDP values (0.71 kPa) were registered when the temperature reached 30 °C and a relative humidity of 40 % (Figure 4). ^{Mielke et al. (2005)} indicated that high VPD values could induce stomatal closure and limit the entry and diffusion of CO₂; thus, the highest VPD near noon can explain the stomatal closure between 11:00 and 14:00 hours and, consequently, the decrease in photosynthetic activity. Similar results were found by ^{Almeida et al. (2014)}, who stated that the reduction in CO₂ assimilation is associated with the stomatal closure by increasing VPD.

Figure 4 Vapor pressure deficit (VPD) at the top of Theobroma cacao plants. *: P ≤ 0.05. 

Gas exchange

All cocoa clones, except for CCN51, reached their highest photosynthetic capacity between 08:00 and 09:00 hours, showing a maximum A of 6.6 μmol CO₂ m^-2 s^-1 (TCS06), 5.72 μmol CO₂ m^-2 s^-1 (TCS13), 4.46 μmol CO₂ m^-2 s^-1 (TCS19) and 3.99 μmol CO₂ m^-2 s^-1 (TCS01), while at noon, a reduction in photosynthetic activity was observed to rates of 2.98 μmol CO₂ m^-2 s^-1 (TCS06), 2.01 μmol CO₂ m^-2 s^-1 (TCS13), 3.12 μmol CO₂ m^-2 s^-1 (TCS19) and 1.57 μmol CO₂ m^-2 s^-1 (TCS01); after this decreasing, clone TSC19 was the only one that remained constant in the rate of CO₂ fixation with values close to 3.12 μmol CO₂ m^-2 s^-1 in the afternoon, unlike clones TCS13 and CCN51, which for the afternoon showed an increase of A, reaching values of 5.20 and 5.12 μmol CO₂ m^-2 s^-1, respectively (Figure 2A).

In general, the photosynthetic rates observed in this study were between 3.0 and 8.8 μmol CO₂ m^-2 s^-1, similar to those reported by several authors ^{(Almeida et al., 2014}; ^{Araque et al., 2012}; ^{Jaimez et al., 2018}; ^{Mielke et al., 2005}; ^{Ribeiro et al., 2016;)}. Cocoa has a relatively low net assimilation rate (A) compared to other crops ^{(Alvim, 1977)}. Cocoa A coincides with those reported for C₃ plants, as the case of coffee (Coffe arabica), with low values of A (3 to 12 μmol CO₂ m^-2 s^-1) and lower than C₄ plants, as maize (Zea mays L.), with A values of 60 to 90 μmol CO₂ m^-2 s^{-1 (Polanía et al., 1982)}.

According to ^{Agudelo et al. (2018)}, clones TSC19 and TCS13 showed the highest photosynthetic rate (4.8 and 4.36 μmoles CO₂ m^-2 s^-1, respectively) between 08:00 and 12:00 hours, compared to other regional clones, such as ICS95. In this study, clones TCS13 and TCS06 stood out by their photosynthetic performance. Regarding g_s and E, a reduction towards noon was observed, similar to the A trend (Figure 2B and 2C). The g_s values ranged from 0.03 (TCS19) to 0.06 mol CO₂ m^-2 s^-1 (TCS06 and CCN51). In terms of E, values of 1.70 (TCS01) to 2.91 mmol H₂O m^-2 s^-1 (TCS06) were recorded, coinciding with the time when the highest leaf temperature occurred (41 °C) (Figure 2E).

Throughout the day, clone CCN51 stood out for presenting the highest values of g_s at 11:00 and 14:00 hours (0.12 and 0.11 mol CO₂ m^-2 s^-1, respectively) and, consequently, higher E (4.25 and 4.52 mmol H₂O m^-2 s^-1, respectively); however, clone CCSN51 did not show the highest photosynthetic activity. Several g_s are reported, ranging from 0.02 to 0.14 mol CO₂ m^-2 s^{-1 (Araque et al., 2012}; ^{Baligar et al., 2008}; ^{De Almeida et al., 2016}; ^{Jaimez et al., 2018}; ^{Mielke et al., 2005)}, while transpiration ranged between 0.39 and 6.0 mmol H₂O m^-2 s^-1. ^{Mielke et al. (2005)}, ^{Baligar et al. (2008)}, ^{Almeida et al. (2014)} and ^{Jaimez et al. (2018)} found a direct relationship between the values of g_s and A, the greater the conductance, the greater photosynthesis (g_s 0.03 mol CO₂ m^-2 s^-1 / A 4 μmoles CO₂ m^-2 s^-1 and g_s 0.13 mol CO₂ m^-2 s^-1 / A 8.5 μmoles CO₂ m^-2 s^-1). On the other hand, ^{Araque et al. (2012)} found that variation in the parameters A, g_s, and E was related to the moisture regime; all cultivars exhibited a 60 % reduction in g_s with drought, with a 73 % decrease in A and E; meanwhile, ^{De Almeida et al. (2016)} reported a positive relationship between the values of g_s and A, but in turn, these were influenced by different water levels.

^{Jaimez et al. (2018)} showed that g_s, unlike E, is influenced by PAR variation, and that high values of g_s resulted in a greater CO₂ availability for carboxylation by Rubisco during the Calvin cycle; in contrast, ^{Baligar et al. (2008)} indicated that the main factor influencing E is VPD, presenting a direct relationship between both parameters.

The daily integral of gas exchange (Area under the curve-AUC)

There were variaritions for cocoa clones (P ≤ 0.01) in the daily integral of A, g_s, and E. The highest values of these parameters were found in clones TCS06, TCS13 and CCN51, while the lowest ones were found in clone TCS19. Clone TCS06 presented 137,300 µmol CO₂ m^-2 day^-1, corresponding to a variation of 44 and 40 % more CO₂ fixed than clones TCS19 and TCS01, respectively. A similar behavior was observed for E; clone CCN51 showed significantly (P = 0.00299) more stomatal conductance and transpiration rates than clone TCS19, reaching values of 108,300 mmol H₂O m^-2 day^-1 compared to 66,730 mmol H₂O m^-2 day^-1 in TCS19. Regarding g_s, clone TCS19 showed the lowest value (1484 mol CO2 m-2 day-1), which was significantly different for the stomatal conductance reached by the remaining four clones. Regarding the efficiency of the use of water and radiation, the clones did not differ significantly (P > 0.05) (Table 1). ^{Agudelo et al. (2018)} reported that clone TCS13 achieved a greater efficient use of water (2.8 μmol CO₂ mmol^-1 H₂O) compared to clones like ICS95, which coincides with what was found in this study, where the clone TCS13 reached the highest values in WUE (1.6 μmol CO₂ mmol^-1 H₂O).

Regarding the selection of forest species used as shade for cocoa trees, it is necessary to consider the climatic, edaphic, and environmental requirements and some morphophysiological characteristics that affect the percentage and durability of the shade on the cocoa canopy; among them, the foliar density and the longevity of the leaves ^{(Chavarría, 2013)}.

^{Agudelo-Castañeda et al. (2018)} stated that trees used in agroforestry systems influence the ecophysiological behavior of cocoa clones, mainly in photosynthesis rates. Abarco has greater benefits, especially in wet and dry seasons, which allows maintaining the physiological activity of cocoa throughout the year; thus abarco is a recommended tree for the studied region, which has two rainy seasons (March-June and September-November) and two periods of less precipitation (December-February and July-August). In this sense, cocoa-producing areas with high radiation intensity and water-deficient seasons, abarco can be helpful in agroforestry arrangements with cocoa; however, in the cocoa-producing areas with high cloudiness and periods without water deficit, abarco can be a forest alternative. Greater spacing between forest components should be considered to guarantee adequate shade and radiation availability to carry out photosynthesis. Similarly, the cocoa-abarco spacing arrangement needs to consider the light requirements of the cocoa genetic material to be used. As observed in this study, clone CCN51 requires more radiation to achieve high photosynthetic rates, unlike clones TCS 01, TCS 06 and TCS 19, which achieved higher values ​​of fixed carbon at lower radiation.