Introduction

Protected agriculture, based on the use of greenhouses and hydroponic systems, represents an effective strategy for reducing risks in agricultural production and improving economic profitability, especially for small producers (Sánchez-del Castillo & Moreno-Pérez, 2017). The implementation of these technologies in crops of high yield and economic value ensures sufficient income for a dignified life and generates a positive impact on local employment, by creating 10 full-time workdays per hectare, which contributes to regional economic development (Pratt & Ortega, 2019).

Tomato (Solanum lycopersicum L.) is the most widely cultivated vegetable in greenhouse and hydroponic systems in Mexico and several countries around the world (Costa & Heuvelink, 2018; Servicio de Información Agroalimentaria y Pesquera [SIAP], 2020). In 2018, Mexico exported 1.68 million tons, equivalent to 48.7 % of the national production estimated at just over 3 million tons. Of this volume, 99.7 % was exported to the United States (Fideicomisos Institudos en Relación con la Agricultura [FIRA], 2019).

Researchers at the Chapingo Autonomous University (UACh, by its acronym in Spanish) developed and commercially validated an alternative system for hydroponic tomato production. This system allows for a 50 % increase in annual yield compared to the conventional greenhouse system, without increasing the cost per ton produced. This has resulted in greater economic benefits for producers who sell primarily in local markets (Sánchez-del Castillo & Moreno-Pérez, 2017; Sánchez-del Castillo et al., 2012; Sánchez-del Castillo et al., 2010).

The developed alternative system (Figure 1) is characterized by two cultivation stages: 1) the seedling stage in the seedbeds (from sowing to transplanting), and 2) the final production stage (from transplanting to the end of the harvest). Unlike the conventional system, which contemplates a single annual cycle, this model allows three cultivation cycles per year. To this end, cultural practices have been implemented, such as delaying transplanting until 45-50 days after sowing (das) and topping plants (removing the terminal bud) above the third inflorescence. Although this strategy reduces individual yield per plant, the smaller leaf area generated allows for an increase in population density to up to 8 plants∙m^-2, which partially offsets this reduction. Furthermore, shortening the transplant-to-harvest period to less than 110 days allows for three crop cycles per year (Sánchez-del Castillo et al., 2012; Sánchez-del Castillo et al., 2017a), with potential annual yields close to 450 t∙ha^-1 (Moreno-Pérez et al., 2022; Sánchez-del Castillo et al., 2021).

For the development and validation of the high population density system, several factors were evaluated, such as obtaining older seedlings at transplant to shorten the productive cycle (López-Valencia et al., 2002; Sánchez-del Castillo et al., 2012), the selection of economical and durable substrates and containers (Caraveo-López et al., 1996; Pineda-Pineda et al., 2019), productivity and precocity depending on different tipping levels (one, two or three clusters per plant), the optimal density within each tipping level (Sánchez-del Castillo et al., 2012; Sánchez-del Castillo et al., 2017a; Sánchez-del Castillo & Ponce-Ocampo, 1998), the evaluation of varieties (considering yield, quality and market preferences) (Sánchez-del Castillo et al., 2017a; Sánchez-del Castillo et al., 1999), vegetative propagation by cuttings to reduce costs due to the high demand for hybrid seed (Juárez-López et al., 2000; Mejía-Betancourt et al., 2023; Moreno-Pérez et al., 2016), and the spatial and temporal arrangement of plants to optimize the interception of solar radiation (Méndez-Galicia et al., 2005; Sánchez-del Castillo et al., 2017b; Vázquez-Rodríguez et al., 2007).

Despite the increased use of seeds, trays, and germination substrate, production costs per cycle remain, on average, lower than those generated with the conventional single-cycle system. In this sense, the cost per kilogram produced is similar, but there is a greater annual profit with three crop cycles (Sánchez-del Castillo et al., 2012; Sánchez-del Castillo et al., 2017b; Sánchez-del Castillo et al., 2021).

In the pursuit of greater annual productivity, early pruning has been evaluated to obtain only one or two clusters per plant, with proportional increases in population density. However, these adjustments have proven impractical due to the decrease in yield per plant, the increase in cultivation tasks, and the higher cost per cycle due to the intensive use of hybrid seeds (Sánchez-del Castillo et al., 2012; Sánchez-del Castillo et al., 2017a).

Recently, at the UACh a new proposal has been developed based on three cultivation stages: 1) seedlings in seedbeds, 2) high-density intermediate stage, and 3) final production stage. By topping to three clusters per plant and including the intermediate stage, which lasts 35 to 45 days, the final production stage has been reduced from 100-110 days to 50-60 days, allowing for up to seven cultivation cycles per year (data publication pending).

The seedbed stage maintains its approximate duration of 45 days. Subsequently, the seedlings are transplanted into bags or pots with hydroponic substrate and carried with wire rod staking in an additional smaller greenhouse, where the plants are managed at high population density (18 to 25 plants∙m^-2) for 35 to 45 days. To ensure economic viability, this intermediate stage should not occupy more than 30 % of the greenhouse surface area designated for the final stage. Between 80 and 90 das (i.e., 35 to 45 days after transplanting) the plants are moved to the final production greenhouse (7 to 8 plants∙m^-2), thus beginning the final production stage, which lasts 50 to 60 days (Figure 2).

The viability of the three-stage system with three clusters is based on the fact that, at 45 days, the plants have a reduced leaf area, which allows their temporary management at high density without the leaf area index (LAI) generating problems of competition for light. The combination of the seedbed and intermediate stages allows the crops to be staggered over time and reach up to seven annual cycles in the final production greenhouse. This represents the potential to double the annual yield compared to the previous three-cycle, two-stage system. However, overextending the duration of the intermediate stage can induce light competition due to increased leaf area and LAI (Jishi, 2018), which could negatively affect plant growth and yield (Taiz et al., 2018). Heuvelink et al. (2018) and Heuvelink et al. (2020) suggest that the optimal LAI to maximize tomato production is between 3 and 4.

An aspect that remains to be resolved, to successfully implement this three-stage production system, is to define the optimal population densities for the intermediate and final stage, which allow maintaining the LAI within adequate ranges to avoid excessive competition and, at the same time, ensure maximum yield per unit area in each cycle.

Based on the above, this study aimed to evaluate the impact of plant population density during the intermediate and final stages of cultivation on the agronomic behavior and yield per unit area in tomato plants topped at the third cluster. The evaluation considered that the area occupied by plants during the intermediate stage should not exceed 30 % of the area required for the final production stage.

Materials and methods

The experiment was established at the facilities of the Horticulture Institute of the Chapingo Autonomous University, Mexico (19° 29’ 35’’ N and 98° 52’ 19’’ W, at 2 250 m a. s. l.), under greenhouse and hydroponic conditions during a 143-day crop cycle (from sowing to the end of harvest), between March and July 2023.

A chapel-type greenhouse with a metal structure and a high-light-dispersion thermal polyethylene roof was used, oriented north-south. The climatization system included LP gas heating, a wet wall cooling system, and ventilation with exhaust fans located on the opposite wall. The floor was covered with white polypropylene ground cover. The three growing stages (seedbed, high-density intermediate, and final production) took place in different spaces within the same greenhouse.

A tomato hybrid Pai-Pai® (Enza Zaden), saladette type and indeterminate growth habit, was used. Seeds were sown in 60-well polystyrene trays, with a volume of 200 cm³ per well and a 5 cm separation between wells. A 1:1 (v/v) peat moss and perlite mixture was used as the seedbed substrate. The seeds were covered with a 0.5 cm layer of vermiculite.

Seedlings were irrigated with a 50 % nutrient solution for the first 30 das. Subsequently, a 100 % nutrient solution was applied, containing the following nutrient concentrations (mg∙L^-1): N = 200, P = 60, K = 250, Ca = 250, Mg = 50, S = 150, Fe = 3, Mn = 1, B = 0.5, Cu = 0.2, and Zn = 0.2. Calcium nitrate, potassium sulfate, 85 % phosphoric acid, magnesium sulfate, ferrous sulfate, manganese sulfate, boric acid, copper sulfate, and zinc sulfate were used as nutrient sources.

The nutrient solution was prepared in a reservoir and distributed using drip lines, with drippers spaced 20 cm apart for the intermediate stage and 30 cm apart for the final stage of production. Irrigation frequency and the amount of nutrient solution applied were adjusted daily to ensure 20 % drainage relative to the volume of solution applied per irrigation in each container. Between four and eight daily irrigations were provided depending on weather conditions and the phenological stage of the cultivation.

The seedlings were transplanted at 45 das in 17-cm-diameter, 30-cm-deep pots with a capacity of 6.8 L of substrate. The substrate consisted of red tezontle sand, with particles between 1 and 3 mm in diameter. A 1.2 m long piece of wire rod was placed in each container to serve as a support for the plant; the plants were secured with plastic rings.

Plants in all treatments were topped (removal of the terminal bud) between 75 and 80 das, leaving two leaves above the third inflorescence. Side shoots were removed as they appeared, leading the plants to a single stem and three clusters each. Plants remained in the intermediate stage for 45 days at the density defined for each treatment. At the end of this stage (90 das), the plants were relocated (with a pot and support) to the location assigned for the final production stage and at the density corresponding to the treatment; this stage lasted 53 days. Harvesting began at 120 das, and successive cuts were made over 23 days.

A randomized complete block experimental design was established with four treatments, four replicates, and 12 plants per experimental unit. Treatments were defined by population density in the intermediate and final stages (Table 1). In all cases, six rows of plants with 50 cm wide corridors were established:

Table 1. Population density of tomato cultivation in two stages of the three-stage system.

To estimate the effect of high density in the intermediate stage on yield and its components, four treatments were included as controls based on two cultivation stages, also managed at three clusters per plant. These treatments consisted of transplanting at 45 days after transplanting and a final stage of 98 days with densities of 8, 7, 6, and 5 plants∙m^-2. The controls were compared only with their equivalent three-stage treatments; that is, with those that shared the same density in the final production stage.

To evaluate the effect of light competition on plant development, growth variables were measured at 70 and 90 das (coinciding with the middle and end of the intermediate stage), and yield variables at the end of the crop cycle. The growth variables evaluated were:

• 1.

  Leaf area per plant (cm²): determined with a leaf area meter (LI-3000A, LI-COR®, USA).

• 2.

  Leaf area index (m²∙m^-2): was obtained from the total leaf area (m²) per m² of covered surface.

• 3.

  Stem diameter (mm): measured with an electronic vernier caliper (model 14388, Truper®, China) between the fourth and fifth node.

• 4.

  Plant height (cm): measured with a flexometer from the base of the stem to the apex.

• 1.

  Number of fruits harvested per plant (fruits∙plant^-1): the fruits harvested in each cut for each experimental unit were counted and divided by the number of plants present per experimental unit.

• 2.

  Number of fruits per unit of surface (fruits∙m^-2 of greenhouse): it was calculated by multiplying the number of fruits per plant by the final population density of each treatment.

• 3.

  Average fruit weight (g): obtained by dividing the total fruit weight per unit area (g∙m^-2 of greenhouse) by the number of fruits harvested from that area (fruits∙m^-2). The fruits were weighed on an analytical balance (MC-173467, Ohaus®, USA).

• 4.

  Yield per plant (g∙plant^-1): was determined from the sum of the weight of the fruits of all the cuts in each experimental unit divided by the number of plants in the mentioned unit.

• 5.

  Yield per unit of surface (kg∙m^-2 of greenhouse): calculated from the yield per plant multiplied by the population density of each treatment at the end of production.

The data obtained were subjected to analysis of variance and comparison of means tests (Tukey, P ≤ 0.05). To estimate the effect of density in the intermediate stage on yield and its components, orthogonal contrasts were performed on pairs of treatments with the same density in the final stage; that is, with and without an intermediate stage. All analyses were performed using the SAS Institute statistical software (2002).

Results and discussion

According to the means comparison test, the treatment with a density of 25 plants∙m^-2 in the intermediate stage presented an average plant height of 87 cm at 70 das, a value significantly higher than that recorded with the lowest density (17 plants∙m^-2), whose plants reached an average height of 76 cm (Table 2).

Table 2. Comparison of means of growth variables of tomato plants at 70 and 90 days after planting managed with different population densities in their intermediate cultivation stage.

Although LAI values were not as high at 70 das, an elongation effect was already observed, probably due to mutual shading between plants generated by the increase in leaf area at high densities (Jishi, 2018; Taiz et al., 2018). Although the differences were not statistically significant, the treatment with 25 plants∙m^-2 in the intermediate stage showed a trend towards a higher LAI (5.51 m²∙m^-2), approximately 20 % higher than the treatment with 17 plants∙m^-2 (4.01 m²∙m^-2).

At 90 das, following the pruning procedure performed at 80 das, the plants no longer experienced an increase in height, so no significant differences were observed between treatments in this variable. No differences were also found in stem diameter and leaf area per plant. However, the LAI was statistically higher (greater than 7 m²∙m^-2) in the two highest-density treatments (30 and 25 plants∙m^-2), compared to the lowest-density treatment (17 plants∙m^-2), with a value of 4.94 m²∙m^-2.

A more detailed analysis of the data in Table 2 reveals that, during the last 20 days of the intermediate stage, plants in the higher density treatments showed an increase of almost 40 % in their leaf area per plant and in their LAI, while this increase was only 25 % in the lower density treatments. Specifically, the treatment with 25 plants∙m^-2 increased its leaf area per plant by 853 cm² (an average of 42.65 cm² per day) and its LAI by 2.1 m²∙m^-2. In contrast, the treatment with 17 plants∙m^-2 only increased its leaf area by 518 cm² (an average of 25.9 cm² per day) and its LAI by 0.93 m²∙m^-2. These results suggest that plants at high densities experienced light competition stress at early stages, as noted by Jishi (2018) and Taiz et al. (2018), which can have negative effects on plant yield (Heuvelink et al., 2020; Papadopoulos & Pararajasingham, 1997; Taiz et al., 2018).

Comparison of the means of the yield variables and their components shows that the number of fruits per plant was similar among the treatments evaluated (Table 3). However, the treatment with densities of 30 plants∙m^-2 in the intermediate stage and 8 plants∙m^-2 in the final stage had the highest number of fruits per unit area. The highest average fruit weight was obtained in the treatment with 17 and 5 plants∙m^-2 in the intermediate and final stages, respectively, which differed statistically from the treatment with the highest density (30 plants∙m^-2 in the intermediate stage and 8 plants∙m^-2 in the final stage). Regarding yield per plant, the treatment with 17 plants∙m^-2 in the intermediate stage and 5 plants∙m^-2 in the final stage showed the highest statistical value (1.51 kg∙plant^-1). However, the four treatments evaluated showed similar yields per unit area.

Table 3. Comparison of means of the yield variables and their components in tomato plants managed with different population densities under the three-stage system.

The results confirmed that the increase in plant density during the intermediate stage has a negative effect on the components of plant yield. Both the number of fruits per plant and the average fruit weight decreased as plant density increased during this stage, resulting in a reduction in individual plant yield. However, the number of fruits per unit area was higher in treatments with higher densities, which partially offset the lower individual fruit weight and resulted in similar yield per unit area across treatments.

Considering that with the three-stage system, up to seven crop cycles per year can be obtained by reducing the final production stage to between 50 and 60 days, any of the evaluated treatments could achieve an annual yield greater than 50 kg∙m^-2∙year^-1 (500 t∙ha^-1∙year^-1).

When comparing the pairs of treatments of the three-stage system with the two-stage system, at the same population density in the final stage (Table 4), it was observed that the high densities in the intermediate stage (30 and 25 plants∙m^-2) of the three-stage system significantly reduced the number of fruits per plant and per m² compared to the two-stage system under the same final density. However, this was not the case at low densities in the intermediate stage (20 and 17 plants∙m^-2). A similar trend was detected for yield per plant and per unit area, although in this case the density of 20 plants∙m^-2 also showed significant differences in favor of the two-stage system. Furthermore, the average fruit weight was negatively affected under the higher density treatment.

Table 4. Orthogonal contrast tests of the means of the yield variables and their components between the three- and two-stage tomato cultivation treatments with the same population density in the final stage.

The observed performance when comparing both systems can be explained by the fact that the greater increase in LAI caused by high densities during the intermediate stage intensified competition for solar radiation among plants, which negatively affected the absorption of photoassimilates per plant and, consequently, reduced daily dry weight gain (Heuvelink et al., 2020). These factors affected the final yield compared to plants that did not have an intermediate stage.

Although yield per cycle was lower in plants managed with three stages, the possibility of achieving more than double the number of crop cycles per year with this scheme considerably increases the potential annual yield, which could translate into a more profitable system for the producer. However, a detailed cost-benefit analysis is needed to support this hypothesis. For example, the two-stage system with a final density of 8 plants∙m^-2 generated a yield of 12.7 kg∙m^-2 per cycle, and allows for up to three cycles per year (Sánchez-del Castillo et al., 2012), which is equivalent to a potential annual productivity of 38.1 kg∙m^-2∙year^-1. In contrast, the three-stage system, with an intermediate stage of 30 plants∙m^-2 and a final stage of 8 plants∙m^-2, yielded 8.8 kg∙m^-2 per cycle, but with the possibility of carrying out up to seven cycles per year, representing a potential annual productivity of 61.6 kg∙m^-2∙year^-1; that is, 62 % more production than the two-stage system. Although this system involves higher production costs, these are likely to be offset by the increase in annual yield.

It should be noted that the yield per unit area per cycle was lower than that reported in other studies that used the three-cluster system in two cultivation stages with similar densities in the final stage. In this regard, Sánchez-del Castillo et al. (2017a) obtained yields of 15 to 18 kg∙m^-2 per cycle with the cultivar ‘Moctezuma’ at a density of 7 plants∙m^-2, while Moreno-Pérez et al. (2022) reported 13 to 14 kg∙m^-2 with the cultivar ‘El Cid’ at a density of 8 plants∙m^-2. In both studies, they used beds filled with tezontle sand with a depth of 30 cm as containers, which provided a volume of 25 to 30 L of substrate per plant, in contrast to the 6.8 L used in the present study.

According to Heuvelink et al. (2018) and Heuvelink et al. (2020), the highest yield can be achieved with an LAI close to 4, which is consistent with the results obtained. Therefore, it is recommended to use population densities of 7 to 8 plants∙m^-2 in the final production stage. Furthermore, it is suggested to explore the use of bags or pots with a greater volume of substrate per plant, since this could improve yield by providing better conditions for root development, as well as improving oxygenation and the retention capacity of the nutrient solution, with fewer fluctuations in pH and electrical conductivity between irrigations (Suazo-López et al., 2014).

Conclusions

In the three-stage tomato production system, higher population densities during the intermediate stage (25 and 30 plants∙m^-2) caused internode elongation at 70 days after planting and increased leaf area index at 90 days, compared to plants established at lower densities in the same stage.

Although the three-stage system presented a lower fruit yield per cycle compared to the two-stage system when the same density was used in the final stage, its greater number of possible cycles per year (up to seven) provides a significantly higher potential annual yield per unit area.

The tomato plants established at a density of 25 plants∙m^-2 in the intermediate stage and 7 plants∙m^-2 in the final stage reached a yield of 8.9 kg∙m^-2 per crop cycle, with the possibility of obtaining seven cycles per year, which represents an annual potential of 62.3 kg∙m^-2 (623 t∙ha^-1∙year^-1). Therefore, it is proposed to use this combination of densities in the production protocol of the three-cluster tomato cultivation system in three stages that is being developed.