Introduction

ydroponic production of bell peppers (Capsicum annuum L.) in greenhouses began in Mexico twenty years ago. Mexican producers have adopted production systems developed in other countries, principally those in Europe with different climatic and socioeconomic conditions. Bell pepper plants that are cultivated in greenhouses show indeterminate growth (^{Jovicich et al., 2004}) and branching in their stems, therefore dividing the principal stem in two or more branches. Each branch develops one to three leaves, it then branches again, and this growth repeats successively; normally in each branching, a fruit is formed. Conventionally, two production systems are used based on this growth. One consists of pruning to maintain each plant with two stems (prune in a V shape, also called the Dutch system). In this system population densities are established in 2 or 3 plants m^-2 (^{Heuvelink et al., 2004}). In the other system, the plant grows freely in a bush-like form (Spanish System) and has 3 plants m^-2 (^{Jovicich et al., 2004}). In both systems, the growing cycle lasts eight to ten months from transplant to harvest, in addition to 45 to 60 more d for the seeding to develop in the seedbed: In this system, it is possible to obtain only one growing cycle per year (^{Heuvelink et al., 2004}).

In the Spanish system, the first fruits developed exert greater competition for photoassimilates with the vegetative and reproductive growth that continues in the indeterminate fashion. The main consequence is an elevated number of aborted fruits at the beginning of its development. After harvesting the first fruits, the new fruits stop aborting. This causes the production to be obtained in cultivation flows separated by a two to three-month time period during the growing cycle; In this system, there are two harvesting flows and a yield around 100 t ha^-1 (^{Cantliffe y Vansickle 2001}; ^{Jovicich et al., 2004}).

Two-stem pruning method, the balanced distribution of the sugars from the photosynthesis between the vegetative and reproductive growth is sought, along with producing only one fruit in each joint where the stems branch. This way, the abortive fruits almost completely avoid one another and the harvest is continuous throughout the year. However, the accumulation of production time is slow and it is difficult programming the production for markets with short time frames of price favorable for the producer. In high-tech greenhouses, yields of up to 250 t ha^-1 are achieved (^{Cantliffe y Vansickle, 2001}; ^{Heuvelink et al., 2004}), although the costs of production are very elevated.

Bell pepper cultivation, from beginning to end of the harvest is long. Generally, the harvest is obtained when the value of the product greatly fluctuates in the market and the best prices are obtained in specific windows, normally well-defined and in a short time frame. Thus, for the producers of this crop, it would be desirable to dispose of production systems that allow the harvest to be concentrated when the prices are highest.

There are regions with extreme climates that hamper the cultivation of the pepper for various months of the year, even in a greenhouse. Having a production system with a short growing cycle and high population density, could enable the concentration of the harvest, thus avoiding these adverse conditions with lower production costs; furthermore, the risk of crop disease could be reduced due to the short cycle (^{Reséndiz et al., 2010}). Occasionally, a system like the one just described, could, in temperate climates, obtain various growing cycles per year, with yields similar to those in the north of Europe, but with lower production costs (^{Cruz et al., 2009}).

The growing cycle of the pepper could be cut three or four months after the transplant with seedlings 60 d old, when the terminal buds are trimmed from all the branches above the third or fourth branch (^{Cruz et al., 2009}; ^{Reséndiz et al., 2010}). The population density increases to compensate for the lower yield of the plant because of the smaller leaf area per plant.

Based on this, the objective of this study was to evaluate the yields of two bell pepper varieties, budding above the third or fourth branch of the stem, established in three population densities and with transplant growing cycles in order to equal or reduce the four months of harvest, thus increasing annual production.

Materials and Methods

The study took place in a greenhouse with the dimensions of: 50 m in length, 11 m in with, and 4 m in height in the experimental field of the Autonomous University of Chapingo, State of Mexico (19° 29’ N, 98° 53’ O y 2250 m above sea level).

Orion and Triple Star are the varieties that were evaluated. The seedlings grew in pots that had 700 cm³ and with volcanic tuff (tezontle, for its name in Spanish) and particles of 1 to 3 mm in diameter that were used as substrate. For the first eight days, they were irrigated with water; a week after their emergence and until their transplant, a nutritive solution with (mg·L^-1): N 100; P 30; K 125; Ca 125; Mg 30; S 110; Fe 1.5; B 0.3; Mn 0.3; Cu 0.05; Zn 0.05, was used. After the transplant and until the end of the cycle, the nutritive solution contained double of what was used for the seedling, with an electrical conductivity of 2.5 dS m^-1 and pH between 6 and 6.5.

The transplant was carried out with 60-day-old seedlings in an open hydroponic system of cultivation beds (1.2 m wide and 0.3 m deep), filled with red volcanic tuff (particles of 3 to 5 mm in diameter) and a drip irrigation system (in strips with transmitters every 20 cm). The plants were supported by a strand of raffia at the base of the stem, fixed to wires along the greenhouse.

The experimental design was a complete random block design with four repetitions and factorial arrangement of treatments. The factors and levels were:

Trimming (TL): elimination of apical buds in the third or fourth branch; Population density (PD): 5, 6.5, and 8 plants per m^-2 of greenhouse; Variety (V): Orion and Triple Star. With the combination of the factors and levels, 12 treatments were carried out. The experimental unit was 1.5 m².

The evaluations included:

Index of leaf area. Leaf area, which was measured with an integrator (LICOR^® 300 Lincoln, Nebraska), was divided between the plant surface per plant in each density.

Fruit set (%FS). The number of fruits per plant was subtracted from the number of flowers in anthesis and was expressed in percent.

Total dry weight of the plant (TDW). The leaves, the stem, the root, and the fruit of two plants for each experimental unit were dried at; the dry weight was expressed in grams.

Median weight of fresh fruit (MFW) was expressed in grams.

Furthermore, the number of fruit per plant (NFP); the number of fruit per m² (NFA), yield (kg) per plant (YP) and yield per area (YA) in kg·m^-2.

The analysis of the data included a variance analysis (ANOVA) and comparison of means (Tukey, p≤0.05).

Results and Discussion

The variance analysis (Table 1) showed highly significant effects for each factor in the majority of the variables, and only in some cases was the interaction significant. The interaction between trimming levels (TL) and the population density (PD) resulted highly significant for the number of fruits per plant and unit of area, whereas the SL with variety (V) resulted highly significant for the percentage of fruit set and the number of fruits per area (Table 1). All the other interactions were not significant. These interactions allow us to explain why with less density, the number of fruits per plant was greater with trimming at the fourth branch than with the third (7.2 and 4.4, respectively). By increasing the density from 5 to 6.5 plants per m^-2, there was a reduction in the number of fruits per plant at both levels of trimming. An explication of these results is that with the pruning at the third branch, the plants formed little leaf area; consequently, upon passing from 5 to 6.5 plants m^-2, the interception of the active photosynthetic radiation (APR) per plant was not notably affected and the number of fruits per plant stayed the same. Contrastingly, with the pruning at the fourth branch, the leaf area per plant was greater than with the pruning at the third branch; for which the competition for the APR certainly increased. There was less photoassimilates per plant and with it, greater abortion. With the increase in density, from 6.5 to 8 plants m^-2, the pressure of the density affected the interception of the APR at both levels of trimming. ^{Reséndiz et al. (2010)} evaluated 17 varieties of bell pepper, with 4 and 6 plants m^-2 and pruning at the fourth branch; and they also observed that with the high density, the number of fruits per plant decreased, although per unit of area there were no differences.

For the trimming at the third branch, with the density of 5 to 6.5 plants m^-2, the number of fruit increased from 35 to 44; with the trimming at the fourth bifurcation, the number of fruit was the same (58). Upon increasing from 6.5 to 8 plants m^-2, there was no increase at any level of trimming.

The main significant effect of the variables was observed in the main factors. Therefore, the variables were analyzed per study factor.

Table 1 Mean squares of the variance analysis of bell pepper plants; Orion and Triple Star varieties, trimmed at the third and fourth branch in three population densities 

SV: Source of variation, DL: degrees of liberty; LAI: leaf area index, %FS: percentage of fruit set, TDW: Total dry weight; MFW: Median Fruit weight; YP: yield per plant; YA: yield per area (m²), NFP: number of fruits per plant; and NFA: number of fruits per m².

Levels of trimming

Trimming above the fourth branch increased LAI, total dry weight (TDW), yield per plant (YP), yield per area (YA), the number of fruits per plant (NFP) and fruits per area (NFA); however, there was a decrease in the percentage of fruit set (%FS) with regards to the trimming at the third branch (Table 2).

Table 2 Comparison of variables means in bell pepper plants trimmed at the third and fourth branch. 

LAI: leaf area index; %SF: percent of fruit set; TDW: total Dry weight; MFW: median fruit weight; YP: yield per plant; YA: yield per area; NFP: number of fruit per plant; and NFA: number of fruit per area.

The LAI of the plants pruned at the fourth branch was between the values signaled as optimum for the pepper (^{Cruz et al., 2005}; ^{Cruz et al., 2009}), which increased the interception of solar radiation per canopy, increased the dry weight per plant and the yield per plant and per area (Table 2).

The greatest yield (6.4 kg m^-2) with the pruning at the fourth branch was also due to the greater number of fruits per plant and per unit of area, without affecting the median weight of the fruit. This higher number of fruits pruned at the fourth branch was 15 (equal to one per bifurcation), while the pruning at the third branch would only obtain seven fruits per plant. The number of fruit per plant with the pruning at the fourth branch in this study is similar to that reported by ^{Reséndiz et al. (2010)}.

The temperature, the relative humidity and the hormones affect the source-demand relations and the fruit set (^{Heuvelink et al., 2004}; ^{Marcelis et al., 2004}; ^{Peil y Galvez, 2005}). Particularly, the cytokinins play an important role in cellular division and the demand of sugars (^{Lambers et al., 2008}); they can have effects on the fruit set, growth of the fruit and in the final yield. However, the low percentage fruits in this study can be attributed to the high quantity of cloudy days during the development of the fruits. This effect was documented by ^{Marcelis et al. (2004)}, ^{Urrestarazu (2004)} and ^{Reséndiz et al. (2010)}. Based on meteorological data from the Autonomous University of Chapingo (data not shown), 75 % of the days of the growing cycle were cloudy or partially cloudy; as such, the photoassimilates were probably insufficient to maintain the fruit on the plant, along with its subsequent development, as the pepper requires high irradiance (^{Turnes y Wien, 1994}; ^{Urrestarazu, 2004}), at least 700 µm m^-2 s^-1 in the day during flowering and fruition (^{Guenkov, 1983}). This photon flow was not reached on the cloudy days of our study.

Population density

With the increase in the population density from 5 to 6 plants m^-2, we would expect to see an increase in the yield per area as well. This effect was not observed, and quite the contrary, it significantly decreased (Table 3). With this population density, TDW, MFW, YP, YA, and NFP also significantly decreased, probably because of the low irradiance (^{Jovicich et al., 2004}) for the high population density and the cloudy days during flowering. The sugars, product of the photosynthesis, could be insufficient to ensure the fruit set when the density increases, for when the contribution of sugars decreases in bell peppers, the absorption of flowers and fruits increase linearly (^{Marcelis et al., 2004}). The decrease in the dry weight of the plant as well as the increase in population density was also documented by ^{Cruz et al. (2005)} and ^{Agarwal et al. (2007)}.

Given that there were no significant differences in the mean weight of the fruit between both trimming systems, it can be deduced that in the environmental conditions during the development of growth, with pruning at the third and fourth branch, the appropriate density was 6.5 and 5 plants m^-2. With the latter density, more fruit per unit area was obtained, even though the yield per m² was the same in both densities. For a blocky-type fruit, with good texture and form, like those in the study, and with mean weight in both densities, these are acceptable for the national market and for export.

Table 3 Comparison of variable means in bell pepper plants for three population densities. 

LAI: leaf area index; %SF: percent of fruit set; TDW: total Dry weight; MFW: median fruit weight; YP: yield per plant; YA: yield per area; NFP: number of fruit per plant; NFA: number of fruit per area; MHSD: minimum honest significant difference.

Varieties

The YA was greater in the Orion variety than in the Triple Star (Table 4). The number of fruits per area was the same in both varieties, for which the difference was due to the fact that the mean weight of the Orion fruit was 50 g per fruit, more than that of Triple Star. Orion also exhibited a greater TDW. Even though the LAI of Triple Star was statistically greater, both values are close to three, which is adequate for the high interception percentage of APR (^{Gardner et al., 1995}).

Table 4 Comparison of the means of bell pepper plants trimmed at the third and fourth branch in three population densities. 

LAI: leaf area index; %SF: percent of fruit set; TDW: total Dry weight; MFW: median fruit weight; YP: yield per plant; YA: yield per area; NFP: number of fruit per plant; NFA: number of fruit per area; MHSD: minimum honest significant difference.

When the trimming occurred at the third branch, the two varieties showed high fruit set percentages (62 % in Orion and 66 % in Triple Star), but when the trimming occurred at the fourth branch, both varieties decreased in %SF, but the great effect was observed in Triple Star (50 % compared to 39 %). That means that only close to half of the fruits that could potentially be developed were obtained. ^{Reséndiz et al. (2010)} also obtained high percentages of aborted fruit, in his case, it was higher than the present study. With the trimming at the third branch, NFA was also similar in both varieties, but with the trimming at the fourth branch, the Orion variety had a greater NFA than Triple star, which depended on the percent of fruit sets.

A growing cycle lasts four months from the transplanting of the seedlings at 60 d and with the trimming at the fourth branch if the fruit is harvested at its final color, and only three months if it is harvested unripe (^{Reséndiz et al., 2010}). As such, a favorable environment, like that achieved in a well-designed and equipped greenhouse, could potentially obtain three and four cycles per year. If a yield per cycle is considered (6.4 kg m^-2), like in our study, the plants pruned at the fourth branch, in three cycles per year, 200 t ha^-1 could be obtained per year. This is similar to the yield recorded in high-tech greenhouses that use long-cycle (^{Paschold y Zengerle, 2000}; ^{Jovicich et al., 2004}).

These short-cycle systems can also allow greenhouse production in places that have extreme climates most of the year, for a growing cycle can be programmed to coincide with the season of the year when the climate is most favorable. Thus, heating or cooling costs for the greenhouse could be avoided and the cost of production would decrease, while at the same time obtaining more innocuous fruits due to reducing pests and diseases as well as reducing the use of chemicals that are applied. Additionally, with the plants pruned in this manner, lower greenhouses can be used, like the one in this study, with lower cost than the greenhouses are normally used when the plant is grown conventionally higher.

Conclusions

The pruning system at the fourth branch produced a greater yield per area, without affecting the mean weight of the fruit, when compared to the trimming at the third branch

With the trimming at the third branch, the appropriate population density is 6.5 plants m^-2 and with the pruning at the fourth branch, it is 5 plants m^-2.

Both levels of trimming and population densities for the Orion variety showed greater yield per unit of area than the Triple Star variety.