Experiment 1: Effect of the flower eliminating intensity upon the growth rate, abortion rate, and size and fruits number

Plants were established at a density of six plants per m², the experimental design was completely randomized design with three replications and experimental units of 16 plants. The treatments were: 1) flowers removal at first node or central fork of the plant (27 ddt), 2) flowers removal at the first two nodes (46 ddt) and, 3) flower removal of from the first three nodes (53 ddt).

In the three central plants of each plot the length and diameter of the fruits were measured weekly, starting once 15 mm were reached. Harvested fruits were weighed and length and diameter were measured. Weekly harvests started at 119 ddt and ended at 319 ddt.

Additional plants were used to remove fruit during the growth, covering all stages up to maturity. They were measured (diameter and length), weighed, oven dried and weighed again. Two regression curves were developed for fresh weight and biomass, according to the product of the fruits dimensions (diameter×length).

Using the measures of the sampled fruits, regression equations of fresh weight and weekly accumulated biomass during growth were estimated. Data in diameter, length, fresh weight and biomass were treated in the same manner and separately, setting the next (logistics) nonlinear model:

where Y: response variable; D: days of observation; φ ₁: asymptotic value or maximum value of Y; φ ₂: time in which half the value of the asymptote is reached; φ ₃: elapsed time between the time half was reached and ³⁄₄ of the asymptote.

As a result of this model, the instantaneous growth rates were estimated using the first derivative based on the daily observation. The aborted and developed reproductive structures up to 15^th node were also counted.

Experiment 2: Influence of the sink strength on the CO ₂ assimilation rate at different plant nodes

The plants were distributed in nine rows inside a greenhouse. Five plants randomly distributed in the central row were weekly used to measure the diameter of the first five fruits: one from the first node (central), two from the second and two from the third node (without floral pruning).

In the same plants gas exchange parameters (CO₂ assimilation (A), stomatal conductance (gs), transpiration (E), radiation (PAR), ambient CO₂ concentration (Ca) and intercellular CO₂ concentration (Ci) were measured with an open type IRGA gas analyzer (Infra-Red gas Analyser ADC, model LCA4). Those leaves that accompanied the first five fruits and an extended apical leave were taken into account. The gas exchange measurements were performed at 109, 126 and 150 ddt.

Chlorophyll a fluorescence was measured with a portable fluorimeter (Model PAM-2100, Heinz Walz GmbH, Germany) in the same leaves used to measure gas exchange. The fluorescence variable evaluated were: quantum yield of photosystem II ΦPSII=Fm´-Fs/Fm´, photochemical extinction coefficient (q _P ) and non-photochemical (q _N ) calculated: qP=Fm´-Fs/Fm´-Fo´, qN=Fv-Fv´∕Fv, and the electron transport rate (J =Φ _PSII ×PAR×0.84×0.5) (^{Genty et al., 1989}).

The chlorophyll a fluorescence was measured three times during the crop cycle (119, 167 and 185 ddt). That is to say, at full fruit production, post-harvest and the end phase of the leaves; these measurements were done at 28 °C temperature and 1000 𝛍mol m^‒2 s of photon flux density (DFFF).

In all cases, Kruskal Wallis ranking test method was applied with p≤ 0.05.

Effect of the intensity of flowers removal on the growth rate, abortion rate, size and fruit number

The yield showed fluctuations in all three treatments. In all cases two periods of high yield (p≤0.05) were recorded. The first of these occurred between 119 and 167 ddt and the second between 230 and 245 ddt (Figure 2). The offset of the start of production between treatments is related to the flower removal intensity. This procedure delayed fruiting. The mismatch disappeared and all plants coincided the following harvest maximum (217 ddt). During periods of high harvest, yield was similar for all treatments (p>0.05).

Figure 2: Dynamics of production (kg per plant) of Capsicum annuum (cv. P1216) cultivated in three conditions of flowers elimination: first node (○), up to second node (□), up to the third node (∆). 

This cyclic behavior seems characteristic of the Capsicum genus, given what is reported for C. annuum cultivars with different fruits sizes (^{Wubs et al., 2009a}) and C. chinense (^{Jaimez et al., 2000}; ^{Jaimez and Rada, 2006}).

In C. annuum cultivation the flowering rate decreases during the fruit development and vegetative growth alternates with the reproductive growth (^{Fontes et al., 2005}). When fruiting rate approaches zero, the growth rate of stem, root and flower buds rapidly increases (^{Clapham and Marsh, 1987}), resulting in different fruiting intensities during the crop cycle. The time between the harvesting maximum varies among cultivars (^{Wubs et al., 2009a}).

In all treatments flowers of the first node remained on the plant and developed mature fruits. In the second node reached 100 % of harvest with the lower flower removal and 99 % was reached in the other two treatments. No other node reached this percentage of high fruiting, regardless of how much the initial sink strength was reduced. Except for the first two nodes, the abortion of reproductive structures was always greater than 40 % in the plants of the two treatments with the lowest flower removal, and greater than 60 % in those with the most intense flower removal (Figure 3). Similar abortion-fruiting patterns were observed in the treatments with flower removal up to the first and second node. After the maximum fruit production in the first two nodes, a total fruiting decrease was observed. This happened in the fifth and sixth node on plants of the first treatment (without flowers on the first node) and the seventh node in the plants of the second treatment (without flowers to the second node). Fruiting percentage increased up to 60 % from the seventh and eighth nodes in the first and second treatment. The highest fruit development was at the 12° and 13° nodes with the removal of flowers on the first node and in the 9° and 10° nodes with the removal of flowers up to the 2^nd node.

Figure 3: Percentage ratio between fruits set and abortions in the first fifteen nodes of Capsicum annuum plants grown in three conditions of flower removal: A) on the first node, B) up to the second node, C) up to third node. 

The most important difference in the behavior of plants with the highest flower removal was the absence of nodes with a 100 % aborted flowers; although a similar pattern was observed in treatments one and two. An area with preponderant flower loss was clearly delimited between the fourth and sixth node along with increased fruit production between the seventh and 11° node; maximum fruiting (40 %) occurred at the 9° and 10° nodes.

Fruit development in lower nodes is related to competition between reproductive structures of the upper nodes and with increased flower abortion. Other studies have shown that with increasing sink strength the source/sink relationship decreases and the overall rate of abortion linearly increases (^{Hall, 1977}; ^{Clapham and Marsh, 1987}). All together, the increase in the overall abortion rate of the plant is related to the high growth rates of the fruits (^{Marcelis et al., 2004}).

The three flower removal treatments did not cause a significantly different effect on yield (p>0.05) (Table 1). This result does not match with the paradigm established the beginning of the studies about since this phenomenon, which states that dry matter accumulation in the root and fruiting efficiency increase with the removal of flowers (^{Hall, 1977}; ^{Clapham and Marsh, 1987}). In other study higher yield was obtained by removing the flowers on the second node (^{Jaimez et al., 2005}).

Table 1: Average production (kg per plant, kg m^‒2 and fruits per plant) of fruits of Capsicum annuum in three treatments of flower removal intensity, cultivated for 320 d in greenhouse. 

♦ Between parenthesis, the standard error. There were no statistical differences between treatments (Kruskal-Wallis; p>0.05).

The fruits of treatments with maximum flower removal reached the highest average weight (158 g±53); in treatments one and two, average fruit weight was lower (140 g±67 and 136 g±62), respectively (p≤0.05). There was no significant difference in the yield of plants with different flower removal treatments (p>0.05).

The quality of the fruits varies in relation to the removal of flowers (p≤0.05) (Table 2). With less flower removal, the fruits distribution between weight classes was more homogeneous. In all treatments fruit production with weights greater than 300 g was low. The difference in the amount of harvested fruits with less than 100 g weight was evident among treatments, and they were the lowest quality. The plants of both treatments with less flower removal had a higher percentage of fruit with these characteristics (30 % and 28 %). In plants with higher flower elimination, fruits of that class accounted for only 11 % of total production.

Table 2: Percentage distribution by classes regards fresh weight (g) of Capsicum annuum fruits. Plants subjected to three flower removal intensities and cultured for 320 days in a greenhouse. 

♦ Chi-square test: pχ42<11.143=0.975.

The fruits of all treatments showed the same growth rate in the previous lapse to the maximum instantaneous fresh weight accumulation rate period. Afterwards, (about 20 d), the growth rate was higher, whereas flower removal increased (Figure 4).

Figure 4: Instantaneous growth rate (g d^‒1) in dry weight of Capsicum annuum fruits until harvest. Plants subjected to three flowers removal intensities (A) on the first node, (B) up to the second node, (C) up to the third node. 

The maximum growth rate of the fruit plants with larger flowers removal took place one or two days after the fruits of the other two treatments (Table 3). This behavior disappears after day 60, when the fruit growth rate is minimal and unchanged in all treatments. This allows linking the fruit high growth rate from the highest flower removal treatments with the overall floral abortion rate in those plants. ^{Wubs et al. (2009b)} reported that cultivars with fruit average greater than 140 g reach the highest growth rates by the twentieth day. However, in the literature reviewed there were no researches linking flowers elimination intensity with fruits classification by weight or with the instantaneous growth rate.

Table 3: Maximum instantaneous growth rates (mm d^‒1 or g d^‒1) and time to reach (d) developing fruits of Capsicum annuum plants submitted to three flower removal intensities. 

We showed that by using the fruits classification according to its fresh weight, after 260 d of cultivation, the fruits quality significantly decreased, regardless of the flower removal treatment applied. This decrease in fruits quality is consistent with results reported by ^{Khah and Passam (1992)}, who found that during the cultivation of C. annuum in greenhouse conditions, the fruit size and number of seeds are gradually reduced. Thus, we propose that in tropical mountain conditions a cultivation period of 260 d in order to optimize the harvests quality and take advantage of the remaining time for new planting. Then, within a year the advantage of a greenhouse structure could be used to develop a full crop and a third from another.

Fruits rapidly increased their diameter and longitude during the first developmental phase After fruits reached 15 mm in diameter, it took 7 to 9 d to increasing greater diameter and 5 to 6 d for the length; whereas the maximum gain in weight occurred between days 18 and 19 (Table 3). The backwardness of the mass gain with respect to the dimensional growth was not associated with treatments of removal flowers.

The initial reduction of the sink strength (by removing flowers) and the concomitant increase in vegetative growth prior to fruiting seems to increase the rate of biomass accumulation during fruit development. However, there is no evidence that this effect is maintained in the plant for indefinite time, because the observations were made up to the 15° node.

Relationship between dimensions and weight of the fruit during its development

The relationship between the fresh fruit weight and the product of the diameter × length is described by the equation: Y=6.473e^‒7X²+0.016X‒4.426 (R²=0.95). The relationship between the dry weight of the fruit (biomass) and the product of the diameter × the length is described by the equation: Y=2.112e^‒8X²+1.673e^‒3X‒1.140 (R²=0.91). Through these equations it was possible to estimate the fresh and dry weight of the fruits still attached to plants. This result concur with those of ^{Bozokalfa and Kilic (2010)}, who showed that the volume of the fruits of C. annuum depends on 95 % in the relationship between diameter, length and fresh weight. The main difference between their results and the ones here presented is that they used the volumetric estimation of the fruit, which is used in the calculation of yield (commercial); and in the present research we used the fresh and dry weight estimation in order to apply a growth model and calculate the daily rate of weight gain.

Influence of the sink strength on photosynthetic traits of leaf at different heights of the plant

The fruit diameter when harvested was almost invariable in all fruit, regardless of the node where it was formed (Figure 5); but the average fresh weight at harvest did present wide variations. There was a tendency to decrease the weight of the fruit with rise of its position along the plant. The fruits developed on the first node were significantly heavier than those from the second and third (p≤0.05).

Figure 5: Final average weight (g) and final average diameter (mm) of fruits of C. annuum in the first three nodes. (*) Indicates significant difference between nodes (Kruskal-Wallis; p≤0.05). Bars indicate standard error. 

With the increase of the fruit diameter the gs of the accompanying leave decreased and A showed a similar behavior (Figure 6) and three sets of leaves, with significant differences defined. The apical leaves, with larger A (14.55 𝛍mol m^‒2 s^‒1); leaves on two and three nodes, with intermediate rates (9.81-10.82 𝛍mol m^‒2 s^‒1) and leaves the first node with A 5.93 𝛍mol m^‒2 s^‒1.

Figure 6: (A) Mena fruit diameter (mm) on C. annum on the first three nodes (white columns) and rate of CO₂ assimilation (𝛍mol m^‒2 s^‒1) (± standard error). (B) Mean stomatal conductance (mol m^‒2 s^‒1) on accompanying and apical leaves (± standard error). Mean irradiance 900 𝛍mol m^‒2 s^‒1 and mean temperature 30 °C. (*) Indicates significant differences among nodes (Kruskal-Wallis; p≤0.05). 

Leaves on nodes with fruits which already was mature or harvest fruits have low A due to decreased organic foliar N and low demand of the fruit; whereas leaves that are near of developing fruits have the highest A (^{González-Real et al., 2009}). Our results agree with this approach, because A decreased with increasing fruit diameter adjacent, in their last phase of development. However, the A in the apical leaves was higher than in leaves that accompanied the developing fruits; which contradicts previous reports (^{González-Real et al., 2009}). The latter was attributed to discrepancies in the characteristics of the apical leaf used in both investigations.

The simultaneous A measurement and fruit growth showed that the assimilation capacity of the leaves declined as the diameter increased (Figure 7). This shows that due to the demand from developing fruits, mobilization of nitrogen compounds take place from the surrounding leaves, which reduces photosynthesis (^{González-Real et al., 2009}). In this regard, the management of turnover and the number of leaves per node can positively impact the sink/source ratio of carbohydrates, and thereby productivity.

Figure 7: Rate of CO₂ assimilation (𝛍mol m^‒2 s^‒1) as a function of fruit diameter (mm) nearest to leave, for two cultivars of Capsicum annuum. Y=^‒0.35X+37.00 (R²=0.86). 

The photochemical apparatus of photosystem II efficiency decreases with the aging of the leaf. That is, as leaves became senescent the photochemical yield and photochemical extinction coefficient (q _p ) decreased and the no-phootochemical extinction coefficient (q _N ) increased (Figure 8). This means that with time, the photoprotection mechanisms increase as the proportion of photon directed to CO₂ fixation is reduced. This trend towards decreased photochemical yield (Φ _PSII ) of the plants was very markedly shown on the lower leaves, which accompanied the harvested fruit. Besides, in the first experiment no relationship was found between the progressive decrease in the weight of the fruit and the flowers removal treatments applied. This suggests that the decrease is due to the aging of the plants and the concomitant reduction in photosynthetic capacity.

Figura 8: Fluorescence parameters of chlorophylla in apical leaves of the third and second node of Capsicum annuum. Values obtained in three stages of cultivation (± standard error). Average radiation 1000 𝛍mol m^‒2 s^‒1 and constant temperature (28 °C). 

This decrease in the photosynthetic capacity during aging is correlated with a decrease in chlorophyll content, leaf nitrogen, leaf proteins and Rubisco enzyme activity (^{Field, 1987}). Under this proposal, ^{Imai et al. (2008)} showed that the Rubisco synthesis with leaf senescence. ^{Olesinski et al. (1989)} stated that the decrease in photosynthetic rates decrease, as leaves get older, responds to the redistribution of resources during which it favors the youngest leaves. This would make senescent leaves act as resources for the other that eventually accompany developing fruits demanding carbohydrates.

The turnover of the leaves on the stem depends on the availability of N in the soil and in the plant and the allocation-distribution carbon rate in the plant. This process of foliar turnover maximizes the efficient use of resources and can determine the success and dominance in the competition and demand between neighboring plant organs (^{Kouki, 2005}). In our study, N supply was continuous during the growing season, so, this would not be a limiting factor; however, ^{González-Real et al. (2009)} showed that even with continuous fertilization, organic foliar N content decrease depending on the leaf senescence.

Consequently, recycling of resources from basal leaves (senescent), distribution of N or change in the rate of synthesis of enzymes responsible for CO₂ fixation might be some of the factors determining the decrease of the photosynthetic capacity of leaves over time.