Plant growth regulation: a method for fine-tuning mepiquat chloride rates in cotton 1

and

According to Ferreira (2014), the total PGR amounts to be applied at high, medium and low plant heights are 85-100 g a.i.ha -1 , 60-85 g a.i.ha -1 and ≤ 50 g a.i.ha -1 , split in four applications (10 %, 20 %, 30 % and 40 % of the total rate, respectively).However, cotton growth and its response to PGR application is highly dependent on temperature and water availability, which are not considered in this recommendation.
When cotton is grown in tropical regions with a long growing season (approximately 160-210 days), there is a high probability of rank growth, and an easily applicable method for fine-tuning mepiquat chloride rates in cotton would help in the decision-making to manage the mepiquat chloride use.Despite the existence of crop growth simulations with good anticipation of mepiquat chloride rates and application time (Gu et al. 2014), an extensive area cultivated with cotton around the world is located in countries where these resources are not readily available to farmers, or where farmers are not prepared to use these resources.Hence, it is important that the mepiquat chloride recommendation method be simple and easy to apply.
Our hypothesis is that cotton growth is highly dependent on its genetics, as well as the interaction between cultivar and environment, and this interaction may increase/decrease the cotton mepiquat chloride requirements as a function of the amount of dry matter accumulation.Thus, the aims of this study were: i) to propose a new method for defining mepiquat chloride rates in cotton fields; ii) compare this method with traditional management practices; iii) evaluate the effect of mepiquat chloride application using the proposed method in cotton cultivars with different growth cycles.

MATERIAL AND METHODS
A preliminary experiment (Experiment 1) was conducted to obtain a dry matter estimate to facilitate field measurements.Cotton (Gossypium hirsutum cv.Delta Opal) plants were grown in 12-L pots, in a greenhouse at the Universidade Estadual Paulista, in Botucatu, São Paulo state, Brazil.Eight pots containing two cotton plants each were grown for 57 days, with two replications, and harvest at 39, 45, 51 and 57 days after emergence (DAE), to determine the shoot dry weight, plant height and leaf area, and a curve was fitted (Figure 1).
Since it is impossible to determine the plant dry matter in the field without harvesting the plant, the shoot dry weight was estimated by a regression model based on plant height.This equation was used to estimate the growth regulator concentration in the plants.These plants did not receive PGR application.The regression model of plant shoot dry weight with plant height is shown in Figure 1.A close and significant (p < 0.05) correlation was found, with a coefficient of determination (R 2 ) of 0.98, suggesting that the plant height measurement at field conditions may be a useful tool to estimate the plant dry weight.
Another experiment (Experiment 2) was conducted to determine the plant growth rate and PGR retention on plant leaves.Mepiquat chloride rates (0 g ha -1 , 7.5 g ha -1 , 15 g ha -1 and 22.5 g ha -1 ) were applied at the pinhead square stage, at 39 DAE, and a solution of Blue Brilliant (FDandC n.1), a color additive, was added to the spray solution at 3,000 mg L -1 .Leaves from treated plants were collected, washed and the water used was analyzed by spectrophotometry.The amount of product retained in the leaves was calculated.
Plant growth rate was inversely proportional to its leaf PGR concentration (Figure 2).Therefore, in order to decrease growth rate, the amount of PGR applied must be increased and, for cotton cultivars with low growth rate, the PGR dose would be lower.
For calculating the PGR rate, we assumed that a high-yield plant should have 25 nodes: 5 of them being vegetative nodes, 5 nodes from the top that are not considered productive because they will bear bolls produced after the cut-out and will not be harvested, and 15 being effective reproductive nodes.Considering the ideal plant height at harvesting of 125 cm, the average internodes length would be 5 cm.On average, one new node appears every 3 days (considering a basal temperature of 15 ºC and an average daily temperature of 25 ºC); hence, the daily growth rate would be 1.6 cm day -1 .
Three scenarios were designed, based on the daily crop growth rate: i) regular: considering an early cycle cultivar of regular vegetative growth with a daily growth rate of 1.5 cm; ii) favorable: considering a mid-late cycle cultivar of vigorous growth with a daily growth rate of 1.25 cm; iii) very favorable: for those vigorous cultivars plus an environmental condition prone to rank growth, such as good soil humidity and high temperatures (above 32 ºC), considering a daily growth rate of 1.0 cm.This scenario can be predicted with an accurate weather forecast one week ahead of the mepiquat chloride application.
The PGR rate to be applied as a function of daily growth rate (1.5 cm day -1 , 1.25 cm day -1 and 1.0 cm day -1 ) took into account the plant height, shoot dry weight (estimated by plant height) and PGR concentration needed per plant and per gram of shoot dry weight.This concentration was obtained from the daily growth rate of plants receiving different PGR rates and their shoot dry weight, leaf area and amount of PGR retained in the canopy.With the plant leaf area and the calculated sprayed volume per unit of leaf area, the amount retained per plant, in mL per plant, for each applied rate (7.5 g ha -1 , 15.0 g ha -1 and 22.5 g ha -1 ), was calculated.For this, we considered the shoot dry weight of plants that received the Blue Brilliant and PGR.
An average of PGR needed to achieve each growth rate was obtained, and an equation was fitted to estimate the concentration of PGR needed (μg g -1 ) to obtain the stipulated daily growth rates (1.5 cm day -1 , 1.25 cm day -1 and 1.0 cm day -1 ).The amount of PGR per plant (μg plant -1 ) was the result of the multiplication of the PGR concentration needed per shoot dry weight unity.Finally, this value was multiplied by 90,000 plants (usual plant population for cotton crop), and then the estimated PGR rate was obtained.Curves were fitted and, according to their equations, it was possible to estimate the amount of PGR replacing the "x" by the measured plant height in the field (Table 1 and Figure 3).Growing conditions: regular: early cycle cultivar with a daily growth rate of 1.5 cm; favorable: mid-late cycle cultivar of vigorous growth with a daily growth rate of 1.25 cm; very favorable: vigorous cultivars plus an environmental condition prone to rank growth, such as good soil humidity and high temperatures (above 32 ºC), considering a daily growth rate of 1.0 cm.The estimated PGR rates, as a function of plant height, are shown in Figure 3, based on the following equations: regular: y = 12.64 -0.538x + 0.0083x 2 ; favorable: y = 24.39-1.007x + 0.0154x 2 ; and very favorable: y = 34.16-1.438x + 0.0224x 2 .All scenarios designed have led to low PGR requirements up to 60 cm of plant height.At this point, the PGR requirement was different in each growth condition, because the growth vigor was higher in the favorable and very favorable environments, as compared with the regular environment.
An experiment was carried out in the field at Primavera do Leste, Mato Grosso state, Brazil (15º33'S, 54º17'W), to assess the efficiency of the Crop Growth Rate (CGR) application method in plant height control, as well as in cotton yield and fiber quality.A combination of eight cultivars (late cultivars FM993, IMA 1318, IMACD-8276, FMT709; mid-late cultivars IMACD6001, IMACD07-6035, IMA108-3869; and early cultivar FM 966) and two PGR managements were arranged in a randomized complete block design, replicated four times.
Cotton was planted in rows (0.90 m apart) on December 15 (2012).The final plant density was 10.7 ± 2.86 plants m -2 .The rates of mepiquat chloride used in the traditional method were based on the manufacturer recommendations, as it follows: 20 g a.i.ha -1 , 20 g a.i.ha -1 , 38 g a.i.ha -1 and 38 g a.i.ha -1 at the first, second, third and fourth applications, respectively (more details are given in Table 3), and the use of the CGR method as described previously in Table 1.
The plant height and average of the top five internodes length were monitored twice a week.When the average of the top five internodes length was greater than 3.5 cm, PGR was applied, as mepiquat chloride, based on equations, as previously mentioned (Table 1 and Figure 3).The rates applied for each cultivar in each treatment are shown in Tables 4 (2012/2013) and 5 (2013/2014).
In both experiments, pearl millet was grown for 85 days and herbicide-dissected (glyphosate 1.92 kg a.i.ha -1 ) before cotton planting.Cotton was fertilized with phosphorus (120 kg ha -1 of P 2 O 5 ) broadcast-applied one week before planting.At harvesting, the plant height, number of bolls, boll weight, fiber yield and fiber quality (HVI) were determined.Anova and regression analyses were performed (SigmaPlot ® ).The average values for each treatment were compared by the Tukey test (p < 0.05).

RESULTS AND DISCUSSION
Plant growth regulator rates were higher under the CGR method, when compared with the traditional growers practice (Table 3), except for the early cycle cultivar FM966.The amount of PGR applied to this cultivar amounted to 43 % of the growers treatment.On the other hand, the PGR rate applied to other cultivars increased, on average, 24 %, ranging 18-36 %, if compared with the growers practice.Most of the cultivars received lower PGR rates under the CGR treatment in the first and third applications.
This difference is the result of the low plant dry matter accumulation at the pinhead square stage (B1) (~40 DAE).Since the CGR method takes into account the plant height and dry matter accumulated, the amount of PGR applied was 70 % lower for the early cycle cultivar and 35-40 % for the other cultivars, as compared with the traditional method.At this stage, shoot development is slower, because the carbohydrates produced by the leaves are used mainly for root growth, and the requirement for PGR is low.Therefore, we argue that the amount of PGR applied using CGR was more consistent with the plants development stage than using the traditional method.In the second application (56 DAE), the PGR rate was lower under the CGR method, as compared to the traditional method, for IMA1318, IMACD8276 and FM966, and higher for the other cultivars, due to the growth scenario at the time of the application.In the third application, the PGR rate was lower in all cultivars and, at the fourth application, was lower for FM966 and higher in the others, when PGR was applied using the CGR method.
There was no interaction of cotton cultivars and PGR application method, except for micronaire (Table 4 and Figure 4).Additionally, the methods of PGR management did not affect cotton growth parameters, yield components and neither fiber length, strength and maturity (Table 3).Plant height at harvesting was affected only by cultivar, and it was the shortest for the early cycle cultivars IMACD-6001 and FM 966 (Table 4).No differences in plant height were observed among the other cultivars.The late maturity cultivar FM 993 had 10 % more nodes than IMACD 6001 (early) and IMACD 8276 (late).
The number of bolls was greater in FMT709.However, the boll weight was greater in IMACD-8276 and IMA108-3869, possibly owing to the lowest number of bolls in such cultivars (Table 4).Gin turnout was significantly lower in the FM966 and Late cultivars: FM 993, IMA 1318, IMACD-8276, FMT 709; mid-late cultivars: IMACD 6001, IMACD07-6035, IMA108-3869; early cultivar: FM 966.* The 1st, 2nd, 3rd and 4th applications were carried out at 40, 56, 74 and 86 days after emergence (DAE) for the growers pattern management, and at 43, 56, 79 and 86 DAE for the crop growth rate method (CGR).Growth scenario: R -regular, considering an early cycle cultivar of regular vegetative growth with a daily growth rate of 1.5 cm; F -favorable, considering a mid-late cycle cultivar of vigorous growth with a daily growth rate of 1.25 cm; V -very favorable, for those vigorous cultivars plus an environmental condition favorable for rank growth, such as good soil humidity and high temperatures (above 32 ºC), considering a daily growth rate of 1.0 cm.

Management Cultivar
Rate of mepiquat chloride per application (g a.i.ha In general, all cultivars showed a good fiber length, strength and maturity (Table 4).Micronaire was lower for IMACD-8276, FM966 and IMA03-1318 under traditional PGR management, if compared with the same cultivars under the CGR method (Figure 4).The new method was efficient in delivering the right PGR rate, specifically for an early cycle cultivar (IMA5675B2RF), which showed a high sensitivity to PGR.In such a cultivar, PGR would not be necessary, as the plant height at harvesting was adequate for mechanical harvesting (above 1.35 m for row spacing of 0.90 m), even in PGR non-treated plants (Figure 7).
In addition, Gwathmey & Craig ( 2003) also reported that the excessive control of vegetative growth on early cycle cultivars with limited potential of growth could result in reduced lint yields.A previous study has shown a high sensitivity of early cultivars to mepiquat chloride (Bogiani & Rosolem 2009).
The proposed method detected these differences in the total amount of applied PGR.For example, in the 2012/2013 experiment, PGR rates ranged 65-156 g a.i.ha -1 and, in the 2013/2014 experiment, 31.6-164.6g a.i.ha -1 , showing that, depending on the cotton cultivar, the requirement may be significantly increased or decreased, reducing costs in early cultivars and preventing PGR losses and yield reductions.
For example, in the 2012/2013 experiment, FM993, FMT709, IMACD3869, IMACD6001 and IMACD 6035 received PGR based on the very   2009), if compared to late cultivars.As the input in the formula is plant height, the amount of PGR applied early in the crop season will always be lower and will increase as plants age.
In addition, higher PGR rates were applied for the CGR method, if compared with the traditional recommendation, most likely due to the amount applied in the fourth application (86 DAE) (Table 3), when PGR rates were almost double the traditional management approach.The possibility of PGR losses due to rainfall occurring after application does exist (Echer & Rosolem 2012) and it could be a disadvantage of this method, once the PGR rate increases as plants grow higher.Johnson & Pettigrew (2006) also observed no interaction between PGR management and cotton cultivars on plant height, as reported in our experiment (2012/2013).On the other hand, the interaction was significant for lint yield, contrary to what we found in this experiment.O 'Berry et al. (2009) suggested that cotton cultivars may respond differently to PGR application, during certain years, even if actual plant height reductions are not significant.Moreover, no difference in plant height between CGR and traditional management practices was observed in the 2012/2013 experiment (Table 5), but some cultivars exhibited micronaire variation (Figure 4).
Plant height was different among cotton cultivars without PGR application at 65 DAE.IMA08-12427 and IMACV22 had the tallest, and IMA5675B2RF the shortest plants, at 110 DAE (Figure 5).Differences in plant height were also observed at 65 DAE in PGR-treated plants, and IMACV22 and IMA076035 had the highest plants, while IMA5675B2RF had the shortest plants at 110 DAE (Figure 5).
There was a significant interaction of cultivars and PGR application for plant height (Figure 6), number of bolls (Figure 7), boll weight (Figure 8) and fiber length (Figure 9).However, no interaction was observed for number of nodes, fiber yield, gin turnout, fiber strength, micronaire and maturity (Table 6).
The final plant height was lower for PGRtreated plants in all cultivars.Additionally, PGR nontreated plants of IMA5675B2RF and IMACV690 had an appropriate plant height for mechanical harvest (~1.30m), showing that PGR would not be necessary for these cultivars (Figure 6).
The number of nodes was dependent on the cultivars, and the early maturing cultivar Late cultivar: IMACD-8276; mid-late cultivars: IMA08-12427, IMA108-3869, IMA5822B2RF, IMA5672B2RF, IMACD07-6035, IMACV22, IMACV690; early cultivar: IMA5675B2RF.* The 1st, 2nd, 3rd, 4th and 5th applications were carried out at 40, 48, 61, 68 and 89 days after emergence.Growth scenario: R -regular, considering an early cycle cultivar of regular vegetative growth with a daily growth rate of 1.5 cm; F -favorable, considering a mid-late cycle cultivar of vigorous growth with a daily growth rate of 1.25 cm; V -very favorable: for those vigorous cultivars plus an environmental condition favorable for rank growth, such as good soil humidity and high temperatures (above 32 ºC), considering a daily growth rate of 1.0 cm.  6).Lint yield was decreased for IMA5672B2RF, if compared with IMACD07-6035 and IMACV690, but no differences among the others cultivars were observed.Micronaire and maturity ratio were not different among cotton cultivars (Table 6).

Cultivar
A 2 % (40.25-39.46)decrease in gin turnout was observed in PGR-treated plants.Otherwise, fiber strength was increased by the CGR treatment (Table 6).The number of bolls for PGR-treated plants was reduced, if compared with non-treated plants of IMA5672B2RF and IMA5675B2RF (Figure 7), what may explain the higher boll weight observed in these cultivars (Figure 8).Finally, the interaction of cultivars and PGR, for fiber length, was significant, and the PGR treatment decreased the length of IMACV22LL and IMACV690 (Figure 9).In two years of evaluation, the PGR management did not affect lint yield (Tables 5 and 6).Studies with two or more years of field experiment or different locations have shown that mepiquat chloride responses in yield are dependent upon environmental conditions (Zhao et al. 2017), such as season length, because the mepiquat chloride treatment increases fruit retention at lower fruiting branches.In years with a short growing season, mepiquat chloride tends to increase yield (Kerby et al. 1986).Otherwise, in a full growing season, the bolls produced at the top may mature, and differences in yield disappear (Zhao & Oosterhuis 2000).
Mepiquat chloride had an inconsistent effect on the lint percentage, with a neutral effect for the 2012/2013 and a negative effect for the 2013/2014 experiment.PGR application has been reported to reduce the lint percentage in a wide range of environments (Cathey & Meredith 1988, Biles & Cothren 2001, Johnson & Pettigrew 2006, Ren et al. 2013), and this decrease has been attributed to an increase in seed weight (Biles & Cothren   Mepiquat chloride-treated cotton plants have also shown inconsistent changes in micronaire (Cathey & Meredith 1988, Siebert & Stewart 2006).In the present study, it was observed, in the 2012/2013 experiment, that the PGR management based on the new method could prevent a decrease in micronaire.Fiber length was unaffected in 2012/2013, as observed by Heilman (1985), and decreased by the mepiquat chloride treatment, in 2013/2014, contrary to data from O 'Berry et al. (2009), who reported an increase in fiber length, as a response to mepiquat chloride rates.Fiber strength was not affected by the PGR management in 2012/2013, corroborating the findings of Heilman (1985), but a positive effect was obtained by the mepiquat chloride treatment in 2013/2014, consistently with previous studies (Cathey & Meredith 1988, Ren et al. 2013), following the application of PGR to cotton.CONCLUSIONS 1.The proposed method is efficient in defining the plant growth regulator rates for cotton cultivars, and results in economy of mepiquat chloride in early cycle cultivars, preserving lint yield and fiber quality; 2. The easy applicability of this method distinguishes it from other methods, because all the user needs to know is the actual plant height, cultivar maturity group and weather forecast for the next week.

Figure 1 .
Figure 1.Correlation between the cotton shoot dry matter production and plant height.

Figure 2 .
Figure 2. Daily growth rate of cotton plants, as a function of plant growth regulator (PGR) concentration on leaf tissue.

Figure 3 .
Figure 3. Cotton plant growth regulator (PGR) estimated rate, as a function of plant height in regular, favorable and very favorable growth scenarios.

Figure 7 .
Figure 7. Number of bolls in cotton cultivars, as affected by plant growth regulator (PGR) treatments, in the 2013/2014 growing season.Capital letters compare PGR treatments in each cultivar (Tukey, p < 0.05).

Figure 8 .
Figure 8. Boll weight of cotton cultivars, as affected by plant growth regulator (PGR) treatments, in the 2013/2014 growing season.Capital letters compare PGR treatments in each cultivar (Tukey, p < 0.05).

Figure 9 .
Figure 9. Fiber length of cotton cultivars treated and non-treated with plant growth regulator (PGR), in the 2013/2014 growing season.Capital letters compare PGR treatments in each cultivar (Tukey, p < 0.05).

Table 1
. Mepiquat chloride rates, as a function of plant height and growing condition.The plant growth regulator rate resulted from the replacement of "x" in the equation by plant height in the first column.Plant growth regulation: a method for fine-tuning mepiquat chloride rates in cotton e-ISSN 1983-4063 -www.agro.ufg.br/pat-Pesq.Agropec.Trop., Goiânia, v. 47, n. 3, p. 286-295, Jul./Sep.2017