Stalk productivity and quality of three sugarcane varieties at the beginning, in the middle, and at the end of the harvest

1 Escola Superior de Agricultura Luiz de Queiroz, Universidade de São Paulo, Av. Pádua Dias, 11, 13418-900, Piracicaba, São Paulo, Brazil. 2 Centro de Ciências Agrárias, Universidade Federal de Alagoas, BR 104, Km 85, s/n, 57100-000. Rio Largo, Alagoas, Brazil. 3 Departamento de Pesquisa, Usina Triunfo, Vila Triunfo, s/n, 57680-000, Boca da Mata, Alagoas, Brazil. 4 Empresa Brasileira de Pesquisa Agropecuária, Centro de Pesquisa Agropecuária dos Tabuleiros Costeiros, Unidade de Execução de Pesquisa de Rio Largo. BR 104, Km 85, s/n, Rio 57100-000. Rio Largo, Alagoas, Brazil. 5 Departamento de Agronomia, Universidade Federal Rural de Pernambuco, Av. Dom Manoel de Medeiros, s/n, 52171900, Recife, Pernambuco, Brazil.


INTRODUCTION
The sugar-alcohol sector in Brazil is one of the most technology intensive in the world. This is the result of research, with technological improvement programs developed in the country over decades making new varieties available with high productive potential. There have been great advances in knowledge of soils and plant nutrition, in cultivation practices, production management, and cane payment for juice quality (Oliveira et al., 2014a;Simões et al., 2015;Rhein et al., 2016). The adoption of improved varieties has also contributed to an increase in cane productivity and crop profitability (Souza et al., 2012;Silva, 2013). Among these new varieties, RB92579, RB98710, and RB961552 stand out. RB92579 is characterized by its high productivity, hydric efficiency, optimal sucrose content, and short flowering period. RB98710 has a high productivity and sucrose content, low fiber content, early maturity, and is recommended for restricted environments, with it rarely flourishing or toppling. RB961552 is highly productive and responds excellently to irrigation (RIDESA, 2010).
Soil and climate conditions have a great impact on sugarcane juice production and quality. For this reason, it is important to carry out studies to determine the productivity and industrial quality of juice, in various production environments, in order to establish cultivation practices that make it possible to exploit sugarcane productive potential to the maximum (Calheiros et al., 2012;Silva et al., 2014).
In evaluations of the productive potential of particular sugarcane varieties, industrialized stalk productivity per hectare (TSH) and juice quality should be evaluated, especially apparent sucrose (AS) and inorganic phosphorus (Pi) levels (Tasso Júnior et al., 2014;Rhein et al., 2016). Sucrose is the raw material in sugar and alcohol production and the Pi concentration in juice is important both for yeast metabolism in alcoholic fermentation and in sugar production (Oliveira et al., 2011;Tasso Júnior et al., 2014;Mohammed et al., 2016). Pi is important in the sugarcane juice clarification process during sugar production. On reacting with the slaked lime (Ca[OH] 2 ), tricalcium phosphate is formed, which through flocculation and sedimentation draws the impurities to the bottom of the decanter (Calheiros et al., 2012;Oliveira et al., 2014b;Mohammed et al., 2016). Pi levels greater than 50 mg L -1 have been prescribed for good alcoholic fermentation to occur and Pi levels greater than 100 mg L -1 for efficient juice clarification (Martins, 2004;Tasso Júnior et al., 2014).
Juice quality varies according to stalk position and sugarcane maturity, with it therefore being important to evaluate juice quality in different parts of the stalk and different harvesting periods in order to identify the best period for each variety. In light of this, the aim of this paper was to evaluate the productive potential of the recently launched RB961552 and RB98710 sugarcane genotype varieties, compared to RB92579, in the caneplant and first regrowth cycles, in cultivations in the Zona da Mata (scrub region) in Alagoas State, as well as juice quality in three harvesting periods in different parts of the de Oliveira et al. 261 stalk.

MATERIALS AND METHODS
The study w as conducted at the Jequiá Plantation, located in the municipality of Anadia, AL, at 09°41'04''S latitude and 36°18'15"W longitude, and belonging to the Triunfo Mill, over the period from August 2011 to January 2014. The evaluations w ere carried out in tw o cycles: cane-plant and first regrow th. The region's climate is rainy tropical w ith dry summers, according to the Koppen classification. The average annual precipitation is 1200 mm, w ith an average annual temperature of 29°C. In 2012, accumulated rainfall in the months of November and December w as only 1.4 mm. In 2013, the volume of rain in the months from January to March w as low er than in 2012 ( Figure 1). The study w as set up in Yellow Dystrophic Latosol (Embrapa, 2013) w ith an average texture. Prior to setting up the study, a chemical analysis of the soil w as carried out at depths of 0 to 20 and 20 to 40 cm. With the results, dolomitic limestone and plaster w ere applied in a proportion of 3:1, and in sufficient quantity to raise base saturation to 60% in the topsoil layer and reduce aluminum saturation on the subsurface, as proposed by Oliveira et al. (2007) and Raij (2011). After 60 days, the soil w as plow ed, harrow ed, and subsequently furrow ed.
Planting w as carried out in August 2011. 500 Mg ha -1 of 09-14-22 chemical fertilizer w as applied at the bottom of the planting furrow s. Table 1 show s the results of the soil analysis after soil fertilization. Three sugarcane varieties w ere planted: RB92579, RB961552, and RB98710, w ith treatments placed in random blocks w ith five repetitions. RB92579 w as chosen as a reference due to it currently being the most w idely-planted cane variety in Alagoas (Ridesa, 2012). RB961552 and RB98710 are promising varieties, but there is little information available in the literature.
The study w as carried out in an random block experimental design w ith five repititions. The plots consisted of 7 furrow s, measuring 8 m in length, w ith 1.0 m spacing and a total area of 56 m 2 . The useful area w as 30 m 2 , composed of 5 central row s, excluding one meter borders. Plant density fluctuated betw een 15 and 18 seedlings per meter of furrow , w hich w ere collected from a nursery at the Triunfo Mill. The seedlings w ere manually covered w ith soil, w ith an approximately 5 cm layer of earth placed over them and Tebutiuron herbicide applied straight afterw ards in doses of 1.0 kg of active ingredient per hectare. Fipronil w as also used to control leafcutter ants. Leafhopper (Mahanarva species) and sugarcane borer (Diatraea species) controls w ere carried out via biological control, w ith the use of Metarhizium anisopliae and Cotesia falvipes, respectively (Benedini, 2006). Fertilization of the cane from first regrow th w as carried out after cane-plant harvesting. 500 kg ha -1 of 20-05-25 fertilizer w as applied by manually spreading the fertilizer. The w eed and pest controls adopted in first cane regrow th w ere the same as those for the cane-plant stage.
The quality of the low er, middle, and upper thirds of the industrialized stalks w ere evaluated at 14, 15, and 17 months of sugarcane age, and at 9, 10, and 12 months in first regrow th, corresponding to the beginning, middle, and end of the harvest, respectively. In the Northeast of Brazil, the beginning of the harvest corresponds to the months of September and October, the middle of the harvest to November and December, and the end of the harvest to January and February (Souza et al., 2012).
In each of the evaluations, five stalks from the second row from  3.0 2.2 left to right w ere collected, separated, clipped, divided into low er (LT), middle (MT), and upper (UT) thirds, and then w eighed. The samples for each third w ere passed through a forage cutter and homogenized. A subsample of 500 ± 1.0 g of chopped stalks w as pressed at 250 kgf cm -2 for 60 s to separate the juice from the pulp (CONSECANA, 2006). The juice obtained w as analyzed for apparent sucrose levels in the juice (AS), juice purity (PUR), total recoverable sugars in the juice (TRS), and inorganic phosphorus (Pi). From the w et pulp, the stalk fiber level (fiber) w as determined. The Pi calculation w as carried out at the Center for A gricultural Sciences Agricultural Chemistry Laboratory (ASAC) of Alagoas Federal University (UFAL), in accordance w ith the methods described by Delgado et al. (1984). The other analyses w ere carried out at the Triunfo Mill Juice Quality Analyses Laboratory , follow ing the methods described by Fernandes (2000) and Lavanholi (2008). In the last evaluation, at 17 months after planting and at 12 months after cane-plant collection, industrialized stalk and sugar productivity w ere determined for each variety. The evaluation w as carried out in plot row s 3, 4, and 5, from left to right. The plants w ere cut close to the soil, separated, clipped, and w eighed, to determine industrialized stalk productivity in tons of sugarcane per hectare (TSH). In a subsample of these stalks, the juice w as extracted and the TRS w as determined (Fernandes, 2000). Gross income per hectare of sugar (GIS) w as obtained by multiplying stalk productivity per hectare by TRS.
The data w ere submitted for variance analysis using the F test and the averages compared using the Scott-Knott test w ith a 5% probability. These analyses w ere carried out w ith the help of the Sisvar softw are (Ferreira, 2011).

RESULTS AND DISCUSSION
In the cane-plant cycle, the AS and TRS levels were significantly influenced by the varieties, by the harvesting period, and by the stalk third (Tabel 2). There was a significant difference in fiber between the stalk thirds and the harvesting periods, while PUR was only affected by the stalk third (Table 2). Pi was influenced by the variety and by the stalk third. In the first regrowth cycle, all of the variables were influenced by the varieties, by the harvesting period, and by the stalk third (Table 2). There was interaction between the harvesting period and the stalk third for all the juice quality variables analyzed in the first regrowth cycle (Table 2) and for AS, PUR, and TRS in the cane-plant cycle. The interaction between the varieties and harvesting period was significant for Pi in the cane-plant cycle and for AS, PUR, and TRS in the first regrowth cycle.
RB961552 presented AS and TRS levels that were 10 and 9% lower than the averages for the RB92579 and RB98710 varieties, in the cane-plant cycle, and 10 and 8% lower in the first regrowth cycle, respectively ( Table  3). The R92579 and RB98710 varieties presented statistically similar averages in the two cycles, for AS and TRS, except for AS in the first regrowth cycle, when RB98710 presented a 2.4% higher average. TRS and AS behaviors are similar, since the TRS variable depends on the sucrose level present in the cane juice (Oliveira et al., 2014b).
In a study conducted by Calheiros et al. (2012), regarding RB867515 and RB92579 cultivated in Rio Largo, in the Zona da Mata in Alagoas, in the cane-plant cycle, AS values were obtained for RB92579 that are similar to those observed in this study. Moreover, Oliveira et al. (2011), in a study conducted in Boca da Mata, AL, observed AS levels fluctuating around 18.0% for RB98710, RB867515, and RB92579, collected at the end of the Alagoan harvest.
AS and TRS levels were always lower at the beginning of the harvest (Table 3). The increase in sucrose in the stalk is associated with the reduction in available soil water and the action of invertase enzymes and sucrose phosphate, as cited by Vieira (1988) and Casagrande and Vasconcelo (2008). In these studies, it was verified that acid invertase activity is high in internodes in elongation, but absent in mature internodes. In the mature stalk internodes, there is an increase in alkaline invertase and sucrose phosphate synthesis. In the caneplant cycle, there was no statistical difference for AS and TRS levels in the middle and at the end of the harvest. However, in the second year of cultivation, the sugarcane collected at the end of the harvest presented a higher AS level. This occurs due to the reduciton in available soil water and consequent hydric restriction in the plant, which influences sugarcane maturation (Toppa et al., 2010). In Brazil, the low temperatures associated with hydric deficit are the main climatic factors responsible for de Oliveira et al. 263 sugarcane maturation. In Alagoas, it is the main climatic factor related to sugarcane maturation (Calheiros et al., 2012).
In the MT and LT, AS and TRS levels did not differ between any crop cycle (Table 3), with an average of 18.93% and 149.70 kg Mg -1 in the cane-plant cycle and 19.47% and 156.75 kg Mg -1 in first regrowth, respectively. The difference between the UT and MT and LT averages was 25.94 and 17.15%, respectively. This difference was lower than that observed by Martins (2004), who while working with the SP823530, SP835073, and RB835486 varieties observed an average difference of 71.43%. This greater difference in sucrose levels between stalks is due to a shorter sugarcane maturation phase, as excess rainfall in the study delayed cane dehydration.
In the interaction between the harvest period and stalk thirds (Tables 4 and 5), it was observed that the UT always presented lower AS and TRS levels. For caneplant, there was no statistical difference for AS and TRS levels in the MT and LT in any harvest period. On the other hand, for first regrowth, the AS and TRS levels increased in the UT, MT, and LT at the beginning of the harvest. As the harvest advanced, the difference in sucrose concentrations between the thirds decreased, showing advanced maturity (Leite et al., 2010;Toppa et al., 2010).
From the analysis of the interaction between variety and harvest period, in the first regrowth cycle (Table 6), it is observed that AS and TRS levels increase progressively as the harvest advances. The RB92579 and RB98710 varieties did not differ between each other in the three harvest periods, and were always greater than RB961552. The percentage difference between the AS level in RB961552 and the average for the other two varieties was 10.07, 15.29, and 14.23%, at the beginning, in the middle, and at the end of the harvest, respectively. On the other hand, RB961552 presented a lower TRS at the beginning and in the middle of the harvest, however was similar to RB92579 and RB98710 at the end of the harvest.
In the cane-plant cycle, there was no statistical difference between the varieties studied, with the average PUR value being 87.15% (Tables 2 and 3). In the first regrowth cycle, RB92579 was the variety that presented the greatest PUR, with 88.75%, this being 1.6% greater than the average for the other varieties. The harvest period only influenced PUR in the first regrowth cycle, when the juice collected at the end of the harvest presented greater purity. In both cycles, the UT presented lower purity than the MT and LT. This lower purity in the UT is the result of the sugarcane maturation process, which occurs from the base to the apex (Segato et al., 2006;Leite et al., 2010;Toppa et al., 2010).
On analyzing the interaction between the stalk thirds and harvest period, it was observed that, in the two cultivation cycles (Tables 4 and 5), the UT had less juice purity, however, at the end of the harvest, there was no Table 2. Values and significance of average squares from variance analyses and variation coefficients of soluble solid percentages (SS), apparent sucrose in juice (AS), purity (PUR), total recoverable sugars (TRS), and inorganic phosphorus (Pi), for three sugarcane varieties, in three harvesting periods (P), in three parts of the stalk (third), in the cane-plant and first regrow th cycles. ns ,*,** represent, respectively, not significant and significant to 5.0% and 1.0% probability using the F test.   Averages followed by the same letter in the column do not differ statistically between each other using the Scott-Knott test w ith a 5% probability. statistical difference between the thirds. Uniformity of purity in the stalk is expected when sugarcane reaches maximum maturity, presenting a maturity index between 0.85 and 1.0 (Toppa et al., 2010). The maturation index is the proportion of apparent sucrose content, determined using polarimetry, from the base to the industrially useable stalk. It is an index used to evaluate sugarcane maturation (Liz et al., 2016). Analyzing the interaction between the varieties and the harvest period (Table 6), in the first regrowth cycle, it is found that in the middle of the harvest, RB92579 presented greater purity, however it did not differ from the other varieties at the beginning and end of the harvest.

Source of variation G.L Average squares AS (%) PUR (%) Fiber (%) TRS (Mg ha
In all of the harvesting periods, all of the varieties presented over 80% purity, which is considered adequate for sugarcane industrialization (Rhein et al., 2016;Rodolfo Junior et al., 2016). In this paper, average purity in the cane-plant and first regrowth cycles was greater than that reported by Oliveira et al. (2011) andSilva (2013). High PUR in sugarcane juice is desired at the time of harvesting, since it implies a higher concentration of sucrose and reduced amino acids, organic acids, starch, reducing sugars, and other color precursor compounds (Rhein et al., 2016;Rodolfo Junior et al., 2016).
The average fiber level in the three varieties for caneplant was 14.34% (Table 3), approximately 11.08% higher than that observed by Oliveira et al. (2011) andSilva (2013). In the first regrowth cycle, RB98710 presented a higher level of fiber, with 15.10%, while the RB92579 and RB961552 varieties did not differ between each other and present an average of 14.80%. In both cycles, the level of fiber increased as the harvest advanced and the UT presented a higher level of fiber than the MT and LT, which did not differ between each other. The difference between the UT and the average for the MT and LT was 5.51 and 16.21% in the cane-plant cycle and first regrowth, respectively. The higher level of fiber in the UT is probably due to the lower accumulation of sucrose compared to the other thirds.
Fiber is important when it comes to industries' energy balance, as pulp is used for obtaining electrical energy; however, a high level of fiber causes resistance to juice extraction (Simões et al., 2015;Rodolfo Junior et al., 2016). To maintain energy balance, a percentage of fiber between 10 and 12.5% has been recommended. However, the Northeast region of Brazil presents greater evapotranspiration than the Center-South region, for which reason sugarcane cultivated in the Northeast has a higher level of fiber at the time of harvesting (Oliveira et al., 2011(Oliveira et al., , 2014b).
RB92579 presented a lower level of Pi in the two cultivation cycles, although it did not differ from RB961552 in the cane-plant cycle (Table 3). On the other hand, RB98710 was the variety that presented the highest level of Pi in the juice. In the cane-plant cycle, the level of Pi in RB92579 juice was lower at the beginning of the harvest, but the concentration rose during the harvest, reaching the same values as RB98710 at the end of the harvest (Table 7). The Pi concentration in the sugarcane cycle was lower than in the first regrowth cycle, probably due to the greater production of biomass in the cane-plant cycle, resulting in dilution of the absorbed phosphorus (Oliveira et al., 2007). The average Table 7. Average values for inorganic phosphorus (Pi) in mg L -1 in the RB92579, RB961552, and RB98710 varieties, collected at the beginning, in the middle, and at the end of harvest, in the cane plant cycle.

Variety
Collection Pi levels in RB92579, RB961552, and RB98710 juice for the two cycles were 73, 77, and 82 mg L -1 of P, respectively. High Pi levels in sugarcane juice during industrialization are desirable for reducing the cost of clarifying the juice, since the addition of exogenous Pi is necessary when juice levels are not adequate for good clarification (below 100 mg L -1 ) (Mohammed et al., 2016).
Pi levels in juice of around 180 mg L -1 were obtained by Oliveira et al. (2011) in studies conducted in the Alagoan Agreste region involving the RB867515 variety. Tasso Júnior et al. (2014) evaluated Pi levels in the CTC9, CT15, and CTC16 cane varieties and did not find any difference between the varieties with regards to Pi in the juice, finding an average juice value of 147 mg L -1 . In the study conducted by Martins (2004), the Pi level was influenced by the variety, with Pi values of 151, 236, and 388 mg L -1 for the SP823530, SP835073, and RB835486 varieties, respectively.
The Pi levels only differed between the harvesting periods in the first regrowth cycle, when the sample corresponding to the middle of the harvest presented a higher Pi level than in the other samples (Table 3). Pi level behavior in the thirds differs between the cycles studied. In the cane-plant cycle, the highest Pi content was observed in LT, however in first regrowth, the UT presented the highest Pi level. By analyzing the interaction between the thirds and the harvesting periods in the first regrowth cycle (Table 5), it is observed that at the beginning of the harvest, the UT presented the lowest Pi level, in the middle of the harvest there was no difference between the thirds, and at the end of the harvest the LT has the highest Pi level. When sugarcane is not yet completely mature, the UT is the most biochemically active part, demanding greater quantities of Pi (Oliveira et al., 2014b;Tasso Júnior et al., 2014). With sugarcane maturity, Pi comes to be required in greater quantities in the LT and MT, where it acts as an energy source in the sucrose accumulation process in the cell vacuoles (Casagrande and Vasconcelos, 2008). Thus, when sugarcane starts the maturation process, Pi migrates from the UT to the MT and LT.  Table 9. RB92579 was the variety that presented the highest TSH, at around 15% more than the other varieties, and consequently the highest GIS. RB98710, despite having similar TRS to RB92579, produced fewer stalks and therefore its GIS was lower. RB961552 was less productive for all the varieties analyzed. The RB92579 variety presented a higher GIS than those observed by Aquino et al. (2016). Ferreira Junior et al. (2014) indicated that RB98710 has a high sugar level and high productivity. When cultivated using drip irrigation, they observed that RB98710 sugar productivity was 17.8 Mg ha -1 (Ferreira Junior et al., 2014), similar to the RB92579 productivity observed in this study. The stalk productivity obtained in this study (118.52 Mg ha -1 ) is considered as average to high for the state. In Alagoas, the maximum sugarcane growth phase occurs on short days, and therefore under low luminosity, unlike in the Center-South of Brazil, where increased luminosity coincides with greater hydric availability. The non-coincidence of maximum hydric availability with luminosity negatively influences photosynthetic rates, resulting in lower cane productivity in Alagoas compared to the Center-South (Oliveira et al., 2011;Calheiros et al., 2012).
Studies carried out in Brazil (Calheiros et al., 2012;Oliveira et al., 2014b) indicate RB92579 as one of the most productive varieties, and this is one of the reasons for which, together with RB867515, it is in expansion. However, the juice from this variety presents high phenolic and flavonoid levels (Oliveira et al., 2011), characteristics that are not contemplated in sugarcane payment for recoverable sugar (TRS), but which negatively contribute to juice color and makes industrialization difficult. Phenolic compounds are substances that negatively influence juice color and consequently that of the sugar, reducing the quality and acceptability of the product (Qudsieh et al., 2002). They  Averages followed by the same letter in the column do not differ statistically betw een each other using the Scott-Knott test w ith a 5% probability.
also have a negative effect on fermentation, especially by reducing the action of invertase excreted by the yeast. The productive superiority of RB92579 was not proven in the first regrowth cycle, as there was no significant difference for any of the variables in this cycle. The averages for TSH, TRS, and GIS in the first regrowth cycle were 66.98 Mg ha  (Table 8). The TRS was similar to that found by Silva (2013), however the TSH and GIS were lower. The decrease in productivity in the first regrowth cycle was high and probably influenced due by the hydric stress in the growth phase in the second cycle of the study.
Hydric deficit in the growth phase is one of the main causes of reduced sugarcane productivity (Rhein et al., 2016;Rodolfo Junior et al., 2016), since it causes morphophysiological defense alterations such as reductions in leaf area and gas exchange. Bueno et al. (2012) studied 10 sugarcane genotypes in the first regrowth cycle collected in different periods in the state of Paraná, where it was observed that hydric deficit in the cane growth phase reduced agricultural production and the accumulation of sugar collected in April, the beginning of the harvest in the region.

Conclusions
RB961552 has lower levels of apparent sucrose, total recoverable sugars, and purity, than RB92579 and RB98710. However, all the varieties have ideal apparent sucrose and purity levels for the samples in the three harvest periods, with these values increasing as the harvest advanced.
RB98710 has higher Pi levels in the juice than RB92579 and RB921552. In the cane-plant cycle, all the varieties have lower than ideal Pi levels for juice clarification, while in the first regrowth cycle, RB98710 and RB961552 have Pi levels within the ideal range.
In the cane-plant cycle, the RB92579 variety has higher stalk and sugar productivity than the other varieties.