Doses and application methods of Azospirillum brasilense in irrigated upland rice

Doses e modos de aplicação de Azospirillum brasilense em arroz de terras altas irrigado R E S U M O Este trabalho, desenvolvido no município de Selvíria, MS, nos anos agrícolas de 2011/12 e 2012/13 teve, como objetivo, avaliar a eficiência de Azospirillum brasilense na fixação de nitrogênio em arroz de terras altas, em função de doses e modos de aplicação do inoculante contendo a bactéria diazotrófica. Utilizou-se delineamento experimental de blocos ao acaso disposto em esquema fatorial 4 x 4 sendo 4 doses de inoculante (testemunha sem inoculação, 100, 200 e 300 mL do produto comercial ha-1) x 4 modos de aplicação (inoculação das sementes, aplicação no sulco de semeadura, pulverização do solo logo após a semeadura e pulverização foliar no início do perfilhamento das plantas) com 4 repetições. Durante a condução do experimento foram avaliados as características agronômicas, os componentes de produção e a produtividade da cultura do arroz. Conclui-se que o inoculante contendo Azospirillum brasilense promove acréscimo (19%) na produtividade de arroz de terras altas irrigado por aspersão, quando usado na dose de 200 mL ha-1, independentemente do modo de aplicação.


Introduction
Despite the reductions in rice production and yield in the 2015/16 season, compared with the previous one, respectively 14.3 and 0.6% (CONAB, 2016), the production chain of this cereal is currently one of the most important for the Brazilian agribusiness, since it is a product with high internal consumption and that represents an expressive volume of the grain production in the country (Zamberlan & Sonaglio, 2011).
The rice plant is characterized for having high demand for nutrients and nitrogen (N) being the main limiting factor to its yield.It is a component of the chlorophyll that increases plant leaf area and, in turn, increases the efficiency in solar radiation interception and photosynthetic rate, thus having high influence on grain yield (Fageria & Stone, 2003).Additionally, it influences the number of panicles per area unit and of spikelets per panicle, fertility of spikelets, grain weight and panicle length (Fageria & Barbosa Filho, 2001).
Analyzing this agricultural scene and given the importance of this nutrient, studies that evaluate N-fixing bacteria, regarding their alternative supply of this nutrient, are an important sustainable method in agriculture (Elbeltagy et al., 2001;Guimarães et al., 2007) and can promote yield increments, which is very important because, for upland rice in the Central-West region of Brazil (3,203 kg ha -1 ), for example, is relatively lower than that of the irrigated system (6,798 kg ha -1 ) obtained in the south of the country (CONAB, 2016).
It is important to point out that the process of biological fixation by these bacteria excretes only part of the N directly fixed to the associated plant, but it is able to meet just part of the N requirements of the plants (Hungria et al., 2010).
The bacteria of the Azospirillum genus, also considered as plant growth-promoting bacteria (PGPB), besides promoting biological N fixation (Perrig et al., 2007), synthesize hormones such as auxin, which stimulates shoot and root growth of various grass crops, including rice (Radwan et al., 2004); however, studies involving this crop with the use of inoculants and the various forms of inoculation are still scarce.
Considering the above, this study was conceived with the objective of evaluating the influence of doses and forms of application of inoculant containing Azospirillum brasilense in upland rice under sprinkler irrigation.

Material and Methods
The study was carried out during the agricultural years of 2011/12 and 2012/13 in an experimental area of the Faculty of Engineering -UNESP, Campus of Ilha Solteira, located in the municipality of Selvíria, MS, Brazil, at approximate coordinates of 51º 22' W and 20º S, with altitude of 340 m.
The local soil is a clayey typic dystrophic Red Latosol (Santos et al., 2013).The mean annual rainfall is 1,370 mm, mean annual temperature is 23.5 ºC and mean annual relative air humidity is between 70 and 80%.
The climatic conditions during the experiment are shown in Figure 1.
The experimental design was randomized blocks, arranged in a 4 x 4 factorial scheme.The treatments consisted of the combinations of doses of inoculant containing Azospirillum brasilense (control without inoculation, 100, 200 and 300 mL ha -1 of the commercial product) and forms of application (seed inoculation, application in the sowing furrow, soil spraying after sowing and foliar spraying at the beginning of plant tillering), with 4 replicates.Plots comprised four 4.5-m-long rows spaced by 0.35 m.The evaluated area consisted of the three central rows, disregarding 0.50 m on each side.
Before installing the experiment, in 2014, soil samples were collected in the layer of 0.00-0.20 m for chemical analysis and the results are presented in Table 1.
The mineral fertilization used in the sowing furrows was calculated according to the soil chemical characteristics taking into consideration the recommendations of Cantarella & Furlani (1996).250 kg ha -1 of the formulation 04-30-10 were used and the top-dressing fertilization was performed at 30 DAE, using ammonium sulfate as the source.The utilized dose was 42 kg ha -1 of N, defined based on the expected yield range and the class of soil response to N, by applying 70% of the recommendation.At first, it was expected that the other 30% were supplied by the diazotrophic bacteria; immediately after application, irrigation was performed with a water depth of approximately 10 mm.
Water was supplied through a fixed sprinkler irrigation system with mean flow rate of 3.3 mm h -1 in the sprinklers.For water management, three crop coefficients (Kc) were used, distributed in four periods, comprehended between emergence and harvest.For the vegetative stage, the value of 0.4 was used; for the reproductive stage, two crop coefficients were used, the initial of 0.70 and the final of 1.00; and, for the maturation stage, these values were inverted, i.e., the initial of 1.00 and the final of 0.70 (Rodrigues et al., 2004).
A fungicide containing trifloxystrobin + tebuconazole (75 + 150 g a.i.ha -1 ) was applied for the control of upland rice diseases, such as blast and helminthosporium, in the entire experimental area at 65 and 78 DAE; the other cultivation practices were performed according to the recommendations for the crop.
At flowering, the N content was evaluated in the leaves, determined through the collection of the leaf blades of 20 flag leaves per plot and following the methodology of Malavolta et al. (1997); plant height (m), during the stage of pasty grains, at 3 random points per plot, measured as the distance between soil surface and the upper tip of the highest panicle; number of panicles per square meter, determined by the count of the number of panicles in 1.0 m of plant row in the evaluated area and then calculated per square meter.
In the crop maturation stage, 100-grain weight was also determined, by randomly collecting and weighing two 100-grain samples of each plot (13% wet basis), as well as hectoliter mass, evaluated by weighing a 0.25-L sample of grains of each plot and then converting to kg 100 L -1 at 13% of moisture.Lastly, grain yield was determined through the weight of hulled grains, from the evaluated area of the plots, correcting moisture to 13% and converting it to kg ha -1 .
The results were subjected to the F test of the analysis of variance, comparing the means of the treatments with Azospirillum brasilense by Tukey test at 0.05 probability level for the forms of application and polynomial regression for the inoculant doses, fitting models of significant linear and quadratic equations by F test (p < 0.01 and p < 0.05).The statistical analyses were performed using the program SISVAR (Ferreira, 2011).

Results and Discussion
In the 2011/12 agricultural year, the accumulated rainfall during the crop cycle was 652 mm (Figure 1A), while in the second year (2012/13) it was 513 mm (Figure 1B), thus in agreement with the crop requirement, which is between 450 and 700 mm during the entire cycle.This behavior depends on the cultivar and on the soil in the cultivation area (Rodrigues et al., 2004).
In the agricultural years of 2011/12 and 2012/13, plant emergence occurred at 6 and 7 DAE and the full rice flowering occurred at 86 and 75 DAE, respectively.The maturation allowed the harvest on February 27, 2012 (107 DAE) and on February 26, 2013 (104 DAE) in the first and second years, respectively.
The obtained results for leaf N content, plant height and number of panicles m -2 are presented in Table 2.
Leaf N contents and the number of panicles m -2 were not influenced by the treatments of doses and forms of application of inoculant, in both seasons.N values in the flag leaf (leaf N) remained always in the range reported by Raij et al. (2001) as adequate (27-35 g kg -1 ).Araújo et al. (2010) and Goes (2012), evaluating Azospirillum brasilense inoculation in rice seeds, also did not find difference in the number of panicles per m 2 , corroborating the results found in the present study.
For plant height (Table 2), there was no significant effect in the first cultivation year; however, for the agricultural year of 2012/13, there was interaction between the doses and forms of application of the inoculant containing Azospirillum brasilense.
According to the follow-up analysis (Table 3) for inoculant doses in each form of application, for the applications in the soil, seed and plant, PH fitted to quadratic functions with maximum estimated points of 183, 135 and 158 mL ha -1 , respectively.For the application in the sowing furrow, the data fitted to a positive linear equation, i.e., the higher the inoculant dose, the higher the PH of upland rice.
As to the follow-up analysis for the forms of application at the inoculant doses, it is observed that, at the dose of 100 mL ha -1 , the application in the sowing furrow showed lower PH, whereas at the dose of 300 mL ha -1 of inoculant the application in the soil showed higher value in comparison to the application in the seed.
According to Bottini et al. (1989), the production of phytohormones by bacteria, as in the case of inoculation with Azospirillum strains, is one of the factors responsible for the stimulating effect observed on plant growth.As found in the present study, rice growth had hormonal effect when plants were inoculated, in relation to the control.
For hectoliter mass and according to the observed means (Table 4), there was no significant difference between the treatments.
For grain yield, there was no difference between treatments in the first year; similarly, Banheza et al. (2012) also did not observe influence of the doses of 0, 100, 200 and 400 mL ha -1 of commercial inoculant applied in the seed on upland rice yield.Gitti et al. (2012), evaluating the effect of seed inoculation with A. brasilense with 200 g of inoculant for 25 kg of seeds, did not find influence on the yield of upland rice under sprinkler irrigation.
The data of the second agricultural year also did not indicate significant differences for the forms of application of the inoculant; however, for the doses, the data fitted to a quadratic equation.The maximum point occurred at the inoculant dose of 192 mL ha -1 , which corresponded to the estimated grain yield of 6,622 kg ha -1 ; therefore, an increment of 9.47% in relation to the control.Thus, it is noted that the inoculation promoted

Figure 1 .
Figure 1.Values of rainfall and temperature in the experimental area in the period from sowing to harvest of rice in the agricultural years of 2011/12 (A) and 2012/13 (B)

Table 4 .
Mean values of 100-grain weight (100GW), hectoliter mass (HM) and grain yield (GY) of upland rice as a function of doses and forms of application of inoculant containing Azospirillum brasilense during 2011/12 and 2012/13 season

Table 1
of the cultivar IAC 202.The seeds were previously treated with fipronil (50 g of a.i. 100 kg of . Results of soil chemical analysis in the layer of 0.00-0.20 m, before experiment installation

Table 3 .
Follow-up analysis of the interaction of inoculant doses x forms of application of Azospirillum brasilense for plant height (PH) of upland rice under sprinkler irrigation during 2012/13 season