Cover crops change the phytosociology of weeds and the banana yield

This work aimed to evaluate the phytosociology of weeds and the fruit yield of banana (cv. Tropical) in different soil coverings. The experimental design used was randomized blocks, with eight treatments and three replications. The treatments consisted of sowing of Canavalia ensiformis, Cajanus cajan, Crotalaria juncea, Stylosanthes 'Campo Grande', Crotalaria spectabilis, Sorghum bicolor, and Phaseolus lunatus between the rows of the banana trees. A control treatment was also used without cover crops. Weed phytosociology was evaluated in four treatments. The number of hands per bunch, the number of fruits per hand, the number of fruits per bunch, and the fruit yield were estimated in all treatments. In the phytosociological survey, 23 species were identified, in 13 families, especially Asteraceae and Poaceae families. Compared to the control, the coverage with C. cajan, C. juncea, and C. ensiformis provided a reduction in weed density of 108%, 71.7%, and 14.6%, respectively. C. ensiformis and C. cajan provided the highest yield of banana fruits, with an increase of 51.2% about the control. It is concluded that C. ensiformis and C. juncea provided the most significant suppression of weeds, while C. cajan and C. ensiformis provided the highest fruit yield.


Introduction
Inadequate weed management can cause significant losses in the production components of different species of economic interest, such as corn and crambe (Coelho et al., 2016;Lima et al., 2018), forest species (Eloy et al., 2014;Faria et al., 2018) and vegetables (Guerra et al., 2016;Vendruscolo et al., 2017), resulting in significant economic losses (Vicensi et al., 2011;Vendruscolo et al., 2017). Thus, for better control of interspecific competition between weeds and crops, herbicides are usually applied, which affects not only the weeds but the ecosystem (Myers et al., 2016;Van Bruggen et al., 2018).
The search for cultivation management that is less aggressive to the environment has culminated in the increase of studies on the use of biological methods for the control of insect pests, diseases, and weeds. For the weed control, the potential of some species to suppress the emergence of unwanted plants is observed, whether through shading, allelopathy, or even the formation of physical barriers (Moura Filho et al., 2015;Coelho et al., 2016).
Simultaneously with the suppression of weeds, the maintenance of cover plants allows the cycling of nutrients through the accumulation in phytomass, mineralization, and availability of these for crops of economic interest (Almeida and Câmara, 2011). In this sense, it appears that this technique, although studied and applied to annual crops, can also be used for fruit production, grown in a consortium between the planting rows (Paulino et al., 2010). Among the species used as cover crops, legumes stand out for having a higher capacity to release nutrients in the short term when compared to grass species due to the lower carbon: nitrogen ratio (Acosta et al., 2014).
Information obtained in scientific studies indicates benefits of using cover crops between the rows of established orchards, either by suppressing the development of weeds, maintaining edaphic quality or by making nutrients available (Partelli et al., 2010;Carvalho et al., 2013;Feitosa et al., 2009). For the banana crop (Musa sp.), positive results were also observed for the intercropped management of cover crops in the Brazilian semiarid region, improving the vegetative and reproductive development of plants (Quaresma et al., 2015), demonstrating their potential for employment in crops in other regions of the country. Therefore, it is essential to know about the species present in each cultivation system since interspecific interactions are observed (Coelho et al., 2016).
The work aimed to evaluate the effect of different cover crops on weed phytosociology and the fruit yield and production components of banana from BRS Tropical cultivar.

Material and Methods
The experiment was carried out in Chapadão do Sul -MS, at coordinates 18º 46'17.8" S, 52º37'27.7" W, and altitude of 813 m. The soil in the area was classified as Latossolo Vermelho Distrófico of clayey texture, or Oxisol (Soil Survey Staff, 2014). The climate of the region is humid tropical (Aw), with a rainy season in summer and a dry season in winter, and an average annual rainfall of 1,850 mm (Cunha et al., 2013).
The banana plantation was implanted in 2014 and is composed of plants from BRS Tropical cultivar, with 2.5 m between rows and plants. The banana suckers were thinned at intervals of 45 to 60 days, avoiding their competition with the mother plant for water and nutrients. The elimination of male flower bud, which consists of the terminal portion of the rachis, when it reaches between 20 and 25 cm away from the last hand, was carried out to improve the health of the fruits through insect control. Fungicide, insecticide, and herbicide applications were carried out aiming at the control of diseases, insects, and excess of weeds outside the useful area of the experiment, to obtain good phytosanitary conditions in the banana plantation.
The experiment consisted of 200 plants, with 96 useful plants and the others serving as a border. Each experimental plot was formed by eight plants, and the four central plants were used as a useful plot. The treatments were distributed in three blocks, composed of eight treatments, totaling 24 plots. The treatments were formed by following cover crops Canavalia ensiformis, Cajanus cajan, Crotalaria juncea, Stylosanthes 'Campo Grande', Crotalaria spectabilis; Sorghum bicolor, Phaseolus lunatus L., and control treatment, without a cover crop.
The plots with cover crops were composed of 2.0 m wide and 3.0 m long. The area between the banana tree rows was desiccated using glyphosate (1.0 kg a.i. ha -1 ) + 2.4D 1.0 (kg a.i. ha -1 ) before the cover crops sowing. The sowing of the cover crops was performed manually in sowing furrows spaced 0.30 m apart.
Phytosociological assessments were performed with the square (0.25 m) randomly launched in each plot. The weeds present within that square were classified and separated between species. The data obtained were used to calculate the phytosociological indices defined by Mueller-Dombois and Ellenberg (1974): a) Density (D): total number of individuals in a population/total number of species; b) Relative Density (Dr): (species density x 100)/total species density; c) Frequency (F): number of squares containing the species/total number of squares obtained (total area); d) Relative frequency (Fr): (species frequency x 100)/total species frequency; Revista de Agricultura Neotropical, Cassilândia-MS, v. 7, n. 1, p. 53-59, jan./mar. 2020. e) Abundance (A): total number of individuals per species / total number of squares that contain the species; f) Relative abundance (Ar): (species abundance x 100)/total abundance of all species; g) Importance Value Index (IVI): relative frequency + relative density + relative abundance (Curtis and Mcintosh, 1950;Mueller-Dombois and Ellemberg, 1974).
In the banana crop, after harvesting the bunches, the number of hands per cluster, the number of fruits per hand, the number of fruits per bunch, and the fruit yield (Mg ha -1 ) were evaluated. Phytosociology data were organized in tables. The data of fruit production components and fruit yield were submitted to analysis of variance. The Tukey test compared the means at 5% probability.

Results and Discussion
Only treatments with Canavalia ensiformis, Cajanus cajan, Crotalaria juncea, and the control was considered in the evaluation of phytosociology. The other cover crops had their emergence and initial establishment hindered by low germination resulting from a prolonged period without rain.
In the phytosociological survey, 23 weed species were identified in 13 families (Table 1). Asteraceae and Poaceae stood out, which participated with 6 and 4 species, 26.09% and 17.39%, respectively. This result corroborates the data verified in the literature, which points out that, among the infesting families, Asteraceae and Poaceae stand out, being found in different areas of production, from grains to sugarcane (Oliveira and Freitas, 2008).
For the treatments evaluated between the rows of the banana trees, the treatment in which the highest density of weeds was obtained was the control treatment ( Figure  1), which did not receive cover crops. This culminated in the absence of suppression, besides that carried out by the banana plants themselves. However, given that the banana plantation was fully established, the competition was not overly effective. It is observed that, for the banana crop, the first five months are limiting to development, requiring constant cleaning to avoid competition for resources (Carvalho and Vargas, 2008). Compared to the control treatment, weed density decreased when C. cajan, C. juncea, and C. ensiformis were implanted. In the plots containing these treatments, a reduction of 51.94%, 41.75%, and 12.77% was observed in the weed density, respectively. These values indicate the potential of the cover crops to restrict the manifestation of the seed bank at the time of cultivation, being extremely important for the producer to take the necessary measures and make the appropriate choices. You should also consider other beneficial effects of the cover crops, such as nitrogen fixation.
The C. ensiformis is a recommended cover crop due to its rapid initial growth, and that it presents efficient control of invasive plants through suppression by shading, caused by the abundant presence of broad leaves, as observed by Coelho et al. (2016). Also, C. cajan, which has vigorous growth and tall plants, was the cover that most effectively suppressed the emergence of weeds, corroborating other studies in which its use as vegetation cover was adequate for the same purpose (Lima et al., 2014;Dantas et al., 2015).
The species Ageratum conyzoides stood out in the four treatments, presenting high IVI (Tables 2, 3, 4, and 5). The Commelina benghalensis species stood out in the treatment without cover crops, while the species Spermacoce latifolia, Richardia brasiliensis, and Alternanthera tenella stood out in the treatment with C. juncea. This shows that the suppression capacity is dependent on the weed species, the species used as cover, and the interaction between species.  A possible technique to aid suppression would be to use cover crop species together, since weeds behave differently, depending on the species used as a cover crop. The higher production of dry mass by area, observed with the use of cover crops cocktails, culminates in a significant stifling of weeds, in addition to increasing the yield of plants of economic interest such as corn (Coelho et al., 2016). About the variables related to the banana crop, it was found that for the number of hands per bunch and the number of fruits per bunch, the treatments composed by the species C. ensiformis, C. juncea, C. spectablis, C. cajan, and P. lunatus were significantly superior to the control treatment. For the number of fruits per hand, only the treatment with C. ensiformis was significantly superior to the control treatment. The yield was benefited using all cover species, except for Sorghum bicolor and P. lunatus (Table 6). Specifically, for yield, it was observed that the treatments composed by C. ensiformis and C. cajan stood out from the others, not differing significantly from the treatment composed by C. juncea (Table 6). These treatments provided increments of 50.60%, 51.79%, and 35.71%, respectively. Parts of these benefits result from the capacity of these plants to perform biological nitrogen fixation and nutrient release after decomposition (Almeida and Câmara, 2011), combined with the good soil cover which they provide, which results in direct interference in the emergence of weeds (Moura Filho et al., 2015;Coelho et al., 2016).
Most of the infesting species have fast germination, short cycle, high production of diaspores, and a high degree of interference, becoming extremely aggressive in competition with cultivated plants. This dynamic capacity of weeds makes the management of the productive system more complex, requiring management actions to reduce the immediate number of weeds capable of competing with the crop. This whole process makes the system more costly (Vendruscolo et al., 2017), in addition to harming the environment more intensely, which can even lead to contamination of the marketed products and the degradation of the soil and water present in the environment (Myers et al., 2016;Van Bruggen et al., 2018).
In this way, given the results obtained, the possibility of using cover crops as a technique to minimize weed infestation is verified, allowing the lesser introduction of herbicides into the production system. Additionally, the maintenance of the cover crops will provide better maintenance of the edaphic conditions, improving the development of the plants.

Conclusions
The presence of cover crops provides suppression of weeds in an established banana plantation.
C. cajan and C. ensiformis, used as a cover crop between the banana rows, increase the components of production and fruit yield.