INTRODUCTION

Carrot (DaucuscarotaL.)is an important crop, grown and consumed in most countries worldwide. China and the United States are the top producers. In Brazil, the carrot is the fifth most cultivated crop, and its average yield ranges from 30 to35tha^-1(^{Paula, 2019}).

Carrot yield and quality rely on many factors related to soil fertility and irrigation. Inadequate growth conditions result in physiological anomalies and reduced root mass, which decrease market value or even cause product losses of up to 30% (^{Nick and Borém, 2016}). Inoculation with plant growth-promoting rhizobacteria (PGPR) is a promising alternative that improves environmental conditions for carrot crops, as previously reported for other agricultural species. The benefits of PGPR in other crops include higher germination rates, increased root and shoot biomass, improved nutrition and resistance to biotic and abiotic stresses (^{Kloepper et al., 1980}; ^{Vessey, 2003}; ^{Aloo et al., 2019}; ^{Fioreze et al., 2020}; ^{Lin et al., 2020}).

Studies on inoculation benefits in carrots are scarce and focus on a small number of measured parameters. Evaluations by ^{Kozusny-Andreani et al. (2014)}, ^{Matsuoka et al. (2016)} and ^{Clemente et al. (2016)} reported the benefits of Pseudomonas and Bacillus as increased root length, fresh mass and dry root mass.

Different PGPR species can be combined in a process known as co-inoculation. The benefits of co-inoculation are well reported for several crops, such as soybean, common bean and corn (^{Molla et al., 2001}; ^{Fukami et al., 2018}; ^{Kumawat et al., 2019}; ^{Zeffa et al., 2020}). The inoculation of soybean with Bradyrhizobium japonicum resulted in a yield of 2,200 kg ha^-1, while co-inoculation with this bacterium and Azospirillum brasilense increased this value by 12% (2,496 kg ha^-1). Common bean yield also benefited from co-inoculation, with production of 1,866kg/ha when Rhizobium tropici was used with Azospirillum brasilense, as compared to inoculation with only Rhizobium tropici, which resulted in 1,543 kg ha^-1 (^{Hungria et al., 2013}).

Only one study by ^{Merriman et al. (1974)} has provided information on the benefits of co-inoculation in carrots. These authors described the benefits of Bacillus subtilis and Streptomyces griseus, which included 17% more root mass than without inoculation. However, all studies on carrots either lack bacterial species designation or do not include evaluations during the entire crop cycle, which limits the understanding of inoculation and co-inoculation benefits. Moreover, the use of PGPR with reduced N levels has not been evaluated in any published study but this practice warrants research considering global dependence on nitrogen fertilizers. The benefits of rhizobacteria in crops such as corn and soybean are often expressed or optimized with reduced nitrogen fertilization, which decreases the costs of crop production (^{Mendes et al., 2003}; ^{Lana et al., 2012}). Therefore, this study tested the hypothesis that inoculation with two PGPR (Bacillus subtilis and Pseudomonas fluorescens) with reduced nitrogen levels improves growth and yield in carrots.

MATERIALS AND METHODS

This experiment was established in Ponte Alta do Norte, Santa Catarina State, Brazil (27º13’29’’ S and50º25’28’ W) at 990 m a.s.l. The climate was classified as Cfb (Köppen’s classification) with average temperatures between 12.4 and22.0ºC. Prior to establishing the experiment, soybean and corn were cultivated in the area for several years.

Soil samples (0-20 cm) were collected to analyze chemical properties and soil granulometry. The results were: pH in H₂O 6.60; Ca 9.65 cmol_c dm^-3; Mg 5.35 cmol_c dm^-3; H+Al 2.00 cmol_c dm^-3; K 0.15 cmol_c dm^-3; P 6.00 mg dm^-3, Na 1.00 mg dm^-3 and K 1.57 mg dm^-3. The CTC at pH 7.0 was 13.99 cmol_c dm^-3, the effective CTC was 19.59 cmol_c dm^-3, the saturation of bases was 90.74%, the organic matter content was 3.30%, the soil C content was 1.91% C, and the clay was 39%.

Correction of the soil pH was not necessary because the value was within the range of requirements for this crop. Fertilization was done with N, P and K as recommended for carrot crops in Brazil. A N dose of 20 kg ha^-1 (as urea) was applied in the furrows before sowing in the control treatment (100% N). This value was reduced to 20% in all other inoculation treatments. 80 kg of N/ha were split-applied at 15, 30 and 40 days after sowing-DAS (16, 24 and 40 kg) in all treatments as recommended for this crop in southern Brazil (^{CQFS, 2016}). Fertilization with P and K (as P₂O₅ and K₂O) was performed in all experiment units. 240 kg ha^-1 of P₂O₅ were applied only at sowing, and 190 kg/ha of K₂O were split-applied at 15, 30 and 40 DAS (38, 57 and 95 kg, respectively).

The carrots were cultivated between 2018 and 2019 in the summer seasons. Seeds of the Livia cultivar (Brasilia group) were used, with 90-100 d required for complete development. The distance between rows was 20 cm, while the distance between seeds was 8 cm. The experiment used completely randomized design with five treatments and five replicates. Each experimental plot was 1 m² and was spaced 0.2 m from the other plots.

Five treatments were tested: 100% N; 20% N; 20% N and inoculation with Bacillus subtilis(20% N+B. subtilis); 20% N and inoculation with Pseudomonas fluorescens (20% N+P. fluorescens); 20% N and co-inoculation with B.subtilisand P. fluorescens(20% N+co-inoculation). All inoculants were provided by Total Biotecnologia (Curitiba, Brazil) and developed in liquid form. Both inoculants had a concentration of 1.0·10⁸ CFU/mL (commercial registration for soybean and maize at Ministry of Agriculture, Livestock, and Supply - MAPA, B. subtilis: PR 9392310112-8; P. fluorescens: PR 9392310104-7). However, there were no recommendations for these inoculants in carrots.

In the 20% N+B. subtilis and 20% N+P. fluorescens treatments, inoculants were applied in the rows at sowing. The inoculants were diluted in water (1 L: 499 L per hectare), and this mixture was sprayed-applied inside the rows before sowing. In the 20% N+co-inoculation treatment, the ratio was adjusted to 1 L of each inoculant, diluted in 498 L of water. Inoculants were also applied in the rows at sowing.

After emergence, weed thinning and controls were performed manually throughout the entire experiment, as recommended by ^{Vieira and Pessoa (2008)}, and no chemical product was used at any stage.

Several parameters were evaluated throughout plant growth and at harvest, as recommended for carrots by ^{CEAGESP (2000)}. The plant height and number of green leaves were evaluated at 38, 45, 52, 66, 80, 95 and 100 DAS.

At harvest, the roots and shoots were separated, and the fresh mass of both components was immediately determined. The root length was measured with a ruler. The root volume was estimated by placing the carrots in a flask with water and measuring the volume of liquid displaced (^{Novoselov, 1960}). The plant material was dried in an air-forced dryer at 65°C until constant weight, and the mass of the dry roots and shoots was determined.

The presence of defects in each root was evaluated (split roots, cracks, purple shoulder, green shoulder and adventitious roots), which was used to calculate the percentage of abnormal roots.

Tests of normality and homogeneity of variances were applied to all data. If the assumptions of ANOVA were fulfilled, then the F test was used. When a significant p value was obtained (P≤0.05), the means were separated with the Scott-Knott test using the ^{R Core Team program (2020)}. The data on root defects did not satisfy any statistical conditions of the ANOVA or non-parametric procedures, even when submitted to several forms of transformation. Therefore, only mean and average values were recorded.

RESULTS AND DISCUSSION

The plant height increased in all inoculation treatments, as compared to 100% N and 20% N from 95 DAS on. The average increase in growth ranged from five to eleven centimeters (Tab. 1). The highest values were observed when the carrots were co-inoculated with B. subtilis and P. fluorescens (73.22 cm), as compared to nitrogen dose (62.26 cm).

Effects of inoculation on the number of green leaves were observed only until 66 DAS (Tab. 2). Inoculation with P. fluorescens was the only treatment that promoted the same number of leaves as 100% nitrogen across all evalutions where statistical differences were observed, namely up to 66 DAS (Tab. 2).

The highest mean fresh shoot mass was detected with 100% N and use of nitrogen, and all inoculation treatments with reduced nitrogen levels decreased shoot biomass production (Tab. 3). On the other hand, root length was increased by inoculation with P. fluorescens and co-inoculation, as compared to 100% N, 20% N and 20% N+B. subtilis. The plants inoculated with both bacteria produced roots that were, on average, four to five centimeters longer than in the plants that were inoculated only with B. subtilis or not inoculated, regardless of the N supply (Tab. 3).

The fresh root mass increased when the plants were inoculated with P. fluorescens and co-inoculated with P. fluorescens and B. subtilis (Tab. 3). Compared to 100% N, co-inoculation promoted an increase of 81.8% in the root mass, whereas inoculation with only P. fluorescens improved the fresh root massby 70%.

More than 50% of the roots were considered normal (Tab. 4). The only treatment in which split roots were not observed was 20% N+co-inoculation. Purple and green shoulders were detected in all treatments except 100% N (Tab. 4).

This is the first study evaluating the effects of inoculation in carrots on developmental parameters throughout the crop entire cycle. This is also the first report on co-inoculation benefits for growth and yield in carrots. These findings contribute to more sustainable agricultural practices linked to economical gains and broadens the range of crop species known to benefit from co-inoculation.

There are no previous records on plant height measurements throughout the entire development cycle in response to inoculation in carrots. Our data suggest that the effects of bacteria on this variable are detected towards the physiological maturity of carrots. The opposite pattern was observed for the number of green leaves. No effects of any treatment were observed after 80 DAS. P. fluorescens was the most efficient bacterium for increasing the number of leaves. Reports on other plant species collaborate the beneficial effects of Pseudomonas on this parameter. ^{Corrêa et al. (2010)}, for example, analyzed the response of lettuce to Pseudomonas chlororaphis and observed 4 to 7% more leaves as a result of inoculation.

All inoculation treatments had smaller shoot mass values. Other studies, however, have shown no significant differences with inoculation. ^{Matsuoka et al. (2016)} inoculated carrots with ACC (1-aminocyclopropane-1-carboxylate) deaminase-producing bacteria isolated from carrots, turnips and sweet peppers (called OFT2, OFT5, and RH7). They observed no effect from inoculation on the fresh leaf mass, with values ranging from 20 to 24 g per plant. ^{Kozusny-Andreani et al. (2014)} tested the effects of 14 unidentified rhizobacteria isolated from Crotalaria spectabilis on carrots and found no difference between the control and inoculation treatments for shoot mass.

Increases in root length in response to inoculation with PGPR have been reported by other researchers. ^{Kozusny-Andreani et al. (2014)} observed a large variation in root length between different isolates (14 - 22 cm) in different treatments, with ten out of 14 bacterial isolates increasing root length as compared to no inoculation. Positive responses in root length are important in carrots because this trait is commercially valuable. Five commercial classes were established in Brazil according to ^{Nick and Borém (2016)}: 10 (>10 and <14 cm), 14 (>14 and <18 cm), 18 (>18 and <22 cm), 22 (>22 and <26 cm) and 26 (>26 cm). Treatments 1, 2 and 3 resulted in values belonging to Class 18. However, only the treatments with inoculation with P. fluorescens and co-inoculation with P. fluorescens and B. subtilis resulted in longer roots, with values belonging to class 22. Therefore, inoculation with these two bacteria may be an alternative for improving the commercial quality of carrots grown with a decreased N supply.

Yield was improved by inoculation with P. fluorescens and co-inoculation with P. fluorescens and B. subtilis.^{Merriman et al. (1974)} first reported the benefits of Bacillus subtilis associated with Streptomyces griseus in carrots, increasing root mass by 17%. ^{Kozusny-Andreani et al. (2014)} tested 14 bacterial strains, and half of them were effective at promoting the fresh root biomass. These authors showed that, while the treatment with only N resulted in 28 g of roots, inoculation with the strain UCCBj-CE1 resulted in more than 60 g of roots. ^{Clemente et al. (2016)} studied the response of carrots to Bacillus subtilis and B. methylotrophicus and observed increased fresh root mass in carrots. Without inoculation, the mean value was 5.9 kg m². Inoculation with B. subtilis increased the fresh root mass (6.5 kg m^-2), as well as inoculation with B. methylotrophicus (7.26 kg m²). On the other hand, ^{Matsuoka et al. (2016)} found no differences in the mass of fresh roots inoculated with two bacterial strains. One of them (RH7) actually reduced the root mass to 37 g, as compared to 48.7 g in the control. Phylogenetic analyses revealed that all strains are related to species that belong to the genus Pseudomonas.

There was no difference between the treatments for dry root mass. However, ^{Kozusny-Andreani et al. (2014)} quantified more dry roots in carrots inoculated with four strains of unidentified rhizobacteria (mean of 5.5 g) and compared them to a nitrogen control (3 g).

The statistical analysis of data on carrot problems (such as split root, cracked roots, purple shoulders, green shoulders or adventitious roots) was not possible. However, co-inoculation was the only condition in which no plants presented split roots. Both biotic and abiotic stresses are known to cause split and cracked roots, often associated with inadequate humidity. Rhizobacteria are known to produce exopolysaccharides that help retain water in the rhizosphere and reduce physical damage in roots (^{Chang et al., 2007}). Rhizobacteria may decrease the occurrence of purple and green shoulders; however, an analysis of the data related to this hypothesis was not possible. According to ^{Patten and Glick (1996)}, rhizobacteria are known to produce indole acetic acid, which promotes root elongation and deeper penetration in soils, resulting in less exposure to the soil surface and sunlight. Further studies should be carried out to better understand the effects of inoculation on reducing carrot defects.

The hypothesis was not confirmed because not all inoculation treatments with reduced N levels were effective in promoting plant development. The benefits of B. subtilis were also less pronounced than those of P. fluorescens. Although co-inoculation was as efficient as inoculation with only P. fluorescens, both treatments were outstanding methods for increasing growth and yield in carrots. Reduced N fertilization to obtain carrots that are classified in commercial classes with a higher aggregate value may motivate farmers to incorporate inoculation as a routine practice when growing carrots.

CONCLUSION

Bacillus subtilis and Pseudomonas fluorescens improve growth and yield in carrots and are plant growth-promoting microorganisms used with reduced levels of nitrogen fertilization.