Virus-free garlic: yield and commercial classification as a function of plant spacing and seed size

Studies on the interaction between garlic plant density and virus-free seed size are scarce in Brazil. Thus, this study was installed to evaluate the effect of plant spacing and seed size on garlic traits and yield for infected and virus-free bulbs. Treatments were arranged in a randomized block design and 2x5x3 factorial combination [infected and virus-free bulbs, five plant spacings (210, 260, 300, 360, and 390 cm 2 per plant), and three bulbous seed sizes (sieve one, two, and three)]. The highest bulb yield was observed for virus-free seeds at a plant spacing of 390 cm 2 plant -1 , while the highest commercial yield was verified for the spacing of 210 cm 2 plant -1 . The combination of virus-free seeds, larger bulbs, and spacing of 332 cm 2 plant -1 promoted the highest leaf area index. For virus-free garlic, lower plant densities resulted in higher yields and garlic bulb quality. The best option for higher yields and improved commercial quality bulbs was the use of medium-sized virus-free bulbils at a plant density of about 300 thousand plants ha -1 .


Research
Horticultura Brasileira 38 (3) July -September, 2020 G arlic (Allium sativum) is propagated exclusively through bulbs and bulbils (clones); however, these may increase the incidence and spread pathogens from one crop cycle to another (Marodin et al., 2019). Viral complexes heavily jeopardize crop growth and yield (Oliveira et al., 2014;Marodin et al., 2019). As a result, seed bulbs from virus-free stock plants have been recommended, as stocks are produced using apical stems from in vitro culture associated with thermotherapy (Oliveira et al., 2014;Vieira et al., 2015). In this sense, virus-free seed bulbs can increase yield, commercial quality of bulbs, and sustainability of garlic production systems (Gimenez et al., 2016;Marodin et al., 2019).
Bulbil plant density is a major factor for garlic production for being able to increase yields of such economicallyimportant vegetable (Rathinakumari et al., 2015). In terms of the final product, small bulbs have less commercial value if compared to larger ones, thus decreasing farmer profits. Moreover, higher garlic yields are observed in systems using high plant densities and larger garlic seeds, especially when combined with high technification (Asgharipour & Arshadi, 2012;Lima et al., 2019).
Planting of larger bulbils can also provide higher yields (Marodin et al., 2019) as they increase vegetative growth for having a surplus of nutritional reserves (Stahlschmidt et al., 1997). Gedamu et al. (2008) observed an interaction between garlic seed size and plant spacing, wherein small bulbils promoted higher yields in denser spacings. On the contrary, less dense spacings only produce larger average weights of bulbs when using larger garlic seeds. Moreover, larger garlic seeds provide greater vegetative growth mainly when associated with increased plant spacing, as it increases light uptake and production of larger bulbs at harvest time (Moravčević et al., 2011).
Despite the above reports, there is a lack of studies on the potential interaction between planting density and size of garlic seeds. Given the non-degeneration, clones free of viral complexes, such as those composed of Allexivirus, Carlavirus, and Potyvirus genera (Fayad-André et al., 2011;Oliveira et al., 2014), provide more vigorous vegetative growth than virus-infected plants (Gimenez et al., 2016). Therefore, plant spacings other than those in systems using virusinfected garlic seeds may be suitable for increased yields.
Given the above, in our study we aimed to evaluate the garlic traits and yield for garlic cropping using virusinfected and -free seeds at different plant densities and seed sizes.

MATERIAL AND METHODS
An experiment was carried out at Embrapa Hortaliças, located in Brasília-DF, Brazil (15°56'S, 48°9'W, 1150 m altitude). According to Köppen's classification, the local climate is classified as Aw type, which stands for a tropical climate with dry winters and rainy summers. The experiment was conducted from May to October 2013, when temperature and relative humidity averages were 21.3°C and 59%, respectively. The local soil is classified as eutrophic Yellow Latosol (Embrapa, 2006), i.e., Oxisol, with a sandy-silty-clayey texture.
The experiment was carried out in a randomized block design, with four replications and split plots. Treatments comprised virus-infected and -free garlic seeds in plots, and plant spacings and seed sizes in subplots. The cultivar Chonan was used as seeds source. Some of these seeds were infected by virus and some were virus-free. Seeds were divided into three size groups (from sieve 1: 15x25 mm; sieve 2: 10x20 mm; and sieve 3: 8x17 mm). Seeds were planted at five plant spacings (210, 260, 300, 360, and 390 cm 2 per plant).
Virus-infected seeds came from the garlic germplasm bank of Embrapa and were maintained under field conditions. Yet, virus-free seeds were obtained using apical stems from in vitro culture, maintained under a controlled environment. Dot-Elisa tests confirmed viral complex presence and absence (Allexivirus, Carlavirus and Potyvirus genera) in the propagation materials. These belong to Embrapa Hortaliças. The cultivation of virus-free materials for a single generation has been proven to not compromise plant development, even if grown close to virus-infected plants (Marodin et al., 2019). The cultivar Chonan is of the noble group and has an average of twelve bulbils per bulb but is susceptible to overgrowth.
Bulbils were planted in beds, and experimental plots were composed of four double rows, totalling 232 plants per plot. The useful area comprised 100 plants in the two central rows, discarding two plants at the end of rows in all plots. Plant spacing varied between single and double rows, as shown in Table 1.
Prior to planting, garlic bulb-seeds were vernalized for 57 days at 3-5°C, in a cold chamber. The average weight of virus-free bulbils from sieves one, two, and three corresponded to 3.75, 2.80, and 1.90 g bulbil -1 , respectively. Yet, the average weight of virus-infected bulbils from sieves one, two, and three corresponded to 3.25, 2.37, and 1.72 g bulbil -1 , respectively.
At 55 days after planting (DAP), water stress was imposed to reduce bulbil sprouting. The plants were kept for 12 days under low water supply. After water stress, two topdressing fertilizations were performed (at 70 and 90 DAP), applying a total of 80 kg N and 80 kg K 2 O ha -1 , using urea and potassium chloride as sources.
Spontaneous plants were controlled by oxadiazon herbicide and manual weeding. Phytosanitary control was performed using iprodione (fungicide) and deltamethrin (insecticide) against Alternaria porri and Thrips tabaci, respectively. Irrigation was performed via conventional sprinkling, as a function of crop water requirements at each garlic cultivation stage, applying 5 mm daily, three times a week.
At 90 DAP, four plants were sampled per plot in the morning. The sampled material was divided into bulbs, pseudostems, and leaves. The area of green leaves was obtained through a bench-top leaf area meter (Area Meter) LI-COR ® , model LI 3100C, and expressed in cm 2 plant -1 . The total fresh weights of bulbs, pseudo-stems, and leaves were measured on a 0.001-g precision scale.
Bulbs were harvested after maturity, during plant senescence, at 130 DAP. Bulb-curing process was performed under shade for 25 days after harvest (Gabriel et al., 2020). The bulbs were cleaned by removing roots and leaves with pruning shears. Then, they were classified into the commercial classes established in Ordinance #242 of 17 September 1992, by the Brazilian Virus-free garlic: yield and commercial classification as a function of plant spacing and seed size.
Commercial bulb production (CBP) was evaluated according to average bulb mass (ABM) and percentages of class-7 (PB7), class-6 (PB6), class-5 (PB5), and class 4+3 (PB4+3) among the bulbs harvested in each plot. Given the absence of 'refuse' bulbs, production was estimated by weighing bulbs from all classes and expressed in t ha -1 . The average weight of produced bulbs was determined by the ratio of total weight to the number of bulbs, with values expressed in g bulb -1 . The rates (%) of class 7, 6, 5, and 4+3 bulbs were determined by the ratio of the number of bulbs in each respective class to the total number of bulbs.
Data were tested for normality and homogeneity of residual variances by the Lilliefors and Bartlett tests, respectively. When the assumptions were met, the data underwent the F-test for analysis of variance, using a joint analysis to compare the effects of using virus-infected and -free seeds. The means of bulb sizes were compared by the Tukey's test (p≤0.05), while the ones of plant spacing were submitted to regression analysis, with derivation for maximum or minimum values. These statistical analyses were performed using the SISVAR software (Ferreira, 2011).

RESULTS AND DISCUSSION
The use of virus-free garlic seeds increased all studied parameters but rates of class 5 (PB5) and class 4+3 (PB4+3) bulbs, regardless of seed size and plant spacing (Tables 2 and 3; Figures 1, 2, and 3).
The planting of sieve-1 virus-free

A B
garlic seeds showed a quadratic response for leaf area as plant spacing increased. In this treatment, a spacing of 332 cm 2 plant -1 reached the largest leaf area (617 cm 2 plant -1 ). Yet, for infected clones, the maximum leaf area (451 cm 2 plant -1 ) was achieved using sieve-1 seeds and a plant spacing of 315 cm 2 plant -1 (Figure 1).
Over the years, virus degeneration can limit vegetative growth of infected plants, mainly when closer spacing and smaller seeds are used (Lima et al., 2019). The larger leaf areas in virus-free materials (Table 2) may have been due to the greater plant vegetative vigour of plants since they need larger spacings to express their maximum growth potential.
For virus-infected garlic, the use of larger seeds may increase vegetative growth due to major nutritional reserve in seed bulbs (Pramesh & Baranwal, 2015). This ensures plants a greater initial growth, higher leaf number and length, which may have resulted in the largest leaf areas (Castellanos et al., 2004). Still, fresh plant weight, average bulb weight, and yield varied with seed size (Table 2), and only sieve-1 bulbs provided higher values for these traits. Bulbs with higher carbohydrate reserves tend to favour plant establishment in the field and hence improve agronomic performance (Lencha & Buke, 2017).
Conversely, for virus-free clones, garlic yield and average bulb weight showed no differences between seed sizes from sieve 1 and sieve 2 ( Table 2). This might have occurred due to a lesser virus degeneration of plants (Marodin et al., 2019). Yet, medium-sized seeds provided vegetative growth potential and bulb yield similar to those of treatments using larger seeds ( Table 2).
The best vegetative growth in virus-free plants was due to better photosynthetic capacity (Moravčević et al., 2011) and nutrient accumulation in these plants (Fan et al., 2017), increasing vegetative vigour, leaf area, and fresh weight compared to the infected clones (Table 2). This corroborates Resende et al. (1999) who observed, in garlic from meristem culture, an increase of 80.3% in shoot dry weight accumulation compared to virus-infected plants. A greater shoot vegetative growth Table 2. Area of green leaves (AGL), fresh plant weight (FPW), commercial bulb production (CP) and average bulb mass (ABM) of garlic, cv. Chonan, virus-free (LV) and infected (IN) in response to three bulbil sizes [sieves 1 (15x25 mm), 2 (10x20 mm) and 3 (8x17 mm  A B C therefore, inversely proportional to plant spacing ( Figure 2). Furthermore, Lima et al. (2019) found that soluble solids, total soluble sugars, and soluble solids/ titratable acidity ratio in bulbs increase as plant spacing decreases, regardless of bulb-seed size and viral infection. From a qualitative point of view, bulbs of higher commercial classes have a better market share (Marodin et al., 2019). Virus-free clones produced more class-7 and -6 and less class-4 and -3 bulbs. Conversely, virus-infected clones produced less class-7 bulbs. The use of virus-free seeds increased by 393% compared to the infected material. Moreover, for virus-free material, largeand medium-sized seeds (from sieve 1 and 2) increased by up to 36% compared to sieve-3 ones (Table 3).
For virus-free garlic, larger spacings promoted a linear increase in class-7 bulbs, reaching a maximum of 36.4% at the largest spacing (390 cm 2 plant -1 ). However, when closer spacings were used, the percentage of class-6 bulbs increased. The excessive increase in plant density reduced the quality and sizes of produced bulbs, mainly for infected seeds, with a significant increase of class 4+3 bulbs (Figure 3).    clones, plant potential expression was enhanced in virus-free ones ( Figure 2). As plant spacing increased, competition for light, nutrients, and water decreased, while photosynthetic efficiency and dry matter accumulation increased. Thus, the vegetative growth was improved, and larger garlic bulbs were produced (Ahmed et al., 2017). Larger plant spacings reduce density per area, decreasing number of produced bulbs. Even with the increase in bulb weight, lower plant densities decreased garlic yield both for virus-free and -infected clones. These traits are, increase photo-assimilates that will be directed to bulbs (Marodin et al., 2019). By contrast, infected plants tend to have reduced vegetative growth, early bulbification, and premature senescence (Fayad-André et al., 2011). The presence of viruses does not always lead to garlic death, but makes plants more susceptible to other biotic stresses and adverse environmental conditions (Marodin et al., 2019).
When virus-free seeds were used, larger plant spacings provided increased plant fresh and average bulb weights. Moreover, if compared to virus-infected Virus-free garlic: yield and commercial classification as a function of plant spacing and seed size.
Since seeds have a high cost and represent a large part of garlic farming investments, the use of medium-sized garlic seeds can decrease planting costs. Moreover, virus-free clones have satisfactory results at larger plant spacings, improving vegetative growth and bulb commercial classification, as well as reducing expenses per area. As a result, using medium-sized virus-free garlic bulbs (sieve 2) as seed material, at a density of 300,000 plants ha -1 , may improve garlic yields and commercial classification of produced bulbs.