ORIGINAL ARTICLES
Enlarging Broccoli (Brassica oleracea L. var. italica) Heads by Extending the Growing Period and Sparse Planting to Increase Floret Yield
2021 Volume 90 Issue 1 Pages 75-84
Details
2021 Volume 90 Issue 1 Pages 75-84
The demand for broccoli (Brassica oleracea L. var. italica) is rapidly growing in Japan, and a new production system to increase broccoli yield is required. During processing, broccoli florets are separated from the head, and floret yield may be increased by enlargement of heads beyond existing retail standards. Therefore, we conducted four cultivation tests over two years and investigated floret yields of 10 cultivars during spring and autumn cropping. We sought to identify high-yielding cultivars when the heads are enlarged via either an extension of the cultivation period or sparse planting. In the control plots, the broccoli heads were harvested at approximately 12 cm in diameter, following normal retail standards. In the extended plots, the cultivation period was extended to the limits of marketable quality. The plant density of these two plots was 3,125 plants per 10 a. In the sparse plots, plant density was half of the other plots, 1,563 plants per 10 a, and the heads were enlarged and harvested as in the extended plots. The floret yield of the control plots was 560–958 kg/10 a, and the floret yield of the extended plots was 905–2,504 kg/10 a. Head diameter of the extended plots was 15–21 cm. In the sparse plots, head diameter was 15–25 cm, and the fresh weight of the florets tended to increase, but floret yield was only 657–1,870 kg/10 a, due to the half number of plants. Across seasons and years, the ‘Grandome’ cultivar had consistently high yields, suggesting that it is suitable for enlarged harvesting. Compared to the ‘Pixel’ cultivar, the photoassimilates of ‘Grandome’ were more concentrated in the main shoot due to the fewer side shoots. Moreover, there were several immature floral shoots in the head of ‘Grandome’ at 12 cm diameter, which may explain the large size. The harvest index also increased with broccoli head size, so the large head production system enabled more efficient production and increased floret yield compared to conventional cropping methods.
The amount of broccoli (Brassica oleracea L. var. italica) consumed in Japan increased 21.6%, from approximately 176,000 t in 2008 to 214,000 t in 2018, and demand is expected to continue to grow (MAFF, 2020 <https://www.maff.go.jp/index.html>). Domestic broccoli production increased 15% during the same period, (approximately 121,000 t to 139,000 t). However, the amount of imported broccoli increased 36%, from approximately 55,000 t to 75,000 t (ALIC, 2020 <https://vegetan.alic.go.jp/vegetan/sch7.do>). Thus, the self-sufficiency ratio for broccoli in Japan decreased from 69% to 65%, indicating that the increase in domestic broccoli production was lower than the increase in demand over this 10-year span. The amount of imported frozen broccoli increased about 2.5 times, from 23,000 t to 57,000 t in the same 10-year span. The increased demand for frozen broccoli is largely due to the food service industry because of its ease of cooking, stable price, and long-term storage capacity (ALIC, 2016 <https://vegetable.alic.go.jp/yasaijoho/senmon/1605/chosa02.html>). Domestic broccoli are predominantly sold in retail outlets because of the high market prices stemming from high fresh vegetable demand for household consumption. The wholesale price of domestic fresh broccoli was 410 JPY·kg−1 at the Tokyo Metropolitan Central Wholesale Market in 2018, while imported frozen broccoli was only 209 JPY·kg−1 (ALIC, 2020 <https://vegetan.alic.go.jp/vegetan/sch7.do>). Therefore, most of the frozen broccoli in Japan is imported, and domestic broccoli is rarely frozen. There is strong growth potential for processing broccoli in Japan, but the market is saturated with imported frozen broccoli. To overcome this problem and promote domestic broccoli for processing use, the production costs of domestic broccoli must be reduced. The broccoli shipping standards for retail in Japan depend on the individual production areas, but broccoli heads 11–13 cm in diameter are usually harvested and prepared 15–20 cm in length from the top to the cut section of the stem. Therefore, the yields and quality are generally evaluated according to this standard (JA Zen-Noh, 2020 <https://www.jazhr.jp/syukkakikaku/>; Nakano et al., 2017, 2020). Several methods to increase broccoli yield were reported, including the harvest of several heads from a single plant to satisfy retail standards (Kodera, 1988; Sato, 2015; Takahashi et al., 2018, 2019). However, broccoli head properties required for processing use are assumed to be different from those in retail. For example, nonpreferred properties for retail such as hollow stem disorder or anthocyanin accumulation may be accepted for processing use if stem parts are not used and floret parts are boiled (the purple color caused by anthocyanin vanishes) before providing to consumers. Dome-shaped heads with good appearance are preferred in retail sales but flat-shaped heads with long floral shoots may be preferred for processing use due to ease in cutting. These evaluation criteria depend on wholesalers or purposes and are unlikely to be established in a uniform manner. However, it is common in Japan that the broccoli for processing use are used as florets, so head diameter does not have to be 11–13 cm at harvest as long as the quality is satisfactory. Without size restriction, it may be possible to increase floret yield per unit area. There are two approaches to increasing floret yield per unit area, increasing the number of heads via dense planting or enlarging the heads via sparse planting, and experiments to identify suitable plant densities and the cultivars were conducted worldwide (Chung, 1982; Damato, 1997; Dufault and Waters, 1985; Schellenberg et al., 2009; Ward et al., 2015; Westcott and Callan, 1990). For example, Jett et al. (1995) found that a single (large) head is preferable to bunching broccoli (many small broccoli heads) for floret production. The authors reported that 3,600 plants per 10 a can produce marketable florets at a rate of 1,080–1,090 kg/10 a and suggested the population should be used for the exclusive production of single-head broccoli with highly marketable floret yields (Jett et al., 1995).
Domestic broccoli cultivars usually harvested according to conventional standards (11–13 cm head diameter) and shipped for retail sales have the potential to be larger, and enlarging the broccoli heads, while ignoring conventional standards, could increase floret yields. To elucidate the effects of enlarged broccoli heads via an extended growth period, we conducted cultivation experiments four times over two years with domestic 10 cultivars to investigate floret yields and identify high-yielding cultivars in spring and autumn cropping. Due to high competition for sunlight and fertilizer between neighboring plants in conventionally planted crops, capability for large growth may be reduced. Therefore, in addition to extending the growth period, we also examined the effects of sparse planting on yields and plant shoot morphology.
The experiments were conducted in 2018 and 2019 at a field in the National Agriculture and Food Research Organization (NARO), Tsukuba City, Ibaraki Prefecture, Japan (36°01' N, 140°06' E). The cultivars used in this study were ‘Pixel’, ‘Ohayo’, ‘Grandome’, ‘Winter dome’ (Sakata Seed Corporation, Yokohama, Japan), ‘Madoka’, ‘Clear’ (Brolead Co., Ltd., Tsu, Japan), ‘Aina’ (Mikado Kyowa Seed Co., Ltd., Chiba, Japan), ‘Yumehibiki’ (Nanto Seed Co., Ltd., Kashihara, Japan), ‘Nozomi #424’ (Nacos Co., Ltd., Tsu, Japan), and ‘Morimidori 180’ (Nozaki Saishujo Ltd., Nagoya, Japan). The ‘Pixel’, ‘Ohayo’, and ‘Yumehibiki’ cultivars are early-maturing, ‘Aina’, ‘Madoka’, ‘Grandome’, and ‘Clear’ are normal-maturing, and ‘Winter dome’, ‘Nozomi #424’, and ‘Morimidori 180’ are late-maturing cultivars. Three growing conditions (control plot, extended plot, and sparse plot) were arranged for each cultivar. In the control plots, the broccoli heads were harvested at 12 cm in diameter, as per conventional retail standards in Japan. In the extended plots, instead of harvesting heads at 12 cm in diameter, we continued growth and enlarged the heads. Heads were finally harvested just before loosening and the loss of marketability. In the sparse plots, plant density was half that in the control and extended plots, and the heads were harvested just before losing marketability, as in the extended plots. Head quality was evaluated subjectively mainly based on the tightness of heads and smoothness of head surface. In the extended and sparse plots, heads with comparable tightness and smoothness to heads at 12 cm in diameter were regarded as marketable.
The dates of seeding and transplanting were February 14 and March 9, 2018, and February 13 and March 13, 2019, respectively. The seeds were sown in cell trays (25 mL × 128 cells) filled with compost (NAPLA type S; YANMAR Co., Ltd., Osaka, Japan). Two weeks after sowing, the seedlings were fertilized with N:P2O5:K2O = 300:300:300 mg in the form of liquid fertilizer (OAT Agrio Co., Ltd., Tokyo, Japan) per cell tray. The seedlings were grown in a greenhouse at > 10°C, and the seedlings were transferred to a plastic greenhouse at >1°C for low-temperature acclimation 3–5 days before transplanting. The centers of the plant beds were 160 cm apart, covered with black mulch film, and with a bed top width of approximately 100 cm. In the control and extended plots, the plants were transplanted into double rows 60 cm apart, with 40 cm spacing between plants. Density was 3,125 plants per 10 a. In the sparse plots, the beds were partitioned as in the control plots, but adjacent planting spots were skipped during transplanting, resulting in an interplant spacing of 80 cm (in double rows 60 cm apart) and a density of 1,563 plants per 10 a. The field was fertilized with N:P2O5:K2O = 40:40:40 kg/10 a, half of which was slow-release fertilizer. After transplanting, the beds were covered with transparent low plastic tunnels (polyethylene film, 0.07 mm thickness). To avoid daytime heat stress and maintain warmth at night, the tunnel was opened in the morning and shut in the evening or kept shut on rainy or cloudy days. The tunnels were fully open from early April onwards. When the apical heads were harvested, the stem was cut to a length of 15 cm from the top of the dome (Fig. 1a). Leafstalks were trimmed to the width of the head diameter, and head diameter and fresh weight (FW) were measured. Florets were cut to about 5 cm in length and FW was measured (Fig. 1b). The head and floret FWs were converted into yields per 10 a. Three plants were used for the measurement from each plot, and there were three independent plots (replications) per cultivar and growing condition in each year. For clarity, we define each part of the head as in Figure 1 (Björkman and Pearson, 1998; Buck, 1956; Francescangeli et al., 2006; Fujime, 1983).
Broccoli head and florets prepared for measurements. (a) The prepared head and (b) cut off.
The dates of seeding and transplanting were August 2 and 29, 2018, and August 2 and 26, 2019, respectively. The seedlings were nursed in rain shelters. Plastic tunnels were not used during this season. All other growing conditions were the same as the spring cropping.
2. Changes in plant shoot morphology (Exp. 2) 1) Changes in shoot weightDuring autumn cropping in 2018, whole aboveground shoots of ‘Pixel’ and ‘Grandome’ were sampled from the control, extended, and sparse plots when the heads reached the respective harvest thresholds. The whole shoot was divided into four parts, the apical head, leaves on the main shoot, side shoots (lateral branches), and the rest of the main stem, and the FW of each part was measured.
2) Changes in the number of nodesIn the spring and autumn cropping of 2019, the number of nodes of ‘Pixel’ and ‘Grandome’ were measured in the control, extended, and sparse plots when the heads reached the respective harvest thresholds. The first internode was directly under the first true leaf, and the last node was directly under the first floral shoot of the apical head.
3. Statistical analysisThe statistical analyses were performed in R (R Core Team, 2015 <http://www.R-project.org.>).
The harvest periods in 2018 and 2019 were from early May to early June. In the extended plots, the days after transplanting (DAT) to harvesting increased approximately 4–9 days compared to the control plots for the early-maturing cultivars ‘Pixel’, ‘Ohayo’, and ‘Yumehibiki’ and the normal-maturing cultivars ‘Aina’, ‘Madoka’, ‘Grandome’, and ‘Clear’ (Table 1). The heads of these seven cultivars enlarged to 17–21 cm in diameter, maintained marketable quality, and the FWs of the heads and florets increased compared to the control plots. The percentage of floret FW in the head also significantly increased in most of the cultivars, from approximately 48–64% in the control plots to 60–79% in the extended plots. The seven cultivars had higher floret yields in the extended plots (1,334–2,282 kg/10 a) than in the control plots (560–935 kg/10 a), and the three highest floret-yielding cultivars were ‘Clear’ (2,282 kg/10 a), ‘Ohayo’ (2,124 kg/10 a), and ‘Grandome’ (2,089 kg/10 a) in 2018 and ‘Grandome’ (2,032 kg/10 a), ‘Madoka’ (1,885 kg/10 a), and ‘Clear’ (1,852 kg/10 a) in 2019. There were no significant differences between years in the extended plots of these cultivars.
Harvest data from spring cropping.
The heads of the late-maturing cultivars ‘Winter dome’, ‘Nozomi #424’, and ‘Morimidori 180’ also enlarged in the extended plots, but their head quality was lower due to heat stress-related disorders. Irregular sized flower buds, bumpy surface and leafy heads were observed in all of the plots (Fig. S1).
2) Autumn croppingThe harvest periods were from late October to early January in 2018 and from early November to late January in 2019. Two typhoons, typhoon #15 (Faxai) on September 9 and typhoon #19 (Hagibis) on October 12, 2019, hit the experimental field and damaged the plants, causing delayed growth and decreased yields in 2019. Heat stress-related disorders for late-maturing cultivars were not observed during this season.
DAT increased in the autumn cropping compared to the spring cropping, regardless of the cultivars or plots (Table 2). In the extended plots, DAT increased approximately 4–30 days compared to the control plots. For most cultivars in the extended plots, the heads enlarged to approximately 15–21 cm in diameter, and the FWs of the heads and florets, the percentage of florets on the head, and the floret yield per 10 a increased significantly. The floret yield increased from 603–958 kg/10 a in the control plots to 1,248–2,504 kg/10 a in the extended plots. The three highest floret-yielding cultivars were ‘Grandome’ (2,504 kg/10 a), ‘Clear’ (2,306 kg/10 a), and ‘Winter dome’ (2,180 kg/10 a) in 2018 and ‘Grandome’ (2,220 kg/10 a), ‘Nozomi #424’ (1,784 kg/10 a), and ‘Clear’ (1,727 kg/10 a) in 2019. There were no significant differences between years in the extended plots of these cultivars.
Harvest data from autumn cropping.
DAT in the sparse plots did not change compared to the extended plots, regardless of year, cropping season, or cultivar (Tables 1 and 2). However, the head diameters of most cultivars tended to increase; in spring 2019, ‘Grandome’ exceeded 25 cm (Table 1). The FWs of the heads and florets tended to increase with increasing head diameter, but the percentage of floret weight on the head did not change compared to the extended plots (Tables 1 and 2). Although individual heads in the sparse plots were larger than in the extended plots, head and floret yields per 10 a in the sparse plots were less than in the extended plots, due to lower plant numbers (Tables 1 and 2).
Heads in the 12, 20, and 24 cm classes of ‘Grandome’ harvested from the control, extended, and sparse plots are shown in Figure 2a. Each head was cut off to the florets, and the flower buds were observed for the upper, middle, and lower parts of the head. There were no clear differences in the shape or size of the flower buds between plots and positions (Fig. 2b, c, d, e).
Heads and florets of ‘Grandome’. (a) The heads harvested from each plot, 1: Control, 2: Extended, and 3: Sparse plots. (b) Excised florets. Florets located on the upper and lower parts of the head were aligned from top to bottom, U: Upper, M: Middle, and L: Lower parts. Close observation of the (c) upper, (d) middle, and (e) lower parts of the florets. Florets marked as U1-3, M1-3, and L1-3 in (b) correspond to the same florets in (c), (d), and (e), respectively.
Whole plants of ‘Pixel’ and ‘Grandome’ from the control, extended, and sparse plots are shown in Figure 3a, b (when the heads reached the harvest thresholds). In the extended plots, the FWs of the shoot, head, and the rest of the main stem for both cultivars significantly increased compared to the control plots (Fig. 3c, d). The percentage of the head within the plant shoot increased from 19.2% to 25.2% in ‘Pixel’ and from 17.0% to 26.7% in ‘Grandome’. In the sparse plots, the FWs of the shoot, head, and the rest of the main stem in both cultivars significantly increased compared to the extended plots. However, the side shoots of ‘Pixel’ vigorously developed (Fig. 3a), resulting in a lower percentage of head within the shoot (22.0%; Fig. 3c). The side shoots of ‘Grandome’ barely developed, even in the sparse plots (Fig. 3b), and the percentage remained low (Fig. 3d). As a result, the FWs of the leaves and heads significantly increased, and the head percentage also increased to 29.0%.
Plant figures and percentages of each shoot part. Whole plant figures at harvest for (a) ‘Pixel’ and (b) ‘Grandome’. All of the leaves were removed to allow easier recognition of the entire structure. As plants were pulled out without care, so the roots in the picture do not reflect the whole roots. The percentages of each shoot part for (c) ‘Pixel’ and (d) ‘Grandome’. The values on the graphs indicate the percentages of the part (%). The same letters within a part were not significantly different (P < 0.05) according to Tukey’s test, using the FW of each part (n = 3).
The number of nodes was approximately 16–17 in ‘Pixel’ and 21–23 in ‘Grandome’ during spring cropping (Table 3). These values significantly increased in both of the cultivars during autumn cropping, approximately 23–24 in ‘Pixel’ and 40–41 in ‘Grandome’. Season (S), cultivar (C), and their interaction (S × C) significantly affected the number of nodes (P < 0.001), but the effects of plot (P) were non-significant.
The effects of season, cultivar, and growing conditions (plots) on the number of nodes.
Observation of the internal structure of the head in the 12 cm class of ‘Pixel’ revealed that all of the floral shoots had risen, flower buds were present at the surface of the head, and there were spaces between peduncles (Fig. 4a). In the 20 cm class, the peduncles were elongated, and the spaces expanded in the head (Fig. 4b). On the other hand, small, light-colored, and tightly packed floral shoots were inside the head in the 12 cm class of ‘Grandome’ (Fig. 4c). The peduncles were elongated, and spaces were observed inside the 20 cm ‘Grandome’ class heads as in ‘Pixel’ (Fig. 4d).
Observations of the internal structure of the heads. The heads of ‘Pixel’ in the (a) 12 cm and (b) 20 cm classes, and ‘Grandome’ in the (c) 12 cm and (d) 20 cm classes. The red arrows in (c) indicate immature floral shoots.
As low density and high nitrogen fertility tend to induce hollow stem (Schellenberg et al., 2009), it may have occurred in some plots, but we considered such a physiological disorder on the stem as of little importance and did not check because only florets would be used for processing in Japan. Moreover, some enlarged heads showed anthocyanin accumulation in autumn cropping due to the effect of lower temperatures resulting from the extended cultivation period, but we did not regard this as degradation on the assumption that broccoli for processing use were usually boiled in Japan.
‘Grandome’ and ‘Clear’ always ranked in the three highest yielding cultivars in the extended plots (Tables 1 and 2), and ‘Ohayo’ and ‘Madoka’ in the spring cropping, and ‘Winter dome’ and ‘Nozomi #424’ in the autumn cropping ranked highly, respectively. There were no significant differences between years among these cultivars in the extended plots. Thus, these results suggest that the cultivars are suitable for the stable production of enlarged heads.
In general, early-maturing cultivars develop and bolt early and come into harvest while plant bodies are relatively small. For example, we previously reported that the FW of the whole shoots and the leaf areas of early-maturing cultivars (e.g., ‘Pixel’ and ‘Yumehibiki’) were significantly lower than for normal-maturing cultivars (e.g., ‘Grandome’ and ‘Madoka’) (Takahashi and Sasaki, 2019). The ‘Ohayo’ cultivar produced a high yield of 2,124 kg/10 a in spring 2018, suggesting that the early-maturing cultivars can also produce numerous florets, depending on the growing condition. However, the heads of early-maturing cultivars are less likely to be large, and extending the cultivation period may have relatively small effects on the overall yield.
Late-maturing cultivars are usually grown during autumn cropping in warm regions in southern and western Japan, transplanted in late summer and autumn, and harvested in winter to early spring (Ogiwara et al., 1990). The plant bodies tend to be large because of the long cultivation period, suggesting the potential to produce large heads. However, normal flower development may be disrupted by the increasing temperatures of spring cropping; broccoli is a green plant vernalization type, and low temperatures are required for late-maturing cultivars in the relatively later stage to induce reproductive transition (Farnham and Björkman, 2011; Fujime, 1983). As expected, the late-maturing cultivar yields were low in spring cropping, and head quality was low (Table 1; Fig. S1). In autumn cropping, which is suitable for late-maturing cultivars, ‘Winter dome’ achieved a high yield in 2018 (2,180 kg/10 a; Table 2). However, the yield was relatively low for the length of growth and below that of the normal-maturing cultivars (e.g., ‘Grandome’ and ‘Clear’; Table 2). Our experimental field may have been too cold for the late-maturing cultivars. ‘Nozomi #424’ had the second highest yield in 2019 (1,784 kg/10 a; Table 2), but this value was still somewhat low. It is possible that the thick stem of ‘Nozomi #424’ allowed it to withstand the strong typhoon winds, while the other cultivars were knocked down and severely damaged. If so, it is not that ‘Nozomi #424’ is a high-yielding cultivar, per se, but that the extent of the damage to ‘Nozomi #424’ was relatively small.
Normal-maturing cultivars require longer growth periods and have larger plant bodies than early-maturing cultivars, so their heads are likely to be larger. Also, the low temperature requirement for reproductive transition is weaker than in late-maturing cultivars, enabling normal head formation during spring cropping and harvesting before the arrival of colder temperatures during autumn cropping. The quality and yields of the normal-maturing cultivars, ‘Grandome’, ‘Clear’, and ‘Madoka’, were consistent across years and seasons, and the percentages of florets for these three cultivars in the extended plots were 68.7–79.3% (Tables 1 and 2). These percentages are high and may explain the high floret yields for these three cultivars; the respective percentages of florets for the early- and late-maturing cultivars were 60.2–71.7% and 53.8–68.4%. The percentage of florets for the ‘Brigadier’ cultivar was 65% (Jett et al., 1995). Although ‘Aina’ is a normal-maturing cultivar and achieved a relatively high floret yield in 2018 autumn cropping (1,749 kg/10 a), the yield for the 2019 cultivation differed (1,267 kg/10 a; Table 2), suggesting inconsistent large head production.
In the extended plots, ‘Grandome’ always achieved high yields, and average head and floret yields were 3,048 kg/10 a and 2,211 kg/10 a, respectively. The average broccoli (head) yield in Japan in 2018 was 999 kg/10 a (MAFF, 2020 <https://www.maff.go.jp/index.html>). From this value, and assuming a floret percentage of 55–75%, the estimated average broccoli floret yield in Japan is 549–749 kg/10 a. This suggests that the head and floret yields of ‘Grandome’ in the extended plots were 3 to 4 times higher than the average yield in Japan. Thus, even if domestic broccoli for processing use were transacted at a price equal to imported frozen broccoli (almost half the price in comparison with domestic fresh broccoli), profits for farmers seem to increase by at least one and a half times as much as conventional shipping for fresh retail markets. Therefore, enlarging the broccoli heads represents a promising way to increase broccoli production in Japan. The harvest index (the percentage of heads in the plant shoots) also increased with head size (Fig. 3d), so enlarging the heads can be an efficient broccoli production system. Suitable cultivars and head size should be considered for each production area, cropping season, and wholesaler quality requirements, however, our results suggest that ‘Grandome’ is the best candidate for stable large head production in spring and autumn cropping.
Effects of sparse planting and the morphological characteristics of ‘Grandome’The plant density of the sparse plots (1,563 plants per 10 a) was low; the recommended plant density for single large head production is 3,000–3,600 plants per 10 a (Jett et al., 1995; Schellenberg et al., 2009) and 3,500–4,000 plants per 10 a is recommended for ‘Ohayo’, ‘Pixel’, and ‘Grandome’ (Sakata seed, 2020 <https://shop.sakataseed.co.jp/layout/sntec/pdf/broccoli_c.pdf>). We allowed large gaps in the sparse planting plots to reduce the competition between neighboring plants, allowing the cultivars to reach their full potential. As a result, the sparse planting enabled further head enlargement and a significant increase in the FWs of the heads and florets of several cultivars compared to the extended plots (Tables 1 and 2). The floret yield of the sparse plots was less than that of the extended plots due to the lower number of plants. However, the floret yields of ‘Grandome’ in the 2018 spring and autumn cropping exceeded 1,800 kg/10 a, which seems much higher than in the control plots. Jett et al. (1995) suggested a possible labor saving advantage from growing fewer plants, and sparse planting may reduce the costs associated with seeds and labor (i.e., nursery seedlings, transplanting, and harvesting). Future studies should investigate the optimal plant density to maximize profit.
DAT, the percentage of florets, or the number of nodes did not differ between the sparse and extended plots (Tables 1, 2, and 3). The lack of a relationship between changes in plant density and changes in DAT is consistent with the report of Jett et al. (1995). Francescangeli et al. (2006) also found that plant density did not influence the vegetative and reproductive periods or the total number of leaves and florets. These previous studies and our results indicate that changes in plant density do not induce developmental changes, such as alterations to the timing of reproductive transition. Therefore, larger broccoli heads in the sparse plots may be due to the increased availability of resources per plant, such as sunlight or fertilizer. However, the percentage increase of head FW in the sparse plots compared to the extended plots differed among cultivars, e.g., a 25.6% increase (898 g to 1,128 g) for ‘Nozomi #424’, 48.3% (735 g to 1,090 g) for ‘Pixel’, and 71.4% (912 g to 1,563 g) for ‘Grandome’ during the 2018 spring cropping (Table 1). The heads of ‘Nozomi #424’ did not develop normally, so we compared the shoots of ‘Pixel’ and ‘Grandome’. The percentage of heads in both cultivars increased in the extended plots compared to the control plots (Fig. 3c, d). In the sparse plots, the percentage of heads for ‘Grandome’ increased compared to the extended plots, but for ‘Pixel’, the percentage of heads decreased, and the percentage of side shoots increased. The side shoots of intact plants are generally non-essential for the life cycle, and excessive branching may be costly (Dun et al., 2006). We have shown that the side shoots of broccoli compete with the main shoot (Takahashi and Sasaki, 2019). In contrast to the ‘Pixel’ cultivar, ‘Grandome’ rarely produces side shoots (Fig. 3b), which may help to concentrate more photoassimilates into the main shoot, resulting in a larger head. There are concerns that head enlargement may be associated with flower bud expansion or that flowering and floret quality may be degraded. However, the ‘Grandome’ flower buds changed little in any of the plots or at any of floret positions (Fig. 2c, d, e). How the head of ‘Grandome’ can be so large without affecting the flower buds was unclear, so we also examined the internal structure of the heads. A comparison of the heads of the two cultivars in the 12 cm class revealed some immature floral shoots inside the heads of ‘Grandome’, but no floral shoots in the heads of ‘Pixel’ (Fig. 4a, c). It appears that the ‘Grandome’ immature floral shoots grew and filled the gaps in the head surface as the head was enlarging, which maintained the tightness of the head. Thus, the number of floral shoots may influence the potential for head enlargement.
Plant shoots and the reproductive organs are composed of a series of repeating units, called phytomers (Pouteau and Albertini, 2011; Schultz and Haughn, 1991). There is a bract at the base of the floral shoot (Fig. 1a) (Buck, 1956; Fujime, 1983), so the floral shoots of broccoli may also be a series of phytomers. Both ‘Grandome’ and ‘Pixel’ had more nodes in autumn cropping than in spring cropping, and ‘Grandome’ had more nodes than ‘Pixel’ in both seasons (Table 3). These results suggest that the ability to produce phytomers during the vegetative phase differs among cultivars (C) and is affected by the cropping season (S). As the interaction effect (S × C) was detected, the impact of cropping seasons on producing phytomers also differed between cultivars. In fact, the number of nodes in ‘Grandome’ increased by 17–19 from spring to autumn cropping, whereas that of ‘Pixel’ increased by only around 7 (Table 3). This difference may be caused by varietal differences in low temperature requirements for reproductive transition, but if such a capability to produce phytomers was maintained and utilized during the reproductive phase for ‘Grandome’, the number of floral shoots would likely increase, resulting in the high floret yield of ‘Grandome’ in autumn cropping.
The number and size of the floral shoots of cultivars used in the 1970s–80s in Japan, for example, ‘Wasemidori’ (Fujime, 1983), are much smaller than the present-day cultivars, indicating that the breeding process has enhanced the increase of floral shoots and flower buds in response to consumer preference for large, dome-shaped broccoli heads. Such breeding processes have gradually enhanced the capacity for head enlargement in broccoli cultivars like ‘Grandome’, but this capacity has been largely overlooked because the heads are harvested according to the conventional retail standards based on old cultivars. The results of this study demonstrated that the enlargement of broccoli heads enabled a large increase in floret yield, while maintaining marketable quality in terms of tightness and surface smoothness. Other qualities, such as shelf life, texture, workability on cutting or frozen characteristics should be investigated for practical application; however, this study provides a useful review of the conventional cropping methods and shipping standards used in broccoli production in Japan.
We thank Mr. Suzuki, Mr. Amagai, Mr. Kuriyama, Ms. Morioka, and Mr. Hattori of the Institute of Vegetable and Floriculture Science, NARO, for growing the plants and supporting the experiments.
