Genotypic variability in broccoli hollow stem as affected by plant density under contrasting seasonal weather conditions

Hollow stem formation can cause substantial quality and yield losses in broccoli production. This physiological disorder is closely associated with weight and stem diameter of the inflorescence. These head traits vary depending on variety and growing conditions. The objective of this study was to assess the susceptibility of different broccoli varieties to hollow stem formation under contrasting seasonal weather conditions and to evaluate increased plant density as a crop management strategy to mitigate this disorder. For this purpose, three field experiments were conducted using twelve commercially available varieties (‘Ares’, ‘Batavia’, ‘Covina’, ‘Ironman’, ‘Larsson’, ‘Malibu’, ‘Naxos’, ‘Neliam’, ‘Parthenon’, ‘SV1002BL’, ‘Triton’, and ‘Vicario’), grown at two plant densities (4.5 and 9.3 plants m -2) and across three seasons (early summer, summer, and fall). Vegetative growth, head traits, and the occurrence of stem cavities were evaluated. Four varieties (‘Ares’, ‘Ironman’, ‘Larsson’, and ‘SV1002BL’) exhibited low susceptibility to pith cracking, whereas three varieties (‘Batavia’, ‘Parthenon’, and ‘Vicario’) were highly affected. Under dry conditions during vegetative development, the proportion of heads exhibiting hollow stem remained below 9%, whereas it increased to 53% under warm and humid conditions. Close plant spacing markedly reduced stem and head width and almost completely suppressed cavity formation. Variety-specific threshold values for head weight and stem diameter were identified, above which the severity increased linearly. Overall, the results demonstrate that both variety choice and plant density represent effective management tools for minimizing hollow stem formation in broccoli production.

Broccoli (Brassica oleracea var. italica) belongs to a morphologically diverse group of crop forms that derive from the wild species B. oleracea through long-term domestication and selection (Li et al., 2024). Within this diversification, broccoli is assigned to the arrested-inflorescence lineage, which is characterized by proliferating, developmentally delayed floral meristems forming a compact, immature head that is harvested prior to anthesis (Cai et al., 2022; Zuluaga et al., 2024). Broccoli-like forms likely arose from Mediterranean landraces in antiquity and were subsequently shaped by targeted selection, particularly in southern Italy (Maggioni et al., 2018; Branca et al., 2020). Sustained breeding efforts have generated a broad diversity of broccoli varieties. In Europe alone, more than 200 varieties are registered, differing in growth characteristics, cultivation period, head development, and resilience to abiotic and biotic stresses (Canaday, 1991; Ordiales et al., 2017; Greer et al., 2023; Zhang et al., 2023, 2024; Akter et al., 2025; European Commission, 2025). These genotypic characteristics determine the choice of varieties for agricultural use. A fundamental distinction among broccoli varieties is based on their intended use, differentiating between types grown for processing and those for the fresh market (Ordiales et al., 2017). Fresh-market broccoli is commonly classified into two main types: single-head varieties, which form a large central head similar to cauliflower, and sprouting broccoli, which produces multiple harvestable spears over time (Dahal et al., 2022). In addition, climatic conditions of the production region and seasonal weather patterns strongly influence variety choice. For broccoli cultivation during cooler seasons in spring and fall, varieties with a short cropping duration are preferred, as they reach harvest more rapidly before unfavorable weather conditions occur. High temperatures during summer, particularly during the harvest period, can adversely affect head quality and yield (Kałużewicz et al., 2009; Sahamishirazi et al., 2018). Under such conditions, varieties with improved tolerance to elevated temperatures during head initiation are required to maintain compact and well-formed heads (Wurr et al., 1995; Farnham and Björkman, 2011). Conversely, declining temperatures in fall may also restrict broccoli development and maturation. At temperatures below 5.0 °C, plant growth is effectively arrested (Diputado and Nichols, 1989; Fyffe and Titley, 1989; Wurr et al., 1995; Tan et al., 2000). Consequently, variety choice represents an important management component in broccoli production, enabling adaptation to climatic conditions and compliance with market-specific requirements regarding head size and quality (Wurr et al., 1995; Farnham and Björkman, 2011; Sahamishirazi et al., 2018).

A widespread quality problem in broccoli production is the hollow stem formation, which regularly leads to high crop losses in many growing regions (Tremblay, 1989; Griffith and Carling, 1991; Babik and Elkner, 2002; San Bautista et al., 2005; Grabowska et al., 2007). The development of this physiological disorder is commonly attributed to mechanical stresses within the pith tissue, caused by radial expansion of the stem axis and internode elongation during inflorescence development (Hejnowicz, 1997;Tan et al., 2000; Boersma et al., 2013). Plant growth has been reported to influence the occurrence of hollow stem (Zink, 1968; Hipp, 1974; Boersma et al., 2009). In particular, leaf number has been shown to be positively correlated with disorder severity, whereas increases in plant height and width appear to play only a minor role (Frieman et al., 2026). However, several studies indicate that vegetative growth parameters are of secondary importance for cavity development. Instead, head traits such as stem diameter and head weight are considered the primary determinants, exhibiting strong variety-specific expression (Shattuck and Shelp 1987; Tremblay, 1989; Hussain et al., 2012; Frieman et al., 2026). Although clear genotypic differences in hollow stem susceptibility have been reported, the molecular basis of this disorder remains poorly understood. Genetic mapping studies suggest that hollow stem is a complex phenomenon influenced by genetic factors and environmental conditions, with several quantitative trait loci (QTLs) identified in broccoli (Yu et al., 2019). Pit cracking is primarily associated with single-head-forming broccoli types, which are characterized by a markedly thick stem and high head weight.

Weather conditions associated with growing region and season influence both plant development and the occurrence of hollow stem. Broccoli prefers mean daily temperatures between 16 and 18 °C and is therefore classified as a cool-season crop (Smith et al., 2024). In addition, a water supply of approximately 25–40 mm per week is required to achieve high yields and marketable head quality (Sanderson and Fillmore, 2010). Exceeding this water supply through excessive rainfall or irrigation has been shown to promote the disorder (Tremblay, 1989; Griffith and Carling, 1991; Babik and Elkner, 2002). Accordingly, hollow stem is reported more frequently in temperate climates, whereas in Mediterranean production regions the disorder predominantly occurs during precipitation-rich spring and fall periods (Tremblay, 1989; Griffith and Carling, 1991; Babik and Elkner, 2002; San Bautista et al., 2005; Grabowska et al., 2007).

An effective approach to mitigate hollow stem formation is increasing plant density. As plant spacing decreases, competition among plants for light, water, and nutrients intensifies, thereby restricting vegetative growth and reducing both head weight and stem diameter. This reduction in head traits is accompanied by a decrease in hollow stem incidence, which under very close spacing can even result in complete suppression of the disorder (Zink, 1968; Cutcliffe, 1972; Griffith and Carling, 1991; Sanderson and Fillmore, 2010; Frieman et al., 2026). However, these beneficial effects are associated with changes in canopy structure that alter the microclimate within the crop stand, potentially increasing the risk of head rot and promoting the development of misshapen heads. In addition, closer plant spacing may reduce head weight to an extent that limits marketability of the product (Sanderson and Fillmore, 2010).

The relationships outlined above indicate that cultivation practices and weather conditions substantially influence the risk of hollow stem formation, particularly through their effects on plant morphological characteristics. However, it remains unclear whether these effects vary among genotypes or follow a uniform pattern across varieties. Therefore, the objective of this study was to compare the susceptibility of twelve broccoli varieties to this physiological disorder under contrasting seasonal weather conditions. In addition, the influence of plant density on plant development and the expression of stem cavitation was assessed. Finally, the study examined whether a variety-independent pattern exists in the occurrence of cavities in relation to specific head traits.

We hypothesized that hollow stem formation can be reduced through the choice of less susceptible varieties and the use of close plant spacing. These effects were assumed to be associated with the expression of head traits. In particular, we expected that cavity formation would occur only after exceeding critical values of head weight and stem diameter, with these thresholds varying among genotypes. In addition, we hypothesized that moist weather conditions promote the development of hollow stem.

Plant material and growing conditions

The study included twelve F 1 hybrid broccoli varieties differing in their preferred growing season and growth rate (Table 1). With the exception of ‘Larsson’, ‘Triton’, and ‘Vicario’, all varieties are considered suitable for summer production. ‘Ares’ and ‘Naxos’ are regarded as particularly heat tolerant and therefore well adapted to warm growing conditions. Under moist weather conditions, however, ‘Naxos’ exhibits an increased susceptibility to head rot and is consequently not recommended for cultivation outside the summer months. ‘Ares’ and ‘Batavia’ are characterized by particularly rapid plant development, whereas the cultivation period of ‘Parthenon’ and ‘Triton’ may extend up to 95 days. The varieties also differ in their head traits and, according to breeder information and practical production experience, show varying susceptibility to hollow stem formation. Varieties known for high head weight and yield include ‘Covina’, ‘Ironman’, ‘Malibu’, ‘Neliam’, ‘Parthenon’, ‘SV1002BL’, and ‘Triton’ (Bayer Vegetables Deutschland, 2026; Bejo Zaden, 2026; Rijk Zwaan, 2026; Sakata, 2026). ‘Naxos’ is characterized by comparatively narrow inflorescences with particularly long floret stalks (Figure 1).

The three field experiments were conducted in 2021 and 2022, with one experiment established per season. Based on the timing of the main growth phase, the experiments are hereafter referred to as “Early Summer” (Exp. ES), “Summer” (Exp. S), and “Fall” (Exp. F) (Table 2). Exp. S was conducted in 2021, while Exp. ES and Exp. F were conducted in 2022. Transplanting took place in mid-June 2021 (Exp. S), late May 2022 (Exp. ES), and early July 2022 (Exp. F), with harvests carried out in late September 2021, August 2022, and early October 2022, respectively (Table 2). The experimental sites were located in northwestern Germany on arable land in Warnstedt (52°45′57.0″ N, 8°01′00.9″ E), Varnhorn (52°51′41.5″ N, 8°20′08.2″ E), and Stapelfeld (52°48′12.3″ N, 7°59′58.9″ E). The maximum linear distance between sites was approximately 23 km. At the three locations soil texture is sandy (Exp. S) or sandy loam (Exp. ES and F), with organic matter contents ranging from 1.8% to 3.7%. Soil pH, measured in 0.01 mol L -1 calcium chloride (CaCl 2), was adjusted to 5.5 using quicklime with 90% calcium oxide (CaO). The mineral nitrogen supply (N min), defined as the sum of nitrate-N (NO 3 - -N) and ammonium-N (NH 4 + -N), was set to 250 kg ha -1. Nitrogen fertilization was split into two applications. A base dressing of 54 kg N ha -1 was applied immediately after transplanting. The top dressing was applied at 23 days after transplanting (DAT) in Exp. ES, 34 DAT in Exp. S, and 33 DAT in Exp. F, considering the N min content measured at the beginning of cultivation, at rates of 174, 197, and 151 kg N ha -1, respectively. Granulated calcium ammonium nitrate (13.5% NO 3 -N/13.5% NH 4 -N, Yara GmbH & Co. KG, Dülmen, Germany) was applied broadcast. Plant available phosphorus (P), potassium (K), and magnesium (Mg) were in the optimal to slightly elevated range (Table 1). Prior to transplanting, all sites received a basal application of Korn-KALI ® (31.5% K, 3.6% Mg, 4.8% S, K+S Aktiengesellschaft, Kassel, Germany) at rates of 300 kg ha -1 (Exp. S) and 400 kg ha -1 (Exp. ES and F). To control fungal diseases and insect pests, the pesticides Benevia ® (cyantraniliprole, Cheminova Deutschland GmbH & Co. KG, Stade, Germany), Bulldock ® (beta-cyfluthrin, Nufarm Deutschland GmbH, Cologne, Germany), Calypso ® (thiacloprid, Bayer CropScience AG, Monheim am Rhein, Germany), Forum ® (dimethomorph, BASF SE, Ludwigshafen, Germany), Karate ® Zeon (lambda-cyhalothrin, Syngenta Agro GmbH, Frankfurt am Main, Germany), Luna ® Experience (fluopyram and tebuconazole, Bayer CropScience AG, Monheim am Rhein, Germany), Signum ® (boscalid and pyraclostrobin, BASF SE, Ludwigshafen, Germany) and Trebon ® 30 EC (etofenprox, Certis Belchim B.V., Hanover, Germany) were applied according to standard commercial practice. In Exp. ES, a UV-stabilized polyethylene net (anti-bird net Type 234; HADI GmbH, Marschacht/Rönne, Germany) with a mesh size of 7 mm x 5 mm and a weight of 35 g m -2 was installed over the crop for the first 30 DAT to prevent feeding damage.

Weather data were recorded directly at the experimental sites using in-field weather stations (model METOS ® Basic 6, Pessl Instruments GmbH, Weiz, Austria). At Varnhorn (Exp. S), relatively stable daily temperatures prevailed throughout the growing period, with a mean of 17.5 °C (max. 24.0 °C/min. 11.1 °C). Total precipitation during the 71-day cultivation period amounted to 145 mm (Supplementary Figure 1A). At Warnstedt (Exp. S), the mean temperature was comparable at 16.8 °C (max. 21.8 °C/min. 11.2 °C). Cumulative precipitation during the 72-day growing period reached 170 mm and was higher than in the preceding year (Supplementary Figure 1B). At Stapelfeld (Exp. F), the growing period was longest at 96 days and was characterized by two distinct phases. During the first 69 DAT, warm and dry conditions prevailed, with an average daily temperature of 19.7 °C (max. 28.4 °C/min. 11.9 °C) and a precipitation of 87 mm. Subsequently, temperatures declined to an average of 12.7 °C (max. 19.9 °C/min. 8.2 °C), accompanied with higher precipitation of 104 mm (Supplementary Figure 1C). In addition, the experiment received supplemental irrigation of 50 mm using well water at 60 DAT. Nitrate concentration in the irrigation water was below 0.5 mg L -1, indicating that irrigation did not have a relevant effect on plant N supply.

Trial set-up

The field experiments were conducted in a randomized split-plot design with four replicates. Seedlings were raised in peat press pots (3.8 cm × 3.8 cm × 3.8 cm) and transplanted at the three-leaf stage. Wide plant spacing was 60 cm between rows and 36 cm within rows, representing common commercial practice. In addition, the effect of closer plant spacing of 30 cm x 36 cm was evaluated.

Measurement of vegetative plant growth parameters

Plant growth was assessed based on plant height, plant width, and number of leaves. For data collection, ten plants per plot were selected. All fully expanded true leaves were counted, while cotyledons were excluded. Plant height was defined as the distance from the soil surface to the highest point of the broccoli plant. Plant width was determined as the maximum horizontal canopy spread of each plant, measured from outer leaf tip to outer leaf tip. Two randomly oriented crosswise measurements were taken per plant, with the mean of both measurements representing plant diameter. Mean daily increases in plant height, plant width, and leaf number were calculated by dividing the respective measurement values by the number of cultivation days. The number of leaves was recorded at five sampling dates up to 45 DAT (Exp. ES), 47 DAT (Exp. S), and 46 DAT (Exp. F). At later stages of cultivation, senescence and abscission of older leaves occurred, which would have confounded further assessments. Plant height and width were measured up to 52, 47, and 60 DAT, respectively. No further increase in size was observed thereafter.

Harvest and evaluation of head characteristics

The broccoli heads were harvested in four (Exp. ES and Exp. S) or five (Exp. F) rounds. In the summer planting of 2021, harvests were conducted from 22 to 29 September. In the following year, the harvest period extended from 28 July to 4 August (Exp. ES) and from 19 September to 5 October (Exp. F). Each experimental plot consisted of 48 plants arranged according to the respective spacing treatment. Owing to the different row spacing treatments, plot area varied between 4.1 m 2 (close spacing) and 6.9 m 2 (common spacing). To avoid border effects, ten plants from the central rows of each plot were used for harvest assessments.

Heads were harvested with a stem length of 8 cm, measured from the harvest cut to the first branching of the inflorescence. Subsequently, harvested heads were evaluated based on predefined head traits, including head diameter, stem diameter, and head weight. Head diameter was calculated as the mean of two crosswise measurements taken at the widest part of the head. For stem diameter determination, broccoli heads were first split longitudinally. Measurements were then taken on the inner side of the stem, 3 cm above the harvest cut.

The incidence and severity of hollow stem were assessed on the inner surface of the longitudinally halved heads. All harvested broccoli heads were included in the evaluation. Severity was rated on a scale from 1 (no hollow stem) to 9 (very severe) (Figure 2). Heads with a rating of 2 or higher were classified as affected by hollow stem. When cavities were visible at the cut surface and exhibited browning (possible from a rating of 4), the broccoli was considered unmarketable. At severity scores exceeding 6, pronounced browning and, in some cases, corking were typically observed, with cavities often extending to the cut surface of the harvested produce.

Soil analyses

For basic soil characterization, soil samples were collected on the day of transplanting from a depth of 0–30 cm using a Pürckhauer drill stick, with ten punctures per site. Additional samples for determination of the N min content were collected from a depth of 0–30 cm and 30–60 cm. All soil samples were processed and analyzed in accordance with the guidelines of the Association of German Agricultural and Research Institutes (VDLUFA, 2016). Mineral N and Mg were extracted using aqueous 0.0125 mol L -1 calcium chloride (CaCl 2 method), P and K were extracted using a solution of 0.05 mol L -1 calcium lactate, 0.05 mol L -1 calcium acetate, and 0.3 mol L -1 acetic acid (CAL method), and B with 0.01 mol L -1 CaCl 2 and 0.002 mol L -1 di-ethyl-ene-tri-amine-penta-acetic acid (CaCl 2 /DTPA method). Subsequent analyses were performed spectrophotometrically (NO 3 - -N + NH 4 + -N) and by inductively coupled plasma optical emission spectroscopy (ICP-OES) for P, K, Mg, and B. Soil pH was determined potentiometrically using a glass electrode in an aqueous suspension containing 0.01 mol L -1 CaCl 2.

Statistical analyses

The effects of variety and plant density on vegetative plant growth and head traits were evaluated using a two-way analysis of variance (ANOVA). Multiple mean comparisons were performed using Tukey’s honestly significant difference (HSD) test at α = 0.05. All measured variables met the assumptions required for ANOVA. Statistical analyses were conducted using R software, version 4.4.2 (R Foundation for Statistical Computing, Vienna, Austria).

To analyze the relationship between head traits and the occurrence of hollow stems, a linear response-plateau (LRP) model was employed. This model reflects the observed trend, where an initial constant phase without hollow stem symptoms transitions into a linear increase once a specific threshold of head traits is exceeded. The calculations were performed using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA) following the approach described by Kuzyakov et al. (1997).

Development of broccoli plants

The period from transplanting to the onset of harvest varied between 63 and 80 days across the three field experiments, depending on weather conditions. The harvest period lasted between 8 and 16 days. Under comparable weather conditions (Supplementary Figure 1A, B), broccoli plants in the early summer (Exp. ES) and summer (Exp. S) experiments developed in a similar manner, with cultivation periods of 71 and 72 days, respectively, and harvest windows of approximately one week. The fall planting (Exp. F), which was initially characterized by dry and warm conditions, exhibited a markedly prolonged cultivation period. The combination of high temperatures and low precipitation totaling only 87 mm during the first ten weeks after transplanting constrained plant development, despite an irrigation event of 50 mm at 60 DAT. Physiological maturity was not attained until the cooler and more humid terminal phase of cultivation, which lasted 27 days and received 104 mm of precipitation.

Across all seasons, broccoli plants developed an average of 15 leaves under common plant spacing. Clear variety-specific differences were observed: ‘Triton’ had the lowest number, with eleven leaves, whereas ‘Ironman’ and ‘Neliam’ developed nearly twice as many leaves (Figure 3A). The enhanced leaf production of these varieties was primarily attributable to increased lateral shoot formation. Mean plant height was 53 cm and varied among varieties from 47 cm (‘Malibu’) to 58 cm (‘Triton’) (Figure 3B). Despite considerable variability across seasons, significant differences in plant height among varieties were detected within individual experiments (Supplementary Table 1). A similar pattern was observed for plant width, which ranged from 53 cm (‘Ironman’) to 65 cm (‘Covina’) (Figure 3C). Increasing plant density from 4.5 to 9.3 plants m -2 reduced leaf number and plant width by an average of 18%, while plant height remained unaffected. The number of leaves decreased by two leaves in ‘Malibu’ and ‘Parthenon’ under closer spacing, whereas ‘Neliam’ and ‘Ironman’ produced four and five fewer leaves, respectively. Regarding plant width, ‘Batavia’ was least affected by increased plant density, showing a reduction of only 9% compared with decreases of 15–20% observed in the other varieties (Supplementary Table 2).

In the drought-affected fall planting, plants remained comparatively compact in both height and width, averaging 37 cm and 48 cm, respectively, under common plant spacing and 36 cm and 40 cm under close spacing. Likewise, leaf number was reduced, with averages of 13 and 10 leaves, respectively, compared with the summer plantings (Supplementary Table 2).

Characteristics of broccoli heads

Head traits were significantly affected by plant density, genotype, and their interaction (Supplementary Table 3). Increasing plant density from 4.5 to 9.3 plants m -2 resulted in an average reduction of stem diameter by 15% and head width by 16% (Figures 4A, B). At the same time, head weight declined by 37% (Figure 4C). Accordingly, heads exceeding 500 g were predominantly harvested under wide spacing, whereas under high plant density less than one third of the harvest reached this weight class (Figure 5).

Under common plant spacing, ‘Ares’ developed the largest and heaviest inflorescences, with a stem diameter of 4.3 cm, a head width of 16.6 cm, and a mean head weight of 576 g. ‘Larsson’, ‘Parthenon’, and ‘Triton’ showed comparable head traits, with stem diameters ranging from 4.1 to 4.2 cm, head widths between 15 and 16 cm, and head weights exceeding 520 g. In contrast, ‘Naxos’, ‘Neliam’, and ‘SV1002BL’ remained distinctly smaller, with stem diameters not exceeding 3.7 cm and head weights below 500 g. Among these, ‘Naxos’ exhibited the narrowest heads at 14 cm and was the lightest variety, with a mean head weight of 417 g. With increasing plant density, varieties differed markedly in their head development. The greatest reductions in head weight were observed for ‘Naxos’ (49%), ‘Parthenon’ (46%), and ‘SV1002BL’ (45%). In ‘Naxos’, this decrease was accompanied by a 19% reduction in stem diameter, whereas in ‘Parthenon’ and ‘SV1002BL’ the reduction primarily affected head width (23–28%). By contrast, ‘Batavia’, ‘Covina’, and ‘Malibu’ were least affected by close spacing, showing reductions in head weight of only 22–30%. In addition, head width in ‘Batavia’ and ‘Covina’ decreased by only 5%.

Differences in weather conditions among the field experiments were also reflected in head development. Under humid early summer conditions, broccoli plants grown at common plant density showed the largest heads, with a mean stem diameter of 4.1 cm and a head width of 17.0 cm. Plants in the fall experiment developed stems of comparable thickness, whereas stem diameter in the summer experiment was approximately 0.5 cm smaller. In contrast, head width was reduced in both later planting dates, averaging about 14.6 cm. Likewise, mean head weight was highest in early summer, reaching 621 g and exceeding values recorded in the later cultivation periods by approximately 28%. Under close spacing, these seasonal differences in head traits were generally less pronounced. Beyond these overall trends, marked variety-specific responses to seasonal weather variation were observed. ‘Larsson’, ‘Naxos’, ‘Parthenon’, and ‘Vicario’ developed the largest stem diameters under the dry conditions of the fall experiment. Among these, ‘Naxos’ proved particularly robust, as stem diameter declined only marginally under closer spacing. Head weight of ‘Naxos’, as well as of ‘Parthenon’ and ‘Larsson’, was likewise only weakly affected by the dry and warm conditions. In contrast, other varieties responded more sensitively to wet and warm weather: head weight of ‘Covina’ and ‘Triton’ was significantly reduced in the summer experiment, whereas ‘Ares’ produced the widest heads under these conditions. Overall, these results indicate that the effects of plant spacing and weather conditions on head traits were strongly variety dependent.

Stem hollowness in broccoli

The severity of stem hollowness was significantly affected by plant spacing, variety, and prevailing weather conditions. Averaged across varieties, close spacing reduced the severity of the disorder by approximately 70% (Figures 6A–C), while the proportion of affected heads declined by 56% (Supplementary Table 5). Varieties differed markedly in their susceptibility: ‘Ares’, ‘Ironman’, ‘Larsson’, and ‘SV1002BL’ consistently exhibited mean ratings below 2 and therefore remained largely free of cavities. In contrast, ‘Parthenon’ and ‘Vicario’ showed pronounced symptoms. Between the two summer experiments, which were conducted under comparable weather conditions, only minor differences in stem hollowness were observed. For instance, ‘Naxos’ showed nearly twice the severity in the early summer planting, with a mean rating of 4.0, compared with the later summer experiment (Figures 6A, B). Conversely, disorder severity in ‘Triton’ and ‘Malibu’ increased by approximately 0.7 rating units from early to late summer. Under the dry conditions of the fall experiment, stem hollowness remained consistently low across all varieties, with severity ratings below 2.0, indicating an absence of relevant damage (Figure 6C). Increasing incidence of affected heads was closely associated with higher disorder severity. Both parameters were strongly correlated (R² = 0.82–0.90), demonstrating a high degree of consistency throughout the growing season despite pronounced variation in weather conditions (Figure 7).

The hollow stem formation was closely associated with head traits, which were strongly influenced by variety and growing conditions. In both summer experiments, the first cavities appeared at head weights ranging from 306 to 710 g and stem diameters between 3.1 and 4.3 cm. Head weight generally showed a slightly stronger association with stem hollowness than stem diameter. ‘Naxos’, ‘Vicario’, and ‘Parthenon’ proved highly susceptible, with threshold values below 400 g, whereas less susceptible varieties such as ‘Larsson’, ‘SV1002BL’, and ‘Ironman’ did not exhibit symptoms until head weights exceeded 580 g (Table 3). Once these thresholds were surpassed, disorder severity increased linearly with further increases in head weight and stem diameter. Representative examples of varieties with high, moderate, and low susceptibility to stem hollowness are shown in Figure 8. In ‘Parthenon’, initial symptoms were already observed at a head weight of 322 g and a stem diameter of 3.3 cm, followed by a steep increase in severity, reaching a rating of 6.1 at 580 g and 4.5 cm, respectively (Figure 8A). In contrast, ‘Malibu’ developed initial cracks in the stem only above a head weight of 420 g and a stem diameter of 3.4 cm. In this variety, the increase in disorder severity was considerably less pronounced and did not exceed a rating of 3.2, even at head weights of up to 679 g (Figure 8B). The variety ‘Ares’ remained largely unaffected by cavities across the entire range of head traits evaluated (Figure 8C). For most varieties, threshold values for head weight and stem diameter were higher in the late summer experiment than in the early summer planting. Exceptions were ‘Batavia’, ‘Covina’, ‘Triton’, and ‘Malibu’, which showed a greater tendency toward stem hollowness under early summer conditions (results not shown).

Based on disorder severity and the head weight at which symptoms first appeared under common plant spacing in the summer experiments, the varieties could be classified into three susceptibility groups. Group I comprised ‘Batavia’, ‘Parthenon’, and ‘Vicario’, which developed stem cavities at head weights below 500 g. These varieties were comparatively strongly affected, with mean severity ratings exceeding 2.5, and were therefore classified as highly susceptible to the disorder. Group II included ‘Covina’, ‘Malibu’, ‘Naxos’, ‘Neliam’, and ‘Triton’. In these varieties, stem hollowness also occurred at relatively low head weights, but symptom expression remained weak. Varieties assigned to Group III, which included ‘Ironman’, ‘Larsson’, and ‘SV1002BL’, showed only a low propensity for cavity formation. Symptoms appeared only at head weights above 500 g and remained minor (Figure 9A). Across all experiments, ‘Ares’ did not exceed a severity rating of 2 and was thus the only variety classified as nearly free of stem hollowness (Figure 9B).

Plant spacing distinctly affected the classification of varieties according to their susceptibility to stem hollowness. Under wide spacing, five varieties were assigned to the highly susceptible Group I, whereas only ‘Covina’, ‘Malibu’ and ‘Larsson’ remained in the moderately susceptible Group II. The composition of the low-susceptibility Group III was unchanged under these conditions (Supplementary Figure 2A). Under close spacing, disorder severity was substantially reduced. Only half of the varieties developed any symptoms, which were generally mild. These were predominantly varieties that had been classified as highly susceptible (Group I) under wide spacing. The remaining varieties remained largely free of cavities and did not develop symptoms even at head weights exceeding 400 g (Supplementary Figure 2B).

Variety-specific differences in plant development and hollow stem susceptibility

The twelve broccoli varieties examined in this study, all of them F 1 hybrids, differed in a range of phenotypic traits that varied in their relevance to the development of stem cavities. Among the vegetative growth parameters, the most pronounced genotypic variation was observed in leaf number, which ranged from 11 to 20 leaves per plant under common spacing. The highest leaf numbers were recorded for ‘Ironman’ and ‘Neliam’, primarily reflecting an increased branching. Although lateral shoot development is characteristic of sprouting broccoli types, it may also occur to a limited extent in single-head fresh-market varieties (Stansell and Björkman, 2020). Consistent with these observations, Sarker et al. (2022) and Bagale et al. (2024) reported only minor variation in leaf number among broccoli varieties lacking pronounced branching, with values typically ranging from 12 to 16 leaves per plant. Genotypic differences were also evident for plant height and width. However, deviations from the variety-averaged mean remained below 10% and were therefore comparatively small. Only ‘Triton’ exhibited a greater plant height of 12%, which can be attributed to its distinctly upright, cup-shaped growth habit. In contrast, ‘Covina’ showed the largest deviation in plant width, exceeding the mean by 13%. Similar patterns of limited variability in plant architectural traits have been reported for modern fresh-market broccoli varieties in previous studies (Prashanthi et al., 2022; Sarker et al., 2022; Bagale et al., 2024).

Among inflorescence-related traits, head weight showed the widest variation, ranging from 417 g in ‘Naxos’ to 576 g in ‘Ares’ across all experiments. However, differences of this magnitude were comparatively rare. For most genotypes, deviations in head weight from the variety-averaged mean remained below 6%. Likewise, head diameter and stem diameter exhibited only limited variation, with values predominantly falling within a range of 10%. An exception was observed for the stem diameter of ‘Ironman’, whose mean deviation was slightly higher at 12%. A similarly high degree of uniformity in head-related traits has been reported in other comparative studies on broccoli varieties (Sánchez et al., 2016; Prashanthi et al., 2022; Bagale et al., 2024). This pattern likely reflects the breeding selection of modern broccoli varieties. Population genetic analyses indicate a markedly reduced allelic diversity in F 1 hybrids compared with traditional landraces, accompanied by a shift towards larger and more compact heads with an increased harvest index (Stansell and Björkman, 2020). Never-the-less, Renaud et al. (2014), comparing 23 varieties grown in spring and fall across two locations in the United States, reported head weight variations of up to 38% relative to the overall mean. In that study, F 1 hybrids were evaluated alongside open-pollinated varieties and inbred lines. Notably, high-performing genotypes largely maintained ranking positions irrespective of the cultivation system.

Under summer growing conditions at common plant density, approximately half of the broccoli heads exhibited stem cavities on average. Incidence varied widely among varieties, ranging from 13% to 82% (Supplementary Table 5). Comparable levels of this disorder under similar cultivation conditions have been reported previously (Gruda and Heine, 2002; Bakker et al., 2009; Frieman et al., 2026). With rising incidence, the severity of internal stem damage also increased (Figure 7). Broccoli with a low proportion of affected heads generally exhibited mild symptoms. At incidence levels below 40%, damage ratings predominantly remained below a score of 2 and were therefore considered non-critical. Only significantly higher incidences led to more severe damage to the stem tissue. This pattern corresponds to the relationships described by Frieman et al. (2026), although more severe symptom expression at comparable incidence levels was occasionally observed in that study. This discrepancy may, at least in part, be attributed to the higher nitrogen supply of 300 kg N ha -1 applied in their experiments, which exceeded the fertilization rate used in the present study by 50 kg N ha -1. Several studies have demonstrated that increasing soil mineral nitrogen availability enhances both the incidence and severity of hollow stem disorders in broccoli (Bélec et al., 2001; Moniruzzaman et al., 2007; Hussain et al., 2012). Severe internal damage markedly compromises marketability. In particular, when cavities are clearly visible at the cut surface and exhibit browning, heads are no longer acceptable for retail marketing. Affected produce is consequently left in the field, resulting in substantial economic losses in commercial broccoli production.

The occurrence of stem cavities was closely associated with head size and weight. Heads whose weight and stem diameter remained below variety specific thresholds were largely free of symptoms. Once these critical values were exceeded, the extent of cavitation increased continuously with increasing head weight and stem diameter. The threshold varied markedly among varieties and ranged from 306 to 710 g for head weight and from 3.1 to 4.3 cm for stem diameter (Table 3). In contrast, vegetative growth parameters such as leaf development as well as plant height and width showed no discernible relationship with the severity of stem defects (Supplementary Figure 3). In addition, early maturing varieties such as ‘Ares’ and ‘Batavia’ did not exhibit a higher susceptibility to this disorder than more slowly developing varieties including ‘Larsson’ and ‘Vicario’ (Supplementary Figure 4). These findings are consistent with earlier studies identifying head-related traits as the primary determinants of hollow stem disorder in broccoli (Zink, 1968; Schellenberg et al., 2009; Hussain et al., 2012; Frieman et al., 2026). From a mechanistic perspective, it is plausible that continued increases in stem diameter and head mass beyond the respective thresholds result in elevated mechanical stresses within the shoot axis. Such internal stresses may promote structural damage to the pith tissue, ultimately leading to the characteristic longitudinal splitting associated with hollow stem (Griffith and Carling, 1991; Hejnowicz, 1997; Tan et al., 2000; Boersma et al., 2013).

The wide range of critical values for head weight and stem diameter at which stem cavitation occurred indicates a pronounced variety-specific variability in susceptibility to this disorder. ‘Parthenon’, ‘Vicario’ and ‘Batavia’ developed initial cavities already at relatively low head weights and stem diameters below 400 g and 3.5 cm, respectively, and exhibited the most severe symptom expression. In contrast, first signs of pith cracking in ‘Larsson’, ‘SV1002BL’ and ‘Ironman’ were observed only at head weights exceeding 550 g and stem diameters above 3.8 cm, and these symptoms remained comparatively weak (Figure 9A). ‘Ares’ proved to be particularly resilient, remaining largely unaffected even at head weights above 700 g and stem diameters of up to 4.6 cm (Figure 9B). Although significant differences in head weight and stem diameter were detected among varieties, no general variety-independent relationship between these traits and susceptibility to internal stem defects could be identified. Varieties characterized by particularly high head weights and thick stems did not exhibit an increased risk of cavitation. These findings contrast with several earlier studies in which genotypes with strongly expressed head traits – especially high head weight – were reported to be more prone to hollow stem formation (Cutcliffe, 1972; Griffith and Carling, 1991; Schellenberg et al., 2009; Frieman et al., 2026). However, those studies were largely based on comparisons of only two varieties, thereby limiting their ability to infer general relationships between head traits and hollow stem susceptibility. By contrast, when a broader genotypic range was considered, Shattuck et al. (1986) and Renaud et al. (2014) found no relationship between the expression of head traits and the occurrence of the disorder. Instead, their results suggested that susceptibility to cavity formation is primarily determined by genotype. Our results support this interpretation and further demonstrate that the relevance of individual head traits is strongly variety-specific (Figures 8A–C).

Influence of plant spacing on plant development and hollow stem formation

Reducing plant spacing from the commercial density of 4.5 plants m -2 to 9.3 plants m -2 had a pronounced effect on broccoli growth. Under close spacing, plants developed fewer leaves, exhibited a more compact growth habit, and produced heads that were both lighter and smaller in diameter (Figure 3 and 4). However, the timing of harvest maturity was not or only marginally affected by this practice (Supplementary Table 6). These observations are consistent with previous studies showing that vegetative growth is restricted at plant densities of approximately 5.6 plants m -2 (Agarwal et al., 2007; Sanderson and Fillmore, 2010; Frieman et al., 2026). Head development may already be limited at densities of 3.3 plants m -2 (Hossain et al., 2011). Despite reductions in individual head weight, increasing the number of plants per unit area can enhance total yield at the field scale (Westcott and Callan, 1990; Jett et al., 1995; Agarwal et al., 2007; Schellenberg et al., 2009; Sanderson and Fillmore, 2010; Hossain et al., 2011). However, experiments conducted at extremely high plant densities of 16.7 plants m -2 indicate substantial delays in head maturation as well as an elevated risk of head rot, ultimately resulting in yield losses (Frieman et al., 2026). Head weight is a key parameter for the German food retail market, in which broccoli is usually sold as single units with a minimum weight of 500 g. At common plant spacing, approximately half of the harvested heads exceeded this threshold when averaged across varieties and experiments, while with closer planting, a large proportion of the harvest remained significantly below this weight. Although marketing as bundled produce is generally feasible, a considerable share of heads would still fail to meet the required minimum package weight.

The decline in head weight with increasing plant density was closely reflected in both the incidence and severity of hollow stem (Figure 6; Supplementary Table 5). Under summer growing conditions at common plant spacing, nearly half of the broccoli heads exhibited stem cavities, whereas this proportion decreased to approximately one fifth under close spacing. At the same time, symptom severity was markedly reduced, so that in most cases no economically relevant damage was observed. While more susceptible varieties reached damage ratings between 4.0 and 4.5 on a scale from 1 (no hollow stem) to 9 (very severe) at common spacing, damage scores under high plant density declined to values below 2 for almost all varieties. These results corroborate earlier findings identifying increased plant density as an effective agronomic strategy to mitigate hollow stem disorder in broccoli (Zink, 1968; Sanderson and Fillmore, 2010; Frieman et al., 2026).

The varieties examined responded differently to changes in plant density. Under close spacing, head weight declined by 22–49%, whereas stem diameter was reduced to a lesser extent, by 10–19%. The most pronounced reductions in head weight were observed for ‘Naxos’, ‘Parthenon’ and ‘SV1002BL’, while ‘Vicario’ and ‘Ares’ exhibited the strongest decreases in stem diameter (Supplementary Table 4). Similarly, the incidence of stem cavitation decreased across varieties. Varieties that were highly susceptible under wide spacing (class I) exhibited only moderate susceptibility (class II) when grown at high plant density. Varieties previously classified as moderately or weakly susceptible (classes II and III) remained largely free of hollow stems under close spacing, with the exception of ‘Covina’ (Supplementary Figure 2). ‘Batavia’ showed only minor changes in head and stem traits in response to increased plant density and, consequently, developed the most pronounced symptoms under these conditions. These findings support earlier evidence for genotypic differences in the response of broccoli to plant density with respect to internal stem disorders (Cutcliffe, 1972; Frieman et al., 2026). Accordingly, varieties with a higher inherent susceptibility should be cultivated at closer spacing than less susceptible genotypes. For commercial broccoli production, plant densities between 4.5 and 6.0 plants m -2 appear to represent a suitable compromise to maintain product quality while ensuring sufficient head weight (Cutcliffe, 1972; Sanderson and Fillmore, 2010; Frieman et al., 2026).

Impact of weather conditions on plant growth dynamics and hollow stem occurrence

The broccoli crops at the three experimental sites were grown on comparable sandy to sandy loam soils. Therefore, the pronounced differences in plant development observed among field experiments can be largely attributed to the prevailing weather conditions during the respective growing periods. The dry conditions during the first ten weeks of the fall planting substantially constrained broccoli growth. During this phase, vegetative growth was at times almost completely arrested and resumed only after water availability increased. As a consequence, the onset of harvest was delayed by an average of 20 days. Drought stress is known to impair fundamental physiological processes in plants. Limited water availability reduces stomatal conductance, restricts carbon dioxide uptake, and lowers photosynthetic and transpiration rates, thereby directly constraining assimilate production and plant growth. In addition, water deficit can damage the photosynthetic apparatus through chlorophyll degradation and oxidative stress caused by elevated formation of reactive oxygen species, resulting in injury to cell membranes and photosystems. Nutrient uptake is further reduced under dry conditions, as low soil moisture limits both the mobility and availability of essential elements such as N, P, and K. Overall, these interacting effects explain why broccoli plants exhibit pronounced growth limitations under drought stress (Seleiman et al., 2021; Akter et al., 2025; Mohan et al., 2025). The genotypes evaluated in the present study responded differently to seasonal weather variability. Under the dry conditions of the fall planting, head weights were generally lowest. Only ‘Parthenon’ and ‘Naxos’ deviated from this pattern. ‘Parthenon’ is known to be sensitive to elevated temperatures during head formation and requires cooler conditions for uniform and orderly maturation (Sakata, 2026). Owing to declining daily mean temperatures to 12.7 °C and increasing water availability during the final four weeks of the cultivation period, favorable conditions for head development of this variety prevailed in the fall experiment. In ‘Naxos’ and ‘Triton’, the cup-shaped growth habit likely provided a protective effect against high temperatures and intense solar radiation, offering a plausible explanation for the observed variety-specific responses and consistent with breeder descriptions (Sakata, 2026). Studies by Akter et al. (2025) demonstrate that certain broccoli varieties, such as ‘Early Green’, exhibit a high degree of drought tolerance and can maintain substantial yields even when water availability is restricted to 45% of field capacity. By contrast, most genotypes examined in the present study appeared to depend on adequate water supply, as they consistently produced the heaviest heads under the warm and wet conditions of the early summer planting. Although cumulative precipitation at this site exceeded that of the late summer planting by only approximately 25 mm, higher soil moisture was already present at transplanting. This elevated moisture level was maintained for an extended period due to cool conditions at the beginning of cultivation and the use of net covering until 30 DAT, thereby promoting vegetative growth and head development. Numerous studies confirm that water availability in the root zone is a key determinant for the formation of marketable heads and high yields in broccoli (Sanderson and Fillmore, 2010; Erken et al., 2013; Akter et al., 2025).

In line with head-related traits, the occurrence of stem cavitation also varied strongly in response to weather conditions, with particularly high levels of damage observed during the summer growing periods. In the early summer planting, an average of 53% of harvested heads were affected, while incidence reached 44% in the late summer planting. Both experiments were characterized by sustained warm and humid conditions. Under comparable weather regimes, Frieman et al. (2026) reported that in some cases the entire harvest was affected by internal stem defects. Several other studies likewise describe an increased prevalence of the disorder under conditions combining elevated air temperatures with high water availability resulting from regular rainfall or irrigation (Griffith and Carling, 1991; San Bautista et al., 2005; Sanderson and Fillmore, 2010). In contrast, hollow stem occurrence remained low under the prolonged dry conditions of the fall planting, where only an average of 8% of heads exhibited symptoms. In agreement with this observation, Babik and Elkner (2002) showed in an irrigation study on broccoli yield and quality that limiting water supply to a maximum of 135 mm almost completely suppressed hollow stem formation, whereas an additional 120 mm of irrigation distinctly increased incidence levels. The dry conditions during the fall planting constrained plant growth prior to the onset of the generative phase. Subsequently, inflorescence development proceeded under cooler and wetter conditions, further slowing head growth. This overall reduction in growth dynamics likely contributed to the lower expression of stem cavitation. Boersma et al. (2013) observed the appearance of initial small cavities approximately 49 DAT, during the late stage of floral induction (head diameter 10–15 mm) shortly before bud formation. This suggests that weather-driven effects on plant growth during this specific developmental window may be particularly critical for the initiation of internal stem disorders. However, the underlying physiological and molecular mechanisms remain insufficiently understood and warrant further targeted investigation (Yan et al, 2026).

Although individual broccoli varieties are capable of combining high head weight with low susceptibility to hollow stems, the cultivation of several genotypes within a single production cycle is advisable from a production-oriented perspective. This strategy effectively mitigates risks associated with variable weather patterns, differences in disease susceptibility, and other difficult-to-predict factors. A site- and season-specific selection of varieties can therefore enhance both productivity and plant health in a sustainable manner (Renaud et al., 2014; Sánchez et al., 2016). The evaluation of individual varieties requires a holistic consideration of agronomic and market-relevant traits. This is exemplified by the variety ‘Ares’, which remained largely free of stem cavitation across all experiments and produced the highest head weights under summer growing conditions. However, once optimal head size was reached, ‘Ares’ exhibited very rapid maturation, increasing the risk of premature flowering and yellowing. Consequently, a higher frequency of harvest passes, and close monitoring of maturity were required. By contrast, ‘SV1002BL’ and ‘Larsson’ were characterized by comparatively slow development, such that head formation during spring planting coincided with summer heat periods, while in late summer planting it overlapped with fall precipitation. Under elevated temperatures, these varieties may develop a flattened head shape, rendering the product unmarketable. Heat-tolerant varieties with a cup-shaped growth habit, such as ‘Naxos’ and ‘Triton’, proved advantageous under dry conditions but exhibited an increased susceptibility to head rot under wet weather, likely due to restricted drying of the heads. Wider plant spacing can mitigate this effect by improving air circulation and reducing infection risk (AAFC, 2025). In addition, varieties with domed head architecture, such as ‘Parthenon’ and ‘SV1002BL’, facilitate water runoff and thereby maintain high quality and yield under rainy conditions (Darling et al., 2000; Herren and Cardinali, 2024). These findings underscore the importance of site- and season-specific variety choice that integrates all relevant agronomic and quality-related traits as a foundation for resilient and economically viable broccoli production.

Overall, hollow stem formation in broccoli can be interpreted as the outcome of interactions between genotypic predisposition, environmental conditions, and crop management factors affecting head development and stem growth (Figure 10).

The results of this study demonstrate that genotypic differences play a key role in the development of hollow stems in broccoli. This physiological disorder was closely associated with head-related traits, for which variety-specific critical values could be identified. Below these thresholds, broccoli heads remained largely free of cavities, whereas symptom expression increased markedly once the thresholds were exceeded. Considerable differences were observed among the genotypes examined: highly susceptible varieties developed cavities already at head weights below 400 g, whereas less susceptible varieties remained unaffected even at weights exceeding 550 g. These findings indicate that, while the severity of internal stem disorders is modulated by head traits, susceptibility is primarily genetically determined. In contrast, plant growth rate had no discernible influence on the formation of stem cavitation.

The combination of variety choice and plant density proved to be an effective strategy for reducing the incidence of hollow stem. Close plant spacing consistently resulted in smaller heads and thinner stems, thereby lowering the risk of cavity formation. However, the associated reductions in head weight, together with an increased susceptibility to fungal and bacterial pathogens, partially constrained marketability. This highlights the need to adjust plant density in a way that minimizes internal stem disorders while maintaining sufficient head size and stable plant health.

Hollow stem formation was strongly influenced by weather conditions. Warm and humid growing periods markedly promoted the development of this disorder, whereas dry phases during vegetative growth almost completely suppressed symptom expression. The results suggest that temporary growth limitations prior to or at the onset of inflorescence initiation may enhance the robustness of broccoli plants against internal stem disorders. To what extent this effect can be deliberately exploited through crop management practices, such as irrigation and nitrogen fertilization strategies, to reduce hollow stem risk in broccoli production should be addressed in future research.