Plant materials and early growing conditions.

Two different experiments (Expts. 1 and 2) were conducted, each repeated twice, once in spring and once in summer. Grafted and nongrafted seedlings were used in Expt. 1, and only nongrafted seedlings were used in Expt. 2. Watermelon cultivar Tri-X 313 (Syngenta Seeds, Boise, ID) and interspecific hybrid squash (Cucurbita maxima × Cucurbita moschata) ‘Strong Tosa’ (Syngenta Seeds, Wilmington, DE) rootstock were used. Seeds were germinated in 98-count plug trays filled with Fafard #2 substrate (Sungro Horticulture, Agawam, MA) and placed in a dark growth chamber at 28 °C for 2 days. Seedlings were then moved to a glass greenhouse (92.9 m² floor area; Columbus, OH) set at 19/24 °C night/day air temperatures. The greenhouse temperature was controlled by pad-and-fan evaporative cooling and hot water radiant heating along the perimeter of the greenhouse. Heating and cooling setpoints were 23 ± 1 °C during daytime, and 19 ± 1 °C during nighttime. Supplemental overhead lighting in the greenhouse was provided for a maximum of 15.5 h·d⁻¹ × 1000 W metal halide lamps (PARsource, Petaluma, CA) set to turn off 30 min before dusk. Supplemental lighting was activated when there was less than 350 W·m⁻² of solar radiation (measured by an outdoor weather station on top of the greenhouse) for 10 min and turned off after 15 min of light above 400 W·m⁻². If activated for the full 15.5 h, supplemental light provided 3.6 ± 0.6 mol·m⁻²·d⁻¹ daily light integral (DLI). Seedlings were subirrigated with water (up to 14th day after seeding) and then with nutrient solution (half strength University of Arizona multicrop growth formula) containing (mg·L⁻¹): 95 NO₃-N, 24 P, 175 K, 100 Ca, 30 Mg, 58 S, 49 Cl, and micronutrients.

Grafted plants for Expt.1 were seeded 7 d earlier than nongrafted watermelon plants to account for the 7-d healing period. Grafting was conducted using a single cotyledon grafting method, and then plants were placed for 7 d inside a sealed, clear container (47-L volume, 45.7 × 66.0 × 22.9 cm) (Rubbermaid, Atlanta, GA) in a growth chamber (Conviron, Winnipeg, MB, Canada). Continuous lighting at 120 µmol·m⁻²·s⁻¹ of photosynthetic photon flux density (PPFD) (measured on top of the container) provided by white fluorescent lamps (F72T12-CW-HO, 85 W; GE, Boston, MA) was used for the 7-d healing except for the first 24 h (in darkness). The chambers targeted 28 °C air temperature inside the healing container. Relative humidity inside the healing container was 100% during the first 5 d and gradually lowered over the final 2 d by opening the container lid a small amount at a time. After 7 d in healing, the seedlings were removed from the container and returned to the glass greenhouse and well-watered, followed by 7 d of hardening with reduced watering before the low-temperature treatments, together with nongrafted seedlings.

Low-temperature treatments.

Low-temperature treatments were applied over a 72-h period in both experiments to simulate chilling stress during transportation. Expt. 1 exposed grafted and nongrafted seedlings to 0, 6, 12, 24, or 48 h of 1 °C chilling, and then were moved to a 12 °C growth chamber for the remainder of the 72-h treatment period. Expt. 2 exposed nongrafted seedlings to 0, 24, 32, 41, 44, or 48 h of selected chilling temperatures in a range between −0.4 and 1.2 °C (Table 1) before transferring to 12 °C for the rest of the 72-h period. These low-temperature treatments were performed inside 31-L or 47-L containers (46 × 31 × 23 cm; or 46 × 66 × 23 cm; Rubbermaid) kept in dark temperature-controlled chambers. Different container sizes were used because of space limitations within the chilling chambers. Respectively, 15 or 30 plants were placed into small or large treatment containers. For temperature control in Expt. 1, two identical reach-in growth chambers were used (Model E15; Conviron). For Expt. 2, two chambers were designated for treatments above 0 °C (Sheldon Manufacturing Inc., Cornelius, OR), and those below 0 °C (ThermoFisher Scientific, Waltham, MA), and one chamber was designated for 12 °C (Model E15; Conviron). The locations of treatment containers inside each chamber were randomized.

Table 1.

Chilling temperatures, duration of acute chilling, and duration of 12 °C storage in Expt. 2 (spring). Trt = Treatment.

View Table

Posttransplanting growing conditions.

Following the 72-h low-temperature treatments, seedlings were placed on the greenhouse bench for 2 d before transplanting. Ten randomly selected seedlings from each treatment were transplanted into 3.8-L black plastic pots (diameter at rim = 16.5 cm; one plant per pot) in the greenhouse). The substrate used was BM-1 All-Purpose mix (Berger, Quebec, Canada). The plants were arranged into five double rows (90 cm apart) with 10 plants per row and plants spaced 30 cm apart in-row (3.3 plants/m²), and rows were oriented North/South. As plant growth progressed, plants were individually strung up by vine clips and horticultural twine to a wire 2.2 m above each row of pots to support vines upright. These practices were chosen for effectively observing differences in early plant growth and flower development among plants grown at this high planting density.

All plants were drip-irrigated with the nutrient solution mentioned earlier. Irrigation run time was 3 min, distributing 110 mL per run per plant with up to six runs per day at the height of vegetative growth. The number of irrigation runs was adjusted according to the percent drainage measured from the pot and targeted 30% drainage of the daily irrigation. Drip solution pH and electrical conductivity (EC) were adjusted to 5.5 to 6.5 and 1.3 to 1.4 dS·m⁻¹, respectively, to achieve efflux drainage between 6.0 and 6.5 pH and under 1.7 dS·m⁻¹ EC.

Data collection and analyses.

Seedlings were assessed for visual quality of all true leaves (>1 cm²) 1 h before the start of the 72-h low-temperature treatment, and 24 h after returning to the greenhouse. Visual assessments were conducted by the same person in the same laboratory condition using the Horsfall-Barratt scale (H-B, Table 2) based on approximate percentage leaf area with chlorosis and/or necrosis. For analysis, the H-B scores were converted to interval midpoints (Table 2), and then used to quantify the difference in damage observed over the time between immediately before and 24 h after treatment (ΔMP) for individual plants. This midpoint transformation is commonly used for H-B scores, as it reflects the variable ranges of area damaged (Bock et al., 2010).

Table 2.

Horsfall-Barratt scale values, ranges, and interval midpoints.^z

View Table

Before and immediately after the 72-h low-temperature treatments (except for the first repeat of Expt. 1, for which this analysis was not performed), chlorophyll fluorescence was evaluated on the newest true leaf with an area larger than 2 cm² using a Ciras 3 Portable Photosynthesis System (PP Systems, Amesbury, MA) with an actinic light (Red/Green/Blue at 38%/37%/25%) at a PPFD of 9000 µmol·m⁻²·s⁻¹. Eight plants before, or six plants following the end of the 72-h low-temperature treatments, were randomly selected per treatment and dark-adapted for 1 h at 28 °C before measuring chlorophyll fluorescence. Chlorophyll fluorescence induction parameter Fv/Fm was selected, as it reflects the maximum potential quantum efficiency of photosystem II (Maxwell and Johnson, 2000) and therefore the ideal measurement for assessing leaf photosynthetic capacity as affected by the chilling stress.

After transplanting, the plants were assessed for vegetative growth weekly. Flower anthesis and nodal position of first male and female flowers were recorded daily. Vegetative growth measurements included leaf count, primary vine length, and internodal lengths. These measurements were then converted to a per-day increase.

Environmental conditions were monitored throughout the experiment. Specifically, air temperature in the greenhouse was monitored using calibrated T-type thermocouples (gauge: 36) spaced centrally above the seedling trays or in each experimental block. Additional air temperature and relative humidity (converted to air-based vapor pressure deficit, VPD) were measured using a fan-aspirated HMP60 probe (Vaisala; Vantaa, Helsinki, Finland) inside a radiation shield hung at a central location in the greenhouse. PPFD of natural light was monitored in the greenhouse attic using an LI-90R quantum sensor (LI-COR, Lincoln, NE). Air temperatures in chambers and containers used for healing and low-temperature treatments were monitored by calibrated T-type thermocouples (gauge: 24) (two thermocouples per chamber). All sensor readings were recorded with a CR10X or CR1000 datalogger (Campbell Scientific, Logan, UT). In addition, HOBO U23 dataloggers (Onset Computer Co, Bourne, MA) were placed inside all healing containers in the growth chamber for monitoring relative humidity, later converted to VPD. The same loggers were also placed in one container in each chamber during the chilling treatments.

Experimental design and statistical analysis.

For Expt. 1, the spring and summer repeats were conducted for 22 Feb. to 20 May 2019 and 10 May to 21 July 2019, respectively. For Expt. 2, the summer and spring repeats were for 9 July to 3 Sep. 2019 and 13 Dec. 2019 to 14 Mar. 2020, respectively. For pretransplanting evaluations (visual quality and Fv/Fm), the container used for low-temperature treatments was considered the experimental unit, with plants inside the container considered sampling units. The experimental model included the container [random; degrees of freedom (DF) = 4 in Expt. 1 and DF = 15 in Expt. 2], the treatment (fixed; DF = 9 in Expt. 1 and DF = 6 in Expt. 2), and the residual error.

For the posttransplanting measurements of vegetative growth and floral development, each plant was considered its own experimental unit, as no container effect was found when analyzing our results from the seedling chilling portion of the experiment. Plants were arranged in a randomized complete block design (RCBD) that divided the five double rows into North and South sections (10 blocks with 10 plants per block, 100 total plants) in the greenhouse. Expt. 1 allowed for two plants (one grafted and one nongrafted plant) of each of the five chilling treatments to be placed in each block, and Expt. 2 contained one to two nongrafted plants of each chilling treatment in each block. Thus, the RCBD blocking factor (10 blocks in the greenhouse) was considered as the only random effect to explain variation within the posttransplanting portion of the experiment. The experimental model included the block (random; DF = 9), the treatment (fixed; DF = 9 in Expt. 1 and DF = 6 in Expt. 2), and the residual error.

All statistical analyses were conducted using two statistical software programs (JMP Pro version 14.0.0; SAS Institute Inc., Cary, NC; RStudio version 1.4.1106; PBC, Boston, MA). Linear or nonlinear regressions were made to correlate CDH and specific plant responses (e.g., days to flower) using the JMP Nonlinear Model. A Wald’s test was used to determine the effects of blocking on the model. A Tukey’s honestly significant difference test and a pairwise t test were used to separate means (P ≤ 0.05).

Growth chamber environment.

Temperature differences measured within the containers during germination and graft healing were negligible between containers in the same growth chambers (data not shown). Graft healing temperatures within Expt. 1 averaged 26.9 ± 1.1 °C and 27.7 ± 1.0 °C for the spring and summer repeats, respectively. VPD was 0.08 ± 0.03 and 0.09 ± 0.03 kPa. Light intensity at the top of the containers were 122.0 ± 2.0 µmol·m⁻²·s⁻¹ in both repeats.

During the low-temperature treatments, once containers were placed in the growth chambers, the internal temperature of the container matched the chamber temperature within 120 min and varied within less than 0.2 °C of the setpoints in Expt. 1. However, temperatures achieved within the containers in Expt. 2 varied greatly from the original setpoints we selected because of growth chamber variation. Therefore, our analyses and CDH calculations were based on the temperatures recorded within the containers of each treatment group (as shown in Table 1) instead of setpoint temperatures. VPD within each container used in the experiments was less than 0.1 kPa (data not shown) for both experiments.

Greenhouse environment.

Table 3 summarizes the environmental conditions of the greenhouse. Average solar DLI during the seedling growth stages ranged from ≈5.6 to 24.4 mol·m⁻²·d⁻¹, and 8.1 to 23.7 mol·m⁻²·d⁻¹ during the posttransplanting stages. Air temperatures did not vary greatly from setpoints conducted during spring months when the greenhouse received lower solar DLI. However, this was not the case in the summer months, which experienced consistently higher temperatures than the setpoints.

Table 3.

Greenhouse environmental conditions of the seedling and posttransplanting environment, means ± standard errors.

View Table

Environmental conditions in the greenhouse had a strong impact on the posttransplanting growth of plants following acute chilling. Specifically, none of the posttransplanting variables measured in the repeats conducted in summer months were significantly affected by the low-temperature treatments (data not shown), whereas the other repeats conducted in spring months showed significant facultative responses to chilling treatments (as described later). A possible reason could be a more advanced reproductive development of seedlings relative to the timing of chilling stress when seedlings were grown under higher temperatures in summer as further discussed in a later section. Another reason may be that the higher light intensities, longer daylengths, and higher day and night temperatures during posttransplanting growth and development may enhance the growth rate enough to mitigate any observable delays in flower anthesis. Summer temperature and light conditions masked the effects of acute chilling stress for both experiments. This aspect of light and temperature effects on crop development needs to be further investigated. Thus, results of the spring repeats are primarily discussed.

Expt. 1: Effects of 1 °C chilling durations on grafted and nongrafted seedlings.

The summer repeat of Expt. 1 did not show significant effects of chilling treatments on posttransplanting plant growth, and developmental parameters (data not shown). Therefore, the following data analysis and discussion mainly focuses on the results of the spring repeat of Expt. 1.

Seedling visual quality.

Positive ΔMPs were recorded in both repeats in this experiment, indicating the visual degradation of visual quality as affected by chilling at 1 °C. The cause of degradation was due to leaf chlorosis (yellowing), necrosis, and wilting, and a significant positive linear correlation was observed between ΔMP and duration of 1 °C exposure (Fig. 1A). When chilled, low temperatures can disrupt typical physiological functions by disrupting membrane proteins and lipids that play critical roles in plant structure and processes (Lyons, 1973). Activation energy for basic metabolic processes (e.g., light harvesting in the photosystem) increases as functional proteins decrease in plant cells (Janmohammadi, 2012), and ion leakage can lead to the accumulation of metabolites such as pyruvate, acetaldehyde, and ethanol, which can damage cells when concentrated in high amounts (Lyons, 1973). As a result, cells and tissues are not able to carry out normal function and can result in damage that leads to leaf yellowing, wilting, or necrosis, as we saw in this experiment. Because of the disruptive nature of chilling, plant tissues have reduced capacity to deal with environmental stress and may suffer increased damage posttransplant due to interference with normal mechanisms used to decrease oxidative stress (Allen and Ort, 2001).

(A) Percent change in leaf tissue damage (ΔMP) of grafted and nongrafted seedlings in Expt. 1 (spring), 24 h after the end of chilling as affected by the duration of exposure to 1 °C. Means ± standard errors (n = 5 treatment containers). (B) Chlorophyll fluorescence parameter variable fluorescence/maximum fluorescence (Fv/Fm) as affected by duration of exposure to 1 °C in Expt. 1 (summer). Data of grafted and nongrafted plants were combined, as there was no significant effect of grafting on Fv/Fm. Means ± standard errors (n = 4 treatment containers). (C) Days from transplant to first female flower anthesis for grafted and nongrafted plants as affected by durations of exposure to 1 °C in Expt. 1 (spring). Means ± standard errors (n = 10 blocks).

Citation: HortScience 57, 8; 10.21273/HORTSCI16557-22

• Download Figure

View Full Size

(A) Percent change in leaf tissue damage (ΔMP) of grafted and nongrafted seedlings in Expt. 1 (spring), 24 h after the end of chilling as affected by the duration of exposure to 1 °C. Means ± standard errors (n = 5 treatment containers). (B) Chlorophyll fluorescence parameter variable fluorescence/maximum fluorescence (Fv/Fm) as affected by duration of exposure to 1 °C in Expt. 1 (summer). Data of grafted and nongrafted plants were combined, as there was no significant effect of grafting on Fv/Fm. Means ± standard errors (n = 4 treatment containers). (C) Days from transplant to first female flower anthesis for grafted and nongrafted plants as affected by durations of exposure to 1 °C in Expt. 1 (spring). Means ± standard errors (n = 10 blocks).

Citation: HortScience 57, 8; 10.21273/HORTSCI16557-22

• Download Figure

(A) Percent change in leaf tissue damage (ΔMP) of grafted and nongrafted seedlings in Expt. 1 (spring), 24 h after the end of chilling as affected by the duration of exposure to 1 °C. Means ± standard errors (n = 5 treatment containers). (B) Chlorophyll fluorescence parameter variable fluorescence/maximum fluorescence (Fv/Fm) as affected by duration of exposure to 1 °C in Expt. 1 (summer). Data of grafted and nongrafted plants were combined, as there was no significant effect of grafting on Fv/Fm. Means ± standard errors (n = 4 treatment containers). (C) Days from transplant to first female flower anthesis for grafted and nongrafted plants as affected by durations of exposure to 1 °C in Expt. 1 (spring). Means ± standard errors (n = 10 blocks).

Citation: HortScience 57, 8; 10.21273/HORTSCI16557-22

• Download Figure

In this experiment, ΔMPs for grafted plants were greater than those for nongrafted plants when the durations of 1 °C exposure exceeds 10 h (Fig. 1A). For example, the seedlings chilled for 48 h reached 42.7% ± 5.7% and 19.8% ± 3.2% surface damage for grafted and nongrafted seedlings, respectively. This trend was similar between both repeats and was also observed in our earlier similar study with 3 °C chilling (Ertle et al., 2021). Likely, this was due to the additional stress of graft healing before the chilling treatments that reduced seedling tolerance to chilling temperatures.

Chlorophyll fluorescence.

Chlorophyll fluorescence Fv/Fm before chilling was 0.72 ± 0.01 and did not differ between grafted and nongrafted plants. The Fv/Fm measured immediately after the 72-h low-temperature treatments decreased with increasing the duration of 1 °C chilling regardless of grafting (Fig. 1B). Seedlings with 48 h of 1 °C chilling had the lowest Fv/Fm, 11% lower than the nonchilled control treatment (72 h of 12 °C storage without 1 °C chilling). These significant reductions in Fv/Fm indicate a reduced photosynthetic capacity of the tissue, and is measurable before visually detectable symptoms (e.g., chlorosis, necrosis) emerge.

Chlorophyll fluorescence has been used widely for assessing physiological injuries caused by abiotic stress on plants, including low temperature. Hou et al. (2016) found that Fv/Fm of watermelon seedlings declined from 0.82 to 0.75 after 3 d under a 12/10 °C day/night temperature and an irradiance of 250 µmol·m⁻²·s⁻¹ in a growth chamber. They also showed that reducing light to 80 µmol·m⁻²·s⁻¹ under the same temperature regimen allowed the seedlings to maintain an Fv/Fm of 0.77. Ding et al. (2011) showed that grafted or nongrafted watermelon seedlings stored in darkness at 15 °C for 6 d reduced Fv/Fm from 0.8 prestorage to 0.77 to 0.78 poststorage. However, these levels of reduction in Fv/Fm, including what we observed in the experiment, may not induce a prolonged impact to posttransplanting growth and development. The physiological damage at the seedling stage temporarily reduces the maximum photosynthetic capacity of the whole plant until new leaves are developed and damaged leaves are removed. For example, Smillie et al. (1988) reported cucumber, tomato, and 13 other species stored at 7 °C for 20 h had reduced Fv/Fm that persisted once plants were returned to 21 °C. Damaged tissues result in reduced capacity to combat oxidative stress, and an individual plant must devote more resources to prevent further physiological damage at the cost of its primary productivity. Regardless, advanced leaf injury, especially when visible on transplanting, would be perceived as low transplant quality for growers, especially when plants were produced from a third party.

Vegetative growth and floral development after transplanting.

Regardless of repeat or grafting, none of the vegetative growth parameters were affected by chilling treatments. Grafting affected the primary vine growth (cm/d) and was seen only in the first repeat (data not shown), resulting in increased length for grafted plants. Because of the greenhouse environmental conditions during spring vs. summer, all vegetable growth variables were more advanced in summer and exhibited greater extension growth (primary vine length and internode length) and leaf count in summer than spring, for both grafted and nongrafted plants (data not shown).

Days from transplant to the first male and female flower emergence was significantly affected by grafting (spring and summer) and chilling duration (spring only) where the number of days to reach first female flower anthesis linearly increased as chilling duration at 1 °C increased (Fig. 1C). Nongrafted plants had a greater extent of delay in first female flower development (i.e., a greater slope in the regression), indicating that they were more affected by chilling than grafted counterparts. Grafted and nongrafted plants exhibited delays in female flowering by up to 3.2 and 6.1 d, respectively, over nonchilled grafted and nonchilled nongrafted plants.

Delays in first flowering correlate well with delays in first harvest. For example, 1-week delay in flowering due to grafting is observed when watermelon is grafted onto bottle gourd rootstocks under nonoptimal growing conditions, and this grafting-delayed flowering can also delay fruit maturity by 1 week (Miguel, 1997). In our experiment, delay in flowering by grafting was observed also. This grafting-induced flowering delay is often attributed to the excessive vegetative growth, and recommendations are typically to manage vegetative vigor by optimizing nitrogen fertilization in the field. In contrast, delay in CIs received during seedling stage before transplanting may not be easily manageable, and therefore chilling stress must be avoided during transportation.

As briefly described earlier, differences between the two repeats in terms of chilling effects may be the difference in plant floral bud developmental stage at the time of chilling exposure. Watermelon seedlings begin to differentiate reproductive tissues in the meristem around the time of transplanting, based on preliminary meristem dissections we performed on seedlings. However, little is known about the impact of chilling on Cucurbitaceae (or Solanaceae) flower buds relative to their developmental stages. In contrast, more information is available for other fruit crops. For example, chilling of strawberry (Fragaria ×ananassa) plants (2 °C for 24 h) can decrease ovule viability, increase nonviable pollen, increase cell death, and disrupt development in four of 13 distinct stages of reproductive bud development (Ariza et al., 2015). In apricot trees, prebloom flower buds become susceptible to ice formation and bud drop (flower abortion), despite tolerating lower temperatures during overwintering (Julian et al., 2007). Based on preliminary flower dissections performed on nongrafted watermelon seedlings grown under similar conditions to spring, we were able to identify anthers, but not stigmas, present in the flower buds of some seedlings between 21 and 25 d after germination. The chilling treatments were conducted ≈28 d after germination in the present experiment, which suggests that differentiated flower parts were likely present at the time of chilling. Increased durations of 1 °C exposure during the seedling stage delayed flower emergence for both grafted and nongrafted plants in spring (Fig. 1C), but not in summer. This indicates that increasing chilling stress can cause disruption in reproductive development but may be due to the specific stage of flower development in the meristem. More research is needed to evaluate this theory, as well as confirming the rate of flower bud development under different environmental conditions and the sensitivity of reproductive developmental stages to chilling stress.

Expt. 2: Effects of varied CDHs at selected BT for plant response.

Like Expt. 1, no significant effects of chilling treatments were found in all posttransplanting growth and developmental parameters when conducted in summer months (data not shown). Therefore, the following data analysis and discussion were made using the repeat conducted during spring months.

Seedling visual quality relative to CDH.

After chilling, ΔMP ranged between a 4.1% and 23.4% increase among treatments (Table 1). Overall, ΔMP increased with either increasing duration of chilling exposure (as also seen in Expt. 1) or reducing the chilling temperature. Treatment III (0.4 °C for 44 h; Table 1) was the most damaging, yielding a ΔMP of 23.4%. When compared at a similar mean temperature (e.g., 0.4 to 0.5 °C, Treatment III vs. IV), longer duration of exposure tends to increase ΔMP. When compared at the same duration of exposure (e.g., Treatment V vs. VI) the temperature effect was not clear, likely due to the relatively large variability of temperature in this experiment.

When correlated with CDH computed with varied BT (3 to 15 °C), ΔMP showed a nonlinear response of increasing ΔMP with increasing CDH. Reduced mean square error (RMSE) found for each nonlinear regression analysis (between ΔMP and CDH) decreased as BT increases (data not shown). When 15 °C was selected as a BT (CDH₁₅), ΔMP increased exponentially after 600 CDH₁₅ (Fig. 2A). This indicates that visible seedling damage occurred at temperatures ≤15 °C but was not notable until after 600 CDH₁₅ is reached. After this point, seedling damage rapidly increased. The most damaging chilling treatment (0.4 °C for 44 h, or 746 CDH₁₅; Table 1) yielded a ΔMP of 23.4% of the true leaf area (Fig. 2A). Kozik and Wehner (2014) found that similar true leaf, growing point, and cotyledon damage to watermelon seedlings was accrued with 24 h of 2 °C storage as 36 h of 4 °C storage, indicating a similar facultative response (i.e., reduction in seedling quality) can be achieved with cumulative chilling at different temperatures and durations. Spalholz (2013) demonstrated two cultivars of watermelon seedlings can be stored at 5 °C with 2 µmol·m⁻²·s⁻¹ PPFD and suffer no mortality after 2 weeks. However, one cultivar suffered 40% mortality after 3 weeks, and 100% of the seedlings from both cultivars had died after the fourth week. In this instance, seedling death resembled an exponential increase after 3 weeks. This study may indicate that a threshold of tolerance for chilling stress exists, but can vary depending on the conditions, cultivar, and measure of plant health (e.g., ΔMP and mortality). Possibly, this could be due to a reduced ability to remove harmful metabolites, enhanced production and reduced removal of reactive oxygen species, or a loss of critical structures that compromise tissues. These would result in increasing seedling injury and, eventually, death.

(A) ΔMP (percent of leaf area newly damaged) of seedlings evaluated 24 h after the chilling treatments with increasing chilling degree hours [base temperature (BT) = 15 °C] in Expt. 2 (spring) (n = 16 treatment containers). Means ± standard errors (n = 10 treatment containers). (B) Variable fluorescence/maximum fluorescence (Fv/Fm) ratios of seedlings across chilling degree hours (BT = 3) in Expt. 2 (spring). Means ± standard errors (n = 16). (C) Days from transplant to first female flower anthesis across chilling degree hours (BT = 4 °C) in Expt. 2 (spring). Means ± standard errors (n = 10 blocks).

Citation: HortScience 57, 8; 10.21273/HORTSCI16557-22

• Download Figure

View Full Size

(A) ΔMP (percent of leaf area newly damaged) of seedlings evaluated 24 h after the chilling treatments with increasing chilling degree hours [base temperature (BT) = 15 °C] in Expt. 2 (spring) (n = 16 treatment containers). Means ± standard errors (n = 10 treatment containers). (B) Variable fluorescence/maximum fluorescence (Fv/Fm) ratios of seedlings across chilling degree hours (BT = 3) in Expt. 2 (spring). Means ± standard errors (n = 16). (C) Days from transplant to first female flower anthesis across chilling degree hours (BT = 4 °C) in Expt. 2 (spring). Means ± standard errors (n = 10 blocks).

Citation: HortScience 57, 8; 10.21273/HORTSCI16557-22

• Download Figure

(A) ΔMP (percent of leaf area newly damaged) of seedlings evaluated 24 h after the chilling treatments with increasing chilling degree hours [base temperature (BT) = 15 °C] in Expt. 2 (spring) (n = 16 treatment containers). Means ± standard errors (n = 10 treatment containers). (B) Variable fluorescence/maximum fluorescence (Fv/Fm) ratios of seedlings across chilling degree hours (BT = 3) in Expt. 2 (spring). Means ± standard errors (n = 16). (C) Days from transplant to first female flower anthesis across chilling degree hours (BT = 4 °C) in Expt. 2 (spring). Means ± standard errors (n = 10 blocks).

Citation: HortScience 57, 8; 10.21273/HORTSCI16557-22

• Download Figure

Chlorophyll fluorescence relative to CDH.

After the 72-h low-temperature treatments, the nonchilled control (72 h of 12 °C) had an Fv/Fm of 0.74 ± 0.01. The Fv/Fm decreased either as increasing duration of chilling exposure (as also seen in Expt. 1) or reducing the chilling temperature. The greatest reduction in Fv/Fm was found in the Treatment VI (1.2 °C for 48 h) to be 0.66 ± 0.04, which is an 11% reduction vs. the nonchilled control (Table 1).

When correlated with CDH, Fv/Fm showed a linear decrease with increasing CDH and had a smaller RMSE at lower BTs (data not shown). When 3 °C was selected as BT, the slope of linear equation showed that Fv/Fm decreased by 0.05 every increase of 100 CDH₃ (Fig. 2B), which can be achieved for 50 h at 1 °C as an example. The same CDH₃ induced similar Fv/Fm regardless of chilling temperature and exposure time. For example, Treatments II and IV (CDH₃ = 87) or Treatments III and V (CDH₃ = 124 or 125) had similar Fv/Fm (Table 1).

The regression analyses for Fv/Fm and CDH indicated the stronger correlations at lower BT in the temperature range we examined. This is an opposite trend observed for visual damage (as measured using ΔMP) where higher BT had a stronger correlation (i.e., BT = 15 °C). Further, although physical damage on true leaves (assessed by ΔMP) increased exponentially after a certain threshold was reached, the photosynthetic capacity of leaf tissue (assessed by Fv/Fm) linearly declined with CDH. Photosynthetic capacity declines as light-harvesting complexes are destabilized by low temperatures, and eventually reach a critical point in chilling accumulation that causes a more destructive tissue response (e.g., necrosis) (Lyons, 1973). Damaged tissues have reduced to no ability to harvest light energy, depending on the severity of damage. Reduction in Fv/Fm and visible seedling damage seen in this experiment was like that exhibited by seedlings in Expt. 1, exposed to 48 h of 1 °C chilling.

Vegetative growth and floral development after transplanting relative to CDH.

Vegetative growth was unaffected by any of the chilling treatments. Plants across all chilling treatments averaged 1.5 ± 0.04 cm/d of growth on the primary vine, 6.0 ± 0.07 cm internode lengths, and produced 0.3 ± 0.02 leaves/d, 0.03 ± 0.01 leaves/d, and 0.35 ± 0.03 leaves/d on the primary vine, secondary vine, and whole plant, respectively (data not shown).

Flower development was affected by the chilling treatments, but only for female flowers. Plants reached first male flower anthesis 42.1 ± 0.6 d after transplanting on average for all chilled treatments. Female flower emergence was delayed by either increasing duration of chilling exposure (as also seen in Expt. 1) or reducing the chilling temperature. The greatest delay was in Treatment III (0.4 °C for 44 h) and reached anthesis 4.3 d later than the control (Treatment VII; 11.8 °C for 72 h) (Table 1). Plants in the control (Treatment VII) reached anthesis before all other treatments, 34.6 ± 1.3 d after transplantation.

Significant linear correlations with CDH were found for days to first female flower. Regression analyses showed that RMSE was generally high and similar among the treatments due to relatively large variability (data not shown), but CDH and days to first female flower were found most highly correlated (lowest RMSE among the BT tested) with a BT of 4 °C. With a BT of 4 °C, increase in days (delay) of anthesis of first female flower was 2.6 d every 100 CDH₄ (Fig. 2C), which can be achieved by 20 h at −1 °C for example. Similar CDH₄ (Treatments III and V accumulating 168 and 169 CDH₄, respectively) induced a similar delay (4.0–4.3 d), whereas the treatment temperature (0.4 vs. 0.8 °C) and duration (44 vs. 48 h) are different.

Previously, we exposed grafted and nongrafted watermelon seedlings to 0 to 48 h of 1 °C (Expt. 1 in this report) or 3 °C chilling (Ertle et al., 2021). There were no significant delays in flower anthesis in nongrafted watermelon plants when treated at 3 °C, whereas there were significant delays up to 6 d at 1 °C. Those two experiments were conducted using different cultivars (Tri-X-313 and SSX-8585) examined at 1 and 3 °C, respectively, and therefore the difference in results could be partially attributed to the possible cultivar difference. Nevertheless, when we attempt to apply the CDH₄ to these earlier results, the longest duration (48 h) of 1 and 3 °C chilling accumulates 144 and 48 CDH₄, respectively. We did not see any significant delay with less than 50 CDH₄ in either Expts. 1 or 2, although accumulation greater than 50 CDH₄ in either experiment correlated with delays in floweranthesis. Using CDH₄, we further validated the model using existing data approximating the delay in female flower anthesis (Fig. 3). As plant development rate is highly affected by growing temperature conditions, the expected delays may need to be expressed with thermal time (e.g., degree days) instead of simple time. Nevertheless, there seems to be a threshold at ≈50 to 70 CDH₄ to cause notable delay in female flower development and anthesis.

Delay (d) in female flower anthesis across chilling degree hours at 4 °C base temperature (CDH₄) observed in four datasets when chilling was induced in the seedling stage. (A) Grafted and nongrafted watermelon plants reported by Ertle et al. (2021) with 0 to 48 h of exposure to 3 °C chilling. (B) Nongrafted watermelon plants from years 1 and 2 reported by Korkmaz and Dufault (2001) with 0 to 81 h of exposure to 2 °C chilling. (C) Grafted and nongrafted plants in the present Expt. 1 with 0 to 48 h of exposure to 1 °C chilling. (D) Nongrafted plants in the present Expt. 2 with 24 to 48 h of exposure to −0.4 to 1.2 °C chilling.

Citation: HortScience 57, 8; 10.21273/HORTSCI16557-22

• Download Figure

View Full Size

Delay (d) in female flower anthesis across chilling degree hours at 4 °C base temperature (CDH₄) observed in four datasets when chilling was induced in the seedling stage. (A) Grafted and nongrafted watermelon plants reported by Ertle et al. (2021) with 0 to 48 h of exposure to 3 °C chilling. (B) Nongrafted watermelon plants from years 1 and 2 reported by Korkmaz and Dufault (2001) with 0 to 81 h of exposure to 2 °C chilling. (C) Grafted and nongrafted plants in the present Expt. 1 with 0 to 48 h of exposure to 1 °C chilling. (D) Nongrafted plants in the present Expt. 2 with 24 to 48 h of exposure to −0.4 to 1.2 °C chilling.

Citation: HortScience 57, 8; 10.21273/HORTSCI16557-22

• Download Figure

Delay (d) in female flower anthesis across chilling degree hours at 4 °C base temperature (CDH₄) observed in four datasets when chilling was induced in the seedling stage. (A) Grafted and nongrafted watermelon plants reported by Ertle et al. (2021) with 0 to 48 h of exposure to 3 °C chilling. (B) Nongrafted watermelon plants from years 1 and 2 reported by Korkmaz and Dufault (2001) with 0 to 81 h of exposure to 2 °C chilling. (C) Grafted and nongrafted plants in the present Expt. 1 with 0 to 48 h of exposure to 1 °C chilling. (D) Nongrafted plants in the present Expt. 2 with 24 to 48 h of exposure to −0.4 to 1.2 °C chilling.

Citation: HortScience 57, 8; 10.21273/HORTSCI16557-22

• Download Figure

Korkmaz and Dufault (2001) exposed diploid watermelon seedlings to 2 °C for between 9 and 81 h, and found that seedling quality and posttransplanting growth in an open field was negatively affected with increased chilling. After transplanting, seedlings exposed to longer durations of 2 °C chilling took a longer time to produce secondary vines, reach flower anthesis, and fruit set. We plotted their reported 2-year data of delayed female flower anthesis against our CDH₄ model (Fig. 3B). We found that in both years, anthesis was delayed with increasing CDH₄, and the trend was mostly linear as CDH₄ accumulated. Linear trendlines over each dataset yielded slopes between 0.023 and 0.050. Thus, for every 50 CDH₄ accumulated, female flower anthesis is delayed between 1.1 and 3.7 d for nongrafted plants, and 1.9 d for grafted plants. The slopes for both years of Korkmaz and Dufault (2001) data fell between the slopes derived from our data for nongrafted plants in Expt. 1 and Expt. 2, indicating our model can predict flowering delays with a reasonable accuracy. Although Korkmaz and Dufault (2001) used diploid seedlings, we found our results using triploid seedlings to exhibit similar flowering delays. No data are available comparing chilling tolerance of diploid vs. triploid watermelon seedlings, but the similarity in trends between our results and their study indicate that there is little difference.