Morphological and anatomical characterization of colchicine-induced polyploids in watermelon

This study aimed to elucidate the effective colchicine concentration to induce polyploidization distinction based on morphological and anatomical features in watermelon. Watermelon seeds were soaked in different colchicine concentrations (0.05, 0.1 and 0.5%) for several durations (24, 48 and 72 h) to induce polyploidization. Putative polyploids were evaluated regarding their morphological and anatomical traits compared to diploids (control). A colchicine concentration of 0.5% for 72 h resulted in the lowest germination percentage with high mortality in putative polyploids compared to in diploids. Morphological traits revealed vigorous growth in putative tetraploids with slower germination speeds, whereas the putative octoploids had suppressed growth compared to tetraploids and diploids. Regarding reproductive biology, the petal number (6), pollen size and viability were remarkably higher in induced polyploids, which confirmed successful tetraploid induction by 0.5% colchicine treatment for 72 h. Similarly, a bigger stomatal size with a lower density was also noticed in induced tetraploids compared to in diploids using the same treatment after anatomical analyses. Meanwhile, a PCA and correlation matrix illustrated that, among the 20 variables, polyploid induction efficiency (PIE), leaf length (LL), guard cell distance (GCD), and pollen viability (PV) were recognized as the most effective morphological and anatomical indicators for successful polyploid induction confirmation with colchicine in watermelon. The present findings provide a basis for distinguishing colchicine-induced polyploids as improved genetic resources to enhance seedless triploid breeding in watermelon.


Introduction
Watermelon (Citrullus lanatus) is one of the most popular and economically viable cucurbitaceous crops in tropical countries including Bangladesh. It is highly relished as a fresh fruit because of its thirst-quenching attribute in addition to many other identified characteristics like size, color, sweetness, nutritional values, etc. (Barai 2016). Global production of watermelon in 2020 was about 101,620,420 tons, with China alone accounting for 59.29% of the total production while secondary producers with more than 1% of the world's production included Turkey, India, Iran, Brazil, Uzbekistan, Algeria, the United States, Russia, Egypt, Mexico and Kazakhstan (FAOSTAT 2021). However, the total watermelon production in Bangladesh hit an all-time high of 1.77 million on 43,347 hectares of land in the 2021 fiscal year (Wardad 2022).
Watermelon is considered a highly attractive fruit whereas one of the main limiting traits of this fruit is the presence of enormous seeds. Usually, consumers prefer to have fruit with few seeds or seedless watermelon because of the unpalatable taste of hard seeds. About 300-500 seeds per watermelon fruit are present depending on the size of fruits (Grant 2020). Therefore, seedlessness is the most desirable trait in watermelon, and seedless cultivars with high fruit quality are available to the consumers of developed countries commanding a high price (Compton and Gray 1993). Polyploid induction can potentially produce watermelons with few seeds. One of the most successful methods for polyploidy induction is seed treatment with colchicine, which has been artificially applied for a range of plant species . Simple methods for early identification of ploidy levels are important for breeding programs, especially when many plants are to be treated. For confirmation of polyploidy, different morphological and physiological traits, particularly pollen diameter, number of chloroplasts, stomatal size and stomatal density, can be studied as an indirect method, and chromosomes can be counted as a direct method. However, flow cytometry (FCM) is considered a more reliable, rapid and simple direct method to analyze a large number of samples in a very short time period (Sattler et al. 2016). But FCM is expensive and facilities are not always available, in which case, morphological, physiological, and cytological characteristics need to be assayed to confirm the conversion of diploidy to tetraploidy (Sabzehzari et al. 2019).
According to the advantages of polyploidy, the current study was designed to obtain colchicine-induced tetraploids as a logical first step for the expansion of genetic resources in further triploid breeding programs of watermelon (C. lanatus). In addition, morphological, physiological and cytological alterations that arise from polyploidization have been documented in many crops. The triploid watermelon was first reported in 1947 by Kihara and Nishiyama in Japan (Kihara and Nishiyama 1947). Since then, improvement efforts have been continued. But unfortunately, limited efforts have been made to produce seedless watermelons in Bangladesh, despite having favorable climatic conditions for the cultivation of watermelon along with an increasing demand in national and international markets. Therefore, the present study addressed the hypothesis that colchicine can induce polyploidization with remarkable changes in morphological and anatomical features that will enable confirmation of polyploids (Fig. 1). The current study aimed to determine the most effective colchicine concentration and exposure duration for polyploid induction. Furthermore, it aimed to identify distinguishable morphological and anatomical traits associated with polyploid induction.

Experimental site
The experiment was conducted at the tissue culture laboratory, experimental field, advanced genetic laboratory of Department of Horticulture and Department of Genetics and Plant Breeding in Bangabandhu Sheikh Mujibur Rahman Agricultural University (BSMRAU), Gazipur-1706, Bangladesh from November 2020 to July 2021. This area belongs to AEZ 28 of Bangladesh called the Madhupur tract (24°09′ N Fig. 1 Schematic of polyploid induction and confirmation with colchicine treatment of watermelon seeds 1 3 latitude and 90°26′ E longitude), with a mean temperature varying from 28 to 32 °C in summer but winter temperatures below 20 °C; the annual rainfall is 1000-1500 m. The soil was clay loam in texture and acidic in nature with a pH of around 5.8 (FAO 1988;Haider et al. 1991).

Plant material and polyploidy induction
Seeds of diploid watermelon (C. lanatus) Thailand-2 variety with a chromosome number of 2n = 22 were used for the polyploidy induction. Fresh, healthy and mature seeds of watermelon were soaked in aqueous colchicine at concentrations of 0 (C 1 ), 0.05 (C 2 ), 0.1 (C 3 ) or 0.5% (C 4 ) (w/v) for 24 (D 1 ), 48 (D 2 ) and 72 h (D 3 ) and kept in an incubator without light to maintain a dark condition. Stock colchicine solutions were prepared by dissolving various colchicine concentrations in sterilized distilled water and 2% (v/v) dimethyl sulfoxide (DMSO), which increases cell permeability and colchicine absorption (Glowacka et al. 2009). The colchicine concentrations and soaking durations were selected based on the findings outlined by Hassan et al. (2020) in pointed gourd. The experiment was replicated three times with 90 seeds per treatment. A similar amount of seeds was also soaked in tap water (without colchicine) for 24, 48 and 72 h as controls. Following the colchicine treatment, the seeds were thoroughly rinsed three times with sterile distilled water and air dried at room temperature. The treated and untreated seeds were planted for germination in polybags with three seeds in each bag. Bags were filled with garden soil and compost mix (2:1). The polybags were cultured under a shaded condition in a nursery, and regular observations were carried out to assess the seed germination and seedling mortality rates of each colchicine and control treatment. At the 3-4 true leaf stage, all seedlings were transferred to small plastic bags and kept in the nursery for further growth and development.

Early screening and transplanting
Early screening was done using surviving 45-day-old germinated seedlings. Colchicine-treated seedlings were compared with control (diploid) seedlings to reduce the workload of further studies and to identify putative polyploids. Polyploid plants were visually determined based on atypical morphology in terms of their cotyledon shape and color and hypocotyl thickness in contrast with diploid plants. After marking deformed seedlings as putative polyploids, seedlings were transplanted into the experimental bed in the main research field. The field layout was maintained in a two-factor randomized complete block design (RCBD) with three replications consisting of 12 treatments with four plants in each treatment. Fertilization and application of pesticides were done at definite time intervals with optimum doses.
Weeding, irrigation, trellising, and pheromone trap setting etc. was also done at appropriate times.

Vegetative attributes
Every germinated and dead seedling was recorded from the beginning of seed sowing up to 40 days after sowing (DAS) to obtain the final germination and mortality percentage. The following equation given by Ajmal Khan and Ungar (1998) was applied to obtain the final germination percentage: where Ni is the number of germinated seeds and N is the number of seeds used. Similarly, the mortality percentage was calculated from the following equation: In addition, the speed of germination was determined using the following formula (Bradbeer 1988): where N 1 , N 2 , N 3 , … N k is the number of germinated seeds observed at time (days) T 1 , T 2 , T 3 , … T k after sowing (this is not the accumulated/cumulative number, but the number of seeds that germinated at a specific time) and k is the total number of time intervals.
After transplanting seedlings, the leaf blade (third leaf from the top) was collected for identifying external differences based on their degree of lobing, apex shape and color according to the guidelines of UPOV (2010), Sashital (2018). At 45 days after transplanting (DAT) the average leaf blade length, width, petiole length and stem diameter were measured with the help of slide caliper. The average plant height was also recorded at 45 DAT from the base to the top of the plant with a measuring tape.

Reproductive characteristics
The number of days required for first flowering in both male and female flowers was recorded. Likewise, the node number of first male and female flower initiation and the number of petals on male flowers were counted. Petal length was measured from the base of the petal to the apex and breadth was calculated at the widest point of the petal with the help of slide calipers. Furthermore, pollen was collected from male flowers for the in-vitro viability test. The pollen collection (1) Germination percentage = Ni∕N × 100 (3) Germination speed (GS) = N 1 ∕T 1 + N 1 + N 2 ∕T 2 + … + N 1 + N 2 + … + T k ∕T k 1 3 was done early in the morning as pollen often loses its viability with increasing temperature. Afterwards, 1-2 drops of 1% acetocarmine solution was added to previously shredded pollen on a slide, and the pollen was observed using a digital microscope (Olympus CX43 equipped with digital camera) at 40× magnification. Pollen grains with a dark red color were considered fertile, and pollen grains that had a lighter color or that were translucent were recorded as sterile. The pollen grains were photographed and images were calibrated using Toupview 3.7 and ImageJ (64bit, Java 1.8.0_172). Final measurements of pollen diameter (µm) were determined using ImageJ.

Anatomical attribute evaluation and putative polyploid estimation
Matured leaves were collected at midday for anatomical observation as it neutralizes the effect of natural opening of stomata (Sabzehzari et al. 2019). A portion of the leaf petiole was soaked in water for water absorption for the clear observation of stomata. The upper epidermal layer of expanded leaves was collected and observed using a digital microscope (Olympus CX43 equipped with digital camera) at 40× magnification for every observation. Three images were taken for each sample to determine the stomatal length, diameter, area, and density; distance between two guard cells; and number of chloroplasts. The stomatal length and diameter and the distance between guard cells were determined using ImageJ as done previously for pollen diameter. The number of chloroplasts was counted from the images and the stomatal area and density were measured according to Paul et al. (2017); the stomatal area was calculated with the equation A = πr 2 , where r is the radius of the stomata. These anatomical features were used to confirm polyploid induction.

Polyploid induction efficiency (PIE)
A low density of stomata per leaf area and high number of chloroplasts in stomata indicate ploidy induction through colchicine. The PIE was calculated using the following equation (Hassan et al. 2020):

Statistical analyses
Means were compared using two-way analyses of variance (ANOVAs) and honestly significant difference (HSD) tests at a 5% probability level were carried out using R (Version 4.1.2). In addition, a correlation matrix, cluster analysis, principal component analysis (PCA) and biplot analysis were performed using the GGally, agricolae, Factoextra, and Corrplot packages in R for further clarification of polyploid induction by colchicine treatment. Data were visualized using Microsoft Excel 2019 and R.

Vegetative morphology
Colchicine concentrations and exposure durations resulted in diversified responses in various vegetative parameters. Cotyledon characteristics such as shape and intensity of green color are presented in Fig. 2. Early screening according to these attributes found that diploids (control treatment) exhibited a narrow elliptic shape and light green color whereas the putative polyploids had a broad elliptic shape and darker green appearance. The hypocotyl was thick under the high colchicine concentration (0.5%) for each exposure duration (Fig. 3). The thickness of this organ was helpful during early screening for polyploid induction. All quantitative characteristics observed at the vegetative stage of watermelon are disclosed in Table 1. At the beginning, colchicine inhibited germination. For the final germination (40 DAS), the highest rate (100%) was found under the control treatment for 72 h and the lowest rate was under the 0.5% colchicine treatment for 48 h (77.8%). However, the mortality rate was high in the field. Results revealed that the highest colchicine concentration (0.5%) for 72 h had the highest influence on mortality rate (42.2%), and the lowest (4) PIE (%) = (Number of induced polyploids) ∕(seedling survival rate (%)) × 100 Fig. 2 Cotyledons of diploids and putative polyploids 1 3 mortality rate was registered in the control treatment for 48 h (6.3%). A similar trend was also noticed regarding the speed of germination, with higher concentrations of colchicine resulting in lower speeds of germination. More specifically, it was revealed that the 0.5% colchicine treatment for 24 h resulted in the lowest speed (466.49). Meanwhile, the highest speed was found under the control treatment for 48 h (563.10). The germination speed ranged from 547.96 to 5470.24 for the other treatments (Table 1).
There were diversified responses in leaf blade phenotypic appearance and quantitative attributes ( Fig. 4 and Table 1).
The leaf green color intensity was darker in diploids (control treatments) and tetraploids (control 0.05, 0.1 and 0.5% with three durations) but was lighter in some of the putative octoploid plants grown under the 0.5% colchicine treatments. Interestingly, the degree of lobing (medium, weak or strong) varied according to the ploidy level (Fig. 4).

Reproductive morphology
Male flowers exhibited differences after colchicine treatments. Differences were noticed in the petal number of male flowers. Generally, the petal number of diploid male flowers is five (zygomorphic) in watermelon plants. However, in the present study, some flowers had six petals in plants treated with 0.5% colchicine (Fig. 5).
Quantitative characteristics varied with the different colchicine concentrations and exposure durations ( Table 2). The petal length under the 0.1% colchicine treatment for 24 h was 17.0 mm; however, the 0.5% colchicine treatments for 24 and 48 h resulted in no flower production for the whole flowering season. Moreover, 0.5% colchicine for 72 h resulted in the lowest petal length (8.5 mm). The control treatments results in a petal width of 6.4-6.7 mm, which was similar to the lowest petal width (5.2 mm) observed under the 0.5% treatment for 72 h. For the rest of the treatments, the average petal width was 9.0-10.5 mm; 10.5 mm for the 0.05% colchicine treatment for 24 h. The highest average node number for male flower initiation was 4.8 under the control treatment for 24 h and under the 0.1% treatment for 48 h; on the contrary, the lowest (3.0) was found under the 0.5% colchicine treatment for 72 h. Early flowering (the lowest number of days required until flowering) was noticed in the plants treated with 0.1% for 72 h (20.7) and 0.5% colchicine for 72 h (21.8). The other treatments required 42.7 to 46.0 days to initiate the first male flower. Pollen features like viability and diameter gave precise outcomes to distinguish diploids and polyploids (Fig. 6). Smaller and larger pollen indicated diploids and tetraploids, respectively, whereas a mixture of different pollen sizes indicated mixoploids. Subsequently, significant differences were observed for pollen diameter, where the highest (5.6 µm) pollen diameter was measured under the 0.5% colchicine treatment for 72 h, and plants under the remaining treatments had similar pollen diameters as those of control plants, ranging from 4 to 5 µm. Furthermore, the lowest viability (57.0%) was observed under the control treatment for 48 h. However, the control treatment for 72 h and 0.1% treatment for 48 h also showed high percentages of viability (95.2% and 94.4%, respectively).

Anatomical attributes to identify polyploids
Diploid plants had 8-9 chloroplasts within the stomatal guard cells of leaf blades; this was observed in control plants. However, putative mixoploids contained 10-12 chloroplasts within the stomata and sometimes chloroplasts were fused together within the guard cells and were not clearly visible. Interestingly, putative polyploids (tetraploids and octoploids) generally had 14-18 chloroplasts (Fig. 7). Usually, plants treated with 0.1% and 0.5% colchicine, regardless of treatment duration, possessed 14-18 chloroplasts. Quantitative anatomical characterization related to stomatal attributes is visualized in Fig. 8. These attributes were found to be beneficial for PIE confirmation. The lowest stomatal length (15.9 µm) was found under the 0.05% colchicine treatment for 24 h (C 2 × D 1 ) and the highest stomatal length was found under the 0.5% colchicine treatment for 48 h (C 4 × D 2 ). Most treatments resulted in stomatal lengths above 20 µm (C 1 × D 2 , C 2 × D 3 , C 3 × D 1 , C 3 × D 2 , C 3 × D 3 , C 4 × D 1 , C 1 × D 3 ). Other treatments resulted in stomatal lengths below or near 20 µm (C 1 × D 1 , C 1 × D 3 , C 2 × D 2 ). The highest concentration 0.5% (C 4 × D 1 , C 4 × D2, C 4 × D 3 ) resulted in an average stomatal diameter ≥ 13.0 µm. The other treatments resulted in average stomatal diameters between 10.6 and 12.8 µm, with the lowest stomatal diameter (10.6 µm) under the control treatment for 72 h (C 1 × D 3 ). The 0.5% colchicine treatment resulted in high average stomatal areas (171.2 µm 2 for 48 h, 152.8 µm 2 for 72 h and 136.4 µm 2 for 24 h). The other treatments had stomatal areas below 130 µm 2 . The lowest stomatal area (88.6 µm 2 ) was found under the control treatment for 72 h (C 1 × D 3 ). In summary, higher colchicine concentrations resulted in larger stomata. Regarding the distance between two guard cells, there were some distinguishable responses found; the greatest distance (4.0 µm) was under the 0.5% colchicine treatment for 72 h (C 4 × D 3 ) and the lowest distance (1.8 µm) was under the control treatment for 24 h and the 0.05% colchicine treatment for 24 h (C 1 × D 1 and C 2 × D 1 ). The distance between guard cells was generally 1.9-2.8 µm. This result indicates stomatal enlargement by colchicine.

Pearson's correlation analysis and dendrogram cluster analysis
Polyploid induction demands changes in various parameters with evident interrelations. The Pearson's correlation matrix showed the extent of both positive and negative correlations among the 20 studied variables. Figure 10A shows negative and positive values indicated by red and blue circles.
Vacant cells indicate that there was no significant relationship among the variables at the 5% level of significance. This analysis revealed that most of the studied variables related to vegetative and reproductive morphology positively correlated with each other. However, stomatal attributes such as SPM and SD were negatively correlated with SL. Also, reproductive variables, namely FFM, NNM, PL (cm) and PW (cm), had negative correlations. PIE was negatively correlated with FFM, LL, PH, PLC, and SD (at p < 0.05). In Fig. 10A, larger circles denote stronger correlations between the two respective variables. A heatmap with a dendrogram cluster was prepared using the 20 studied dependent variables, and two major clusters were identified (10B). Cluster one included SL, PIE, GCD and MF. Several small clusters were also identified, including one with SPM and GS. The multiple clusters in one major cluster indicated that there was variation among the variables even though they were correlated with each other to some extent.

Principal component analysis (PCA)
The data presented in Table 3 shows that the first two principal components (PC) explained 59% of the variation among plants. Individually, PC1 and PC2 explained 46.68% and 12.09% of the total variation. From the PCA (Table 4), it is evident that FFM, LL, PH, LW, PLC, SD, NNM, PL (cm), and PW (cm) had the highest positive loadings on PC1. On the other hand, for PC2, only LL had a high positive loading. SL had the most negative loading value for PC1, and MF, PIE, and GCD also had negative loadings. The biplot of the PCA based on colchicine concentration illuminated that the 0.5% (C 4 ) concentration treatment resulted in a distinct cluster, including variables of SL, PIE, MF, GCD which had negative loading values (Fig. 11A). Also, the 0.1% treatment (C 3 ) was intercept to some extent with C 4 for the same variables of SL, PIE, MF and GCD than that of the other studied dependent variables. Therefore, it has been revealed that though SL, PIE, MF and GCD are affected by the 0.5% and 0.1% treatments but PIE and GCD are strongly influenced by 0.5% treatment. Meanwhile, the other two treatments (C 1 , C 2 ) did not affect SL, PIE, MF and GCD variables, that confirmed the highest influence of 0.5% treatment on PIE. Similarly, the 72 h duration (D 3 ) treatments resulted in a distinct cluster, with most parameters, including PIE, having negative loadings (Fig. 11B).

Discussion
The experiment tested two factors, colchicine concentration and duration, as these are the most important factors for polyploid induction through seed treatment with colchicine (Moore and Janick 1983;Hassan et al. 2020). Colchicine is effective only in the cell division stage; therefore, both colchicine concentration and duration of seed soaking can impact polyploid induction (Emsweller 1961). Morphological traits were recorded initially to identify putative polyploids; these parameters can be used to identify the ploidy level, however, they are not always completely reliable (Compton et al. 1996;Norman et al. 1995). Stomatal attributes can be used to identify polyploids more precisely (Sabzehzari et al. 2019).
At the germination stage, there were higher germination percentages and lower mortality percentages in the control treatments than in the colchicine treatments. This is likely due to the inhibitory effect of colchicine on living parts, causing seedling death (Zlesak et al. 2005). Likewise, definite variations were also found in morphological characteristics and reproductive biology (Figs. 5, 6, Table 2). Usually, tetraploid plants tend to be larger plants, with larger leaves, flowers, and pollen, compared to diploid plants. However, in this experiment the leaf size, plant height, and other plant organs remained stunted after three months of germination. This is likely due to the colchicine, which can slow down mitotic cell division (Stebbins 1971). Similarly, stunted growth was also reported in Trichosantes dioica Roxb. until four months after germination (Hassan et al. 2020). However, the hypocotyl was thicker in colchicine-treated seedlings than in control seedlings. These findings are concurrent with the findings of Hassan et al. (2020), which showed that organ thickness might occur because of a larger number of chromosomes, maintaining the balance between cytoplasm and nuclear volume, and higher protein accumulation, which is the result of higher gene expression due to a higher allelic number (Hassan et al. 2020).
Regarding reproductive biology, the petal number was higher in 0.5% colchicine-treated plants than in control plants.  (Ning et al. 2009). In the current study, the petal length and width of male flowers were low in the colchicine-treated plants; but previous studies reported that flowering parts are always larger in tetraploids of basil and myrtle (Omidbaigi et al. 2010;Zhang et al. 2009). On the other hand, the pollen diameter of was greater in colchicine-treated plants than in control plants, similar to results for Plantago psyllium (Calvalho et al. 2005), Allium sativum (Cheng et al. 2012), Arachis paraguariensis (Aina et al. 2012), and Ziziphus jujube (Cui et al. 2017). Flowering initiation of both male and female flowers occurred earlier under higher colchicine concentrations, while significant variations were recorded among the node numbers of the first flower position. In contrast, flowering time was not significantly affected by colchicine in Jatropha curcas (Niu et al. 2016). Thus, chromosome doubling might cause these kinds of variations in external morphology in several species (Zhang et al. 2019). Control plants and plants treated with 0.05 and 0.1% colchicine produced flowers (male and female). This may have been due to lower photosynthetic product production (Rekika et al. 2013). The stomatal length and diameter and the distance between two guard cells were greater in colchicine-treated plants than in control plants (diploids). As a result, the stomatal density was lower in colchicine-treated plants than in control plants. These findings support results reported for P. psyllium (Sabzehzari et al. 2019;Niu et al. 2016), which had similar features to those of plants in the current study. The chloroplast number was higher in the putative polyploids than in the diploids. The chlorophyll content may increase when the number of chloroplasts in stomatal guard cells increases; thus, most of the polyploid leaves had dark leaves . Most putative polyploids were detected under the highest colchicine concentration (0.5%) for 72 h. This result agrees with a previous report on T. dioica Roxb. (Hassan et al. 2020). Considering the duration effect more putative polyploids were detected after 24 h colchicine treatment than after 48 h colchicine treatment. This result is also similar to those of previous studies on Platanus acerifolia and Ocimum basilicum (Liu et al. 2007;Omidbaigi et al. 2010).
The present findings suggest that colchicine causes substantial morphological and anatomical changes and that significant correlations occur among the most important dependent variables. Suppression of initial growth and changes to anatomical and morphological attributes under colchicine treatment could be due to the effect of colchicine on spindle fiber formation (Hassan et al. 2020). Distinguishable reproductive traits, both in male and female flowers, were observed; however, no female flowers were produced in plants treated with the highest colchicine concentration (0.5%) for 72 h. Diploid and putative polyploids were characterized morphologically; however, final classification of putative polyploid induction efficiency (PIE) was based on the stomata structure analyses. The stomata with higher guard cell distances and lower densities were classified as putative polyploids. Likewise, the pollen diameter was noticeably greater in tetraploids than in diploids. Therefore, these morphological and anatomical traits could be used as an alternative to expensive flow cytometry (FCM) for the validation of polyploid induction through colchicine in watermelon.

Conclusion
Putative polyploid induction was highest after treatment with 0.5% colchicine for 72 h. Significant morphological differences were observed between diploids and putative polyploids, with putative polyploids having thicker hypocotyls, stunted growth with thinner stems and lower leaf areas. Regarding reproductive biology, under the 0.5% colchicine treatment, male flowers had six petals, and greater pollen diameters and viability compared to diploid flowers, confirming polyploid induction. The stomatal length, diameter and area and the distance between two guard cells were correlated with each other, particularly in putative polyploids. Further research should be carried out with various colchicine concentrations for successful polyploid induction and cytology analysis for validation. Colchicine-induced polyploid resources could be used for breeding of seedless watermelon fruit.