Plant material and plant identification

Field studies and plant sample collections were carried out throughout Thailand from April 2014 to April 2024. A large collection of 88 accessions covering 14 recognized species, one variety, and one undescribed species belonging to Kaempferia subg. Protanthium were included in the present study (Table 1), representing 14 of the 15 recognized species worldwide. An accession number was assigned to refer to the geographical location (district and province names) of the individual population sampled. Two accessions of K. rotunda (accession NNSB600–1 and NNSB600–2) were obtained from the Chatuchak plant market in Bangkok, although these plants were initially introduced from their natural habitat in S Laos. One cultivated accession of K. rotunda (accession NNSB602) was collected from a private preservation area of the fifth author. Rhizomes of all samples were collected when the plants were blooming. During the present study, the living plant samples have been maintained at the nurseries of Mahidol University in Nakhon Pathom, Thailand and the National Science Museum Arboretum (NSM Arboretum) in Pathum Thani, Thailand. Some replicates of the living specimens have been grown at the Ginger collection nursery in Queen Sirikit Botanic Garden (QSBG), Chiang Mai, Thailand. The species identification was based on the identification key to species according to the taxonomic revision in the Flora of Thailand (Jenjittikul & al. 2023) and the most recent publication (Nopporncharoenkul & Jenjittikul 2024; Nopporncharoenkul & al. 2024), using both floral and vegetative morphological characters, phenological character and distribution information. Protologues and herbarium specimens of species in the subgenus held at BK, BKF, QBG, SING and SLR herbaria (for herbarium abbreviations see Thiers 2023+) and several online herbarium specimen databases, especially the Kew Herbarium Catalogue ( http://apps.kew.org/herbcat/navigator.do) and the Chinese Virtual Herbarium (CVH;  https://www.cvh.ac.cn/) were also extensively examined. An undescribed species is provisionally named as Kaempferia sp. A distribution map of the accessions included in the present study is displayed as Fig. 1. The flowers of all species investigated are shown in Fig. 2. The specimens were prepared and deposited in the BKF, QBG or SLR herbaria (see  Appendix S1 in Supplemental content online (wi.54.54201_Appendices_S1-S3.pdf)).

Mitotic chromosome study

Of the overall 88 accessions, 42 accessions representing 14 recognized species, one variety, and one undescribed species of Kaempferia subg. Protanthium were subjected to chromosome number investigation. The 2n chromosome number of each accession was analysed from at least 20 cells per plant, and three to five individual plants per accession. Mitotic chromosome preparation was performed using the enzymatic squash technique, according to the protocols of Mandáková & Lysak (2016) and Chow & al. (2020) with modifications. Actively growing root tips were excised from living plants, which were grown from rhizomes originally collected from their natural habitats and in cultivation. The root samples were immediately pre-treated with saturated para-dichlorobenzene solution at 4 °C for 16–18 hours in darkness. To stop all cellular activities and reactions, the samples were fixed in Farmer's fluid (3: 1 v/v of absolute ethanol: glacial acetic acid) at 4 °C for 10 min, then transferred and preserved in 70% ethanol at 4 °C until further use. To soften the root tips, the fixed roots were rinsed in a citrate buffer (4 mM citric acid monohydrate and 6 mM trisodium citrate dihydrate) at room temperature twice for 10 min, and then incubated in an enzyme mixture (citrate buffer with added 10% w/v of cellulase [Onozuka R10, Saint Louis, USA] and 12% v/v of pectinase [Sigma P-4716, Saint Louis, USA]) at 37 °C for 15–20 min. Afterward, each softened root tip was carefully rinsed with 45% acetic acid on a clean microscopic slide at room temperature twice for 2 min. The meristematic cells were gently separated in a drop of 45% acetic acid using dissecting needles and smeared in 2% w/v of aceto-orcein stain. Finally, the fine cell suspension was covered with a coverslip and tapped vertically with dissecting needles to squash the cells flat. The chromosomes were investigated at 1000× magnification under an Olympus CX23 light microscope (Tokyo, Japan). The spread chromosomes were photographed with an Olympus DP73 digital camera (Tokyo, Japan) attached to the microscope. The somatic chromosome number was determined from the well-spread chromosomes in metaphase cells.

Genome size estimation

Of the overall 88 accessions, 84 accessions (276 individual plants) representing 14 recognized species, one variety and one unidentified taxon were included in the present genome size analysis. In order to obtain the intraspecific variation of genome size, the samples analysed were obtained from at least three accessions (different geographical locations) per species, and one to five individual plants were subjected and analysed for each accession. Each individual plant was re-analysed three times on different days. However, data for species with a small population distributed in a restricted area, including Kaempferia caespitosa, K. grandifolia, K. sipraiana and Kaempferia sp., were obtained from only one to two populations (Table 1). Leaf samples were harvested from living plants and immediately used for analysis on the same day. The fresh young, unfolded leaves without diseases and pests were selected for analyses. In this study, Glycine max (L.) Merr. cv. Polanka (obtained from the Institute of Experimental Botany, Olomouc, Czech Republic, 2C = 2.5 pg: Doležel & al. 2007) and Musa serpentina Swangpol & Somana ‘SS&JS 246 clone’ (2C = 1.36 pg: Rotchanapreeda & al. 2016) were used as the reference standards (Moonkaew & al. 2020).

Genome size (2C value) was estimated using propidium iodide flow cytometry, according to the two-step protocol described by Doležel & al. (2007), with minor modifications. For nuclei extraction, the leaves of both sample and standard were concurrently chopped using a new sharp razor blade in a petri dish with 1 ml of fresh ice-cold Otto's nuclear-isolation buffer I (0.1 M citric acid and 0.5% Tween 20). The homogenate nuclei suspension was mixed by pipetting and then filtered through a 42-µm nylon mesh. The nuclei were pelleted by centrifugation at 3500 rpm for 5 min, and the supernatant was carefully removed. Afterward, the nuclear pellet was resuspended in 200 µl of ice-cold Otto's buffer I by gentle shaking. Thereafter, 400 µl of Otto II solution (0.4 M disodium hydrogenphosphate) supplemented with 50 µg/ml of propidium iodide (PI), 50 µg/ml of RNase A and 2 µl/ml of β-mercaptoethanol was applied to each sample tube with nuclei suspension in Otto's buffer I. The nuclei suspension was subsequently incubated at room temperature for 30 min in the dark. Each sample was analysed using the BD FACSCalibur Flow Cytometer (BD Biosciences, California, USA). Histograms of the relation between PI fluorescence intensity (PI-A, X-axis) and number of nuclei (event, Y-axis) were generated and 2C peaks of sample and standard were gated with a coefficient of variation lower than 3% using the BD FACSDiva version 6.1.1 software (BD Biosciences, California, USA). The estimated genome size was calculated using the linear relationship between the fluorescent intensity from stained nuclei of sample and internal standard, according to the following formula: Genome size of sample (pg) = (sample G0/G1 mean peak/reference standard G0/G1 mean peak) × Standard genome size (pg). In addition, the putative ploidy levels of the accessions which were excluded in chromosome analysis were inferred based on comparison of the genome sizes to those of the accessions obtained with both genome sizes and chromosome counts.

Fig. 2.

Floral morphology of Kaempferia subg. Protanthium in Thailand – A: K. albiflora (NNSB-546); B: K. aurora (NNSB-713); C: K. caespitosa (NNSB-733); D: K. calcicola (NNSB-903); E: K. graminifolia (NNSB-686); F: K. grandifolia (NNSB-519); G: K. jenjittikuliae (NNSB-760); H: K. lopburiensis (NNSB-541); I: K. noctiflora var. noctiflora (NNSB-554); J: K. noctiflora var. thepthepae (NNSB-928); K: K. rotunda (NNSB-534); L: K. simaoensis (NNSB-676); M: K. sipraiana (NNSB-656); N: K. subglobosa (NNSB-749); O: K. takensis (NNSB-697); P: K. udonensis (NNSB-508). – All photographs by N. Nopporncharoenkul.

Species description

The morphological and phenological characters of an undescribed species were investigated, measured, photographed and described from the living specimens in its natural habitat and from material cultivated in the nursery of NSM Arboretum. The morphological terminology used in the species description followed Beentje (2016). The diagnostic characters were discussed in relation to the morphologically closest similar species. The conservation status was assessed following the guidance to the IUCN Red List Categories and Criteria, version 15.1 (IUCN Standards and Petitions Subcommittee 2022). The extent of occurrence (EOO) and area of occupancy (AOO) were calculated using GeoCAT (Bachman & al. 2011). Type specimens with duplicates were prepared and will be deposited in the BK, BKF, QBG, SLR and SING herbaria.

Statistical analysis

The genome size data were analysed using descriptive statistics (mean ± standard deviation), Kolmogorov-Smirnov normality test (K-S test), and non-parametric statistical test, Kruskal-Wallis ANOVA, with the software IBM SPSS Statistics for Windows Version 21.0 (IBM Corp., New York, USA) (Ostertagová & al. 2014). Pairwise comparisons of species were also conducted using Kruskal-Wallis one-way ANOVA test at a significant level of p-value < 0.05 to test the difference in genome size among the species. In addition, box and dot plots of genome size variation in Kaempferia subg. Protanthium were also created using IBM SPSS Statistics version 21.0 (Spriestersbach & al. 2009; Sen & Yildirim 2022).

Chromosome numbers

The mitotic chromosomes of all 14 recognized and one undescribed species belonging to Kaempferia subg. Protanthium from Thailand were successfully investigated. The 2n chromosome numbers were found to be 22, 33, and 44 (Table 1; Fig. 3). The chromosome results unequivocally clarified x = 11 as the base chromosome number of K. subg. Protanthium. All species investigated, including Kaempferia sp., were diploid (2n = 2x = 22) or included diploid accessions. Two species, K. rotunda and K. takensis, provided included with different ploidy levels. Kaempferia rotunda included diploid (Fig. 3K), triploid (2n = 3x = 33: Fig. 3L) and tetraploid (2n = 4x = 44: Fig. 3M–N) accessions, whereas K. takensis included diploid (Fig. 3R) together with tetraploid (Fig. 3S) accessions. None of the species examined in the present chromosome study substantiated presence of aneuploid.

Genome size variation

In the present genome size study, we thoroughly examined 84 accessions of the species belonging to the Kaempferia subg. Protanthium collected from Thailand and Laos. The mean genome size with the standard deviation (S.D.) of each accession were summarized in Table 1, while those of individual plants analysed were reported in  Appendix S2 (wi.54.54201_Appendices_S1-S3.pdf). The putative ploidy levels of the accessions without chromosome number information were inferred based on comparison of the genome sizes with those of the accessions which were successfully clarified both genome sizes and 2n chromosome numbers. Among the diploid accessions, the genome sizes of the subgenus were found to range from 3.687 ± 0.052 pg in K. simaoensis accession NNSB676 to 6.412 ± 0.070 pg in K. albiflora accession NNSB741. Surprisingly, Kaempferia sp. had the highest mean genome size of 6.255 ± 0.097 pg, although it was not statistically different from that of K. albiflora (Table 1). However, it should be noted that genome size of Kaempferia sp. was analysed in only five individual plants from a single accession only. The range of genome sizes of all species was analysed and performed in the box and dot plots as shown in Fig. 4, 5, respectively.

After initially testing the normal distribution of genome sizes using the K-S test, the genome sizes of several species were not normally distributed (see  Appendix S3 (wi.54.54201_Appendices_S1-S3.pdf)). Therefore, a non-parametric statistical test using Kruskal-Wallis one-way ANOVA was performed to analyse the differences between genome sizes of the studied species. The statistical result indicated significant difference between means of genome sizes of the studied species. Moreover, statistical pairwise comparisons were also reported to categorized group between the analysed species based on the estimated genome sizes (Fig. 4;  Appendix S3 (wi.54.54201_Appendices_S1-S3.pdf)).

Regarding the species representing accessions with different ploidy levels, the genome sizes of 17 accessions of Kaempferia rotunda collected from different geographic localities were extensively examined. The results displayed as the box and dot plots clearly revealed the significant difference in three ranges of genome size (Fig. 4, 5A). Consequently, three ploidy level ranges were assigned: diploid (range 4.071–4.296 pg: 7 accessions), triploid (range 6.787–7.156 pg: 3 accessions) and tetraploid (range 8.165–9.172 pg: 7 accessions). The mean genome sizes of diploids, triploids and tetraploids were 4.193, 6.983 and 8.543 pg, respectively. The mean genome size values of triploid and tetraploid were 1.67 and 2.04 times the diploid mean value. The increase in genome size of the tetraploid K. rotunda is linearly proportionate to the increase in ploidy level, while the triploids had an approximately 11% larger than expected genome.

Kaempferia takensis is another species representing different ploidy levels among the accessions investigated, and the box and dot plot analyses revealed discontinuous ranges of genome sizes (Fig. 4, 5C). The diploids had a range of 4.579–5.100 pg (9 accessions) and tetraploids exhibited a range of 7.959–8.341 pg (4 accessions). The mean genome size of the tetraploids was up to 1.7-fold larger than that of a diploid. The increase in genome size of tetraploid K. takensis is not in linearly proportionate to the increase in ploidy level, but intermediate between 1.5 times the theoretical diploid size in triploids and 2 times in tetraploids.

Fig. 3.

Mitotic metaphase chromosomes of selected accessions of Kaempferia subg. Protanthium in Thailand – A: K. albiflora, 2n = 22 (NNSB-634); B: K. aurora, 2n = 22 (NNSB-713); C: K. caespitosa, 2n = 22 (NNSB-733); D: K. calcicola, 2n = 22 (NNSB-903); E: K. graminifolia, 2n = 22 (NNSB-684); F: K. grandifolia, 2n = 22 (NNSB-519); G: K. jenjittikuliae, 2n = 22 (NNSB-836); H: K. lopburiensis, 2n = 22 (NNSB-335); I: K. noctiflora var. noctiflora, 2n = 22 (NNSB-640); J: K. noctiflora var. thepthepae, 2n = 22 (NNSB-928); K: K. rotunda, 2n = 22 (NNSB-567); L: K. rotunda, 2n = 33 (NNSB-629); M: K. rotunda, 2n = 44 (NNSB-642); N: K. rotunda, 2n = 44 (NNSB-703); O: K. simaoensis, 2n = 22 (NNSB-535); P: K. sipraiana, 2n = 22 (NNSB-656); Q: K. subglobosa, 2n = 22 (NNSB-749); R: K. takensis, 2n = 22 (NNSB-697); S: K. takensis, 2n = 44 (NNSB-524); T: K. udonensis, 2n = 22 (NNSB-752). – Scale bars: A–T = 10 µm. – Photographs: A–C, E–I, K–T by N. Nopporncharoenkul; D, J by W. Sukseansri.

Identification of undescribed species

An undescribed species, Kaempferia sp., was found on the top of limestone hills and cliffs in Khon Kaen province, NE Thailand (Fig. 6). After intensive morphological study and comparison with protologues and herbarium specimens were conducted, we could not taxonomically identify this plant with any of the existing species belonging to K. subg. Protanthium. However, an undescribed species can be recognized and differentiated from the others by the combination of the following diagnostic characters: (1) well-developed pseudostems above ground, (2) elliptic, elliptic-oblong to lanceolate-oblong leaf blade with a long petiole, (3) flat type floral plane, consisting of horizontal to slightly arcuate lateral staminodes and labellum, which laid in the same plane and parallel to the ground, (4) bilobed labellum with an incision c. 3/5 of its length, (5) an involute labellum base, loosely enclosing the anther, (6) a subsessile stamen with an extremely short filament and 3–4(–5) mm long anther thecae, (7) an obreniform, broadly ovate, obovate to obdeltoid anther crest with an irregular undulate to crenate apex and (8) the anther crest extends backward and positioned nearly perpendicular (c. 90 degree) to the anther connective.

Taxonomy of Kaempferia subg. Protanthium in Thailand

Thailand is regarded as the biodiversity hotspot of the genus Kaempferia (Jenjittikul & al. 2023). The recent taxonomic studies revealed 14 accepted species and one variety belonging to K. subg. Protanthium distributed throughout Thailand (except the peninsular region), including 10 strictly endemic taxa (Nopporncharoenkul & Jenjittikul 2024; Nopporncharoenkul & al. 2024). However, the taxonomic circumscription of several recognized and unidentified taxa still remains unclear, leading to the problems with delimitation of Kaempferia species. Taxonomically, the presence of both inflorescences and leafy shoots is extremely necessary for accurate species-specific identification in the genus (Sirirugsa 1989). However, the reproductive and vegetative parts of the plants in K. subg. Protanthium cannot be observed at the same time (see introduction). Although the flowering period of the subgenus is generally from March to June, it is very short with only two to three weeks observed in each population. In addition, after the end of the flowering period, growth of the leafy shoot mostly coincides with the beginning of the rainy season, and the shoot dies off and goes into dormancy for several months during the dry season. Moreover, the vegetative part of Kaempferia is highly variable within the species, especially in the length of the petiole, leaf blade size and shape, the presence of the variegated patterns on the adaxial side of leaf blade, and the presence of indumentum. Furthermore, several species of K. subg. Protanthium imply morphological similarities of vegetative and reproductive parts among the species, even with other genera in Zingiberaceae, particularly Boesenbergia Kuntze and Curcuma L. (Larsen & Larsen 2006; Techaprasan & al. 2010). For example, the leafy shoot (up to 80 cm tall) of K. simaoensis, consisting of a prominent, well-developed pseudostem, long petioles (up to 15 cm long) and lanceolate, elliptic to ovate leaf blades (up to 50 by 30 cm), usually slightly plicate and sometimes with a red to purplish red patch along the midrib adaxially (Jenjittikul & al. 2023), is morphologically similar to that of several species of Boesenbergia and Curcuma. Regarding within K. subg. Protanthium, K. simaoensis collected in Thailand had previously been recognized as a variation of K. rotunda (referred to K. rotunda accession TT15732 and TT16426, Techaprasan & al. 2010) owing to sharing the similarity in both leafy shoot and inflorescence characters. The morphological diversifications distinguishing between the species can be found in the anther crest and the presence of two prominent yellow bands on the labellum base toward incision. However, the taxonomic status of K. simaoensis was clarified and subsequently recognized as another species based on molecular phylogenetic analysis of ITS2 sequences (Nopporncharoenkul & al. 2016). According to the ambiguity in morphological and phenological variations, it is extremely difficult to identify or differentiate the species based on the investigation of morphological characters alone, especially in the absence of inflorescences.

The predominantly floral morphological characteristics for accurate identification of the species in Kaempferia subg. Protanthium include the floral plane type, the colouration and incision depth of the labellum, the length of the filament, and anther crest shape and size. Regarding the floral plane, two main types are classified for the genus Kaempferia, namely perpendicular type and horizontal (flat) type (Nopporncharoenkul & al. 2021). The flowers representing the perpendicular type are characterized by having upright to slightly arcuate lateral staminodes and a deflexed in distal half labellum. The labellum base is often flat and an incision is around or less than 1/2 of the labellum length. A filament is noticeable and flat. On the other hand, the flowers with the flat floral plane type (T shape formed) are characterized by having horizontal to slightly arcuate lateral staminodes and labellum, which are arranged in the same plane and usually parallel to the ground. The labellum base is conspicuously involute, tightly enclosing the anther connective and thecae. A labellum incision is around or more than 2/3 of its length. The stamen of the flat floral type is mostly subsessile with an extremely short filament (Nopporncharoenkul & al. 2021). Additionally, the presence of glandular hairs on the filament and anther connective is another floral characteristic, supporting taxonomic differentiation among the species sharing the close similarity in the flowers (Nopporncharoenkul & al. 2024). Also, anthesis time, referred to the period during which a flower is fully open and functional, is a phenological characteristic classifying Kaempferia into two distinct groups (Nopporncharoenkul & Jenjittikul 2017; Jenjittikul & Ruchisansakun 2020): nyctanthous (nocturnal anthesis) and hemeranthous (diurnal anthesis). Although anthesis time is considered as a species-specific phenological character, the most recent study unveiled that K. noctiflora can produce both nocturnal and diurnal flowers (Nopporncharoenkul & al. 2024). Biogeographically, K. noctiflora var. noctiflora and K. noctiflora var. thepthepae are endemic to Chiang Mai province of N Thailand, but their populations are distributed allopatrically. However, both varieties can be differentiated based on anthesis time and the colouration on the labellum. Remarkably, K. noctiflora var. noctiflora produces the nocturnal (night-blooming) flowers with a white labellum having a pale yellow patch on the labellum base toward incision. Conversely, K. noctiflora var. thepthepae represents the diurnal flowers, which start to open in the morning (around 6 a.m.) and wither around noon. The labellum of K. noctiflora var. thepthepae is white to pale light purple labellum with the central white to cream white patch basally surrounded by two light purple stripes from base toward centre of lobes (Nopporncharoenkul & al. 2024).

Fig. 4.

Boxplot representing range of genome sizes (in picograms) of species of Kaempferia subg. Protanthium in Thailand based on 84 accessions (276 individual plants) from 15 species and one variety. Lines extending from boxes (whiskers) indicate variability outside upper and lower quartiles. Different letters above each box indicate statistically significant difference between means of genome sizes (p < 0.05). – N = number of accessions included in flow cytometry analysis. – 2x = diploid; 3x = triploid; 4x = tetraploid. Mean genome size with standard deviation data of individual plants follows  Appendix S2 (wi.54.54201_Appendices_S1-S3.pdf).

In this study, an undescribed species (Kaempferia sp.) was found in Khon Kaen province, NE Thailand. It was morphologically classified into the species producing the diurnal flowers with the flat type floral plane. The flowers consist of white to pale light pink lateral staminodes and a white to pale light pink labellum with the central white to cream white patch basally surrounded by two light pink to pale purple stripes from base toward centre of lobes, resembling those of K. lopburiensis and K. takensis (Picheansoonthon 2010; Boonma & al. 2020; Jenjittikul & al. 2023). The anther connective and filament are glabrous dorsally and laterally, as also observed in K. rotunda and K. takensis (Jenjittikul & al. 2023; Nopporncharoenkul & al. 2024). The anther crest is remarkable large, obreniform, broadly ovate, obovate to obdeltoid in shape with an irregular undulate to crenate apex, extending backward and positioned nearly perpendicular to the connective, which is similar to that of K. lopburiensis and K. udonensis (Picheansoonthon 2010; Phokham & al. 2013; Jenjittikul & al. 2023). Considering the leafy shoots, an undescribed species has a prominent, well-developed pseudostem, long petioles and elliptic, elliptic-oblong to lanceolate-oblong leaf blades that resembled those of K. rotunda and K. takensis (Boonma & al. 2020; Jenjittikul & al. 2023). As mentioned above, the species having the flat type floral plane represent the labellum with an incision around or more than 2/3 of its length. Interestingly, the labellum incision of an undescribed species is approximately 3/5 of the labellum length. In addition, the labellum base of an undescribed species is slightly involute and loosely encloses the anther connective and thecae. According to the morphological characters, we could not taxonomically identify it with any of the existing species, suggesting that it could be recognized as another species new to science. The morphological characters of an undescribed species are clearly compared and discussed below with the morphologically closest alliances, K. lopburiensis, K. rotunda and K. takensis and also shown in Table 2.

Chromosome number and genome size variation

In this study, we conducted extensive characterization of cytogenetic characters, including the 2n chromosome number and genome size, in order to better understand the species circumscription and support the taxonomic status of the species belonging to Kaempferia subg. Protanthium. The study included 88 accessions belonging to 14 recognized species, one variety, and one undescribed species from various regions throughout Thailand, except the peninsular part. Regarding the accessions analysed, 84 and 42 accessions were subjected to flow cytometry and chromosome investigation, respectively. The 2n chromosome numbers of the species in K. subg. Protanthium apart from K. grandifolia and K. rotunda were reported here for the first time, varying from 2n = 22, 33 to 44. In those species for which chromosome numbers had been determined previously, in the present cytogenetic analyses we obtained chromosome numbers in plant materials from other geographical locations. According to the mitotic chromosome results, the somatic chromosome 2n = 22 was observed in all species analysed, including K. grandifolia which is congruent with the number reported in the previous study (Saensouk & Jenjittikul 2001). Regarding K. rotunda, the 2n chromosome numbers revealed in the present mitotic analysis well agreed with those from the previous studies: 2n = 22 (Saensouk & al. 1999; Chandrmai & al. 2012; Nopporncharoenkul & al. 2017; Saensouk & Saensouk 2021b; Saensouk & al. 2023), 2n = 33 (Chakravorti 1948; Mahanty 1970; Eksomtramage & Boontum 1995; Nopporncharoenkul & al. 2017), and 2n = 44 (Ramachandran 1969; Omanakumari & Mathew 1984; Sadhu & al. 2016). However, we did not detect aneuploid number among 42 accessions included in the present chromosome analysis, but other previous studies published 2n = 30 and 2n = 54 from materials collected from Thailand (Saenprom & al. 2018) and India (Raghavan & Venkatasubban 1943), respectively. The unusual chromosome numbers may have originated from unbalanced gametes through irregular chromosome segregation during unequal meiotic division of the odd ploidy levels. However, some unbalanced gametes can take part in fertilization to produce aneuploid progeny (Wang & al. 2017). With the present and previous cytogenetic results, it is possible to hypothesize that the somatic chromosome numbers 2n = 30 and 54 have possibly arisen from the triploid (2n = 3x = 33) and pentaploid (2n = 5x = 55) ancestors with some chromosome eliminations, suggesting 3x - 3 and 5x - 1 respectively.

During the mitotic chromosome analysis, we did not measure the chromosome length because high fluctuation in chromosome size was observed from the cells of root tips collected from the individual plant. Variation in chromosome size may have resulted from the patterns of chromatin condensation, varying from the different responses of the meristematic tissue in each root material during the pretreated step (Pitaktharm & al. 2024). As the length of chromosome depends on several uncontrollable factors, we therefore extensively analysed the genome sizes of the samples, which reflect the correlation with the chromosome numbers and morphology and ploidy levels over evolutionary time, using flow cytometry. In addition, the putative ploidy levels of the accessions were inferred based on comparison of the genome sizes to those of the accessions obtained with both genome sizes and chromosome numbers. In the present study, we uncovered the genome sizes of the species of Kaempferia subg. Protanthium, except K. rotunda, for the first time. The combined results from mitotic and genome size analyses indicate that 74 accessions from all analysed species are putative diploid, whereas the others are putative polyploid, including putative triploid (three accessions) and putative tetraploid (11 accessions). The cytogenetic evidence clearly indicates that diploids with 2n = 2x = 22 predominate as the most common in K. subg. Protanthium, which is congruent with the previous cytogenetic analyses in K. subg. Kaempferia (Nopporncharoenkul & al. 2017). The genome sizes among the diploid accessions ranged from 3.687 ± 0.052 pg in K. simaoensis accession NNSB676 to 6.412 ± 0.070 pg in K. albiflora accession NNSB741. A 1.74-fold range in genome size was observed among the diploid species having 2n = 22 of K. subg. Protanthium analysed here.

Interestingly, high intraspecific variation in genome size among the diploid accessions was obtained from several species, especially in Kaempferia albiflora, K. lopburiensis, K. takensis and K. udonensis (Fig. 4). Regarding K. takensis, the genome sizes of the species were found to range from 4.579–5.100 pg with an approximately 11.4% variation. According to the geographic distribution, K. takensis could be classified into two populations: Chiang Rai province population and other provinces (type) population (Fig. 5F). The plants imply morphological overlap among the populations. Notably, the accessions belonging to the type population generally produce pale light pink flowers with two deep pink to light reddish spots at the centre of the labellum (see Fig. 2O), which is the same colour as the flowers of the plants in the type locality (Boonma & al. 2020). Although the plants collected from Chiang Rai province also produce pale pink to pale light purple flowers, but they have two purple to deep purple marks at the centre of the labellum which resemble the flowers of K. xiengkhouangensis in Laos (Phokham & al. 2013). Interestingly, the collecting sites of accessions in Chiang Rai province are the same latitude as the type locality of K. xiengkhouangensis, but they are c. 550 km apart. The morphological differences between K. takensis and K. xiengkhouangensis are found only in the length of petiole and the presence of variegated patterns on the leaf blade adaxially. Kaempferia takensis has the leaves with the prominent petioles (up to 5 cm long) and usually represents the variegated patterns on the leaf blade adaxially (Boonma & al. 2020; Jenjittikul & al. 2023), whereas K. xiengkhouangensis has sessile green leaves (Phokham & al. 2013). As the plants distributed in Chiang Rai province represent the petiolate leaves while K. xiengkhouangensis has not been reported in Thailand, we therefore provisionally identified the accessions collected from Chiang Rai province as K. takensis (referred to accessions NNSB531 and NNSB696, Table 1). Cytogenetically, the genome sizes (4.579–4.649 pg) of K. takensis accessions collected from Chiang Rai population were smaller, but not statistically significant, than those of the type population (4.742–5.100 pg) (Fig. 5E). With the present genome size and geographic distribution data it is possible to postulate that intraspecific variation in genome size between the geographically distant populations of K. takensis may have resulted from either (1) the difference in genome structure through the divergent evolutionary processes, such as mutations, natural selection, genetic drift, genetic hitchhiking and/ or gene flow, in each individual population or (2) the cryptic species may be included within the accessions in the present study. However, K. xiengkhouangensis from the type locality in Laos was not included in the present cytogenetic analyses. The species is therefore extremely necessary and will be subjected to further studies in order to clarify the species circumscription of the K. takensis complex.

Besides Kaempferia takensis, K. udonensis also provided high variation in genome size with an approximately 19.4% (range 4.057–4.844 pg). The species could be classified into two populations: a northeastern (NE) population and a southwestern (SW) population, based on the distinct collecting sites (Fig. 5H). Although the plants from both populations are distributed allopatrically, the plants do not only grow in the same habitat type of a mixed deciduous forest usually with bamboos, but also represent the same morphological characters. Interestingly, K. udonensis collected from NE Thailand displayed the larger genome sizes with no significant than those of the accessions from SW Thailand (Fig. 5G). According to intraspecific genome size variation observed in K. udonensis, we imply that the plants which are distributed in severely fragmented areas have been precluded opportunities for gene flow between genetically distant populations by geographic discontinuities, contributing to high genetic difference between the populations (Choudhuri 2014). Consequently, the reason can also explain for the occurrences of intraspecific variation in genome sizes of K. albiflora and K. lopburiensis as the populations have been observed in severely fragmented localities geographically.

Although the underlying evolutionary mechanisms involving intraspecific variation in genome size of genus Kaempferia remain unknown, we propose that it may be influenced by variation in heterochromatin levels and chromosome sizes via chromosomal rearrangements, duplications, deletions or translocations through retrotransposon or repetitive DNA element expansion, which play an important role in plant adaptation (Ortiz-Barrientos & al. 2016). Actually, variation in chromatin levels generally exists at the diploid level in several plants, for example, 1.7-fold in Cirsium Mill. (Bureš & al. 2004), 2.8-fold in Streptocarpus Lindl. (Möller 2018), and 4-fold in Lactuca L. (Doležalová & al. 2002) and Trifolium L. (Vižintin & Bohanec 2008). In addition, genome size variation within the ploidy level is also associated with evolutionary constraints on plant development, phenology or ecological performance (Vesely & al. 2012; Greilhuber & Leitch 2013). The recent study on the correlation between genome size and habitat type of the plants belonging to subfamily Zingiberoideae with dormancy period revealed that the species having small genome sizes tend to be more frequent in dry habitats since they enable faster growth, which is important especially at the beginning of rainy season. Conversely, the species which exist in shady habitats have significantly larger genome sizes than those occurring in full sun to partial shaded areas (Záveská & al. 2024). Furthermore, a sexual reproduction through seed production has been recognized as the mechanism for maintaining high genetic diversity within the species. In family Zingiberaceae, Záveská & al. (2011) revealed that diversity of Nei's gene in sexually-reproducing diploid Curcuma is significantly greater than in vegetatively-reproducing taxa. During the present study, we found all diploid accessions analysed were fully fertile as good seed sets have been observed in natural habitats and/or in cultivation. Consequently, the viable seed producing evidence is considered as one of the factors resulting intraspecific morphological and genome size variation in several Kaempferia species.

Polyploidy

Polyploidy plays a crucial role in plant evolution and speciation (De Storme & Mason 2014). In genus Kaempferia, polyploidy has been continually reported (Chakravorti 1948; Ramachandran 1969; Mahanty 1970; Omanakumari & Mathew 1984; Eksomtramage & Boontum 1995; Sadhu & al. 2016; Nopporncharoenkul & al. 2017; Saenprom & al. 2018; Záveská & al. 2024) while the first chromosome number evidence of polyploids, K. galanga (2n = 54) and K. rotunda (2n = 54) from India, was published by Raghavan & Venkatasubban (1943). A comprehensive chromosome number investigation of K. subg. Kaempferia from Thailand and Laos revealed ploidy variation within the species, ranged from diploid (2x), triploid (3x), tetraploid (4x) to pentaploid (5x) (Nopporncharoenkul & al. 2017). The present cytogenetic study of K. subg. Protanthium also unveiled polyploidies in two species, namely K. rotunda and K. takensis, whereas other species investigated were diploidy. Regarding K. rotunda, the investigated accessions can be classified into three ploidies based on the 2n chromosome numbers: diploid (2n = 22), triploid (2n = 33) and tetraploid (2n = 44), which are congruent with the number reported in the previous cytogenetic study (Nopporncharoenkul & al. 2017). However, we did not encounter any pentaploid K. rotunda, which was previously reported by Raghavan & Venkatasubban (1943) based on plant samples from India. The present genome sizes of K. rotunda are mostly consistent with those of the previous studies, although some discrepancies occur. In particular, the estimated genome sizes of diploid K. rotunda (4.071–4.296 pg) fit well into the previous diploid range of 3.468–4.43 pg reported by Chandrmai & al. (2012) and Basak & al. (2018), as well as the genome sizes of triploid K. rotunda (6.787–7.156 pg) which fully support the range (6.307–7.291 pg) revealed in the most recent study of Záveská & al. (2024). However, genome sizes of tetraploid K. rotunda accessions (8.165–9.172 pg) obtained in the present study were greater than those of tetraploid materials in the previous genome size report (7.45 pg, 2n = 44: Sadhu & al. 2016). Although the evolutionary causes involving in high intraspecific variation in genome size of K. rotunda are the topic of ongoing debate and still remain unclear, we propose that the difference in genome size at the same ploidy level (1) may result from disturbing effects of phenolic compounds from the leaf samples, such as deep purple pigments from the leaf blades abaxially (Jenjittikul & al. 2023) or secondary metabolites of plant materials with potential seasonal fluctuation (Walker & al. 2006); (2) may be influenced by the differences in measurements among different laboratories and protocols, and/or errors of instruments and methodologies (Doležel & al. 1998); (3) may potentially refer to chromosomal heterogeneity (aneuploidy) and/or variation in repetitive elements (non-coding regions) through evolutionary time of the plants distributed in geographic discontinuity as discussed before; or (4) may be mistaken from the taxonomic heterogeneity of plant materials analysed because K. rotunda provides a high degree of intraspecific morphological variation and also implies morphological overlap among Kaempferia species (Jenjittikul & al. 2023), possibly contributing to misidentification between K. rotunda and the cryptic species.

In triploid Kaempferia rotunda, three analysed accessions having the same chromosome number of 2n = 33 represented genome size ranged of 6.787–7.156 pg with c. 5.4% intraspecific variation. Interestingly, the triploid plants had an approximately 11% larger than expected genome due to genome increasing of 1.67-fold compared to the diploids. The genome size expansion in polyploids can be mostly explained by duplications of repetitive elements, such as heterochromatin, microsatellites and retrotransposon expansion, which are less likely to cause phenotypic changes (Blommaert 2020). Morphologically, the leafy shoots and inflorescences of both diploid and triploid K. rotunda are very similar to each other. Notably, the diploid K. rotunda generally produces a short, ovoid to subglobose rhizome with a single leafy shoot, whereas the triploid plants colonize via a clump of large moniliform-like rhizomes usually with several leafy shoots. Plausibly, genome expansion and intraspecific genome size variation in triploids may be related to a long-term cultivation, due to the plant improvement purposes, such as targeted selection of desirable external features and massive production of high chemical amounts in their rhizomes (Leong-Škorničková & al. 2007). Actually, the triploid K. rotunda has been cultivated commercially and commonly sold as traditional herb and attractive ornamental plant in plant markets throughout Thailand, referring to accession NNSB166 in Nopporncharoenkul & al. (2017) and accession NNSB602 in the present study originally obtained from the plant markets in Thailand. Moreover, the plants have been widely cultivated and used in several countries in S and SE Asia, which are concordant with the triploid materials obtained from India and Laos detected in the previous genome and/or chromosome studies (Chakravorti 1948; Mahanty 1970; Záveská & al. 2024). Unsurprisingly, no fertile seed was found in all triploid K. rotunda accessions during the present study, implying that triploid plants are not expected to be sexually fertile. Fundamentally, triploid K. rotunda having chromosome number 2n = 33 tends to generate unbalanced gametes owing to abnormalities in meiotic chromosome pairing. In Zingiberaceae, the irregularities in meiotic configuration comprising an assortment of univalents, bivalents and/or trivalents were observed in several previous meiotic studies of triploid Kaempferia (i.e. K. elegans, n = 11III: Nopporncharoenkul & al. 2017) and Curcuma (i.e. C. comosa Roxb., n = 21III and C. latifolia Roscoe, nearly regular synapsis: Puangpairote & al. 2015). The meiotic figure evidences most likely indicate that the triploid plants in family Zingiberaceae scarcely produce a fertile seed. However, they predominantly reproduce vegetatively by expansion and fragmentation of rhizomes. The producing of bigger rhizomes of triploid plants may indicate that they can store more nutrient, water and secondary metabolites, allowing more effective survival during a dormant period of dry season (Leong-Škorničková & al. 2007; Puangpairote & al. 2015).

Regarding tetraploid Kaempferia rotunda with chromosome number 2n = 44, the analysed accessions provided an approximately 12.3% variation in genome size (range 8.165–9.172 pg). The mean genome size was 2.04-fold compared the diploid mean, indicating that the increase in genome size is linearly proportionate to the increase in ploidy level. Accordingly, the present cytogenetic evidences imply that analysed plants are recent autotetraploids, displaying the complete whole genome duplication through polyploidization process, without genome downsizing observed (Möller 2018). The autotetraploid could be generated by three potential pathways (Ramsey & Schemske 1998): (1) the union of unreduced gametes (n = 2x) of diploid progenitors, (2) the union of reduced gametes (n = x) of diploids followed by chromosome doubling, and (3) the union of reduced and unreduced gametes to generate triploids (2n = 3x) and subsequently backcrossing to diploids or crossing to triploids. However, the pathway passing the triploid bridge seems to rarely occur in nature as the coexistence between triploid and tetraploid K. rotunda within the same populations was not encountered during the study. Geographically, tetraploid plants have been found only in N Thailand, whereas the diploids are widespread throughout SW and C Thailand (Fig. 5B). Consequently, it can be explained that after polyploidization, tetraploid K. rotunda has proceeded the physiological adaptation, survived in cooler and drier habitats, and distributed covering the entire areas in N Thailand. The adaptation in polyploid K. rotunda well agree with the previous reports that polyploid plants display a better adaptability to different ecological niches, increasing their chance for successful establishment through natural selection (Pelé & al. 2018; Van de Peer & al. 2021; Islam & al. 2022).

In tetraploid Kaempferia takensis, geographic distribution of tetraploid plants relates to distribution of the diploid populations, as discussed before (Fig. 5D). Accordingly, we classified tetraploid K. takensis into two populations: western and northern populations. Regarding the western population, the tetraploid K. takensis accession NNSB526–2 was collected from Kamphaeng Phet province which is the same locality as the diploid K. takensis accession NNSB526–1. The genome size of this tetraploid accession (range 8.315–8.362 pg, mean 8.331 ± 0.018 pg) was an approximately 1.7-fold compared the diploid K. takensis accessions from the type population (4.742–5.100 pg; Fig. 5F), indicating c. 15.3% genome downsizing. In the northern population, three accessions of tetraploid K. takensis were obtained from Phrae province of N Thailand. The genome size of tetraploid accessions from the northern population (range 7.883–8.387 pg, mean 8.152 ± 0.165 pg) was an approximately 1.77-fold compared the diploid accessions from the Chiang Rai population (4.579–4.649 pg; Fig. 5F), displaying an approximately 11.7% genome downsizing. Consequently, tetraploid accessions of K. takensis represent genome downsizing with 11.7–15.3% compared to their diploid relatives. During the study, all tetraploid accessions of K. takensis produce numerous viable seeds in both natural habitats and in cultivation, suggesting high fertility and productivity. Based on genome downsizing and viable seed setting evidences, tetraploid K. takensis could be postulated as either autotetraploid or allotetraploid which has been long evolutionary history and proceeded genome reorganization through the diploidization process. Genome downsizing in autopolyploids occurs rapidly at or in early polyploid generations after the polyploidization owing to genetic instability resulted from additivity of DNA amounts of the diploid progenitors (Eilam & al. 2010; Wang & al. 2021). In general, the recent autopolyploids usually show a high frequency of multivalent meiotic configuration and represent some degree of sterility, especially in unbalanced gametes (Parisod & al. 2010). After polyploidization, polyploid plants have proceeded genome reorganization through DNA elimination over evolutionary time until complete restoration of diploid-like behaviour, representing a high percentage of bivalents, via the diploidization process, contributing to genome downsizing in diploidized autopolyploids and allopolyploids (Eilam & al. 2010; Song & Chen 2015). Alternatively, massive DNA losses are more continuous process over evolutionary time, such as limiting the damaging activity of repetitive DNA (Wang & al. 2021). The effect of genome downsizing is well established in numerous species of flowering plants. For example, a decrease in gene number and a 10–25% reduction in genome size were observed in Triticum spp. (Feldman & al. 1997), Brassica napus L. (Gaeta & al. 2007) and Tragopogon L. (Buggs & al. 2012), while a remarkable genome downsizing of up to 44.4% was detected in polyploid Streptocarpus Lindl. (Möller 2018).

Fig. 5.

Dot plots showing intra- or interspecific variation in genome sizes (A, C, E, G, I, K) and distribution maps of plant samples (B, D, F, H, J, L) – A, B: intraspecific variation of Kaempferia rotunda: diploid (blue dots), triploid (red dots) and tetraploid (green dots); C, D: intraspecific variation of K. takensis: diploid (blue dots) and tetraploid (red dots); E, F: intraspecific variation of diploid K. takensis: Chiang Rai population (blue dots) and Type population (red dots); G, H: intraspecific variation of K. udonensis: SW Thailand population (blue dots) and NE Thailand population (red dots); I, J: intraspecific variation of K. noctiflora: K. noctiflora var. noctiflora (blue dots) and K. noctiflora var. thepthepae (red dots); K, L: interspecific variation of K. calcicola and morphologically similar species: K. rotunda (blue dots), K. takensis (red dots), K. lopburiensis (green dots) and K. calcicola (purple dots). – A, C, E, G, I, K: each dot represents mean genome size of individual plant based on re-analysis three times. – B, D, F, H, J, L: each dot represents collecting site of individual accession.

However, the underlying evolutionary mechanisms involving intraspecific variation in genome size and polyploidization of Kaempferia rotunda and K. takensis still remain unclear. Further studies of meiotic configuration, male gamete chromosome and karyotype are necessary, as well as accessing additional polyploid materials throughout Thailand. These efforts will unveil and fully understand the mechanisms of genome origin and evolution in polyploids.

Cytogenetic characters support species circumscription

In the present study, the combined data of 2n chromosome number and genome size for taxonomic purposes seems to be rather limited owing to much overlap between the species (Fig. 4). Moreover, the limitation in plant materials (only one to three individual plants analysed for each accession or species) might be inexact intraspecific genome size variation in some species, particularly in Kaempferia caespitosa (only three plants analysed) and K. sipraiana (only three plants analysed). However, we reveal the difference in the range of genome sizes, together with 2n chromosome numbers, can be used as supportive taxonomic markers for understanding species circumscription and elucidating differences between some species. Notable cases are observed in the species differentiation among K. rotunda and other closely morphologically similar species. The study of Nopporncharoenkul & Jenjittikul (2018) described K. graminifolia, distinguishing it from K. rotunda by having linear grass-like to narrowly lanceolate-oblong leaf blades, usually less than 5 cm wide. Moreover, the floral morphological diversification is found in the anther crest shape. Remarkable, K. graminifolia has broadly obdeltoid to broadly obovate anther crest with a bifid to crenate-bifid apex, whereas K. rotunda represents ovate-oblong anther crest with a bilobed apex, usually with 2–3 small teeth between lobes (Nopporncharoenkul & al. 2024). According to the results obtained from the present cytogenetic analyses, K. graminifolia is diploid with 2n = 22 and represents the genome size ranged 5.927–6.084 pg, which is significantly higher than that of K. rotunda at the diploid level (range 4.071–4.296 pg). The cytogenetic evidence clarifies and strongly supports the taxonomic status of K. graminifolia that it is not a morphologically variation of K. rotunda.

Besides Kaempferia graminifolia, K. aurora also shares the morphological characters of both leafy shoot and inflorescence with K. rotunda. The obvious differences are the presence of the anther crest with a tridentate to undulate-truncate apex and 6–12 mm long epigynous glands in K. aurora (Nopporncharoenkul & al. 2020b), whereas K. rotunda has the anther crest with a bilobed apex and shorter epigynous glands (2–5 mm long). As mentioned previously in Nopporncharoenkul & al. (2020b), we applied the genome size data as a taxonomic marker for supporting the differentiation of K. aurora from K. rotunda. In this study, the genome size range of K. aurora (5.205–5.402 pg) with 2n = 22 is significantly higher than that measured in the diploid K. rotunda, supporting the taxonomic status and species circumscription of K. aurora.

Regarding Kaempferia grandifolia, it is endemic species to the area surrounded by the Phu Wiang mountains of Khon Kaen province, NE Thailand (Jenjittikul & al. 2023). Due to the previous molecular phylogenetic analyses of Techaprasan & al. (2010) and Nopporncharoenkul & al. (2016), the accessions of K. grandifolia were clustered among the K. rotunda accessions, suggesting both species are phylogenetically closely related. However, K. grandifolia and K. rotunda show morphological diversifications in both leafy shoot and flower characters. Apparently, K. grandifolia has orbicular, suborbicular to ovate leaf blades, often appressed to the ground, and produces nocturnal flowers with the flat type floral plane (Saensouk & Jenjittikul 2001). Conversely, K. rotunda has upright lanceolate-oblong, elliptic to ovate leaf blades and represents diurnal flowers with the perpendicular floral plane type (Jenjittikul & al. 2023). Our cytogenetic results unveil that the genome size (range 5.634–5.731 pg) of K. grandifolia is larger than that of diploid K. rotunda (range 4.071–4.296 pg), well supporting the taxonomic status of K. grandifolia which was differentiated from K. rotunda (Saensouk & Jenjittikul 2001).

As discussed previously, based on morphological and phenological characters, Kaempferia noctiflora is taxonomic classified into two varieties, namely K. noctiflora var. noctiflora and K. noctiflora var. thepthepae (Nopporncharoenkul & al. 2024). The present mitotic chromosome investigation uncovers that both varieties are diploid with 2n = 22. Moreover, the range of genome size of K. noctiflora var. noctiflora (4.616–4.777 pg) mostly overlaps with that observed in K. noctiflora var. thepthepae (4.625–4.869 pg) (Fig. 5I), possibly implying the closely related genomes. Consequently, these cytogenetic evidences support the taxonomic status of K. noctiflora var. thepthepae that it cannot be recognized as an individual species differentiating from K. noctiflora despite the morphological and phenological diversifications.

Interestingly, another notable case using cytogenetic characters for supporting the taxonomic status is found in an undescribed species. Kaempferia sp. is a diploid species with the somatic chromosome number 2n = 22, which is the same as that obtained from the other diploid Kaempferia species (Nopporncharoenkul & al. 2017). Remarkably, the plants represent the highest mean genome size among the diploid accessions of the subgenus. Moreover, the range of genome size of this species (6.136–6.354 pg) does not overlap and is also higher than that of morphologically similar species: K. rotunda (4.071–4.296 pg), K. takensis (4.579–5.100 pg) and K. lopburiensis (4.884–5.445 pg) (Fig. 5K), supporting that it belongs to another species. On the basis of these findings, morphological characters (as discussed previously) together with cytogenetic evidence unequivocally clarify the taxonomic status of an undescribed species that it deserves recognition as a species new to science, which is taxonomically described below as K. calcicola Noppornch.

The present study emphasizes that characterization of cytogenetic characters, 2n chromosome number and genome size, is not crucial only in species discrimination between morphologically similar species but can also support the taxonomic description of new species. However, further studies focusing on integration of karyotype, genome size and molecular systematic analyses will be conducted, as well as accessing additional plant materials and populations covering geographic distribution throughout Asia in order to clearly understand the mechanisms involving in chromosomal and genome evolution and relationship of the species within Kaempferia subg. Protanthium.