The morphological study was based on living plants in the forest, cultivated plants in the nursery of the Forest Research Institute Malaysia and specimens in the herbaria of AAU, BKF, C, FI, E, K, KEP, KLU, PSU, SAN, SAR and SING. A total of 372 herbarium specimens were examined in this study which includes 29 Scaphochlamys species and four Borneocola species (the types of another four Borneocola species were not yet deposited in the herbaria).The morphological characters examined in the study were habit; position of the rhizome, thickness and colour; height of leafy stem, its base (whether swollen to form a bulbous base); distance between leafy stems; characters of bladeless sheath such as colour, indumentum, number and length; ligule length, indumentum and shape; petiole length, indumentum, whether channelled or rounded in cross section; number of leaves per leafy stem; lamina colour on both surfaces, size, shape, venation, texture, indumentum, apex and base; length of the inflorescence and infructescence, arrangement of the floral bracts on the rachis, characters of floral bracts and bracteoles (colour, indumentum, texture, shape); size, colour and shape of calyx, floral tube, corolla lobes, labellum, staminodes, stamen, ovary.

Fresh leaves from the cultivated plants or silica-dried materials from plants collected in the field were used for genomic DNA extraction.

For the ITS, the genomic DNA was extracted using the DNeasy Plant Mini Kit (Qiagen, Valencia, California, USA) following the manufacturer’s protocol. Two primers, ITS 5P (5’-GGAAGGAGAAGTCGTAACAAGG-3’) and ITS 8P (5’-CACGCTTCTCCAGACTACA3’) (Moller and Cronk 1997) were used to amplify the ITS region during the polymerase chain reaction (PCR). The thermal cycle of PCR for the amplification of the ITS sequences is initial denaturation at 94°C for 2 minutes, 40 cycles of denaturation at 94°C for 30 seconds, primers annealing at 48°C for 2 minutes, an extension at 72°C for 45 seconds and final extension at 72°C for 7 minutes. The PCR products were then purified using MinElute Gel Extraction Kit (Qiagen, Valencia, California, USA).

For the matK, the protocols for DNA extraction, condition, purification and DNA sequencing were described previously by Takano and Nagamasu (2007). The PCR and sequencing primers for matK (cpDNA) were TA-240f (5’-GGGAAA GGATGGGGTCTCCCG-3’), TA-150r (5’-CTCAAGGAGTTTTGTGGTTC-3’), TA-470F (5’-CCCTCTCCCGTCCATATGGA-3’) (all three were designed in the present study), matK8 (Steele and Vilgalys 1994), m5r (Kress et al. 2002), matK8r (Ooi et al. 1995), trnK2621 (Liston and Kadereit 1995), TA-10F, TA-05R, TA-02F and TA-02R (all from Takano and Okada 2002).

Raw sequence data were assembled and edited manually using BioEdit software ver. 7.2.5 (Hall 1999). DNA sequences were aligned with the CLUSTALW 1.83 software package, with default settings and multiple alignments (Thompson et al. 1994). Alignments of the matK sequences of cpDNA and the ITS sequences of nrDNA were combined. Gaps were deleted.

A total of 100 individuals including 54 taxa of Scaphochlamys and allied species were used. The three Siphonochilus species were used as an outgroup (Kress et al. 2002). Materials, accession numbers for the sequences, vouchers and references to the literature are presented in Table 1 at the end of this paper. Three datasets which comprise ITS, matK and ITS+matK combined, each containing 82, 78, and 61 taxa, were constructed. These three datasets were analysed using three methods: maximum parsimony, maximum likelihood and Bayesian analysis. A maximum parsimony (MP) analysis was performed with MEGA 6 (Tamura et al. 2013). Heuristic searches were conducted with RANDOM addition, SPR branch swapping and MULPARS options. Support for each branch was estimated with a bootstrap analysis, with 1000 replications (Felsenstein 1985), in a heuristic search with RANDOM addition and TBR branch swapping. The maximum likelihood (ML), based on the Tamura-Nei model (Tamura and Nei 1993), was also determined with MEGA 6 (Tamura et al. 2013). Neighbor-Join and BioNJ algorithms were applied to a matrix of pairwise distances estimated with the maximum composite likelihood approach; then, the topology that had the best log likelihood value was selected. Bootstrap analysis under the MP criterion was conducted with “fast” stepwise, addition searches, with 1000 replicates. In addition, a Bayesian analysis was carried out with MrBayes software ver. 3.1.2 (Huelsenbeck and Rohnquist 2001; Rohnquist and Huelsenbeck 2003). The best fitting substitution model (the GTR+G model for nrDNA datasets, the GTR+G model for cpDNA datasets and the GTR+I+G model for cpDNA+nrDNA datasets) was selected for Bayesian analysis based on a series of hierarchical likelihood ratio tests, implemented in MrModeltest software ver. 2.3 (Nylander 2004). The analysis was performed with the selected model and two simultaneous runs of two million generations with four chains, sampling every 100 generations. Each analysis reached stationarity (i.e. when the average standard deviation of split frequencies between runs was ≤ 0.01) well before the end of the run. Burn-in trees were discarded and the remaining trees and their parameters were saved. A 50% majority rule consensus tree was constructed. The results of the Bayesian analysis were reported as the posterior probability (PP; Huelsenbeck and Rohnquist 2001), which is equal to the percentage of phylogenetic trees sampled when a given clade was resolved. Only PP scores above 50% are shown.

The ITS datasets for 82 individuals with 29 taxa of Scaphochlamys and 6 taxa of Borneocola contained 786 characters after alignment, which decreased to 769 after gaps were deleted; 319 of these were parsimony-informative. Likelihood analysis resulted in a ML tree with –lnL = 10438.212. Parsimony analysis produced three parsimonious trees with 1865 steps, a consistency index (CI) of 0.391 and retention index (RI) of 0.609. The ML, MP and Bayesian trees had similar topology; the ML tree is shown with bootstrap (BS) and MP-BS, and Bayesian Posterior Probability (PP) support in Figure 1 below.

Figure 1. 

Molecular phylogenetic analysis of the ITS sequence data by the Maximum Likelihood method. Numbers above branches indicate bootstrap values of ML and MP and posterior probability of Bayesian Analysis.

Scaphochlamys formed a well supported clade (ML-BS/MP-BS/BA-PP support, 84/87/1.00). Each Myxochlamys and Borneocola consisted of a well supported subclade and became sisters to each other and they also became sistersto the Scaphochlamys clade. Distichochlamys species formed a well supported subclade and became sister to the Myxochlamys + Scaphochlamys + Borneocola clade (ML-BS/MP-BS/BA-PP 99/100/1.00).

The matK datasets for 78 individuals including 25 taxa of Scaphochlamys and 7 taxa of Borneocola contained 1,599 characters after alignment; 182 of these were parsimony-informative. Likelihood analysis resulted in a ML tree with -lnL = 5952.438. Parsimony analysis produced ten parsimonious trees with 557 steps, a consistency index (CI) of 0.613 and retention index (RI) of 0.080. The ML, MP and Bayesian trees had similar topology; the ML tree is shown with BS and MP-BS, PP support in Figure 2 below.

Figure 2. 

Molecular phylogenetic analysis of the matK sequence data by the Maximum Likelihood method. Numbers above branches indicate bootstrap values of ML and MP and posterior probability of Bayesian Analysis.

Each of the two Myxochlamys species and seven Borneocola species formed a strongly supported subclade and became sisters to each other. Scaphochlamys became sister to them, but bootstrap or probability support was weak. The Distichochlamys species formed a well supported subclade, but all the genera that belong to subfamily Zingiberoideae became sisters to Scaphochlamys + Myxochlamys + Borneocola clade and not only to Distichochlamys.

The combined ITS and matK datasets for 61 individuals including 13 taxa of Scaphochlamys and 6 taxa of Borneocola, resulted in 2,336 characters, 488 of these were parsimony-informative (Figure 3 below). Likelihood analysis resulted in a ML tree with lnL = 16671.531. Parsimony analysis produced the most parsimonious trees with 2247 steps, a CI of 0.440 and a RI of 0.635. The ML, MP strict consensus and Bayesian trees had almost the same topology; the ML tree is shown with MP-BS, ML-BS and BA/PP support in Figure 3.

Figure 3. 

Molecular phylogenetic analysis of the ITS+matK sequence data by the Maximum Likelihood method. Numbers above branches indicate bootstrap values of ML and MP and posterior probability of Bayesian Analysis.

Two Myxochlamys species and six Borneocola species formed a strongly supported subclade each and became sisters to each other. Scaphochlamys became sister to them and the bootstrap or posterior probability support was moderate. Distichochlamys species formed a well supported subclade and became sister to Scaphochlamys + Myxochlamys + Borneocola clade.

The Borneocola and Scaphochlamys species look similar in their vegetative morphologies. They are mostly small-sized gingers without the conspicuous pseudostem, with one to several leaves arranged spirally and tightly on a very short stem at the base. So far, all the Borneocola species examined are unifoliate. Similarly, most of the Scaphochlamys species also bear one leaf except for several species which have leafy shoots composed of multiple leaves, for example, S. grandis, S. lanceolata, S. kunstleri, S. malaccana and S. minutiflora. The basal part of the leaves is covered with a few bladeless sheaths which are rather different for both groups in terms of their texture and colour. For Scaphochlamys, the sheaths are coriaceous, green, green with a red tinge or red and mostly persistent until the end of flowering (Figure 4A, B). On the other hand, the sheaths of Borneocola are thinner in texture with a lighter shade of green or brown. The thin sheaths normally dry up early (Figure 4C) and sometimes they are completely shredded during the time of flowering.

Figure 4. 

Bladeless sheaths A Green and coriaceous in Scaphochlamys klossii (Peninsular Malaysia) B Red and coriaceous in Scaphochlamys abdullahii (Peninsular Malaysia) C Papery and marcescent in Borneocola calcicola (Sarawak). (Photographs by Y.Y. Sam).

The inflorescences of Borneocola and Scaphochlamys are terminal, stalked and consisted of few to many floral bracts. The differences lie in the characteristics of the floral bracts and flowers. Borneocola species have thin, translucent, early decaying and marcescent floral bracts. The colours of the bracts can be pink, pale brown, pale or light green (Figure 5A). On the contrary, the bracts of Scaphochlamys are coriaceous and sometimes hard in texture. They are usually green, green tinged red, red or reddish brown and remain fresh throughout the flowering (Figure 5B, C).

Figure 5. 

A–C Floral bracts A Green and coriaceous in Scaphochlamys klossii B Red and coriaceous in Scaphochlamys pusilla C Scarious and marcescent in Borneocola petiolatus D–F Variegation on labellum D White labellum with purple lines beside the median band in Scaphochlamys malaccana E White labellum with red streaks beside the band in Scaphochlamys concinna F Lilac labellum without coloured streaks beside the band in Borneocola petiolatus. (Photographs by Y.Y. Sam)

Besides the characteristics of the floral bracts, the variegation on the labellum can give a quick guide to the two genera. Most Scaphochlamys have white flowers with a yellow median band and lilac, purple, red streaks or patches flanking the band on the labellum (Figure 5D, E). However, there is no such variegation on the labellum of Borneocola (Figure 5F). The whole labellum of Borneocola is pale pink, lilac, violet or white with a light yellow or greenish yellow median band.

Both Borneocola and Scaphochlamys have a long slender floral tube which is mostly puberulent externally in Borneocola (except for B. calcicola) but glabrous for Scaphochlamys. Another marked difference observed is in the stigma shape. Scaphochlamys has a funnel-shaped or beak-like stigma (Figure 6A, B) while it is almost oblate with two dorsal knobs in Borneocola (Figure 6C).

Figure 6. 

Stigma A Funnel-shaped in Scaphochlamys endauensis B Beak-like in Scaphochlamys biloba C Capitate in Borneocola petiolatus. (Photographs by A & C N.M. Aidil, B Y.Y. Sam)

Scaphochlamys anomala (Hallier f.) R.J.Searle, Edinburgh J. Bot. 67: 85 (2010).

Kaempferia anomala Hallier f., Bull. Herb. Boissier 6: 357 (1898). Gastrochilus anomalum (Hallier f.) K.Schum. in Engler, Pflanzenr. IV, 46 (Heft 20): 92 (1904). Boesenbergia anomala (Hallier f.) Schltr., Repert. Spec. Nov. Regni Veg. 12: 315 (1913).

Gastrochilus hallieri (Hallier f.) Ridl., J. Straits Branch Roy. Asiat. Soc. 32: 109 (1899), nom. illegit.

Type. INDONESIA. Cultivated in Bogor, originally from Liang Gagang, Kalimantan Borneo, Hallier s.n. (original material: BO, specimen lost; lectotype (designated by Searle, 2010) Figure drawn from original Hallier’s material and published as t. IX, fig. 3, Bull. Herb. Boissier 6: 357 (1898).

Notes. The type, the only specimen ever collected, was lost. However, Hallier (1898) gave a very detailed description and drawing of the plant and this has convinced Searle (2010) to place it in the genus Scaphochlamys. The drawing, which is based on the type specimen and designated by Searle as the lectotype, is the only material that gives a glimpse of the appearance of the species. In the drawing, the flower and spirally arranged floral bracts are typical of both Scaphochlamys and Borneocola. Until another specimen is collected and is available for close examination, we prefer to retain this imperfectly known species in Scaphochlamys.