Revision on the Genus Paris in Thailand, with a New Species Paris siamensis

The genus Paris is an important and confusing taxon due to high variation within species, and differences between species are sometimes difficult to delimit. Thus, the status of some taxa has changed over time. To clarify the status of Paris species for plant conservation and effective management of this genus in Thailand, we performed an intensive survey in northern Thailand, studied morphological characteristics, and constructed a molecular phylogenic tree, which we compared to recently published results of this genus. Our results indicate that there are two species in Thailand: P. yunnanensis and a new species, P. siamensis. Detailed descriptions, illustrations, and the phylogenetic position of these two species are provided here.


Introduction
Plants in the genus Paris L. [1] are important Chinese medicinal plants [2]. In recent years, these plants have been collected in Thailand and exported to China, causing them to become extremely threatened [2]. For the conservation and effective management of this genus, intensive taxonomic studies are required. The morphological characteristics of this genus have been reported to be distinct from other genera [2]. However, broad variation within species (e.g., in the shape and size of each plant part) has been observed, resulting in difficulties when identifying plants at the species level.
In the flora of Thailand [3], only one species of the genus Paris is reported in Thailand: P. chinensis Franch. [4] (=Daiswa polyphylla var. chinensis (Franch.) M. N. Tamura. in the book). However, in the taxonomic revision of Paris, Ji [2] mentioned that two species were found in Thailand: P. chinensis and P. caobangensis [5]. From our revision using an intensive morphological and molecular examination of fresh specimens in Thailand, we found that P. yunnanensis was distributed in many provinces in Northern Thailand. However, populations in Nan province seemed to differ from other populations but were unable to be confidently identified as any of the species reported in the recent revision [2]. Hence, both morphological and molecular methods were applied to identify and confirm the status of Paris in northern Thailand [6][7][8].
Phenology. Flowering April-August. Fruiting July-February. Distribution. To date, endemic to Nan province, Thailand (may be found in Laos), 1050-1216 m alt.

Phylogenetic Analysis
Most samples of Paris in Thailand collected from Chiang Mai, Chiang Rai, Mae Hong Son, and Lampang provinces ( Figure 4) were classified in the same clade as P. yunnanensis with a low support value of 0.74 ( Figure 5). This clade was sister to P. yanchii with high support (PP = 0.98). All samples collected from Nan province, Thailand, were monophyletic with high support (PP = 1:ML-MP = 95). This clade was sister to P. liiana with high support (PP = 1:ML-MP = 95).

Phylogenetic Analysis
Most samples of Paris in Thailand collected from Chiang Mai, Chiang Rai, Mae Hong Son, and Lampang provinces ( Figure 4) were classified in the same clade as P. yunnanensis with a low support value of 0.74 ( Figure 5). This clade was sister to P. yanchii with high support (PP = 0.98). All samples collected from Nan province, Thailand, were monophyletic

ANOSIM Analysis
The results of the similarity analysis (ANOSIM) showed that morphological characters between P. yunnanensis and P. siamensis were significantly different in terms of both flowering state (R-value = 0.3461, p-value = 0.0087) and non-flowering state (R-value = 0.0988, p-value = 0.0442).

ANOSIM Analysis
The results of the similarity analysis (ANOSIM) showed that morphological characters between P. yunnanensis and P. siamensis were significantly different in terms of both flowering state (R-value = 0.3461, p-value = 0.0087) and non-flowering state (R-value = 0.0988, p-value = 0.0442).

Principal Component Analysis (PCA) of Flowering State
The Kaiser-Meyer-Olkin (KMO) measure was used to verify sampling adequacy for the analysis. The obtained KMO value for the flowering state was 0.639, indicating that the sample size was adequate for factor analysis. Analysis of Bartlett's test of sphericity (χ 2 (351)) was 2959.810 with a p-value < 0.001, suggesting that correlations between items were sufficiently large for PCA.
The first two Principle Component axes explained 40.35% of the variation ( Figure 6). PC1, which explained 12.40% of the variation, seemed to identify lamina width. PC2, which explained 27.95% of the variation, seemed to identify petiole length, lamina length, the number of sepals, sepal length, the number of petals, anther length, anther width, and ovary length. Axis PC2 was unable to separate the two species, while axis PC1 showed separation with some overlap. The values of P. siamensis along PC1 were greater than those of P. yunnanensis, indicating that P. siamensis generally differs from P. chinensis in terms of its greater lamina width. The Kaiser-Meyer-Olkin (KMO) measure was used to verify sampling adequacy for the analysis. The obtained KMO value for the flowering state was 0.639, indicating that the sample size was adequate for factor analysis. Analysis of Bartlett's test of sphericity (χ 2 (351)) was 2959.810 with a p-value < 0.001, suggesting that correlations between items were sufficiently large for PCA.
The first two Principle Component axes explained 40.35% of the variation ( Figure 6). PC1, which explained 12.40% of the variation, seemed to identify lamina width. PC2, which explained 27.95% of the variation, seemed to identify petiole length, lamina length, the number of sepals, sepal length, the number of petals, anther length, anther width, and ovary length. Axis PC2 was unable to separate the two species, while axis PC1 showed separation with some overlap. The values of P. siamensis along PC1 were greater than those of P. yunnanensis, indicating that P. siamensis generally differs from P. chinensis in terms of its greater lamina width.

Principal Component Analysis (PCA) of Non-Flowering State
The Kaiser-Meyer-Olkin (KMO) value for the non-flowering state was 0.697, which is greater than 0.5, indicating that the sample size was adequate for factor analysis. Bartlett's test of sphericity (χ 2 (21) = 344.429; p-value < 0.001) indicated that correlations between items were sufficiently large for PCA.
The first two Principle Component axes explained 78.84% of the variation (Figure 7). PC1, which explained 61.81% of the variation, seemed to identify stem length, tuber length, petiole length, lamina length, lamina width, and the number of leaves. In addition, PC2, which explained 17.03% of the variation, appeared to identify the leaf width/length ratio. Axis PC1 was unable to separate the two species, while axis PC2 divided these two species with little overlap. The values of P. siamensis along PC2 were less than those of P. yunnanensis, indicating that P. siamensis generally differed from P. yunnansis in terms of their smaller leaf width/length ratio.

Principal Component Analysis (PCA) of Non-Flowering State
The Kaiser-Meyer-Olkin (KMO) value for the non-flowering state was 0.697, which is greater than 0.5, indicating that the sample size was adequate for factor analysis. Bartlett's test of sphericity (χ 2 (21) = 344.429; p-value < 0.001) indicated that correlations between items were sufficiently large for PCA.
The first two Principle Component axes explained 78.84% of the variation (Figure 7). PC1, which explained 61.81% of the variation, seemed to identify stem length, tuber length, petiole length, lamina length, lamina width, and the number of leaves. In addition, PC2, which explained 17.03% of the variation, appeared to identify the leaf width/length ratio. Axis PC1 was unable to separate the two species, while axis PC2 divided these two species with little overlap. The values of P. siamensis along PC2 were less than those of P. yunnanensis, indicating that P. siamensis generally differed from P. yunnansis in terms of their smaller leaf width/length ratio.

Differences between P. yunnanensis and P. siamensis
The results from t-tests showed that these two species are different with respect to several characteristics in both the flowering and non-flowering stages (Figure 8). In the flowering stage, lamina width, lamina width/length ratio, sepal width/length ratio, etc., were significantly different between the two species. In the non-flowering stage, the lamina width/length ratio and stem length were also significantly different.

Differences between P. yunnanensis and P. siamensis
The results from t-tests showed that these two species are different with respect to several characteristics in both the flowering and non-flowering stages (Figure 8). In the flowering stage, lamina width, lamina width/length ratio, sepal width/length ratio, etc., were significantly different between the two species. In the non-flowering stage, the lamina width/length ratio and stem length were also significantly different.