06/04/2026
EP2: Let’s continue by taking a closer look at a phylogenetic tree from the study: Phylogenetic position of Vanda coelestis (Rchb.f.) Motes (Orchidaceae; Aeridinae), published in 2024 by Meng-Yao Zeng, Jin-Liao Chen, Ming-Zhong Huang, and Ming-He.
In Figure 1, you see the first phylogenetic tree from this study, using DNA sequence data from chloroplast DNA to assess evolutionary relationships. Figure 2 shows a second phylogenetic tree from the same study. This one combines DNA sequence data from both the cell nucleus and chloroplasts to evaluate how closely related the species are.
From these figures, it appears that only one orchid species in this study was represented by more than one sample for chloroplast DNA analysis. This suggests that the multiple samples were probably not used as a positive control, as seen in other studies. In contrast, Figure 2 does not include any species or varieties represented by more than one sample. Instead, it seems that species within the same section, indicated by the same text color, may have been used as a kind of positive control to help validate the results.
When we look more closely at Figure 3, a portion of the figure 1, we see that V. coerulescens is grouped together with V. concolor and V. subconcolor of the section Vanda. V. coerulescens also appears fairly close to a cluster that includes V. brunnea, Rhynchostylis coelestis and V. coerulea. Even this alone raises concerns, suggesting that the study’s research methodology may not be robust or fully reliable from a scientific standpoint.
And when we turn to Figure 4, a portion of the figure 2, we find that…
Set 1: V. coerulescens appears to be closely related to V. bensonii and V. liouvillei,–members of the section Testacea. This is unexpected when you consider their overall plant shape, leaf structure, flower structure, and inflorescence structure, which are quite different. On the other hand, the finding that V. coerulescens is closely related to V. lilacina, unlike the result in the EP1, actually fits well with their overall similarities in overall plant shape, leaf structure, flower structure, and inflorescence structure.
Note: In my opinion, V. coerulescens and V. lilacina should be closely related and likely belong in the same section. They differ only slightly in structure and share overlapping natural habitats. In contrast, grouping V. coerulea together with V. coerulescens in the same section based mainly on flower color and inflorescence length seems inappropriate. Flower color is a trait that can mutate or vary easily, while inflorescence length depends heavily on climate and the plant’s overall health, factors that taxonomists working mainly with herbarium specimens may not fully aware.
Set 2: Rhynchostylis coelestis appears closely related to V. flabellate, a relationship I’ll come back to later. More striking, however, is that V. miniata (from Section Ascocentrum) clusters with a mixed group that includes V. bensonii and V. liouvillei ( section Testacea), V. coerulescens, V. lilacina, and Rhynchostylis coelestis. This is puzzling, since these species not only differ greatly in overall plant shape, leaf structure, flower structure, and inflorescence structure, but also occupy habitats that do not overlap.
Set 3: V. coerulea is shown as closely related to species like. brunnea and V. denisoniana (both in Section Obtusiloba), which seems unlikely given their clear overall plant, leaf, flower and inflorescence structural differences. This set also places V. parviflora (Section Testacea), which has a plant form and leaves resembling V. curvifolia, within the same evolutionary group as V. concolor and V. subconcolor (Section Vanda), whose overall appearance is more similar to the Obtusiloba-type orchids
Set 4: V. curvifolia and V. tessellata appear embedded within a group containing V. flavobrunnea, V. alpina and V. cristata (Section Cristatae), despite having very different overall plant, leaf, flower and inflorescence structures.
Set 5: V. ampullacea and V. christensoniana, two of four members of Section Ascocentrum, are shown as closely related to species in Section Neofinetia. This is also difficult to reconcile, given their major differences in overall plant, leaf, flower and inflorescence structures.
From all the points discussed above, it’s clear that the study’s research methodology has significant weaknesses. The results do not match what we observe from plant morphology, nor do they align with the species’ natural habitats and geographic distribution.
The current orchid DNA analysis and phylogenetic tree construction, mostly employed a few sample per taxon that do not represent the full diversity or habitat range and ignored the difference in natural mutation frequency among genes, should not be used as the main criteria for reclassification. Morphological traits should remain the primary basis for field identifying and classifying orchid species and genera. DNA phylogenetic tree should be used only as preliminary supporting evidence.
Because of this, selectively using parts of these findings to revise classifications carries a high risk of error. For example, the study’s claim that Rhynchostylis coelestis is closely related to V. flabellata is not convincing and should not be considered reliable.
