Even among orchids in the same genus, the scent compounds released by flowers differ greatly. Have these differences spread by chance over a long evolutionary history? Or do relationships with insects and other pollen carriers constrain the range of possible changes?
A study published by Artuso and colleagues in 2026 compared floral scents in 41 Malagasy species of Bulbophyllum. Beyond describing their chemical diversity, the researchers used relationships among the plants to examine how scent had changed, comparing several evolutionary models.
297 volatile compounds from flowers of 41 species
The study included 50 plant samples representing 41 species: 30 formally recognized species and 11 undescribed species that may be new to science. Forty-one samples came from plants cultivated in botanical gardens in Vienna or Salzburg, and nine were collected in the wild in Madagascar.
The researchers enclosed inflorescences in bags, collected scent released into the surrounding air, and analyzed it by gas chromatography–mass spectrometry. They also collected ambient-air and leaf samples as controls to distinguish compounds emitted by flowers.
They detected 297 volatile compounds in total. The number found in a single species ranged widely, from 1 to 80; 30 of the 41 species had fewer than 10. Bulbophyllum henrici, by contrast, yielded 80. A higher compound count does not rank how pleasant or strong a scent seems to people.
Of the 297 compounds, 225, or about 76%, were detected in just one of the species examined. This indicates species-specific features within this survey, not that those compounds occur nowhere else in the world.
Familiar compounds and a first report from floral scent
Combinations also varied. Some groups were rich in fatty-acid-derived compounds; some species featured terpenes such as linalool and limonene, while others contained abundant nitrogen-bearing compounds such as indole. Compositions differed among related groups, yet some species in different groups had similar combinations.
In Bulbophyllum cardiobulbum in particular, N,N-dimethylleucine O-methyl ester made up a large proportion. The researchers also used nuclear magnetic resonance to confirm its structure and described this as the first report of the compound in floral scent. It has been reported in other natural products, so this was not the first discovery of a previously unknown molecule.
Small flies have been observed visiting this orchid, but whether the compound attracts them remains untested. Its presence and its ecological role need to be investigated separately.
Moreover, only about half of all the compounds could be structurally identified. Many remain unnamed, leaving substantial room for further chemical investigation of orchid scents.
Comparing scent differences with evolutionary models
The evolutionary analysis used 32 species whose relationships were known. It was further restricted to 59 compounds shared by at least two of these 32 species, comparing their relative proportions within the overall scent. The researchers did not simply put all 297 compounds into the evolutionary models.
Three models were compared. In the first, changes accumulate randomly and differences increase over time. In the second, large changes occur early and subsequently slow down. The third includes a tendency to return toward a central value despite ongoing change.
The third is called the Ornstein–Uhlenbeck, or OU, model. It can represent evolution in which scent characteristics are unlikely to move far outside a certain range. It does not mean that plants set goals and deliberately improve their scent.
For the scent data combining the 59 compounds, the OU model fitted better than the other two. The central finding is that even large differences among species do not necessarily mean that change expands without limits.
Close relatives do not necessarily smell alike
A separate test of whether closely related species had more similar scents found no clear statistical evidence for that tendency. This did not, however, prove that evolutionary relationships have no influence whatsoever.
That test and the model comparison address different questions. Even without a strong relationship between relatedness and scent similarity, overall scent change might tend to remain within a certain range. The researchers considered the two results together as evidence that diversity and evolutionary constraints may coexist.
When compounds were analyzed individually, the OU model also fitted best for 41 of the 59. In some cases, however, differences from other models were small, so it cannot be concluded that every compound evolved in the same way.
The authors suggest that different compounds may change differently because their roles vary, for example in attracting insects or defending flowers against organisms that eat them. This also makes it meaningful to examine the histories of entire scents and individual molecules separately.
Relationships with small flies are an important hypothesis
Small flies are thought to be involved in pollination in this group of Malagasy orchids. The authors interpret the results as suggesting that attracting insects to flowers and guiding them to where pollen transfer occurs may constrain scent evolution. The model results are consistent with that hypothesis.
However, the study did not directly establish which insects prefer which compounds or whether those preferences led to species divergence. A good fit to the OU model alone cannot identify pollinators as the cause of the constraints.
Many species were represented by few samples, and cultivation conditions and sampling times differed between botanical gardens and the wild. These influences were not tested independently. How much within-species individual and environmental variation contributes to apparent differences among species remains another question.
The authors propose detailed studies of wild populations of closely related species, comparing insect behavior, flowering time, flower color and shape, and genetic differences together. Understanding scent’s role in species divergence requires investigation of actual flower–insect relationships alongside chemical analysis and modeling.
Summary
The 41 Malagasy orchid species had scents that differed greatly in both compound numbers and combinations. Meanwhile, the evolution of compounds shared by multiple species within the 32-species subset was better explained by a model favoring bounded change than by one in which differences expand through random change alone.
Rich diversity and evolutionary constraints can coexist. By reconsidering chemical individuality in relation to pollinators, defense, and a long evolutionary history, this study sought to understand the significance of the wide variety of floral scents.
References
Artuso, S., Gamisch, A., Sieder, A., Fuchs, R., Gindorf, S., Schubert, M., Dötterl, S., & Comes, H. P. (2026). Diverse floral scents in Malagasy Bulbophyllum evolve in a bounded fashion. Frontiers in Plant Science, 17, 1785538. DOI: 10.3389/fpls.2026.1785538 / Full text. Original article: © 2026 Artuso et al., CC BY 4.0.
