# Source paper archive - Source title: Floral scent diversity and evolution in deceptive Cypripedium, and post-pollination effects on floral scent and its biosynthesis in C. henryi - Author(s): Chazalon, J.-B., Pfanzelt, S., Siadjeu, C., Hu, E. Y., Kotagal, S., Doetterl, S., Kadereit, G. - Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.04.760698v1 - PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.04.760698v1.full.pdf - PDF SHA-256: 6e717bebd71be783d01002cc9a76bb2e2ac002391fc6134749fef01b2693fc60 - Converter: pymupdf4llm 0.0.27 - Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version. --- [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1 ORIGINAL ARTICLE 2 **Floral scent diversity and evolution in deceptive** _**Cypripedium,**_ **and post-** 3 **pollination effects on floral scent and its biosynthesis in** _**C. henryi**_ 4 5 6 Jean-Baptiste Chazalon [1], Simon Pfanzelt [1], Christian Siadjeu [3], Emy Yue Hu [4], Suvrat Kotagal [4], 7 Stefan Dötterl [2] and Gudrun Kadereit [1,4] 8 9 1- Staatliche Naturwissenschaftliche Sammlungen Bayerns (SNSB): Botanische 10 Staatssammlung München und Botanischer Garten München-Nymphenburg, Munich, 11 Germany 12 2- Department of Environment & Biodiversity, Plant Ecology Lab, University of Salzburg, 13 Austria 14 3- Institut für Organismische und Molekulare Evolutionsbiologie (imoE), Faculty of Biology, 15 Johannes Gutenberg University of Mainz, Mainz, Germany 16 4- Prinzessin Therese von Bayern-Lehrstuhl für Systematik, Biodiversität & Evolution der 17 Pflanzen, Ludwig-Maximilians-Universität München, Munich, Germany 18 19 **Corresponding author:** 20 Jean-Baptiste Chazalon (chazalon@snsb.de) 21 22 23 24 25 26 27 28 1 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) ### 29 Abstract 1 **Background and Aims** Floral scent is a primary driver of pollinator attraction and 2 reproductive isolation in deceptive orchids. Yet, little is known about the variation and 3 evolution of scent in a phylogenetic framework and how pollination affects the biosynthesis 4 and emission of scents in these lineages. This study characterizes scent diversity across 5 _Cypripedium,_ a genus of deceptive orchids, examines phylogenetic constraints, and 6 investigates post-pollination effects on scent composition in one of the species. 7 **Methods** Floral volatile organic compounds (VOCs) were sampled from 31 _Cypripedium_ taxa 8 using dynamic headspace techniques, and analysed by gas chromatography coupled to mass 9 spectrometry (GC/MS). Uni- and multivariate methods were used to analyse intra- and 10 interspecific scent variation and test for phylogenetic signal in floral scent. Post-pollination 11 effects on scent composition were experimentally tested in _C. henryi_ using four pollination 12 treatments, combined with RNA-seq analyses of floral tissues to examine transcriptional 13 responses in VOC biosynthetic pathways. 14 **Key results** We detected 258 floral VOCs, revealing a high chemical diversity dominated by 15 terpenoids, aliphatics, and aromatic compounds. Floral scent composition was highly species 16 specific, and closely related taxa often shared only few compounds. Scent composition 17 differed among pollination groups, but these groups explained only a limited proportion of 18 overall scent variation. We detected a significant but weak phylogenetic signal, suggesting 19 limited evolutionary constraint in floral scent. Pollination had limited effect on _C. henryi’_ s 20 floral scent and volatile biosynthesis gene expression, despite a significant decrease in total 21 scent emission after pollinium deposition. Transcriptional changes predominantly occurred in 22 the petals rather than the labellum and staminode. 23 **Conclusion** Floral scent evolution in _Cypripedium_ is characterized by high chemical 24 diversity, and a weak phylogenetic constraint, consistent with variable selection in food 25 deceptive systems. The pollination study further suggests the absence of a strong, directed 26 post-pollination effect on the floral scent of _C. henryi._ 27 **Keywords:** _Cypripedium_ ; deceptive pollination; floral scent; volatile organic compounds 28 (VOCs); phylogenetic signal; linalool; post-pollination effects; transcriptome profiles; 29 differential gene expression; floral organs 2 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) ### 30 Introduction 31 Floral volatile organic compounds (VOCs) strongly influence plant reproductive success 32 (Ramya _et al.,_ 2020) and play a key role in both short- and long-distance pollinator attraction 33 (Knudsen 2002; Dötterl and Jürgens 2005; B. Song _et al.,_ 2015; Dötterl and Gershenzon 34 2023). More than 1700 floral VOCs have been identified to this date, with Orchidaceae being 35 the most extensively studied plant family (Knudsen _et al.,_ 2006; Dötterl and Gershenzon 36 2023). Through specialized olfactory receptors, insects can interpret floral volatiles as 37 informative chemical cues, guiding them toward specific resources ranging from nutritive 38 rewards like nectar and pollen, to non-nutritive benefits such as brood sites, fragrance 39 compounds or shelter. Over time, foraging insects can learn to associate specific volatile 40 profiles with the presence of food rewards, adapting their foraging behaviour to maximize 41 energetic gain (Haverkamp _et al.,_ 2018). Some plants have evolved to exploit these 42 established sensory biases and attract visitors through deceit, using olfactory and/or visual 43 cues without offering the expected reward most often associated with these signals. 44 Orchidaceae contains more deceptive taxa than any other plant family, with roughly one-third 45 of its species relying on some form of deceit, predominantly generalized food deception 46 followed by sexual deception, to achieve pollination (Jersáková _et al.,_ 2006). A prominent 47 example is the subfamily Cypripedioideae, commonly known as “slipper orchids”. Species of 48 this orchid lineage are characterized by a distinctive slipper-shaped labellum, which functions 49 as a one-way trap for pollinators. To exit the flower, trapped insects must climb out through 50 the back of the labellum, thereby touching first the stigma and subsequently the pollinia. 51 When the visitors are of appropriate size and shape, and carry a pollen smear on their thorax, 52 this process results in effective pollination (Cribb 1997). 53 _Cypripedium_ L. is a monophyletic genus within Cypripedioideae (J. Li _et al.,_ 2011) 54 comprising 54 recognized species (World Checklist of Selected Plant Families, accessed June 55 2026). This genus is a compelling system for studying scent evolution in deceptive plants as 56 its high diversity of floral VOCs is complemented by increasingly clarified phylogenetic 57 relationships. Cribb and Frosch (2012) recognized 13 sections within _Cypripedium_ (Figure 1), 58 all of which have been shown to be monophyletic in phylogenetic studies employing nuclear 59 and chloroplast markers (Fatihah _et al.,_ 2011; J. Li _et al.,_ 2011). The monophyly of these 60 sections, as well as the topological position of most of the sections, was further supported by 61 incorporating plastome data into a more extensive nuclear loci dataset (41 loci) (Liao _et al.,_ 62 2024). Lagou _et al._ (2024) applied a phylogenomic approach using 913 nuclear loci to resolve 3 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 63 remaining uncertainties in the phylogeny of _Cypripedium_ . Their results confirmed earlier 64 findings, including sectional monophyly, the position of sect. _Irapeana_ as the earliest 65 diverging lineage, and incongruence between the two traditionally recognized subsections 66 within sect. _Cypripedium_ and the two clades actually recovered. Despite the expanded dataset, 67 support along the backbone of the phylogeny remained low, which was attributed to 68 incomplete lineage sorting associated with rapid radiation events, and possible reticulation 69 (Lagou _et al.,_ 2024). 70 Research on the genus _Cypripedium_ has resulted in documented pollinator records for more 71 than half of the known species. Most species within the genus are pollinated by solitary bees, 72 which likely visit the flowers in search of edible rewards. A notable exception is _C._ 73 _fasciculatum_ Kellogg (sect. _Enantiopedilum_ ), which is pollinated exclusively by _Cinetus sp._ 74 wasps (Ferguson and Donham 1999). These parasitic wasps are likely attracted to the flower 75 while searching for their fungus gnat hosts. Beyond Hymenoptera, dipteran pollination 76 represents a significant evolutionary strategy within the genus. Sections _Trigonopedia_ and 77 _Sinopedilum_ have transitioned exclusively to myophily, exhibiting specialized morphological 78 and olfactory traits. These include mottled, dull-coloured foliage and floral volatiles that 79 mimic fermentation or decaying organic matter to lure specific fly families. Recorded 80 pollinators include hoverflies (Syrphidae), fruit flies (Drosophilidae), dung flies 81 (Scathophagidae), blowflies (Calliphoridae), and flat-footed flies (Platypezidae). Dipteran 82 pollination has also been documented in sect. _Obtusipetala_ (Vogt 1990; Bänziger _et al.,_ 2008; 83 Zheng _et al.,_ 2011), sect. _Subtropica_ (Jiang _et al.,_ 2020), _C. calceolus_ of sect. _Cypripedium_ 84 (Braunschmid _et al.,_ 2017), and has been proposed for _C. debile_ of sect. _Retinervia_ 85 (Stoutamire 1967; Cribb 1997). Pollinator assemblages may vary geographically: species with 86 broad distributions, such as _C. calceolus_ and _C. reginae_, exhibit a greater diversity of 87 pollinators. For example, _C. reginae_ has been observed being pollinated by bees in Ontario 88 (Canada) and Missouri (USA), but by hoverflies in Vermont (USA) (Guignard 1886; Vogt 89 1990; Edens-Meier _et al.,_ 2011). 90 The genus _Cypripedium_ is thought to rely primarily on generalized food deception as its 91 pollination strategy (Nilsson 1979; Pemberton 2013), displaying general food related cues to 92 attract naive foragers rather than mimicking a specific model species. However, alternative 93 deceptive mechanisms have been hypothesized for certain species of the genus **.** For instance, 94 _C. macranthos_ var. _rebunense_ Kudô may exhibit floral mimicry of _Pedicularis schistostegia_ 95 Vved., potentially exploiting pollinators that typically visit that nectar-rewarding species 96 (Sugiura 2002). Similarly, _C. tibeticum_ King ex Rolfe is believed to mimic nesting sites of 4 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 97 _Bombus_ queens, in addition to employing food deception as an attractant mechanism to target 98 the same visitors (P. Li _et al.,_ 2006). A rewarding mimicry system was previously described 99 for _C. subtropicum_ where hairy tufts on the labellum did not only act as visual lures but also 100 served as food rewards for visiting hoverflies, with a nutrient content comparable to the 101 honeydew secreted by aphids (Jiang _et al.,_ 2020). In recent years, guild mimicry has been 102 proposed in Orchidaceae as an intermediate pollination system between generalized food 103 deception and Batesian floral mimicry, where deceptive species do not imitate flowers of a 104 single specific species but rather a group of often distantly related co-flowering species 105 (Jersáková _et al.,_ 2016; Scaccabarozzi _et al.,_ 2018). Etl _et al._ (2025) proposed that _C._ 106 _calceolus_ employs a mixed pollination strategy, in which guild mimicry plays a significant 107 role. In addition to producing common floral volatiles and bee pheromones, the species 108 emitted more specialized floral compounds also present in the fragrances of multiple co 109 flowering species. 110 Although floral scent diversity in Cypripedioideae remains only moderately explored, 111 _Cypripedium_ as a genus is comparatively well studied. Early work focused on the European 112 species _C. calceolus_ and several North American taxa (Nilsson 1979; Bergstrom _et al.,_ 1992; 113 Barkman _et al.,_ 1997). Nilsson (1979) proposed that an acetate-rich floral fragrance, 114 comprising compounds such as octyl, decyl, and hexyl acetate, mimics bee volatiles to 115 function as a landing-control pheromone. Bergström _et al._ (1992) confirmed the high acetate 116 proportion in the floral scent of _C. calceolus_ but also showcased very different floral 117 fragrance composition of the very closely related _C. parviflorum_ var. _parviflorum_, and _C._ 118 _parviflorum_ var. _pubescens_, with different terpenoids together with 1,4-dimethoxybenzene 119 dominating the former and 1,3,5-trimethoxybenzene dominating the latter. Studies on Chinese 120 _Cypripedium_ species further illustrate high chemical diversity. In _C. tibeticum_, ethyl acetate, 121 perceived as a faint fruity scent, was suggested to attract bees via food deception (P. Li _et al.,_ 122 2006). The fragrance of _C. plectrochilum_ is dominated by 3-methyl-decen-2-one (P. Li _et al.,_ 123 2008a), while _C. henryi_ emits mainly fatty acid derivatives, such as dodecanal, undecane, 124 nonanal, and decanal (P. Li _et al.,_ 2008b). More recent work by Braunschmid _et al_ . (2017, 125 2020) analysed the scent of _C. calceolus_ very thoroughly identifying over 80 floral 126 compounds, making it the species with the best characterized floral scent in the genus. 127 Collectively, available studies highlight pronounced interspecific variation in floral scent 128 composition within _Cypripedium_ . 129 While the effect of pollination on different variables, such as flower life span (Proctor and 130 Harder 1995), colour change (Weiss 1995) and scent emission (Schiestl and Ayasse 2001), 5 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 131 have been studied, little is known about the link between pollination-induced changes in VOC 132 composition and the biosynthetic and regulatory pathways involved in scent production and 133 emission. One way to address this question is to examine the transcriptional response of 134 plants by assessing differential gene expression under various pollination treatments. This 135 allows for a better understanding of the molecular basis of how plants mediate their 136 relationships with pollinators and helps to elucidate the processes involved in the evolution of 137 deceptive pollination. For example, the floral scent of sexually deceptive _Ophrys sphegodes_ 138 (Orchidaceae) chemically simulates the sex pheromones of its specific pollinator, with 139 alkenes constituting important compounds shared by both the pheromones and the floral 140 scent. It was found that expression of a gene involved in alkene biosynthesis increases in 141 association with increased alkene production (Schlüter _et al.,_ 2011). Mutation in such a gene 142 may result in a change in scent composition and thus alter pollinator attraction. Since 143 pollination in _Ophrys_ is highly species-specific, this could lead to reproductive isolation and, 144 finally, speciation. 145 The extensive floral scent diversity within the genus _Cypripedium_ offers a unique opportunity 146 to study trait evolution, yet the gene expression profiles that regulate these signals during the 147 floral lifespan remain largely uncharacterized. By characterizing this broad chemical diversity 148 alongside the gene expression patterns triggered by pollination, we can better understand how 149 both historical phylogeny and immediate physiological responses shape the deceptive signals 150 of this genus. Additionally, combining the comprehensive volatile data collected in this study 151 with available species-level phylogenies allows us to evaluate the extent of phylogenetic 152 signal in floral scent phenotypes, which we expect to evolve rapidly in a general food 153 deceptive lineage such as _Cypripedium._ Here, we generate data of floral VOCs for 31 154 _Cypripedium_ taxa and use these data together with the available phylogeny (Lagou _et al.,_ 155 2024) and pollinator data of the genus to answer the following research questions: (1) How 156 diverse are the floral VOC profiles across the genus _Cypripedium_ ? (2) How extensive is 157 interspecific variation in floral scent across the genus, and what insights can be gained into its 158 intraspecific variation? (3) Does interspecific scent variation follow a predictable 159 phylogenetic signal, or does it align more closely with shifts in pollinator groups? (4) Finally, 160 does pollination trigger tissue-specific changes in floral volatile production and gene 161 expression within the genus _Cypripedium_ ? To address this final question, we analysed post 162 pollination shifts in VOC composition in _C. henryi_ across four treatments and linked these 163 chemical changes to underlying gene expression patterns using RNA-seq data from distinct 164 floral tissues. 6 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) ### 165 Material and Methods 166 **Plant material** 167 All plant material utilized in this research originated from the living collection of 168 _Cypripedium_ maintained and curated at the Oberhof Garden, Eurasburg, a facility associated 169 with the Botanical Garden Munich-Nymphenburg, Germany. A complete list of all studied 170 taxa, with details regarding their specific provenance, is provided in Supplementary Data 171 Table S1. Plants were cultivated in pots either in a greenhouse or outdoors depending on the 172 species requirements. For the genus-wide comparative study, a total of 31 _Cypripedium_ taxa, 173 including 24 different species, three natural hybrids and three varieties each of _C. macranthos_ 174 and _C. parviflorum_, were included. For the pollination study, two multiflowered clonal 175 individuals of _C. henryi_ were sampled (Accession No. 2023/1153). 176 **Literature review on pollinator species of the genus** _**Cypripedium**_ 177 To establish a comprehensive dataset of _Cypripedium_ pollinators, we updated the checklist 178 provided by Pemberton (2013). Our revised dataset includes pollinator entries for six 179 additional _Cypripedium_ species previously undocumented in the 2013 review (Supplementary 180 Data Table S2). The review was carried out primarily using Google Scholar, complemented 181 by searches on ResearchGate and JSTOR, and was completed in December 2025. Literature 182 searches were conducted using the keywords “ _Cypripedium_ pollination” and “ _Cypripedium_ 183 pollinators.” 184 **Scent collection** 185 Scent collection for the comparative study was conducted _in situ_ at the Oberhof Garden 186 workspace outside the greenhouses. Plant volatiles were collected using a dynamic headspace 187 method following Dötterl _et al_ . (2005). Prior to sampling for comparative analyses across the 188 genus, flowers were examined to confirm the presence and intactness of both pollinia. Leaf 189 samples serving as controls were collected for each floral sample (from the same individual) 190 and ambient air samples were also taken on each collection day. Single flowers and leaves 191 were enclosed in polyester oven bags (Toppits R©), and following a scent accumulation of 5 192 min, samples were collected for 5 min at a flow rate of 0.15 - 0.2 l/min into thermal 193 desorption tubes (TD-tubes) using a pump (Rotary Vane Pump Series G1, Thomas R©). 194 Subsequently, a second sample was collected from the same flower and leaf using the same 195 protocol, but with an extended accumulation and collection time of 15 min each. These two 196 different accumulation and collection times were used as the amounts of scents released from 197 the flowers were unknown and to increase the likelihood of obtaining samples that contain 7 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 198 scent amounts well suited for GC/Ms-analyses (see below). Scent collection for the 199 comparative study was carried out between April 27th and May 23 [rd], 2022, between 9 am and 200 3 pm. The average ambient temperature during scent collection was 18.6 ± 3.1 °C (mean ± 201 SD), with temperatures ranging from 14 to 24 °C. The age of the flowers from which the 202 scent was collected ranged from two to nine days. The flowers of _Cypripedium_ last for 203 between two to five weeks (Cribb 1997). 204 To test for post-pollination effects of scent in _C. henryi_, scent was collected from May 7th to 205 May 16th, 2023, between 10 am and 2 pm under controlled conditions in a greenhouse at the 206 Munich-Nymphenburg Botanical Garden. Here, accumulation and collection time were set at 207 6 minutes each, based on data obtained in the previous analyses. Polyester oven bags of 208 identical dimensions (15x15 cm) were used for bagging to ensure consistency across all 209 samples. The opening of the flower buds was monitored and the first day the flower opened 210 was set as day zero. A flower was considered as open when the lateral petals, the top sepal 211 and the synsepalum had spread away from the labellum and reached their final position. Plant 212 volatiles of both flower and leaf were collected on day two for the first time. Pollination 213 treatments (see below for details) were carried out directly after the first scent collection. A 214 second scent collection was performed on day four, i.e., two days after the pollination 215 treatments. The average ambient temperature during the first scent collection was 22.2 ± 1.6 216 °C (mean ± SD), with temperatures ranging from 19 to 25 °C, whereas during the second 217 scent collection it was 21.1 ± 2.0 °C (mean +- SD), ranging from 19 to 24 °C. 218 **Scent analysis** 219 Scent samples were analysed through Gas Chromatography-Mass Spectrometry (GC/MS). An 220 automatic thermal desorption (TD) system (TD-20, Shimadzu, Japan) was used, coupled to a 221 Shimadzu GC/MS-QP2010 Ultra equipped with a ZB-5 fused silica column (5% phenyl 222 polysiloxane; 60 m, i.d. 0.25 mm, film thickness 0.25 μ m, Phenomenex). When loading the 223 samples onto the column, split ratios varied between 1:1 to 1:25 (Supplementary Data Table 224 S1), with a split ratio of 1:1 being used for the _C. henryi_ samples of the pollination study. 225 Mass spectra were taken at 70 eV (EI mode) from 34 to 350 m/z. To process the GC/MS data, 226 the GCMSolution package Version 4.50 was used (Shimadzu Corporation 1999-2018). For 227 additional GC/MS setting details see Braunschmid _et al_ . (2017). In the comparative study, 228 samples with the shorter accumulation and collection time (5 min) were analysed first. 229 Samples with the longer accumulation and collection time (15 min) were only analysed if the 230 sample with shorter collection time showed a low amount of fragrance compounds. If both 8 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 231 samples were analysed, the one with the higher scent amounts (typically the sample with the 232 longer collection time) was selected for downstream analyses. Plant volatiles were identified 233 by database comparisons (retention index, mass spectra), using the databases ADAMS, 234 ESSENTIALOILS-23P, FFNSC 2, W9N11, and a database created by the Plant Ecology 235 Laboratory of the University of Salzburg using available synthetic standards. A compound 236 was identified as a floral volatile if it was absent in the leaf sample, or if the peak area (total 237 ion current) in the gas chromatogram was at least five times higher in the flower sample than 238 in the associated leaf sample. In the experimental pollination study with _C. henryi_, floral 239 compounds with a single occurrence were excluded from the analyses. 240 **Pollination treatments of** _**Cypripedium henryi**_ **flowers** 241 Four flower-level treatments were conducted using 12 flowers from two plant individuals. (1) 242 Removal of both pollinia (removal); (2) deposition of a pollinium from the same clonal 243 accession on the stigma (deposition); (3) both the removal and deposition treatment combined 244 (deposition+removal); (4) control group with no artificial pollination performed (control). To 245 minimize potential plant-individual bias, flowers assigned to each treatment were distributed 246 as evenly as possible across the two sampled plants. Due to a delayed onset of volatile 247 emission, two flowers (one each from the control and removal group) were excluded from the 248 scent analysis. Consequently, these samples were also omitted from the differential 249 expression analysis and all subsequent steps within the transcriptomic pipeline. 250 **RNA isolation and Transcriptome Sequencing** 251 The labellum, staminode and lateral petals were collected in liquid nitrogen immediately 252 following the second scent sampling (i.e. 48 h (± 2 h) after pollination treatment). Total RNA 253 was isolated using the InnuPREP Plant RNA Kit (Innuscreen GmbH). For each flower, the 254 whole staminode (around 15 mg), one longitudinal half of the labellum (around 60 mg) and a 255 whole lateral petal (around 30 mg) were used for RNA extraction. Prior to the extraction, the 256 tissue samples were crushed on dry ice using a rotor-stator homogenizer. Total RNA was 257 eluted in 20 µl RNAse-free H 2 O. RNA quality was assessed via Agilent ScreenTape, with the 258 RIN threshold of samples ranging between 5.8 and 8.4. Both transcriptome library preparation 259 and sequencing were performed by BMKGENE (Münster, Germany). Libraries were 260 constructed using a poly(A)-selection strategy and a strand-specific (dUTP) protocol, 261 followed by sequencing on an Illumina NovaSeq 6000 platform to generate 150 bp paired-end 262 reads. 9 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 263 **Transcriptome assembly** 264 Pre-assembly of transcriptomic data was performed following the workflow of Siadjeu and 265 Kadereit (2024). The following tools were used: FastQC v0.12.1 266 (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and MultiQC v1.17 (Ewels _et_ 267 _al.,_ 2016) for sequence quality control assessment, Rcorrector v1.0.6 (L. Song and Florea 268 2015) for correction of sequencing errors, TranscriptomeAssemblyTools 269 (https://github.com/harvardinformatics/TranscriptomeAssemblyTools) for removal of 270 uncorrectable reads, TrimGalore v0.6.10 271 (https://github.com/FelixKrueger/TrimGalore/releases) for filtering of adapters and low 272 quality reads, and finally Bowtie2 2.5.2 (Langmead and Salzberg 2012) and SILVA v-138 273 rRNA database for removal of rRNA. 274 De novo transcriptome assembly was performed using Trinity v2.15.1 (Grabherr _et al.,_ 2011) 275 with the following parameters: Trinity –seqType fq –max_memory 150G –CPU 54. The 276 quality of the transcriptome assembly was checked using two different methods. First read 277 support was checked by mapping the cleaned reads to the assembled transcriptome using 278 Bowtie2 v2.5.2. Second BUSCO completeness was checked by using the software BUSCO 279 v5.5.0 (Manni _et al.,_ 2021). 280 **Transcript quantification and differential expression analysis** 281 The assembled reads were clustered at 98% similarity with CD-HIT v4.8.1 (Fu _et al.,_ 2012) to 282 remove duplicates and obtain the final transcripts. For subsequent analysis, transcripts with an 283 ORF were extracted using Transdecoder v.5.5.0 284 (https://github.com/sghignone/TransDecoder). Transcript quantification was performed using 285 Salmon 1.10.2 (Patro _et al.,_ 2017). Only transcripts with an occurrence of at least ten in three 286 different samples were kept for further analysis. Differential expression analysis was 287 performed in R Statistical Software v4.3.3 (R Core Team 2025) using the DESeq2 package 288 (Love _et al.,_ 2014). To account for inter-individual variation between the two sampled plants, 289 Plant ID was incorporated into the generalized linear model as a blocking factor (design = ~ 290 Plant_ID + Treatment). Transcripts were considered differentially expressed if they exhibited 291 a false discovery rate (FDR) of <5% (adjusted p-value <0.05) and a ∣ log 2 FC ∣ >1. To evaluate 292 the impact of pollination on volatile-related gene expression, each treatment group was 293 compared against the control across three distinct floral tissues: petals, staminodes, and 294 labella. For the purpose of this analysis, "petals" refers exclusively to the lateral petals, as 295 distinct from the labellum, which represents a morphologically modified third petal. 10 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 296 **Gene annotation and candidate gene selection** 297 Metabolic pathways and annotations of differentially expressed transcripts were assigned via 298 the tool Mercator4 v6.0 (Schwacke _et al.,_ 2019). Both swiss-prot protein sequences database 299 and prot-scriber were included to improve the annotations. Subsequent analysis focused on 300 genes involved in floral volatile biosynthesis in _C. henryi_, specifically those associated with 301 the fatty acid (FA) and terpenoid biosynthesis pathways. 302 **Data visualisation and statistical analysis** 303 To visualize scent sample (dis)similarity, NMDS plots were created using the R packages 304 vegan v2.6-4 (Oksanen _et al.,_ 2024) and ggplot2 v4.0.1 (Wickham 2016). For the genus-wide 305 dataset, we additionally utilized the package chemodiv v0.2.0 (Petrén _et al.,_ 2023) to compute 306 Generalized UniFrac distances, which account for structural and biosynthetic similarity 307 between volatiles. Specifically, Generalized UniFrac (GenUniFrac) dissimilarities were 308 calculated with the parameter α set to 1, thereby weighting compound dissimilarities in direct 309 proportion to their relative abundances. To visualize the distribution of floral VOCs across all 310 sampled _Cypripedium_ taxa, a heatmap based on relative abundances was generated using the 311 R package pheatmap (Kolde 2010). To simplify the visualization and focus on major 312 constituents, the dataset for this heatmap was reduced to only include floral VOCs whose total 313 relative quantity across sampled taxa exceeded an arbitrary value of 8%. 314 Next to determine the proportion of variance in the floral scent composition explained by 315 taxonomic section and pollinator group, we conducted a PERMANOVA (Permutational 316 Multivariate Analysis of Variance) with 10,000 permutations, based on chemically informed 317 GenUniFrac dissimilarities. Dissimilarity matrices were generated using the chemodiv 318 package, and multivariate testing was performed in vegan. We categorized the _Cypripedium_ 319 taxa in our dataset into three main pollinator syndrome categories: (1) bees, (2) bees and 320 floricolous flies, (3) drosophilid flies. These assignments were derived from primary literature 321 or, in the absence of empirical records, through phylogenetic inference while also considering 322 the visual phenotype of the flower. To assess the correlation between floral scent composition 323 and phylogenetic relationships, we performed a Mantel test across the 31 sampled taxa. This 324 analysis compared the GenUniFrac dissimilarity matrix against a matrix of pairwise 325 phylogenetic distances, which were derived from the Lagou _et al_ . (2024) phylogeny using the 326 ape v5.8 package. 327 To assess the effect of pollination treatment on the floral scent of _C. henryi_ while accounting 328 for individual variation, we conducted a Type-I sequential PERMANOVA. Treatment type 11 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 329 was used as a fixed factor, with Plant ID entered first in the model to partition out inter 330 individual variance. Again, this analysis was based on GenUniFrac dissimilarities calculated 331 with chemodiv. Additionally, paired t-tests were conducted to assess variation within 332 individual flowers between the first and second scent collection events for both the total scent 333 emission and the relative abundances of individual compounds. Since our analysis did not 334 include internal calibration standards, total scent emission was estimated as the sum of all 335 volatile peak areas. Although absolute emission rates were not quantified, total peak area 336 serves as a reliable proxy for total scent output, allowing valid intra-individual comparisons 337 given our standardized sampling conditions. For compound-level analyses, p-values were 338 corrected for multiple testing using the Bonferroni method. Expression profiles of the 339 differentially expressed candidate genes identified across pollination treatments in _C. henryi_ 340 were visualized using the R package pheatmap (Kolde & Kolde, 2015). ### 341 Results 342 **Floral fragrance composition in** _**Cypripedium**_ 343 In the 44 analysed scent samples collected from 31 _Cypripedium_ taxa, a total of 258 floral 344 volatiles were found of which 174 were (tentatively) identified. These included 64 aliphatic 345 compounds, 63 terpenoids, 34 aromatic compounds, eight C5-branched compounds, three N 346 bearing compounds, as well as one miscellaneous cyclic compound and one spiroacetal 347 (Supplementary Data Table S3). The total number of floral compounds detected per flower 348 ranged from 1 to 42 compounds, with an average of 14.1 ± 12.7 (SD) per sample and 15.2 ± 349 12.8 (SD) at the taxon level. Floral scent samples of _C. calceolus_ and _C. californicum_ held the 350 highest number of floral volatiles with 42 and 39 compounds, respectively. The floral volatile 351 with the highest occurrence was the monoterpenoid linalool (Figure 2), which was identified 352 as a floral compound in nine taxa of sect. _Cypripedium_ and in _C. californicum_ from the 353 monotypic sect. _Californica_ . Linalool derivatives like linalool oxides (furanoids, pyranoids) 354 and lilac compounds, were also detected in smaller amounts than linalool, and only occurred 355 in samples that also included linalool with the exception being _C. froschii_ . The other most 356 often occurring floral VOCs were the aliphatic compounds ( _Z_ )-3-hexen-1-ol (8 taxa) and 1 357 hexanol (7 taxa), the aromatic compounds benzyl alcohol (7 taxa) and 2-phenylethanol (6 358 taxa) and the monoterpenoid ( _E_ )- β -ocimene (7 taxa). Identification was limited ( ≤ 75% of total 359 scent) for five of the 31 taxa studied. Most notably, _C. flavum_ exhibited a particularly 360 complex profile of unknowns, with 20 compounds remaining unidentified (comprising 27% 361 of its total scent). In contrast, the unidentified portion of the _C. passerinum_ bouquet (29%) 12 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 362 consisted of only two components. Two further unidentified compounds (unk_739 and 363 unk_1097) contributed significantly to the profiles of _C. froschii_ (9% and 21.4%), _C._ 364 _macranthos_ var. _hotei-atsumorianum_ (7.1% and 10.5%), and _C. macranthos_ var. _rebunense_ 365 (8.1% and 10.5%). Detailed RI and mass spectral data for all unidentified components are 366 available in Supplementary Data Table S4. 367 **Comparison of floral scent across** _**Cypripedium**_ **sections** 368 The average number of detected floral volatiles per taxon (n = 31; comprising 24 species and 369 6 varieties) as well as the chemical nature of compounds varied notably among the different 370 sections of _Cypripedium_ covered in this study (Figure 3). Section _Arietinum_ exhibited the 371 lowest floral volatile diversity, with only two compounds, epoxy-oxoisophorone (89%) and 4 372 oxoisophorone (11%), both terpenoids, detected in the single sampled taxon, _C._ 373 _plectrochilum_ . Section _Bifolia_ averaged 10.7 floral volatiles per sampled taxon and consisted 374 of aliphatics, terpenoids and aromatic compounds. Only two floral compounds were identified 375 in _C. yatabeanum_, the terpenoid dihydro-ß-ionone (53%) and the aliphatic compound ( _Z_ )-3 376 hexen-1-ol (47%). The floral scents of the other taxa of sect. _Bifolia_ were more complex: _C._ x 377 _alaskanum_ was dominated by the aromatic compounds benzyl alcohol (78%) and 378 benzaldehyde (15%), with minor aliphatic and terpenoid components, while _C. guttatum_ had 379 the most diverse profile, including the aliphatic compound methyl octanoate (58%), the 380 aromatic compounds benzyl alcohol (21%), 1,4-dimethoxybenzene (4%) and methyleugenol 381 (3%), and the terpenoid β -citronellol (2%). The highly complex scent of _C. californicum_ of 382 the monotypic sect. _Californica_ was composed almost exclusively of aromatics and 383 terpenoids with benzyl alcohol (25%), geranyl acetone (15%) and methoxymethylbenzene 384 (14%) being the most prevalent compounds. Section _Cypripedium_, to which the bulk of the 385 sampled taxa belonged, averaged 8.8 floral volatiles per taxon. Aliphatics, aromatics, and 386 terpenoids were about equally prevalent in this section, alongside a notable proportion of 387 unidentified compounds. Linalool was the most prominent VOC, occurring in nearly half of 388 the section's taxa (9 out of 19) and acting as the dominant scent constituent in five of those 389 cases. Floral scent of sect. _Flabilinervia_ consisted of only six different compounds, with 390 exclusively aliphatics and aromatics making up the scent of the two species of this section. 391 Floral scent of _C. japonicum_ was dominated by 3-octanone (68%) and _p_ -anisaldehyde (17%) 392 while the scent of _C. formosanum_ flowers was more evenly distributed around 3-octanol 393 (37%), ( _Z_ )-3-hexen-1-ol (36%) and 4-methylanisole (27%). Flowers of sect. _Obtusipetala_ 394 exhibited the chemically most diverse scent profile with representatives of most major 13 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 395 compound classes, and also notable quantities of C5-branched chain compounds, N-bearing 396 compounds and a significant proportion of unknown volatiles. This diversity was reflected by 397 the dominant compounds of the three sampled taxa: the aliphatic pentadecane (63%) in _C._ 398 _passerinum_, the aromatic methyl salicylate (31%) in _C. reginae_, and the N-bearing δ -1 399 pyrroline (15%) in _C. flavum_ . Finally, both sampled species of sect. _Sinopedilum_, _C._ 400 _micranthum_ and _C. bardolphianum_, had floral scents dominated by aliphatic compounds, with 401 _C. micranthum_ primarily emitting acetoin (80%) and the floral scent of _C. bardolphianum_ 402 being characterized by ethyl butyrate (48%) and the aromatic compound 2-methoxyphenol 403 (24%). 404 **Floral scent variation within** _**Cypripedium**_ 405 Floral scent varied also within each section of _C_ _ypripedium_ (Figure 4). We found substantial 406 diversity in floral volatiles among _Cypripedium_ species, with many compounds being unique 407 to a single taxon out of the 31 investigated. Notably, acetoin, pentadecane, methyl 408 dodecanoate, and dodecanal constituted respectively 80%, 63%, 59%, and 47% of the total 409 floral fragrance in individual species, while not being identified as floral volatiles in any of 410 the other taxa examined. The PERMANOVA results indicated that the taxonomic sections 411 accounted for 15% of the total scent variation, though this effect was not statistically 412 significant (F 5,23 = 0.95, p = 0.586). In contrast, pollinator groups significantly explained 10% 413 of the variance in floral scent composition (F 2,23 = 1.55, p = 0.049). We also detected 414 substantial intraspecific variation of floral scent profiles among the three sampled 415 infraspecific taxa of _C. parviflorum_ . Only three floral VOCs were shared across the taxa 416 examined, and none of these occurred in all three. 1,4-Dimethoxybenzene dominated the 417 floral fragrances of both _C. parviflorum_ var. _pubescens_ f. _planipetalum_ (59%) and _C._ 418 _parviflorum_ var. _makasin_ (32%). The only shared compound between _C. parviflorum_ var. 419 _makasin_ and _C. parviflorum_ var. _pubescens_ f. _pubescens_ was indole (5% and 2%, 420 respectively). Finally, benzaldehyde was a major constituent in _C. parviflorum_ var. _pubescens_ 421 f. _planipetalum_ (26%) and _C. parviflorum_ var. _pubescens_ f. _pubescens_ (18%). In comparison, 422 the three sampled taxa of _C. macranthos_ exhibited a higher degree of volatile overlap. This 423 was especially the case for _C. macranthos_ var _rebunense_ and _C. macranthos_ var. _hotei-_ 424 _atsumorianum_, which shared a similar overall floral scent profile (Figure 4). The floral scent 425 profile of _C. macranthos_ var. _macranthos_ consisted of only four VOCs, with linalool being 426 the sole compound shared with the other two sampled infraspecific taxa of the species. 427 **Phylogenetic signal analysis in floral scent of** _**Cypripedium**_ 14 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 428 A Mantel test revealed a significant positive correlation between scent profile dissimilarity 429 and phylogenetic distance (p = 0.048), indicating that phylogenetically closely related species 430 share more similar scent compositions. However, the effect size measured as the Pearson 431 correlation was low with r = 0.161. 432 **Post-pollination effect on floral scent of** _**Cypripedium henryi**_ 433 In the 20 analysed floral scent samples (ten each pre- and post-treatment), 28 of the 33 434 detected floral volatiles could be identified (Table 1). The identified floral volatiles were 435 mostly aliphatics (19 compounds) followed by terpenoids (5 compounds) and aromatic 436 compounds (4 compounds). Five floral volatiles could not be identified but represented a 437 small proportion of less than 1 % of the total floral fragrance. Floral scent samples prior to 438 pollination treatments were dominated by aliphatic compounds, specifically FA derivatives. 439 Dodecanal was the most abundant compound, accounting for 57–77% of the total volatile 440 emission, followed by undecane (4.8–16.8%). Nonanal and decanal were generally the next 441 highest concentrated VOCs in floral scent samples, but were only intermittently identified as 442 floral VOCs, as their abundance did not consistently exceed the five-fold threshold relative to 443 leaf control samples. In addition to these dominant volatiles, a suite of other FA-derived 444 compounds was consistently present in all samples, regardless of treatment. These included 445 hexanoic acid, 1-decanol, and several C12 to C14 derivatives (Table 1). Next to these 446 aliphatic compounds, the terpenes linalool and 4-oxoisophorone were also consistently 447 present, with linalool contributing 2.1–13.3% and 4-oxoisophorone about 1% to the floral 448 scent. 449 Changes in floral scent emission following treatment were predominantly quantitative rather 450 than qualitative, as the chemical composition remained largely consistent across all samples. 451 Total scent generally decreased between collection events; while the groups control, removal, 452 and deposition+removal showed non-significant reductions of 1.9-fold, 2.9-fold, and 3.6-fold, 453 respectively (p > 0.05), the deposition treatment exhibited a significant 3.3-fold reduction (p = 454 0.003). The decline in absolute amounts was consistent for all major volatiles across all 455 treatment groups. At the individual compound level, several relative shifts were observed 456 within the different treatment groups (see Table 1). However, none of these relative 457 differences remained significant after Bonferroni correction for multiple comparisons. 458 Although not statistically significant, linalool showed a downward trend across all groups, 459 being ubiquitous in the first scent collection but only present in three post-treatment samples 460 (one replicate each from control, removal, and deposition+removal). 15 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 461 NMDS ordination of _C. henryi_ floral scent samples showed substantial overlap between post 462 pollination treatment groups (Figure 5). While samples from the deposition treatment formed 463 a loose grouping post-pollination, a similar pattern was also evident in the corresponding pre 464 treatment samples. Type-I sequential PERMANOVA of post-treatment scent samples 465 revealed that the pollination treatments had no significant effect on the relative floral scent 466 composition (F 3,5 = 1.36, p = 0.282). In contrast a highly significant ‘Plant ID’ effect was 467 detected (F 1,5 = 5.97, p = 0.006), which accounted for 39.7% of the observed variation in 468 scent profiles. 469 **Transcriptome analysis** 470 RNA sequencing yielded between around 31M and 43M reads per sample. In total, 471 2,647,661,174 reads were processed with Rcorrector, resulting in the correction of 472 987,472,741 bases. The de novo transcriptome assembly generated a total of 563,487 473 transcripts. This study was conducted using de a novo transcriptome assembly that did not 474 have reliable gene models. As a result, we performed expression quantification and 475 differential expression analysis at the transcript level instead of the gene level. Therefore, 476 throughout the manuscript, we will refer to significantly differentially expressed features as 477 differentially expressed transcripts (DETs). Evaluation with BUSCO (n = 1,614, 478 _embryophyta_odb10_ ) indicated high completeness (C: 95.3%), with 19.6% single-copy and 479 75.7% duplicated BUSCOs, alongside low fragmentation (2.9%) and few missing genes 480 (1.8%). Read support analysis further confirmed the robustness of the assembly, with 98.91% 481 of reads aligning back to the transcriptome. Using TransDecoder, 124,929 coding transcripts 482 were identified. After filtering for low-abundance transcripts and excluding the samples from 483 the two flowers with delayed volatile emission (one each from the control group and the 484 removal treatment), the dataset used for differential expression analysis consisted of 70,804 485 transcripts. 486 **Differential expression analysis** 487 To identify treatment-induced transcriptional shifts, the transcriptomes from treated flowers 488 were compared against the control group within each tissue type. Overall, the petals exhibited 489 a markedly higher number of DETs than the labellum or staminode (Supplementary Table 490 S5). Functional annotation with Mercator4 revealed 19 DETs linked to FA biosynthesis, and a 491 single DET associated with terpenoid synthesis. Consistent with the global expression trends, 492 the transcriptional response of identified candidate genes was more extensive in the petals 16 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 493 than in the labellum or staminode (Figure 6). Expression profiles revealed almost no patterns 494 restricted exclusively to the two treatments involving pollination (deposition and 495 deposition+removal). The sole exception was the chloroplastic pyruvate dehydrogenase E1 496 component subunit alpha-3 ( _PDH-E1-_ α ), involved in acetyl-CoA generation, which was 497 slightly but significantly up-regulated in petals. Furthermore, treatment-specific patterns 498 shared across tissues were rare, with the exception of 3-ketoacyl-CoA synthase 20 ( _KCS20_ ). 499 This gene, involved in FA elongation, was down-regulated following the deposition treatment 500 in both the staminode and labellum, but remained stable in the petals. Several DETs 501 associated with FA biosynthesis, exhibited uniform tissue-specific responses across all 502 treatments. For instance, the up-regulation of _ZC3H55_ (encoding zinc finger CCCH domain 503 containing protein 55) was restricted to the staminode. In contrast, petal-specific up 504 regulation was observed for two acyl-coenzyme A oxidase paralogs ( _ACX3_ and _ACX1_1_ ) and 505 an isoform of biotin carboxylase 1 ( _BCCP1_2_ ), while the peroxisomal long-chain acyl-CoA 506 synthetase 6 ( _LACS6_ ) was consistently down-regulated in the petals across all treatments. 507 Regulation of _BCCP1_ was found to be isoform-specific within the petals, where the two 508 identified isoforms showed opposing directions of differential expression. The only identified 509 terpenoid-related DET was 3-hydroxy-3-methylglutaryl-CoA reductase 3 ( _HMGCR3_ ), a key 510 enzyme in the mevalonate (MVA) pathway. _HMGCR3_ exhibited a highly tissue-specific 511 response profile: while it was constitutively up-regulated across all treatments in the petals, its 512 activation in the staminode was restricted to both the removal and deposition group. Notably, 513 in the labellum, _HMGCR3_ was uniquely up-regulated only under the combined 514 deposition+removal treatment. Detailed statistical values, including log 2 FC and p adj, for these 515 transcripts are provided in Supplementary Table S6. ### 516 Discussion 517 **Floral scent diversity in the genus** _**Cypripedium**_ 518 _Overview of floral scent in the genus_ Cypripedium 519 In the comparative study we discovered a substantial diversity of floral VOCs in 520 _Cypripedium_ . Across 31 sampled taxa and covering 24 out of the 54 accepted species of the 521 genus, we identified 258 floral VOCs, predominantly terpenoids, aliphatics, and aromatics. 522 While some floral VOCs were detected across multiple taxa, many were exclusive to a single 523 taxon. To our knowledge, this study provides one of the most extensive genus-wide samplings 524 of floral volatiles conducted thus far. The only broader genus-level analyses to date are those 525 by Füssel _et al._ (2007) on 34 _Salix_ species and Cao _et al._ (2026) on 32 _Ficus_ species, both 17 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 526 belonging to substantially larger genera (~475 and ~892 species, respectively; World 527 Checklist of Selected Plant Families, accessed June 2026). Notably, our _Cypripedium_ dataset 528 exhibited greater chemical diversity than similarly sized datasets of the highly fragrant, 529 euglossine bee-pollinated orchid genera _Stanhopea_ (Williams and Whitten 1999), _Gongora_ 530 (Hetherington-Rauth and Ramírez 2016) and _Catasetum_ (Milet-Pinheiro and Gerlach 2017). 531 However, it is worth noting that aforementioned studies captured a smaller percentage of 532 genus-wide variation, as these genera contain more species than _Cypripedium_ . In our study, 533 several taxa such as _C. californicum_ and _C. calceolus_ exhibited very complex floral scent 534 profiles. While the number of individual VOCs varied substantially between species, the high 535 cumulative diversity observed within _Cypripedium_ may be driven by the broad range of 536 pollinators exploited across the genus. The monoterpene alcohol linalool appears to be a 537 central component of the _Cypripedium_ chemical repertoire, particularly within section 538 _Cypripedium_ where it frequently dominates the scent bouquet. Such widespread occurrence 539 aligns with the broader role of linalool as one of the most ubiquitous floral volatiles in 540 angiosperms (Knudsen _et al.,_ 2006; Farré-Armengol _et al.,_ 2020), likely serving a dual 541 function in these orchids as both a primary pollinator attractant and a defence against 542 herbivory (L. Zhang _et al_ ., 2023) and floral pathogens (Shimada _et al.,_ 2014). 543 _Floral scent variation in_ Cypripedium _: Pollinator influence and chemical relationships_ 544 Although pollinator group significantly influenced scent composition (p = 0.049), it 545 accounted for only 10% of the total variance. This relatively low explanatory power aligns 546 with a strategy of generalized food deception hypothesized for the majority of our sampled 547 taxa. Unlike Batesian mimicry or sexually deceptive pheromone systems, generalized food 548 deception lacks the stringent evolutionary pressure for high chemical specificity. The floral 549 scents of taxa within the bees and floricolous flies pollination group contained several well 550 known bee attractants, including the aromatic methyl salicylate and the common terpenes 551 linalool and ( _E_ )- β -ocimene (Williams and Whitten 1983; Gerlach and Schill 1991; Farré 552 Armengol _et al.,_ 2017), which are likely associated with the presence of nectar and pollen by 553 these insects. Electrophysiological studies support the functional relevance of these 554 compounds to bees; linalool, in particular, has been shown to elicit strong antennal responses 555 in known pollinator species of _Cypripedium_, specifically within the key pollinator genera 556 _Lasioglossum_, _Andrena_, and _Apis_ (Luo _et al.,_ 2013; Braunschmid _et al.,_ 2017). As linalool is 557 a dominant constituent in numerous bee-pollinated lineages (Gong _et al.,_ 2015; Krug _et al.,_ 558 2018; Cordeiro and Dötterl 2023), its prevalence in our sampled taxa likely reflects an 18 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 559 exploitation of conserved sensory biases in these hymenopterans. Among the three species in 560 our dataset with documented bee and fly pollination, _C. calceolus_ and _C. reginae_ are 561 associated with hoverflies, whereas _C. flavum_ is visited by blowflies (Table S2). Interestingly, 562 ( _E_ )- β -farnesene, an aphid alarm pheromone known to attract hoverflies in search of an 563 oviposition site (Jiang _et al.,_ 2020; B. Wang _et al.,_ 2022), was present in our overall dataset 564 but absent from the two hoverfly pollinated species _C. calceolus_ and _C. reginae_ . The absence 565 of ( _E_ )- β -farnesene or other aphid-related VOCs in the floral fragrance of _C. calceolus_ 566 combined with the fact that the species is predominantly pollinated by generalist polylectic 567 bees, suggests that hoverflies likely visit the flowers in search of nutritive rewards rather than 568 oviposition sites. In contrast, _C. reginae_ may indeed employ brood site mimicry, given that its 569 most abundant floral VOC methyl salicylate (31%) and ethyl salicylate (7%) are well-known 570 herbivore-induced plant volatiles (HIPVs). Methyl salicylate in particular, is frequently 571 produced by plants in reaction to aphid infestation, which consequently exerts a strong 572 attraction on hoverflies in search of an aphid colony for oviposition (Primante and Dötterl 573 2010; Rodriguez-Saona _et al.,_ 2011; Salamanca _et al.,_ 2019). Although a multimodal strategy 574 integrating both visual and olfactory cues is likely, hoverflies often exhibit a strong reliance 575 on visual floral cues (Laubertie _et al.,_ 2006; Mishra _et al.,_ 2025). This is exemplified by the 576 tropical slipper orchid _Paphiopedilum barbigerum_, which is also hoverfly pollinated but lacks 577 floral volatiles. In this species, the yellow-coloured staminode was hypothesized to be the 578 primary visual attractant, leveraging the known attraction of hoverflies to this colour (Shi _et_ 579 _al.,_ 2009). A comparable visual cue exists in _C. reginae_, whose staminode also exhibits 580 distinct yellow spots. The bimodal volatile profile of _C. flavum_ suggests an adaptive dual 581 signalling strategy adapted to the olfactory preferences of its pollinators, bumblebees and 582 blowflies. The floral VOC δ -1-pyrroline and the two C5-branched compounds, which are 583 associated with decay or fermentation odours, could increase the attractiveness to the 584 pollinating blowfly species, while the presence of the terpenes ( _Z/E_ )- β -Ocimene, α -Copaene, 585 and β -Caryophyllene, likely serves as the primary olfactory attractant for bumblebees (Dötterl 586 and Gershenzon 2023). The drosophilid-pollinated _C. micranthum_ and _C. bardolphianum_ 587 emitted scent profiles that were markedly different from the rest of the genus, characterized 588 predominantly by aliphatic alcohols and a distinct lack of shared major VOCs. Although the 589 fragrances of these two species were perceived as similarly sweet and alcoholic, their 590 chemical compositions differed substantially. The floral scent of _C. bardolphianum_ was 591 defined by the aliphatic esters ethyl butyrate and ethyl propanoate, along with the phenolic 592 compound 2-methoxyphenol. While the former two esters likely mimic the chemical signature 19 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 593 of fermenting or overripe fruit attractive to most drosophilid species, 2-methoxyphenol is 594 known to function not only as a floral attractant for moths (Knudsen _et al_ ., 1993) but also as a 595 volatile marker of microbial (Tahir _et al_ ., 2019) and insect-driven (Revadi _et al_ ., 2021) 596 degradation of plant tissues. In contrast, the profile of _C. micranthum_ was dominated by the 597 fermentation-associated aliphatic compounds acetoin and 2,3-butanediol. These latter 598 compounds were also found in high concentrations in the drosophilid-pollinated _Aristolochia_ 599 _baetica_ (Rupp _et al.,_ 2024). The additional detection of β -cyclocitral, a known fruit fly 600 attractant (Piñero _et al.,_ 2019), provides further evidence for a highly specialized drosophilid 601 attraction mechanism in _C. micranthum_ . 602 _Phylogenetic signal in_ Cypripedium _’s floral scent_ 603 Our analyses reveal a complex evolutionary pattern of floral scent variation within 604 _Cypripedium_ . At the sectional level, we did not detect a significant effect of phylogenetic 605 proximity on floral fragrance similarity. This lack of significance may be attributed to the fact 606 that taxonomical sections represent relatively coarse categories and high intra-sectional 607 variability likely obscured scent-based differences between sections. Furthermore, several 608 sections were represented by small sample sizes, partly due to our dataset composition and 609 partly because sections such as _Flabellinervia_ (n = 2), _Arietinum_ (n = 2), and _Californica_ (n = 610 1) are inherently species-poor. At the interspecific level however, the Mantel test revealed a 611 significant, although weak, correlation between phylogenetic distance and floral scent 612 dissimilarity (r=0.161). While this indicates the presence of a phylogenetic signal, the 613 explanatory power was remarkably low (r [2] =0.026), suggesting that phylogenetic relatedness 614 accounts for only ~2.6% of the observed variation. This interpretation is further supported by 615 the observation that even closely related species in our dataset often shared only a small 616 number of floral scent compounds. A weak phylogenetic signal is consistent with the view of 617 floral scent as a highly labile and rapidly evolving trait. Floral scent is often subject to intense 618 selection by pollinators and local ecological conditions, which limits its usefulness for 619 phylogenetic reconstruction (Williams and Whitten 1999; Levin _et al.,_ 2003). The 620 predominant role of ecology over ancestry is supported by our finding that, while sectional 621 taxonomy had no effect, pollinator groups significantly influenced scent composition in 622 _Cypripedium_ . While phylogenetic relationships influence floral scent variation in some 623 Orchidaceae lineages (Steiner _et al.,_ 2011; Joffard _et al.,_ 2020), the generalized food 624 deception strategy employed by most _Cypripedium_ taxa could contribute to the weak signal 625 observed here. Because these orchids often target a broad pollinator spectrum, stabilizing 20 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 626 selection on specific VOC profiles may be relaxed. This allows for high chemical plasticity 627 and rapid divergence between species. The degree of phylogenetic signal in floral scent can be 628 a product of pollinator specificity. For instance, Prieto � Benítez _et al_ . (2016) reported a 629 stronger phylogenetic signal in nocturnal scents compared to diurnal scents in _Sileneae_, 630 attributing this pattern to the more taxonomically constrained nature of the nocturnal 631 pollinator assemblage compared to the diurnal one. Nevertheless, one must consider that 632 distinct pollinator groups carry distinct evolutionary and sensory limitations. These pollinator 633 constraints also directly influence the variability of floral scent bouquets over evolutionary 634 time. To conclude, the broad and often overlapping pollinator assemblages across 635 _Cypripedium_ likely account for this pattern, where a significant but minimal phylogenetic 636 signal is present but where ancestral constraints are secondary to the selective pressures of 637 pollinator attraction, resulting in a scent profile that is functionally driven rather than 638 phylogenetically conserved. 639 _Intra-specific variation of floral scent in_ Cypripedium 640 High intraspecific variation was uncovered within the three sampled infraspecific taxa of both 641 _C. macranthos_ and _C. parviflorum_ . In the latter, the variation was so prominent that the two 642 sampled forms of var. _pubescens_ shared only a single common floral VOC. Such chemical 643 heterogeneity aligns with the polyphyletic nature of the var. _pubescens_ complex recently 644 identified by Lagou _et al_ . (2024). These results provide further evidence of the systematic 645 confusion surrounding the var. _pubescens_ complex, highlighting the critical need for a 646 taxonomic revision of this group. Variation in floral VOCs was also pronounced among the 647 sampled infraspecific taxa of _C. macranthos_, but to a bit smaller amount. In the phylogenetic 648 analysis by Lagou _et al_ . (2024), which utilized a significantly more comprehensive sampling 649 of _C. macranthos_ infraspecific taxa than previous studies (Fatihah _et al.,_ 2011; J. Li _et al.,_ 650 2011; Szlachetko _et al.,_ 2021), both _C. franchetii_ and _C. yunnanense_ were recovered as nested 651 within the _C. macranthos_ clade. While both species were included in the present study, our 652 floral scent data neither contradict nor explicitly support this placement; neither species 653 exhibited a profile particularly similar to the sampled varieties of _C. macranthos_ . Specifically, 654 _C. franchetii_ shared only linalool, while _C. yunnanense_ possessed no volatile organic 655 compounds (VOCs) in common with the _C. macranthos_ varieties. Interestingly, within this 656 broader ' _macranthos_ ' clade, i.e., including _C. franchetii_ and _C. yunnanense_, floral scent 657 exhibited high functional lability even at shallow taxonomic scales. For instance, the scent 658 profiles of _C. macranthos_ var. _rebunense_ and _C. macranthos_ var. _hotei-atsumorianum_ were 21 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 659 the most similar (Figure 4), despite not being closely related within the species complex 660 (Lagou _et al.,_ 2024). Comparison of the floral scent profiles generated in this study with 661 published volatile data for _Cypripedium_ species revealed both expected similarities and 662 several unexpected discrepancies. For _C. calceolus_, _C. reginae_, and _C. guttatum_, our samples 663 largely confirm previously reported scent compositions, with our dominant core compounds 664 matching published profiles and differences confined primarily to lower abundance VOCs 665 (Bergstrom _et al.,_ 1992; Barkman _et al.,_ 1997; Braunschmid _et al.,_ 2017, 2020). In contrast, 666 published data for _C. parviflorum_ var. _pubescens_ already show marked variation. Bergström 667 _et al_ . (1992) and Barkman _et al_ . (1997) reported strikingly different floral fragrance 668 compositions, with the sample of the latter study lacking the otherwise highly dominant 669 compound 1,3,5-trimethoxybenzene as well as most terpenoid VOCs. Our sample closely 670 resembled the profile reported by Barkman _et al_ . (1997), characterized by 2-phenylethanol as 671 the major constituent, followed by benzaldehyde and phenylacetaldehyde. Surprisingly, the 672 floral scent profiles we obtained for several species stands in stark contrast to previously 673 published floral VOCs data, though similar methods were used among the studies. For 674 instance, our findings failed to confirm the dominant compounds previously reported for _C._ 675 _kentuckiense_ (Barkman _et al_ ., 1997), _C. tibeticum_ (P. Li _et al_ ., 2006), _C. plectrochilum_ (P. Li 676 _et al_ ., 2008a), _C. flavum_ (Zheng _et al_ ., 2011), _C. micranthum_, and _C. bardolphianum_ (P. Li, 677 2006). Most notably, in the cases of _C. tibeticum_, _C. flavum_, and _C. plectrochilum_, there was 678 no overlap between the floral VOCs identified in our samples and the previously published 679 profiles. Altogether, these observations demonstrate that floral VOC profiles in _Cypripedium_ 680 can be highly variable within species, suggesting that population-specific adaptive forces 681 alongside genetic drift and gene flow shape floral VOC emission across a species' range. 682 Concurrently, the marked discrepancies between our findings and published data underscore a 683 key methodological limitation of single-individual sampling. Fully capturing the spectrum of 684 intraspecific chemical diversity across populations will require future studies to implement 685 broader sampling regimes incorporating multiple individuals per species. 686 **Case study in** _**C. henryi**_ **: post pollination effects on floral scent production** 687 _Floral scent of_ Cypripedium henryi 688 The floral scent of _C. henryi_ was found to be strongly dominated by aliphatic compounds, 689 which generally accounted for more than 90% of the total floral bouquet (Table 1). Analysis 690 of the 24 floral scent samples revealed 33 VOCs, a substantially greater number than the four 691 aliphatic volatiles (dodecanal, undecane, nonanal, and decanal) identified in the earlier study 22 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 692 by Peng Li _et al_ . (2008). Nonanal and decanal were not consistently identified as floral 693 volatiles in our study, as they did not reach the required threshold of being at least fivefold 694 more abundant in floral samples than in the corresponding leaf controls. Consistent with Peng 695 Li _et al_ . (2008), dodecanal and undecane emerged as the major components of the floral scent 696 in our samples. The fatty aldehyde dodecanal was by far the most abundant compound in the 697 floral fragrance of _C. henryi_ . The presence of this compound, along with the co-detected C 12 698 metabolic derivatives 1-dodecanol and dodecanoic acid, suggests a potential dual function 699 beyond pollinator attraction, as these molecules are known to possess significant 700 antimicrobial and antifungal activities (Togashi _et al.,_ 2007; Davis � Hanna _et al.,_ 2008; Hall 701 _et al.,_ 2011; Giovagnoni _et al.,_ 2022). Given that dodecanal was the heavily dominant 702 compound in both our analysed samples and published literature, its high emission levels are 703 unlikely to be linked to artifacts or stress responses induced by our experimental setup. 704 Ultimately, this dual functionality of floral VOCs likely represents an evolutionary adaptation 705 that simultaneously facilitates pollinator attraction and floral defence. Undecane was the 706 second most prevalent compound, and although it has shown to act as an attractant for various 707 types of insects (Y. Wang _et al.,_ 2017; X. Zhang _et al.,_ 2025) and more specifically as an alert 708 pheromone for ants (Hölldobler and Wilson 1990; Xu _et al.,_ 2023), its ecological role in 709 _Cypripedium_ pollination remains uncertain. The detection of terpenoid volatiles in the floral 710 fragrance of _C. henryi_ represents a new finding compared to previous studies. Specifically, 711 the compounds linalool, 4-oxoisophorone, and 2,4,4-trimethyl-3-carboxaldehyde-5-hydroxy 712 2,5-cyclohexadien-1-one were consistently present across all floral samples analysed. 713 Notably, the chemical composition of _C. henryi_ ’s floral scent differed greatly from that of its 714 sister taxon _C. segawai_, which lacked aliphatic compounds and was dominated by terpenoids. 715 _Effects of pollination on floral fragrance composition in_ C. henryi 716 Pollination treatment (pollinium removal and/or deposition) revealed only subtle effects on 717 the floral scent of _C. henryi_, with the inter-individual variation between the two sampled 718 plants contributing substantially to the observed variance (Figure 5). While we observed a 719 significant reduction in total scent emission following cross-pollination (i.e. deposition 720 treatment), the more pronounced 3.6-fold decrease observed in flowers of the 721 deposition+removal treatment did not reach statistical significance. This difference in 722 significance could very well be attributable to high inter-replicate variability rather than 723 antagonistic physiological effects caused by the pollinia removal. A reduction in total scent 724 emission represents the most widely documented post-pollination phenomenon in flowering 23 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 725 plants. This response has been described across various Orchidaceae, including _Platanthera_ 726 _bifolia_ (Tollsten 1993) and _Ophrys sphegodes_ (Schiestl _et al.,_ 1997), as well as in diverse taxa 727 such as _Cirsium_ spp (Theis and Raguso 2005) and _Silene latifolia_ (Dötterl _et al.,_ 2005; 728 Muhlemann _et al.,_ 2006). In _S. latifolia_, this downregulation is particularly robust, with scent 729 emission decreasing four-fold (Muhlemann _et al.,_ 2006) or even ceasing entirely (Dötterl _et_ 730 _al.,_ 2005). In contrast, our results for _C. henryi_ remain ambiguous due to high inter-replicate 731 variability and a general, although non-significant, decline in volatile emissions across all 732 groups. The general decrease in the absolute abundance of all major VOCs across all post 733 treatment samples suggests a broad senescence-related decline in _C. henryi_ that may mask 734 more subtle, specific post-pollination responses. Furthermore, we found no evidence of VOC 735 upregulation acting as a pollinator repellent, a mechanism previously identified in _O._ 736 _sphegodes_ (Schiestl and Ayasse 2001). One prevailing assumption suggests that post 737 pollination changes are only initiated once pollen tubes reach the ovary, a process taking less 738 than 48 hours in _Antirrhinum_ and _Petunia_ (Negre _et al.,_ 2003), but which may vary 739 significantly across species. While our results, alongside unpublished data from _C. calceolus_ 740 (S. Dötterl _et al_ ., unpublished data), might suggest that the genus _Cypripedium_ lacks strong 741 post-pollination scent cues, we caution against over-generalization. Post-pollination responses 742 can vary significantly between congeneric species as it was found for _Cirsium_ (Theis and 743 Raguso 2005). Finally, we acknowledge the limitations imposed by our small sample size. 744 Future studies with increased replicates, would be necessary to confirm if pollinated flowers 745 truly exhibit a more substantial drop in total scent than unpollinated controls. Furthermore, 746 given the late onset of scent production observed in _C. henryi_, we recommend delaying the 747 initial sampling period to 72 hours post-anthesis. 748 _Effects of pollination on floral volatile gene expression levels in_ C. henryi 749 Our results indicate that while transcriptional responses to pollination in _C. henryi_ are 750 detectable, they remain largely ambiguous. Expression profiles for FA and terpenoid 751 biosynthetic genes lacked clear treatment-specific signatures; no identified genes exhibited 752 uniform reactions across all tissue types, and almost none showed consistent patterns across 753 the two treatments involving pollinia deposition. The only exception to this last point was the 754 up-regulation of _PDH-E1-_ α in petals. Despite its role in providing Acetyl-CoA for FA 755 biosynthesis, its expression trajectory did not align with the volatile data. As FA-derived 756 volatiles decreased across all groups and showed no treatment effect, the increased expression 757 of _PDH-E1-_ α was likely counterbalanced by downstream regulatory mechanisms or metabolic 24 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 758 activity of other floral tissue. _BCCP1_1_ and _BADC_ displayed distinct localized responses 759 being downregulated primarily in petals and only in treatments with pollinia removal 760 (removal and deposition+removal). This pattern potentially reflects metabolic consequences 761 of mechanical disturbance triggered by pollinia removal, independent of pollination. The 762 parallel expression of these two genes is particularly notable given their antagonistic roles 763 within the ACCase complex, the primary gatekeeper of FA biosynthesis (Page _et al_ ., 1994; 764 Balbuena _et al_ ., 2025). While _BCCP1_ provides the essential biotin-carrying subunit to drive 765 the reaction, _BADC_ acts as a competitive inhibitor to suppress it (Salie _et al_ ., 2016). Their 766 synchronized downregulation could suggest a coordinated recalibration of the ACCase 767 complex in response to pollinia removal. Furthermore, the divergent regulation of a second 768 _BCCP1_ isoform, which exhibited an antagonistic expression pattern to the first, underscores 769 the sophisticated and multi-layered control of the ACCase complex in these tissues. _LACS6_ 770 encodes the enzyme responsible for activating free FAs into fatty acyl-CoA (Zhao _et al_ ., 771 2021), an obligatory step for their entry into peroxisomal β -oxidation pathway in which long 772 chain precursors are sequentially shortened to produce the aldehydes and alkanes that 773 constitute FA-derived volatiles (Schwab _et al_ ., 2008). In our study, _LACS6_ exhibited 774 consistent down-regulation in petals of flowers having undergone a treatment, pointing to a 775 coordinated, petal-specific shift in lipid metabolic balance. While a downregulation of _LACS6_ 776 offers a plausible molecular mechanism for a reduction in FA-derived volatiles (such as 777 dodecanal and undecane), this transcriptional signal observed in treated flowers was not 778 mirrored in the scent profiles. There was no statistical evidence nor any discernible trend in 779 the scent data suggesting that absolute levels of _LACS6_ -derived compounds were lower in 780 treated flowers compared to controls. This divergence suggests that the metabolic shift may 781 be more subtle than the transcriptional data implies, and may point to post-transcriptional 782 regulations or compensatory pathways that stabilize volatile emission despite localized gene 783 suppression. The only terpenoid-related gene identified through differential expression 784 analysis was _HMGCR3,_ which encodes the rate-limiting enzyme of the MVA pathway. This 785 gene exhibited the most complex regulatory profile with its upregulation across treatments 786 being highly tissue-specific (Figure 6). This upregulation is noteworthy since the MVA 787 pathway is primarily associated with the synthesis of sesquiterpenes (Rohmer 1999; 788 Lichtenthaler 1999), a class of compounds entirely absent from the floral scent of _C. henryi_ . 789 Consequently, these changes in MVA-pathway gene expression are probably linked to other 790 metabolic functions, including sterol production or defence-related compounds, rather than to 791 scent emission itself. 25 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 792 **Conclusion** 793 Our study uncovered a high chemical heterogeneity in the floral fragrances across the genus 794 _Cypripedium_, revealing a broad spectrum of scent bouquets that varied greatly in both the 795 number and identity of floral volatiles. The pronounced interspecific variation revealed here, 796 alongside existing volatile data, suggests that floral scent in this genus is a highly labile trait 797 capable of rapid evolutionary divergence. Within this diversity, we uncovered a significant 798 but weak phylogenetic signal in the floral scent profiles of the sampled dataset. On a broader 799 scale, pollinator groups significantly explained some of the variance in floral scent 800 composition, whereas taxonomic sections did not. This finding aligns with the genus's 801 primary strategy of generalized food deception, where the high lability of floral scents enables 802 these deceptive orchids to break away from ancestral patterns and optimize their chemical 803 profiles to exploit the pre-existing biases of local pollinator communities. The investigation 804 into _C. henryi_ provided limited evidence for a directed post-pollination effect on floral 805 fragrance, as sample variation was explained far more by individual plant clone than by the 806 pollination treatments. Instead, a general treatment-independent decline in floral volatiles 807 across all flowers suggested a reaction induced by floral aging. While total scent emission did 808 decrease significantly after cross-pollination only, this result should be interpreted with 809 caution given the low statistical power and the fact that the finding was not confirmed in the 810 treatment involving pollinia removal alongside cross-pollination. Furthermore, differential 811 expression analysis of candidate genes related to volatile biosynthesis aligned with the 812 volatile data in lacking clear treatment-specific signatures. 813 **List of Supplementary Data** 814 Table S1. Overview of sampled _Cypripedium_ taxa, sample sizes, material origin, sectional 815 classifications and used gas chromatography (GC) split ratios. 816 Table S2. Summary of documented _Cypripedium_ pollinators. 817 Table S3. Relative amounts (%) of floral volatiles detected in 31 taxa of _Cypripedium_ . 818 Table S4. Unidentified compounds in floral scent of _Cypripedium_ . 819 Table S5. Global summary of total significantly differentially expressed transcripts (padj < 820 0.05, ∣ log2FC ∣ >1) across all treatment comparisons in _C. henryi_ . 821 Table S6. Differentially expressed transcripts (DETs) of candidate genes involved in terpene 822 and fatty acid biosynthetic pathways in _C. henryi._ 823 **Funding** 26 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 824 This work was supported by the Carl Friedrich von Siemens Foundation. 825 **Conflict of interest** 826 No conflicts of interest to declare 827 **Author contributions** 828 Conceptualization: JBC, GK 829 Methodology: JBC, SP, SD, CS, EYH, GK 830 Resources and Supervision: GK, SP, SD 831 Formal analysis: JBC, SD, EYH, CS, SK 832 Investigation and Interpretation: JBC, SP, SD, CS, EYH, SK, GK 833 Writing - Original Draft: JBC, EYH, GK 834 Writing - Review & Editing: JBC, SP, SD, CS, EYH, SK, GK 835 **AI assistance acknowledgement** 836 AI driven language processing tools were used during the drafting of this manuscript to assist 837 with linguistic editing and the development of code for figure generation. 838 **Availability of Data and Materials** 839 RNA data generated within this study will be deposited in the European Nucleotide Archive 840 (ENA) at EMBL-EBI. The accession numbers will be provided prior to final publication. 841 **General acknowledgements** 842 We are grateful to Sebastian Urban and his team for their horticultural assistance and 843 maintenance of the _Cypripedium_ collection. 844 **References** 845 **Balbuena TS, Jorge GL, Da Silva Nascimento JR, Thelen JJ** . **2025** . Novel effector 846 proteins of the plant heteromeric acetyl-CoA carboxylase (D Allen, Ed.). _Journal of_ 847 _Experimental Botany_ . https://doi.org/10.1093/jxb/eraf509 848 **Bänziger H, Sun H, Luo Y-B** . **2008** . Pollination of wild lady slipper orchids _Cypripedium_ 849 _yunnanense_ and _C. flavum_ (Orchidaceae) in south-west China: why are there no hybrids? 850 _Botanical Journal of the Linnean Society_ **156** : 51–64. https://doi.org/10.1111/j.1095 851 8339.2007.00755.x 27 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 852 **Barkman TJ, Beaman JH, Gage DA** . **1997** . Floral fragrance variation in _Cypripedium_ : 853 Implications for evolutionary and ecological studies. _Phytochemistry_ **44** : 875–882. 854 https://doi.org/10.1016/S0031-9422(96)00614-0 855 **Bergstrom G, Birgersson G, Groth I, Anders Nilsson L** . **1992** . Floral fragrance disparity 856 between three taxa of lady’s slipper _Cypripedium calceolus_ (Orchidaceae). _Phytochemistry_ 857 **31** : 2315–2319. https://doi.org/10.1016/0031-9422(92)83271-Y 858 **Braunschmid H, Guilhot R, Dötterl S** . **2020** . Floral scent and pollinators of _Cypripedium_ 859 _calceolus_ L. at different latitudes. _Diversity_ **13** : 5. https://doi.org/10.3390/d13010005 860 **Braunschmid H, Mükisch B, Rupp T, et al., Dötterl S** . **2017** . Interpopulation variation in 861 pollinators and floral scent of the lady’s-slipper orchid _Cypripedium calceolus_ L. 862 _Arthropod-Plant Interactions_ **11** : 363–379. https://doi.org/10.1007/s11829-017-9512-x 863 **Cao L, Joffard N, Segar ST, ... Proffit M. 2026.** Phylogenetic history shapes the 864 composition of floral scents in a specialized pollination mutualism. _New Phytologist_ 250: 865 3428-3443. https://doi.org/10.1111/nph.71133 866 **Cordeiro GD, Dötterl S** . **2023** . Floral scents in bee-pollinated buckwheat and oilseed rape 867 under a global warming scenario. _Insects_ **14** : 242. https://doi.org/10.3390/insects14030242 868 **Cribb PJ** . **1997** . _The genus Cypripedium_ . London: Royal Botanic Gardens, Kew. 869 **Davis** � **Hanna A, Piispanen AE, Stateva LI, Hogan DA** . **2008** . Farnesol and dodecanol 870 effects on the _Candida albicans_ Ras1 � cAMP signalling pathway and the regulation of 871 morphogenesis. _Molecular Microbiology_ **67** : 47–62. https://doi.org/10.1111/j.1365 872 2958.2007.06013.x 873 **Dötterl S, Gershenzon J** . **2023** . Chemistry, biosynthesis and biology of floral volatiles: roles 874 in pollination and other functions. _Natural Product Reports_ **40** : 1901–1937. 875 https://doi.org/10.1039/D3NP00024A 876 **Dötterl S, Jürgens A** . **2005** . Spatial fragrance patterns in flowers of _Silene latifolia_ : lilac 877 compounds as olfactory nectar guides? _Plant Systematics and Evolution_ **255** : 99–109. 878 https://doi.org/10.1007/s00606-005-0344-2 28 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 879 **Dötterl S, Wolfe LM, Jürgens A** . **2005** . Qualitative and quantitative analyses of flower scent 880 in _Silene_ _latifolia_ . _Phytochemistry_ **66** : 203–213. 881 https://doi.org/10.1016/j.phytochem.2004.12.002 882 **Edens-Meier R, Arduser M, Westhus E, Bernhardt P** . **2011** . Pollination ecology of 883 _Cypripedium reginae_ Walter (Orchidaceae): size matters. _Telopea_ **13** : 327–340. 884 https://doi.org/10.7751/telopea20116024 885 **Etl C, Etl F, Guilhot R, Braunschmid H, Gross K, Dötterl S** . **2025** . Deceptive 886 _Cypripedium calceolus_ shares more floral scent compounds with co-flowering rewarding 887 species than those species share among each other. _Frontiers in Plant Science_ **16** : 888 1627890. https://doi.org/10.3389/fpls.2025.1627890 889 **Ewels P, Magnusson M, Lundin S, Käller M** . **2016** . MultiQC: summarize analysis results 890 for multiple tools and samples in a single report. _Bioinformatics_ **32** : 3047–3048. 891 https://doi.org/10.1093/bioinformatics/btw354 892 **Farré-Armengol G, Fernández-Martínez M, Filella I, Junker RR, Peñuelas J** . **2020** . 893 Deciphering the biotic and climatic factors that influence floral scents: a systematic review 894 of floral volatile emissions. _Frontiers_ _in_ _Plant_ _Science_ **11** : 1154. 895 https://doi.org/10.3389/fpls.2020.01154 896 **Farré-Armengol G, Filella I, Llusià J, Peñuelas J** . **2017** . β -Ocimene, a key floral and foliar 897 volatile involved in multiple interactions between plants and other organisms. _Molecules_ 898 **22** : 1148. https://doi.org/10.3390/molecules22071148 899 **Fatihah NHN, Fay MF, Maxted N** . **2011** . Molecular phylogenetics of _Cypripedium_ L. 900 (Orchidaceae) based on plastid and nuclear DNA sequences. _Journal of Agrobiotechnology_ 901 **2** : 35–51. 902 **Ferguson CS, Donham K** . **1999** . Pollinator of clustered lady’s slipper orchid _Cypripedium_ 903 _fasciculatum_ (Orchidaceae) in Oregon. _North American Native Orchid Journal_ **5** : 180–184. 904 **Frosch W, Cribb P** . **2012** . _Hardy_ Cypripedium _: species, hybrids and cultivation_ . Richmond: 905 Kew pub. 29 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 906 **Fu L, Niu B, Zhu Z, Wu S, Li W** . **2012** . CD-HIT: accelerated for clustering the next 907 generation sequencing data. _Bioinformatics_ **28** : 3150–3152. 908 https://doi.org/10.1093/bioinformatics/bts565 909 **Füssel U, Dötterl S, Jürgens A, Aas G** . **2007** . Inter- and intraspecific variation in floral scent 910 in the genus _Salix_ and its implication for pollination. _Journal of Chemical Ecology_ **33** : 911 749–765. https://doi.org/10.1007/s10886-007-9257-6 912 **Gerlach G, Schill R** . **1991** . Composition of orchid scents attracting euglossine bees. _Botanica_ 913 _Acta_ **104** : 379–384. https://doi.org/10.1111/j.1438-8677.1991.tb00245.x 914 **Giovagnoni G, Tugnoli B, Piva A, Grilli E** . **2022** . Dual antimicrobial effect of medium 915 chain fatty acids against an italian multidrug resistant _Brachyspira hyodysenteriae_ strain. 916 _Microorganisms_ **10** : 301. https://doi.org/10.3390/microorganisms10020301 917 **Gong W** � **C., Chen G, Vereecken NJ, Dunn BL, Ma Y** � **P., Sun W** � **B.** **2015** . Floral 918 scent composition predicts bee pollination system in five butterfly bush ( _Buddleja_, 919 Scrophulariaceae) species (A Dafni, Ed.). _Plant_ _Biology_ **17** : 245–255. 920 https://doi.org/10.1111/plb.12176 921 **Grabherr MG, Haas BJ, Yassour M, et al., Regev A** . **2011** . Full-length transcriptome 922 assembly from RNA-Seq data without a reference genome. _Nature Biotechnology_ **29** : 644– 923 652. https://doi.org/10.1038/nbt.1883 924 **Guignard J** . **1886** . Insects and orchids. _Annual Report of the Entomological Society of_ 925 _Ontario_ **16** : 39–48. 926 **Hall RA, Turner KJ, Chaloupka J, et al., Mühlschlegel FA** . **2011** . The quorum-sensing 927 molecules farnesol/homoserine lactone and dodecanol operate via distinct modes of action 928 in _Candida albicans_ . _Eukaryotic Cell_ **10** : 1034–1042. https://doi.org/10.1128/EC.05060-11 929 **Haverkamp A, Hansson BS, Baldwin IT, Knaden M, Yon F** . **2018** . Floral trait variations 930 among wild tobacco populations influence the foraging behavior of hawkmoth pollinators. 931 _Frontiers in Ecology and Evolution_ **6** : 19. https://doi.org/10.3389/fevo.2018.00019 30 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 932 **Hetherington-Rauth MC, Ramírez SR** . **2016** . Evolution and diversity of floral scent 933 chemistry in the euglossine bee-pollinated orchid genus _Gongora_ . _Annals of Botany_ **118** : 934 135–148. https://doi.org/10.1093/aob/mcw072 935 **Hölldobler B, Wilson EO** . **1990** . _The ants_ . Berlin, Heidelberg: Springer Berlin Heidelberg. 936 https://doi.org/10.1007/978-3-662-10306-7 937 **Jersáková J, Johnson SD, Kindlmann P** . **2006** . Mechanisms and evolution of deceptive 938 pollination in orchids. _Biological_ _Reviews_ **81** : 219–235. 939 https://doi.org/10.1017/S1464793105006986 940 **Jersáková J, Spaethe J, Streinzer M, et al., Johnson SD** . **2016** . Does _Traunsteinera_ 941 _globosa_ (the globe orchid) dupe its pollinators through generalized food deception or 942 mimicry? _Botanical_ _Journal_ _of_ _the_ _Linnean_ _Society_ **180** : 269–294. 943 https://doi.org/10.1111/boj.12364 944 **Jiang H, Kong J, Chen H, et al., Lee Y** . **2020** . _Cypripedium subtropicum_ (Orchidaceae) 945 employs aphid colony mimicry to attract hoverfly (Syrphidae) pollinators. _New Phytologist_ 946 **227** : 1213–1221. https://doi.org/10.1111/nph.16623 947 **Joffard N, Arnal V, Buatois B, Schatz B, Montgelard C** . **2020** . Floral scent evolution in the 948 section _Pseudophrys_ : pollinator � mediated selection or phylogenetic constraints? (A Dafni, 949 Ed.). _Plant Biology_ **22** : 881–889. https://doi.org/10.1111/plb.13104 950 **Knudsen JT, Tollsten L, Bergström LG. 1993.** Floral scents—a checklist of volatile 951 compounds isolated by head-space techniques. _Phytochemistry_ **33** : 253–280. 952 https://doi.org/10.1016/0031-9422(93)85502-I 953 **Knudsen JT** . **2002** . Variation in floral scent composition within and between populations of 954 _Geonoma macrostachys_ (Arecaceae) in the western Amazon. _American Journal of Botany_ 955 **89** : 1772–1778. https://doi.org/10.3732/ajb.89.11.1772 956 **Knudsen JT, Eriksson R, Gershenzon J, Ståhl B** . **2006** . Diversity and distribution of floral 957 scent. _The_ _Botanical_ _Review_ **72** : 1–120. https://doi.org/10.1663/0006 958 8101(2006)72[1:DADOFS]2.0.CO;2 31 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 959 **Kolde** **R** . **2010** . pheatmap: pretty heatmaps.: 1.0.13. 960 https://doi.org/10.32614/CRAN.package.pheatmap 961 **Krug C, Cordeiro GD, Schäffler I, et al., Alves-dos-Santos I** . **2018** . Nocturnal bee 962 pollinators are attracted to guarana flowers by their scents. _Frontiers in Plant Science_ **9** : 963 1072. https://doi.org/10.3389/fpls.2018.01072 964 **Lagou LJ, Kadereit G, Morales-Briones DF** . **2024** . Phylogenomic analysis of target 965 enrichment and transcriptome data uncovers rapid radiation and extensive hybridization in 966 the slipper orchid genus _Cypripedium_ . _Annals of Botany_ **134** : 1229–1250. 967 https://doi.org/10.1093/aob/mcae161 968 **Langmead B, Salzberg SL** . **2012** . Fast gapped-read alignment with Bowtie 2. _Nature_ 969 _Methods_ **9** : 357–359. https://doi.org/10.1038/nmeth.1923 970 **Laubertie EA, Wratten SD, Sedcole JR** . **2006** . The role of odour and visual cues in the 971 pan � trap catching of hoverflies (Diptera: Syrphidae). _Annals of Applied Biology_ **148** : 173– 972 178. https://doi.org/10.1111/j.1744-7348.2006.00046.x 973 **Levin RA, McDade LA, Raguso RA** . **2003** . The systematic utility of floral and vegetative 974 fragrance in two genera of Nyctaginaceae (R Mason-Gamer, Ed.). _Systematic Biology_ **52** : 975 334–351. https://doi.org/10.1080/10635150390196975 976 **Li P** . **2006** . Pollination biology of _Cypripedium_ (Orchidaceae) in Huanglong, Sichuan. 977 **Li J, Liu Z, Salazar GA, et al., Luo Y** . **2011** . Molecular phylogeny of _Cypripedium_ 978 (Orchidaceae: Cypripedioideae) inferred from multiple nuclear and chloroplast regions. 979 _Molecular_ _Phylogenetics_ _and_ _Evolution_ **61** : 308–320. 980 https://doi.org/10.1016/j.ympev.2011.06.006 981 **Li P, Luo Y., Bernhardt P, Kou Y, Perner H** . **2008a** . Pollination of _Cypripedium_ 982 _plectrochilum_ (Orchidaceae) by _Lasioglossum_ spp. (Halictidae): the roles of generalist 983 attractants versus restrictive floral architecture. _Plant Biology_ **10** : 220–230. 984 https://doi.org/10.1111/j.1438-8677.2007.00020.x 32 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 985 **Li P, Luo YB, Bernhardt P, Yang XQ, Kou Y** . **2006** . Deceptive pollination of the lady’s 986 slipper _Cypripedium tibeticum_ (Orchidaceae). _Plant Systematics and Evolution_ **262** : 53–63. 987 https://doi.org/10.1007/s00606-006-0456-3 988 **Li P, Luo YB, Deng Y-X, Kou Y** . **2008b** . Pollination of the lady’s slipper _Cypripedium_ 989 _henryi_ Rolfe (Orchidaceae). _Botanical Journal of the Linnean Society_ **156** : 491–499. 990 https://doi.org/10.1111/j.1095-8339.2007.00767.x 991 **Liao M, Zhang J-Y, Feng Y, Ren Z-X, Deng H-N, Xu B** . **2024** . Phylogenomic insights into 992 the historical biogeography, character-state evolution, and species diversification rates of 993 Cypripedioideae (Orchidaceae). _Molecular Phylogenetics and Evolution_ **199** : 108138. 994 https://doi.org/10.1016/j.ympev.2024.108138 995 **Lichtenthaler HK** . **1999** . The 1-deoxy-d-xylulose-5-phosphate pathway of isoprenoid 996 biosynthesis in plants. _Annual Review of Plant Physiology and Plant Molecular Biology_ 997 **50** : 47–65. https://doi.org/10.1146/annurev.arplant.50.1.47 998 **Love MI, Huber W, Anders S** . **2014** . Moderated estimation of fold change and dispersion 999 for RNA-seq data with DESeq2. _Genome Biology_ **15** : 550. https://doi.org/10.1186/s13059 1000 014-0550-8 1001 **Luo C, Huang ZY, Li K, Chen X, Chen Y, Sun Y** . **2013** . EAG responses of _Apis cerana_ to 1002 floral compounds of a biodiesel plant, _Jatropha curcas_ (Euphorbiaceae). _Journal of_ 1003 _Economic Entomology_ **106** : 1653–1658. https://doi.org/10.1603/EC12458 1004 **Manni M, Berkeley MR, Seppey M, Simão FA, Zdobnov EM** . **2021** . BUSCO update: 1005 novel and streamlined workflows along with broader and deeper phylogenetic coverage for 1006 scoring of eukaryotic, prokaryotic, and viral genomes (J Kelley, Ed.). _Molecular Biology_ 1007 _and Evolution_ **38** : 4647–4654. https://doi.org/10.1093/molbev/msab199 1008 **Milet-Pinheiro P, Gerlach G** . **2017** . Biology of the Neotropical orchid genus _Catasetum_ : A 1009 historical review on floral scent chemistry and pollinators. _Perspectives in Plant Ecology,_ 1010 _Evolution and Systematics_ **27** : 23–34. https://doi.org/10.1016/j.ppees.2017.05.004 1011 **Mishra A, Jain A, Iyer PS, Suryanarayanan A, Nordström K, Olsson SB** . **2025** . Innate 1012 floral object identification in a solitary pollinator employs a combination of both visual and 1013 olfactory cues. _The Science of Nature_ **112** : 18. https://doi.org/10.1007/s00114-025-01965-5 33 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1014 **Muhlemann JK, Waelti MO, Widmer A, Schiestl FP** . **2006** . Postpollination changes in 1015 floral odor in _Silene latifolia_ : adaptive mechanisms for seed-predator avoidance? _Journal_ 1016 _of Chemical Ecology_ **32** : 1855–1860. https://doi.org/10.1007/s10886-006-9113-0 1017 **Negre F, Kish CM, Boatright J, et al., Dudareva N** . **2003** . Regulation of methylbenzoate 1018 emission after pollination in snapdragon and petunia flowers. _The Plant Cell_ **15** : 2992– 1019 3006. https://doi.org/10.1105/tpc.016766 1020 **Nilsson LA** . **1979** . Anthecological studies on the lady’s slipper, _Cypripedium calceolus_ 1021 (Orchidaceae). _Botaniska Notiser_ **132** : 329–347. 1022 **Oksanen J, Simpson G, Blanchet F, et al., Weedon J 2022.** _vegan_ : community ecology 1023 package. R package version 2.6-4, https://CRAN.R-project.org/package=vegan. 1024 **Page RA, Okada S, Harwood JL** . **1994** . Acetyl-CoA carboxylase exerts strong flux control 1025 over lipid synthesis in plants. _Biochimica et Biophysica Acta (BBA) - Lipids and Lipid_ 1026 _Metabolism_ **1210** : 369–372. https://doi.org/10.1016/0005-2760(94)90242-9 1027 **Patro R, Duggal G, Love MI, Irizarry RA, Kingsford C** . **2017** . Salmon provides fast and 1028 bias-aware quantification of transcript expression. _Nature Methods_ **14** : 417–419. 1029 https://doi.org/10.1038/nmeth.4197 1030 **Pemberton RW** . **2013** . Pollination of slipper orchids (Cypripedioideae): a review. 1031 _Lankesteriana_ . https://doi.org/10.15517/lank.v0i0.11539 1032 **Petrén H, Köllner TG, Junker RR** . **2023** . Quantifying chemodiversity considering 1033 biochemical and structural properties of compounds with the R package CHEMODIV . _New_ 1034 _Phytologist_ **237** : 2478–2492. https://doi.org/10.1111/nph.18685 1035 **Piñero JC, Barrett BA, Bolton LG, Follett PA** . **2019** . β -cyclocitral synergizes the response 1036 of adult _Drosophila suzukii_ (Diptera: Drosophilidae) to fruit juices and isoamyl acetate in a 1037 sex-dependent manner. _Scientific Reports_ **9** : 10574. https://doi.org/10.1038/s41598-019 1038 47081-z 1039 **Prieto** � **Benítez S, Millanes AM, Dötterl S, Giménez** � **Benavides L** . **2016** . Comparative 1040 analyses of flower scent in _Sileneae_ reveal a contrasting phylogenetic signal between night 1041 and day emissions. _Ecology_ _and_ _Evolution_ **6** : 7869–7881. 1042 https://doi.org/10.1002/ece3.2377 34 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1043 **Primante C, Dötterl S** . **2010** . A syrphid fly uses olfactory cues to find a non-yellow flower. 1044 _Journal of Chemical Ecology_ **36** : 1207–1210. https://doi.org/10.1007/s10886-010-9871-6 1045 **Proctor HC, Harder LD** . **1995** . Effect of pollination success on floral longevity in the orchid 1046 _Calypso bulbosa_ (Orchidaceae). _American Journal of Botany_ **82** : 1131–1136. 1047 https://doi.org/10.1002/j.1537-2197.1995.tb11584.x 1048 **R Core Team** . **2025** . R: a language and environment for statistical computing. 1049 **Ramya M, Jang S, An H-R, Lee S-Y, Park P-M, Park PH** . **2020** . Volatile organic 1050 compounds from orchids: from synthesis and function to gene regulation. _International_ 1051 _Journal of Molecular Sciences_ **21** : 1160. https://doi.org/10.3390/ijms21031160 1052 **Revadi SV, Giannuzzi VA, Vetukuri RR, Walker WB, Becher PG. 2021.** Larval response 1053 to frass and guaiacol: detection of an attractant produced by bacteria from _Spodoptera_ 1054 _littoralis_ frass. _Journal of Pest Science_ **94** : 1105–1118. https://doi.org/10.1007/s10340 1055 021-01352-9 1056 **Rodriguez-Saona C, Kaplan I, Braasch J, Chinnasamy D, Williams L** . **2011** . Field 1057 responses of predaceous arthropods to methyl salicylate: A meta-analysis and case study in 1058 cranberries. _Biological_ _Control_ **59** : 294–303. 1059 https://doi.org/10.1016/j.biocontrol.2011.06.017 1060 **Rohmer M** . **1999** . The discovery of a mevalonate-independent pathway for isoprenoid 1061 biosynthesis in bacteria, algae and higher plants. _Natural Product Reports_ **16** : 565–574. 1062 https://doi.org/10.1039/a709175c 1063 **Rupp T, Oelschlägel B, Berjano R, et al., Dötterl S** . **2024** . Chemical imitation of yeast 1064 fermentation by the drosophilid-pollinated deceptive trap-flower _Aristolochia baetica_ 1065 (Aristolochiaceae). _Phytochemistry_ **224** : 114142. 1066 https://doi.org/10.1016/j.phytochem.2024.114142 1067 **Salamanca J, Souza B, Kyryczenko-Roth V, Rodriguez-Saona C** . **2019** . Methyl salicylate 1068 increases attraction and function of beneficial arthropods in cranberries. _Insects_ **10** : 423. 1069 https://doi.org/10.3390/insects10120423 35 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1070 **Salie MJ, Zhang N, Lancikova V, Xu D, Thelen JJ** . **2016** . A family of negative regulators 1071 targets the committed step of de novo fatty acid biosynthesis. _The Plant Cell_ **28** : 2312– 1072 2325. https://doi.org/10.1105/tpc.16.00317 1073 **Scaccabarozzi D, Cozzolino S, Guzzetti L, et al., Phillips RD** . **2018** . Masquerading as pea 1074 plants: behavioural and morphological evidence for mimicry of multiple models in an 1075 Australian orchid. _Annals of Botany_ . https://doi.org/10.1093/aob/mcy166 1076 **Schiestl FP, Ayasse M** . **2001** . Post-pollination emission of a repellent compound in a 1077 sexually deceptive orchid: a new mechanism for maximising reproductive success? 1078 _Oecologia_ **126** : 531–534. https://doi.org/10.1007/s004420000552 1079 **Schiestl FP, Ayasse M, Paulus HF, Erdmann D, Francke W** . **1997** . Variation of floral 1080 scent emission and postpollination changes in individual flowers of _Ophrys sphegodes_ 1081 subsp. _sphegodes_ . _Journal_ _of_ _Chemical_ _Ecology_ **23** : 2881–2895. 1082 https://doi.org/10.1023/A:1022527430163 1083 **Schlüter PM, Xu S, Gagliardini V, et al., Schiestl FP** . **2011** . Stearoyl-acyl carrier protein 1084 desaturases are associated with floral isolation in sexually deceptive orchids. _Proceedings_ 1085 _of_ _the_ _National_ _Academy_ _of_ _Sciences_ **108** : 5696–5701. 1086 https://doi.org/10.1073/pnas.1013313108 1087 **Schwab W, Davidovich** � **Rikanati R, Lewinsohn E. 2008.** Biosynthesis of plant � derived 1088 flavor compounds. _The Plant Journal_ **54** : 712–732. https://doi.org/10.1111/j.1365 1089 313X.2008.03446.x 1090 **Schwacke R, Ponce-Soto GY, Krause K, et al., Usadel B** . **2019** . MapMan4: a refined 1091 protein classification and annotation framework applicable to multi-omics data analysis. 1092 _Molecular Plant_ **12** : 879–892. https://doi.org/10.1016/j.molp.2019.01.003 1093 **Shi J, Luo Y** � **B., Bernhardt P, Ran J** � **C., Liu Z** � **J., Zhou Q** . **2009** . Pollination by 1094 deceit in _Paphiopedilum barbigerum_ (Orchidaceae): a staminode exploits the innate colour 1095 preferences of hoverflies (Syrphidae). _Plant_ _Biology_ **11** : 17–28. 1096 https://doi.org/10.1111/j.1438-8677.2008.00120.x 1097 **Shimada T, Endo T, Fujii H, Rodríguez A, Peña L, Omura M** . **2014** . Characterization of 1098 three linalool synthase genes from _Citrus unshiu_ Marc. and analysis of linalool-mediated 36 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1099 resistance against _Xanthomonas citri_ subsp. _citri_ and _Penicilium italicum_ in citrus leaves 1100 and fruits. _Plant Science_ **229** : 154–166. https://doi.org/10.1016/j.plantsci.2014.09.008 1101 **Siadjeu C, Kadereit G** . **2024** . C4-like _Sesuvium sesuvioides_ (Aizoaceae) exhibits CAM in 1102 cotyledons and putative C4-like + CAM metabolism in adult leaves as revealed by 1103 transcriptome analysis. _BMC Genomics_ **25** : 688. https://doi.org/10.1186/s12864-024 1104 10553-2 1105 **Song L, Florea L** . **2015** . Rcorrector: efficient and accurate error correction for Illumina 1106 RNA-seq reads. _Gigascience_ **4** : s13742-015-0089-y. https://doi.org/10.1186/s13742-015 1107 0089-y 1108 **Song B, Niu Y, Stocklin J, et al., Sun H** . **2015** . Pollinator attraction in _Cornus capitata_ 1109 (Cornaceae): the relative role of visual and olfactory cues. _Journal of Plant Ecology_ **8** : 1110 173–181. https://doi.org/10.1093/jpe/rtv012 1111 **Steiner KE, Kaiser R, Dötterl S** . **2011** . Strong phylogenetic effects on floral scent variation 1112 of oil � secreting orchids in South Africa. _American Journal of Botany_ **98** : 1663–1679. 1113 https://doi.org/10.3732/ajb.1100141 1114 **Stoutamire WP** . **1967** . Flower biology of the lady’s slipper. _The Michigan Botanist_ : 159– 1115 175. 1116 **Sugiura N** . **2002** . Bumblebee pollination of _Cypripedium macranthos_ var. _rebunense_ 1117 (Orchidaceae); a possible case of floral mimicry of _Pedicularis schistostegia_ 1118 (Orobanchaceae). _Plant Systematics and Evolution_ **235** : 189–195. 1119 **Szlachetko DL, Górniak M, Kowalkowska AK, Kolanowska M, Jurczak-Kurek A,** 1120 **Archila Morales F** . **2021** . The natural history of the genus _Cypripedium_ (Orchidaceae). 1121 _Plant Biosystems - An International Journal Dealing with all Aspects of Plant Biology_ **155** : 1122 772–796. https://doi.org/10.1080/11263504.2020.1785963 1123 **Tahir AA, Mohd Barnoh NF, Yusof N, et al., Hara H. 2019.** Microbial Diversity in 1124 Decaying Oil Palm Empty Fruit Bunches (OPEFB) and Isolation of Lignin-degrading 1125 Bacteria from a Tropical Environment. _Microbes and Environments_ **34** : 161–168. 1126 https://doi.org/10.1264/jsme2.ME18117 37 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1127 **Theis N, Raguso RA** . **2005** . The effect of pollination on floral fragrance in thistles. _Journal_ 1128 _of Chemical Ecology_ **31** : 2581–2600. https://doi.org/10.1007/s10886-005-7615-9 1129 **Togashi N, Shiraishi A, Nishizaka M, et al., Inoue Y** . **2007** . Antibacterial activity of long 1130 chain fatty alcohols against _Staphylococcus aureus_ . _Molecules_ **12** : 139–148. 1131 https://doi.org/10.3390/12020139 1132 **Tollsten L** . **1993** . A multivariate approach to post � pollination changes in the floral scent of 1133 _Platanthera_ _bifolia_ (Orchidaceae). _Nordic_ _Journal_ _of_ _Botany_ **13** : 495–499. 1134 https://doi.org/10.1111/j.1756-1051.1993.tb00088.x 1135 **Vogt CA** . **1990** . Pollination in _Cypripedium reginae_ (Orchidaceae). _Lindleyana_ **5** : 145–150. 1136 **Wang Y, Chen Q, Guo J, et al., Ren B** . **2017** . Molecular basis of peripheral olfactory 1137 sensing during oviposition in the behavior of the parasitic wasp _Anastatus japonicus_ . _Insect_ 1138 _Biochemistry_ _and_ _Molecular_ _Biology_ **89** : 58–70. 1139 https://doi.org/10.1016/j.ibmb.2017.09.001 1140 **Wang Bing, Dong W, Li H, et al., Wang G** . **2022** . Molecular basis of (E)- β -farnesene 1141 mediated aphid location in the predator _Eupeodes corollae_ . _Current Biology_ **32** : 951 1142 962.e7. https://doi.org/10.1016/j.cub.2021.12.054 1143 **Weiss MR** . **1995** . Floral color change: a widespread functional convergence. _American_ 1144 _Journal of Botany_ **82** : 167–185. https://doi.org/10.1002/j.1537-2197.1995.tb11486.x 1145 **Wickham H** . **2016** . _ggplot2: Elegant Graphics for Data Analysis_ . Springer-Verlag New 1146 York. 1147 **Williams NH, Whitten WM** . **1983** . Orchid floral fragrances and male euglossine bees: 1148 methods and advances in the last sesquidecade. _The Biological Bulletin_ **164** : 355–395. 1149 https://doi.org/10.2307/1541248 1150 **Williams NH, Whitten WM** . **1999** . Molecular phylogeny and floral fragrances of male 1151 euglossine bee � pollinated orchids: A study of _Stanhopea_ (Orchidaceae). _Plant Species_ 1152 _Biology_ **14** : 129–136. https://doi.org/10.1046/j.1442-1984.1999.00016.x 38 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1153 **Xu W, Zhao M, Tang L, Ma R, He H** . **2023** . Chemical components of dufour’s and venom 1154 glands in _Camponotus japonicus_ (Hymenoptera, Formicidae). _Insects_ **14** : 664. 1155 https://doi.org/10.3390/insects14070664 1156 **Zhang L, Su Q, Wang L, Lv M, Hou Y, Li S. 2023.** Linalool: A ubiquitous floral volatile 1157 mediating the communication between plants and insects. _Journal of Systematics and_ 1158 _Evolution_ **61** : 538–549. https://doi.org/10.1111/jse.12930 1159 **Zhang X, Han W, Zhang F, Wang J, Liu S, Wang X** . **2025** . Attraction of _Nicotiana_ 1160 _benthamiana_ to _Bemisia tabaci_ is related to a chemical signal in plant volatile, undecane. 1161 _Pest Management Science_ **81** : 7359–7369. https://doi.org/10.1002/ps.8411 1162 **Zhao H, Kosma DK, Lü S** . **2021** . Functional role of long-chain acyl-CoA synthetases in 1163 plant development and stress responses. _Frontiers in Plant Science_ **12** : 640996. 1164 https://doi.org/10.3389/fpls.2021.640996 1165 **Zheng G, Li P, Pemberton R, Luo Y** . **2011** . Mixed bumblebee and blowfly pollination of 1166 _Cypripedium flavum_ (Orchidaceae) in Sichuan, China. _Ecological Research_ **26** : 453–459. 1167 https://doi.org/10.1007/s11284-010-0798-8 1168 1169 **Table 1. Mean relative amount of** _**C. henryi**_ **floral volatiles per treatment group before** 1170 **and after pollination treatment application.** **Control:** no pollination; **removal:** pollinia 1171 removal; **deposition:** pollination; **dep. + rem.:** pollination and pollinia removal. The standard 1172 error is indicated for values higher than one. Arrows indicate statistically significant 1173 differences in relative amounts before and after the pollination treatment (paired t-test p < 1174 0.05), prior to the Bonferroni correction. RI Compound Control Removal Deposition Dep.+ rem. (n=2) (n=2) (n=3) (n=3) before after before after before after before after Aliphatic compounds 968 Hexanoic acid 8 ± 4 7 ± 0.1 7 ± 2 8 ± 2 4 ± 0.3 2 ± 0.3 ↓ 7 ± 3 7 ± 3 985 ( _E_ )-3-Hexenoic acid - <1 - - <1 - ↓ <1 1065 Heptanoic acid - <1 - - <1 - <1 1100 Undecane 13 ± 4 18 ± 8 7 ± 2 17 ± 5 7 ± 0.3 19 ± 1 10 ± 2 18 ± 3 1105 Nonanal - - - - 4 ± 2 - 2 ± 2 1207 Decanal 3 ± 3 - - - 5 ± 1 - ↓ 3 ± 2 1272 1-Decanol <1 <1 <1 <1 <1 <1 <1 <1 1288 6-Tridecene <1 <1 ↑ <1 <1 <1 <1 <1 <1 1300 Tridecane 1 ± 0.6 1 ± 0.3 <1 <1 1 ± 0.1 3 ± 0.3 1 ± 0,2 2 ± 0.4 (n=2) (n=2) (n=3) 39 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) RI Compound Control Removal Deposition Dep.+ rem. (n=2) (n=2) (n=3) (n=3) before after before after before after before after 1309 Undecanal <1 1 ± 1 <1 - <1 2 ± 1 2 ± 1 1358 Decanoic acid - <1 - - - <1 <1 1412 Dodecanal 66 ± 3 65 ± 6 74 ± 3 67 ± 7 66 ± 2 69 ± 2 59 ± 0.1 66 ± 5 1471 ( _E_ )-2-Dodecenal <1 <1 <1 <1 <1 <1 <1 <1 1475 1-Dodecanol <1 1 ± 0.1 <1 <1 <1 <1 <1 <1 1485 Tetradecatriene isomer <1 - - - <1 <1 <1 1514 Tridecanal <1 <1 <1 <1 <1 <1 <1 <1 ↑ 1530 (2 _E_,4 _E_ )-2,4- <1 - <1 <1 <1 - <1 <1 Dodecadienal 1556 Dodecanoic acid <1 <1 <1 <1 <1 <1 <1 <1 1616 Tetradecanal 1 ± 0.4 1 ± 0.3 1 ± 0.8 <1 1 ± 0.2 <1 <1 <1 **TOTAL** **94.46** **89.20** **92.31** **95.26** **92.13** **98.28** **87.81** **95.83** Terpenoids 1045 Lavender lactone <1 - - <1 <1 <1 <1 <1 1101 Linalool 4 ± 1 3 ±3 4 ± 1 2 ± 2 2 ± 0.1 - ↓ 5 ± 2 2 ± 2 1150 4-Oxoisophorone <1 1 ± 0.8 <1 2 ± 2 <1 <1 <1 1 ± 1 1179 ( _Z_ / _E_ )-Linalool oxide - <1 - - <1 <1 ↑ <1 <1 pyranoid (n=2) (n=2) (n=3) <1 - <1 <1 <1 - <1 <1 1403 2,4,4-trimethyl-3carboxaldehyde-5hydroxy-2,5cyclohexadien-1-one <1 <1 <1 <1 4 ± 1 <1 ↓ 2 ± 1 <1 **TOTAL** **5.06** **4.49** **4.92** **4.52** **7.02** **1.41** **7.75** **3.91** Aromatic compounds 1073 Acetophenone - 5 ± 5 3 ± 3 - - - 3 ± 3 1078 Methyl benzoylformate - <1 - - - - <1 1120 2-Phenylethanol <1 <1 <1 - <1 <1 ↑ <1 <1 1670 cf 1,4-Dimethylindanyl acetate <1 <1 <1 - <1 - ↓ <1 **TOTAL** <1 **6.06** **2.67** **-** **0.35** **0.05** **3.97** **0.04** **TOTAL identified** **compounds** 1175 **List of figure legends** **99.62** **99.81** **99.84** **99.77** **99.51** **99.74** **99.54** **99.78** 1176 **Figure 1. Section-level Cladogram of the genus** _**Cypripedium**_ **.** The cladogram is adapted 1177 from the phylogenetic analysis of Lagou et al. (2024), with sectional classification following 1178 Frosch and Cribb (2012). Numbers in parenthesis indicate the total number of species 1179 currently recognized within each section according to Plants of the World Online (POWO). 1180 Sections covered in this study are highlighted in bold. Representative flower morphology for 1181 each section is illustrated to the right of the section name: _C. parviflorum_ var. _pubescens_ f. 1182 _planipetalum_ (sect. _Cypripedium_ ), _C. guttatum_ (sect. _Bifolia_ ), _C. flavum_ (sect. _Obtusipetala_ ), 1183 _C. lentiginosum_ (sect. _Trigonopedia_ ), _C. micranthum_ (sect. _Sinopedilum_ ), _C. acaule_ (sect. 1184 _Acaulia_ ), _C. californicum_ (sect. _Californica_ ), _C. palangshanense_ (sect. _Enantiopedilum_ ), _C._ 1185 _japonicum_ (sect. _Flabellinervia_ ), _C. subtropicum_ (sect. _Subtropica_ ), _C. plectrochilum_ (sect. 1186 _Arietinum_ ), _C. debile_ (sect. _Retinervia_ ), _C. irapeanum_ (sect. _Irapeana_ ). Photo credits (from 40 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1187 top to bottom): Forest Service of the Northern Region USA, L. Chen, J-B. Chazalon, S. 1188 Urban, V. Steindl, V. Steindl, S. Urban, S. Urban, L. Chen, W. Frosch, J-B. Chazalon, S. 1189 Urban, M. Béhar. 1190 **Figure 2. Floral volatile occurrence across sampled** _**Cypripedium**_ **spp.** The blue colour 1191 gradient represents the relative proportion of specific floral VOCs within each species scent 1192 profile. For better visualization the dataset was reduced to only include floral VOCs whose 1193 relative quantity across samples exceeded 8%. 1194 **Figure 3. Mean relative abundance (%) of floral volatile compound classes across** 1195 **sampled** _**Cypripedium**_ **sections.** The number of taxa sampled within each section is provided 1196 in parentheses. Values above the bars represent the rounded mean number of total floral 1197 volatiles identified per taxon for that respective section. Representative flowers for each 1198 taxonomic section are shown above the bars: _C. plectrochilum_ (sect. _Arietinum_ ), _C. guttatum_ 1199 (sect. _Bifolia_ ), _C. californicum_ (sect. _Californica_ ), _C. parviflorum_ var. _pubescens_ (sect. 1200 _Cypripedium_ ), _C. japonicum_ (sect. _Flabellinervia_ ), _C. flavum_ (sect. _Obtusipetala_ ) and _C._ 1201 _micranthum_ (sect. _Sinopedilum_ ). See Fig. 1 for credits of the flower images. 1202 **Figure 4.** **Non-metric multidimensional scaling (NMDS) showing the differences in** 1203 **relative amounts of floral scent components among the 31 sampled** _**Cypripedium**_ **taxa.** 1204 Samples are labelled with abbreviations corresponding to their respective taxa. 1205 **Figure 5.** **NMDS analysis visualizing the semi-quantitative floral scent data obtained** 1206 **from floral scent samples of** _**C. henryi**_ **before and after pollination treatments. Control:** 1207 no pollination treatment applied; **Removal:** pollinia removal; **Deposition:** pollination; **Dep.+** 1208 **Rem.:** pollination and pollinia removal. Open symbols indicate samples collected prior to 1209 treatment, whereas filled symbols represent post-treatment samples. Numbers indicate the 1210 plant ID of the two sampled plants. 1211 **Figure 6.** **Heatmap showing differential expression of terpenoid and fatty acid** 1212 **metabolism genes of** _**C. henryi**_ **in treatments compared to control.** Colour intensities 1213 represent log 2 FC, with red indicating upregulation and blue indicating downregulation. 1214 _HMGCR3_ : 3-hydroxy-3-methylglutaryl-coenzyme A reductase 3; _RABGAP22_ : Rab GTPase 1215 activating protein 22; _ICL_ : Isocitrate Lyase; _ACX3_ : Acyl-coenzyme A oxidase 3; _KCS20_ : 3 1216 ketoacyl-CoA synthase 20; _ENR_ : Enoyl-[acyl-carrier-protein] reductase; _ACX1_ : Acyl 1217 coenzyme A oxidase 1; _GlyMFP-_ α : Glyoxysomal fatty acid beta-oxidation multifunctional 41 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1218 protein MFP- α ; _PDH-E1-_ α _3_ : Pyruvate dehydrogenase E1 component subunit alpha-3; 1219 _BCCP1_ : Biotin carboxylase 1; _PMPCB_ : Mitochondrial-processing peptidase subunit beta; 1220 _KCS2_ : 3-ketoacyl-CoA synthase 2; _ZC3H55_ : Zinc finger CCCH domain-containing protein 1221 55; _TNT1-94 Pol_ : Retrovirus-related Pol polyprotein from transposon TNT 1-94; _BADC_ : 1222 Biotin lipoyl attachment domain-containing protein; _FAD4_ : Fatty acid desaturase 4; _PKA_ : 1223 Pyruvate kinase isozyme A; _LACS6_ : Long chain Acyl-CoA synthase 6. 1224 1225 1226 Figure 1 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1227 1228 1229 1230 Figure 2 1231 1232 1233 1234 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1235 1236 1237 Figure 3 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1238 1239 Figure 4 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1240 1241 Figure 5 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.760698; this version posted November 4, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.11.04.760698) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-ND 4.0 International license.](http://creativecommons.org/licenses/by-nd/4.0/) 1242 1243 1244 Figure 6 1245 1246 1247 1248 1249 1250 1251 1252 1253 _**Cypripedium**_ _(21)_ _**Bifolia**_ _(2)_ _**Obtusipetala**_ _(3)_ _Trigonopedia_ _(6)_ _**Sinopedilum**_ _(3)_ _Acaulia_ _(1)_ _**Californica**_ _(1)_ _Enantiopedilum_ _(2)_ _**Flabellinervia**_ _(2)_ _Subtropica_ _(2)_ _**Arietinum**_ _(2)_ _Retinervia_ _(2)_ _Irapeana_ _(7)_ Bees Drosophilids Bees & floricolous flies ### _Pass_ # **Section** ## Arietinum Bifolia Californica Cypripedium Flabellinervia Obtusipetala Sinopedilum ### _Micr_ ### _Cal_ ### _Flav_ _Hen_ _Ken_ ### _Form Ven_ ### _Fro_ _Macr_hot_ ### _Macr_reb_ # **Stress = 0.143** ### _Farr_ ### _Fascio_ ### _Jap_ _Gut_ ### _Bard_ _Reg_ ### _Cord_ ### _Fran_ ### _Macr_ _Yata_ ### _Plec_ ### _Seg_ ### _Tib_ ### _Par_mak_ ### _Par_pub_plan_ ### _Alask_ ### _Parv_pub_pub_ ### _Cali_ _Col_ _Yun_ 0.2 0.1 0.0 -0.1 Removal (before) Deposition (before) Dep.+ Rem. (before) **Treatment groups** Control (before) Control (after) Removal (after) Deposition (after) Dep.+ Rem. (after) -0.2 -0.1 0.0 0.1 0.2 0.3 NMDS1