Taxonomic description and pollination ecology of Cyrtorchis okuensis (Orchidaceae, Angraecinae), a new species endemic to the Cameroon Volcanic Line

Background and aims – A recent taxonomic and molecular study of the genus Cyrtorchis revealed three new species. Here, we describe one of these new species, endemic to the Cameroon Volcanic Line (CVL), and compare it to its closest relatives. Its conservation status is provided along with detailed information on its ecology and reproductive biology. Material and methods – A detailed examination of 21 herbarium specimens was performed using standard practices of herbarium taxonomy. The conservation status of the new species was assessed using the IUCN Red List Categories and Criteria. Furthermore, we investigated the reproductive biology of the new species in an easily accessible subpopulation in Cameroon and tested the hypothesis of hawkmoth pollination in the genus Cyrtorchis . Key results – The newly discovered species, here named C. okuensis , is restricted to the montane forest (1600 to 2500 m elevation) of the CVL and is currently assessed as Near Threatened (NT). Among angraecoid orchids in which hawkmoth pollination is recorded (~16 species), C. okuensis has the shortest nectar spur (19–40 mm). In situ observations revealed hawkmoths with short proboscis of the genus Hippotion ( H. celerio , H. eson , and H. osiris ) as the main pollinators. The rate of pollination and fruiting under natural conditions was low and among the 448 surveyed flowers only 38.8% (174/448 flowers) had their pollinia removed, 19% produced fruits (87/448 flowers), and 16% (70/448 flowers) set mature fruits. Conclusion – Cyrtorchis okuensis is clearly differentiated from C. submontana by its longer and wider floral bracts, longer anther caps, and wider triangular dorsal sepal.


INTRODUCTION
A recent study combining morphological and molecular data on the African angraecoid genus Cyrtorchis tested the monophyly of its two sections and re-assessed the circumscription of each described species (Azandi et al. 2021). This study revealed three novelties, and among these, one is endemic to the Cameroon Volcanic Line (CVL). Some of the specimens ascribable to this taxon have been wrongly identified as C. guillaumetii (Pérez-Vera) R.Rice, C. brownii (Rolfe) Schltr., or C. submontana Stévart, Droissart & Azandi because of their similarities in leaf shape and size and in inflorescence and spur lengths.
African angraecoid orchids include a high proportion of species characterised by white, long-spurred, nectariferous, and night-fragant flowers, features which are typical of hawkmoth pollination or sphingophily (Martins & Johnson 2007;Micheneau et al. 2009). These pollination syndromes or combination of floral traits are known to have independently evolved to attract hawkmoths (Ashworth et al. 2015). In tropical Africa, several studies on this pollination syndrome have been performed in its eastern part (see Martins & Johnson 2007, 2013, its southern part (see Luyt & Johnson 2001;Johnson & Raguso 2016;Johnson et al. 2017), and in Madagascar (see Nilsson et al. 1985Nilsson et al. , 1987Nilsson et al. , 1992Nilsson 1992Nilsson , 1998Wasserthal 1997); yet such studies are missing for Central Africa. Most of these previous studies on angraecoid orchids pollination have demonstrated that the long nectar spurs of orchids and the proboscis of pollinators are involved in evolutionary processes resulting in the remarkable angraecoid floral specialization. The reproductive success of these orchids is known to be pollinator-mediated and generally results in low fruit set due to low visiting frequency of hawkmoths (Micheneau et al. 2009;Amorim et al. 2014). In-depth knowledge about the pollination ecology of threatened species would enable the identification and description of the factors limiting their reproduction and allow proposing adequate conservation measures for both plants and their pollinators (Senapathi et al. 2015).
During our field trips in Cameroon, we discovered a large and accessible subpopulation of the new species in Oku National Park (North-West Region). That subpopulation provided us with the opportunity to test Cribb's (1989) suggestion that the considerable resemblance of the flowers in the genus Cyrtorchis derives from a common hawkmoth pollination syndrome (i.e. sphingophily), a hypothesis that remains unverified to date.
The present contribution intends to describe the new Cyrtorchis species revealed by the recent study combining morphological and phylogenetic data on Cyrtorchis spp. and to describe, for the first time in Central Africa, aspects of reproductive biology of an angraecoid orchid genus.

Taxonomy
Material and morphological examination -Taxonomic work is based on a detailed examination of 21 specimens of the novelty, which have been collected in different locations in Cameroon, Bioko (Equatorial Guinea), and Nigeria, 15 of which were collected since 1997 by our team (see Stévart et al. 2020). Specimens include 10 dried and 11 spirit-preserved samples and are deposited in the following herbaria: BRLU, FHI, K, MA, P, WAG, and YA (acronyms according to Thiers continuously updated). For stereomicroscopic observations, flowers from dried materials were initially boiled in water. Morphological investigations were performed using a Zeiss stereomicroscope Stemi SV11, both for spirit-preserved and herbarium materials. Morphology of seeds was observed under a light microscope Olympus BX51; quantitative parameters (length, width, and shape of viable seeds and embryos, and number of testa cells) were measured in μm with the image acquisition and processing software ArchiMed microvision v.6.1.3. Measurements, colours, and other details given in the description below are based on living material, spirit-preserved, herbarium specimens, and data derived from field notes. Description, geographical distribution, and conservation status assessment -The terminology used to describe structures and shapes follows the Systematics Association Committee for Descriptive Biological Terminology (1962a, 1962b) and the Kew Plant Glossary (Beentje 2010). The geographical distribution was determined from data given on the herbarium sheets. The distribution map was prepared using the software ArcGIS v.10.5.1 (ESRI 2017). The risk of extinction of the novelty was assessed by applying the International Union for Conservation of Nature (IUCN) Red List Categories and Criteria (IUCN 2012, 2019) and using GeoCAT (Bachman et al. 2011) to calculate the area of occupancy (AOO) and extent of occurrence (EOO).

Ecology and reproductive biology
Study site -Flowering phenology and floral visitors of the novelty were investigated in the Mount Oku National Park in May 2017. Mount Oku, the second highest mountain on the West African mainland, with an elevation of about 3,011 m a.s.l. (Asanga 2002), is part of the Bamenda Highlands, in the North-West Region of Cameroon (6°12′N, 10°32′E). Mount Oku is located within the Kilum-Ijim forest massif, which is the largest remnant of West African upper montane forests (Maisels et al. 2000). The vegetation includes a mosaic of montane forest, montane grassland, and sub-alpine habitats Asanga 2002).
Our ecological survey was conducted in 2017 in an area of wooded savannah, at the upper montane forest-grassland edge, located on the north-western part of Mount Oku crater lake, which rises to about 2,400 m elevation ( fig. 1A). This site was chosen because of the abundance of mature individuals of the novelty within reach, to explore their ecology and reproductive biology. We planned to return to complete our data with more and improved camera traps. However, since mid-2017, troubles in this part of Cameroon prevented all access to the study area. No improvement of the situation seems conceivable in the short-and mediumterm. Ecology, phenology, and pollination reproductive success -We studied the ecology of the new Cyrtorchis species in about one hectare of natural vegetation. Phytosociological surveys of 2 m length each (see Stévart 2003) were made on branches of labelled host trees of 33 mature Cyrtorchis individuals bearing one to four inflorescences (one individual per survey) (supplementary file 1). Herbarium vouchers were collected for each of the host trees along with the epiphytes growing with the new species for subsequent identification. To study its reproductive biology, monitoring of flowering from anthesis (opening of the first flower on an individual) was conducted in situ from 5 to 29 May 2017 (25 days). We daily recorded (a) the proportion of individuals with open flowers, (b) the number of open flowers and their opening date, and (c) the proportion of flowers visited by insects and fruit set (pollinia removal from the gynostemia for male fitness and pollinia load on stigmas or fruit set for female fitness). Pollinator observation and syndromes -To identify the pollinators of this new species, we used the intervalprogramming function of a waterproof digital camera (Camera Bushnell Natureview HD MAX), using the settings to take pictures at 1-minute intervals, and also by motion detection in the field of vision on five individuals located at least 10 m apart from each other (see details in supplementary file 1). Furthermore, during two nights (on 6-7 May 2017 from 6 pm to midnight, and on 8-9 May 2017 from 7 pm to 2 am), light traps with mercury vapor lamps (Sylvania HQL bulb, 250 W), installed near the surveyed population, were used to investigate and collect potential pollinators (i.e. hawkmoths and settling moths). Captured moth specimens were examined to detect pollinia on their body. A subsequent analysis relating the pictures recorded with the cameras and the daily survey of each flower was used to determine whether a specific species is a passive visitor or an effective pollinator of the novelty. To further confirm the potential pollinator identity, the relationship between the proboscis and spur length was assessed for hawkmoths seen on the camera footage foraging the flowers. The proboscis length of individual hawkmoths was measured to the nearest millimetre (from the base near the head to the tip) with a ruler (Martins & Johnson 2007;Peter & Venter 2017). The spur length of fully developed flowers was measured from BRLU herbarium specimens (n = 9). The captured hawkmoths specimens were sent to Rodolphe Rougerie, a sphingid specialist at the Muséum national d'histoire naturelle in Paris for identification. Breeding system -To test the ability of the new species to reproduce by autonomous self-pollination, we bagged three individuals (with a total of five inflorescences and with 20 closed flower buds) with fine nets prior to flower anthesis and we monitored their development in natural condition until the end of flowering. Hand pollinations were set up ex situ in the Elak village, to investigate the new species' breeding system, by performing 114 hand-pollinations including 58 cross-pollinations (20 individuals; 58 flowers) and 56 selfpollinations (19 individuals; 56 flowers) on 39 cultivated specimens. The Shapiro-Wilk Normality test (Royston 1982;Dytham 2011) was performed prior to statistical difference testing to evaluate if fruit set followed a normal distribution pattern. To statistically evaluate the difference between the fruit set according to the pollination treatment (self-or crosspollination) a Wilcoxon rank-sum test was performed, which is recommended for the comparison of two groups of nonparametric datasets (fruit set for self-and cross-pollination treatments), as observed here (e.g. McDonald 2014). These analyses were performed using R v.3.4.4 (R Core Team 2020). Diagnosis -Cyrtorchis okuensis is similar to C. submontana but differs in having wider floral bracts (7-12.5 × 8-13.7 mm vs 4-6.5 × 5-7 mm), longer anther caps (2.5-3 vs 1-2 mm), and wider triangular dorsal sepal (4-5 mm vs 2-3 mm). These species also differ by their habitat; C. submontana is found in submontane and evergreen moist forest (between 550 and 1,200 m), whereas C. okuensis is restricted to the upper montane vegetation (between 1,600 and 2,500 m) along the CVL. Description -Epiphytic herb, erect, monopodial, with short to long unbranched stems, 3-18.5 cm long bearing 3-13 alternate leaves with a 16-21.5 cm leaf span. Roots 15.5-20.8 cm long and 0.3-0.4 cm diameter, pale green and covered by mosses. Leaves closely elliptic, distichous, subimbricate, fleshy, olive green, margins entire, 4.6-11.3 × 1.2-3 cm, unequally bilobed at the apex, with rounded lobes. Inflorescences up to 4, axillary or below the leaves, 4.4-9 cm long, 3-12-flowered, flowers 3.3-4.6 cm long with floral node ranging from 0.5-0.9 cm long. Bracts broadly ovate, 7-12.5 mm long and 8-13.7 mm apart, dark brown apiculate at the apex. Flowers white, fading to orange, 3-4.7 cm long and 1-1.8 cm in diameter. Dorsal sepal triangular, 8-13 × 4-5.5 mm. Lateral sepals lanceolate, acute at the apex, 8-15.5 × 3-5.2 mm. Petals similar to lateral sepal but shorter and usually acuminate at the apex, 8-13 × 3-4 mm. Lip triangular with acuminated fleshy apex, 8.5-12 × 5-6 mm; spur incurved sometime sigmoid, 1.9-4 cm long. Anther cap whitish, turning brown with senescence, obovate in the upper part and with narrowed lower part, slightly toothed in the margins and covering the base of the rostellum 2.5-3 × 1.5-2 mm. Pollinia 2, oblate from side rarely ovoid. Column stout and short, 1-2 mm long. Viscidium saddle-shaped, with a uniformly hyaline texture, bifid at the basal apex, 2-4 mm. Stipites spathulate or slightly clavate with acute apices pointing towards the middle side, insertion point at the rear third of the length of the viscidium, 1.5-2.5 mm long. Pedicel with ovary 10.5-19.5 mm long; ovary triquetrous. Rostellum trifid, 2.5-4 mm long, midlobe straight, acute, lateral lobes linear, about three times as long as the midlobe, pendulous and parallel to each other with rounded apices. Fruits narrowly ellipsoid capsules, triquetrous with three smooth sides, narrowly ellipsoid, turning yellowish-green at maturity, 12.9-24.5 × 4.5-10.8 mm. Seeds fusiform, 149-194 × 70-93 μm, rarely ellipsoid with ovoid to ellipsoid embryo, 117-141 × 56-76 μm, with two testa cells in an individual seed coat along the longitudinal axis of seed. Position -Cyrtorchis okuensis belongs to section Homocolleticon Summerh. based on the structure of its viscidium, uniformly hyaline in texture, and its thin apices of the stipites. Distribution -Endemic to the Cameroon Volcanic Line: Nigeria, Cameroon, and Bioko (Equatorial Guinea) ( fig. 3). Habitat and ecology -Restricted to montane vegetation (moist forest, forest-grassland edge, and wooded savannah from 1,600 to 2,500 m), Cyrtorchis okuensis grows on mossy branches of Gnidia glauca (Fresen.) Gilg (Thymelaeaceae), Albizia gummifera (J.F.Gmel.) C.A.Sm. (Fabaceae), and Prunus africana (Hook.f.) Kalkman (Rosaceae). On these phorophytes, C. okuensis is usually found growing with mosses, lichens, ferns, and other orchid species (see details below in "Ecology and reproductive biology"). Phenology -The flowering period occurs between April and June with a flowering peak in May. Fruit maturation lasts about 230 days on average (from June to January). Etymology -The epithet okuensis refers to Oku, a subdivision in the North-West Region of Cameroon where the large, surveyed subpopulation was discovered around Lake Oku. The term Oku also refers to the people who live in this area and their native language. Red list conservation status -The species is given a Red List status of Near Threatened [NT]. The species is known from 21 specimens in Nigeria, Cameroon, and Bioko (Equatorial Guinea), which represent 21 occurrences. The extent of occurrence (EOO) of Cyrtorchis okuensis is estimated to be 38,319.1 km², far exceeding the upper limit of the Vulnerable category (20,000 km²) under the subcriterion B1, while its area of occupancy (AOO) is estimated to be 68 km², which falls within the limits of the Endangered category under the subcriterion B2.

Cyrtorchis okuensis
In Nigeria, C. okuensis has been collected in the Boshi Extension Forest Reserve. In Bioko, the species is known from the Pico Basilé National Park and the Reserva Científica de la Caldeira de San Carlos (at Lago Moka). In Cameroon, the species is recorded in three protected areas: the Mount Oku and the Mount Manengouba National Parks (where some occurrences are below the lower altitudinal limit of effective protection) and the Bafut Nguemba Forest Reserve. In unprotected sites, the species habitat is mainly subject to small-scale shifting agriculture, the logging activities on Prunus africana, and the fuelwood collection for small-scale subsistence.
Based on the most serious plausible threat, which is shifting agriculture, the 21 occurrences represent 16 locations (sensu IUCN 2019), which is more than 10, the upper limit of the Vulnerable category under the condition 'a' of subcriterion B2. Cyrtorchis okuensis is not severely fragmented and there are no extreme fluctuations in its parameters (EOO, AOO, number of locations or subpopulations, and number of mature individuals). The projected ongoing loss of its habitat leads us to predict a continuing decline in the number of subpopulations in four Regions in Cameroon (in Bangem near the Banyang-Mbo Wildlife Sanctuary in the South-West Region, in Bambuto in the West Region, in Mount Manengouba near Nkongsamba in the Littoral Region, in Nkambe and the Bafut Nguemba Forest Reserve in North-West Region), in mature individuals, in its AOO and its EOO. Since two (a and c) of the three conditions under subcriterion B2 are not met, C. okuensis cannot be regarded as a threatened species. Considering that shifting agriculture will lead to the loss of four locations in the near future, reducing the number of locations to 12, Cyrtorchis okuensis is assigned a IUCN Red List status of Near Threatened (NT

Ecology and reproductive biology
Ecology, phenology and pollination reproductive success -In the shrubby savannah surveyed around Lake Oku, the main host trees were Gnidia glauca with 32 host trees and Prunus africana with only one host tree. Gnidia glauca, a small bushy tree growing in a large, monodominant subpopulation in the study area, was always associated with a high number of individuals of C. okuensis, bearing up to 13 individuals ( fig. 1B). Specimens of C. okuensis were found growing on branches of 4 to 30 cm in diameter, under the canopy of both host trees at 1 to 4 m above ground level. On the 33 phytosociological branch surveys, the species was found growing in association with mosses such as In natural conditions, the rate of pollinarium removal averaged 38.8% (174 flowers out of 448 flowers) for the 33 individuals. Successful pollen deposit on the stigma, indirectly measured by fruit set rate, was 19.4% (87 fruits/448 flowers), and 15.6% reached maturity (70 matured fruits/448 flowers) after an average of 230 days (table 1). Pollinator observation and syndromes -A total of 72,476 pictures (more than 1207 hours of observation) were recorded with four camera traps, showing visits of at least seven different insects, one spider, and two bird species (the northern double-collared sunbird (Cinnyris reichenowi Sharpe, 1891) and the oriole finch (Linurgus olivaceus Fraser, 1843)). During daytime, flowers were mostly visited by birds and flies, while hawkmoths, settling moths, and grasshoppers were the main visitors at night. Apart from birds, which ate the perianth parts and with possible nectar robbery, and hawkmoths, which only foraged on flowers for nectar, the other insects were passive visitors.
The opening of the flowers occurred only at night. A slight floral scent was detected during daytime when getting near to the flowers. At dusk, from about 6 pm, a jasmine-like scent began to emanate from the flowers and was detectable a few meters away. Scent release began from about 7:30 pm and remained strong throughout the night. Hawkmoth

In situ Pollinators excluded (bagged flowers)
3 20 ---  visits mostly occurred at dusk between 6 pm and 1 am with a peak of visits being between 10 pm and 1 am (  5B-D). Unfortunately, none of the recorded nor captured hawkmoth specimens were caught with the pollinia attached on any part of their bodies. Nevertheless, based on pollinia removal and/or deposit observed during daily monitoring, as well as fruit set after their visits, we infer that one or more of these hawkmoth species is the effective pollinator of C. okuensis in the Mount Oku area. Morphologically, the mean spur length of C. okuensis was 28.2 mm (sd = 0.64, n = 9). This value matches the length of the proboscis of all nine captured Hippotion specimens (H. celerio (5 individuals), H. eson (3), and H. osiris (1)), which ranged between 20 and 42 mm, with a mean of 31 mm (sd = 0.86, n = 9). Based on these spur and tongue lengths, the majority of these species of hawkmoth are likely to forage on the flowers in this subpopulation. There was no significant difference in the spur and proboscis length (p = 0.89). Breeding system assessment -None of the flowers tested for autonomous self-pollination (bagged inflorescence) produced fruits, suggesting that C. okuensis requires a pollinator to set fruit (table 1). The species is however self-compatible as the self-pollination treatments resulted in 64.2% fruit set. Nevertheless, cross-pollination shows a higher success rate with a fruit set of 88%. These two controlled pollination treatments did however not differ significantly for fruit set success (W = 138, p > 0.05).

Cyrtorchis okuensis is the angraecoid orchid with the shortest spur that is pollinated by hawkmoths
Our formal description of Cyrtorchis okuensis constitutes an additional step towards a global revision of the genus Cyrtorchis. The current study not only confirms pollination by hawkmoths in the genus Cyrtorchis, as predicted by  Pl. Ecol. Evol. 154 (3), 2021 Cribb (1989), but also brings down the lower limit of spur lengths associated with hawkmoth pollination within the angraecoid group (~800 species; see Simo- Droissart et al. 2018;Farminhão et al. 2021). Cyrtorchis okuensis exhibits clear floral adaptations to hawkmoth pollination, such as the white night-scented flowers with a "relatively" long nectarrewarding spur (Martins & Johnson 2007, 2013Vogel 2012;Johnson & Raguso 2016;Johnson et al. 2017). These white long-spurred orchids with night-scented flowers are common in Africa (Dressler 1981), with data supporting that about 50% of the African orchids are pollinated by hawkmoths, based on the high frequency of this floral syndrome (Martins & Johnson 2007, 2013Johnson & Raguso 2016;Johnson et al. 2017). However, the spur length of C. okuensis appears to be relatively short (1.9-4 cm vs 4.4-24 cm long) compared to most of the other known African continental angraecoids (~16 species) in which hawkmoth pollination syndromes are documented. Luyt & Johnson (2001) previously reported this pollination syndrome for Mystacidium venosum Harv. ex Rolfe, whose spur length ranges from 2.5 to 5.5 cm and Martins & Johnson (2007) later described it for Aerangis confusa J.Stewart with an average spur length of 4.5 cm.
As observed for other short-spurred angraecoid species (i.e. M. venosum and A. confusa), there is a good match between the spur length in C. okuensis and the proboscis length of the pollinators. This might suggest that a strong selective pressure of the proboscis on spur length might have played a role. Three different species of Hippotion (H. celerio, H. eson, and H. osiris) were identified as effective pollinators of C. okuensis. These results are in line with previous studies, as H. eson was also observed feeding on flowers of Mystacidium venosum (Luyt & Johnson 2001) and H. celerio was reported to be a pollinator of Aerangis confusa in Kenya (Martins & Johnson 2007). Although it was difficult to identify pollinators at the species level, the confirmed presence of the three captured Hippotion spp. and their quite similar proboscis length advocate in support of them being effective pollinators of C. okuensis. This was previously found in a study of Alexandersson & Johnson (2002) on Gladiolus longicollis Baker, in which moths with matching proboscis lengths were effective agents of pollen removal and deposition. Additionally, more recent studies suggested that hawkmoths are highly polyphagous and readily feed on flowers that have tubes much shorter than their proboscis lengths, while plants adapted to hawkmoths are more specialized, particularly when they are long tubed (Martins & Johnson 2013;Amorim et al. 2014). Given the slight difference in size between the average length of the pollinators' proboscis and the spur of C. okuensis, and the fact that pollinia in the genus Cyrtorchis are generally placed along the proboscis (Laura Azandi, pers. obs. on Cyrtorchis letouzeyi and C. chailluana), we can assume that there is a selection pattern for short spurs as described by Ellis & Johnson (2010) in their sex-specific selection model. This model states that if selection is made through the male, this would favour an evolution towards shorter spurs than proboscis because pollinia export is potentially highest in short-spurred species as pollinia are placed far from the base of the proboscis and could reach both the stigmas of long and short-spurred flowers. Our autonomous self-pollination experiment provides strong evidence that C. okuensis fruit production depends on external pollinators. Under natural conditions, hawkmoth visitation events on C. okuensis flowers seem relatively rare (see fig. 4 and table 2) and resulted in low reproductive success (pollen removal and fruit set) as only 19.4% of the 448 flowers we surveyed in May 2017 produced fruits. Amorim et al. (2014) and Micheneau et al. (2009) also reported low visitation and fruit set rates of orchid flowers when pollination is accomplished by hawkmoths.

Importance and conservation of the Cameroon Volcanic Line
The CVL includes a chain of isolated volcanic or plutonic mountain peaks that covers ~40,877 km 2 , stretching from Annobón Island in the Gulf of Guinea to the Mandara Mountains in the Far North Region in Cameroon. It includes a broader continental part and a smaller oceanic portion in the form of four major islands (Ayonghe et al. 1999;Frodin 2001;Sainge et al. 2017). The CVL is known for its rich and diverse flora with high levels of endemism combined with many threats (Sosef et al. 2017;Droissart et al. 2018), which led to its classification of a biodiversity hotspot Myers et al. 2000;Barthlott et al. 2005). Plant diversity comprises about 4000 species . The orchid family represents one of the most diverse plant groups of the CVL (Cable & Cheek 1998; Cheek et al. 2000Cheek et al. , 2004Onana & Cheek 2011;Harvey et al. 2004) including 33 strictly endemic taxa, distributed in the submontane (17 taxa), montane (12 taxa), and both submontane and montane strata (4 taxa) (Droissart 2009). The new orchid described here further stresses the need for additional fieldwork in these mountains, specifically for sites that have so far been little or not explored. This seems particularly the case for the forests of the Cross River National Park, the Boshi Extension Forest Reserve, and the Gashaka-Gumti National Park in Nigeria, and for the Rumpi Hills area in Cameroon. Additional material from African herbaria (e.g. Forest Herbarium Ibadan (FHI)) should be carefully examined to identify putative unidentified or misidentified orchid material.
With over 200 plant species considered as threatened Onana & Cheek 2011) and with more than 80 endemic species Franke 2004;Sainge et al. 2005Sainge et al. , 2010Sainge 2012Sainge , 2016, the CVL is endangered by increasing threats due to human activities. Land use change and deforestation have resulted in degraded forest areas, fragmented populations, and the CVL is therefore impoverishment on all three levels of biodiversity (i.e. genetic, species, and ecosystem diversity). It has been estimated that over 96% of the original forest cover of the Bamenda highlands above 1,500 m elevation has been lost . Hence, a regional synthesis, including recent species diversity checklists coupled with Important Plant Area assessments (Darbyshire et al. 2017), is urgently needed to better manage and conserve the remaining biodiversity of the CVL.

Threats to the host trees of Cyrtorchis okuensis
Our results identified Gnidia glauca and Prunus africana as main host trees of C. okuensis. This disparity in the number of host trees identified in this region could be closely related to the fact that our study was conducted at a grassland edge of a natural and monospecific G. glauca forest (Momo et al. 2017), but also to the strong threat to P. africana over the past several years by unsustainable harvesting in many countries where it occurs. Indeed, this species is used by pharmaceutical companies to manufacture a drug used in treating benign prostatic hyperplasia and prostate gland hypertrophy (Stewart 2003). Cunningham (2005) reported that over the last 40 years, P. africana bark harvest for its medicinal properties has shifted from subsistence use to large-scale commercial use for international trade. Cameroon comprises large parts of the current species distribution, and in the Kilum-Ijim area, P. africana is traditionally used to treat malaria and stomach ache (Nsom & Dick 1992). Currently listed as Vulnerable on the IUCN Red List (IUCN 2020), the species is also included in Appendix II of the Convention on International Trade in Endangered Species of Flora and Fauna (CITES) (Betti 2011). Therefore, its overexploitation could have led to a drastic reduction of its populations.
Although G. glauca seems less threatened, it is not exempt of risks from future overexploitation as it is also widely used in traditional pharmacopoeia, in nanomedicine, in domestic service in the form of rope or yarn made from bark fibres, and as an insecticide or piscicide (Ghosh et al. 2012;Avana-Tientcheu et al. 2018). A recent study (Avana-Tientcheu et al. 2018) conducted in the Oku area revealed that G. glauca was more vulnerable in forests than in savannah, because of the proximity of forest stands to neighbouring villages makes them more easily accessible than savannah stands. The IUCN conservation status of G. glauca has not yet been assessed and would help to propose effective conservation measures. As with P. africana (Avana 2006), domestication should be also considered for G. glauca to be able to meet potential high future demands (Avana-Tientcheu et al. 2018).

SUPPLEMENTARY FILE
Supplementary file 1 -Detailed information about the inflorescences and flowers per surveyed individual and about the camera survey sessions. The IDs followed by a * are those surveyed with cameras. https://doi.org/10.5091/plecevo.2021.1823.2579

ACKNOWLEDGEMENTS
We express our gratitude to the Institut de Recherche pour le Développement (IRD) for financial support provided for field activities in Oku. We are also grateful to the American Orchid Society (AOS) and the Académie de Recherche et d'Enseignement Supérieur (ARES) for funding the PhD activities of the first author in Cameroon and her stay at the herbarium of the Université Libre de Bruxelles, and to the Fonds de la Recherche Scientifique (FNRS-FRIA) for the grant to Tania D'haijère. We also express our gratitude to the National Geographic Society (Grant C303-15, Vincent Droissart as PI) who supported ex situ conservation activities (orchid living collection and seedbank) in Cameroon. We thank all the conservation staff of the Mount Oku National Park and the traditional authorities for allowing us to work in the reserve and access to their facilities. We are grateful to M. Samuel Bilack for his support during the field trip and Roger Kamgang who initiated us to insect light trapping. We gratefully acknowledge the curators of various herbaria (BR, BRLU, K, P, WAG, and YA) for making their collections available for direct examination and loan. We express our sincere gratitude to the curator of the Lepidoptera at the Muséum national d'histoire naturelle in Paris, Dr Rodolphe Rougerie, for his helpful comments on moths' collection and management, and for the identification of hawkmoths collected in the field. We also acknowledge Dr Vincent Deblauwe for reporting and providing pictures of herbarium sheets of Cyrtorchis okuensis from the Forest Herbarium of Ibadan in Nigeria. We are grateful to Nicolas Texier for his help in providing the distribution map and to João N.M. Farminhão for his constructive comments on the description of the species and for reviewing an earlier version of the manuscript. We finally wish to express our gratitude to Dr Elmar Robbrecht, Dr Thierry Pailler, and an anonymous reviewer who provided a thoughtful and helpful review of the manuscript.