The subtribe Malaxidinae includes approximately 1250 species across 14 genera, with a predominantly tropical and subtropical distribution (Kolomeitseva et al. 2024; Zeng et al. 2024). Malaxis is one of the most diverse genera within the subtribe, comprising ca 300 species found worldwide, especially in tropical and subtropical regions (Chinchilla et al. 2022; Jermakowicz et al. 2022). Most species are terrestrial, bearing fibrous superficial roots and tunicate pseudobulbs. Plants typically have one or two leaves per sympodial unit and produce inflorescences as spikes, racemes, or corymbs. The flowers are small, non-resupinate, usually pale green, and lack accessory structures in their pollinia (i.e. naked pollinia) (Dressler 1981; Cameron 2005). In the Americas, a total of 143 Malaxis species have been documented and distributed from North America to Argentina (Ulloa Ulloa et al. 2017; Chinchilla et al. 2022). Mexico is recognised as the most diverse country for the genus, hosting 71 species. Brazil harbours 11 species (Santos and Smidt 2023), while eight species are recorded in Peru (Ulloa Ulloa et al. 2017).

Here, we examined two Malaxis species. Malaxis parthoni (Fig. 1) and M. excavata (Fig. 2) are terrestrial orchids with comparable plant heights, ranging from 14.52 to 39.33 cm. Both species typically produce two leaves. In our study sites, M. parthoni displayed a highly dispersed distribution, with scarce and isolated individuals, whereas M. excavata occurred in dense aggregations. The leaves of M. parthoni are elliptic with acute apices (Fig. 1A), while those of M. excavata are elliptic-lanceolate, also ending in acute tips (Fig. 2A). The main difference between the two species lies in the shape and size of the labellum: in M. parthoni, the labellum is relatively flattened (Fig. 1F), whereas in M. excavata it is concave with a truncate apex (Fig. 2I).

Figure 1. 

Vegetative and floral characteristics of Malaxis parthoni. A. Habit. B. Upper view of the corymbiform inflorescence. C. Frontal view of a flower. D. Back view of a flower. E. Longitudinal section of a flower. F. Detached labellum and column. G. Cavity of the labellum with visible nectar. H. Naked pollinia. I. Lateral view of the labellum and column, showing the stigmatic surface (white arrowhead).

Figure 2. 

Vegetative and floral characteristics of Malaxis excavata. A. Habit. B. Upper view of the corymbiform inflorescence. C. Frontal view of the inflorescence. D. Frontal view of a flower. E. Back view of a flower. F. Naked pollinia. G. Column showing the stigmatic surface. H. Stigmatic surface with pollinia attached (white arrowhead). I. Labellum with visible nectar (blue ellipse). J. Developing fruit.

Malaxis parthoni was investigated in southern Brazil, specifically in Morro Santana (Porto Alegre municipality; 30°04’01.59”S, 51°07’29.69”W; 79 m) and Morro São Pedro (Viamão municipality; 30°10’55.82”S, 51°05’55.59”W; 91 m), both located in the state of Rio Grande do Sul, near the southern limit of the Atlantic Rainforest Biome (Mata Atlântica) (Fig. 3A, C, E). The climate in this region is characterised by a mean annual temperature ranging from 18 to 20°C and an average annual precipitation of 1,300 to 1,500 mm, with no distinct dry season throughout the year (Overbeck et al. 2005).

Figure 3. 

Study sites for pollinator observations and breeding system experiments. The map shows the country boundaries in South America (A), highlighting the Peruvian department of Ayacucho, where Malaxis excavata was studied (B), and the Brazilian state of Rio Grande do Sul, where M. parthoni was studied (C). Satellite imagery illustrates the study area in Peru (D), as well as the two in Brazil (E).

Malaxis excavata was studied in the Andean region of Ayacucho, Peru, at an elevation of 3,500 m (13°00’10.25”S, 74°08’46.84”W) (Fig. 3A, B, D). According to Holdridge’s life zone classification, this area corresponds to the Subtropical Montane Humid Forest, characterised by a mean annual temperature of 12.9°C and a maximum average total precipitation of 1,190 mm (INRENA 1995). Unlike the Brazilian sites, the Ayacucho region exhibits a clearly defined wet and dry season, with the wet season usually extending from late October to March and the dry season from April to early October. In this area, M. excavata grows within a monospecific forest dominated by Alnus acuminata Kunth (Betulaceae). Maps were developed using QGIS v.3.22.2. Elevation maps were obtained from WorldClim 2.1 (Fick and Hijmans 2017).

Given that floral traits are considered key to understanding pollination processes, and floral attractants (Woodcock et al. 2014; Assis 2023), floral and inflorescence features including flower colour, scent, size, sexual organ position, reward type, and inflorescence morphology were described and measured from fresh anthesis-stage flowers of five individuals per species (6 to 10 flowers in M. parthoni and 6 to 8 flowers in M. excavata). As hypothesised, Diptera may serve as potential pollinators. In orchids pollinated by this group of insects, fragrance emission is commonly associated with pollinator attraction (Tan et al. 2002; Jermakowicz et al. 2022). To test this, the timing of fragrance release was recorded through hourly olfactory inspections of inflorescences, all conducted by the same observer (Calderon-Quispe and Singer 2024). Flower longevity was also recorded, as it determines the period during which flowers remain available to attract and receive visits (Primack 1985). To quantify longevity, flowers were isolated from pollinators using tulle bags (Calderon-Quispe and Singer 2024). In total, 20 flowers of M. parthoni (from 10 individuals) and 30 flowers of M. excavata (from 10 individuals) were monitored.

During pollinator observations, we also noted the presence of a watery secretion on the labellum. However, the volume was minimal, hence a qualitative test for sugar content was performed to determine whether the secretion could be classified as nectar. Using a glucose monitoring system, Bioland G-245-3, a drop of distilled water was deposited on the labellum to wash the surface, and the resulting liquid was then absorbed using a reactive glucose test strip, allowing for qualitative assessment of sugar presence (n= 10 flowers per species) (Power et al. 2018). The voucher specimens of M. excavata were deposited in the Herbario Sur Peruano, Instituto Científico Michael Owen Dillon (HSP), and those of M. parthoni in the Herbário do Instituto de Ciências Naturais, Universidade Federal do Rio Grande do Sul (ICN).

Diurnal and nocturnal observations were conducted for both species. Diurnal observations took place between 06:00 and 18:00 h, while nocturnal observations were carried out from 18:00 to 24:00 h. Observations of M. parthoni were conducted in June and July of 2023 and 2024, whereas observations of M. excavata were performed in January 2024. For each species, a total of 80 h per species was conducted over five days. Only insects that effectively removed and inserted pollinia were considered pollinators (Adams and Lawson 1993). Pollinators of M. parthoni were photographed and filmed using a Nikon D5300 camera and a Nikon AF-S VR 105-mm macro lens, while those of M. excavata were documented with a Sony DSC-HX400V 215-mm camera. Video recordings were used to describe visitor behaviour. Fungus gnats (Mycomya Rondani, 1856) visiting the flowers were collected using an aspirator. Two individuals per species were sampled and preserved in 70% ethanol. The Mycomya species collected in Peru was deposited in the Museo Sur Peruano, Instituto Cientifico Michael Owen Dillon (MSP), while the voucher specimens collected in Brazil were deposited in the Museu de Ciências Naturais, Secretaria do Meio Ambiente e Infraestrutura (MCN), Porto Alegre, Brazil.

Ten individuals were isolated from pollinators using tulle bags, through which air and light could pass, thereby minimising effects on the flowers and plants (Sagili et al. 2025). Four treatments, previously applied to Neotropical orchids following the methodologies of Calderon-Quispe and Singer (2024), Sanguinetti and Singer (2014), and Buzatto et al. (2022), were conducted on each individual: intact flowers, to test for autonomous autogamy (i.e. whether flowers can set fruit without any pollinator intervention); emasculation, to test for apomixis (i.e. fruit development in the absence of pollen and consequently fecundation), which consists of removing the pollinia; manual self-pollination, to assess self-compatibility (i.e. the capacity to produce fruits and seeds from pollen of the same flower); and manual cross-pollination, to evaluate fruit set when pollen comes from a different individual. For M. parthoni, five flowers per treatment were applied to each of 10 individuals (50 flowers per treatment in total), whereas for M. excavata, only three flowers per treatment could be used per individual (30 flowers per treatment in total). No statistical comparisons were applied, since only one treatment produced fruits.

Pollination efficiency was assessed using Nilsson’s male efficiency factor, calculated as the ratio of the percentage of pollinated flowers (flowers per inflorescence on which pollinia were deposited on the stigmatic surface) and pollen donor flowers (flowers per inflorescence from which pollinia were removed) (Nilsson et al. 1992; Buzatto et al. 2022). Accordingly, we recorded the number of pollinated flowers, and pollen donor flowers were evaluated in 10 individuals of Malaxis parthoni and 21 individuals of M. excavata. The number of pollinated flowers was statistically compared to the number of donor flowers using the non-parametric Wilcoxon signed-rank test, after assessing normality with the Shapiro-Wilk test (p < 0.05).

In the same localities where pollinator observations were carried out, 10 individuals per species were collected just at the end of the flowering period (i.e. upon fruit maturation). Plants were monitored weekly to determine the fruit maturation. For each individual, fruit set success was calculated by dividing the number of fruits formed by the number of flowers per inflorescence (Castro et al. 2022; Calderon-Quispe and Singer 2024). Statistical comparisons of Nilsson’s male efficiency factor and fruiting success between species were performed using the non-parametric Mann-Whitney U test, after testing for normality with the Shapiro-Wilk test (p < 0.05). All statistical comparisons were done in RStudio v.2023.12.1.402 (R Core Team 2024), and the graph was generated using the R package ggplot2 v.3.5.2 (Wickham 2016).

Both species produce corymbiform inflorescences (Figs 1B, 2B). Statistically, M. parthoni produces significantly more flowers than M. excavata (Mann-Whitney U test, p < 0.05), with floral counts ranging from 98 to 206 in M. parthoni, and from 18 to 102 in M. excavata (Table 1). Flowers are greenish, non-resupinate, and pedicellate. Pedicel lengths vary from 0.53 to 0.91 cm in M. parthoni, and from 0.71 to 1.02 cm in M. excavata (Table 1). Sepals are oblong-lanceolate in M. parthoni and elliptic-oblong in M. excavata. The lateral petals are linear and convolute in both species. The labellum of M. parthoni is flattened, trilobate with an entire margin, an acute apex, and inconspicuous, nearly rounded lateral lobes. In contrast, the labellum of M. excavata is trilobate, truncate, and concave, featuring a longitudinal crest on its abaxial surface. Both species secrete drops of liquid containing sugars, as confirmed by a qualitative assay, which likely serves as a floral reward for pollinators. This secretion can therefore be regarded as nectar, released through two cavities at the base of the labellum (Figs 1G, 2I) and also from the fleshy anterior region of the middle lobe, which is more robust and visibly nectariferous in M. excavata (blue ellipse in Fig. 2I) compared to M. parthoni. The column in both species is dorso-ventrally compressed (Figs 1F, 2D) and bears two naked pollinia at the subapical dorsal position (Figs 1H, 2F). The stigma surface is bilobed and positioned apically on the column (Figs 1I, 2GH). The ovary is green and glabrous in both species (Figs 1D, 2E), as are the fruits (Fig. 2J). Overall, the flowers of M. excavata are larger than those of M. parthoni (Table 1).

The flowering period of M. parthoni extends from June to August, while that of M. excavata occurs between December and February. Flowers are long-lived, ranging from 8 to 21 days in M. parthoni and from 14 to 17 days in M. excavata, with significantly longer floral duration in the latter species (Mann-Whitney U test, p < 0.05; Table 1). In both species, the flowers are considered long-lived. The floral scent of M. parthoni resembles that of fresh fish and was detected between 14:30 and 18:00 h, with peak intensity around 17:00 h. In contrast, M. excavata emits an acrid scent, detected between 16:00 and 20:00 h, peaking around 19:00 h.

Fungus gnats of the genus Mycomya (Mycetophilidae) acted as pollinators of both species (Suppl. material 1 and 2). Individuals of a species provisionally designed as Mycomya sp.1 pollinated Malaxis parthoni, while those of another species of the same genus (Mycomya sp.2) pollinated Malaxis excavata. A total of 59 pollinator interactions were recorded for M. parthoni and 83 for M. excavata. In M. parthoni, pollinator activity occurred between 13:30 and 18:00 h, with a peak in visitation between 16:00 and 17:00 h (Fig. 4). During each visit, pollinators interacted with 2 to 23 flowers, spending between 2 and 22 seconds per flower, and between 10 and 65 seconds per inflorescence (Table 2). In M. excavata, interactions were recorded between 16:00 and 20:30 h, with a peak between 18:00 and 18:30 h (Fig. 4). Pollinators visited 1 to 23 flowers per visit, spending 12 to 313 seconds per flower and 87 to 1032 seconds per inflorescence (Table 2).

Figure 4. 

Visits by Mycomya species to Malaxis parthoni (studied in southern Brazil) and M. excavata (studied in the Peruvian Andes) flowers, recorded at 30-minute intervals for a total of 80 hours of observation per species of Malaxis.

The pollination process was similar in both Malaxis species. Both exhibit corymbiform inflorescences (Figs 1B, 2B) that serve as a landing platform for pollinators, as the flowers are considerably smaller than the body size of the visiting insects (Figs 5A–D, 6A–C). Occasionally, more than two individuals of Mycomya were observed arriving at the inflorescences simultaneously (Figs 5B, 6A). Pollinators typically begin by feeding on nectar secreted along the margin of the labellum (Figs 5D, 6B). In M. excavata, this behaviour is more conspicuous due to a pronounced thickening along the labellar margin, where a greater volume of nectar is secreted (Suppl. material 2). Notably, pollinators of M. excavata were frequently observed feeding exclusively on this marginal nectar, without necessarily accessing the nectar accumulated in the cavities of the labellum. In contrast, M. parthoni has a more flattened labellum, making nectar secretion along the margin less prominent (Suppl. material 1). After feeding at the labellar margin, Mycomya individuals in both species may proceed to feed on nectar secreted by the cavities, where nectar accumulates in greater quantity. To access this nectar, the insects must insert their head into the inner part of the labellum (Figs 5C, 6C). It is at this point that the ventral part of the prothorax and the area behind the mouthparts come into contact with the pollinia that are located at the subapical dorsal position of the column (Figs 1H, 2F), which then adhere to these regions (Figs 5D–F, 6D–F). When the insect visits another flower and feeds on nectar from cavities of the labellum, the pollinia are deposited onto the stigma, which is located at the apex of the column.

Figure 5. 

Pollination process of Malaxis parthoni. A. Mycomya sp.1 (Mycetophilidae) feeding on nectar from the auricles of the labellum. B. Two individuals of Mycomya sp.1 arrive simultaneously at the inflorescence. C. Lateral view of Mycomya sp.1 feeding on nectar from the labellar cavity. D. Mycomya sp.1 carrying pollinia (white arrowhead) after visiting the flower shown in panel C, and feeding on nectar from the labellar margin. E. Ventral view of Mycomya sp.1 under a stereomicroscope, showing pollinia attached behind the mouthparts. F. Lateral view of Mycomya sp.1 under a stereomicroscope, showing the same attachment of pollinia behind the mouthparts.

Figure 6. 

Pollination process of Malaxis excavata. A. Two individuals of Mycomya sp.2 (Mycetophilidae) feeding on nectar at the same inflorescence. B. Mycomya sp.2 feeding on nectar secreted along the margin of the labellum. C. Mycomya sp.2 feeding on nectar from the cavity of the labellum. D, E. Lateral views of Mycomya sp.2 carrying pollinia (white arrowheads). F. Ventral view of Mycomya sp.2 under a stereomicroscope, showing pollinia attached behind the mouthparts.

No fruit development was observed in bagged intact flowers or emasculated individuals of either Malaxis species, indicating that both are pollinator-dependent for fruit set. Likewise, no fruit was produced under manual self-pollination, suggesting that both species require cross-pollination to achieve successful reproduction (Table 3).

In both species, the number of flowers acting as pollen donors was significantly higher than the number of pollinated flowers per inflorescence (Wilcoxon signed-rank test; p < 0.05; Table 2). In M. excavata, the number of pollinated flowers tended to be higher than in M. parthoni (Table 2). Nilsson’s male efficiency index ranged from 0.18 to 0.50 in M. parthoni (Table 2), while in M. excavata it was notably higher, ranging from 0.30 to 1.00 (Table 2), indicating greater pollination efficiency in the latter species (Mann-Whitney U test, p < 0.05; Table 2). Fruiting success in M. parthoni was consistently low, ranging from 4.03% to 35.00% (Table 2). In contrast, M. excavata showed higher values, with fruiting success ranging from 5.00% to 85.71% (Table 2).

The floral features of nectar-rewarding plants pollinated by fungus gnats are generally characterised by dark red to greenish, actinomorphic, and flat-shaped flowers, with short stamens and exposed nectaries (Okuyama et al. 2008; Mochizuki and Kawakita 2018; Mochizuki et al. 2023). Our findings show that both studied Malaxis species conform to these floral features: their flowers are greenish, with M. parthoni displaying a more flattened morphology, and in both species, nectar is secreted superficially on the labellum. A comparable pattern is seen in Neottia cordata (L.) Rich., a terrestrial nectar-rewarding orchid also pollinated by fungus gnats, which exhibits green to purple, flattened flowers with superficial nectar secretion (Ackerman and Mesler 1979). Typically, nectar-rewarding plants pollinated by fungus gnat produce very low volumes of nectar, less than 2 μL (Hayashi et al. 2025). In our study species, we observed the production of nectar; however, based on the minimal volume, we only confirm a qualitative presence of sugar in the solution. Field observation made by Claessens and Kleynen (2011) found droplets of liquid in the labellum of Malaxis monophyllos. In contrast, Jermakowicz et al. (2022) mentioned that this species does not produce nectar but instead exhibits high metabolic and secretory activity throughout the labellum, suggesting the release of volatiles and nutritional rewards such as lipids. Among Neotropical Malaxis species, the production of nectar was not found in M. massonii (Aragon and Ackerman 2001), whereas in some Mexican species, the presence of liquid droplets on the flowers has been reported, which may potentially represent nectar (Kite and Salazar 2008). To date, no information on nectar sugar concentration is available for any Malaxis species, primarily because the secreted volumes are too low to be sampled reliably (Hayashi et al. 2025). In general, the nectar utilised by most flies is characterised by high sugar concentrations and is hexose-rich (Woodcock et al. 2014; Basith and Richard 2025). The limited information currently available underscores the need to expand knowledge on nectar concentration and its potential role in pollination mediated by fungus gnats in orchids. Future studies should specifically investigate nectar traits in these systems to better understand their ecological significance.

Several authors have hypothesised that floral scents play a key role in fungus gnat pollination systems (Okamoto et al. 2015; Katsuhara et al. 2017; Mochizuki and Kawakita 2018). Most plants pollinated by fungus gnats produce unpleasant floral odours, often described as resembling fermented dairy products or fish (Mochizuki and Kawakita 2018). Our study species also emitted similar odours: in Malaxis parthoni, flowers released a fresh, fish-like scent, whereas in M. excavata the odour was unpleasant and acrid. Comparable floral odours have been reported in other Malaxis and Liparis species (Kite and Salazar 2008; Kaiser 1993), with the exception of M. rzedowskiana R.González, which produces violet-like floral notes (Kite and Salazar 2008). Detailed analyses of scent composition in M. monophyllos revealed high levels of aliphatic compounds that elicit strong responses in Diptera (Jermakowicz et al. 2022). In this species, additional visual signals are also involved: raphides along the labellum margin fluoresce under UV light and may help guide pollinators (Jermakowicz et al. 2022). Similarly, in other orchid groups pollinated by fungus gnats, the release of unpleasant odours has been consistently reported (Ackerman and Mesler 1979; Borba et al. 2011). In our study species, the timing of scent emission coincided with peak pollinator activity, suggesting that these odours play an important role in pollinator attraction and thereby contribute to the higher fruiting success observed in both species.

Finally, the flower longevity has only been determined in a few Malaxis species. In our study, M. parthoni exhibited an average floral lifespan of 13.75 days, whereas M. excavata showed a longer average of 15.86 days. Flowers lasting more than three days are generally considered long-lived (Primack 1985; Fonseca et al. 2015). Such long-lived flowers are commonly reported in plants that rely on specialised pollinators, such as orchids (Primack 1985; Ashman and Schoen 1994). In orchids, floral longevity has been interpreted as an adaptation to increase the likelihood of cross-pollination and thereby enhance reproductive success (Fonseca et al. 2015). Moreover, plants growing at high elevations often exhibit extended floral lifespans, a pattern associated with reduced pollinator activity and scarcity at such elevations (Steinacher and Wagner 2010; Calderon-Quispe and Singer 2024). Consistent with this, M. excavata, which occurs at higher elevations, presented significantly longer flower longevity than M. parthoni, which inhabits lower elevations. On the other hand, inflorescences with a higher number of flowers generally show reduced floral longevity (Ashman and Schoen 1994). In line with this trend, M. parthoni produced significantly more flowers per inflorescence than M. excavata, but exhibited markedly shorter flower longevity. Comparable floral longevities have been reported in other genera within the subtribe. For example, Reeves and Reeves (1984) found that in Hammarbya paludosa, the first flowers lasted about 3 to 4 weeks, while later ones persisted only a few days. In Crepidium acuminatum and Dienia ophrydis, floral longevity ranged between 7 and 15 days (Nuammee 2018), whereas a much shorter lifespan was recorded for Liparis grandiflora Ridl., averaging just 5 to 6 days (Qi et al. 2024). Due to the fact that most species studied so far grow at low elevations, it remains unclear whether high-elevation species of Malaxis and related genera within the subtribe exhibit the same pattern of extended floral longevity as observed in M. excavata.

In both Malaxis species, only fungus gnats of the genus Mycomya (Mycetophilidae) were seen performing effective pollination, specifically through the removal and transfer of pollinia. In our observations, these insects visited the flowers primarily to obtain nectar, a behaviour that mirrors what has been reported for Liparis gigantea, where Mycomya also acts as a pollinator by feeding on floral nectar (Qi et al. 2024). We did not observe any behaviour indicative of sexual or reward deception. However, a deceptive mechanism (whether sexual or reward-based) has been suggested for Malaxis massonii, as nectar was not found as a potential reward for pollinators in this species (Aragon and Ackerman 2001; Argue 2012). In other Malaxis species from North America, Europe, and Asia, floral size and phenology suggest that fungus gnats and gall midges may serve as potential pollinators, although this has not yet been confirmed (Argue 2012, 2014).

What has been confirmed, at least for the species of Malaxis whose pollinators and other floral visitors have been documented so far, is their strong association with Diptera (Claessens and Kleynen 2011; Argue 2014; Jermakowicz et al. 2022). Different taxonomic groups within this order have been observed actively visiting the flowers (Jermakowicz et al. 2022). As mentioned in the previous section (see Floral features), this association is likely mediated by the emission of floral scents that are highly attractive to dipteran pollinators (Kite and Salazar 2008; Jermakowicz et al. 2022). In this sense, our observations provide the first confirmed case of pollination mediated by fungus gnats in Neotropical Malaxis species.

Fungus gnats are generally most active in the early morning and late afternoon, although some species exhibit strictly nocturnal behaviour (Jakovlev 2012). In our study species, visitation peaked in the late afternoon (16:00–17:00 h) in M. parthoni, whereas in M. excavata, activity extended further into the evening (18:00–20:30 h). A comparable visitation pattern has been reported in Liparis gigantea (Qi et al. 2024). Nocturnal pollinator activity has also been documented in Pleurothallis marthae Luer & R.Escobar (Duque Buitrago et al. 2014). However, contrary to this general tendency, some orchids, such as Pterostylis sanguinea D.L.Jones & M.A.Clem., as well as unrelated plant families pollinated by fungus gnats, show diurnal visitation patterns (Phillips et al. 2014; Goldblatt et al. 2004; Okuyama et al. 2004). Such variation in pollinator activity may reflect the combined influence of floral traits (e.g. peaks in scent emission and nectar availability) and abiotic factors such as temperature, relative humidity, and light intensity, to which fungus gnats are particularly sensitive (Jakovlev 2012; Duque Buitrago et al. 2014; Basith and Richard 2025). In both studied species, pollinator visitation peaks coincided with periods of maximum floral scent emission and ceased once fragrance production ended, demonstrating and reinforcing the importance of floral scents as key attractants in these interactions. During the pollination process in both studied species, pollinia were attached to the anteroventral part of the thorax and behind the mouthparts of the fungus gnat. These observations are consistent with previous reports in Malaxis monophyllos (Claessens and Kleynen 2011), Hammarbya paludosa (Reeves and Reeves 1984), and Crepidium acuminatum. This last species is pollinated by a fungus gnat of the Sciaridae family (Nuammee 2018). Other genera of the subtribe Malaxidinae exhibit different pollinia attachment sites. For example, in Liparis gigantea, the pollinia were attached to the dorsal position of the thorax (Qi et al. 2024). While in Oberonia japonica, the pollinia adhere to the heads of gall midges from the Cecidomyiidae family (Sunakawa et al. 2024). In the case of C. acuminatum, the pollinia have also been observed attached to the proboscides and legs of Syrphidae species. Similarly, in D. ophrydis, the pollinia were attached to the proboscis of a bug of the family Miridae. However, this species is also self-pollinated with rainy assistance, similar to that documented in some Liparis species (Catling 1980; Nuammee 2018; Suetsugu 2019).

Our results show that Malaxis parthoni and M. excavata are self-incompatible and rely on pollinators for fruit sets. These findings are consistent with those reported for M. massonii in Puerto Rico (Aragon and Ackerman 2001) and in contrast to reports of self-compatibility in Asian and North American species (Argue 2014). Additionally, this study examined two Malaxis species from different elevations. Despite this difference and the common expectation that high-elevation orchids tend to be autonomously self-pollinated due to limited pollinator availability (Jacquemyn et al. 2005; Ackerman et al. 2023), we demonstrate that, regardless of elevation, both species are pollinator-dependent and do not exhibit autonomous self-pollination. In other orchid species pollinated by dipterans, particularly within species of the subtribe Pleurothallidinae and species of the genus Bulbophyllum Thouars, self-incompatibility has been reported alongside high levels of genetic variability. This pattern is noteworthy given that fly behaviour generally promotes self-pollination, as these insects tend to remain on flowers for extended periods and visit multiple flowers within the same inflorescence. The elevated genetic variability observed in these species is primarily explained by the presence of self-incompatibility mechanisms and inbreeding depression, which prevent effective self-fertilisation despite the pollinators’ behaviour (Borba and Semir 2001; Barbosa et al. 2009; Borba et al. 2011; Zhang et al. 2024).

In our observations, most flowers acted as pollen donors, while a smaller proportion received pollen, approximately half as many (see Table 2). A similar pattern was noted by Darwin (1862) and later by Reeves and Reeves (1984) in Hammarbya paludosa, where the majority of the flowers exhibited pollinia removal. In Malaxis monophyllos, a high percentage of flowers with removed pollinia has also been reported, ranging from 39.8% to 83.6%. However, the fruit set in this species was low, suggesting significant pollen loss (Jermakowicz et al. 2015, 2022). These reported values closely resemble those observed in both Malaxis parthoni and M. excavata in our study, further supporting a consistent pattern of high pollen removal but limited effective pollination in fly-pollinated Malaxis species. On the other hand, high values of Nilsson’s male efficiency index were recorded principally in Malaxis excavata, underscoring the effectiveness of fungus gnats as pollinators, which may explain the relatively high values of fruit success in both studied species. A comparable pattern has been observed in Neottia cordata, where fungus gnats contribute to a high fruit set, likely due to their pollination efficiency and high population density in the study area (Ackerman and Mesler 1979). This species is also self-compatible (Ackerman and Mesler 1979; Mesler et al. 1980), a trait that likely enhances its fruiting success (Calderon-Quispe and Singer 2024). Nonetheless, it is important to note that our study species also occur at different elevations, and thus the observed variation in fruit set could be influenced by diverse factors such as population density, elevation, or other ecological factors potentially associated with reproductive success. In our study, disentangling the relative contribution of these variables represents a limitation in fully understanding the drivers of fruiting success. Therefore, future research should explicitly evaluate these factors, especially considering that the wide distribution of the genus Malaxis makes it a particularly suitable group for testing such hypotheses.

In the species studied, the mean fruiting success was 11.24% in Malaxis parthoni and 36.01% in M. excavata. Low fruit production is common within the genus, with fruit set rarely exceeding 25% under natural conditions, as reported in M. monophyllos (Jermakowicz et al. 2015, 2022). Some species show even lower fruiting success; for example, M. massonii has been reported to exhibit fruit set as low as 1.48%. In contrast, other species such as Hammarbya paludosa can reach substantially higher values, with fruit set rates exceeding 50% (Darwin 1862). Notably, in our study, M. excavata reached exceptionally high values, with fruit set up to 85.71%, far surpassing what is typically reported for the genus and many self-incompatible orchids. This level of reproductive success is remarkable, especially considering that low fruit set—often below 10%—is characteristic of self-incompatible species, particularly those lacking floral rewards and relying on deceptive pollination strategies (Neiland and Wilcock 1998; Tremblay et al. 2005; Castro et al. 2022; Ackerman et al. 2023). In contrast, both Malaxis species studied here offer nectar as a floral reward, which may contribute to their relatively high fruiting success. When compared to other species in the genus and typical self-incompatible orchids, the reproductive output of M. parthoni and, especially, the High Andean M. excavata stands out as unusually high.

A statistically significant difference in fruiting success was observed between Malaxis excavata and M. parthoni, with M. excavata exhibiting notably higher success. This disparity could also be influenced by plant density, as individuals of M. parthoni were considerably more distanced from each other. Similar trends have been documented in M. massonii, where reproductive success was influenced by population density (Aragon and Ackerman 2001). Numerous studies have demonstrated that low population density can negatively impact pollination success, particularly in self-incompatible species that rely on cross-pollination for fruit and seed development (Schmitt 1983; Kunin 1997; Metcalfe and Kunin 2006). Even among self-compatible orchids such as Neottia cordata, reduced seed production has been reported in spatially isolated individuals, emphasising the broader implications of plant density on reproductive outcomes (Meléndez-Ackerman and Ackerman 2001).