The floral biology of the Spiranthinae was stated to be “much more complex than cursory comparisons suggest” (Salazar et al. 2018: 277). A recent review of the floral biology of Pelexia clade (Buzatto et al. 2022) states that orchids in this clade are pollinated mainly by bees, especially those from the family Halictidae, which are attracted by the nectar produced as a reward (Table 2). The flowers are, in general, small for the family, white, with floral parts covered with carpets of glandular hairs. Cyclopogon guayanensis conforms very well to this pattern but stands out due to the wide taxonomic diversity of both pollinators and visitors, and lack of visible nectar.
Bee visitation observed in Cyclopogon guayanensis closely follows the description given by Singer and Cocucci (1999) for C. diversifolius (Cogn.) Schltr., with pollinia possibly adhering to the underside of the bee labrum. These authors hypothesized that C. diversifolius, as well as the other two non-related species they studied [Campylocentrum aromaticum Barb.Rodr. and Prescottia densiflora (Brong.) Lindl.] are pollinated mainly by Halictid bees and well-adapted for “the particular swivel proboscis mechanism of the halictid bees for pollinaria removal and deposition” (Singer and Coccuci 1999: 112). In C. guayanensis, although halictid bees were present, their role as pollinators was overshadowed by the Apidae, i.e. Exomalopsis, Tetrapedia and Nomada (observed bearing pollinaria) and by stingless bees as visitors. This did not seem to be associated to any significant differences in bee body size, pollinaria attachment, or behaviour. We suggest this could be, at least in part, a product of the different bee fauna in our study area. Halictidae are highly abundant, surpassing the Apidae in several surveys (Martins et al. 2013) in subtropical and temperate regions of South America (where other studies of Cyclopogon pollinators have been done; e.g., Singer and Coccuci 1999; Singer and Sazima 1999; Galetto et al. 1997;Benitez-Vieyra et al. 2006).
Buzatto et al. (2022) noted changes in the adherence position of pollinaria on the bee labrum in the phylogenetic tree of species in the Pelexia clade using ancestral character state reconstruction. These authors hypothesized that the ancestor of the Pelexia alliance had ventrally viscous pollinia and noted that the two most basal species Coccineorchis cernua (Lindl.) Garay and Sauroglossum elatum Lindl. (respectively pollinated by hummingbirds and lepidopterans) retained this ancestral character. The shift to hymenopteran pollination was accompanied by a shift from ventral to dorsal pollinia that adhere to the ventral surface of the bee labrum in Cyclopogon comosus (Rchb.f.) Burns-Bal. & E.W.Greenw., Pachygenium bonariensis (Lindl.) Schltr., Pelexia adnata (Sw.) Poit. ex Rich., Sarcoglottis acaulis (Sm.) Schltr., and with a reversal to ancestral ventral pollinia in Brachystele unilateralis (the type species of the genus), i.e., a member of Brachystele sensu stricto. Thus, if adherence of pollinia to the inner surface of the bee labrum in C. guayanensis is confirmed in future captured bees, its transfer to Cyclopogon first suggested by molecular markers (Salazar et al. 2018) will be further strengthened and will also sustain the evolutionary scenario proposed by Buzatto et al. (2022) for this character.
When commenting on the position of Cyclopogon guayanensis in the Spiranthinae phylogenetic tree, Salazar et al. (2018: 296) noted the difference in flower size between the C. guayanensis (then Brachystele guayanensis) and Veyretia spp. and stated that: “upon close examination, it is evident that its flowers have the two-chambered nectary of Veyretia (although more shallowly so in C. guayanensis, in proportion to its noticeable reduction in flower size). This shallow two-chambered nectary was not seen.
Abundant nectar is reported for some Cyclopogon species (Singer and Cocucci 1999; Wiemer et al. 2009), but this was not the case of C. guayanensis, in which neither the capillary tube and nor paper filter methodologies were capable of collecting nectar. Nectar is secreted by the papillae on the inner surface of the labellum, particularly along its central groove, in minute quantities. Similar papillae have been detected in C. apricus (Adachi 2015) and C. dutrae (Galetto et al. 1997). Nectar secreting papillae on the inner surface of the labellum have also been found in the ghost orchid, Epipogium aphyllum Sw., subtribe Epipogiinae, tribe Gastrodieae, subfamily Epidendroideae, by Swięczkowska and Kowalkowska (2015), a northern temperate mycotrophic orchid pollinated by Bombus bees. Nectar is secreted along the central groove of the labellum of another mycotrophic orchid, European Epipactis atropurpurea Raf. (Pais and Figueiredo 1994), tribe Neottiae, subfamily Epidendroideae (Chase et al. 2015), which has, however, epidermal nectariferous tissue along the groove and not papillae.
Nectariferous tissue is composed of epidermal, specialized parenchymatic cells present on the surfaces of plant tissue that are usually either elevated or sunken (Fahn 1988; Galetto et al. 1997). In many orchids, the nectariferous tissue contains starch grains in the pre-secretory stage, which is the energy source to produce the nectar (Pais and Figueiredo 1994; Stpiczyńska and Matusiewicz 2001). Starch grains are frequent in angiosperm nectaries, and vascular bundles also can occur, although not necessarily (Fahn 1988, 1989). Nectar-secreting papillae (secreting minute quantities of nectar) have also recorded the Dactylorhiza maculata complex, tribe Orchidieae, subfamily Orchidoideae, but these are in the spur (Naczk et al. 2018) not on the labellum. Papillae secreting “floral rewards” (a mixture of carbohydrates, lipids and proteins that feed its midge pollinators - Ceratopogonidae, Diptera) have also been recorded on the labellum of the neotropical genus Trichosalpinx Luer, subtribe Pleurothallidinae, tribe Epidendreae, subfamily Epidendroideae (Bogarín et al. 2018). Thus, the exact combination of nectariferous papillae on the inner surface of a labellum that does not form a spur (Fig. 4M-N, arrows), associated to the confirmed presence of nectar sugars is, as far as we are aware, a novelty in subfamily Orchidoideae. Papillae conclusively shown to produce nectar have been recorded in the Orchidoideae, but these are all within spurs, i.e., in several genera of subfamily Orchidoideae, subtribe Orchidinae (Bell et al. 2009). In spurless Cyclopogon apricus, similar papillae have been detected and these were interpreted as nectaries (Adachi 2015). Field studies on the reproductive biology of C. apricus have been done and have shown that it secretes an exsudate considered by the authors to be nectar (Sazima and Cocucci 1999). We consider it highly likely that the secretion produced by C. apricus is nectar, and that its papillae are analogous to the nectariferous papillae of C. guayanensis.
The early literature stated that floral nectar is mostly composed of varying proportions of sucrose, glucose and fructose (Percival 1961; Baker and Baker 1983). This composition has been widely supported by later research (Chalcoff et al. 2017), although other components, such as amino acids, can also be important in the plant-pollinator interaction (Gottsberger et al. 1984; Petanidou et al. 2006; Brzosko and Mirski 2021). Baker and Baker (1983) proposed a classification of floral nectar into four types based on an “r” ratio of sucrose/fructose + glucose: 1) sucrose dominant (r > 0.99); 2) sucrose rich (r = 0.5–0.99); 3) hexose rich (r = 0.1–0.5); 4) hexose dominant (r < 0.1). Floral nectar sugar composition (and its associated ecological drivers) were recently reviewed for the angiosperms by Chalcoff et al. (2017) and for the Orchidaceae by Brzosko and Mirski (2021).
The nectar of Cyclopogon guayanensis is the second record of a hexose-dominant nectar (type 4 in the system of Baker and Baker 1983) in the Orchidaceae. The first published record of a hexose dominant nectar in the Orchidaceae is very recent, of Prasophyllum innubum D.L. Jones (Hayashi et al. 2024). P. innubum shows remarkable similarities to C. guayanense (Table 3). A taxonomically updated list of nectar types in the Orchidaceae, mostly based on Chalcoff et al. (2017) and Brzosko and Mirski (2021) with some additional records, is presented in Table 2.
Table 2
Orchidaceae with known nectar composition and their pollinator groups ordered by subfamily and tribe. Nectar composition in average values when more than one individual or flower was sampled per taxon. EP = subfamily Epidendroideae; OR = subfamily Orchidoideae; VA = subfamily Vanilloideae; Coll = tribe Collabieae; Cran = tribe Cranichideae; Cymb = tribe Cymbidieae; Dend = tribe Dendrobieae; Diur = tribe Diurideae; Neot = tribe Neottieae; Orch = tribe Orchideae; Vand = tribe Vandeae; Vani = tribe Vanilleae. * name updated to follow Plants of the World Online (https://powo.science.kew.org); r = sucrose/(fructose + glucose); nectar class follows Baker & Baker (1983).
Adapted from Brzosko & Mirsky (2021) and Chalcoff et al. (2017) with additions.
Higher taxa
|
Species, r value and nectar class
|
Visitors
|
References
|
EP, Coll
|
Calanthe angustifolia (Blume) Lindl.
r = 3.27 = sucrose dominant
|
Unknown
|
Freeman et al. (1991)
|
EP, Cymb
|
Eulophia alta (L.) Fawc. & Rendle
r = 0.8 = sucrose rich
|
Hymenoptera
(Bees)
|
Jürgens et al. (2009)
|
EP, Cymb
|
Eulophia cochlearis Lindl. *
r = 2.33 = sucrose dominant
|
Hymenoptera
(Bees)
|
Peter & Johnson (2009)
|
EP, Cymb
|
Maxillaria anceps Ames & C.Schweinf.
r = 70.86 = sucrose dominant
|
Hymenoptera
(Bees)
|
Davies et al. (2005)
|
EP, Cymb
|
Ornithidium fulgens ((Rchb.f.) L.O.Williams
r = 13.7 = sucrose dominant
|
Hummingbirds
|
Lipińska et al. (2022)
|
EP, Dend
|
Dendrobium spp (34 species from SE Asia)
sucrose dominant
|
Hymenoptera
(Bees)
|
Jia & Huang (2022)
|
EP, Neot
|
Epipactis helleborine (L.) Crantz
r (natural areas) = 1.5–2.5 = sucrose dominant
r (disturbed areas) = 0.7 = sucrose rich
|
Hymenoptera (mainly)
|
Brzosko et al. (2023)
|
EP, Neot
|
Epipactis atrorubens (Hoffm.) Besser *
r = 0.27 = hexose rich
|
Insects
(several orders)
|
Pais & Neves (1980)
|
EP, Neot
|
Neottia ovata (L.) Hartm.
r = 0.14–0.3 = hexose rich
|
Insects
(several orders)
|
Brzosko et al. (2021)
|
EP, Vand
|
Cleisostoma lecongkietii Tich & Aver. *
r = 20.89 = sucrose dominant
|
Hymenoptera
(Bees)
|
Ponert et al. (2016)
|
EP, Vand
|
Rhipidoglossum caffrum (Bolus) Farminhão & Stévart *
r = 4.26 = sucrose dominant
|
Lepidoptera
(Moths)
|
Luyt (2002)
|
OR, Cran
|
Cyclopogon dutrae Schltr.
r = 8.17 = sucrose dominant
|
Hymenoptera (Halictidae)
|
Galetto et al. (1997)
|
O, Cran
|
Cyclopogon guayanensis (Lindl.) B.M. Carvalho & Meneguzzo
r = 0.06 = hexose dominant
|
Hymenoptera, Diptera
(Apidae, Halictidae) (Syrphidae)
|
This study
|
OR, Cran
|
Goodyera repens (L.) R.Br.
r = 0.18–0.44 = hexose rich
|
Hymenoptera
(Bombus)
|
Brzosko et al. (2023)
|
OR, Cran
|
Sacoila lanceolata (Aubl.) Garay
r = 0.50 = hexose rich
|
Trochilidae (Hummingbirds)
|
Galetto et al. (1997)
|
OR, Diur.
|
Caladenia arenaria Fitzg.
r = 19.0 = sucrose dominant
|
Hymenoptera
(Wasps)
|
Reiter et al. (2019)
|
OR, Diur.
|
Caladenia colorata D.L.Jones
r = 19.0 = sucrose dominant
|
Hymenoptera
(Wasps)
|
Reiter et al. (2019)
|
OR, Diur.
|
Caladenia nobilis Hopper & A.P.Br.
r = 1.0 = sucrose dominant
|
Hymenoptera
(Wasps)
|
Phillips et al. (2020)
|
OR, Diur.
|
Caladenia paludosa Hopper & A.P.Br.
r = 2.37 = sucrose dominant
|
Hymenoptera
(Wasps)
|
Phillips et al. (2020)
|
OR, Diur
|
Caladenia robinsonii G.W.Carr.
r = 58.5 = sucrose dominant
|
Hymenoptera
(Wasps)
|
Phillips et al. (2024)
|
OR, Diur.
|
Caladenia versicolor G.W.Carr
r = 19.0 = sucrose dominant
|
Hymenoptera
(Wasps)
|
Reiter et al. (2019)
|
OR, Diur.
|
Caladenia xanthochila D.Beards. & C.Beards.
r = 100 = sucrose pure
|
Hymenoptera
(Wasps)
|
Reiter et al. (2023)
|
OR, Diur.
|
Prasophyllum innubum D.L.Jones
r = 0.01 = hexose dominant
|
Insects
(several orders)
|
Hayashi et al. (2024)
|
OR, Orch
|
Bonatea cassidea Sond.
r = 67.5 = sucrose dominant
|
Lepidoptera
(Butterflies)
|
Balducci et al. (2019)
|
OR, Orch
|
Bonatea polypodantha (Rchb.f.) L.Bolus
r = 4.34 = sucrose dominant
|
Lepidoptera
(Moths)
|
Balducci et al. (2020)
|
OR, Orch
|
Dactylorhiza incarnata (L.) Soó
r = 2.63 = sucrose dominant
|
Hymenoptera
|
Naczk et al. (2018)
|
OR, Orch
|
Dactylorhiza maculata var. fuchsii (Druce) Hyl.
r = 5.06 = sucrose dominant
|
Coleoptera,
Diptera
|
Gutowski (1990); Naczk et al. (2018)
|
OR, Orch
|
Dactylorhiza maculata (L.) Soó var. maculata
r = 4.82 = sucrose dominant
|
Hymenoptera
|
Naczk et al. (2018); Koivisto, Valius & Salonen (2002)
|
OR, Orch
|
Dactylorhiza majalis (Rchb.) P.F.Hunt & Summerh.
r = 3.63 = sucrose dominant
|
Diptera, Coleoptera Hymenoptera
|
Naczk et al. (2018)
|
OR, Orch
|
Disa cooperi Rchb.f.
r = 0.89 = sucrose rich
|
Lepidoptera
(Moths)
|
Johnson (1995); Johnson (2006)
|
OR, Orch
|
Elleanthus brasiliensis (Lindl.) Rchb.f.
r = 27.99 = sucrose dominant
|
Trochilidae (Hummingbirds)
|
Nunes et al. (2013)
|
OR, Orch
|
Gymnadenia conopsea (L.) R.Br.
r = 4.2 = sucrose dominant
|
Lepidoptera
(Butterflies)
|
Gijbels et al. (2014)
|
OR, Orch
|
Habenaria hieronymi Kraenzl.
r = 0.19 = hexose rich
|
Hymenoptera
|
Galetto et al. (1997)
|
OR, Orch
|
Habenaria gourlieana Gillies ex Lindl.
r = 5.9–22.26 = sucrose dominant
|
Lepidoptera (Hawkmoths)
|
Galetto et al. (1997)
|
OR, Orch
|
Habenaria obtusa Lindl. *
r = 27.5 = sucrose dominant
|
Lepidoptera
|
Gottsberger et al. (1984)
|
OR, Orch
|
Pelexia bonariensis (Lindl.) Schltr.
r = 1.70 = sucrose dominant
|
Hymenoptera
(Bombus)
|
Galetto et al. (1997); Buzatto et al. (2022)
|
VA, Vani
|
Vanilla hartii Rolfe
r = 5.64 = sucrose dominant
|
Hymenoptera
(Euglossa)
|
Watteyn et al. (2023)
|
Table 3
Comparative characteristics of the only two species of orchids known to produce hexose-rich nectar
Functional trait
|
P. innubum
|
C. guayanense
|
Comparison
|
Geographic distribution
|
Narrow endemic
|
Widespread
|
different
|
Habit
|
Geophyte
|
Geophyte
|
equal
|
Habitat
|
Grasslands and peatlands by streams
|
Highland grasslands, savannas and pastures
|
similar
|
Flowering populations
|
Dense
|
Dense
|
similar
|
Flowering plant size
|
30-90cm
|
9-40cm
|
overlapping
|
Flowers per spike
|
6–20 flowers
|
(9–)30–120(–160)
|
overlapping
|
Flower colour
|
Petals white or purplish
Labellum white or pink
|
Petals creamy white
Labellum yellowish cream
|
similar
similar
|
Flower size
|
6-9mm across;
Labellum 7–9 x 3-3.5mm
|
4–6 across;
Labellum 3.5–4.3 x 1.8-2.8mm
|
overlapping
different
|
Flower resupination
|
Not resupinate
|
Resupinate
|
different
|
Visitors observed with pollinaria
|
Bees: Apidae, Halictidae, Wasps (1 species)
|
Bees: Apidae, Halictidae
|
similar
|
Hexose-dominant nectar is rare in orchids (Table 2), but is the same true outside of the orchids? To address this question, we segregated all the species with hexose-dominant nectars (but at least 1% sucrose) that were herbs, shrubs or subshrubs, and pollinated by either bees or generalists from the list of 1214 angiosperm with data on nectar sugars compiled by Chalcoff et al. (2017). This segregate list recorded 45 species in ten botanical families that were widely scattered across the angiosperm phylogenetic tree. We then investigated these 45 species in Plants of the World (POWO 2024) and reliable identifications in the I-naturalist site and found that they followed a loose pattern (with some exceptions). White, cream, yellow or orange were the common colours, while pink, violet or purple were rare, and red was absent. Most of them had small, densely aggregated flowers, but there were some exceptions to this norm (e.g., Tecoma stans (L.) Juss. ex Kunth). Virtually all were temperate or subtropical plants, although the plants listed by Chalcoff et al. (2017) were mostly tropical. In a semi-desertic, frost-prone scrubland in Patagonia, presumably pollinator hostile, most plants had hexose-dominant nectars and were bee-pollinated (Bernardello et al. 1999). We propose the term ‘modest’ pollination syndrome be applied to orchids that are: geophytic, have small, pale flowers (probably below 1 cm), offer imperceptible quantities of hexose-dominant nectar on an exposed labellum and attract a wide range of insect visitors.
Considering the high species richness of the Orchidaceae, nectar records are surprisingly few (Brzosko et al. 2021). Sucrose-dominant nectar predominates both in the angiosperms (56.8% of 1214 species; Chalcoff et al. 2017) and in the Orchidaceae (80.5% of 43 species; Brzosko and Mirski 2021). Pollinator group was statistically confirmed in both studies as the main ecological driver, followed by climate, i.e., latitudinal climatic zone (for the angiosperms) or climate type (for the orchids).
Baker and Baker (1983), studying nectar composition in 765 species of angiosperms, found that 44% of species pollinated by short-tongued bees produced hexose-dominant nectar, while only 6% of species pollinated by long-tongued bees had this type of nectar. Wilmer (2011) has argued that sucrose/hexose ratios may have more to do with maintaining optimum nectar concentration and viscosity than pollinator preference. Pure sucrose nectar dries out faster than pure hexose nectar at the same humidity levels (Corbet et al. 1979). Therefore, hexose-rich nectar might be preferentially produced by exposed nectaries (pollinated by short-tongued insects), and sucrose-rich nectar by deep, hidden nectaries (pollinated by long-tongued insects). However, Brzosko and Mirski (2021) found no statistical differences of sugar composition between orchids that offer spur nectar and open nectar, although these authors noted that generalist species were under-represented in the available data. It is possible that when nectar is produced not only by exposed nectaries, but in minute quantities in plants that form dense populations (such as was recorded in Cyclopogon guayanensis and Prasophyllum innubum), natural selection will favour high-hexose nectar that is more resistant to desiccation. It is worth mentioning in this context that C. guayanensis flowers in early wet season when mean daily humidity is very high and is also a generalistic species, with a wider taxonomic range of bee pollinators than any other species in either the Pelexia or the Cyclopogon clades sensu Salazar et al. (2018) with known pollinators (Supporting information, Table S2).
The glandular trichomes that abound on Cyclopogon guayanensis inflorescences and outer surfaces of the flowers are lipid secreting (Fig. 4L); these lipids may be acting as aromatic attractants as recorded in C. elatus (Sw.) Schltr., by Wiemer et al. (2009) or as rewards for oil-collecting or scent-collecting bees. Pachygonium pteryganthum (Rchb. f. & Warm.) Schltr., a member of the Pelexia clade, is pollinated by oil-collecting female Centris (Melacentris) bees (Buzatto et al. 2022). In our study area, male Tetrapedia bees are known to collect oil from Malpighiaceae flowers (Cappellari et al. 2012) and we observed a male Tetrapedia bee visiting the flowers of C. guayanensis, but no such oil-collecting activity was observed. The Tetrapedia bee was foraging for nectar, and is presumed to be a pollinator, as it gathered three pollinaria during visits, which it subsequently tried to remove with front legs. Volatile chemicals (scent) can also be collected for mating purposes by male Neotropical orchid bees, i.e. Euglossini (Carvalho-Filho 2010). Despite the high density of glandular trichomes on the C. guayanensis inflorescences, there was no evidence of this kind of reward collecting by pollinators in our study; Euglossini visitors were not observed in the field and glandular trichomes were intact, not destroyed by visitors, with the cuticle on the heads undamaged (Fig. 4G, J–L).
A sweet scent was reported in the flowers of Cyclopogon congestus (Vell.) Hoehne (Singer and Sazima 1999), C. diversifoIius (Cogn.) Schltr. (= C. apricus (Lindl.) Schltr.) (Singer and Cocucci 1999), C. elatus (Wiemer et al. 2009) and Veyretia hassleri (Cogn.) Szlach. (Supporting Information, Table S2). In C. congestus, however, the scent, which was perceived from the morning to the evening hours, was produced by apical callosities on the inner surface of the labellum and lateral sepals (Singer and Sazima 1999), and is thus not analogous to the multicellular, glandular trichomes observed on the outer surface of the corolla in C. guayanensis.
In Cyclopogon elatus, osmophores are reported as trichomes on the outer surface of the labellum (Wiemer et al. 2009) such as was recorded by us in C. guayanensis. However, in C. elatus these trichomes are unicellular, while in C. guayanensis they are pluricellular, with stalk and head, but also occur on the outer surface of the labellum (as well as on other inflorescence surfaces). Also, in C. elatus, only the osmophores of open flowers release scent, which is absent in buds and withered flowers (Wiemer et al. 2009). Thus, C. guayanensis also differs from C. elatus in this aspect, since secretion appears to persist in senescent flowers (Fig. 4D).
The combined evidence suggests that the glandular trichomes on the inflorescence and outer surface of the flowers of Cyclopogon guayanensis are probably osmophores. However, the scent detected in C. guayanensis was extremely faint and was in fact only perceived by one of us; therefore, the function and composition of the lipids secreted by these trichomes should be the subject of further investigation. The external position on the flower is unusual for osmophores; the inner surface of the labellum is the most widespread scent-releasing surface in orchids (Wiemer et al. 2009). In fact, very few angiosperms have osmophores on the outer surface of the flowers, as seems to be the case in Cyclopogon; it was also reported in Cotylolabium lutzii (Pabst) Garay, another genus of the Spiranthinae that is sister taxon to the rest of the Subtribe (Borba et al. 2014). Perhaps the presence of osmophores on the outer surface of the flowers is a common characteristic of this subtribe; this hypothesis would require further investigation.
The cuticle allows both scent diffusion in osmophores and secretion releasing in nectaries; the cuticle expands and there is secretion accumulation in the periplasmic space; this phase was clearly recorded in the nectariferous papillae of Cyclopogon guayanensis in our study (Fig. 4M–N, arrows). In the secretory labellum tissue of Epipactis atropurpurea, the nectar is released by disruption of the epidermal cell cuticle in the secretory stage (Pais and Figueiredo 1994). In the nectariferous papillae of cotton (Gossypium hirsutum L., Malvaceae), however, a two-layered cuticle is detached from the wall in the secretory papillae and the nectar crosses this structure (Eleftheriou and Hall 1983); further studies are needed to clarify the nectar releasing mechanism in C. guayanensis.
Raphid-rich idioblasts have been presumed to be herbivore deterrents (Molano-Flores 2001). Raphid containing idioblasts are not widespread in dicotyledons (Cutler et al. 2008) but are found in many monocotyledons (Fahn 1988) and were observed in the bracteoles (Fig. 4F–G), sepals and petals (Fig. 4H) of Cyclopogon guayanensis. Raphid-rich idioblasts occur in several species of orchids, both in subfamilies Orchidoideae and Epidendroideae. In the Orchidoideae, they are found in the floral parts of Habenaria gourlieana Gillies ex Lindl. (Galetto et al. 1997), and in the hypochilum of Cotylolabium lutzii (Borba et al. 2014) – also a member of subtribe Spiranthinae and were noted to be common in this subtribe by Adachi (2015), based on studies of floral anatomy of Cyclopogon apricus, Mesadenella cuspidata (Lindl.) Garay, Sarcoglottis fasciculata, and Sauroglossum elatum. In subfamily Epidendroideade, they occur in the ghost orchid, Epipogium aphyllum (Swięczkowska and Kowalkowska 2015). However, the density of idioblasts is noticeably higher in the sepals and petals of C. guayanensis (Fig. 4H, arrowheads) than in most of these orchids. In many orchids of the subtribe Pleurothallidinae, prismatic crystals occur as idioblasts in the floral parts and have been associated to visual attraction of pollinators (Bogarín et al. 2018), but herbivore dissuasion is another possible role.
The fact that herbaceous or shrubby plants with hexose-dominant nectar and generalistic pollination systems are usually confined to temperate or subtropical environments (Chalcoff et al. 2017) but appeared in a tropical savanna grassland species is interesting. In the highly seasonal Cerrado biome where our study was carried out, 90% of annual precipitation occurs between the months of October and April, followed by a dry season lasting between 4 and 7 months (Bustamente et al. 2012); in the Distrito Federal, c. 50% of annual precipitation occurs between December and February (Anunciação et al. 2014). We tentatively suggest that a short favourable reproductive season is the driver behind this suite of characters. Its similarity to Epipogium aphyllum suggests convergence with these orchids (Swiwczkowska and Kowalkowska 2015).
Our theory for the unusual characteristics of C. guayanensis is that, as a therophytic orchid that inhabits fire-prone, highly seasonal savanna grasslands, it will presumably be under two selective forces: 1) lack of competition, since grasslands are potentially resource-poor landscapes for pollinating insects and will tend to have low pollinator density, therefore favouring a “take what you can get” generalist pollination strategy offering scanty rewards; 2) a limited growing and flowering season (during the short wet season) which would put a premium on economy of energy, i.e., favour small plants with small flowers, small fruits, and minimalistic pollinator rewards.
The floral biology (position of the viscidium and pollinator behaviour) and micromorphological data gathered in this study (external glandular hairs, raphid-rich idioblasts, internal nectariferous papillae) support the transfer of Brachystele guayanensis to Cyclopogon (Meneguzzo et al. 2024). Our hypothesis is that the C. guayanensis – Veyretia clade is a lineage within Cyclopogon (a predominantly forest genus well adapted to pollination by halictid bees) that has moved into open habitats with accompanying pollinator shifts (i.e., from Halictidae to Apidae and possibly from Halictidae to Megachilidae, respectively; see Table S2), although this would need further investigation of the floral biology of Veyretia. Studies of species of Brachystele s.s. would obviously also be desirable to confirm parallel evolution.
To conclude, on a more general note, we would suggest further testing for trace nectar in orchids that apparently offer no rewards. We cite verbatim a paragraph from the study by Shrestha et al. (2020: 5) on rewardlessness in orchids that reflects our beliefs: If neural mechanisms of reward detection can be exploited and are widespread among floral visitors, the use of such stingy rewards by orchids lacking visible nectar might also be widespread. Pollinator manipulation by trace rewards might still be considered largely deceitful, but the nature of the deceit would be more complex and more subtle than in the case of complete absence of floral sugars. A paucity of reward may actually encourage appropriate flower-handling by visitors.