Small Leaves, Big Diversity: Citizen Science and Taxonomic Revision Triples Species Number in the Carnivorous Drosera microphylla Complex (D. Section Ergaleium, Droseraceae)

Simple Summary A novel taxonomic treatment is provided for the Drosera microphylla complex, which is a group of closely related carnivorous plants endemic to southwest Western Australia. The species that comprise this group are generally rare, micro-endemic, and are potentially threatened by habitat destruction and illegal collection. Resolving the taxonomy and systematics of this complex has been critical to the accurate assessment of its component species under conservation legislation. Following two decades of fieldwork in Western Australia, studies of preserved plant collections, and crucial contributions by citizen scientists and social media, we establish here that the Drosera microphylla complex comprises nine distinct species, three times the number previously recognised. Four species are here described and illustrated as new to science. Two previously described varieties are here re-circumscribed as distinct species in light of their rediscoveries via social media posts, allowing them to be studied for the first time since they were described more than 100 years ago. We provide examples from the genus Drosera for the impact of social media and citizen science on taxonomic work and biological conservation. This work demonstrates the great potential that citizen science has in supporting rapid advances in taxonomic knowledge in the face of extinction crises worldwide. Abstract The carnivorous Drosera microphylla complex from southwest Western Australia comprises a group of rare, narrowly endemic species that are potentially threatened by habitat destruction and illegal collection, thus highlighting a need for accurate taxonomic classification to facilitate conservation efforts. Following extensive fieldwork over two decades, detailed studies of both Australian and European herbaria and consideration of both crucial contributions by citizen scientists and social media observations, nine species of the D. microphylla complex are here described and illustrated, including four new species: D. atrata, D. hortiorum, D. koikyennuruff, and D. reflexa. The identities of the previously described infraspecific taxa D. calycina var. minor and D. microphylla var. macropetala are clarified. Both are here lectotypified, reinstated, and elevated to species rank. A replacement name, D. rubricalyx, is provided for the former taxon. Key morphological characters distinguishing the species of this complex include the presence or absence of axillary leaves, lamina shape, petal colour, filament shape, and style length. A detailed identification key, comparison figures, and a distribution map are provided. Six of the nine species are recommended for inclusion on the Priority Flora List under the Conservation Codes for Western Australian Flora and Fauna.

Etymology: The specific epithet is derived from the Latin atratus (=blackened) and refers to the very dark red to blackish red flower colour of this species, which is the darkest petal colour known in the genus Drosera (under some lighting conditions appearing almost black).
Taxonomic notes: Drosera atrata is arguably the most morphologically distinct species within the D. microphylla complex, exhibiting a unique leaf and inflorescence morphology not found in any other species in this group. While the presence of axillary leaves in the axils of all cauline leaves (thus having groups (sometimes incorrectly called "whorls") of 2-5 leaves at each node) is paralleled in other species of D. section Ergaleium (e.g., in D. macrantha Endl. and D. menziesii R.Br. ex DC.), its occurrence in D. atrata is unique within the D. microphylla complex. Drosera hortiorum, D. macropetala, and D. rubricalyx feature similar (but relatively shorter-petioled) axillary leaves in their uppermost nodes, but, in these species, they are never present in all nodes. In D. atrata, carnivorous axillary leaves are frequently found even in association with the uppermost cataphylls ( Figure 1G).       The petal colour of Drosera atrata is the darkest within the genus with only the tropical rainforest species of north-eastern Australia (D. section Prolifera C.T.White) producing similarly dark red or dark-pink flowers, in particular certain forms of D. adelae F.Muell. While some Australian pygmy sundews (D. section Bryastrum Planch.) and several members of the South African D. section Ptycnostigma Planch. produce almost completely black petal bases, these are always paired with a relatively bright colour that comprises the largest part of the petal [28]. In contrast, the dark colour of D. atrata is relatively uniform across the entire petal. However, this colour often appears even darker (almost black) when viewed at certain angles. Given the deeply concave petal shape, this often results in the appearance of especially dark areas near the petal margins ( Figure 1C,E,F,L) or bases ( Figure 1D; here, the strongly reflexed and deeply concave petals result in the bases being viewed at an angle while the margins are viewed ± perpendicular).
The seeds of D. atrata differ from all other species of this affinity except D. calycina by their falcate to allantoid shape (narrowly obovate, narrowly clavate, or narrowly obtrullate with truncate upper end (=nail-or pin-shaped) and more or less terete in the remainder of species from the D. microphylla complex), often with the funiculus still attached to the funicular seed end as a pale brown discus ( Figure 5). Only seeds of D. calycina are somewhat more similar to those of D. atrata in being slightly falcate and flattened. Seed morphology has already been shown to be a reliable taxonomic tool for species delimitation in some other species complexes of tuberous Drosera [1].
Distribution and habitat: Drosera atrata is known from eleven locations between Warradarge in the north and Badgingarra in the south ( Figure 3). It occurs in low kwongan heath on the upper slopes of lateritic hills in poorly drained sandy clay with laterite.
Phenology: Flowering has been recorded from May to August. Conservation status: Recommended for listing as Priority Three (poorly known species) under Conservation Codes for Western Australian Flora and Fauna (Western Australian Herbarium 1998-, https://florabase.dpaw.wa.gov.au/ (accessed on 6 December 2022)). It is assessed as Vulnerable (VU) under IUCN criterion D1 following IUCN [29]. The populations of D. atrata frequently comprise extremely small population sizes of just 1-30 plants. Only a single larger population of ca. 200 plants is known from an unprotected road reserve near Warradarge. Six of the eleven known locations occur on land managed by the Western Australian Department of Biodiversity, Conservation and Attractions (DBCA). Unlicenced collectors and illegal commercial/horticultural trade could pose a threat to D. atrata in the future given its extremely small population sizes and the tendency for poachers to target rare carnivorous plant species to supply a demand driven by the horticultural market and carnivorous plant collectors in particular [16]. Further surveys are recommended to gain a better understanding of this taxon's biology, distribution, number and size of populations, and to identify additional potential threats.
Etymology: The specific epithet is derived from the Latin calycinus (having a welldeveloped calyx) and was selected by Planchon [17] to refer to the very large sepals/calyx of this species.
Taxonomic notes: Drosera calycina can easily be distinguished from the remainder of the D. microphylla complex (especially in herbarium material) by the combination of ± straight petioles (which are only arched near the tip; Figure 6C,E) with the absence of any axillary leaves. It is morphologically similar to D. microphylla, D. hortiorum, D. macropetala, and D. rubricalyx. It can be distinguished from D. microphylla by (contrasting characters in parentheses) (1) its lamina shape, which is reniform or orbiculate with flattened, often truncated upper margin (lamina orbiculate or sometimes orbiculate with very slightly flattened upper margin); (2) its comparatively large flowers with a corolla diameter of 14-20 mm (corolla diameter 8-15 mm); (3) its petal colour, which is deep red in inner half transitioning to purplish red in outer half (petals reddish orange with deep red bases); (4) its stamen length, which reaches 4.0-6.5 mm (stamens 2.5-3.5 mm long); and (5) its styles, which do not extend laterally beyond the filaments and have deep red stigmas (styles laterally extending beyond the filaments with reddish purple stigmas).
Drosera calycina is further distinguished from D. hortiorum, D. macropetala, and D. rubricalyx by (contrasting characters in parentheses): (1) its solitary leaves (leaves of upper 1-9 nodes in groups of 2-5 due to the presence of usually two shorter axillary leaves); (2) its lamina shape, which is reniform or orbiculate with flattened, often truncated upper margin (lamina orbiculate or orbiculate with slightly flattened upper margin); and (3) its petal colour, which is deep red in inner half transitioning to purplish red in outer half (petals deep red in inner half, dark purplish red in outer half in D. hortiorum; white with deep red bases in D. macropetala; or deep red in inner half, deep pink in outer half in D. rubricalyx; Figure 4). The four species are also ecologically and geographically well separated ( Figure 3). While D. calycina has been observed growing within a few hundred metres of D. hortiorum near Glen Forrest, just east of Perth, they likely do not co-occur due to their different habitat requirements. In this area, Drosera calycina is restricted to laterite soils in Jarrah forests while D. hortiorum grows in clay loam around granite slopes and boulders.
The distinctive lamina shape of Drosera calycina ( Figure 6E) has been described as "subtruncate" (Planchon, in annot. K000215039), "suborbiculate-lunate" by Planchon [17], or "crescent-shaped and/or broadly reniform" by Lowrie [1] (the latter likely included both D. hortiorum and D. rubricalyx in his description of D. calycina). In all other species but D. atrata, the lamina is usually entirely orbiculate or orbiculate with a slightly flattened upper margin but not truncated. Only D. atrata also often produces reniform or truncated laminae but that species is readily distinguished by the presence of axillary leaves in all nodes, its sparsely branching styles extending laterally beyond the filaments, and its very dark red to blackish-red petals ( Figure 4). Drosera atrata and D. calycina additionally share a falcate to allantoid (i.e., slightly depressed and curved) seed shape, which is another taxonomically informative character to distinguish these two species from the remainder of the D. microphylla complex, which mostly have straight, pin-shaped, or bone-shaped seeds ( Figure 5).
Drosera calycina was previously illustrated by Erickson [22] (p. 40, drawing 2) as "D. microphylla var. macropetala". While both D. calycina and D. macropetala indeed are very tall, large-flowered plants, the petals of D. calycina are never "drying palish" as stated by Erickson [22] [12] (p. 65)) is evidently based on several different specimens. While most of the illustration matches the cited specimen A. Lowrie 3043 (PERTH 08988110!, MEL 2443236A!), the presence of axillary leaves on the habit drawing A is puzzling. Crucially, none of the individuals of the A. Lowrie 3043 collection feature axillary leaves. The specimens also lack tubers, in contrast with the illustration. It is therefore possible that Lowrie's illustration also incorporates specimens of either D. hortiorum, D. macropetala, or D. rubricalyx, all of which have axillary leaves in their upper parts, i.e., the author seems to have used some artistic licence. Phenology: Flowering has been recorded from August and September. Conservation status: Not eligible for Conservation Code listing and Least Concern (LC) following Cross [31]. Drosera calycina is relatively common in its preferred Jarrah forest habitat and at least twenty localities have been recorded, most of which are on land managed by the Western Australian Department for Biodiversity, Conservation and Attractions (DBCA). While frequently occurring in small population sizes of <50 plants, at least two large populations of >200 plants are known to exist. Unlicensed collection by plant collectors may represent a threatening process but regular monitoring of several populations between 2019 and 2022 indicated there are no current threats to this taxon (T. Krueger pers. obs.).
Notes on the lectotypification: Lectotypification of D. calycina is required as Planchon [17] did not select as the type a single specimen out of Drummond's gathering (J. Drummond n. 1). He cited two duplicates that he had studied (constituting syntypes), namely "Drummond in herb. Hook. et Soc. Linn. Londres" [17] (p. 299), which is the specimen from Herbarium Hookerianum (K000215039) and the one from the Linnean Society of London (Herbarium LINN, some specimens have been transferred to BM [32], and some apparently also to K, see below). However, the second specimen could not be found either at LINN or at BM and it may indeed be lost. However, it is also possible that this second specimen was transferred to K when the "Herbarium Australiense" specimens of the Linnean Society of London herbarium were included in the Kew collections in 1915, including Drummond material ("Herbarium Australiense, presented by the Linnean Society, 1915"; Anonymous in annot. K000843361 photo!). As no other matching specimen could be found at K, this would likely mean that this second specimen was added to the same sheet, which is now K000215039, and that the two plants represented there are indeed the two syntype specimens cited by Planchon, housed at different herbaria at the time. While it might seem counterintuitive to combine specimens this way, the practice was not uncommon at that time and another Drummond collection, J. Drummond n. 282, which belongs to the Drummond V collection and represents a different species, D. microphylla, was even added to this same sheet at a later date.
These two specimens of J. Drummond n. 1. cited by Planchon [17] by definition constitute syntypes, hence lectotypification is required (ICN Arts. 7.11 and 9.17 [26]), even if both are found mounted together on the same herbarium sheet today. K000215039 holds a handwritten personal annotation by Planchon, which represents the sketch of a differential diagnosis noted by the author: " Drosera  Marchant et al. [23] unnecessarily selected an "isotype" at K for D. calycina, referring to Planchon's type. In addition, Marchant incorrectly annotated a different specimen at Montpellier Herbarium (MPU1254140) as the "holotype" (Marchant 1985 in sched.) but this was never effectively published (as required by ICN Art 7.10 [26]). Choosing MPU1254140 would have been an incorrect type designation in any case because this specimen is not original material of Drosera calycina. It was not cited by Planchon in 1848 [17] and it was not ascribed to the name D. calycina by Planchon himself (evident from the label on MPU1254140 in Planchon's hand, which reads, "Drosera calycina ? Planch."), and it was annotated by Planchon after he had described D. calycina in 1848 (Planchon became assistant professor at Montpellier in 1853 and director of MPU in 1881, while from 1844-1848 he was based at Kew [32]). Thus, MPU1254140 cannot constitute a type for D. calycina and, in fact, represents a different species (likely D. rubricalyx, see "Notes on Drummond's type collection" under D. rubricalyx). Lowrie [1], simply referring to Marchant et al. [23], also incorrectly lists the MPU specimen as the "holotype".
It should also be noted that J. Drummond n. 1 was likely the only gathering of the D. microphylla complex collected by James Drummond that was available to Planchon in 1848 for his revision of Droseraceae. The other specimens J. Drummond coll. V n. 282, J. Drummond coll. VI n. 109, and J. Drummond coll. VI n. 110 were collected later (collected in 1847 or 1848 and dispatched to Europe in 1849 for coll. V; and collected in 1850 or 1851 for coll. VI [30]) and, thus, were not considered by Planchon in his 1848 taxonomic treatment of Drosera [17].
Notes on Drummond's type collection: Unfortunately, neither the type material nor Drummond's scarce publication records provide any evidence for where exactly in the former Swan River colony the type collection was made. The contemporary botanist Diels [33] (p. 50, literally translated) has already asserted that, "in short, one will never know [exactly] where Drummond's plants were collected; and just in rare cases it can be achieved by the aid of literature to pinpoint at least the approximate habitat". Two of these cases, for which the authors of the present work could trace back the locus classicus from Drummond's historic notes [18,19], are D. macropetala and D. rubricalyx (see "Notes on Drummond's type collection" under the headings of the two respective species).

Etymology:
The specific epithet refers to the Esperance region of southern Western Australia where this species is endemic.
Taxonomic notes: Drosera esperensis is morphologically similar to D. koikyennuruff, D. microphylla, and D. reflexa. It is distinguished from these three species by (contrasting characters in parentheses): (1) its tendency to form dense, clonal, mat-like colonies (plants not colony forming or only forming relatively sparse [not mat-like] colonies); (2) its ± linear filament shape (filaments increasing in width towards apex); (3) its petal colour, which is white with a pale purplish red base (dark red in D. koikyennuruff, reddish orange with deep red base in D. microphylla, or purplish pink with deep red base in D. reflexa; Figure 4); and (4) its style and filament colour, which is white with red or purplish red base (styles and filaments red, deep red, purplish red, or reddish purple). The distinctive white petal colour of D. esperensis is paralleled in D. macropetala, from which it can be distinguished by (contrasting characters in parentheses): (1) its mostly solitary leaves (leaves of upper 1-9 nodes in groups of 3 due to the presence of two shorter axillary leaves); (2) its tendency to form dense, clonal, mat-like colonies (plants not colony forming); (3) its ± linear filament shape (filaments dilated towards apex); and (4) its style colour, which is white with red or purplish red base (styles very dark red).
Drosera esperensis is geographically the most isolated species of the D. microphylla complex, occurring ca. 350 km east of the nearest confirmed population of D. microphylla ( Figure 3; for discussion of the more proximate collection from Hopetoun, see Taxonomic notes under D. microphylla).
Plants from the Cape Arid area have been observed to frequently produce axillary leaves. Further studies of these populations are recommended to determine whether they represent a taxon distinct from D. esperensis (the type of which was collected from the Cape Le Grand area, where this species almost never produces axillary leaves).
Drosera esperensis was previously illustrated by Gibson [14] (p. 41) and Lowrie [1] (p. 435). Distribution and habitat: Only known to occur within the Cape Le Grand and Cape Arid National Parks, east of Esperance ( Figure 3). Grows in wet, mossy areas on and near granite hills in sandy clay or peat.
Phenology: Flowering has been recorded from August to October. In exceptionally wet habitats or seasons, flowering has been observed to continue until at least December (T. Krueger pers. obs.).      (1) its filament shape, which are only slightly dilated towards apex, 0.3-0.5 mm wide near the apex (filaments strongly dilated towards apex, 0.5-0.9 mm wide near apex); (2) its tentacle stalk colour, which is greenish yellow (tentacle stalks red in lower half, greenish yellow in upper half, or red throughout); and (3) its filament colour, which is deep red (filaments deep red in lower half, white, or sometimes red in the upper half). Drosera hortiorum further shares morphological similarities with D. calycina Planch., from which it is distinguished by (contrasting characters in parentheses): (1) the presence of two smaller axillary leaves in the axils of the upper 1-7 cauline leaves (all cauline leaves solitary); (2) its lamina shape, which is orbiculate or orbiculate with a slightly flattened upper margin (lamina reniform or orbiculate with flattened, often truncated upper margin); (3) its filament shape, which only slightly dilated towards apex, 0.3-0.5 mm wide near apex (filaments strongly dilated towards apex, 0.5-1.1 mm wide near apex); and (4) its straight, pin-to bone-shaped seeds (seeds flattened, slightly falcate to slightly allantoid).
Etymology: The specific epithet honours Fred Hort (1937-) and Jean Hort (1952-), enthusiastic field botanists, nature photographers, and volunteers at the Western Australian Herbarium who found this species at the Wandoo National Park type location in 1987 and brought it to the attention of the authors of the present work. Their prolific collections from the eastern Darling Range have led to the recognition of many new species, several of which have already been named in their honour (e.g., [35][36][37][38]).
Taxonomic notes: The presence of axillary leaves in the upper parts of the stem, as well as seed characters ( Figure 5), link D. hortiorum to the morphologically similar D. macropetala and D. rubricalyx. However, its corolla is of a much smaller size and its distinctive dark purplish red petal colour easily distinguishes it from these two species (Figure 4). In addition, all three species are geographically well separated, with D. macropetala and D. rubricalyx occurring well north of Perth while D. hortiorum is only known from areas to the east and south-east of Perth ( Figure 3).
Despite its usually much smaller size, the corolla shape and colour of D. hortiorum closely resembles that of D. calycina. Both species further occur in close geographic proximity ( Figure 3). However, D. hortiorum is easily distinguished from D. calycina by the presence of axillary leaves in the upper parts of the stem (D. calycina has solitary leaves and always lacks axillary leaves). While D. hortiorum has been observed growing within a few hundred metres of D. calycina near Glen Forrest, they do not co-occur syntopically due to their different habitat requirements. In that area, D. calycina is restricted to laterite soils in Jarrah forests while D. hortiorum grows in clay loam around granite slopes and boulders.
A photograph of D. hortiorum was published in 1987 by Lowrie [12] (p. 67) who, at the time, treated all taxa of the complex under D. microphylla. In his 2014 taxonomic treatment, Lowrie likely included D. hortiorum under D. calycina, as he described the presence of axillary leaves for this species ("sometimes forming leaves in groups of 2 to 3 in the upper parts" [1] (p. 354)). Drosera hortiorum is further illustrated in Drosera of the World [15] (p. 228), but with an erroneous location description (Badgingarra). The pictured plant actually represents a specimen cultivated by G. Bourke and originated from the late Allen Lowrie.
Distribution and habitat: Known from Glen Forrest, Wandoo National Park (near York, east of Perth) and two additional sites in the wheatbelt region near York and Wickepin ( Figure 3). In the western part of its range, D. hortiorum appears to be associated with low granite outcrops and granite slopes where it grows in poorly drained clay loam with Borya sp. In the eastern part of its range, D. hortiorum has been recorded from within and near shallow drainage channels and moist sandplains in sandy clay.
It is curious to note that D. hortiorum has been observed in such a wide range of different habitats, as this is unusual for the complex. Only D. microphylla is also known from very different types of habitat.
Phenology: Flowering has been recorded from June to September. Conservation status: Recommended for listing as Priority Two (poorly known species) under Conservation Codes for Western Australian Flora and Fauna (Western Australian Herbarium 1998-; https://florabase.dpaw.wa.gov.au/ (accessed on 6 December 2022)). Data deficient (DD) following IUCN [29]. Diagnosis: Drosera koikyennuruff is morphologically most similar to D. microphylla Endl., from which it is distinguished by (contrasting characters in parentheses): (1) its much earlier flowering time from June to July (flowering from August to October); (2) its dark red petal colour (petals reddish orange with deep red bases); (3) its deep red stigma colour (stigmas reddish purple), (4) its yellowish green tentacle stalk colour (tentacle stalks red or red in lower half with upper half yellowish green); and (5) its preference for relatively dry sandy habitats in open Mallee woodlands (mossy wet habitat areas on and near granite outcrops, seasonally wet swamps, or rocky mountain slopes). It is further distinguished from the morphologically similar D. reflexa G.Bourke & A.S.Rob. by (contrasting characters in parentheses): (1) its sparse populations, which are not colony-forming (plants forming dense populations via adventitious stolons); (2) its petal shape, which is narrowly obovate to broadly spathulate (petals obovate to very broadly obovate); (3) its dark red petal colour (petals purplish pink with deep red base); (4) its yellowish brown to yellowish green sepal colour (sepals red to purplish red); and (5)    Etymology: The specific epithet refers to koikyennuruff, the Noongar Aboriginal name for the Stirling Range, where this taxon occurs. The name means "mist over hills" [39].
Taxonomic notes: The overall habit as well as petiole, lamina, and style shape of D. koikyennuruff indicate that it is morphologically most similar to D. microphylla. Both species grow in close proximity at sites in Stirling Range National Park but favour a different habitat type. While D. koikyennuruff grows in low-lying areas with sandy soils in open Mallee woodlands, D. microphylla appears (in this area) to be restricted to the middle and upper slopes of the Stirling Range mountains where it typically grows in rocky or lateritic soils. In addition, the two species also differ phenologically and thus are reproductively isolated by non-overlapping flowering times, with D. koikyennuruff flowering from June to July while D. microphylla flowers from late August to October. Drosera koikyennuruff is easily distinguished from D. microphylla by its dark red petal colour (D. microphylla has reddish orange petals with deep red bases). The flower colour of the type specimen A. Rose 1029 (PERTH 05812402!) is denoted as "burgandy" (burgundy), which is an apt description for the distinctive dark red petal colour of this species.
Diels 3009 (Plantagenet: westlich des Sucky Peeks [west of "Sucky Peek" =Sukey Hill], B 10 0755996!) represents an intriguing collection from near Cranbrook (marked with "1" in Figure 3). The exceptionally small and consistently single-flowered plants were collected in late May, which is potentially within the flowering time of D. koikyennuruff. In addition, the petal colour is described by Diels as "dunkelkarmin" (dark carmine/dark crimson), which might match the dark red flower colour of D. koikyennuruff. However, the plants are overall much smaller and appear to have shorter styles. Since this population could not be re-located by the authors prior to submission, it is not currently known whether it represents D. koikyennuruff or a closely allied, undescribed species. It is thus not included under D. koikyennuruff in the present work.
Distribution and habitat: Drosera koikyennuruff is only known from two locations, one in Stirling Range National Park and one from nearby Woogenellup (Figure 3). It grows in low heath amongst Mallee scrub in sandy clay soils.
Etymology: The specific epithet, from the Greek macro-(=large) and petalum (=petal), refers to the comparatively large petals of this species. Indeed, with a length of up to 11 mm and a width of up to 7.5 mm, D. macropetala produces the largest petals known in the D. microphylla complex. Only D. calycina sometimes produces similar-sized (but usually distinctly narrower) petals.
Taxonomic notes: Drosera macropetala is morphologically similar to D. calycina, D. hortiorum, and D. rubricalyx from which it can be quickly and reliably differentiatedeven in herbarium material-by its petal colour, which is white with a deep red base (petals deep red in inner half transitioning to purplish red in outer half in D. calycina; deep red in inner half transitioning to dark purplish red in outer half in D. hortiorum; deep red in inner half transitioning to deep pink in outer half in D. rubricalyx). It is further distinguished from D. calycina by (contrasting characters in parentheses): (1) the presence of two smaller axillary leaves in the axils of the upper 1-9 cauline leaves (all cauline leaves solitary); (2) its lamina shape, which is orbiculate or orbiculate with slightly flattened upper margin (lamina reniform or orbiculate with flattened, often truncated upper margin); and (3) its straight, pin-like seeds (seeds flattened, falcate in D. calycina). It is further distinguished from D. hortiorum by (contrasting characters in parentheses): (1) its much larger corolla diameter of 13-22 mm (corolla diameter 8-11 mm) and (2) its filament shape, which is markedly dilated towards apex, 0.5-0.9 mm wide near apex (filament width only slightly dilated towards apex, 0.3-0.5 mm wide near apex). Drosera macropetala is further distinguished from D. rubricalyx by (contrasting characters in parentheses): (1) its broader petals, which are 4.5-7.5 mm wide (petals 3.3-4.8 mm wide); (2) its usually much longer peduncles, which are 1.3-4.2 cm long (peduncles 0.8-2.2 cm long); (3) its filament shape, which is markedly dilated towards apex, 0.5-0.9 mm wide near apex (filaments only slightly dilated towards apex, 0.4-0.6 mm wide near apex); and (4) its yellowish brown or red sepals, which are not strongly contrasting the yellowish green or red stem colour (sepals red, strongly contrasting the yellowish green stem). The distinctive white petal colour of D. macropetala is paralleled in D. esperensis, from which it can be distinguished by (contrasting characters in parentheses): (1) the presence of two smaller axillary leaves in the axils of the upper 1-9 cauline leaves (all cauline leaves usually solitary); (2) plants not colony forming (plants forming dense, mat-like colonies); (3) its filament shape, which is markedly dilated towards apex, 0.5-0.9 mm wide near apex (filaments ± linear, 0.3-0.6 mm wide near apex); and (4) its style colour, which is very dark red (styles white with red base).
Gibson [14] erroneously lists the petal colour of D. macropetala as "purple" even though Diels [8] (p. 121) clearly states "petala [ . . . ] siccata pallida (non atropurpurea)" (dried petals pale white [not dark red]) in his description of D. microphylla var. macropetala. However, the wording of Diels also may indicate that he was not sure of the petal colour in their living state and thus only stated that they are white in the dried condition. The deep red, purplish red, reddish orange, or deep pink petal colours found in other members of the complex are usually well-preserved even in 100+ year-old specimens, provided they have been stored under favourable conditions. Many of the J. Drummond coll. VI n. 109 specimens of D. macropetala still clearly show the reddish inner part of their petals. It is therefore astonishing that the unique flower colour pattern escaped the notice of Diels, the taxon's author, and indeed of later botanists who studied the widely available material; the gathering J. Drummond coll. VI n. 109 consists of numerous duplicates (the herbarium sheets studied by the authors of the present study [see "Types"] comprise ca. 140 individuals of that taxon) and it was apparently distributed to several major European herbaria by Drummond at the time (syntypes were found in twelve herbaria, see "isolectotypes").
The initial collector, James Drummond, referred to D. macropetala as "[t]he most beautiful of the genus Drosera" [18,19] and he also clearly described the distinctive flower colour pattern of this taxon as "flowers [ . . . ] white with a crimson eye, and they are beautifully variegated with crimson veins" [18,19]. These earlier mentions were however apparently overlooked by Ludwig Diels when he described this distinctive Drosera as new to science in 1906 [8].
Distribution and habitat: Drosera macropetala is known from the Dandaragan Plateau between Dandaragan and Mogumber, about 100-150 km north of Perth ( Figure 3). The species appears to be restricted to the upper slopes of lateritic hills where it grows in low heath in poorly drained, sandy clay with laterite. Possibly also occurs in open Eucalyptus ("white gum") forest [18,19].
Phenology: Flowering has only been recorded in August.

Conservation status: Recommended for listing as Priority One (poorly known species) under Conservation Codes for Western Australian Flora and Fauna (Western Australian
Herbarium 1998-; https://florabase.dpaw.wa.gov.au/ (accessed on 6 December 2022)). Endangered (EN) under IUCN Red List criteria B1ab(iii,iv,v)+2ab(iii,iv,v) and C2a(i) following IUCN [29]. The extent of occurrence (EOO) and area of occurrence (AOO) of D. macropetala is estimated at ca. 200 km 2 and 16 km 2 , respectively. These numbers assume that the Mogumber population, last documented in 1904 by Alexander Morrison (see "Specimens examined"), still exists today. Given the extensive vegetation clearing in this area after that year [40], it is possible that this population has been destroyed, in which case both EOO and AOO would be <10 km 2 , meeting the Critically Endangered (CR) criteria [29]. Drosera macropetala is not known to occur on any land managed by the Western Australian Department of Biodiversity, Conservation and Attractions (DBCA) and is thus potentially threatened by future vegetation clearing.    This species was re-located by Declared Rare Flora monitors Gail and Dannielle Reed in August 2020 near Dandaragan, having not been recorded or documented since 1904. Their photographs were uploaded to Facebook and the plants depicted were immediately recognised as an unknown taxon by the authors. Subsequent targeted surveying of this area during 2020, 2021, and 2022 located a total of four (sub-)populations in a single, very narrow strip of unprotected remnant roadside vegetation (T. Krueger pers. obs.). These narrow, linear vegetation corridors, which transect completely cleared agricultural and urban areas, are highly susceptible to road maintenance and construction, altered hydrology, and weed infestation [16]. Population sizes of these four (sub-)populations vary from 2 to ca. 200 mature individuals and the total population in this area is estimated to consist of ca. 500 mature individuals. The historically reported population(s) from near Mogumber (which is ca. 50 km south-east of Dandaragan) have not yet been re-located by the authors as of this publication and their population size and persistence is currently unknown. Unlicenced collectors and illegal commercial/horticultural trade could pose an additional threat to D. macropetala in the future given its extremely small population sizes and the unfortunate tendency for poachers to target rare carnivorous plant species [16]. Further surveys are strongly recommended to gain a better understanding of this taxon's distribution, number and size of populations, and to identify further potential threats.
Notes on the lectotypification: Lectotypification of D. macropetala is required as Diels [8] did not select a type specimen from the duplicates of J. Drummond coll. VI n. 109, which all constitute syntypes. Marchant et al. [23] designated an "isotype" (which is not an inadvertent lectotypification following ICN Arts. 7.11 and 9.10 [26]) and, while K000659191 was labelled as the "holotype" by Marchant in 1985, this was not effectively published and also does not constitute a lectotypification (ICN Arts. 7.10, 7.11, and 9.10 [26]). Even if it had been validly published, the K specimen was incorrectly selected by Marchant as it cannot be the holotype (i.e., the material consulted by the taxon author for the description); Diels visited K to annotate the specimen after the publication of his D. microphylla var. macropetala in 1906 (evident from the fact that his annotation slip on the K specimen-in contrast with that on the B material-does not bear the label head "bearbeitet für das "Pflanzenreich"" [seen for Diels's taxonomic revision of Drosera, i.e., [8]]), which Diels made strict use of for all specimens he examined for his Drosera monograph [8]. Lowrie [1] (p. 354) erroneously assumed that both Bentham's D. calycina var. minor and Diels' D. microphylla var. macropetala are based on the same sheet (K000215038), which is J. Drummond coll. VI n. 110, even though Diels [8] clearly states in his description of D. microphylla var. macropetala that it is based on J. Drummond coll. VI n. 109.
Specimen B 100755976 features an identification slip in the hand of the taxon's author Ludwig Diels and, also given that he worked at B, represents the obvious choice for a lectotype.
At KFTA herbarium, a Drummond specimen has been indicated as "type material" of D. macropetala (KFTA0003370 photo!; identified as an "isotype" of "Drosera macrosepala" [sic.!] in 2013), however, this specimen is not J. Drummond coll. VI n. 109, nor does it agree with the locus classicus for the taxon. Rather, it corresponds to Drummond s.n., a collection of D. menziesii (the original label reads "Swan River Drummond", to which a pencil-written "n. 109" has been added in error later).
Notes on Drummond's type collection: The syntypes of J. Drummond coll. VI n. 109 only provide the rough locality information "Western Australia, between Moore River and Murchison Rivers" (locality not indicated on the lectotype specimen in B). However, more precise information on the locus classicus comes from Drummond's newspaper contributions "The Botany of the North-western Districts of Western Australia" [18], republished by Hooker [19]. There, he describes a Drosera species with white flowers with crimson centres, large glabrous sepals exceeding the petals in size, and flowers that close at night or during rainy weather, a description that exactly matches D. macropetala. Drummond [18,19] mentions that this species "[...] grows abundantly in a White Gum forest about four miles to the north of Dundaragan [Dandaragan]", a locality very close to where it still can be found today (T. Krueger pers. obs.). As J. Drummond coll. VI n. 109 is the only collection of D. macropetala provided by Drummond, it is safe to conclude that this is the collection locality. Additional support for this comes from Barker [41], who evidenced that Drummond's newspaper contribution [18] is referring to Drummond's VI collection series, i.e., the series containing the type collection of D. macropetala. This means that the year of collection (not given on any of the syntype specimens) is 1850 or 1851, as for all specimens comprising the VI collection [30].
Etymology: The specific epithet is derived from the Greek micros (=small) and phyllon (=leaf), referring to the small leaves of this species.
Taxonomic notes: Drosera microphylla is morphologically similar to D. koikyennuruff and D. reflexa, from which it is distinguished by (contrasting characters in parentheses): (1) its reddish orange petal colour (petals dark red in D. koikyennuruff or purplish pink with deep red base in D. reflexa); (2) its tendency to detach leaves during and prior to anthesis, often leaving large leaf-free gaps along stem, except in populations from the Stirling Range and Mt. Lindesay (leaves do not detach before or even after anthesis); (3) its late flowering time from August to October (from June to early September); and (4) its reddish purple stigma colour (stigmas deep red or dark red). It is further distinguished from D. koikyennuruff by: (1) its tentacle stalk colour, which is red or red in lower half with upper half yellowish green (tentacle stalks yellowish green), and (2) its preference for mossy wet areas on and near granite outcrops, seasonally wet swamps, or rocky mountain slopes (relatively dry sandy habitats in open Mallee woodlands).
Drosera microphylla has historically often been confused with D. calycina, from which it can be distinguished by (contrasting characters in parentheses): (1) its lamina shape, which is orbiculate or sometimes orbiculate with very slightly flattened upper margin (lamina reniform or orbiculate with flattened, often truncated upper margin); (2) its comparatively small flowers with a corolla diameter of 8-15 mm (corolla diameter 14-20 mm); (3) its petal colour, which is reddish orange with deep red base (petals deep red in inner half transitioning to purplish red in outer half); (4) its stamen length, which reaches only 2.5-3.5 mm (stamens 4.0-6.5 mm long); and (5) its styles, which extend laterally beyond the filaments and have reddish purple stigmas (styles not laterally extending beyond the filaments, with deep red stigmas).
The leaves of D. microphylla frequently detach from near the petiole bases, mostly shortly before and during anthesis, resulting in large, leaf-free "gaps" along the stem that are distinctive for this species (Figure 14B,F). In some cases, mature flowering individuals have been observed with only two or three widely separated leaves still attached to the stem, the remainder having been shed. This is also very apparent in the majority of herbarium specimens of D. microphylla. For example, Cranfield & Ward 25110 (PERTH 08507929!) has at least ten leaf nodes but only three with leaves still attached to them. However, in populations from Stirling Range National Park and Mt. Lindesay, the leaves do not appear to detach at all (T. Krueger pers. obs.).
Leaf detachment in most D. microphylla populations appears to serve a role in clonal propagation. In at least four populations near Denmark and Walpole, a red, prostrate, adventitious dropper shoot was observed to emerge from near the centre of the adaxial (tentacle-bearing) lamina surface, directly opposite the point of petiole attachment on the abaxial side (so-called epiphyllous budding; Figure 14D). This is congruent with observations made of naturally occurring asexual regrowth from both basal rosette and stem leaves in the tuberous Drosera auriculata Backh. ex Planch. and D. peltata Thunb., both likewise from D. section Ergaleium (depicted and described in detail by Vickery [42]) and from artificially detached leaf cuttings reported for other erect tuberous Drosera in cultivation [43] (A. Fleischmann and G. Bourke pers. obs.). Since the detached leaves of D. microphylla fall on soils that are still wet at that time of year (from July to September), these adventitious droppers can form a new tuber from their tips once they penetrate the soil, prior to the onset of the dry summer conditions. Leaf detachment thus appears to be a strategy for D. microphylla to quickly colonise the bare mossy or sandy soils of its preferred open habitats and may have arisen in connection with the usually very wet seepage soils this species grows in.
The ability to propagate clonally by epiphyllous budding from the leaves mirrors that of perennial Drosera in South Africa and Latin America, where only those species growing in rather wet habitats have the capacity to readily multiply asexually via leaf cuttings, something closely related species from drier habitats are incapable of [44]. Gibson [43] observed that, in cultivation, some erect tuberous Drosera do multiply through adventitious epiphyllous budding from artificially detached leaves, while others do not; again, there seems to be some connection with whether the species' natural habitats are wet or not. Although clonal propagation by budding from leaves still attached to the basal rosette and stem has been reported to occur naturally in D. auriculata and D. peltata [42], D. microphylla thus far is the only tuberous Drosera known to employ its stem leaves for vegetative propagation after shedding them from the mother plant and, generally, it seems to be the only tuberous Drosera species with caducous leaves.
Drosera microphylla is a very variable species and at least two collections exist that potentially represent additional undescribed taxa, the precise status of which could not be determined during this work. Coffey 103 A (PERTH 08468338!) comprises a collection from near Hopetoun, about 200 km east of the confirmed distribution area of D. microphylla (this collection is marked with "3" in Figure 3). While the specimen appears to match D. microphylla in morphological detail and could indeed represent an outlying population of this species, additional in situ observations are required to confirm its identity. Despite multiple attempts between 2019 and 2022, the authors of this work failed to re-locate this population.
Newbey 4204 (PERTH 00666459!; marked with "2" in Figure 3) represents a collection from near Boxwood Hill, collected in June 1974. While the flowering time would match that of D. koikyennuruff, the petal colour appears to be orange, while the overall habit more closely resembles that of D. microphylla. Further in situ observations of this population are required to determine whether it represents D. microphylla, D. koikyennuruff, or an undescribed taxon.
In addition, red-flowered plants with a different seed shape are known to co-occur with D. microphylla at sites along the south coast [15] (G. Bourke pers. obs.). While these have been included within D. microphylla in the present work, additional studies are recommended to evaluate whether these represent intraspecific variation of D. microphylla or a distinct taxon.
Distribution and habitat: Drosera microphylla occurs from Walpole to Cheynes Beach and throughout the Stirling Range (Figure 3). It grows in mossy wet areas on and near granite outcrops, seasonally wet seepage slopes near large swamp systems, rocky mountain ridges, and slopes. Soil usually sandy clay or peat, sometimes with laterite gravel or rocks.
Phenology: Flowering has been recorded from August to early October. Conservation status: Not eligible for Conservation Code listing and Least Concern (LC) according to the IUCN Red List Criteria, following Cross [45]. Not threatened, locally common, and widespread, occurring across many reserves managed by the Western Australian Department of Biodiversity, Conservation and Attractions (DBCA).

Notes on the type collection:
The type collection of D. microphylla, Hügel s.n. (W 0009732!), comprises a single plant collected from near Albany (King George Sound). The relatively long, straight (i.e., not fractiflex) stem and narrow petals indicate that it does indeed belong to the orange-flowered plants from the area and not to those newly described here as D. koikyennuruff and D. reflexa. While the type specimen appears to feature leaf-free gaps along the stem, it is unknown whether this is a result of the distinctive leaf detachment habit or whether they detached later in the brittle herbarium material. The large, separated leaf lying next to the lower part of the stem is a leaf of D. macrantha, which often co-occurs with D. microphylla in its preferred granite outcrop habitat around Albany where this collection was made.    (1) its filament shape, which is strongly dilated towards apex, 0.5-0.9 mm wide near apex (filaments ± linear, 0.3-0.6 mm wide near apex); (2) its petal colour, which is purplish pink with a deep red base (petals white with pale purplish red base); and (3) its flowering time from June to early September (flowering late August to October, sometimes until December). It differs from D. microphylla by (contrasting characters in parentheses): (1) its leaves, which remain attached to the stem even post-anthesis (leaves often detaching during and prior to anthesis, leaving large leaf-free gaps along stem, except in populations from the Stirling Range and Mt. Lindesay); (2) its stamen length of 3.8-4.4 mm (stamens 2.5-3.5 mm long); (3) its petal colour, which is purplish pink with deep red base (petals reddish orange with deep red base); (4) its stigma colour, which is dark red (stigmas reddish purple); and (5) its flowering time from June to early September (flowering from August to October). It differs from D. koikyennuruff by (contrasting characters in parentheses): (1) its tendency to form dense populations via adventitious stolons (plants in sparse populations, not colony-forming); (2) its petal shape, which is obovate to very broadly obovate (petals narrowly obovate to broadly spathulate); (3) its petal colour, which is purplish pink with deep red base (petals dark red); (4) its red to purplish red sepals (sepals yellowish brown to greenish yellow); and (5)  Leaves solitary on each node, irregularly alternate, 7-18 present in flowering individuals; internodes 2-8 mm. Petioles terete, semi-erect, arcuated abaxially (downwards) along whole length, occasionally arching increasing gradually towards the lamina, glabrous to microscopically punctate, 3-8 mm long, 0.3-0.5 mm wide above slightly thickened base, tapering to 0.1-0.2 mm in diameter towards the lamina, yellowish green to red, often darker near lamina. Lamina peltate, orbiculate, occasionally with very slightly flattened adaxial lateral margin, shallowly concave, adaxial surface facing outwards or slightly downwards, 1.5-3.5 mm long, 1.8-3.5 mm wide; lamina adaxial surface covered with stalked, carnivorous, secretive capitate glands (tentacles); tentacles 1.7-4.0 mm long at margin, decreasing in size towards centre of lamina, stalk red throughout or red in lower half with greenish yellow upper half; lamina abaxial surface glabrous but microscopically, sparsely punctate. Inflorescence a 1-3(-5)-flowered scorpioid cyme, terminal, simple, single-sided, 1.0-3.2(-4.9) cm long. Peduncles terete, 0.3-1.9(-4.0) cm long, 0.2-0.5 mm in diameter, microscopically glandular (appearing glabrous), reddish orange to red or rarely yellowish green. Pedicels terete, semi-erect or erect in fruit, 6-18 mm long in fruit, 0.2-0.4 mm in diameter, spaced by 2-9 mm along rhachis, microscopically glandular (appearing glabrous), usually more reddish than peduncle. Bracts spathulate or narrowly obovate, concave, arcuated adaxially (upwards), apex entire or crenulate, 1.7-3.0 mm long, 0.6-1.0 mm wide, abaxial surface minutely glandular (appearing glabrous). Sepals 5, narrowly obovate to narrowly elliptic, arcuated adaxially (upwards), slightly concave, often reflexed during anthesis, apex entire or crenulate, 5.2-8.3 mm long, 2.2-3.3 mm wide, abaxial surface minutely glandular (appearing glabrous), red to purplish red. Corolla 9-15 mm in diameter. Petals 5, purplish pink with large, deep red blotch towards the base, obovate to very broadly obovate, rarely broadly spathulate, deeply concave and slightly arcuated adaxially (upwards), margins entire, 4.2-7.0 mm long, 2.5-4.3 mm wide. Stamens 5, 3.8-4.4 mm long. Filaments dilated towards apex, 0.2-0.3 mm wide at base, 0.5-0.9 mm wide near apex, deep red. Anthers bithecate, retrorse, 0.9-1.1 mm wide, thecae pale yellow to orangey yellow. Pollen yellow to orangey yellow. Ovary obovoid, 3-carpellate, fused, 1.5-1.9 mm in diameter, deep red. Styles 3, divided into a few filiform segments just above the base, style segments again divided into many terete to distally flattened segments, forming a crowded tuft, extending laterally just beyond filaments, 1.5-1.8 mm long, dark red. Stigmas simple or shortly branched, at tips of style segments, papillose, ca. 0.2 mm long, dark red. Seeds narrowly elliptical to narrowly obovate, outline more or less narrowly rectangular, rarely curved, funicular (upper) end truncate or with shallow funicular disc, basal (chalazal) end acute to obtuse, 1.3-1.8 mm long, 0.3-0.4 mm wide, dark brown, appearing black, tips pale brown; testa more or less longitudinally reticulate, with anticlines thin and only shallowly raised.
Etymology: The specific epithet is derived from the Latin reflexus (=turned back or away) and refers to the often strongly reflexed (by up to ca. 140-170 • with respect to the floral axis) sepals and petals.
Taxonomic notes: The overall habit as well as petiole, lamina, and style shape of D. reflexa indicate that it is morphologically most similar to D. esperensis, D. koikyennuruff, and D. microphylla. In living specimens, the petal colour can be easily used to differentiate the species within this group (purplish pink with deep red base in D. reflexa, white with pale purplish red base in D. esperensis, dark red in D. koikyennuruff, and reddish orange with red base in D. microphylla). Drosera reflexa is found in close proximity to D. microphylla (ca. 200 m) but the two taxa do not co-occur (no syntopic occurrence) despite their very similar habitat preferences. No hybrids or intermediates between the two species have been observed. This taxon was first mentioned by Lowrie et al. [15] (p. 266) under D. microphylla as "a diminutive form with bi-coloured red and pink flowers".
Distribution and habitat: Kentdale (between Walpole and Denmark near the south coast of Western Australia; Figure 3). Occurs in shallow decomposed granitic soils over granite lenses in mosses.
Phenology: Flowering has been recorded from June to early September. Conservation status: Listed as Priority Two (poorly known species) under Conservation Codes for Western Australian Flora and Fauna (Western Australian Herbarium 1998-; https://florabase.dpaw.wa.gov.au/ (accessed on 6 December 2022)), under the phrasename "Drosera sp. Kentdale (G.J. Bourke 458)". It is Critically Endangered (CR) under IUCN Red List criteria B1ab(iii,v)+2ab(iii,v) following IUCN [29]. Drosera reflexa is only known from a single population that is partially located on land managed by the Western Australian Department of Biodiversity, Conservation and Attractions (DBCA). Targeted surveys in the region in 2002, 2017, 2019, 2020, 2021, and 2022 were unable to identify any additional populations despite considerable areas of apparently suitable habitat being surveyed (G. Bourke and T. Krueger pers. obs.). Suitable habitat on nearby private land was not surveyed and may yield additional remnant populations. The total population size is estimated at ca. 1000 mature individuals. Damage to the habitat by recreational vehicles has been observed (G. Bourke pers. obs.) and significant invasive weed infestation is apparent in parts of the habitat (T. Krueger pers. obs.). Further surveys are recommended to gain a better understanding of this taxon's distribution, number and size of populations, and to identify additional potential threats. Pollen yellow. Ovary obovoid, 3-carpellate, fused, 1.3-1.9 mm in diameter, deep red. Styles 3, divided into a few filiform segments just above the base, style segments again divided into many terete style segments, forming a crowded tuft, not extending laterally beyond filaments, 1.2-1.7 mm long, very dark red. Stigmas simple or shortly branched, at tips of style segments, papillose, ca. 0.2 mm long, very dark red. Seeds narrowly obtrullate to narrowly obovate, outline narrowly conical with slight ellipsoid swelling in the upper (funicular) half, funicular (upper) end truncate (=pin-or nail-shaped seeds), basal (chalazal) end pointed with short conical to fusiform and often slightly curved appendage, 1.8-2.4 mm long, 0.3-0.5 mm wide, testa black-brown, chalazal and funicular ends pale brown; testa longitudinally reticulate, with anticlines thin and only shallowly raised.
Etymology: The specific epithet is derived from the Latin ruber (=red) and calyx in reference to the red sepal colour of this species that provides a distinct colour contrast to the yellowish green stem and peduncle.
Taxonomic notes: The distinctive characters distinguishing D. rubricalyx from the morphologically most similar taxa (D. hortiorum and D. macropetala) are detailed under those respective subheadings. The unique petal colour combination of deep red and deep pink usually quickly allows identification of D. rubricalyx in the field (Figure 4). Only D. calycina may occasionally produce a similar petal colour, although that species always has solitary cauline leaves, while D. rubricalyx consistently produces two smaller axillary leaves in the axils of the upper 2-10 cauline leaves. Additionally, D. calycina has a different lamina shape, which is reniform to orbiculate with flattened, often truncated, upper margin vs. lamina orbiculate or orbiculate with slightly flattened upper margin in D. rubricalyx.
Bentham [21] distinguished D. calycina var. minor (=D. rubricalyx) from D. calycina based on its smaller leaves and flowers. Indeed, the petals of D. rubricalyx are usually much smaller, especially in width, compared with those of D. calycina and D. macropetala (the latter species was included in Bentham's description of D. calycina as he cites J. Drummond coll. VI n. 109, the type of D. macropetala). Additionally, Bentham [21] already noted that D. rubricalyx has "rather less dilated" filaments when compared to these two species, which is indeed a reliable distinguishing floral feature. However, it is notable that Bentham was able to detect this feature in the dried herbarium specimens he studied, as the filaments in dried specimens of D. microphylla complex species often considerably shrink in width (T. Krueger pers. obs.).
The infrataxon epithet of D. calycina var. minor, published by Bentham in 1864 [21], cannot be elevated to species rank as an older homonym with nomenclatural priority exists (Drosera minor Schumach. & Thonn. in Schumach., published in 1827 and generally treated as a synonym of D. indica L.). Therefore, a new name, D. rubricalyx, had to be coined for D. calycina var. minor at the rank of species.

Endemism and Species Conservation in the Drosera microphylla Complex
The species of the Drosera microphylla complex are among the most narrowly endemic and most threatened species of D. section Ergaleium (tuberous sundews). Six of the nine species of the complex have known distribution areas with a maximum diameter of less than 100 km (D. atrata, D. calycina, D. koikyennuruff, D. macropetala, D. reflexa, and D. rubricalyx; Figure 3). By contrast, only 3 of the 68 remaining species of D. section Ergaleium occur across such small areas of distribution (these are D. graniticola N.G.Marchant, D. orbiculata N.G.Marchant & Lowrie, and D. prostratoscaposa Lowrie & Carlquist; [1]; https://florabase. dpaw.wa.gov.au/ (accessed on 22 December 2022)). Additionally, most members of the D. microphylla complex are uncommon even within their distribution areas, being highly localised and often present in very small populations of fewer than 100 individuals (or even fewer than ten individual plants, as is often the case for D. atrata). These very small distribution areas and population sizes, in combination with the threats of habitat loss (including presumed reductions in gene flow as a result of habitat fragmentation) and illegal collection-identified as threats for seven of the nine species-indicate a strong necessity for targeted conservation efforts to ensure their long-term survival in nature.
Drosera macropetala and D. reflexa are here assessed as the most threatened members of the D. microphylla complex, with a recommended Western Australian Conservation Code status of Priority One and Priority Two, respectively (Western Australian Herbarium 1998-; https://florabase.dpaw.wa.gov.au/ (accessed on 22 December 2022)); and an IUCN category of Endangered (EN) and Critically Endangered (CR), respectively [29]. Despite considerable survey efforts, both species are currently each only known from a single roadside location (with D. macropetala being historically also recorded from a second location further south; Figure 3). Such roadside habitats are particularly vulnerable to threats from road maintenance and construction, altered hydrology, and weed infestation [16].
Drosera atrata and D. rubricalyx are both assessed as Vulnerable (VU) [29] given the available data, with a Western Australian Conservation Code status of Priority Three and Priority Two, respectively. While both are known from multiple locations, their distribution areas are very small (ca. 20-50 km; Figure 3).
Drosera hortiorum is unusual within the D. microphylla complex as it has a relatively large distribution area spanning at least 160 km but is only known from four locations (Figure 3), each with a population size of fewer than 50 individuals. It is recommended for a Western Australian Conservation Code status of Priority Two, but insufficient survey efforts for this species means it cannot yet be assessed under IUCN criteria (=Data Deficient, DD) [29]. Similarly, D. koikyennuruff (which is only known from two locations near the Stirling Range; Figure 3) could not be assessed under IUCN criteria given the lack of available survey data, although it is assessed as Priority Two under the Western Australian Conservation Code.
The only species of the D. microphylla complex that are not currently assessed as potentially threatened are D. calycina, D. esperensis, and D. microphylla. All three species are known from numerous sites and are generally quite common within their preferred habitat (i.e., they tend to form relatively large populations). In addition, D. esperensis and D. microphylla have relatively large distribution areas extending over >100 km (Figure 3).
Knowledge of the distributions and range extensions of the species of the D. microphylla complex for this study were not only gained through field studies and herbarium research, but also from photographs from citizen science and social media (see Section 4.3). This highlights the importance of citizen scientist contributions for nature conservation (see Section 4.3 and [47][48][49]).

Flower Biology of the Drosera microphylla Complex
The members of the D. microphylla complex rank among the comparatively few sundew species that have non-ephemeral flowers, i.e., flowers that last for longer than one day. Within that group, they represent a smaller group still of species whose flowers close every night until they finally fade after about 3-5 days (A. Fleischmann pers. obs.). The daily opening and closing of the flowers is achieved by one-sided petal growth (common in many other plants with flowers that cyclically open and close), as evidenced by the fact that the petals of all members of the D. microphylla complex increase in size during anthesis as they slightly enlarge with each new opening event. The non-ephemeral nature of the flowers of species from this affinity, as well as their nocturnal closure, was first reported by James Drummond [18,19]. The initial opening of the flower from bud takes some time, as both sepals and petals must spread open on the first occasion. Once a flower has fully opened for the first time, the concave petals close during the late afternoon (see Figure 1I), the sepals also subsequently close at night or during unfavourable weather such as on cold and/or rainy days, covering the reproductive organs. The re-opening of an individual flower is light-dependant, but also strongly temperature-dependant, and covering individual, closed flowers with plastic bags on cold sunny days will often induce them to open within just a few minutes (F. Hort, J. Hort, and T. Krueger pers. obs. for D. hortiorum). The flowers of members of the D. microphylla complex are non-fragrant (A. Fleischmann pers. obs.; Gibson [14] for D. esperensis), which is an exception among members of D. section Ergaleium, which usually have strongly fragrant flowers [50] (A. Fleischmann pers. obs.). The combination of non-fragrant, non-ephemeral flowers that close each night is unique among Australian Drosera and restricted to the D. microphylla complex-it represents an ecological apomorphy of that complex. The strongly concave petals, which are shorter than the large sepals, are likewise an apomorphy for this complex. Within the genus, both characters are only paralleled in the very distantly related neotropical Drosera biflora Willd. from D. section Drosera [51].
It is interesting to note that the three species with exceptionally dark flower colours (D. atrata, D. hortiorum, and D. koikyennuruff ) have unusually early flowering times, much earlier than most other species of the complex (only D. reflexa has also been observed to flower as early as June, G. Bourke pers. obs.) and also much earlier than most other erect tuberous Drosera species. This might be an adaptation to certain pollinators that are active during this period, both as temporal reproductive isolation from sympatric taxa and possibly a favoured or easy-to-spot colour among pollinators active at this time of year.

The Role of Citizen Science and Social Media in Taxonomy and Species Discovery-An Example from Drosera
The present work serves as another excellent example of citizen science, social media, and online photo platforms facilitating or even driving improvements in taxonomic and ecological knowledge. These networks have in several cases alerted botanists to new discoveries, resulted in the relocation of 'lost' or poorly known taxa, or simply extended the ranges of known species beyond those distributions established through herbarium and museum records alone [48,49,[52][53][54][55]. Drosera macropetala, D. rubricalyx, D. hortiorum, and D. koikyennuruff were first (re-)discovered via online photographs or significant contributions to their range and distribution gained from images shared on the iNaturalist website and app and other social networking platforms. Carnivorous plants such as Drosera are usually well-represented on naturalist photographic databases and platforms as these peculiar plants are frequently photographed [56]. For example, iNaturalist (Research-Grade observations only) hosts georeferenced photographs of 236 (89.7%) out of the ca. 260 Drosera species known to science to date, representing 62,344 individual observations (most of which consist of several photographs) made by 17,477 observers (https://www.inaturalist.org/observations?place_id=any&taxon_id=51935 (accessed 21 December 2022)). As of December 2022, the Global Biodiversity Information Facility (GBIF; www.gbif.org (accessed on 14 January 2023)) provides 453,239 occurrence records of 263 species of Drosera (=100% taxon coverage); 79.4% of these records are based on citizen science observations ("human observation"), as opposed to 18.5% that come from georeferenced, databased herbarium specimens ("preserved specimen") (GBIF.org (accessed on 9 January 2023) GBIF Occurrence Download https://doi.org/10.15468/dl.h4fwxq).
This coverage is not restricted only to widespread or common species (the most commonly observed carnivorous plant on GBIF is Drosera rotundifolia L. [56], which is as of 9 January 2023 represented in that database by 200,188 occurrences, 89.8% of which are citizen science observations (GBIF.org (accessed on 9 January 2023) GBIF Occurrence Download https://doi.org/10.15468/dl.9w2ded). For some formerly rarely encountered Drosera species, more "photo vouchers" exist as geographic records than are available as herbarium specimens. For example, the South African Drosera xerophila was known only from three herbarium specimens and seven records on iNaturalist at the time of its description (all of them cited in Fleischmann [44]); the number of herbarium specimens has not changed since the publication of the species in April 2018, but the number of observations for that species on iNaturalist had risen to 307 by January 2023, made by 131 different observers (iNaturalist community. Observations of Drosera xerophila. Exported from https://www.inaturalist.org (accessed on 9 January 2023)).
This illustrates very well the scientific value of these image repositories for taxonomic and biogeographic work and for nature conservation [47,54,57]. In other organismic groups that are frequently photographed and uploaded to citizen science social networks, such as iNaturalist, by enthusiasts, these records represent by far the largest contribution to our knowledge of the distribution of these taxa. This is particularly true for "charismatic animals", such as mammals and birds, for which 70% (mammals) and 87% (birds) of the total records on GBIF in 2016 comprised online citizen science records [48]. As the global citizen science naturalist networks continuously and quickly increase their amount of data (e.g., in 2019 alone, the total number of observations on iNaturalist doubled from 25 million to 50 million [55]), the rich dataset of occurrence records provided by citizen science by far outweighs those gained annually from herbarium specimens and literature revisions, though of course the latter usually provide persistent, high-quality taxonomic data along with physical reference specimens that are additionally suitable for DNA extraction and genetic analyses, microscopic examinations, and digitised associated metadata.
Regarding species discoveries and range extensions through social media and citizen science, the present taxonomic revision provides an excellent example. Four out of the six species newly described or newly classified as species here were initially (re-)discovered on social media, with only D. atrata and D. reflexa (co-)discovered in situ by the authors of the present work. Citizen science and social media networks have also provided the first known photographs of D. koikyennuruff, D. macropetala, and D. rubricalyx, which were previously only known from decades-old herbarium collections. Another relatively well-publicised example of a Drosera species discovered on social media is D. magnifica, which was first spotted and identified as a species new to science from photographs posted on Facebook in 2014 [27]. An example for significant range extensions in Drosera provided by citizen science is D. biflora, the first record of which from Colombia was made by photographs posted on iNaturalist; the rediscovery of that species in Brazil was also facilitated by photographs posted online (republished in Gonella et al. [51]). These, at the same time, also represented the rediscovery and first known photographs of this species, which was previously only known from 200-year-old herbarium collections made in Venezuela. The citizen scientist photographs of living D. biflora specimens also provided additional unique biological and morphological details for this species (curiously, including the unique character of patent to reflexed sepals, a character this species shares with the unrelated species from the D. microphylla complex treated here), which could not be discerned from the historic herbarium material and helped to increase the knowledge about the taxonomy and relationships of this species [51].
Even organismic interactions can be revealed or documented for the first time through social media photographs and citizen science [58], such as plant-pollinator relationships. There are numerous examples where floral visitors and pollinators of Drosera species have been first documented by citizen scientists via photographs shared online. Carnivorous plant-prey relationships can also be documented by citizen science, because the different interacting organisms can often be identified from photographs (Drosera prey, for exam-ple, can often be identified from photographs [59]). These platforms effectively connect taxonomic experts from different fields, such as entomologists and botanists. An example involving Drosera is an iNaturalist photograph that was used as a voucher in a citizen science approach for mosquito species monitoring in Australia (published as Figure 4B in Braz Sousa et al. [60]). It shows the mosquito Aedes camptorhynchus captured on a leaf of Drosera planchonii Hook.f. ex Planch. in South Australia (based on an iNaturalist photograph by observer "frank_prinz": https://www.inaturalist.org/observations/54139335 (accessed on 21 December 2022)). While the Drosera expert or trained botanist would have been able to correctly identify the targeted plant taxon in the field and the mosquito expert will name the insect, the advantage of the social media platforms is that both experts, and additionally the observer, are linked via a photograph showing different target taxa of interest. Many of the citizen scientist photographs do not just represent single-species observations but in fact are (often unnoticed) documents of species interactions [58].
An increasing number of applications are being developed to use citizen science and social media data for biodiversity exploration and flora monitoring, linking them with taxonomy (e.g., [56,61]). Machine learning approaches are frequently improved to automate species recognition and thus enhance data mining for images suitable for taxonomy (e.g., [62,63]).
Potential negative effects of providing locality data, especially of rare flora, online in social networks arise in plant groups that are of horticultural interest, such as cacti, orchids, or carnivorous plants [64,65]. For some carnivorous plants, pitcher plants in particular, populations in the wild are mainly threatened by overcollection and poaching for the illegal trade to meet the horticultural demand [16]. However, Drosera species, in particular tuberous Drosera from Western Australia, have also been and continue to be heavily poached (including from protected areas such as nature reserves) to be sold illegally on social media or internet marketplaces to carnivorous plant enthusiasts and growers worldwide [16,65]. This is a major potential threat, particularly to rare, micro-endemic taxa such as the majority of species from the D. microphylla complex treated in this paper. iNaturalist automatically obscures the geographic information of observations of threatened taxa, which greatly helps with mitigating this threat (https://www.inaturalist.org/pages/help#geoprivacy (accessed on 23 December 2022)). For this reason, exact localities have been withheld for conservation purposes for the species of the D. microphylla complex recommended to be listed as Priority under the Western Australian Conservation Codes and the authors of the present work generally do not share locality information for threatened flora.

Conclusions
This study highlights the importance of both citizen science and careful herbarium examination for taxonomic research and conservation efforts. Of the six species newly recognised here, four were (re-)discovered on social media and all but D. reflexa had already been represented in herbaria for many decades. Crucially, these six species are rare, narrowly endemic, and potentially threatened, thus the accurate taxonomic classification provided here is expected to contribute to their conservation.