Introduction

The anthericolous smut fungi on hosts in the Caryophyllaceae are among the longest-studied species of smut fungi. The name Uredo violacea Pers., the basionym of Microbotryum violaceum (Pers.) G. Deml & Oberw., s. str., was described by Persoon (1797), making this species one of the oldest known smut fungi (Vánky 2004). On the basis of inoculation tests, Liro (1924) recognized nine species in this complex of anthericolous fungi. Subsequent lumping and splitting treated this complex as one to many species (cf. Fischer 1953; Savile 1953; Lindeberg 1959; Nannfeldt 1959; Deml & Oberwinkler 1982, 1983; Scholz & Scholz 1988; Vánky 1985, 1994). For a long period, these species were considered as members of Ustilago (Pers.) Roussel and placed in the Ustilaginaceae, until Deml & Oberwinkler (1982) reinstated Microbotryum Lév. (with M. violaceum as the type species) for seven anthericolous species on caryophyllaceous hosts. Thereafter, the concept for this genus was altered several times. Vánky (1998), in his monographic treatment of Microbotryum, extended the definition of this genus and considered it in a broad sense, including 76 species, eight of which occur in the anthers of Caryophyllaceae. Later, on the basis of ultrastructural characteristics and molecular analyses, it was demonstrated that the species of Microbotryum were more closely related to the rust fungi (Pucciniomycotina) than to the true smut fungi (Ustilaginomycotina) (Bauer & al. 1997, 2006; Begerow & al. 1998). Over the past two decades, phylogenetic and genomic approaches have considered the anthericolous species on hosts in the Caryophyllaceae as a model system for studying host-pathogen interactions in natural populations (e.g. Antonovics & al. 2002, 2003; Garber & Ruddat 2002; Uchida & al. 2003; Kazama & al. 2005; Lutz & al. 2005, 2008; Le Gac & al. 2007; Giraud & al. 2008; Refrégier & al. 2008; Sloan & al. 2008; Bernasconi & al. 2009; Denchev & al. 2009; Hood & al. 2010; Piątek & al. 2012, 2013; Perlin & al. 2015; Petit & al. 2017; ).

The anther-smut fungi of Microbotryum on hosts in the Caryophyllaceae cause formation of teliospores instead of pollen in the anthers of bisexual flowers. When female flowers of dioecious species (e.g. Silene latifolia Poir. and S. dioica (L.) Clairv.) are infected, suppression of stamen development does not occur, and development of spore-bearing anthers is induced (Kazama & al. 2005). These smut fungi infect mainly perennial plants (Thrall & al. 1993; Hood & al. 2010). The teliospores are transmitted from diseased to healthy plants mostly by insects that normally serve as pollinators (Fontaine & al. 2013).

The most widely studied anthericolous smuts are those in the anthers of Silene L. It is a group of highly host-specific fungi. The intensive taxonomic and phylogenetic studies of this group yielded resurrection of several species from the catch-all species Ustilago violacea and additional description of new species (Deml & Oberwinkler 1982; Chlebicki & Suková 2005; Lutz & al. 2005, 2008; Denchev 2007a, b; Le Gac & al. 2007; Refrégier & al. 2008; Denchev & al. 2009; Denchev & Denchev 2011; Piątek & al. 2012, 2013).

Sorus morphology divides the recognized 16 species of Microbotryum on Silene into two groups. The first group comprises 13 species that cause typical anther infection: M. bardanense Chleb. & Suková, M. chloranthaeverrucosum M. Lutz & al., M. coronariae (Liro) Denchev & T. Denchev, M. heliospermae Piątek & M. Lutz, M. lagerheimii Denchev, M. lychnidisdioicae (DC.) G. Deml & Oberw., M. silenesacaulis M. Lutz & al., M. silenesdioicae T. Giraud & al., M. silenesinflatae (DC. ex Liro) G. Deml & Oberw., M. silenessaxifragae M. Lutz & al., M. violaceoirregulare (Brandenb. & Schwinn) G. Deml & Oberw., M. violaceoverrucosum (Brandenb. & Schwinn) Vánky and M. violaceum (Pers.) G. Deml & Oberw. The development of their sori is restricted to the anthers. The second group has three species: M. adenopetalae M. Lutz & al., M. majus (J. Schröt.) G. Deml & Oberw. and M. savilei Denchev. The development of their sori results in formation of swollen and deformed flowers, completely filled with spore mass. The sori of these species are usually formed not only in the anthers but also in the filaments, which at a more advanced stage of soral development leads to destruction of the stamens.

In the course of a study of the smut fungi of Greenland, carried out by two of the authors (TTD & CMD), the caryophyllaceous collection in the Greenland Herbarium, which is part of the herbarium of the Natural History Museum of Denmark, University of Copenhagen (C; herbarium code according to Thiers 2019+), was examined for presence of smut fungus infection. Four specimens of Silene uralensis subsp. arctica (Th. Fr.) Bocquet were found infected by a Microbotryum species that caused typical anther infection. On the same host, another species of Microbotryum, M. savilei, has already been described from Canada (Denchev 2007b), but it causes atypical anther infection. For the first time, presence of two kinds of infection of the flowers of S. uralensis subsp. arctica was observed by D. B. O. Savile during his field trips to the Canadian High Arctic islands in 1958–1959 and the early 1960s. Savile never published this observation but the herbarium packets of the specimens, collected during these expeditions, are thoroughly annotated inside. However, he published information about another, strange type of infection of the same host plant, a stigma infection, and supplied a table with comparative data about the morphology of the healthy plants, those with stigma infection, and those with typical anther infection, based on examination of a Somerset Island collection no. 3628 (Savile 1959: 983–984). Our revision of this specimen did not confirm presence of a peculiar infection and we consider the smutted plants in this specimen (DAOM 02-01000869745) as infected by the typical anther infection.

It was found that the above-mentioned specimens on Silene uralensis subsp. arctica from Greenland belong to a species morphologically distinct from Microbotryum savilei. Additionally, ITS and LSU rDNA sequences of the Microbotryum species from Greenland were obtained and compared with sequences of Microbotryum species on caryophyllaceous hosts, available in the NCBI nucleotide database. The morphological differences and molecular phylogenetic analyses revealed the existence of a new smut fungus on S. uralensis subsp. arctica, which is described and illustrated here.

Material and methods

Results

Taxonomy

Re-examination of the type of Microbotryum savilei requires an emended description of its soral morphology.

Microbotryum savilei Denchev in Mycol. Balcan. 4: 72. 2007. – Holotype: on Silene uralensis subsp. arctica, Canada, Nunavut, Kivalliq Region, Southampton Island, Coral Harbour, 16 Aug 1959, D. B. O. Savile 3952, J. A. Calder & I. Kukkonen (DAOM 02-01000571594/DAOM 66879!). – Paratype: on same host, Canada, Nunavut, Kivalliq Region, Southampton Island, Coral Harbour, 16 Aug 1959, D. B. O. Savile 3953, J. A. Calder & I. Kukkonen (DAOM 02-01000868580/DAOM 66880! [white-flowered plants]). – Fig. 2, 3.

Fig. 1.

Best tree of the Maximum Likelihood analysis of caryophyllaceous anther smuts in the genus Microbotryum based on a concatenated ITS and LSU dataset. The tree was rooted with M. anomalum (J. Kunze ex G. Winter) Vánky (based on Kemler & al. 2006). Bootstrap values ≥ 50 are indicated above branches.

Fig. 2.

Microbotryum savilei on Silene uralensis subsp. arctica. – A–D: habit (holotype); note deformed and swollen flower (enclosed by corolla; black arrow in D shows petals that remain within closed calyx), destroyed stamens (white arrows), and aborted ovary and styles (arrowhead); E: habit (paratype, on white-flowered plants); F, G: spores in LM, in median and surface view, respectively (holotype). – Scale bars: A–E = 0.5 cm; F, G = 10 µm.

Fig. 3.

Microbotryum savilei on Silene uralensis subsp. arctica. – A–C: spores in SEM. – Scale bars: A, B = 5 µm; C = 1 µm.

Description — Infection systemic. Sori in deformed and swollen flowers (Fig. 2A, B, E), enclosed by corolla (petals remain within closed calyx) (Fig. 2C, D), affecting anthers and filaments; stamens may or may not be completely destroyed; ovary and styles aborted (Fig. 2D). Calyx of infected flowers erect and closed at throat, broadly ovoid or very broadly ovoid (with approximate height/width ratio 0.8–1.2) (Fig. 2A, B, E). Spore mass filling affected flower completely, enclosed by corolla; pulverulent, dark livid or dark purple (Fig. 2C). Spores subglobose, globose or broadly ellipsoidal, sometimes ellipsoidal, (5.5–)6–9(–10) × (5–)5.5–8(–9) (7.4 ± 0.7 × 6.7 ± 0.6) µm (n/₁ = 300), length/width ratio 1.08–1.10, light vinaceous; wall reticulate, 1–1.7(–2) µm thick (including reticulum); meshes (4 or)5–8(or 9) per spore diameter, polyhedral or irregular, 0.4–1.8(–2.5) µm long; muri (15 or)16–20(–22) on equatorial circumference, to 0.4(–0.5) µm high; meshes rugulose on bottom in SEM (Fig. 3A–C). Spore germination unknown.

Host plant and distribution — On Caryophyllaceae: Silene sect. Physolychnis (Benth.) Bocquet: S. uralensis subsp. arctica, North America (Canadian Arctic). Known only from the type locality (Fig. 5).

Remarks — The paratype was collected by the same collectors and at the same site as the holotype, but, according to Savile's note (placed inside the herbarium packet), the infected plants of the paratype are white-flowered.

Based on our phylogenetic and morphological studies, we propose a new species of Microbotryum on Silene uralensis subsp. arctica.

Microbotryum arcticum T. Denchev, Denchev, Kemler & Begerow, sp. nov. – Fig. 4.

Index Fungorum number: IF 555874.

Holotype: on Silene uralensis subsp. arctica, Greenland, Peary Land, 10 km NW of Mudderbugt, just S of Ndr. Ladegårdså, 82°29–30′N, 21°30–35′W, 7 Aug 1991, B. Fredskild 91433 (SOMF 29 999).

Diagnosis — Microbotryum arcticum differs morphologically from M. savilei in soral development, restricted to the anthers, with presence of intact filaments (vs formation of swollen and deformed flowers, completely filled with a spore mass, and soral development not only in the anthers but also in the filaments, which are morphologically affected; Table 3), and by having spores with lower maximum and mean values of the spore length and width, to 7.5(–8.5) µm long (6.4 ± 0.5) and to 6.5(–7.5) µm wide (5.9 ± 0.4) vs to 9(–10) µm long (7.4 ± 0.7) and to 8(–9) µm wide (6.7 ± 0.6) for M. savilei (Table 2). Microbotryum arcticum is distinguished from the other anther-infecting Microbotryum species by molecular data, based on ITS sequence (deposited in the NCBI nucleotide database with accession no. MK474659) and LSU sequence (accession no. MK474658) derived from the holotype, and ITS sequence (accession no. MK474660) derived from a paratype (SOMF 29 998).

Description — Infection systemic. Sori in considerably swollen anthers, filling pollen sacs with a pulverulent, dark livid or livid vinaceous spore mass. Spores subglobose, globose, broadly ellipsoidal or ovoid, sometimes ellipsoidal or slightly irregular, (5–)5.5–7.5(–8.5) × (4.5–)5–6.5(–7.5) (6.4 ± 0.5 × 5.9 ± 0.4) µm (n/₁₄ = 1600), length/width ratio 1.06–1.12, pale vinaceous; wall reticulate, 0.8–1.3(–1.5) µm thick (including reticulum); meshes 5–8(or 9) per spore diameter, polyhedral or irregular, 0.3–1(–1.5) µm long; muri (15 or)16–21(–23) on equatorial circumference, to 0.4 µm high; meshes smooth or rugulose on bottom in SEM (Fig. 4H–J). Spore germination (Parmelee, unpublished notes and photographs in herbarium packet of DAOM 93735) results in a 4-celled basidium, separating during basidiospore formation as a 3-celled basidium (1 cell remaining attached to teliospore) and producing basidiospores laterally and terminally (Fig. 4F, G).

Fig. 4.

Microbotryum arcticum on Silene uralensis subsp. arctica. – A: habit (DAOM 66883); note nodding, open and not deformed flowers; B: habit (holotype); note petals emerging from calyx (black arrow) and affected anther (white arrow); C: stamens of healthy plants of S. uralensis subsp. arctica (top, DAOM 93735), plants infected by M. arcticum (middle, DAOM 66883) and plants infected by M. savilei (bottom, holotype); D, E: spores in LM, in median and surface view, respectively (holotype); F, G: germination of spores (unpublished photographs by J. A. Parmelee in herbarium packet of DAOM 93735; used by permission of J. Parmelee [17 Jan 2019] and S. Redhead, Curator of DAOM); H–J: spores in SEM (H: Inglefield Land; I, J: Washington Land). – Scale bars: A–C = 0.5 cm; D, E = 10 µm; H–J = 5 µm.

Fig. 5.

Geographic distribution of Microbotryum arcticum and M. savilei (generated with SimpleMappr, Shorthouse 2010).

Host plant and distribution — On Caryophyllaceae: Silene sect. Physolychnis: S. uralensis subsp. arctica, North America (Canadian Arctic and Greenland, with localities at 82°30′N–61°35′N) (Fig. 5).

Additional specimens examined (paratypes) — On Silene uralensis subsp. arctica. — CANADA: Nunavut, Qikiqtaaluk Region, Ellesmere Island, 2 mi NNW of Hazen Camp, 81°49′N, 71°21′W, alt. 1400 ft, 17 Jul 1962, D. B. O. Savile 4656 (DAOM 02-01000869753/DAOM 92911; in Savile & Parmelee 1964: 711 as “U[stilago]. violacea var. violacea”); Ellesmere Island, 6 mi WSW of Hazen Camp, 81°49′N, 71°21′W, alt. 1500 ft, 10 Jul 1962, D. B. O. Savile 4601 (DAOM 02-01000869752/ DAOM 92910; in Savile & Parmelee 1964: 711 as “U. violacea var. violacea”); Axel Heiberg Island, Jacobsen-McGill Base Camp, on margins of Colour Lake, 79°25′N, 90°30′W, 7 Aug 1961, J. A. Parmelee 2215 (DAOM 02-01000869751/DAOM 93735; in Savile & Parmelee 1964: 711 as “U. violacea var. violacea”); Somerset Island, Aston Bay, 73°39′N, 94°45′W, 8 Aug 1958, D. B. O. Savile 3729 (DAOM 02-01000869750/DAOM 60120; in Savile 1959: 983 as “U. violacea var. violacea”); Somerset Island, Four Rivers Bay, 72°50′N, 95°30′W, 31 Jul 1958, D. B. O. Savile 3717 (DAOM 02-01000869749/ DAOM 60135; in Savile 1959: 983 as “U. violacea var. violacea”); Somerset Island, 2 mi NNW of Base Camp, 72°07′N, 94°12′W, 27 Jul 1958, D. B. O. Savile 3628 (DAOM 02-01000869745/DAOM 60165; in Savile 1959: 983 as “U. violacea var. violacea”); Somerset Island, N slope of Hazard Inlet, 72°04′N, 94°10′W, 22 Jul 1958, D. B. O. Savile 3583 (DAOM 02-01000869746/DAOM 59984; in Savile 1959: 983 as “U. violacea var. violacea”); Baffin Island, Frobisher Bay, 63°45′N, 68°32′W, 10 Jul 1948, H. A. Senn 3868 & J. A. Calder (DAOM 02-01000869747/DAOM 60136); ibid., Frobisher Bay, 10 Aug 1959, D. B. O. Savile 3925, J. A. Calder & I. Kukkonen (DAOM 02-01000869748/DAOM 66883, as U. violacea var. stellariae). — Greenland: Warming Land, GGU (Grønlands Geologiske Undersøgelse) Base Camp, 81°32′N, 51°31′W, 13 Aug 1985, C. Bay 85434 (C-Greenland herb. s.n.); Washington Land, Cass Fjord, Nygaard Bugt, 80°06′N, 65°10′W, alt. 10 m, 5 Aug 1976, P. Frykman & B. Fredskild s.n. (C-Greenland herb. s.n.); Inglefield Land, central inland, plain plateau, 78°40′N, 68°18′W, alt. 450 m, 16 Aug 1999, J. Feilberg 534 (SOMF 29 998 ex C-Greenland herb. s.n.); Wollaston Forland, Landingsdalen, 74°27.5′N, 19°03.1′W, 28 Jul 1929, J. Vaage s.n. (O-V-688113).

Literature record (specimen not seen) — On Silene uralensis subsp. arctica. – CANADA: Nunavut, Qikiqtaaluk Region, Hudson Bay, Mansel Island (located at c. 61°35′–62°25′N), 30 Aug 1884, R. Bell 2621 (CAN 54784, n.v., host as “Lychnis nesophila Holm”; in Savile 1959: 994 as “U[stilago]. violacea”) (Fig. 5).

Etymology — The epithet refers to the distribution of the species in the High Arctic of Greenland and in the Canadian Arctic Archipelago.

Table 1.

Microbotryum herbarium specimens and NCBI nucleotide database accession numbers used for phylogenetic analysis (newly generated sequences indicated in boldface).

Continued

Discussion

For the first time from Greenland, anther-smut fungal infection of Silene uralensis subsp. arctica was reported by Hagen (1947, as Ustilago violacea on Melandrium apetalum (L.) Fenzl), based on the following collections:

Greenland: Sabine Island, near Germania Havn (on the south side of the island, c. 74°32.2′N, 18°49.9′W), 22 Jul 1932 & 16 Aug 1932, S. Aandstad s.n. (n.v.), and 21 Jul 1933, A. Hagen, the Norwegian Expedition to NE Greenland 1933 s.n. (n.v.); Gael Hamke Bugt, Jackson Island, c. 73°55′N, 11 Aug 1933, A. Hagen, the Norwegian Expedition to NE Greenland 1933 s.n. (n.v.); Hold with Hope, Stormdalen, c. 73°29.5′N, 20°46.9′W, 9 Aug 1933, A. Hagen, the Norwegian Expedition to NE Greenland 1933 s.n. (n.v.); Troldsøen (as “Trollvatnet”), c. 73°29′N, 20°39′W, 9 Aug 1933, A. Hagen, the Norwegian Expedition to NE Greenland 1933 s.n. (n.v.) (Fig. 5); as well as on the last specimen in the above-mentioned list of paratypes of Microbotryum arcticum.

For six of these collections, no voucher specimens were found in the herbarium in Oslo (O) (K.-H. Larsson, pers. comm.), although we assume that they represent Microbotryum arcticum, as Hagen's spore measurements (5–8 µm in diam.) correspond to our data.

The host plant of the Canadian specimens of Microbotryum arcticum listed here was originally identified by D. B. O. Savile and J. A. Parmelee (coll. no. 2215) as “Lychnis apetala var. arctica (Th. Fr.) Cody”. The infected plant specimens listed from North Greenland (deposited in C) were originally identified as “Melandrium apetalum subsp. arcticum (Th. Fr.) Hultén” or “Silene uralensis subsp. apetala (L.) Bocquet” (coll. no. 534). All of these are considered here as S. uralensis subsp. arctica, in accordance with Elven & al. (2011).

The taxonomic treatments of this plant vary considerably. Silene uralensis (Rupr.) Bocquet (S. sect. Physolychnis) is a very variable species complex (Morton 2005) with not completely clarified specific and infraspecific delimitation. In Bocquet's treatment of S. sect. Physolychnis (1967), four subspecies were recognized within S. uralensis, namely subsp. uralensis, subsp. apetala, subsp. arctica and subsp. porsildii Bocquet (the last one a tetraploid taxon). The populations in Svalbard were treated as an endemic subsp. arctica, while those in Scandinavia and the islands of the Bering Sea were recognized as subsp. apetala (S. wahlbergella Chowdhuri). The remaining populations were referred to subsp. uralensis (with a northern circumpolar distribution). Hultén (1968) accepted two subspecies: Melandrium apetalum subsp. arcticum, mapped by him with a circumpolar distribution, and subsp. apetalum from Scandinavia. In Flora nordica (Jonsell 2001), however, the Fennoscandian plants were treated as a distinct species, S. wahlbergella, and, accordingly, only the arctic plants from North America and Asia were considered related to S. uralensis. In the Silene treatment for Flora of North America (Morton 2005), three subspecies were recognized within the S. uralensis complex: subsp. uralensis, widespread with an arctic circumpolar distribution; subsp. porsildii, in Yukon, Alaska and Arctic Asia; and subsp. ogilviensis (A. E. Porsild) D. F. Brunt. from the Canadian Low Arctic. Of these, only subsp. uralensis was given as present in Greenland and the eastern Canadian Arctic Archipelago. Elven & al. (2011) disagreed with Bocquet's view that subsp. arctica was restricted to Svalbard, and recognized three subspecies of S. uralensis: subsp. uralensis, with a circumpolar distribution (NE Europe, Arctic Asia, Bering Sea islands, Alaska, Canada and W and S Greenland); subsp. arctica, also with a circumpolar distribution (Arctic Far East of Russia, northernmost Alaska and Canada, Greenland and Svalbard); and subsp. ogilviensis.

Table 2.

Comparative morphological data for Microbotryum arcticum and M. savilei, causing anther infection and atypical infection, respectively, on Silene uralensis subsp. arctica.

Table 3.

Effect of Microbotryum arcticum and M. savilei on flower morphology of Silene uralensis subsp. arctica.

Two subspecies of Silene uralensis are present in Greenland (according to the taxonomic scheme of Elven & al. 2011). Silene uralensis subsp. uralensis is characterized by a calyx that is not strongly inflated and usually longer than broad and by petals slightly emerging from the calyx, less so than in subsp. arctica (Elven & al. 2011). The calyx of S. uralensis subsp. arctica is inflated, in the flowering stage c. 1.5 times as long as broad (Alsos & al. 2018). In Greenland, there is an overlap in the ranges of the northern subsp. arctica and more southern subsp. uralensis at 70–71°N, but there are no obvious transitional plants (Elven & al. 2011). In the first half of the last century, the High Arctic entity in Greenland, subsp. arctica, was referred to as “Melandrium apetalum” (e.g. Kruuse 1905; Ostenfeld & Lundager 1910; Hartz & Kruuse 1911; Ostenfeld 1926; Porsild 1926). The southernmost localities of subsp. arctica are at 69°42′N in West Greenland (Porsild 1926) and 69°30′N in East Greenland (Kruuse 1905), while northward it reaches 83°06′N (Maguire 1950). Considering that subsp. uralensis is distributed only on the western and southern coasts of Greenland, the host plant of the Microbotryum specimens recorded by Hagen (1947) on “Melandrium apetalum” from East Greenland (at 73°29′–74°32′N) should be accepted as belonging to subsp. arctica.

The above-mentioned specimen from Mansel Island was examined by Savile (1959). The plant specimen is a type of Lychnis nesophila Holm (Holm 1907; see Aiken & al. 2007a). As was noted by Savile (1959: 983), this type is based on material of Silene uralensis subsp. arctica infected with a smut fungus causing typical anther infection. This smut fungus was not revised in the course of this study, but, based on the symptoms described by Savile, it may be presumed that it belongs to Microbotryum arcticum. This locality is the southernmost for M. arcticum.

Other Canadian specimens of “Ustilago violacea” on “Melandrium apetalum f. arctica” [sic] were recorded by Lind (1910), from Nunavut, King William Island (in the vicinity of Gjoa Haven, as “Gjöa Harbour”, 68°37′38″N, 96°23′40″W, 2 Aug 1905, C. H. Ostenfeld s.n.), and by Liro (1924) and Savile (1957), from Nunavut, Ellesmere Island (Goose Fjord, c. 76°30′N, 88°35′W, 15 Aug 1901, H. G. Simmons s.n., H). These records were not taken into account in this treatment, but probably they also belong to Microbotryum arcticum.

Microbotryum arcticum seems to be a common species on Silene uralensis subsp. arctica in the Canadian Arctic Archipelago. Under the name Ustilago violacea var. violacea, it was reported as a widespread smut fungus on this host at Queen Elizabeth Islands (Savile & Parmelee 1964) and throughout Somerset Island (Savile 1959).

On the basis of available information in Savile (1959), Kevan & Parmelee (1972), Aiken & al. (2007b), Elven & al. (2011), and Alsos & al. (2018), as well as our observations on the cited Microbotryum specimens from DAOM, contrastive morphological characters of the healthy flowers of Silene uralensis subsp. arctica are compared with those of flowers of the same plant with anther infection, and with flowers with atypical infection (Table 3).

Microbotryum arcticum causes typical anther infection, leading to the development of smutted anthers (Fig. 4B, C), which are filled with teliospores instead of fertile pollen grains.

Because the soral development of Microbotryum savilei has not yet been studied, the infection caused by this species is provisionally referred to here as atypical infection. A suite of characters distinguishes both species. Microbotryum savilei can be easily differentiated from the newly described species by the soral morphology and changes in the affected flowers (Table 3). The healthy plants of S. uralensis subsp. arctica (in the flowering stage) and the plants infected by M. arcticum have nodding flowers, whereas the plants infected by M. savilei have erect flowers. The flowers infected by M. arcticum are not deformed; they are open and not completely filled with a spore mass. The flowers infected by M. savilei are closed, deformed and swollen (more so than in the healthy flowers), with a broadly ovoid or very broadly ovoid calyx (with height/width ratio 0.8–1.2), and are completely filled with a spore mass. In the infected flowers, the petals remain within the closed calyx and are never emergent; the calyx is erect and closed at the throat; not only the anthers but also the filaments are affected (Fig. 2D, 4C); the stamens may or may not be completely destroyed; the ovary and styles remain rudimentary and no seeds are produced.

As transmission of the fungus is mediated by insects that are both vectors of the disease and pollinators of the plant (Kevan & Parmelee 1972; Biere & Honders 1996; Fontaine & al. 2013), the flower modifications due to Microbotryum arcticum (specifically, the deeply pigmented flower, dark-coloured spore mass in the smutted anthers, and presence of a more elongated calyx that is more open at the throat, and more exserted petals) could result in more insect visits, which is extremely important under the High Arctic conditions.