Powdery mildews on crops and ornamentals in Canada: a summary of the phylogeny and taxonomy from 2000 – 2019

Abstract

Powdery mildew diseases on crops and ornamental plants are common and cause significant economic losses, by reducing the yield and quality of crops and downgrading the value of ornamentals. In surveys published annually in the Canadian Plant Disease Survey, powdery mildews were reported frequently. In recent years, the taxonomy of powdery mildew fungi has changed tremendously as the result of the increased application of molecular phylogenetic analyses. Several generic concepts have been re-adjusted. Many traditionally considered species turned out to be a complex of several phylogenetic species. In order to facilitate effective communication in the scientific community, it is important to apply the current correct names for field identifications. In this document, we summarize the powdery mildew fungi reported in the Canadian Plant Disease Survey from 2000 to 2019, address the taxonomic and nomenclatural issues associated with the listed species, and tentatively suggest correct names based on host plants wherever possible. Nevertheless, a reliable identification of the powdery mildew must depend on DNA sequence analyses and morphological examinations.

Résumé

Les maladies des cultures et des plantes ornementales causées par l’oïdium sont courantes et entraînent d’importantes pertes économiques en réduisant le rendement et la qualité des cultures ainsi qu’en dépréciant la valeur des plantes ornementales. Dans les études publiées annuellement dans l’Enquête phytosanitaire nationale, les oïdiums sont souvent mentionnés. Au fil des dernières années, la taxinomie des champignons responsables de l’oïdium a changé substantiellement à cause de l’utilization accrue des analyses phylogénétiques moléculaires. Certains concepts génériques ont été modifiés. Plusieurs espèces traditionnelles ont émergé en tant que complexe de plusieurs espèces phylogénétiques. Afin d’établir une communication efficace entre scientifiques, il importe d’utilizer les noms courants exacts pour identifier les espèces sur le terrain. Dans cet article, nous faisons une synthèse des champignons responsables de l’oïdium mentionnés dans l’Enquête phytosanitaire nationale de 2000 à 2019, nous traitons des problèmes associés à la taxinomie et à la nomenclature des espèces listées et nous suggérons provisoirement, dans la mesure du possible, des noms corrects procédant des plantes hôtes. Néanmoins, une identification fiable de l’oïdium doit dépendre des analyses des séquences de l’ADN et des examens morphologiques.

Keywords:

• Ascomycota
• Erysiphaceae
• Helotiales
• identification
• nomenclature
• taxonomy

Mots clés:

• Ascomycètes
• Erysiphaceae
• Helotiales
• identification
• nomenclature
• taxinomie

Introduction

Powdery mildew fungi were historically classified as a family, Erysiphaceae, or as an order, Erysiphales (see the detailed historical account in Braun and Cook 2012). A recent class-wide multi-gene phylogeny of Leotiomycetes indicated Erysiphales is a member of Helotiales (Johnston et al. 2019). They are commonly observed causing infections on leaves, young branches, and occasionally on fruits, of various agricultural and ornamental plants. They can be easily recognized by the powdery-looking patches or a diffusive layer of white, yellow, brown or greyish mycelia and conidiophores on plant surfaces, at times decorated with globose, needle-tip sized, yellow-, brown- to black-coloured sexual reproduction structures (ascomata, more specifically chasmothecia, Agrios 1997). The significant economic impacts of powdery mildew to agriculture and horticulture, the conspicuous development of unique infection structures on the plant surfaces, and diverse plant-parasite interactions have made them favoured model organisms for research on biotrophic pathogens and their interactions with their host plants (Bushnell 2002). Research reports in diverse fields, i.e. disease control, ecology and evolution, epidemiology, fungicide resistance and molecular genetics are copious and were summarized periodically in comprehensive treatises and reviews (Spencer 1978; Bélanger et al. 2002; Jarvis et al. 2002; Bindschedler et al. 2016; Bourras et al. 2018; Vielba-Fernández et al. 2020; Scott 2021).

The taxonomy of powdery mildews has evolved significantly since de Candolle erected the genus Erysiphe and described eight species within it (de Candolle 1805). Léveillé (1851) introduced a six-genera classification based on the shape and structure of chasmothecial appendages and number of asci per ascomata, i.e. Podosphaera and Sphaerotheca were characterized as containing one ascus per ascoma, but the appendages of the former were ramoso-dichotomae (dichotomously branched), and the latter floccosae (hair-like, mycelioid). Different from these two, the ascomata of the other four genera contained multiple asci, and can be separated by the features of appendages as follows: Phyllactinia aciculatae (erroneously applied for acicularis = needle-like), Uncinula uninatae (hooked), Calocladia ramoso-dichotomae (Microsphaera was applied to replace Calocladia on page 381 in the same publication, because the name Calocladia had been previously used for algae), and Erysiphe floccosae. This generic classification was handy and widely accepted by mycologists in North America and Eurasia for about 150 years (Salmon 1900; Blumer 1933; Homma 1937; Braun 1987; Chen et al. 1987), until molecular phylogenies revealed that the feature of unbranched mycelioid appendage was not synapomorphic as to separate Sphaerotheca from Podosphaera, or Erysiphe from Microsphaera and Uncinula, i.e. species of the genera concerned are phylogenetically not differentiated and form large monophyletic clades. As a result, Braun and Takamatsu (2000) merged Sphaerotheca with Podosphaera, and Microsphaera and Uncinula (and numerous small, segregated genera, such as Medusosphaeria and Uncinuliella) with Erysiphe. All species in Sphaerotheca and Microsphaera, Uncinula, etc., were made into new combinations, i.e. Podosphaera spp. and Erysiphe spp., respectively.

Historically, North American and Eurasian researchers had two divergent species level classification systems. In North America, influenced by Salmon’s (1900) system, broad species concepts were often applied as opposed to narrower concepts in Eurasia. As an example, Erysiphe polygoni was considered a species occurring on plants in seven families by Salmon and North American researchers (Parmelee 1977), while only on Polygonaceae by European researchers (Blumer 1967; Junell 1967; Braun 1987). Likewise, many other species held this type of duel concepts including Golovinomyces cichoracearum, Leveillula taurica, Podosphaera macularis, etc. In light of molecular phylogenetic studies, many morphological species with a broad concept, and even some narrow concept species, proved to be species complexes, which can be divided into several phylogenetic species.

In recent decades, numerous species complexes within several powdery mildew genera have been successfully re-examined based on molecular approaches: (1) Blumeria genus revision (Liu et al. 2021); (2) Erysiphe – Erysiphe sect. Erysiphe and sect. Microsphaera basic phylogenetic examination (Takamatsu et al. 2015a) and Erysiphe sect. Uncinula (Takamatsu et al. 2015b), phylogenetic revisions of the Erysiphe alphitoides complex (Takamatsu et al. 2007), and additionally Erysiphe spp. on Carpinus spp. (Braun et al. 2006), on Papaveraceae (Pastirčáková et al. 2016), on Viburnum spp. (Bradshaw et al. 2020b), on Lonicera spp. (Bradshaw et al. 2021c), on Corylus spp. (Bradshaw et al. 2021d), on Salix and Populus spp. (Darsaraei et al. 2021), on Fraxinus spp. (Yamaguchi et al. 2021); (3) Golovinomyces – phylogeny at the genus level (Matsuda and Takamatsu 2003; Takamatsu et al. 2013), the G. biocellatus complex (Scholler et al. 2016), G. chrysanthemi and allied species on Asteraceae (Bradshaw et al. 2017), G. cynoglossi complex on Boraginaceae (Braun et al. 2018), the plurivorous G. orontii complex (Braun et al. 2019b), the G. ambrosiae, G. circumfusus and G. spadiceus complex on composites, including G. latisporium on Helianthus spp. (Qiu et al. 2020); (4) Neoerysiphe – phylogenetic revision at genus level Takamatsu et al. 2008a; (5) Phyllactinia – phylogeny at the genus level (Takamatsu et al. 2008b, 2016); and (6) Podosphaera – the phylogeny of species on rosaceous hosts (Takamatsu et al. 2010), of sect. Sphaerotheca subsect. Magnicellulatae (Ito and Takamatsu 2010), North American Podosphaera species on Prunus spp. (Moparthi et al. 2019), phylogenetic revision of the P. tridactyla complex (Meeboon et al. 2020).

In Canada, the taxonomy of powdery mildews was studied and recorded in the regional flora of Manitoba and Saskatchewan (Bisby et al. 1938), Ontario (Parmelee 1977), and the Maritimes (Wehmeyer 1950), and also summarized in a compendium of plant diseases (Conners 1967) and the Canadian Plant Disease Survey (CPDS) since the 1920s, a serial research report of the Plant Research Institute (Research Branch), Department of Agriculture Canada (currently, Science and Technology branch, Agriculture and Agri-Food Canada). This summary only focuses on the incidence of powdery mildew reported on crops and ornamental plants in the CPDS over the past two decades. Due to historical reasons (explained earlier), many identification records in the CPDS have been based on broad species concepts, and could be adjusted to be consistent with the up-to-date molecular-based classification. Here, we discuss varied species concepts and up-to-date research results for particular reported species. However, it is beyond the scope of this summary to provide a reliable identification for individual cases, which should depend on careful molecular and morphological re-examination.

In the following paragraphs, most species are listed as they were reported in the CPDS, with the exception of Blumeria on Hordeum and on Avena, which were assumed to be B. hordei and B. avenae (Liu et al. 2021). Sphaerotheca spp. were corrected to Podosphaera spp. with Sphaerotheca spp listed as a synonym; in the same manner Microsphaera and Uncinula spp. were listed as Erysiphe spp. The provinces where these reports originated are abbreviated as follows: AB (Alberta), BC (British Colombia), MB (Manitoba), NB (New Brunswick), NS (Newfoundland & Labrador), ON (Ontario), QC (Quebec), and PEI (Prince Edward Island). Detailed bibliographic references of the particular powdery mildew names are not given, but they are available via Index Fungorum and MycoBank (http://www.indexfungorum.org/names/names.asp and https://www.mycobank.org/) as well in taxonomic standard references, such as Braun and Cook (2012).

1. Blumeria graminis (DC.) Speer 1975 [1973–1974]), s. str. (emend., see Liu et al. 2021)

≡ Erysiphe graminis DC. 1815

Powdery mildews on wheat (durum, spelt, spring and winter wheat) were reported almost each year from 2001 to 2019 in Ontario (Xue 2002b; Xue et al. 2003b, 2004b, 2004c, 2005b, 2006b, 2007b, 2009; Xue and Chen 2010b, 2011b; Tamburic-Ilincic and Sparry 2012; Xue and Chen 2012b, 2013b; Shan and Sabourin 2014a, 2014b; Tamburic-Ilincic et al. 2014; Xue and Chen 2014b; Tamburic-Ilincic 2015; Xue and Chen 2015b; Melzer and Shan 2016; Tamburic-Ilincic 2016; Xue and Chen 2016b; Xue et al. 2017b; Melzer and Shan 2018; Xue and Chen 2018, 2019b; Melzer and Shan 2020; Xue and Chen 2020b). Likewise, they were frequently reported from Manitoba (Desjardins et al. 2001, 2003, 2007, 2008, 2009, 2012, 2013, 2014, 2015; Pradhan et al. 2016, 2017, 2018, 2019, 2020), and Quebec (Rioux and Comeau 2003; Gilbert et al. 2009, 2010, 2016; Rioux 2014, 2015, 2016, 2017, 2018, 2019; Rioux and Copley 2020); occasionally in Alberta (Orr and Turkington 2001, 2002, 2003; Xi et al. 2018; Chang et al. 2019), Prince Edward Island (Clark and Cheverie 2008; Clark and Beaton 2009; Clark 2010; Clark and MacLeod 2019), and Saskatchewan (Holzgang and Pearse 2002, 2007; Morrall and Fernandez 2002; Dokken-Bouchard et al. 2017). The incidence and severity varied from low to medium.

Notes: Powdery mildew on cereal crops and grasses was considered as a single species since de Candolle (1815) first described it as Erysiphe graminis, which was later moved to a new genus Blumeria (Speer 1975 [1973–1974]). Biological races, forms and formae speciales of varied host specificity were recognized within species based on inoculation experiments (Marchal 1902; Mains and Dietz 1930; Oku et al. 1985; Troch et al. 2014). Molecular phylogenetic analyses of multiple genes as well as whole genomes further confirmed the intra-specific genetic subdivision, and also shed light on the varied evolutionary trajectories of different lineages (Inuma et al. 2007; Menardo et al. 2017). Only recently was a formal taxonomic treatment published that recognized eight species, which are B. graminis sensu stricto (s. str.), B. americana, B. avenae, B. bromi-cathartici, B. bulbigera, B. dactylidis, B. graminicola and B. hordei (Liu et al. 2021). Although minute morphological variations and host ranges can separate certain species from others, molecular evidence based on multi-gene phylogenies played a crucial role (Liu et al. 2021). Powdery mildew on wheat predominantly belonged to B. graminis s. str.

• (2) Blumeria hordei M. Liu & Hambl. 2021 in Liu et al. 2021

In Ontario, powdery mildew on barley was reported as frequently as on wheat (Xue 2002a; Xue et al. 2003a, 2004a, 2005a, 2006a, 2007a, 2017a; Xue and Chen 2009, 2010a, 2011a, 2012a, 2013a, 2014a, 2015a, 2016a, 2019a, 2020a). Localized incidences of barley powdery mildew were also reported in Prince Edward Island (Clark and Beaton 2009, 2011; Clark 2010; Clark and Driscoll 2016) and Quebec (Rioux and Comeau 2003; Rioux 2015, 2016, 2018, 2019); and occasionally reported from Alberta (Orr and Turkington 2003, 2004), Manitoba (Desjardins et al. 2002; Pradhan et al. 2020), and Saskatchewan (Holzgang and Pearse 2006). The incidence and severity were from low to medium.

Notes: It was noted that barley powdery mildew contains more physiological forms than wheat powdery mildew (Mains and Dietz 1930; Mains 1933). Molecular markers coupled with population genetic analyses also indicated enormous variation among populations (Wolfe and McDermott 1994). In our recent molecular taxonomic study, predominant samples of powdery mildew on barley belonged to B. hordei, and the type specimen was on Hordeum vulgare from QC. It is likely that the pathogens causing powdery mildew on barley in Canada belong to B. hordei. However, there is still a possibility that they could belong to other species, i.e. B. graminis s. str., B. americana and B. dactylidis (Liu et al. 2021). The following minute morphological differences between B. hordei and B. graminis could help identification: (1) the development of primary mycelia and asexual morph of B. graminis s. str. is usually halted by the onset of secondary mycelia, while for B. hordei both primary and secondary mycelia co-exist; and (2) the colour of mycelia of B. graminis s. str. is yellowish brown to rusty reddish brown, while B. hordei is purplish brown or greyish brown, and the secondary mycelia are much lighter in colour. For the confirmation of their identities, DNA sequence analyses is required or at least urgently recommended.

• (3) Blumeria avenae M. Liu & Hambl. 2021

Reported in 2018 and 2019 in Manitoba (Pradhan et al. 2019, 2020).

Notes: Powdery mildew on oats is very likely B. avenae, which proved to be a coherent lineage/species restricted on Avena spp. (Inuma et al. 2007; Liu et al. 2021). In the Canadian National Mycological Herbarium (DAOM), there are two samples from Ontario collected in the 1960s on A. barbata (DAOM 147852) and A. sativa (DAOM 159514), respectively. DAOM 147852 was included in analyses and belonged to B. avenae (Liu et al. 2021). Incidences were also reported in BC, ON, and QC from the Canadian Plant Diseases Survey before the1980s (Conners 1967; Ginns 1986).

• (4) Erysiphe adunca (Wallr.) Fr. 1829

= Uncinula adunca (Wallr.) Lév. 1851

On willow (Salix sp.) in Quebec in 2019 (Breton et al. 2020).

Notes: Besides Erysiphe adunca, Braun and Cook (2012) recorded five other powdery mildew species on willows, viz. Podosphaera schlechtendalii Lév., Phyllactinia populi (Jacz.) Y.N. Yu, Pleochaeta salicicola R.Y. Zheng & G.Q. Chen, E. capreae DC. ex Duby and E. pseudoregularis U. Braun. Genus-specific characteristics can distinguish E. adunca s. str. from the three species in genera Podosphaera, Phyllactinia and Pleochaeta easily. The E. adunca complex has recently been phylogenetically and taxonomically revised and split into several species (Darsaraei et al. 2021): Erysiphe adunca s. str. is confined to Populus spp. as hosts and is also widespread and common in North America; E. salicis and E. capreae (including E. pseudoregularis as a synonym) are two common species on Salix spp. So far, only E. capreae has been confirmed phylogenetically for North America (USA). Erysiphe capreae and E. salicis can been distinguished by different host preferences, and the asexual and sexual morphological features. Erysiphe capreae is on Salix caprea and allied host species belonging to Salix subgen. Vetrix sect. Vetrix; conidiophores very long, 40–200 μm, number and arrangement of cells variable, but often with a shorter foot-cell followed by a longer second cell; chasmothecial appendages numerous, about (33–)50–400, around the middle of the chasmothecium (equatorial) or somewhat in the upper half, forming a dense circle around the chasmothecium, when fully mature, pointing upwards at an angle of about 40–75°, uniformly short, (0.3–)0.5–1.2(–1.5) times the chasmothecial diam., 5–12-μm wide throughout. On the other hand, E. salicis is mainly on Salix spp. belonging to Salix subgen. Salix; conidiophores shorter, 35–110 μm, mostly a longer foot-cells followed by shorter cells or cells of about the same length; chasmothecial appendages longer, 0.5–2 times as long as the chasmothecial diam., mostly slimmer, horizontally spread, i.e. without tendency to point upwards (Darsaraei et al. 2021). However, the entire E. adunca complex requires comprehensive morphological and molecular re-examinations, especially in North America.

• (5) Erysiphe azaleae (U. Braun) Braun and Takamatsu (2000)

≡ Microsphaera azalea U. Braun (1982), nom. cons. (see Braun 2013)

= Oidium ericinum Erikss. 1885

On Azalea sp. (current name: Rhododendron sp. (Joshi and Jeffries 2011), Rhododendron spp. (Elmhirst 2013; Elmhirst and Karlsson 2014, 2015, 2016) in British Columbia.

Notes: Braun and Cook (2012) recorded Phyllactinia enkianthi Z.Y. Zhao and five Erysiphe spp. on Rhododendron, viz. Erysiphe azaleae, E. digitata (A.J. Inman & U. Braun) A.J. Inman & U. Braun, E. izuensis (Y. Nomura) U. Braun & S. Takam., E. rhododendri J.N. Kapoor, and E. vaccinii Schwein. In addition, U.S. National Fungus Collections Fungus-Host Database (accessed on Feb 18, 2021, abbreviated as BPI Fungus—Host database in the following text) also recorded E. cruciferarum Opiz ex L. Junell and E. polygoni DC on Rhododendron. Among the seven Erysiphe spp. on Rhododendron, E. digitata, E. izuensis, and E. rhododendri are distributed in Asia or/and Europe. All specimens of E. cruciferarum on Rhododendron were from Scotland or England. Excluding these species with restricted geographic ranges, the remaining are E. azaleae, E. vaccinii and E. polygoni. The records of ‘E. polygoni’ referring to the species name of Salmon’s (1900) extremely wide circumscription can be ruled out, because E. polygoni (s. str.), i.e. the true E. polygoni, is confined to hosts of the Polygonaceae (Braun and Cook 2012), which has also been confirmed by phylogenetic analyses (Takamatsu et al. 2015a). Erysiphe azaleae is morphologically readily identifiable and easily distinguishable from E. vaccinii by having uniformly short chasmothecial appendages with apices 4–6 times dichotomously branched and recurved tips (Inman et al. 2000; Braun and Cook 2012), as opposed to much longer appendages, 1.5–7 times as long as chasmothecial diam., with straight or knob-like tips of the ultimate branchlets in E. vaccini (Braun and Cook 2012).

• (6) Erysiphe begoniicola Braun and Takamatsu (2000)

= Microsphaera begoniae Sivan. 1971

On Begonia sp. in British Columbia in 2008 (Joshi and Jeffries 2009).

Notes: Species on Begonia include Golovinomyces orontii (Castagne) Heluta, E. begoniae R.Y. Zheng & G.Q. Chen, and E. begoniicola (= Microsphaera begoniae Sivan.). Erysiphe begoniicola produces significantly larger conidia of more varied shapes, (25–)35–60(–70) × (12–)14–22.5(–25) µm, than E. begoniae, mostly subcylindrucal-doliiform 23–38 × 13–16 µm. The chasmothecial appendages of E. begoniicola (Erysiphe sect. Microsphaera) are branched dichotomously (Sivanesan 1971; Braun and Cook 2012), whereas, those of E. begoniae (Erysiphe sect. Erysiphe) are mycelioid and mostly un-branched (Chen et al. 1987; Braun and Cook 2012).

• (7) Erysiphe cruciferarum Opiz ex L. Junell (1967)

On Wasabia japonica in British Columbia (Joshi et al. 2014; Joshi and Jeffries 2016) (Betz and Punja 2017); on canola (Brassica napa) in Ontario (Melzer and Shan 2019); on Cléome épineux (Cleome spinosa) in Quebec (Breton et al. 2018).

Notes: Erysiphe cruciferarum was recorded on a wide range of host plants in five families, the Brassicaceae, Capparidaceae, Fumariaceae, Papaveraceae, and Resedaceae (Braun and Cook 2012). It is very likely a species complex. Pastirčáková et al. (2016) reported that E. cruciferarum on Papaveraceous hosts represents distinct lineages from that on Brassicaceous hosts based on ITS and 28S DNA sequence analyses, but did not provide intra-specific classification.

Paul and Thakur (2006) noted the morphological differences of the specimens on different host plants in India and recognized three varieties, viz. E. cruciferarum var. chandra, E. cruciferarum var. cleomes and E. cruciferarum var. indica. However, they failed to include the word ‘typus’ or ‘holotypus’ to indicate the type specimen, therefore, these names were considered invalid according to International Code of Nomenclature for algae, fungi, and plants Article 40.6 (Turland et al. 2018). Gorter (1988) introduced E. cruciferarum var longispora on Lepidium africanum in South Africa based on longer conidia. However, Braun and Cook (2012) reduced this name to synonymy with E. cruciferarum because the conidial size of this variety falls into the morphological variation of E. cruciferarum.

• (8) Erysiphe diffusa (Cooke & Peck) Braun and Takamatsu (2000)

= Microsphaera diffusa Cooke & Peck 1873

On soybean (Glycine max (L.) Merr.) in ON in 2000 and 2006 (Anderson and Tenuta 2001; Sabourin et al. 2007). On lentil (Lens culinaris Medik) in SK in 2003 (Banniza et al. 2004). This is the first record of powdery mildew on lentils in Canada.

Notes: Erysiphe diffusa, E. glycines F.L. Tai and Podophaera xanthii (Castagne) U. Braun & Shishkoff (Braun and Cook 2012) were recorded on Glycine max. Erysiphe glycines, belonging to E. sect. Erysiphe, can be distinguished from E. diffusa (E. sect. Microsphaera) by the unbranched or rarely irregularly branched appendages arising from the lower part of chasmothecia (Chen et al. 1987), whereas those of E. diffusa are dichotomously branched usually arising equatorially. The conidia of E. glycines can be much larger (25–50 × 13–25 µm) than those of E. diffusa (25–35 × 11–17.5 µm). Erysiphe diffusa is possibly a complex, but comprehensive phylogenetic examinations, including sequences retrieved from E. diffusa on its type host, Desmodium canadense, are not yet available.

For the first report of powdery mildew on lentil in SK in 2003, Banniza et al. (2004) tentatively identified it as Erysiphe diffusa (=Microsphaera diffusa) based on the branch pattern of the chasmothecium appendages, but also noted that the identification is not beyond doubt. In the literature, the causal agents of powdery mildew on lentil were also reported as Erysiphe polygoni DC. (Bayaa and Erskine 1998), and E. trifolii Grev. (Attanayake et al. 2009). In the recent monograph, Braun and Cook (2012) used the name E. trifoliorum (Wallr.) U. Braun, which has priority over E. trifolii according to the nomenclature code. Concepts of these species need to be clarified in the context of molecular phylogenetic analyses.

• (9) Erysiphe elevata (Burrill) Braun and Takamatsu (2000)

= Microsphaera elevata Burrill 1874

On woody ornamentals Catalpa bignonioides Walter in BC in 2019 (Joshi and Berlakoti 2020).

Notes: Braun and Cook (2012) recorded Erysiphe catalpae Simonyan, E. elevata, Fibroidium hiratae (U. Braun) U. Braun & R.T.A. Cook, Leveillula catalpae U. Braun, Neoerysiphe galeopsidis (DC.) U. Braun, and Phyllactinia catalpa U. Braun on Catalpa bignonioides, and Podosphaera catalpa (Z.Y. Zhao) U Braun on C. macrocarpa. Erysiphe elevata differs from E. catalpae in that E. elevata has dichotomously branched chasmothecial appendages, equatorial, long and flaccid, (1.5–)3–5(–7) times as long as the diameter of chasmothecium, and the foot-cells of conidiophores sometime are twisted, whereas the chasmothecial appendages of E. catalpae are unbranched and shorter, 0.4–2 times as long as chasmothecium diam., and the foot-cells of conidiophores are often straight. Neoerysiphe galeopsis (and any species in Neoerysiphe) can be distinguished by the presence of catenescent conidium chains, and the lack of maturity of asci before overwintering.

• (10) Erysiphe heraclei DC. 1815

On carrots (Daucus carota subsp. sativus (Hoffm.) Schübl. & Martens) in British Columbia (Joshi and Jeffries 2006), and Manitoba (Desjardins et al. 2003).

Notes: Potential species on Daucus carota include Leveillula braunii Simonyan & Heluta, L. lanuginosa (Fuckel) Golovin, and E. heraclei. Species in Leveillula can be distinguished from Erysiphe spp. by the narrow hyphae partially growing inside of plant leaves, and conidiophores arising from internal hyphae through stomata, significantly larger conidia (30–115 × 8–30) often dimorphic (Braun and Cook 2012). Paul and Thakur (2006) recognized two varieties, E. heraclei var. heraclei and E. heraclei var. himalayensis Paul and Thakur (2006), however, the latter name is invalid (nom. inval., Art. 40.6).

• (11) Erysiphe liriodendri Schwein. 1832

On tulip tree (Liriodendron sp.) in Ontario (Melzer and Shan 2020).

Notes: Two species were known on Liriodendron sp., viz. Phyllactinia liriodendri U. Braun and Erysiphe liriodendri Schwein. Both species are common in North America (Braun and Cook 2012). However, they can be easily distinguished by the generic characteristics of Phyllactinia, i.e. the bulbous swelling at the base of chasmothecium appendages, presence of the specialized penicillate cells at the apical region of chasmothecia, conidia of varied shapes (clavate, lanceolate, angular-rhombiform etc.).

• (12) Erysiphe ludens (E.S. Salmon) Braun and Takamatsu (2000)

On faba bean (Vicia faba L.) in SK in 2003 (Banniza et al. 2004).

Notes: In the 1970s, this species was also reported on faba beans in MB and SK as Microsphaera penicillata var. ludens (E.S. Salmon) W.B. Cook (McKenzie and Morrall 1975; Kharbanda and Bernier 1977), which is now a taxon considered as a synonym of E. ludens (Braun and Cook 2012).

In addition to E. ludens, other species reported on faba bean include E. viciae-unijugae (Homma) U. Braun on Lathyrus spp. and Vicia spp. distributed in Asia, E. pisi DC (= E. polygoni sensu Salmon (1900)) on numerous genera in Fabaceae, with a worldwide distribution, Leveillula papilionacearum (Kom.) U. Braun (= L. taurica s. lat.) on legumes. The chasmothecial appendages of E. ludens are dichotomously branched at the apices, regularly 4–6(–7) times, a characteristic distinct from E. pisi (seldom irregularly branched).

• (13) Erysiphe magnifica (U. Braun) Braun and Takamatsu (2000)

One report on Magnolia × soulangiana in ornamental nurseries in BC in 2012 (Elmhirst 2013).

Notes: Braun and Cook (2012) recorded four species on Magnolia spp. including Erysiphe bulbosa (U. Braun) U. Braun & S. Takam., E. magnifica, E. magnoliae (Sawada) U. Braun & S. Takam., and Phyllactinia magnolia Y.N. Yu & S.J. Han. Erysiphe bulbosa is characterized by the bulbous swelling at the base of chasmothecium appendages similar to Phyllactinia, however, it differs from Phyllactinia by the branching patterns at appendage apices, and lack of the specialized penicillate cells at the apical region of chasmothecia. Erysiphe magnoliae was described in Japan, and its known distribution is limited to Japan and Korea. Morphologically, the apices of branched chasmothecium appendages in E. magnoliae are always straight, whereas they are distinctly recurved, even circinate in E. magnifica. Another species on Magnolia from Korea and Japan, E. magnoliicola S.E. Cho, S. Takam. & H.D. Shin, described based on ITS tree and morphological characteristics, is closely related with E. magnolia. However, it is distinct from E. magnifica and E. magnoliae by producing rather short conidiophores and with non-inflated foot-cells (Cho et al. 2014)

• (14) Erysiphe necator Schwein. 1832 [1834]

= Uncinula necator (Schwein.) Burrill (1892)

Frequently reported on grape vines (Vitis sp.) in BC (Sholberg and Haag 2003; Joshi 2004; Joshi and Jeffries 2011, 2016; Elmhirst 2013; Joshi et al. 2013; Elmhirst and Karlsson 2014; Joshi and Berlakoti 2020) and QC (Gilbert et al. 2007, 2008, 2014; Breton et al. 2019, 2020); occasionally in MB (Desjardins et al. 2009) and ON (Ker and McFadden-Smith 2002a).

Notes: Erysiphe necator can infect several genera in the grape family (Vitaceae), i.e. Ampelopsis, Cissus, Parthenocissus, and Vitis. Salmon (1900) noted its occurrence on Actinidia (Actinidiaceae), which refers, however, to a powdery mildew nowadays known as Erysiphe actinidiae (Miyabe ex Jacz.) U. Braun & S. Takam. Peck (1872) recognized U. ampelopsis on woodbine (possibly Parthenosissus sp.). Salmon (1900) considered U. ampelopsis as a synonym of E. necator as the morphological differences between the two species were not clear-cut. Braun (1983) observed that the appendages of the mildew on Parthenosissus sp. were shorter and stiff compared with E. necator although their length may overlap, thus accepted the taxon as the variety, E. necator var. ampelopsidis. Phylogenetic examinations are still necessary to establish the relation between Erysiphe on Ampelopsis and Vitis.

Conversely, other mildews can infect grape vine as well, such as Phyllactinia ampelopsidis Y.N. Yu & Y.Q. Lai.

• (15) Erysiphe penicillata (Wallr.) Link (1824), s. lat. (sensu Salmon 1900; Parmelee 1977)

≡ Microsphaera penicillata (Wallr.) Sacc. 1882

= ‘Microsphaera alni (Wallr.) G. Winter’ (see notes in Braun and Cook 2012, p. 492)

Sporadically reported on Magnolia sp. in BC (Joshi and Jeffries 2017), on nanny berry (Viburnum lentago L.) in MB (Desjardins et al. 2003), on Euonymus sp. in ON (Sabourin et al. 2006), on Amelanchier sp. and Sambucus sp. in QC (Gilbert et al. 2007), and on Lilac (Syringa vulgaris L.) in SK (Holzgang and Pearse 2007).

Notes: These reports refer to the old, outdated taxonomy introduced by Salmon (1900), with an extremely wide species concept, which was used for a long time in North America, such as in Parmelee (1977). Erysiphe penicillata s. str. was described on the basis of a German powdery mildew collection on Alnus glutinosa and represents a species confined to multiple Alnus spp. In recently published phylogenetic studies, various species and species complexes, previously referred to as M. penicillata or to the wrongly cited and applied name ‘M. alni’, have been phylogenetically re-examined and in some cases even demonstrated to be complexes of several species, including cryptic undescribed species (Takamatsu et al. 2007; Bradshaw et al. 2020b, 2021c, 2021d). Hence, all reports cited above are unclear in terms of identities, and in urgent need of re-examination based on sequencing data.

• (16) Erysiphe pisi DC. 1805

On field pea crops (Pisum sativum L.), high level incidences were reported occasionally in AB (Su et al. 2002; Chang et al. 2004, 2005b; Wang et al. 2007), low to medium incidences were reported frequently from 2000 to 2012 in MB (Desjardins et al. 2001, 2006; Xue and Yager 2001; Yager and Conner 2002; Yager et al. 2003; McLaren et al. 2004, 2005, 2006, 2007, 2009, 2010, 2013); and similar or lower incidence in SK (Holzgang and Pearse 2001, 2002, 2006, 2007; Chongo et al. 2003; Dokken-Bouchard et al. 2010, 2011).

Notes: In the literature, the most common species recorded on pea (Pisum sativum L.) is Erysiphe pisi (previously referred to E. polygoni s. lat, sensu Salmon 1900). Erysiphe pisi has a wide host and geographic range on multiple legumes. According to Parmelee (1977), Blumer (1967) applied narrower species concepts and separated E. pisi from E. polygoni (as well as several other species) by the fewer and rarely branched appendages and by host ranges, i.e. E. pisi on legumes, while E. polygoni on the Polygonaceae. Having found the morphological characters overlapping among the specimens collected in Canada from coast to coast, Parmelee (1977) applied Salmon’s broad species concept of E. polygoni, thus species on Pisum were recognized as E. polygoni sensu Salmon. Braun and Cook (2012) accepted both as separate species, E. pisi on numerous genera in Fabaceae, E. polygoni on Polygonaceae. Bradshaw et al. (2021b) verified E. pisi as a monophyletic group, in that the ITS sequences retrieved from E. pisi worldwide formed a clade distinct from all allied taxa. In addition, Braun and Cook (2012) recognized a variety, E. pisi var. cruchetiana, on Ononis from Asia and Europe, which differs from E. pisi var. pisi by producing more abundant, irregularly branched appendages. Recently, Schmidt and Braun (2020) examined the asexual morphs of E. pisi var. pisi and var. cruchetiana in detail, including conidial germination, and found significant differences, which led to the conclusion and recommendation to maintain the Ononis Erysiphe as species, E. cruchetiana, i.e. as originally introduced. Boerema and Verhoeven (1979) noted host specialization among legume hosts, and separated E. pisi into four formae speciales, among which E. pisi f. sp. pisi was considered specialized on field and garden peas. A second powdery mildew, Erysiphe trifoliorum, has to be taken into consideration as causative agent of pea powdery mildew. It has recently been found in the USA, and confirmed by DNA sequence analyses (Attanayake et al. 2010). Bradshaw et al. (2021b) confirmed the occurrence of E. trifoliorum on Pisum sativum in Europe and the USA.

• (17) Erysiphe polygoni DC. 1805 s. lat. (sensu Salmon 1900; Parmelee 1977)

Reported on ornamentals (Hydrangea sp., Hylotelephium spp., Sedum sp.) and on beets (Beta vulgaris L.) in BC (Joshi and Jeffries 2009, 2016; Joshi et al. 2014; Joshi and Berlakoti 2020); on coriander (Coriandrum sativum L.) and clover (Trifolium pratense L.) in SK (Peluola et al. 2014, 2015); on clover (Trifolium sp.) in ON (Sabourin et al. 2006).

Notes: Erysiphe polygoni was first described by de Candolle (1805) as Erysiphe of knotweed (Polygonum). Salmon (1900) considered the morphological features of 30 or so other species that overlapped with it, and listed them as synonyms, including E. heraclei, E. liriodendri, E. pisi (as discussed in an earlier paragraph) and E. trifolii Grev., etc. The host range, as summarized by Parmelee (1977), includes 22 genera in seven families viz. Polygonaceae, Ranunculaceae, Fabaceae (Leguminosae), Onagraceae, Hypericaceae, Euphorbiaceae, and Lamiaceae (Labiatae). Represented by Blumer (1967), Junell (1967), and Braun (1987, 1995), European botanists/mycologists preferred to accept several different species in place of E. polygoni sensu Salmon, and applied the strict sense of E. polygoni (s. str.) on Polygonaceae only. Numerous phylogenetic studies clearly showed that E. heraclei, E. pisi, E. trifolii, etc., belonged to different clades, representing species distinct from E. polygoni s. str. (Takamatsu et al. 2002, 2015a; Abasova et al. 2018). If E. polygoni s. str. is applied, the pathogens reported on Hydrangea sp. are very likely Pseudoidium hydrangeae (Shin et al. 2019), on Hylotelephium spp. and Sedum sp. are likely Erysiphe sedi (Shin et al. 2019; Bradshaw et al. 2020a), on Beta vulgaris from BC are likely E. betae (Braun and Cook 2012), on Coriandrum sativum from SK are likely E. heraclei (Braun and Cook 2012), on Trifolium from SK and ON are likely E. trifoliorum (Bradshaw et al. 2021b). In addition, several Brassica spp. were reported as hosts of E. polygoni by Parmelee (1977) and Amano (1986). Based on the publications listed above, the pathogens likely belong to E. cruciferarum. The ‘E. polygoni’ on Aquilegia vulgaris and Delphinium spp. reported by Parmelee (1977) is likely E. aquilegiae (Table 1). Morphological re-examinations and molecular analyses are needed to confirm the above-assumed identities of Canadian collections.

• (18) Erysiphe sedi U. Braun 1981

Table 1. Additional previous records of powdery mildews on cultivated hosts in Canada and their putative current identities

Host	Previous identity	Reference	Putative current identity	Reference
Anthirrinum majus	Erysiphe cichoracearum	Amano 1986	Golovinomyces orontii (Castagne) Heluta	Braun et al. 2019b
Chrysanthemum morifolium	Erysiphe cichoracearum	Parmelee1977	Golovinomyces chrysanthemi (Rabenh.) M. Bradshaw, U. Braun, Meeboon & S. Takam.	Bradshaw et al. 2017
Cosmos bipartitus	Erysiphe cichoracearum	Parmelee1977	Identity unclear*
Dahlia variabilis	Erysiphe cichoracearum	Parmelee1977	Golovinomyces ambrosiae (Schwein.) U. Braun & R.T.A. Cook	Qiu et al. 2020
Hydrangea macrophylla	Oidium hortensiae	Amano 1986	Pseudoidium hortensiae (Jørst. ex S. Blumer) U. Braun & R.T.A. Cook	Shin et al. 2019, Bradshaw et al. 2020a
Linum usitatissimum	Oidium lini	Amano 1986	Podosphaera lini (Zvetkov) U. Braun & S. Takam.	Braun et al. 2019a
Lonicera tatarica	Microsphaera penicillata	Parmelee1977	Erysiphe ehrenbergii (Lév.) U. Braun, M. Bradshaw & S. Takam.	Bradshaw et al. 2021c
Ribes spp.	Sphaerotheca mors-uvae (Schwein.) Berk. & M.A. Curtis)	Parmelee1977	Podosphaera mors-uvae (Schwein.) U. Braun & S. Takam.	Braun and Cook 2012
Ranunculaceae (Aconitum spp., Clematis spp., Delphinium spp.)	Erysiphe polygoni	Parmelee 1977 Amano 1986	Erysiphe aquilegiae	Braun and Cook 2012
Raphanus raphanistrum	Erysiphe cruciferarum	Amano 1986	Erysiphe cruciferarum Opiz ex L. Junell	Braun and Cook 2012
Rudbeckia spp.	Erysiphe cichoracearum	Parmelee1977	Golovinomyces latisporus P.L. Qiu & S.Y. Liu**	Qiu et al. 2020
Zinnia elegans	Erysiphe cichoracearum	Parmelee1977	Golovinomyces ambrosiae (Schwein.) U. Braun & R.T.A. Cook	Qiu et al. 2020

*Braun and Cook (2012) listed a single powdery mildew species on Cosmos, viz., Podosphaera xanthii. Erysiphe cichoracearum on Cosmos spp. has been recorded from numerous countries worldwide (Amano 1986), but this powdery mildew has not yet been phylogenetically proven. The genus Cosmos (tribe Coreopsideae) is allied to Dahlia, so that infections by G. ambrosiae are to be expected.

**Golovinomyces on Rudbeckia agrees morphologically with G. latisporus (see Braun and Cook 2012, as G. ambrosiae). However, the phylogenetic confirmation by sequencing is still pending.

Two incidences reported on Sedum spp. in BC in 2016 (Elmhirst 2017; Joshi and Jeffries 2017).

Notes: Braun and Cook (2012) recorded three species on Sedum, viz Golovinomyces orontii (Castagne) Heluta, Erysiphe sedi and E. umbilici (Kom.) U. Braun & S. Takam. Golovinomyces orontii has been split recently into various plurivorous species (Braun et al. 2019b). Erysiphe sedi can be distinguished from G. orontii s. lat. by the generic characteristics, i.e. Erysiphe asci bear more ascospores (2–)3–8, conidiophores bear a single conidium; whereas Golovinomyces asci bear 2(–4) ascospores, conidia are in catenescent chains. Erysiphe umbilici was only reported from Asia, and differs from E. sedi by the dichotomous branches of appendages. Erysiphe sedi belongs to the Erysiphe aquilegiae complex, composed of numerous species that are insufficiently resolved solely based on ITS data (Takamatsu et al. 2015a; Shin et al. 2019; Bradshaw et al. 2020a). Phylogenetic multilocus approaches are necessary to reach better resolutions in this complex. In the interim, species within this complex can tentatively be identified based on combinations of morphological examinations supplemented by analyses of ITS data (see discussion and recommendation in Bradshaw et al. 2020a).

• (19) Erysiphe symphoricarpi (Howe) Braun and Takamatsu (2000)

Several incidences reported on Symphoricarpos sp. in BC (Elmhirst 2013; Elmhirst and Karlsson 2014).

Notes: Parmelee (1977) reported powdery mildew on Symphoricarpos albus and S. orbiculatus from Canada under Microsphaera diffusa. Erysiphe symphoricarpi morphologically very much resembles to E. diffusa, especially the loose (diffuse) branching patterns of the appendages, convincing Salmon to list it as synonym of E. diffusa (Salmon 1900) although he noted that E. symphoricarpi tends to produce smaller chasmothecia with fewer and longer appendages. Braun and Cook (2012), however, treated them as separate species, and commented that E. diffusa is restricted on the Fabaceae, and has richer branches at appendage apices. The two species are also phylogenetically not closely allied. Sequences retrieved from E. symphoricarpi form a sister clade to sequences of the new species E. flexibilis, and both are closely allied to E. lonicerae (Bradshaw et al. 2021c).

• (20) Erysiphe syringae Schwein. 1832

On lilac (Syringa vulgaris L.) in BC (Joshi and Jeffries 2006, 2018; Elmhirst 2013) and MB (Desjardins 2010; Desjardins et al. 2014). Very common in the national capital region (Ottawa, personal observation)

Notes: According to Braun and Cook (2012), this species is very common in North America and was introduced to South America, Asia and Europe. Morphologically, it is very similar to Erysiphe syringae-japonicae (= Microsphaera syringae-japonicae), a species originally described by Braun (1982) on Syringa amurensis var. japonica in Japan. Erysiphe syringae-japonicae produces evanescent mycelia, 6–8-spored asci and appendages pigmented up to the middle, while E. syringae was considered to produce more persistent mycelia, usually 3–6-spored asci, and appendages pigmented at the base only. However, more morphological variations were found in E. syringae, which can overlap with E. syringae-japonicae (Braun and Cook 2012). The geographic range of E. syringae-japonicae has expanded to Europe since the 1990s (Seko et al. 2008; Mieslerová et al. 2020). Nonetheless, the genetic differentiation was supported by ITS DNA sequence analyses that recovered two haplotypes with only 94% identity, i.e. S-type and K-type corresponding to E. syringae and E. syringae-japonicae, respectively (Seko et al. 2011). Erysiphe syringae-japonicae has not yet been reported from North America, but its introduction is possible and should be kept in mind.

• (21) Erysiphe trifolii Grev. 1824 (Current name: E. trifoliorum (Wallr.) U. Braun)

= Microsphaera trifolii (Grev.) U. Braun 1981

One incidence reported on alfalfa (Medicago sativa L.) in QC (Gilbert et al. 2015).

Notes: As mentioned above, Salmon (1900) treated this species as one of the synonyms of E. polygoni, however, it appeared as a separate species from E. polygoni s. str. based on the phylogenies of rDNA ITS sequences (Takamatsu et al. 2002, 2015a). In a study of powdery mildews on legumes, Braun et al. (2010) examined the original specimen of Alphitomorpha trifoliorum Wallroth, and concluded it is the same species as E. trifolii. Because A. trifoliorum is the oldest valid name for the taxon, he made a new combination, Erysiphe trifoliorum (Wallr.) U. Braun (= E. trifolii Grev.). Reports of E. polygoni on Trifolium hybridum, T. pratense and T. procumbens from Canada in Parmelee (1977) undoubtedly refer to E. trifoliorum. Phylogenetically (solely based on ITS sequences, which provide insufficient resolution), E. trifoliorum constitutes a complex containing various morphologically distinct species with different host ranges, including E. baeumleri and E. hyperici (Bradshaw et al. 2021b).

• (22) Golovinomyces asterum (Schwein.) U. Braun 2012

One incidence on ornamental Aster dumosus L. (current name: Symphyotrichum dumosum (L.) G. L. Nesom) in BC (Elmhirst 2018).

Notes: The name Erysiphe asterum was originated by Schweinitz (1834) to circumscribe species on Aster. Salmon (1900) included E. asterum as a synonym of G. cichoracearum (=E. cichoracearum DC). The broad species concept of G. cichoracearum was applied by Canadian plant pathologists and mycologists (more discussion follows in the paragraph on Golovinomyces cichoracearum). Powdery mildews on Artemisia herriotii Rydb. from AB (McArthur 1966), Aster lindleyanus, A. novae-angliae from MB, and SK (Bisby et al. 1938), Aster sp. from BC (Lowe 1969), Aster spp from NB and NS (Wehmeyer 1950), and a number of species in the Compositae, as summarized by Parmelee (1977), were identified as E. cichoracearum. Blumer (1933) recognized powdery mildew on Aster spp. as one of the 14 formae speciales, E. cichoracearum f. sp. asteris. Braun, in Braun and Cook (2012), considered this taxon a species, applied the name originated by Schweinitz, and made the new combination G. asterum. The phylogenetic position of G. asterum as a well-supported species, forming a clade of its own, was confirmed by Takamatsu et al. (2013).

• (23) Golovinomyces biocellatus (Ehrenb.) Heluta 1988, s. lat.

Several incidences on ornamental Monarda didyma L. in BC (Elmhirst 2017, 2018) and Monarda sp. in ON (Melzer and Shan 2020).

Notes: Golovinomyces biocellatus was circumscribed as a species complex on Lamiaceae (Braun and Cook 2012). Based on rDNA phylogenies and morphology, Scholler et al. (2016) separated the species complex into four species, viz. G. monardae mainly on Monarda, Mentha, Thymus etc., G. neosalviae on Salvia officinalis and S. lavandulifolia, G. salviae on S. pratensis and G. biocellatus s. str. The incidence on M. didyma in BC and Monarda sp. was undoubtedly caused by Golovinomyces monardae (G.S. Nagy) M. Scholler, U. Braun & Anke Schmidt.

• (24) Golovinomyces cichoracearum (DC.) Heluta 1988 s. lat.

Reported on a wide range of crops including vegetables, oil seeds, legumes and ornamentals.

Some incidences of low severity on oil seed sunflowers (Helianthus annuus L.) were reported almost annually from 2000 to 2010 in MB (Rashid et al. 2001, 2002, 2003, 2004, 2005, 2006, 2007; Rashid and Desjardins 2009, 2010, 2011). Vegetables and special crops include cucumber (Cucumis sativus L.) in MB (Pradhan et al. 2017, 2018), NB (Tesfaendrias 2016) and QC (Gilbert et al. 2007), lettuce (Lactuca sativa L.) in BC (Joshi et al. 2015; Joshi and Jeffries 2016) and ON (Tesfaendrias and McDonald 2011, 2013, 2014; Dyk et al. 2015, 2016; Melzer and Shan 2018), lupin (Lupinus angustifolius L.) in AB (Chang et al. 2005a, 2006), melon (Cucumis sp.) in BC (Joshi 2002), mint (Mentha sp.) in ON (Dyk et al. 2016), squash (Cucurbita argyrosperma C. Huber) in BC (Joshi and Berlakoti 2020), NB (Tesfaendrias 2016) and QC (Gilbert et al. 2007, 2008). Ornamentals include Echinacea, Phlox and Polemonium in BC (Joshi 2004; Joshi et al. 2012; Joshi and Jeffries 2016), Coreopsis and Rudbeckia in QC (Gilbert et al. 2008; Breton et al. 2020).

Notes: Erysiphe cichoracearum is a commonly encountered species name in the literature, yet with varied species concepts implied. The specimens based on which de Candolle (1805) described this name were on Scorzonera hispanica and Tragopogon porrifolius (Asteraceae). Salmon (1900) expanded the species concept to include most powdery mildews producing two-spored asci, thus on plants in numerous genera. Parmelee (1977) applied Salmon’s concept when he examined Canadian specimens, and summarized the host genera in 13 families, viz. Begoniaceae, Berberidaceae, Boraginaceae, Compositae (Asteraceae), Geraniaceae, Gesneriaceae, Hydrophyllaceae, Labiateae (Lamiaceae), Plantaginaceae, Polemoniaceae, Rubiaceae, Solanaceae, and Verbenaceae (some of the records could be tentatively re-identified as other species based on recent molecular studies, see Table 1). Blumer (1933) observed varied levels of host specialization among different forms. Based on his own observations and earlier reports by others (Jaczewski 1927), Blumer separated E. cichoracearum into about 10 species (Blumer 1933). Further inoculation tests provided more information on host specialization (Schmitt 1955). Braun (1987) considered E. cichoracearum as a species on Asteraceae only. This separation was supported by RFLP evidence (Zeller and Levy 1995). Molecular evidence indicated that E. cichoracearum sensu Braun (1987) still represents a species complex (Matsuda and Takamatsu 2003). Based on the updated molecular evidence, Braun and Cook (2012) refined the concept of G. cichoracearum s. str. as only on Asteraceae subfam. Cichorioideae (among the listed incidences above, only lettuce belong to this G. cichoracearum sensu Braun and Cook (2012)). However, the Heliantheae clade comprised two morphologically differentiated species, G. ambrosiae and the plurivorous G. spadiceus, which could not be differentiated according to phylogenetic analyses solely based on ITS data. Qiu et al. (2020) clarified this complicated complex with a multilocus approach and demonstrated that G. ambrosiae (syn.: G. spadiceus) is the correct name for the plurivorous species with narrower conidia and typical Euoidium type pattern conidial germination. For the Helianthus powdery mildew with characteristically broad conidia and Euoidium type longitubus pattern of the conidia germination, the new combination Golovinomyces latisporus (U. Braun) P.-L. Qiu & S.-Y. Liu has been introduced. Bradshaw et al. (2021e) conducted host range studies of G. latisporus on a range of Asteraceae species, many of which are native to North America.

On the ITS and 28S phylogenetic trees generated by Matsuda and Takamatsu (2003), samples from Asteraceae, Cucurbitaceae, Lamiaceae, Polemoniaceae etc., formed a heterogeneous group with low statistic support, and five subgroups were recognized within. Subgroup 1 with strong supports on the ITS tree included samples on the Lamiaceae, including Agastache, Monarda, and Salvia, but also Delphinium (Ranunculaceae). This clade corresponds to G. biocellatus (Ehrenb.) Heluta 1988, a species accepted by Braun and Cook (2012). Scholler et al. (2016) further divided G. biocellatus into four species (see above paragraph). Species on Mentha likely belong to G. monardae. Subgroup 4 included samples on Phlox, corresponding to the species G. magnicellulatus (U. Braun) Heluta 1988 (also see Notes in later paragraph).

Golovinomyceous powdery mildews of the Cucurbitaceae were assigned previously to either G. cucurbitacearum (R.Y. Zheng & G.Q. Chen) Vakal. & Kliron. 2001 or G. orontii (Castagne) Heluta 1988 s. lat. Braun and Cook (2012) stated that morphologically G. cucurbitacearum is distinct by its short and narrow conidia, short foot cells, and scanty chasmothecia. Nevertheless, these features are rather inconsistent, so that it is often morphologically difficult to differentiate between G. cucurbitacearum and G. orontii, a species that was often confused with G. cichoracearum. The phylogeny and taxonomy of Golovinomyces on cucurbits has recently been clarified in comprehensive phylogenetic-taxonomic examinations of the G. orontii complex (Braun et al. 2019b; Qiu et al. 2020). Additional discussion regarding G. orontii s. lat. is presented in later paragraphs.

Lupin (Lupinus angustifolius, Fabaceae) was not listed as a potential host even under E. cichoracearum sensu Salmon. Therefore, this record is doubtful and in urgent need of confirmation by means of sequencing. An accidental infection caused by one of the species of the G. orontii s. lat. complex would be conceivable. According to Parmelee (1977), two species are likely accounted for powdery mildews on lupin, i.e. E. pisi and E. trifolii (current name: E. trifoliorum). In a phylogenetic-taxonomic revision of Erysiphe on lupins, Bradshaw et al. (2021b) confirmed three powdery mildew species on Lupinus spp. in North America, viz., Erysiphe diffusa s. lat., E. intermedia (U. Braun) U. Braun 2010, and the new species E. lupini M. Bradshaw 2021. Erysiphe trifoliorum is the prevailing Erysiphe species on lupins in Europe.

• (25) Golovinomyces lycopersici (Cooke & Massee) L. Kiss 2019

= Oidium lycopersici Cooke & Massee 1888

One incidence reported on tomato (Solanum lycopersicum L.) in QC (Gilbert et al. 2015).

Notes: This report was undoubtedly incorrect. As noted by Braun et al. (2019b), Oidium lycopersici was a misapplied name, until Kiss et al. (2001) clarified and neotypified it as a species inhabited only in Australia. Several studies on molecular phylogenetics, morphology, and pathogenesis later indicated that it and a closely related species (O. longipes) belong to Golovinomyces (Kiss et al. 2001, 2008; Cunnington et al. 2005; Braun 2013). The new combinations were made: G. longipes (Noordel. & Loer.) L. Kiss, and G. lycopersici (Cooke & Massee) L. Kiss 2019. Golovinomyces longipes occurs on diverse Solanaceous hosts: Nicotiana sp., Petunia ×hybrida, several Solanum spp. (Kiss et al. 2008; Braun and Cook 2012; Toome et al. 2015), and Calibrachoa hybrids (Brielmaier-Liebetanz et al. 2015), but it was also recorded from Germany under glasshouse conditions on Verbena ×hybrida, Verbenaceae (Brielmaier-Liebetanz et al. 2015) and on Matricaria chamomilla, Asteraceae (Götz and Braun 2021). The main host of G. lycopersici is Solanum lycopersicum, but it could also infect Cyphomandra, Nicotiana, and Petunia (Cunnington et al. 2005) in Australia. The powdery mildew of tomato in other parts of the world was attributed to another species, Pseudoidium neolycopersici (discussed later). Gilbert et al. (2015) listed the Canadian tomato powdery mildew under the name Oidium lycopersici without any further information and description. It is very unlikely that this Australian powdery mildew was the causative agent of the Canadian tomato powdery mildew. This infection has rather been caused by Golovinomyces longipes, which is not uncommon on tomato in Europe and has also been found in the USA on petunia (Braun and Cook 2012). Golovinomyces longipes and G. lycopersici are morphologically easily distinguishable (Braun and Cook 2012), although ITS sequences cannot separate them (Takamatsu et al. 2013). The possibility of Pseudoidium neolycopersici, however, also cannot be excluded.

• (26) Golovinomyces magnicellulatus (U. Braun) Heluta 1988

One incidence reported on Phlox sp. in ON (Melzer and Shan 2020).

Notes: Parmelee (1977) reported Erysiphe cichoracearum s. lat. on Phlox divaricata, P. drummondii, and P. paniculata from Canada. In general, G. magnicellultus resembles G. cichoracearum. Braun (1978) treated it as a separate species based on the observation that it produced mostly larger chasmothecia with evidently larger peridium cells, and the biological specialization on Phlox demonstrated by previous inoculation experiments (Schmitt 1955; Schüepep and Blumer 1963). A number of species of Polemonium were also recorded as hosts (Braun and Cook 2012). In the Canadian Host-Pathogen Database (http://hpdb.biodiversity.agr.gc.ca/hostPathogens, as accessed on march 24, 2021), powdery mildews on Phlox were all recorded as E. cichoracearum (s. lat.). However, G. cichoracearum s. str. is confined to powdery mildew on Scorzonera and Tragopogon spp. Sequences obtained from powdery mildew on these hosts formed a small clade within lineage XI in Takamatsu et al. (2013), whereas sequences retrieved from G. magnicellulatus on Phlox spp. from Asian, European and North America formed a separate clade (lineage X, in Takamatsu et al. 2013), which genetically confirmed the Golovinomyces on Phlox as a species of its own.

• (27) Golovinomyces orontii (Castagne) Heluta 1988

= Erysiphe orontii Castagne. 1851

Incidences reported on tomato (Solanum lycopersicum L.) in QC (Gilbert et al. 2007, 2008, 2010, 2011).

Notes: The pre-molecular concepts of Golovinomyces orontii (= E. orontii) had a wide host range and overlapped with G. cichoracearum s. lat. Oidium lini Škorič 1926 and O. lini Bondartsev 1913 were previously cited as synonyms of G. orontii (Braun and Cook 2012), which is, however, incorrect (Oidium lini Škorič is a synonym of Podosphaera lini, and O. lini Bondartsev is an unpublished name, see Braun et al. 2019a). Braun et al. (2019a) clarified the delimitation of these two species in the context of rDNA phylogenetic trees, morphological characteristics and taxonomic recount. As G. cichoracearum s. str. is now confined to Golovinomyces on Scorzonera and Tragopogon spp. (Takamatsu et al. 2013), G. orontii s. str. was recognized as a monophyletic group on ITS and 28S trees including the sample on the type host, Misopates orontium (Braun et al. 2019b). Besides M. orontium (Plantaginaceae), the hosts of G. orontii s. str. also include Arabidopsis, Capsella, Camelina, Neslia (Brassicaceae), Cucurbita (Cucurbitaceae), Papaver (Papaveraceae), Penstemon (Scrophulariaceae), Valerianella, Valeriana (Valerianaceae), and Viola (Violaceae). The powdery mildews on Solanum reported from QC could be Pseudoidium neolycopersici (see the Notes for Golovinomyces lycopersici and Pseudoidium neolycopersici), or G. longipes.

• (28) Golovinomyces salviae (Jacz.) M. Scholler, U. Braun & Anke Schmidt 2016

One incidence reported on sage (Salvia officinalis L.) in ON (Melzer and Shan 2020).

Notes: According to Scholler et al. (2016), the species on S. officinalis likely belongs to G. neosalviae (also see the notes for G. biocellatus).

• (29) Leveillula taurica (Lév.) G. Arnaud (1921)

= Oidiopsis sicula Scalia 1902

Several incidences reported on pepper (Capsicum sp.) in BC and ON (Joshi and Jeffries 2007; Melzer and Shan 2017).

Notes: Species in Leveillula can be readily identified by conidiophores arising through plant stomata from internal hyphae, large dimorphic conidia produced singly, and large chasmothecia with myceloid appendages. However, the distinction between species can be very challenging. Arnaud (1921) first introduced the genus Leveillula with only one species, L. taurica, previously described as Erysiphe taurica by Léveillé (1851). In Salmon’s monograph (1900), E. taurica was another highly variable species (like E. polygoni and E. cichoracearum) with a long list of synonyms and a wide range of hosts. In fact, the heterogeneous nature of E. taurica was evident in Léveillé’s first description that listed nine genera in six families as hosts. Since the late 1970s, efforts were made to separate E. taurica s. lat. into more coherent species using the morphology of conidia including detailed surface structures by scanning electron microscopy, host specificity and DNA sequence analyses (Braun 1980; Durrieu and Rostam 1984 [1985]; Geljuta and Simonjan 1988; Khodaparast et al. 2007). In a recent monograph, Braun and Cook (2012) accepted 40 Leveillula species, among which 26 were connected with L. taurica s. l. by having at least one form (or forma specialis) of L. taurica listed as synonyms. Leveillula taurica s. str. was circumscribed as having narrowly lanceolate primary conidia. Yet with a wide host range of 14 families, it was still suspected to be a species complex that needs to be further resolved (Braun and Cook 2012). This hypothesis was later supported by the analyses of rDNA ITS sequences with extended samples that showed high levels of intraspecific variation (Khodaparast et al. 2012). Nevertheless, Capsicum was recorded as the host of L. taurica s. str., and the morphology of Leveillula from Capsicum is consistent with the current concept of this species.

• (30) Neoerysiphe hiratae Heluta & S. Takam. 2010

Reported on tomato (Solanum lycopersicum L.) in QC (Breton et al. 2018, 2019).

Notes: Neoerysiphe hiratae is a species described from Japan by Heluta et al. (2010), while they studied Eurasian Neoerysiphe spp. on Asteraceae in comparison with the North American species N. cumminsiana. Considering the host and geographic range of N. hiratae, it is surprising to find it reported on tomato in Quebec. In any case, the Canadian collection is in need of molecular confirmation.

• (31) Pseudoidium neolycopersici (L. Kiss) L. Kiss 2012

= Oidium neolycopersici L. Kiss 2001

Reported on tomato (Solanum lycopersicum L.) recurrently in QC (Gilbert et al. 2012, 2013, 2014, 2015; Breton et al. 2018, 2019, 2020), also in NB (Tesfaendrias 2016, 2017) and ON (Sabourin et al. 2006; Melzer and Shan 2020).

Notes: Kiss et al. (2001) identified two taxa causing powdery mildew diseases on tomato worldwide. Oidium neolycopersici was distinguished from O. lycopersici, by producing non-catenate conidia and having a global distribution, while O. lycopersici produces catenate conidia and has so far only been found in Australia. Based on several morphological features, i.e. conidia produced singly (or with false chains), germ tubes longitubus, alobate or weakly lobate, it was reclassified as Pseudoidium neolycopersici. Phylogenetically, this species pertains to the Erysiphe aquilegiae cluster (Takamatsu et al. 2015a), which is not sufficiently resolved at the species level in phylogenetic analyses solely based on ITS data (Bradshaw et al. 2020a). The sexual morph of this species is still unknown, and the position of this anamorphic powdery mildew as a species of its own has not yet been confirmed by a phylogenetic multi-locus approach. Therefore, this name has not yet been formally transferred to Erysiphe.

• (32) Phyllactinia guttata (Wallr.) Lév. s. str. 1851

Reported on Hazelnuts (Corylus sp.) in recent years in ON (Melzer and Shan 2015, 2016, 2017).

Notes: Phyllactinia guttata is another one of the oldest powdery mildew names with controversial species concepts. It can be traced back to Erysiphe coryli and E. fraxini coined by de Candolle (1805) for powdery mildews on coudrier (French for Corylus) and ash (Fraxinus). Wallroth (1819b) included E. coryli DC., E. fraxini DC., E. suffulta Rebent and several other names as synonyms of Alphitomorpha guttata, a new name he generated and which was an illegitimate name (nom. illegit. [superfl.], Art. 52.1) when introduced. Link (1824) applied Wallroth’s names and made the combination of E. guttata, and accepted three varieties E. guttata var. coryli, var. fraxini, var. ulmorum. Fries (1829) accepted Erysiphe guttata, so that this name is now, according to the current Code (ICN), a sanctioned name (Art. F.3.1 and F.3.7), i.e. this name has a status like a conserved name. When Léveillé (1851) erected genus Phyllactinia, he accepted three species Ph. guttata, Ph. candollei, and Ph. schweinitzii. Phyllactinia guttata was circumscribed as a species with appendiculate appendages with a bulbose swelling at the base and two-spored asci. Salmon (1900) accepted only one name in Phyllactinia, viz., Ph. corylea. Blumer (1933) accepted 10 species, however, neither Ph. guttata nor Ph. corylea were on the list, and instead were considered as synonyms of several different species he accepted. Homma (1937) recognized nine species in Japan and Ph. guttata was considered a synonym of Ph. corylea and Ph. fraxini. Chen et al. (1987) recognized 26 species to replace the previous concept of Ph. corylea, Ph. guttata and Ph. suffulta. Braun (1987) designated ‘Erysiphe coryli Hedw. fl. fr. 730’, ex herb. De Candolle (G) as lectotype of Ph. guttata, and circumscribed it as a species on a wide range of hosts (50 families). This lectotypification determined the application of the name Ph. guttata (type host = Corylus avellana), and the status as sanctioned name rules its priority. Ph. corylea is a heterotypic synonym and Erysiphe coryli is a homotypic synonym of Ph. guttata. A recent phylogenetic study by Takamatsu et al. (2008b) proposed eight species based on the rDNA phylogeny of 98 sequences of Phyllactinia on various hosts. The samples on Corylus spp. formed a clade (1A) with 20 other samples from over 10 host families, which was considered as Ph. guttata. However, these first phylogenetic analyses revealed that the broad concept of Ph. guttata was not tenable any longer and that this species in its previously very broad circumscription had to be divided into numerous species with narrower concepts. Based on these new insights, Braun and Cook (2012) re-examined the morphology of the whole Ph. guttata complex and found obvious differences between Phyllactinia on hosts of different plant families, above all in the previously little examined asexual morphs and details of the penicillate cells of the chasmothecia. As one of the results of this re-examination, Ph. guttata (s. str.) was confined to Phyllactinia on Corylus spp. The Canadian Phyllactinia collections can currently be assigned to Ph. guttata only tentatively. Morphological re-examinations and confirmations by sequence analyses of Canadian specimens are necessary.

• (33) Podosphaera aphanis (Wallr.) Braun and Takamatsu (2000)

Repeatedly reported in BC on ornamentals, Geum macrophyllum Willd., Physocarpus opulifolius (L.) Maxim., Potentilla fruticosa L. (current name: Dasiphora fruticosa (L.) Rydb. subsp. fruticosa), Rubus spectabilis Pursh, and straw berries (Fragaria spp.) in BC (Elmhirst 2013; Elmhirst and Karlsson 2014, 2016), and in AB (Zuzak et al. 2018).

Notes: Wallroth (1819a) described a species on Aphanes and Alchemilla (Rosaceae) as Alphitomorha aphanis. Braun (1982) moved it to Sphaerotheca, a genus erected by Leveille to accommodate species producing one ascus per chasmothcium with myceloid appendages (Léveillé 1851), becoming Sphaerotheca aphanis. With the evidence of rDNA phylogeny that Podosphaera and Sphaerotheca were intermingled on the phylogenetic tree, Braun and Takamatsu (2000) reduced Sphaerotheca as a section, Podosphaera sect. Sphaerotheca. Consequently, a new combination Podosphaera aphanis was made. This species has an ambiguous separation with Podosphaera macularis and varied species concepts exist in the current literature. Braun and Cook (2012) applied a strict species concept of P. macularis (s. str.) considering it a species restricted on Humulus, while P. aphanis was the species on a wide range of hosts including strawberries and blackberries. Three varieties: P. aphanis var. aphanis, P. aphanis var. hyalina, and P. aphanis var. physocarpi were recognized (Braun and Takamatsu 2000). In Canada, a broad species concept of P. macularis (s. lat.) was largely applied (see more discussion later under Podosphaera macularis). The name P. aphanis appeared in the literature only in recent years, as previously species on strawberries and blackberries were recorded as P. macularis. The phylogenetic analyses using ITS sequences of selected samples from DAOM and from GenBank showed that both species are closely allied, with insufficient resolution in analyses solely based on ITS data (personal observation). Furthermore, the entire P. aphanis/macularis complex is in urgent need of a phylogenetic/taxonomic revision. Podosphaera aphanis as hitherto circumscribed by Braun and Cook (2012) is, in any case, a complex of cryptic species. For the time being, all samples on berries should currently be assigned to P. aphanis (s. lat.). The Podosphaera on Physocarpus opulifolius has recently been phylogenetically revised and confirmed to be a species of its own, P. physocarpi (U. Braun) U. Braun 2012, clustering in a sister position to P. amelanchieris Maurizio (1927) (Bradshaw et al. 2020a).

• (34) Podosphaera clandestina (Wallr.) Lév. s. lat. 1851

Repeatedly reported on small fruit and ornamental plant service berry/saskatoon (Amelanchier alnifolia (Nutt.) Nutt. ex M. Roem.) in BC (Elmhirst 2013; Elmhirst and Karlsson 2014), MB (Desjardins et al. 2002, 2003, 2005; Desjardins and Bisht 2011) and ON (Melzer and Shan 2018); on cherry (Prunus avium (L.) L.) in QC (Gilbert et al. 2007).

Notes: In Wallroth’s (1819b) description of Alphitomorha clandestina, Erysiphe oxyacanthae DC. was listed as a synonym, therefore, Braun (1987) designated a lectotype using ‘Erysiphe oxyacanthae, belong des routes pres Baden’ ex herb. De Candolle (G), on Crataegus sp. (cf. oxyacanthae). Noted in Braun and Cook (2012), however, Maurizio (1927) treated North American Po. clandestina s. lat. as a different species, viz., Po. amelanchieris, based on the observation that Po. amelanchieris (the North American species) produces numerous and very long appendages. This observation and treatment were agreed upon by Blumer (1967) and Braun and Cook (2012). Up-to-date molecular phylogenetic analyses (Takamatsu et al. 2010) also supported this classification in that North American Po. clandestina s. lat. samples on Prunus did not group with Po. clandestina s. str. on Crataegus, but clustered with an Asian sample on Spiraea as a separate clade. Recently, Bradshaw et al. (2020a) confirmed the status of Po. amelanchieris as a species of its own, phylogenetically distant from Po. clandestina s. str. on Crataegus. European and North American sequences retrieved from powdery mildew on Amelanchier spp. formed a single clade as sister to Po. physocarpi. According to this classification, the incidences on Amelanchier in BC, MB and ON were very likely Po. amelanchieris. The report of Po. clandestina on Prunus avium in QC very likely belongs to Po. cerasi Moparthi et al. (Moparthi et al. 2019). The latter work provided a phylogenetic-taxonomic revision of Podosphaera on Prunus spp. in North America. However, in-depth morphological and molecular investigations based on Canadian collections of Po. cladestina s. lat. are needed for further clarification. Besides Amelancier spp. and Crataegus sp., Parmelee (1977) listed Prunus cerasus, Pr. pumila, Pr. nigra, Pr. virginiana, Spiraea alba, S. latifolia and S. tomentosa as hosts of Po. clandestina s. lat. Confirmation of the true identities of the powdery mildews on these hosts also requires phylogenetic-morphological re-examination.

• (35) Podosphaera euphorbiae (Castagne) Braun and Takamatsu (2000)

One incidence reported on ornamental Euphorbia sp. in BC (Elmhirst and Karlsson 2014).

Notes: Braun and Cook (2012) noted the holotype of Podosphaera euphorbiae was a specimen on Euphorbia peplus collected in France, and the recorded geographic regions were Asia, and Europe. Two other similar species on Euphorbia are Po. euphorbiae-helioscopiae (Tanda & Y. Nomura) U. Braun & S. Takam., and Po. euphorbiae-hirtae (U. Braun & Somani) U. Braun & S. Takam. The former was distinct by the significantly long, wide and thick-walled appendages, and recorded in Japan only, whereas the latter was reported in various locations including the Caribbean island of Puerto Rico. It was suggested that most of the incidences on Euphorbia likely belong to Po. euphorbiae-hirtae. However, with the analyses of ITS sequences, the sequences of Po. (= Sphaerotheca) euphorbiae-hirtae belonged to the same haplotype with samples on Compositae, and placed in the clade corresponding to Po. xanthii. Braun et al. (2001) proposed to reduce it to a synonym of Po. xanthii.

• (36) Podosphaera fuliginea (Schltdl.) Braun and Takamatsu (2000)

= Sphaerotheca fuliginea (Schltdl.) Pollacci 1911

Repeatedly reported on cucumber (Cucumis sativus L.) in ON (Sabourin et al. 2007; Shan and Sabourin 2014a; Melzer and Shan 2018) and QC (Gilbert et al. 2014), on squash (Cucurbita argyrosperma C. Huber) in BC (Joshi and Jeffries 2006), ON (Tesfaendrias and McDonald 2013) and QC (Gilbert et al. 2009, 2010), on zucchini (Cucurbita pepo L.) in BC (Joshi et al. 2014; Joshi and Jeffries 2016) and ON (Shan and Sabourin 2014b), on watermelon (Citrullus lanatus (Thunb.) Matsum. & Nakai) in BC (Joshi and Jeffries 2016), on ornamental plant Kalanchoe blossfeldiana Poelln. in BC (Joshi 2019), and Veronica sp. (Breton et al. 2018, 2020).

Notes: When Salmon (1900) revised Sphaerotheca humuli as a species on a wide range of hosts, he recognized a variety Spherotheca humuli (DC) Burr. var. fuliginea based on the observation that the chasmothecial peridium cells of this variety were larger. Blumer (1933) retained Sphaerotheca humuli (DC) Burr. var. fuliginea (Schlcht.) Salm. (in part) as a species Sph. fuliginea, and recorded the host in 35 genera from over 10 families, i.e. Compositae (alt. Asteraceae), Dipsaceae (current Caprifoliaceae subfam. Dipsacoideae), Cistaceae, Cruciferae (alt. Brassicaceae), Cucurbitaceae, Papilionaceae (current Fabaceae), Plantaginaceae, Ranunculaceae, Scrophulariaceae (the genera listed under this family currently have been classified to Plantaginaceae or Orobanchaceae) and Solanaceae. According to Braun et al. (2001), Junell (1966) was the first who separated Po. (=Sphaerotheca) fuliginea into several smaller units. Following Junell’s (1966) concept, Braun (1987) circumscribed Po. fuliginea as occurring only on species of Veronica, which was supported later by the ITS phylogeny, showing the taxon on Veronica was evidently separated from Po. fusca and P. xanthii (Hirata et al. 2000; Braun et al. 2001; Braun and Cook 2012). With the refined species concept based on combined molecular and morphological evidence, the species on Cucurbitaceae likely does not belong to Po. fuliginea, but Po. xanthii (Hirata et al. 2000; Braun et al. 2001); see more discussion in the notes for Po. xanthii).

• (37) Podosphaera leucotricha (Ellis & Everh.) E.S. Salmon (1900)

Reported on apple (Malus domestica (Suckow) Borkh.) repeatedly in recent years in NB (Tesfaendrias 2016, 2019, 2020), one incidence in ON (Ker and McFadden-Smith 2002b), one in QC (Breton et al. 2019), occasionally reported on the native Pacific Northwest apple species, Malus fusca (Raf.) C. K. Schneid. in BC (Elmhirst and Karlsson 2016).

Notes: This species was recorded as producing matted patches on apple branches, fruits, petioles or rarely on leaves and described by Ellis and Everhart (1888) as Sphaerotheca leucotricha. Burrill (1892) changed the name to S. mali considering it as the same species as Erysiphe mali by Duby. However, Salmon (1900) pointed out that E. mali Duby was in fact Phyllactinia corylea (now Ph. mali (Duby) U. Braun 1978). After examination of the specimen sent by Prof. Ellis, Salmon (1900) considered the occasionally branched appendages of the present species was different from other Sphaerotheca spp., and therefore moved it to Podosphaera with the recorded hosts including Malus and Pyrus. This species was first known in North America, then reported from Europe, some under the name S. mali (Magnus 1898; Grout 1899; Salmon 1900), and Australia (Cobb 1892). Sholberg et al. (2004) developed PCR and hybridization assays and proved these assays were useful for monitoring this pathogen.

• (38) Podosphaera macrospora (U. Braun) U. Braun & V. Kumm. 2008

One incidence on forage plant Lotus corniculatus L. in QC (Breton et al. 2018).

Notes: Podosphaera macrospora was elevated from Po. alpina f. macrospora U. Braun based on the observation that the samples on Saxifraga tricuspidata from Canada had significantly larger ascospores in thin-walled asci, in contrast with the European Po. alpina [U. Braun & V. Kummer in (Ale-Agha et al. 2008)]. Podosphaera macrospora was circumscribed as a species on a wide range of host in family Saxifragaceae, whereas the European Po. alpina was limited to Saxifraga spp., viz. S. nivalis, and S. rotundifolia (Braun and Cook 2012). The incidence of Po. macrospora on Lotus corniculatus (French: lotier; fam. Fabaceae) in QC is unexpected and doubtful, and it very likely belongs to another Podosphaera species. It is urgently necessary to re-examine the specimens on Lotus corniculatus morphologically and phylogenetically in the context of molecular analyses of specimens of Po. macrospora and similar Podosphaera spp. on legumes, such as Po. astragali (L. Junell) Braun and Takamatsu (2000) and the plurivorous Po. xanthii (Castagne) U. Braun & Shishkoff 2000 (syn.: Po. phaseoli (Z.Y. Zhao) Braun and Takamatsu 2000). However, no reliable sequences from the country of origin on the type host are available.

• (39) Podosphaera macularis (Wallr.) Braun and Takamatsu (2000)

= Sphaerotheca macularis (Wallr.) Magnus 1899

On strawberry annually reported in QC in last 10 years (Gilbert et al. 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015; Breton et al. 2018, 2019), also repeatedly reported in BC (Joshi 2002; Joshi and Jeffries 2006, 2016, 2017; Joshi et al. 2015), NB (Tesfaendrias 2016, 2017, 2018, 2019, 2020), ON (Melzer and Shan 2018, 2019) and PEI (Clark and Driscoll 2017; Clark and MacLeod 2018), one incidence from MB (Desjardins et al. 2001). Less frequently reported on raspberry (Rubus sp.) or blackcurrant (Ribes nigrum L.) in BC (Joshi and Jeffries 2016), MB (Desjardins et al. 2012), QC (Gilbert et al. 2016), and SK (Northover et al. 2011). On hops (Humulus lupulus L.) in BC (Joshi and Jeffries 2017, 2018; Joshi 2019). On Malus domestica and Physocartpus sp. in QC (Gilbert et al. 2009).

Notes: The powdery mildew on hops was first described by de Candolle (1815) as Erysiphe humuli. In 1819, Wallroth introduced Alphitomorpha macularis, a species described on Humulus and Epilobium (Onograceae), and cited E. humuli as a synonym. Although Wallroth’s name was an illegitimate, superfluous name (Art. 52.1) when introduced, Fries (1829) accepted it and introduced the combination Erysiphe macularis, which is, therefore, a sanctioned name, with a nomenclatural status like conserved names (Art. F.3.1 and F.3.7). Blumer (1933) separated species on Humulus and Epilobium again, and accepted three species, viz. Sphaerotheca epilobii (Wallr.) Sacc. on Epilobium, S. humuli (DC) Burrill on Humulus, and S. macularis on various genera in Rosaceae including Comarum, Fragaria, Geum, Potentilla, and Rubus. In recent monographs, Braun (1987) and Braun and Cook (2012) accepted Podosphaera aphanis as the species on Rosaceae (see more discussion in an earlier paragraph), and Podosphaera macularis on Humulus (= Sphaerotheca macularis) and listed S. humuli as a synonym of P. macularis (Braun 1987; Braun and Cook 2012). The application of the name Po. macularis was determined in Braun (1987) via lectotypification with a collection in de Candolle’s herbarium (G) on Humulus lupulus deposited as Erysiphe humuli.

In North America, despite the fact that Salmon (1900) included Erysiphe macularis as one of the over 50 synonyms of E. polygoni, a broad species concept of Podophaera macularis (= Sphaerotheca macularis) was commonly applied. Parmelee (1977) listed host genera including Rhus (Anacardiaceae), Humulus (Cannabinaceae), Brassica (Cruciferae), Geranium (Geraniaceae), Agrimonia, Filipendula, Fragaria, Geum, Physocarpus, Potentilla, and Rubus (Rosaceae). The Canadian Host-Pathogen Database recorded hosts in 29 genera of 15 families. A phylogenetic analysis using ITS sequences of selected P. macularis samples from DAOM and from GenBank showed that samples on Humulus were intermingled with some samples on Rosaceae (personal observation), which does not support the concept that powdery mildews on Humulus belong to one coherent species with a restricted host range. Instead, the hop powdery mildew can occur on plants in Rosaceae. However, embedded in the paraphyly of Humulus, the samples on berries (strawberries and blackberries) and plants of other families formed a monophyletic group, corresponding to P. aphanis (personal observation). In addition, the samples on Brassica were distantly related with other samples, indicating they belong to another species, possibly E. cruciferarum. Confirmation with molecular and morphological evidence is needed.

• (40) Podosphaera mors-uvae (Schwein.) Braun and Takamatsu (2000)

= Sphaerotheca mors-uvae (Schwein.) Berk. & M.A. Curtis

Several incidences on blackcurrant (Ribes nigrum) in AB (Chang et al. 2003; Howard et al. 2005), MB (Desjardins et al. 2007), and QC (Gilbert et al. 2007, 2008; Breton et al. 2019). On ornamental Ribes spp. in BC (Elmhirst 2012; Elmhirst and Karlsson 2014).

Notes: Schweinitz (1834) described Erysiphe mors-uvae on ‘Grossulariae uvae crispae’. Currently, species of Grossularia were reclassified in Ribes L. Other species reported on Ribes spp. in North America include E. grossuariae (Wallr.) de Bary (Shaw 1973; French 1989; Braun and Cook 2012), Leveillula taurica (Lév.) G. Arnaud. (Braun and Cook 2012), Phyllactinia ribis (Jacz.) Z.Q. Zhao (Braun and Cook 2012), Ph. guttata s. lat. (Shaw 1973; French 1989), Sphaerotheca macularis (Cash 1953) and S. fuliginea (Gardner et al. 1970).

• (41) Podosphaera pannosa (Wallr.) de Bary 1870

= Sphaerotheca pannosa (Wallr.) Lév. 1851

On rose (Rosa spp.) repeatedly reported in BC (Elmhirst 2012, 2013, 2017; Elmhirst and Karlsson 2014, 2016; Joshi et al. 2014, 2015; Joshi and Jeffries 2017), occasionally in MB (Desjardins et al. 2003) and NB (Tesfaendrias 2017). A few incidences on peach (Prunus persica (L.) Batsch) in ON (Ker and McFadden-Smith 2002b; Sabourin et al. 2007).

Notes: Other species reported on Rosa in North America include Phyllactinia guttata s. lat. (Shaw 1973), Sphaerotheca fuliginea (Gardner et al. 1970) and S. macularis (Shaw 1973). Braun and Cook (2012) included several species on Rosa in Asia, i.e. Erysiphe simulans (E.S. Salmon) U. Braun & S. Takam., E. rosae (Golovin & Gamalitsk.) U. Braun & S. Takam. and E. karisiana U. Braun & S. Takam.

In addition to Podosphaera pannosa, other species found on Prunus in North America include Golovinomyces orontii, Phyllactinia mali, Po. leucotricha, Po. prunicola and Po. tridactyla. Podosphaera pannosa can be distinguished from the other three Podosphaera spp. by the morphology of the appendages. Podosphaera prunicola and Po. tridactyla have dichotomously branched appendage apices, Po. leucotricha has mostly setiform appendages with unbranched or sometime bifurcate apices, which are all different from Po. pannosa with sinuous mycelium-like unbranched appendages. However, the differences between Po. prunicola and Po. tridactyla are very subtle (see more in the following notes).

• (42) Podosphaera tridactyla (Wallr.) de Bary 1870

One incidence on ornamental Prunus sp. in BC (Joshi and Jeffries 2017).

Notes: Podosphaera tridactyla and Po. prunicola can be found on different Prunus spp, however they share some host species as well, i.e. P. armeniana, P. cerasus, P. domestica, P. persica, etc. The morphological differences between them, as described in Braun and Cook (2012), are very subtle, but easily discernable when carefully observed. Po. tridactyla produces fewer appendages, (1–)2–6(–9), compared with 6–20(–25) in Po. prunicola, which arise from the apex of the chasmothecium in a loose fascicle (vs. more or less equatorially arising from the chasmothecium, i.e. around the middle). The appendages of Po. tridactyla are unequal in length (1–6 times of chasmothecium diam.), and the primary branch of apices are often elongated compared with those of Po. prunicola, which are 1–3.5 times the chasmothecium and tightly dichotomously branched. Braun and Cook (2012) hypothesized that Po. tridactyla is likely a species complex due to its highly variable morphological features. This hypothesis was supported by the ITS phylogenetic tree, on which the samples of Po. tridactyla were grouped into 2–4 clades, representing several separate phylogenetic species (Takamatsu et al. 2010). A comprehensive phylogenetic-taxonomic revision of Po. tridactyla s. lat., carried out by Meeboon et al. (2020), confirmed this assumption and led to a splitting of Po. tridactyla s. lat. into numerous species. However, North American specimens of Po. tridactyla s. lat. were underrepresented in these analyses, i.e. collections from Canada and the USA are in urgent need of a phylogenetic re-examination.

• (43) Podosphaera xanthii (Castagne) U. Braun & Shishkoff 2000

Occasional incidences on cucumber (Cucums sativus) and squash (Cucurbita argyrosperma) in NB (Tesfaendrias 2016, 2020), on Zucchini (Cucurbita pepo) in AB (Zhou et al. 2020) and QC (Breton et al. 2019).

Notes: Historically, the species boundaries among Po. fuliginea, Po. fusca and Po. xanthii were not very clear. The species Erysiphe xanthii was described on Xanthium spinosum (Castagne 1845). Blumer (1933) considered E. xanthii as a synonym of Sphaerotheca fuliginea on a wide range of hosts. Braun (1987), however, treated Sph. fuliginea as a species restricted on Veronica (as noted earlier), and included E. xanthii as one of the numerous synonyms of Sph. fusca. Braun and Takamatsu (2000) commented that the features found by Shishkoff, i.e. the size of the ascomata and particularly the diameter of oculi (the thin-walled apical part of asci), were informative for the taxonomy of Sph. fusca s. lat. and allied species, and divided Sph. fusca sensu Braun (1987) into two species based on these features: Po. (Sphaerotheca) xanthii is the species with larger ascomata and broader ascus oculi in contrast with Po. (Sph.) fusca with smaller ascomata and narrower ascus oculi. This classification aligned well with the results of ITS rDNA sequence phylogenetic analyses by Hirata et al. (2000). Species on Cucumis, Cucurbita, Gynostemma, Melothria and Trichosanthes (Cucurbitaceae) shared a haplotype with other samples on Helianthus, Zinnia (Asteraceae), Saintpaulia (Gesneriaceae), Lycopus (Lamiaceae), Verbena (Verbanaceae), Crotalaria, Glycine, Vigna (Fabaceae), Abelmoschus (Malvaceae) and placed in a clade (group III), which was recognized as Po. xanthii (Hirata et al. 2000; Braun et al. 2001). In any case, Po. xanthii is proven a plurivorous species with wide host range and distribution (Meeboon et al. 2016; Yeh et al. 2021). In the course of a comprehensive re-examination of powdery mildews worldwide, in connection with the preparation of a new powdery mildew monograph, Braun and Cook (2012) further split the old ‘Po. fusca’ (s. lat.) and confined the genuine Po. fusca (s. str.) to collections on species of its type genus (Doronicum) in Europe, based on obvious morphological differences compared to other species of the Po. xanthii complex. Collections on the Compositae with small oculi of the asci, previously also subsumed under Po. (Sph.) fusca are now placed in Po. erigerontis-canadensis (Lév.) U. Braun & T.Z. Liu 2010, which is also a phylogenetically confirmed species. Morphologically similar collections on hosts of the Orobanchaceae and Scrophulariaceae are referred to as Podosphaera phtheirospermi (Henn. & Shirai) U. Braun & T.Z. Liu 2010 (= Sphaerotheca melampyri L. Junell). However, it has to be emphasized that hosts of the latter families can also be infested by Po. xanthii (Yeh et al. 2021).

• (44) Sawadaea bicornis (Wallr.) Homma (1937)

Repeatedly reported on maple (Acer macrophyllum Pursh) in BC (Elmhirst 2013; Elmhirst and Karlsson 2014, 2016).

Notes: According to Homma (1937), the maple powdery mildew was separated from Uncinula and included in the new genus Sawadaea by Miyabe in Sawada 1914. Homma (1937) noted the key teleomorphic characteristic that separates Sawadaea from Uncinula is that a variant portion of appendages was branched (either dichotomously or trichotomously) in Sawadaea. The feature was also used to differentiate the three species recognized in Japan, i.e. S. bicornis, S. tulasnei, and S. negundinis. Homma considered that the appendages of S. bicornis were almost always branched, whereas only 2/3 of them were branched for S. tulasnei and S. negundinis. Sawadaea negundinis was reduced as a synonym of S. bicornis (Braun 1995), but later recognized as a genetically and morphologically differentiated species of its own (Braun and Cook 2012). The confusion was caused by the occurrence of two species on Acer negundo, viz., S. bicornis and S. negundinis. In addition, there is a unique asexual morph that differentiates Sawadaea from all other powdery mildew genera, namely, the formation of dimorphic conidiophores and conidia, differentiated in macro- and micro-conidiophores and conidia (Braun 1987; Braun and Cook 2012). On the ITS tree with extended samples on various host species and from different geographic regions, S. bicronis, S. polyfida and S. tulasnei appeared as clearly separated species, and the subdivisions in the clade of S. bicornis and S. tulasnei suggested these two might still be species complexes (Hirose et al. 2005). In a recent study on the global genetic diversity of S. bicornis, the phylogenetic tree showed the subdivision of haplotypes into two groups that agreed with previous studies (Bradshaw et al. 2021a). The pathogenicity tests on 10 Acer spp. showed A. macrophyllum is significantly more susceptible to S. bicornis than any other Acer spp. (Bradshaw et al. 2021a).

• (45) Sawadaea tulasnei (Fuckel) Homma (1937)

Repeatedly reported on maple trees (Acer platanoides L. and Acer sp.) in BC (Joshi et al. 2013, 2014).

Notes: Braun and Cook (2012) noted that both S. bicornis and S. tulasnei could cause infections on A. platanoides. The two species are morphologically readily distinguishable. Mycelium and chasmothecia of S. tulasnei are usually epiphyllous, whereas hypophyllous in S. bicornis. Furthermore, there are numerous additional differences in the asexual and sexual morphs (Braun and Cook 2012, p. 173).

A doubtful species

Erysiphe communis was reported on Phlox paniculata from Quebec (Gilbert et al. 2014). Nomenclaturally, it is inappropriate to use the name E. communis, and taxonomically ambiguous. When Erysiphe communis (Wallr.) Fr. 1829 was first described by Wallroth (1819) as Alphitomorpha communis, he cited various earlier published valid names, therefore Braun and Cook (2012) considered it nom. illeg. (nom. superfl.) when introduced. However, since Fries (1829) recognized this taxon as E. communis, it became a sanctioned name in the sense of the current Code (ICNafp). Therefore, this name was later rejected via a proposal and is now listed in ‘Appendix V, Suppressed Names’ of the Code. Throughout history, a number of authors recognized numerous forms and varieties. Notably, Jaczewski (1927) recognized over a hundred forms, many of which proved to be synonyms of distant species, e.g. E. pisi DC., Blumeria graminis (DC.) Speer, Podosphaera fugax (Penz. & Sacc.) U. Braun & S. Takam., Golovinomyces biocellatus (Ehrenb.) Heluta, etc. Species on Phlox could possibly be G. magnicellulatus (U. Braun) Heluta var. magnicellulatus or P. xanthii (Castagne) U. Braun & Shishkoff, as recorded in a recent monograph (Braun and Cook 2012). In addition, another four taxa on plants of family Polemoniaceae were G. magnicellulatus var. robustus (R.Y. Zheng & G.Q. Chen) U. Braun, P. collomiae (U. Braun) U. Braun & S. Takam., P. polemonii (L. Junell) U. Braun & S. Takam. and Leveillula taurica (Lév.) G. Arnaud. However, L. taurica is also a poorly defined species. In any case, the Canadian collection reported as ‘E. communis’ on Phlox is in urgent need of re-examination and clarification.

In addition to the species discussed above, of the over 700 powdery mildew incidences recorded in the CPDS from 2000 to 2019, one third was identified only at generic level. The causal agents of those incidences are in urgent need of identification using molecular phylogenies and morphological features. With the increased evidence that many genetically distinct species are morphologically indistinguishable, it is essential to develop well-maintained reference DNA sequences for type specimens to be compared with for identification purposes.

Acknowledgements

We thank the editors and two anonymous reviewers for their careful reviews and meticulous edits to improve the manuscript. This research was supported by Agriculture and Agri-Food Canada STB research funds for fungal and bacterial biosystematics J-002272.

Disclosure statement

No potential conflict of interest was reported by the author(s).