Hymenoscyphus fraxineus vs. Hymenoscyphus albidus – A comparative light microscopic study on the causal agent of European ash dieback and related foliicolous, stroma-forming species

Five species of Hymenoscyphus that fruit on black stromatized parts of dead leaves of deciduous trees are presented, giving details on their morphological and ecological characteristics. Several of these species have previously been misplaced in rutstroemiaceous genera because of the presence of a substratal stroma. However, the heteropolar, scutuloid ascospores with an often hook-like lateral protrusion at the rounded apex and the ascus apical ring of the Hymenoscyphus-type represent two reliable morphological characteristics that, together with molecular data, provide clear evidence for their placement in the genus Hymenoscyphus (Helotiaceae). Among the species treated is Hymenoscyphus fraxineus (=Hymenoscyphus pseudoalbidus), the causal agent of the European ash dieback disease. Since 1992 this species started within Europe to replace the rather uncommon Hymenoscyphus albidus, which is likewise confined to leaves of Fraxinus. Hy. fraxineus has been recorded already since 1990 in Eastern Asia (Japan, Korea, northeast of China), where it had been initially misidentified as Lambertella albida (≡Hy. albidus). In these regions, it occurs as a harmless saprotroph on Fraxinus mandshurica and Fraxinus rhynchophylla, suggesting that those populations are native while the European ash dieback disease has a recent Eastern Asiatic origin. The distinctly higher genetic diversity found in Japanese Hy. fraxineus in contrast to European Hy. fraxineus supports this view. Genetic similarities between Japanese Hy. fraxineus and European Hy. albidus suggest that also Hy. albidus might be a descendant of Asian Hy. fraxineus, though having invaded Europe much earlier. However, consistent genetic deviation between European and Asian Hy. fraxineus at two nucleotide positions of the ITS region indicates that the European ash disease originates from a region different from the presently known areas in Eastern Asia. Our results underline the importance of detailed morphological studies in combination with molecular work. Hy. fraxineus was described from Europe as a cryptic species that differed from Hy. albidus by molecular data alone. However, the Hy. albidus/Hy. fraxineus species complex represents one of many examples within the ascomycetes in which subtle microscopic differences between closely related species, in this case the presence or absence of croziers at the ascus base, are strictly correlated with molecular characteristics. Two species that closely resemble Hy. albidus and Hy. fraxineus form pseudosclerotia in Aesculus leaves and again differ from each other mainly in the ascus base: Hymenoscyphus aesculi on Aesculus hippocastanum from Europe lacks croziers, whereas Hymenoscyphus honshuanus from Japan on Aesculus turbinata possesses croziers. Other taxa treated here include Hymenoscyphus vacini, a European species growing on stromatized net veins of skeletonized leaves of Acer, and Hymenoscyphus torquatus, a Chinese species on unidentified herbaceous stems. An equivalent stroma-forming North American species on leaves of Fraxinus, Rutstroemia longipes (Rutstroemiaceae), is discussed and compared. A key to the Hymenoscyphus species that form a dark stroma on leaves of Acer, Aesculus, Fraxinus, and Picrasma is provided.


Introduction
Ash dieback is a serious disease in Europe causing death of European ash. The disease is caused by Hymenoscyphus fraxineus (T. Kowalski) Baral et al. (≡Chalara fraxinea T. Kowalski, =Hymenoscyphus pseudoalbidus Queloz et al.), which was recently introduced from Asia and has rapidly replaced the non-pathogenic European species Hymenoscyphus albidus. The present study outlines the introduction and cause of the disease, and compares closely related harmless species on ash and other hosts.

Hymenoscyphus albidus, a saprotroph on leaves of European ash
In late summer and early autumn, Hy. albidus (Roberge ex Gillet) W. Phillips forms white stipitate apothecia on fallen, previous year's rachises (here used for the entire main axis including the basal petiole) and leaflet veins of the pinnate leaves of ash (Fraxinus excelsior, Oleaceae, Lamiales). During the past 150 years, the species has infrequently been recorded in temperate and montane Europe. It can be assumed that the ascospores infect living leaves, and the mycelium grows endophytically inside them, similar as it is known in Hy. fraxineus. Quite a long time after the leaves have fallen, and after the leaf blades have disappeared, the fungus forms a very thin black stromatic layer on the surface of the remaining rachises and veins, which is referred to as 'pseudosclerotial plate'. This layer becomes slowly apparent during the winter months, and might serve as a protection of the underlying hyphae against ultraviolet light, or to avoid invasion into the pseudosclerotium by hyphae of other species. The terms pseudosclerotium and pseudosclerotial plate were adopted by Kowalski and Holdenrieder (2009) and Queloz et al. (2011) in the Hy. albidus complex, although they are somewhat ambiguously defined (see Gross & Holdenrieder 2013).
Due to the black pseudosclerotium, the fungus was placed by some authors in the genus Lanzia Sacc. or Lambertella Höhn. (Sclerotiniaceae, now Rutstroemiaceae). However, the type of ascus apical ring and the heteropolar (scutuloid) ascospores refer it to the genus Hymenoscyphus Gray (Baral & Krieglsteiner 1985, p. 121). This placement was meanwhile confirmed by molecular data (Queloz et al. 2011).

Appearance of a second species of Hymenoscyphus
Recent research on the current epidemic ash dieback in Europe revealed the asexual state Ch. fraxinea as the causal agent of ash disease (Kowalski 2006), which, by molecular comparison, proved to be connected to a sexual morph (Kowalski & Holdenrieder 2009), though at first misidentified as Hymenoscyphus albidus. A molecular study by Queloz et al. (2011) on a variety of teleomorph collections, mainly from Switzerland, revealed that in fact two different ascocarp-forming species fruit on the blackened rachises of ash, which could not be distinguished morphologically. One of them (Hy. albidus) proved to be rather harmless, whereas the other represented the serious pathogen. The latter was described as a new species, Hy. pseudoalbidus Queloz et al., based mainly on the deviating rDNA sequence. It formed slightly longer ascospores and larger fruit bodies, but was otherwise thought to concur completely with Hy. albidus and was, therefore, referred to as a 'cryptic species' (Queloz et al. 2011).
Apothecia of Hy. fraxineus, as the holomorph of Ch. fraxinea/Hy. pseudoalbidus is now called, have been collected in Europe at least since 2006 (two collections dating from 1978 and 1987 as listed in the specimens examined by Queloz et al. 2011 were based on confused DNA sequences and concern Hy. albidus, Queloz et al. 2012). The symptoms of leaf wilting were first observed in the northeastern part of Poland in 1992 , and the first isolate of its asexual state dates from December 2000 and represents the type culture of Ch. fraxinea (Kowalski 2006).
Between 2000 and 2010, the invasive fungus spread to the west and south of Central Europe, to southern parts of Northern Europe, and to Eastern and Southeastern Europe (Kirisits 2010), and during 2008-2013 it invaded Great Britain (Hendry 2013). In the last years, Hy. albidus could still be found in high-montane areas of Central Europe (see Queloz et al. 2011), Southern and Southeastern Europe (N. Matočec & I. Kušan personal communication), in Atlantic lowland regions of Western Europe and in boreal to continental regions of Northern and Eastern Europe, where Hy. fraxineus did so far only as an exception expand or now occurs sympatric. Hy. albidus appears to be extinct all over Central Europe north of the Alps.
Besides the here treated Hymenoscyphus honshuanus Baral, Hymenoscyphus aesculi (Velen.) Baral & E. Rubio, and Hymenoscyphus vacini (Velen.) Baral & E. Weber, a further species was recently described, Hy. albidoides (Zheng & Zhuang 2014). It was found fruiting in Eastern China on blackened veins of leaves of Picrasma quassioides (Simaroubaceae, Sapindales). In its morphological characters, this species can hardly be distinguished from Hy. fraxineus, except for the shape of the crystals in the stipe base, but its molecular data significantly deviate. Another species was recently detected in Japan growing on rachises of Fraxinus platypodia, which will be described in a separate paper as Hymenoscyphus linearis Hosoya, A. Gross & Baral. It concurs microscopically quite well with Hy. albidus but forms very narrow, linear stromata of considerable length. We have included both species in our dichotomous key.

Purpose of study
The observation by the first author of striking and unequivocal deviations in the ascus base within European populations collected in 1988European populations collected in -1991European populations collected in (simple septate) and 2006European populations collected in -2010 raised the question whether this character is correlated with the observed molecular markers and, as a consequence, Hy. albidus can be distinguished from Hy. fraxineus by morphological methods. In course of this study, a large number of fresh as well as herbarium specimens were examined in order to verify the presence or absence of croziers and to compare the result with the available molecular data. A detailed documentation of teleomorph morphology was made in order to clarify whether further distinguishing features exist. Japanese herbarium material of H fraxineus [originally identified as Lambertella albida (Roberge ex Gillet) Korf] was included in this study in order to find out whether differences exist to European Hy. fraxineus.

Materials and methods Microscopy
Collections were examined preferably in the living state, but also from rehydrated herbarium material, using a Zeiss Standard 14 and a Zeiss Standard KF microscope equipped with achromatic and plan-apochromatic objectives. Tap water (H 2 O) was used as a standard medium, and viability of cells was evaluated in that medium according to Baral (1992). The iodine reaction was tested with Lugol's solution (IKI =~1% I 2 , 2% KI, in H 2 O), without (rarely with) pre-treatment by potassium hydroxide (KOH 3-5%). Brilliant Cresyl Blue (CRB,~1% in H 2 O) added to a water mount was used for testing the presence of gel, also for vital staining of the refractive vacuoles (VBs). For testing the presence of croziers, fresh apothecia were sectioned free-hand, and sections transferred to the slide in a drop of H 2 O; any pressure on the cover slip was avoided during preparation. In the case of herbarium specimens, hymenial fragments (mainly sections) were placed in a drop of H 2 O, to which a small drop of KOH and often also one of aqueous Congo Red (CR) was added. Alternatively, CR SDS (SDS = sodium dodecyl sulphate) was applied to a water mount. Dissolution of the oil drops (lipid bodies, LBs) by ethanol was conducted as follows: a dry apothecium was rehydrated and a fragment placed in 90% ethanol for ½ min; then the ethanol was removed and a drop of 3% KOH and one of 1% phloxine was added. Waterman blue-black ink was applied for a better visibility of ascospore sheaths. Photographic images (macro-and microphotos) were obtained using a Nikon Coolpix E4500 and a Nikon Coolpix 5000 (Nikon Corporation, Tokyo, Japan). All drawings are free-hand.
Abbreviations * = living state, † = dead state, n.v. = non visus (specimen or documentation not seen by us), d.v. = documentum visus (photos/drawings/descriptions seen by us), sq. = sequence of ITS1-5.8S-ITS2 region, vid. = examined, ø = unpreserved, # = not tested for the ascus base. Values in curled parenthesis {} refer to the number of collections that were examined or, after the host plant and the associated taxa, the number of certain/uncertain records. Countries are abbreviated according to the ISO 3166 two-letter standard code.
Nomenclature. Korf (1982) and Hosoya et al. (1993) cited Desmazières's exsiccatum 'Plant. Crypt. N. France, #2004, 1850' as basionym of Peziza albida Roberge in Desm., whereas other authors considered the first valid publication to be that of Desmazières (1851, p. 323). A description was only supplied in 1851, where Desmazières referred to the exsiccatum as 'Rob. in herb.'. The label on the exsiccatum comprises notes on the substrate (on blackened areas of half-rotten rachises of F. excelsior) and phenology (summer), as well as a comparison with Peziza inflexa (=Cyathicula coronata), which is said to resemble in its pale colour and cyathiform shape. Since the Code does not require a minimum text for a diagnosis, Desmazières' label from his widely distributed edition of exsiccata of Pl. Crypt. N. France meets the requirements for a valid publication of the new taxon. The date of 1850 is somewhat uncertain, however. According to Stafleu and Cowan (1976), the title pages of the fascicles of Pl. Crypt. N. France were preprinted in some cases before distribution and may give a too early date, thus the exact citation of Peziza albida remains unsettled.
Nevertheless, the name Peziza albida is illegitimate because it is a homonym of several older binomials. Therefore, the first legitimate name of the fungus at the species level is Phialea albida Roberge ex Gillet (1881) or Phialea albida Gillet, and this binomial is treated as a nomen novum attributed to Gillet. Likewise, He. albidum (Roberge ex Gillet) Pat. is illegitimate because of homonymy to older binomials. Dennis (1956) created the epithet robergei in order to avoid such homonymy within the genus Helotium Pers. The name He. robergei was in use until Helotium was abandoned by Dennis (1964) in favour of Hymenoscyphus, because of the older Helotium Tode (=Omphalina Quél.). Sharma (1976) and Svrček (1979) combined the epithet robergei in Hymenoscyphus, which is impermissible because here no older competing homonym with the epithet albidus exists.
Some workers have cited the authors of Hymenoscyphus albidus as 'Roberge ex Desm.', which would imply that Desmazières validated an invalid diagnosis of Roberge. On the label of Desmazières 'Plantes Cryptogames Nord du France' (1850) and in his 'Notice sur les Plantes Cryptogames' (1851), the authorship of the diagnosis and discussion to each species is actually difficult to know. We arrived at the conclusion that Desmazières considered himself as the author of a taxon only in those cases of his 1851 paper where 'Desm.' or 'Nob.' is mentioned after the fungal name. This is not the case in Peziza albida where he wrote 'Rob. in herb.'. Also we believe that the author of the printed diagnoses of 1850 and 1851 is Desmazières, unless we would gain knowledge of a manuscript or correspondence proving different authorship. The remark 'Desmaz.' after the ecology in the 1851 paper is found in all those cases where the remark 'Rob. in herb.' or 'Nob. in herb.' appears after the fungal binomial, irrespective of whether or not an exsiccatum is cited. We interpret the remark 'Desmaz.' as an indication of the authorship of the diagnosis but not of the collection. The phrases 'Rob.' or 'Rob. in herb.' after the binomial on the printed label of  Shortly after the demise of Roberge and Desmazières, authors were still quite consistent in their citation. Indeed, many of those in the nineteenth century followed Desmazières by citing the author of the taxon as 'Rob.', e.g., Roumeguère (1870, p. 130), Berkeley and Broome (1878, p. 29), Patouillard (1885, p. 173), Phillips (1887, p. 138), Rehm (1893, p. 797), and Massee and Crossland (1905, p. 285). Consistent with this citation is the variant 'Rob. in Desm.', which can earliest be found in Phillips (1887), and was adopted by Korf (1982) and Hosoya et al. (1993). In contrast, Roumeguère (1891Roumeguère ( , 1892 cited the authors as 'R. et D.' (Roberge & Desm.), hence he considered Desmazières as a coauthor. This version was followed by Boudier (1907, p. 111, 1908, p. 287), Schröter (1893, p. 74), Migula (1913, p. 1167, Dennis (1956), Arendholz (1979), andCarpenter (1981).
We are inclined to adopt here the version 'Roberge in Desm.', because Desmazières attributed the authorship to Roberge. In any case, the original binomial was invalid due to homonymy; therefore, the correct author citation for combinations other than Peziza albida is Roberge ex Gillet or merely Gillet.
History of type studies. In or shortly before 1850 a rich collection was made, on which the description of Hy. albidus in Desmazières (1851, p. 323) is based. The sample was divided already in 1850 in a number of duplicates, and disposed as two exsiccata series in Pl. Crypt. N. France (Ed. 1, n°2004;Ed. 2, n°1604). For each of the two exsiccata, one duplicate was deposited in the herbarium of CN (Caen, Basse-Normandie), one came in Desmazières' herbarium which is preserved at PC, and others are now to be found in herbaria such as FH, GENT, and K.
Desmazières (1851) mentioned the two exsiccata in his description, but without specifying which is the type. Apart from this rich collection, a herbarium specimen exists in K that bears Roberge's handwriting and is apparently not part of the two distributed exsiccata series, but a different, later collection by Roberge, as it bears the note 'Peziza albida Roberg. (Desm. 2004)'. Nylander (1868, p. 40) examined two duplicates of Hy. albidus, possibly those from PC, since he was working at the Muséum national d'histoire naturelle, Paris in 1850-1858, and finally moved to Paris in 1863. He mentioned this only in a note under Peziza albidula (Hedw.) P. Karst.
[as '(Hdw.) Whlnb.', ≡ Phialea cyathoidea var. albidula (Hedw.) Rehm], in which he gave data on ascus and spore dimensions different from those of Desmazières. Karsten (1871, p. 112) made a description, apparently from type material (presumably in H), because he stated that the species was not yet collected in Finland. Cooke (1876, p. 132, pl. LXV fig. 297) presented an uncommented microscopic drawing based on Desmazières n°2004. Massee (1895, p. 260) supplied a description of an authentic specimen of Roberge, and also examined Desmazières n°2004 which he considered conspecific (both in K). Like Nylander and Cooke, also Massee did not select a type. In the nineteenth century, it was not common usage to specify a single specimen as type when erecting a new taxon.
The American mycologist White (1944, p. 608, fig. 19-24) appears to be the first in the twentieth century who re-examined the type collection of Hy. albidus. He considered the two syntypes in FH as conspecific, and gave a thorough and precise redescription. His detailed illustration was based on n°2004, and in the legend he wrote 'all from type material, Desm. Pl. Crypt. Fr., Ed. I, 2004 (FH)'. We believe that this statement can be taken as a lectotypification (Art. 7.11,9.21 ICBN).
When Arendholz (1979, p. 79) studied three duplicates of the two exsiccata (n°2004 in K and PC, n°1604 in K), he stated that n°2004 is the 'type', apparently following White's lectotypification. Also Carpenter (1981, p. 187) examined the two exsiccata in PC (one erroneously citing as n°2005). By overlooking White's lectotypification and unaware of Arendholz' study, he designated n°1604 (PC) as lectotype.
When erecting the new species Hy. pseudoalbidus, Queloz et al. (2011, p. 141) examined those two duplicates of Hy. albidus deposited in GENT. Unaware of the previous designations, the authors designated n°2004 in GENT as lectotype.
Of course, the two latter designations must be rejected. According to Laundon (1979), the personal herbarium of J.B.H.J. Desmazières (1786Desmazières ( -1862 is deposited in PC. Although the Code recommends to select material as lectotype that was in the personal herbarium of the author (Recommendation 9A.4., ICBN), White's selection of a duplicate in FH must be followed.
Desmazières' (1850) two exsiccata appear to be part of a single collection. This can be concluded from the numbers in each of the two editions, which refer over a large range consecutively to the same taxa. Desmazières (1850Desmazières ( , 1851 did not indicate date and location of the collection (except for 'summer'). However, the handwritten label of Roberge's specimen preserved at K bears the note 'Caen ' [dépt. Calvados, France], the city where Roberge lived. Although no locality is mentioned for the type specimen, it seems probable that Roberge collected also this in the vicinity of Caen. Husson et al. (2011) surprisingly reported more detailed specimen data: 'collected by Roberge in 1850 from the Bois de Lébisey, close to Caen'. However, these data are actually an assumption which J.P. Rioult (director of the Départerment de Botanique, Mycologie et Biotechnologies, University of Caen) conveyed to the authors (C. Husson, personal communication) and also to us, reasoned from his personal experience about the most frequent origin of Roberge's collections. Indeed, Roberge mostly collected in the parc de Lébisey, where he discovered his most important cryptogams (Morière 1866).
Misidentified hosts reported for Hymenoscyphus Albidus. Declercq (personal communication) reported collections of Hy. albidus from Belgium on Juglans (unpreserved) and Pyrus (B.D. 94/108). However, re-examination of the latter collection revealed the host to be Fraxinus. Two unpreserved records included in our description of Hy. albidus were thought by the first author to be on Acer. Both showed asci without croziers and unquestionably concern Hy. albidus: in that from Luxembourg (24. IX.1990, 'on blackened petiole of Acer pseudoplatanus') druses of crystals were noted in the stipe, and in that from Tübingen (3.IX.1988, 'on blackened primary veins of skeletonized leaf of ?Acer') the spores contained large LBs. These notes on the presence of crystals or large LBs exclude Hy. vacini. Based on the general observation of a strict occurrence of Hy. albidus on Faxinus, we assume that the host was misidentified here.
Records on Fraxinus leaves misidentified as Hymenoscyphus albidus. Among the records on Fraxinus, possibly all those can be excluded in which the apothecia grew verifiably on unblackened parts of the rachises. Hy. albidus was actually confused in the years before ca. 1950 with Cyathicula fraxinophila (Svrček) Baral. According to Arendholz (1979, p. 80), almost all German records found in herbaria differ from British specimens in a textura oblita and represent at least partly a Cyathicula. Cy. fraxinophila resembles Hy. albidus in external view and is also confined to rachises (including petioles) of Fraxinus. Based on personal observations of the first author, this species deviates in a strongly gelatinized ectal excipulum (textura oblita) covered by rhomboid crystals, much narrower asci and spores, the former arising from croziers and with a more conical apex with an amyloid apical ring that resembles the Calycina-type, the latter homopolar and with a low lipid content that consists of minute LBs [spore size *(10-) 14-18(-20) × (2.2-)2. 5-2.8(-3) µm], also in multiguttulate paraphyses (VBs smaller and always globose), and particularly in the absence of a black stromatic tissue on the rachis and stipe base. The whitish apothecia also deviate in having a concave disc (0.6-1.2 mm diam.) with a finely crenulate margin.
Indeed, Rehm's (1893, p. 797) description under the name He. albidum, with spores 15-18 × 3 µm with one oil drop at each end, undoubtedly refers to Cy. fraxinophila. Rehm mentioned the species from three localities: Bayern (Franken), Switzerland, and Südtirol; yet, he excluded the collection from Südtirol (leg. G. Bresadola) as deviating from the others, although he noticed here only the characteristic black stroma described by Desmazières. Rehm mentioned the Bresadola specimen also under Helotium virgultorum (p. 782), a lignicolous taxon which he described as growing on blackened wood surface, and he was sure that it belongs here. So only this specimen seems to represent true Hy. albidus.
Without description, Le Gal (1938) reported He. albidum from Île-de-France (Villecresnes, Bois de la Grange et de l'Etoile) and stated the species to be very common in summer. It is possible that she was dealing in fact with Cy. fraxinophila. Grelet (1949, p. 35) mentioned this record, but his diagnosis is more or less a copy of Boudier's, and it seems that Le Gal's material has never been re-examined. Also Jaap's (1910) report on Fraxinus rachises is without description; the stated association with Cy. coronata suggests that he was dealing with Cy. fraxinophila, a species which we found three times in association with Cy. coronata, whereas Hy. albidus grew always alone on the rachises. However, during survey in Czechia in 2013 Hy. fraxineus and Cy. fraxinophila were found sometimes on the same petiole (O. Koukol, personal communication). Also Gillet (1881) might have treated Cy. fraxinophila, because he did not mention a dark stroma.
This confusion includes also the leaf-inhabiting Hymenoscyphus caudatus (P. Karst.) Dennis. White (1944) stated that Hy. albidus is 'known with certainty only from the material of Desmazières', whereas he found all the many other European records under this name to be largely based upon specimens better referred to Hy. caudatus or He. scutula (Pers.) W. Phillips. This statement is in contradiction to Arendholz (1979), who mentioned only the genus Cyathicula as misidentification.
History about the detection of Hymenoscyphus fraxineus. Judging from the list of specimens examined by Queloz et al. (2011), Hy. fraxineus was genetically recorded in Europe since 2000, with two exceptions dating back to 1978 and 1987 which later turned out to concern Hy. albidus (see below). Although the ash dieback disease was first observed in Northern Poland in 1992, the oldest isolate of the anamorph state dates from 6.XII.2000, which represents the type culture of Ch. fraxinea (Kowalski 2006 (Kowalski et al. 2013), and totally replacing Hy. albidus in these regions. During invasion of Switzerland, Hy. fraxineus did not surpass an altitude  of 850 m while Hy. albidus was still present at higher altitudes (up to 1690 m), according to data in Queloz et al. (2011). However, there is evidence that both species exhibit a similar tolerance to low temperatures. Apothecia of Hy. fraxineus were presently recorded in the Alps at 1410 m and in Japan at 1325 m as a maximum, and disease symptoms were seen in the Alps up to 1600 m. When randomly checking fresh collections of Hy. albidus (sensu lato), the first author observed differences in ascus development among the studied samples: apothecia found during 1988-1991 in the south of Germany (Tübingen) and in Liechtenstein possessed asci arising from simple septa, whereas in those found during 2006-2010 in Sachsen (Chemnitz), Mecklenburg (Rehna), and Denmark (Sjaelland) the asci consistently arose from croziers. Because there were no other deviating features to be found in the 2006-2010 specimens, the observed deviation was first thought to represent variation within a single species. When ash disease invaded Europe and the Hy. albidus sexual state was identified as the causal agent, the first author became suspicious that the two variants of Hy. albidus that he had observed, with differing types of ascus development, were the two 'cryptic' species reported by molecular methods.
In the present study, 13 sequenced specimens from ZT and 1 from HMIPC, which were identified by Queloz et al. (2011) based on their rDNA sequences, were re-examined for the ascus base. Further four specimens were sent to V. Queloz by the first author to be sequenced. The result confirms the above hypothesis: Hy. fraxineus can be distinguished from Hy. albidus by the presence of croziers: six out of seven collections identified by molecular methods as Hy. pseudoalbidus (=Hy. fraxineus) turned out to have croziers, whereas seven collections identified as Hy. albidus were found to have simple-septate ascus bases (Tables 1 and 2, HMIPC, ZT).
However, the result for one of those specimens identified by molecular data as Hy. pseudoalbidus (Luzern, Aesch, collected in 1978, NMLU 0409-78) behaved contradictory: the asci were found to be simple-septate. Since also the year of collection argues against Hy. fraxineus, the sequence taken from this sample was thought to originate from DNA of another specimen. The same suspicion arose concerning a collection made in 1987 (Zürich, Sihlwald, ZT 87-236) which has not been re-examined for croziers. Due to their age, the asserted identity of these two samples confused also other workers, e.g., Husson et al. (2011). Because of this and the observed discrepancy concerning the ascus base in the specimen from Aesch, both were resequenced for ITS and calmodulin by V. Queloz in 2012 and, indeed, turned out to represent Hy. albidus . Their sequences were subsequently replaced in GenBank, and their identity corrected.
Different theories have been proposed concerning the origin of the causal agent of ash dieback in Europe. These include a virulent mutant, either of the native Hy. albidus (Kowalski and Holdenrieder 2009, p. 307) or of the introduced Hy. fraxineus (Queloz et al. 2011, p. 141), but even the invasion of Hy. fraxineus from outside Europe without associated mutation (Queloz et al. 2011, p. 141). The latter authors and Bengtsson et al. (2012) were the first who assumed an Asian origin for the pathogen. When searching for reports from outside Europe, the only reliable record on Fraxinus rachises that came to our notice was that from Japan by Hosoya et al. (1993), who reported the species under the name Lambertellinia albida, with three specimens on rachises of Fraxinus mandshurica var. japonica collected during 1990-1992 in Hokkaido and Nagano Prefectures. Re-examination of five specimens from the same prefectures collected during 2005-2011, kindly supplied by T. Hosoya and referred to Hy. fraxineus (as Hy. pseudoalbidus) by Zhao et al. (2013) based on their DNA sequence data, proved indeed to have asci arising from croziers (Table 2, TNS-F; see also Zhao et al. 2013, fig. 4). Likewise, several specimens on rachises of F. rhynchophylla and F. mandshurica from South Korea showed croziers at the ascus base, and a sequence that sits in the Hy. fraxineus clade ( Table 2,  These results, which suggest an Eastern Asian origin of the European ash dieback disease, underline the importance of detailed morphological studies in combination with molecular work. In the present case this concerns an unequivocal character state of the ascus base within the genus Hymenoscyphus, which is either + or -, though the elucidation of this character requires a certain skill and mounting technique. The species complex Hy. albidus/Hy. fraxineus represents one of many examples of subtle, currently neglected microscopic differences between closely related species within the ascomycetes. The present example is extraordinary because of the great economic importance of one of the two involved species. The first report of dieback of young shoots of ash within Europe concerns regions in Northwestern Poland and dates from the year 1992 (Kowalski & Czekaj 2010). Other countries announced the disease in subsequent years (see, e.g., Timmermann et al. 2011, fig. 1;Husson et al. 2011). During the past decades Hy. albidus s.l. was not continuously collected and examined by the first author under the microscope, for two reasons: (1) the taxon was apparently not frequent at that time and (2) prior to the first record of croziers in 2006 it seemed to be a species that is already macroscopically well characterized by its occurrence on blackened rachises of ash. Hence, herbarium specimens were rarely kept, and a gapless elucidation of the first appearance of croziers in the ascocarps of this species aggregate cannot be accomplished.
Molecular studies performed on the two taxa by various authors strongly suggest that the causal agent of the disease is always Hy. fraxineus, never Hy. albidus. Our comprehensive microscopic examination of specimens collected since 2011 in Central Europe revealed exclusively apothecia in which the asci arise from croziers. Together with the consistent occurrence under trees that showed severe symptoms of the disease, it can be concluded that all these records represent Hy. fraxineus. In colline regions of Central Europe, this species starts fruiting already in May-June, distinctly earlier than Hy. albidus which started fruiting in July in that region. Undoubtedly, Hy. fraxineus has totally replaced Hy. albidus there, perhaps due to its high pathogenicity to the European ash and the advantage of early fruiting.
Earlier reports of ash dieback. The term 'ash dieback' is not a recent construct that emerged with the appearance of the disease caused by Ch. fraxinea and thus named in French more precisely 'Chalarose du Frêne'. There were other events before with similar features as dieback of young shoots in the crown of the trees but with different causes. In the 1960s and 1970s such occurrences were reported from the US (Croxton 1966;Hibben and Silverborg 1978) and in the 1980s (Pawsey 1983) and even in the late 1990s (Wiltshire et al. 1996) from England. But they were ascribed to climatical conditions rather than to fungal infections.
Nomenclature. According to the new nomenclatural rules (Art. 59, ICN Melbourne, McNeill et al. 2012), the name Hymenoscyphus pseudoalbidus needs to be changed, because the specific epithet of Ch. fraxinea T. Kowalski 2006 is older than that of Hy. pseudoalbidus Queloz et al. 2011, while the generic name Hymenoscyphus Gray 1821 predates that of Chalara (Corda) Rabenh. 1844. Gams et al. (2012, p. 495) recommended to generally accept the epithet which is in the prioritized genus: 'When a binomial in a prioritized genus has a younger epithet than the corresponding name in the suppressed genus, only priority of extant names in the prioritized genus should count.' However, this proposal was later not accepted. Since molecular data on Hy. albidus and Hy. pseudoalbidus support that these species are closely related to the type species of Hymenoscyphus (Queloz et al. 2011), and because Ch. fraxinea yielded in online search engines about 5 times as many hits as Hy. pseudoalbidus (52.200 vs. 10.100, Google Web Search, 22. V.2014), Baral et al. (2014) proposed the new combination Hymenoscyphus fraxineus.
Molecular studies have shown that the genus Chalara in its large sense concerns anamorphs of very different ascomycetes (Cai et al. 2009): most of the investigated species were found to belong in the Helotiales, but a few are closely related to Xylariales. Species of Chalara s.l. connected to Ceratocystis Ellis & Halst. (Microascales) were transferred to the genus Thielaviopsis Went already by Paulin-Mahady et al. (2002).
The type species of Chalara, Ch. fusidioides (Corda) Rabenh., was originally described from Czechia on bark of Pinus, but later records were made on very different substrates such as Fragaria (leaves), Podocarpus, Vitis, and Mycosphaerella (perithecia), according to a redescription by Nag Raj and Kendrick (1975, p. 119, fig. 30F), who presented an illustration of the holotype. A single sequence under the name Ch. fusidioides exists in GenBank (18S rDNA, from Picea, no source given). When best-matched sequences are searched using BLAST in GenBank (http://blast.ncbi.nlm.nih. gov/Blast.cgi), only helotialean taxa are yielded, for instance members of the genera Hyaloscypha Boud. or Bulgaria Fr. but no true Hymenoscyphus under the first 100 hits. In the phylogenetic analysis by Réblová et al. (2011), Ch. fusidioides clusters in a clade with Calycina citrina (as Bisporella citrina). Whether this Chalara strain was correctly identified remains uncertain, as no morphological documentation was available to us. Within the relationship of Calycina, chalara-like anamorphs are known, e.g., in Calycina claroflava (Grev.) Baral et al. (Johnston 1988, as Bisporella discedens, anamorph Bloxamia Berk. & Broome), Calycellina ulmariae (Lasch) Korf (Baral 1989, pl Since the identity of type material of old teleomorph names is often easier to settle compared to anamorph names on the basis of a morphological reinvestigation, teleomorph genera should be given preference in regard to nomenclatural priority in those groups where teleomorphs exhibit clearer specific characteristics than anamorphs. As a consequence, we here consider Chalara as a form genus of anamorphs. Although the type of Chalara resembles Ch. fraxinea Figure 10. Hymenoscyphus fraxineus (Europe). a 1 , b, c, d 1 , e 1 , g, h 3. ascospores (freshly ejected, containing LBs); h 4 . do. (LBs deteriorated); a 2 , j. paraphyses (containing VBs); e 2 , i, k. mature asci; f, h 1 . overmature ascospores (septate, partly with olive-brown wall); u. ascus apices with euamyloid apical ring; remaining figures: ascogenous hyphae with crozier formation. -Living state: a 1 -e 1 , a 2 , f 1, g (in Waterman blue-black ink), h 1 , h 3 , i, j (in CRB), k 1 , r-s, v; dead state: k 2 (in H 2 O), a 3 -a 4 , d 2 -d 4, h 2 (in CR); e 2 -e 3 , l-n, p-q (in CR SDS ), f 2 (in KOH), o, t (in KOH+CR), h 4 (in ethanol + KOH + phloxine), u 1 -u 2 (in IKI). morphologically, the presently available molecular data do not support a close relationship with a phylogenetic group presented in an analysis in , which can be taken as a representation of the family Helotiaceae in a restricted sense, comprising the genera Hymenoscyphus, Phaeohelotium Kanouse, Cudoniella Sacc., and Cyathicula De Not.
Characterization. Hy. honshuanus is not easily separable from Hy. fraxineus by its micromorphology, but the deviating substrate and the apothecia with a yellowish disc and longer stipes appear to be diagnostic. As already mentioned, the only difference between Hy. honshuanus and Hy. fraxineus worth mentioning seems to be the longer apothecial stipe and a more yellowish disc in the former species, apart from the deviating substrate. The excipular cells are a bit longer and narrower than in Hy. fraxineus. The lower width is to some extent due to the shrinking effect in the dead state, however. Moreover, the width of the ectal excipular cells in Hy. fraxineus varies rather strongly. Regrettably, a section of the blackened petioles and the stipe base of Hy. honshuanus was not studied, neither by Korf and Lizoň nor in the present study; therefore, the occurrence of a demarcation line beneath the sclerenchyma as seen in Hy. aesculi remains unclear, and also the presence of crystals in the stipe base, being characteristic of Hy. fraxineus and Hy. albidus, remains to be examined.
Nomenclature. Hy. honshuanus was originally described by Korf and Lizoň (1994) under the name Lambertellinia scutuloides from Japan on blackened petioles of Aesculus turbinata and is known with certainty only from the type collection. A transfer of Lambertellinia scutuloides to Hymenoscyphus is blocked by the different caulicolous taxon Hy. scutuloides Hengstm.; therefore, a new name is here proposed. The species was presented already by Korf and Lizoň (1993) on a poster at the Annual Meeting of the Mycological Society of America, with reference to an earlier volume of Mycotaxon (47, 1993), but it did not actually appear there.
Type study and circumscription. According to the present redescription of the holotype, which was briefly mentioned also in Galán and Baral (1997), the taxon deviates from H. aesculi and H. albidus in the presence of croziers and some other features (see also under those species). While the ascospores of H. aesculi lose their characteristic central swelling in the dead state, it is unknown whether living spores of H. honshuanus might also show such swelling. Korf and Lizoň (1994) thought that Lambertellinia scutuloides resembles the genus Lambertella Höhn. in its black substratal stroma and the ascospores getting finally light brown, but the authors placed it in a new genus because of its long-celled excipular structure, the assumed connection to an Idriella anamorph, and probably also because of the deviating scutuloid spore shape, to which the original species epithet refers. Regrettably, no microscopic features were illustrated in the protologue, except for a photograph showing a section of the excipulum.
Korf and Lizoň believed that collections on leaves of Acer rubrum from North America and Berberis vulgaris from Germany are conspecific, though they examined only that on Berberis. The collection on Acer rubrum was treated by Kimbrough and Atkinson (1972) under the name Hymenoscyphus caudatus. Unlike Hy. fraxineus, a strain from this collection formed an Idriella anamorph in pure culture, with slightly falcate conidia measuring ?15 × 2 µm (evaluated from photos), formed in sporodochia in a sympoduloconidial type of development. The original material was not available to Korf and Lizoň, who were sure about its conspecificity with Lambertellinia scutuloides because of the reported brown colouration of the 1-2-septate ascospores when 'mature'. However, Kimbrough and Atkinson did not observe stromatic tissues around the inhabited midribs and larger veins, although they saw the stipe base to arise from darkened host tissue. Regrettably, Kimbrough and Atkinson did not report the presence or absence of croziers, though they stated that the features of their fungus were essentially the same as in Hy. caudatus as described by White (1943) and others. Korf and Lizoň (1994) characterized the ectal excipulum of Lambertellinia scutuloides as a hyaline textura porrecta consisting of gelatinized hyphae 3-6.5 µm wide, externally covered by a thin layer of straw-or golden-yellow pigmented hyphae 0.5-1.5 µm wide, and internally delimited by a light yellow layer. The authors considered this construction of the ectal excipulum as different from typical members of Hymenoscyphus. However, the present re-examination of the excipular texture ( Figure 12(g)) revealed no substantial differences from, e.g., Hy. albidus, Hy. fraxineus, and Hy. aesculi.
The description given above is based on the present reexamination of the holotype, but it includes also some data of the protologue, mainly in regard to macroscopical features. The petioles from which the apothecia of Hy. honshuanus arise show a dark blackish-grey colour. Korf and Lizoň referred to it as a 'black substratal stroma'. Similar as in the other species treated in the present paper, a black ring was often seen at the outermost base of the stipe (Figure 12(d)). Contrary to Korf and Lizoň, who stated the asci to be simple-septate, the present examination of two apothecia showed that they arise from croziers ( Figure 12(b)). Since the authors did not illustrate the feature, their report of absent croziers remains mysterious. Apart from the abundant hyaline ascospores, only a single spore with a pale ochre-brown wall was seen in the present examination. It must be stated, however, that the asci, spores and paraphyses were distinctly wider than indicated by Korf and Lizoň, who measured them as follows: asci (82-)84.5-100(-116) × (5.6-)6.6-7.5(-8.3) µm, spores (13.8-)15.4-16.9(-18.5) × (2.3-)3.1-3.5(-3.8) µm, paraphyses 1.5 µm wide. The protologue appears to be based exclusively on the Japanese type specimen. Data of the specimen from Germany (on Berberis) were not incorporated, which is obvious from the mentioned larger spore size in the Berberis specimen (Korf & Lizoň 1994, p. 171).
We assume that Hy. honshuanus was incorrectly described by Korf and Lizoň (1994) concerning the ascus base, though it cannot be excluded that the holotype collection represents a mixtum of two species with deviating ascus base characters as well as ascus and spore measurements. In any case, the present data indicate that the here described specimen is not conspecific with either Hy. aesculi or Hy. caudatus. Hy. caudatus was redescribed by White (1943, p. 151, fig. 8), who observed the absence of croziers in most (though not all) of the many examined specimens, including the holotype. Regrettably, White did not specify the aberrant material in which he saw croziers.
Because of the striking morphological similarity with Hy. fraxineus, it would be interesting to determine the ITS sequence of Hy. honshuanus. However, this approach requires careful examination of single apothecia for the presence of croziers, in order to avoid confusion in the case of a mixtum. Actually, a further Japanese record (on petioles of Aesculus sp. kindly sent by T. Hosoya, TNS-F-12758) differs from H. honshuanus in the ascus base, indicating that two different species exist in Japan: the asci arise here from simple septa (Figure 15), and the species is, therefore, included in Hy. aesculi.
With the present data, Korf and Lizoň's conclusion that Lambertellinia scutuloides has an Idriella anamorph needs to be reinvestigated. Experiments with cultures of the common and plurivorous Hy. caudatus, but also Hy. vacini, a taxon specific to leaves of Acer, kept under varying conditions as proposed by Kimbrough and Atkinson (1972), could clarify whether or not any of these species are connected to an Idriella anamorph.
The German specimen on Berberis (Sydow, Mycotheca Marchica no. 1576, 1887) remains of unclear identity. The apothecia are said to arise from stromatized veins and petioles, but the presence or absence of croziers was not stated by Korf and Lizoň. When dry, the apothecia were 'dark brown (almost blackish)'. The spores measured 16.2-19.8 × 3.6-4.8 µm and were 'cylindric-clavate, scutuloid, deep brown to hyaline'. The exsiccatum was issued under the name Helotium berberidis Syd., but without a diagnosis on the label (Korf & Lizoň 1994).
Hy. honshuanus is known with certainty only from the type collection on Aesculus turbinata. Reports under the names He. robergei or Hymenoscyphus albidus on fallen petioles of Aesculus indica in Northern India and Nepal might well concern Hy. honshuanus (or Hy. aesculi), but the ascus base was not studied for the presence of croziers (see discussion under Hy. albidus).
The type locality of Hy. honshuanus is not exactly known, and the collector, Daisuke Shimizu, who was a zoologist specialized in vegetable wasps, passed away in 1998. The small town Oguni(-machi) belongs in the Nishiokitama district and lies 55 km WSW of Yamagata. The village Tsugawa could not be located (it was included in Oguni in 1960), but its altitude was found to be 294 m (T. Hosoya personal communication). Oguni has an altitude of about 140 m; therefore, the type locality must be in one of the adjacent mountain ranges.  Figure 13).
Apothecia. Moist (0.3-)0.6-1.5(-2) mm diam.; disc whitish(-greyish) or pale cream, slightly concave to flat, margin distinct, not protruding, smooth, exterior concolorous; stipe (0.25-)0.5-1.7(-2) × (0.08-)0.15-0.35(-0.42) mm, white, glassy-translucent or non-translucent, slightly pubescent, lower half gradually darker, near base black, erumpent from sclerenchyma-phloem layer; hymenium dry light to dark greyish-to blackish-brown, exterior pale olivaceous-grey. Asci   Characterization. Hymenoscyphus aesculi resembles Hy. albidus in many respects, including simple-septate ascus bases. Like Hy. albidus, it occurs wide-spread within Europe, where it appears to be specific to leaves of Aesculus hippocastanum. Yet, a single Japanese record indicates a wider distribution area for Hy. aesculi. The species differs from Hy. albidus in several characteristics: (1) the living ascospores consistently show a central swelling (bulge); therefore, they look fusiform rather than cylindrical to fusoid as in Hy. albidus; (2) the internal tissue of the stipe base is entirely devoid of crystals; (3) Hy. albidus generally shows a thin yellowish layer of exudate on the entire exterior, especially on the stipe, which is absent in Hy. aesculi; (4) senescent apothecia turn dark olivaceous-grey to blackish-brown whilst those of Hy. albidus turn cream to orange though sometimes also dark brown. These differences hold true also for a comparison with Hy. fraxineus, the latter differing from the other two in having croziers.
In comparison with the here presented description, these measurements fit only concerning spore length and ascus width, whereas apothecial size, spore width and ascus length are extraordinary large and perhaps erroneous, even when taking living elements into consideration.
Svrček (1989) did not describe the type material of He. aesculi, but he stated that it fully concurs with some fresh specimens studied by him. This is astonishing, because he measured the asci in one of his collections with 65-85 × 7-10 µm without discussing the discrepancy to the protologue. He also noted (p. 74) that 'the original Velenovský diagnosis is not accurate if compared with V.
Vacek's [the finder of the holotype] handwritten notes'.
No such handwritten notes were seen when the holotype was re-examined in the present study. It revealed an ascus size (see Figure 13) compatible with Svrček's above-mentioned measurement which likewise undoubtedly refers to dead asci. The contents of many paraphyses were bright olivaceous-brown, and in cross section of a petiole a distinct black stroma could be seen which surrounds the sclerenchyma on its outer as well as inner face. Svrček (1989, pl. 2 fig. 4) illustrated a personal collection from near Praha, which he considered conspecific with the type of Helotium aesculi. Concerning his reasons to transfer the species to Lanzia, Svrček (1985) referred to Figure 17. Hymenoscyphus aesculi: a-b. fresh apothecia; j-l, p. rehydrated apothecia (colour was originally white except for base of stipe); f-g, o. asci and paraphyses(o: with olive secondary pigment especially in paraphyses); h. ascus apices with euamyloid apical ring; i: simple-septate ascus base; c-e. ascospores; m-n, q-s. cross section through cortical region of petioles, with black demarcation line surrounding the sclerenchyma on outer and inner face (cortical parenchyma above and phloem beneath being entirely decomposed). a paper in press which, however, does not treat this species (see also under Hy. vacini). Probably he proposed the combination because of the blackish stroma, but in 1989 he mentioned merely the dark stipe base, not the substratal stroma. The drawn spores look as if they were in the living state, according to the regular lipid content. Svrček (1989, p. 20) described them as 'fusiform, clavate, often strongly tapering below and curved, filled with many small and larger guttules', 15-24.5 × (3.5-)4-5 µm, i.e., rather variable in length. The spores lack the central swelling, however, and the LBs are drawn distinctly smaller than in any of the here studied collections. Svrček examined also a further specimen from Southern Bohemia and one that H. Engel sent to him from Weidhausen near Coburg. The latter collection was also examined in the present study and found to have predominantly strongly fusiform spores with a central bulge and four rather large LBs, unlike the sketch by Engel and Hanff (1989, p. 40) that shows no central bulge and only rather small LBs.
In all collections of Hy. aesculi studied by the first author in the fresh state, the characteristic central swelling of the living spores was noted in a major part of the freshly ejected spores (see Figures 16(i), 16(m); 18(n)). A central swelling is also obvious in some or all spores of fresh collections drawn by B. Grauwinkel (personal communication) from Fallingbostel (Niedersachsen), B. Fellmann (personal communication) from Starnberg (Oberbayern), B. Declercq (personal communication) from Wachtebeke (Eastern Flanders), and Beyer (1998) from Bayreuth (Oberfranken). This swelling is further evident in almost every spore of a British record photographed by C. Yeates (personal communication, Figure 18 (p)), and in one from Asturias by E. Rubio (personal communication, Figure 17(c)-(d)).
However, the peculiar spore shape of Hy. aesculi is state-dependent and well recognizable only in living spores. The loss of the spore bulge under treatment with KOH is illustrated on Figure 18(n)-(o). Apart from the holotype, two specimens of Hy. aesculi were studied in the dead state only (Figure 14-15), and these were first thought to deviate from typical Hy. aesculi in ascospores devoid of a central swelling. From the observed influence of spore turgor on spore shape; however, we assume that they also possessed a central swelling in the living state. In  the holotype of Hy. aesculi (Figure 13), a few of the fusoid spores still showed the median swelling characteristic of the living spores. In conclusion, one of the most diagnostic markers to separate Hy. aesculi from Hy. albidus is partially or entirely obscured in herbarium material.
The medullary excipulum in the central part of the receptacle of Hy. aesculi was found to be composed of a rather dense, upwards oriented textura prismatica-porrecta, unlike Hy. albidus and Hy. fraxineus for which we have always observed a loose t. intricata. However, this feature was only tested in one collection (H.B. 9701) and needs further study.
All presently known European records of Hy. aesculi grew on leaves of Aesculus hippocastanum. The species fruits between July and October, predominantly in August. A distribution map under the name Lanzia aesculi for W-Germany is presented in Krieglsteiner (Krieglsteiner 1993, pl. 744), with 13 grid squares distributed in Niedersachsen, Bayern, and Baden-Württemberg (including Northern Switzerland). One of the two examined Japanese specimens on blackened Aesculus petioles (as Lanzia sp.) is here referred to Hy. aesculi ( Figure 15, 17 (p)-(s)) because of the absence of croziers and the lack of crystals in the stipe base; the other is the type of Hy. honshuanus (≡Lambertellinia scutuloides) and deviates from Hy. aesculi in having croziers.
Nomenclature, misidentification as Hymenoscyphus albidus. Authors appear to have previously merged Hy. aesculi under Hy. albidus. This is undoubtedly true for reports on petioles of Aesculus from Great Britain under that name (FRDBI). However, we here provide arguments that two different species of Hymenoscyphus occur on that substrate. They were formerly treated in the Sclerotiniaceae under the names Lan. aesculi (Velen.) Svrček (1985) and Lambertellinia scutuloides Korf and Lizoň (1994), but are transferred to Hymenoscyphus in the present paper based on their morphological similarity to Hy. albidus and their overall characteristics suggesting this genus. The latter is here given the new name Hy. honshuanus, and differs from Hy. aesculi in the presence of croziers, also in wider ascospores, yellowish apothecia that do not turn dark with age, longer stipes, and narrower ectal excipular cells. Phillips (1887, p. 138) distinguished between typical Hymenoscyphus albidus (on petioles of Fraxinus) and Hy. albidus var. aesculi (on petioles of Aesculus), based on a collection from Shobdon Court, Herefordshire, on the basis of larger, more frequently clavate spores (20-23 × 4-5 µm). Phillips' brief diagnosis does not mention a darkened substrate, but his remark 'the stem is often brown at the base' in Hy. albidus suggests that this character was also present in his specimen on Aesculus. Massee (1895, p. 260) and Saccardo (1889, p. 254) merely copied Phillips' protologue. The type material has apparently never been re-examined. Judging from the clavate spores this might well be a synonym of Velenovský's He. aesculi. A spore length around 23 µm was only very rarely seen in the present study, but falls in the wide range given by Svrček. Records on petioles of Aesculus under the name Hymenoscyphus albidus in the database of the British Mycological Society very probably also belong to Hy. aesculi. Also collections on leaves of Aesculus indica in Northern India and Nepal reported by Thind and Singh (1969) and others might belong to Hy. aesculi (or Hy. honshuanus), but the ascus base was not studied for the presence of croziers (see discussion under Hy. albidus).
He. scutula var. aesculicarpa Syd. was described from capsules of Aesculus and probably represents a member of the Hymenoscyphus fructigenus-complex. Dennis (1964, p. 64) examined the type material and also further British specimens on Aesculus capsules and petioles. He concluded that all of them belong in the scope of the collective species He. scutula, a herbicolous taxon which includes, in his opinion, also the foliicolous Hy. caudatus. Since Dennis did not mention any blackening of the substrate, the studied specimens on petioles seem to belong to Hy. caudatus rather than Hy. aesculi.
Hymenoscyphus vacini (Velen.) Baral  Apothecia. Moist (0.5-)0.7-1.5(-2)((-3)) mm diam.; disc whitish or pale cream, turing yellowish-ochraceous, slightly concave to flat, margin thin, ± smooth, exterior whitish to pale cream, often densely covered by very fine brownish fibres; stipe 0.2-0.5(-0.8) × 0.2-0.4 mm, concolorous with receptacle, partly translucent, mostly strongly blackened at the base, erumpent from sclerenchyma-phloem layer; dry yellowish-ochraceous. Asci *(90-) 100-115(-122) × (9.5-)10. 5-12(-12.8 Characterization. Hymenoscyphus vacini differs from Hy. albidus s.l. in rather large ascospores that contain a few medium-sized and many small LBs in the living state, also in apothecia with brownish external fibres, a comparatively short stipe with a dark brown base, and a strict occurrence on entirely skeletonized leaves of Acer. Nomenclature and circumscription. Spooner (in Kirk & Spooner 1984, p. 570, fig. 11) gave a detailed description based on three British records on leaves of Acer pseudoplatanus. For one of them, he stated the substrate to be skeletonized. In a section of the stipe base and the vein beneath, he figured a small but distinct, black internal stroma that surrounds the vascular bundle and the sclerenchyma of the vein. Although Spooner noted that 'there is no clear stromatic development visible on the substratum surface', the dark brown external appearance of the veins in the here reported specimens is clearly due to the black internal stroma, which is located below the cortical parenchyma. With its demarcation line surrounding both sides of the sclerenchyma (Figure 22), this stroma corresponds to the type of pseudosclerotium observed in Hy. aesculi. Spooner saw a close resemblance in spore shape with typical species of Hymenoscyphus, but he followed the current opinion to consider a dark stroma, though superficially invisible, as a family character of the Sclerotiniaceae (today Rutstroemiaceae). Because of the ectal excipulum with its thin-walled prismatic cells covered by a cortical layer of brown-and rough-walled hyphae and the hyaline spores, Spooner transferred the species to Lanzia. Independently hereof, also Svrček (Svrček 1985, p. 182) placed the species in Lanzia. When doing this superfluous combination, Svrček referred to a manuscript in press on that species ('New or less known Discomycetes. 14. Čes. Mykol. 40', 1986).
However, neither Lan. vacini nor Lan. aesculi, to which he also refers, is mentioned there. Maybe he refrained from this report when being aware of Spooner's contribution to Lan. vacini. His illustration of the holotype (Svrček 1985, pl. 17 fig. 5) fits very well the present specimens.

Remarks on teleomorph and anamorph features of the treated species
Apothecial colour Except for the blackish base of the stipe and sometimes some brownish dots on the receptacle, fresh apothecia of Hy. albidus, Hy. fraxineus, and Hy. aesculi appear as macroscopically pure white, which is reflected by the specific epithet of Hy. albidus. The white colour is astonishing since, under the microscope, the surface of the receptacle and stipe in Hy. albidus and Hy. fraxineus was often covered by a pale to bright golden-yellow-ochraceous, though very thin exudate.
In the dry state, the apothecia attain also macroscopically a pale to bright yellowish-cream to orange colour, which is partly due to this exudate, but also to oxidation of the VBs inside the paraphyses and in the cortical hyphae that cover the ectal excipulum. For instance, Svrček (1985, p. 132) stressed that fresh apothecia of Hy. albidus turn rust-coloured when bruised, a colour change for which these vacuolar contents are responsible.
In Hy. aesculi a striking colour change of the receptacle, particularly the hymenium, towards olivaceous-grey to blackish was observed in rehydrated material ( Figures  16(g), (j)-(k), 17(j)-(l), 18(o)-(p), 18(c)-(d)), which was not found to be mentioned in the literature, except for those specimens from Aesculus indica misidentified as Hy. albidus by Sharma (1991). Svrček (1989, p. 73) merely stated the pure white apothecia did not turn reddish. The bright olivaceous-brown pigment (Figure 11(f)) is located in the cytoplasm, especially in the upper part of the paraphyses, and likewise seems to originate from an oxidative chemical change of the VBs in the dead state.

Size
Apothecial disc diameters were found to vary strongly within a species. In those on leaves, the diameter seems Table 3. Survey of some selected ecological and morphological features of the five here treated foliicolous species of Hymenoscyphus with a black stroma (ascus and spore size in Asian Hy. fraxineus after description in Hosoya et al. (1993)   to depend mainly upon whether the apothecia emerge from a thick petiole or from thin veins. Although Hy. vacini grows on thin net veins of skeletonized leaves, this species is able to produce rather large apothecia.
In Hy. fraxineus a tendency towards larger apothecia in comparison with Hy. albidus is confirmed in the present study (see Tables 1 and 2). A clear trend towards larger apothecia in Hy. fraxineus (maximum 3-7 mm diam.) was reported by Kowalski (2012) in a table, while Queloz et al. (2011) stated that they could not observe any difference in apothecial size between the two species. The disc diameter of Hy. albidus is given in the literature (fresh or rehydrated) as 1-2 mm (Desmazières), up to 2 mm (White, Dennis), 1-2.5 mm (Arendholz), 1-3 mm (Breitenbach & Kränzlin), 1.5-3 mm (Boudier), or 2-4 mm (Srvček). Stipe length is reported as about 1-1.5(-2) mm. This is in concordance with our observations, and gives evidence that this species rarely exceeded 4 mm in diam. Only Patouillard indicated a larger size (3-5 mm), and also   fig. 1) observed diameters of up to 5 mm in a collection from the Bretagne. Hy. fraxineus may reach a size of about 8 mm as a maximum, but is often also found with rather small apothecia.
Online photos of Slovenian records under the name Hy. albidus (by N. Ogris, Ljubljana, http://www. zdravgozd.si/prirocnik/zapis.aspx?idso=319accessed 23 May 2014) show exceptionally long stipes with redbrown lower half. Possibly, the long stipes result from keeping the collection in a moist chamber in a dark room.

Stroma
A character which has all here treated species in common concerns the dark brown or black, melanized pseudosclerotial plate outside of the sclerenchyma of petioles, rachises, and veins that they inhabit. This stromatized host tissue was reported already in the type of Hy. albidus by Desmazières (1851), who thought that it might be an Asteroma, an anamorph connected mainly to members of the family Valsaceae (Diaporthales). However, there is no doubt that the stroma or pseudosclerotium belongs to the Hymenoscyphus.
Sections through petioles show that the apothecial initials are formed from mycelium inside the pseudosclerotia (Figures 3(a), 3(e), 8(f)). In pure culture on modified Seed agar medium, 'at first a whitish-grey mycelium was formed, which changed soon to dark brown or black and pseudostromatic areas with a hard surface were developed' (Pham et al. 2013, p. 447). Contrary to Svrček (1985), who stated that not all apothecia of Hy. albidus emerge from the blackened areas ('epidermis'), we have never seen apothecia inserted on unblackened parts of the substrate in any of the here treated species.
What appears dark brown to black by macroscopical view is a thin layer (demarcation line or pseudosclerotial plate) in the outer region of the host tissue, which completely surrounds the infected parts beneath, and which separates them laterally from uninfected regions. This dark pigment of the pseudosclerotial plate completely fades when placed in eau de Javelle for about 5 min, as is generally observed with melanized pigments.
In all these foliicolous species, the pseudosclerotial plate develops beneath the epidermis within the largecelled cortical parenchyma, often at the border to the smaller-celled, thick-walled sclerenchyma. During winter the epidermis and cortical parenchyma and also the phloem are degraded and may entirely disappear ( Figure  17(m)-(n), 17(s)), but remnants of brown cortical parenchyma or phloem are partly still present when apothecia are formed (Figures 11(k), 19). Only in rare cases the lightcoloured epidermis can still be found loosely attached to the dark stroma beneath by covering large areas of the rachis (Figure 11(c)). The dark stromatic colour originates from black-brown hyphae that form a plectenchyma (textura epidermoidea) on both sides of the host cell walls of the large-celled cortical parenchyma and the outer cells of the sclerenchyma (Figure 3(e)). The lumina of the host cells are densely filled with hyaline or pale brown hyphae (Figures 3(d), 3(f)-(g), 8(i)-(k), 11(h)). The same is true in the host tissue below the apothecial stipe (Figures 3(a), 3 (e), 18h). Cells of the sclerenchyma distant from the demarcation line are only sparsely filled with hyphae, lining the inner face of their walls (Figure 8(g), 18(i)).
In Hy. aesculi and Hy. vacini also the inner border of the sclerenchyma is frequently delimited by a black pseudosclerotial plate when viewed in cross section (Figures 17(m), 17(q)-(s), 19(b), 22). However, also those cells below the inner pseudosclerotial plate contain small amounts of hyphae.
Demarcation lines that delimit the pseudosclerotium from uninfected tissue are infrequently seen in Hy. albidus and Hy. fraxineus when studying cross-cuts of rachises. Especially when several infections took place within a single leaf, the black lines frequently cross outer or inner parts of the mark parenchyma (Figure 11(i)). Also Gross, Zaffarano et al. (2012) observed up to eight different infections and mating type locus (MAT) genotypes in a single rachis colonized by Hy. fraxineus, although it looked homogeneously blackened from outside. A limited (insular) pattern of pseudosclerotia more obviously permits to conclude that a multiple infection of a single rachis took place.
The present study suggests that differences in the extension of the pseudosclerotium occur between Hy. albidus and Hy. fraxineus. The apothecia of Hy. albidus often grow on sharply isolated (insular) black areas on the otherwise pale straw-coloured rachis, though in some specimens, including the type, the stroma covers a main part of the rachises. Hy. fraxineus often stains more or less the entire rachis of a leaf in black, or a major part of it, though also not rarely only insular areas. In Hy. aesculi more or less the entire petiole was found to be occupied by the pseudosclerotium.
Above the black stipe base a constriction is often seen where it intergrades with the whitish main part of the stipe (Figures 7(n), 8(c), 8(f)). The black basal part of the stipe represents the primordium, from which the development of the stipe and later the receptacle starts. The primordia can be seen as minute blackish spikes on the stromatized rachises. Apparently they appear only in late spring shortly before apothecia are formed.
The black pseudosclerotial plate is not yet present when the wilted leaves fall to ground in autumn but is formed prior to the winter season (see Gross & Holdenrieder 2013). Since the fallen rachises lie loosely on the forest floor, they completely dry down during dry weather. The authors found that the mycelium in the pseudosclerotia survives at least 92 days in the dry state, while in nature under repeated drying and rewetting the pseudosclerotia can endure at least 2 years.
Croziers and simple septa Differences in the ascogenous hyphae among closely related species play an important role in the taxonomy of ascomycetes. A couple of papers have drawn attention to this peculiarity, such as Huhtinen (1990) for Hyaloscypha, Baral (1996, p. 255) and Hengstmengel (1996, p. 192) for Hymenoscyphus. Nevertheless, croziers and simple septa are still being frequently neglected in many groups.
Literature reports on the mode of ascus base development in Hymenoscyphus species with a pseudosclerotium are sparse. According to White's (1944, p. 608, figs. 19-24) precise illustration of Pl. Crypt. N. France, éd. I. n°2004 (FH), the asci in the lectotype of Hy. albidus arise from simple septa. White's statement 'not originating from croziers' in the description allows to assume that he saw this feature also in the examined isolectotype collection in FH (éd. II. n°1604). Besides Arendholz (1979), White's report is the only known to us from the past century, which mentions the type of ascus development in this species.
V. Ruiz-Badanelli (in Hairaud 2010) and Zhao et al. (2013) presented photographs of croziers of Hy. fraxineus (as Hy. pseudoalbidus). Hairaud gave also a photo of the simple-septate ascus base in Hy. albidus for his collection from July 2012, following personal communication with H. O. Baral. However, his note that he saw the feature also in the 2007 sample must be in error. A statement by Otto and Wagner (2011) about the difference in the ascus base between the two species originates also from such personal communication with the first author of the present paper.
Hy. honshuanus (≡Lambertellinia scutuloides) is here found to possess croziers (Figure 12(b)), although the protologue states the contrary. Korf and Lizoň (1994) observation was not accompanied by an illustration; however, it remains open whether this observation was erroneous, or the type collection represented a mixtum.
Homothallism, monokaryon, relative DNA content of nuclei The cytological background of the presence vs. absence of croziers is still little understood. In Hy. albidus the absence of croziers is correlated with homothallism and with the absence of an anamorph state (see below under 'Anamorph and life cycle' section). Data on the karyological situation in the ascogenous hyphae of the here treated pseudosclerotium-forming species of Hymenoscyphus were presently not available. Berkson (1966) observed in Chaetomium that species with croziers had dikaryotic ascogenous hyphae whereas others with simple septa were monokaryotic. Also Weber (1992, p. 68) observed monokaryotic ascogenous hyphae in species of Helotiales that lack croziers, although only in a smaller part of them, while in others monokaryotic and dikaryotic cells occurred mixed in the same ascogenous system. Nuclei of monokaryotic cells thereby had the double DNA content compared to those of dikaryotic cells of the same species. In species in which the ascogenous hyphae possess croziers, Weber observed mainly dikaryotic cells. Zickler et al. (1995) reported 'uninucleate croziers' in mutant strains of Podospora anserina which normally were dikaryotic. However, their microphotos (Figure 4(a)-(d)) show simple-septate ascogenous hyphae with a partial presence of hook-like protuberances which did not fuse with the cells below. The complete croziers drawn for the uninucleate mutants on their diagrammatic survey ( Figure 5(b)-(c)) are not seen on their micrographs.
In her study on the relative DNA content of nuclei in vegetative cells, Weber (1992) revealed for Hy. albidus a value of 4× of the basic value (1×), the lowest found in the Helotiales. Within Hymenoscyphus, this value of 4× was not rarely encountered in her study (e.g., in He. scutula and Hy. vacini), but values of 2×, 3×, 6×, and rarely 1× also occurred. A correlation of these values with the presence or absence of croziers could not be established; however, species with croziers as well as simple septa occurred in quite equal frequency within each value, except for those species with values of 1× (simple-septate) and 2× (croziers), which were too sparsely encountered in order to draw a conclusion. In any case, it would be interesting to measure the relative DNA content of the two so far not investigated species of Hymenoscyphus with a pseudosclerotium, Hy. fraxineus and Hy. aesculi. A survey on the known genome size values in fungi are found in Kullman et al. (2005).
A correlation between croziers and apothecial size similar as in Hy. albidus and Hy. fraxineus was noted in two further species pairs of the Helotiales (Baral in Weber 1992, p. 110, 118): (1) Calycina herbarum (without croziers) and C. aff. herbarum (with croziers and larger apothecia), a very similar species which was currently confused with Ca. herbarum, and which inhabits a comparable spectrum of herbaceous stems; (2) Heyderia pusilla (without croziers, on needles of Pinus) and Hy. cucullata (=Hy. abietis, with croziers and larger fruit bodies, on needles of Picea). Within both species pairs the relative DNA content did not differ (Calycina 2×, Heyderia 6×).

Apical ring
The morphology of the ascus apical thickening in the dead expanded state when stained with iodine is usually very similar among the true members of the large genus Hymenoscyphus. Its characteristic morphology was addressed by Baral (in Baral & Krieglsteiner 1985, p. 119), and later referred to as Hymenoscyphus-type (Baral 1987, p. 126, fig. 10-12, Verkley 1993. The ascus apex in the Sclerotiniaceae and Rutstroemiaceae (Sclerotinia-type) sharply differs hereof and easily permits distinction with the light microscope between sclerotiniaceous/rutstroemiaceous and hymenoscyphoid taxa in most cases (see Baral 1987). White (1944) and Svrček (1985) illustrated the Hymenoscyphus-type of apical ring in Hy. albidus. However, it was perhaps more because of a similar spore shape that White believed in a very close relationship to He. scutula and Hy. caudatus. Ascospore size and shape Heteropolar, subapically rounded and with a lateral hooklike protrusion, basally ± pointed ascospores are typical of many of the true members of Hymenoscyphus. This characteristic spore type was called 'scutuloid' by Baral (in Baral & Krieglsteiner 1985) according to the situation in the representative species He. scutula, but was rarely also referred to as virguliformis, which means in Latin commashaped, e.g., by Patouillard (1885, p. 173), who erroneously wrote in French 'spores virgultiformes' instead of 'virguliformes'. The spore size in the type material of Hy. albidus was given as minimum 15 µm by Desmazières, 11-17 × 3.5-4 µm by Nylander (1868), 15-18 × 3.5-4 µm by Karsten (1871), 15-18 × 3-4 µm by Massee (1895), and 13-17 × 4-5 µm by White (1944).
Hy. fraxineus is said to have slightly longer ascospores, but variation and overlap in this feature forbid recognition of the two species on Fraxinus from spore length alone. Queloz et al. (2011) gave a graphic representation of length values of some selected specimens. Accordingly, spore length ranges were  µm in Hy. fraxineus, both possibly evaluated from dead spores. The tendency to longer spores in Hy. fraxineus is confirmed from the specimens studied by us. Our spore length values are about 0.5-1 µm over those of Queloz et al. (2011), probably because of a slight shrinking effect between living and dead spores. The longest freshly ejected spores that we measured in Hy. fraxineus were almost 25 µm long.
The characteristic spore shape of Hy. aesculi with a swollen middle part is only present when living spores are studied in a water mount. After releasing the spore turgor by heating or applying killing agents, the spores shrink and more or less completely loose this inflation (see Figure 18(n)-(o)). Also inside the turgescent asci, the spores are distinctly narrower and less swollen in their middle part because of the ascus turgor. When discharged in a water mount, the spores swell to their typical shape within a few seconds.

Ascospore contents
Published illustrations of living spores of the here presented species are rather rare. Particularly freshly ejected living spores, but also those inside fully turgescent asci, show a regular pattern of larger and smaller oil drops (LBs). Hy. albidus, Hy. fraxineus, and Hy. aesculi possess quite the same pattern of rather large, globose LBs (2-3.5 µm diam. in Hy. albidus and Hy. fraxineus, 1.5-2.5 µm in Hy. aesculi), about 1-4 in each half, surrounded by small ones. The spores of Hy. vacini consistently contain only medium-sized LBs (~1-2 µm diam.) among many small ones. Patouillard's (1885, pl. 382) drawing, though rather small and inexact, appears to be almost the only published documentation of living mature ascospores within the Hy. albidus aggregate, showing about four rather large globose drops in each spore. Two recent illustrations of living spores are found in Hairaud (2010) and Gross, Holdenrieder et al. (2014, fig . 1h). Even Boudier (1908, pl. 492), who consequently studied ascomycetes in the fresh state, illustrated the living spores of Hy. albidus with only small-to medium-sized drops, the latter about 1.2-1.8 µm diam., though with a rather regular pattern. Perhaps theses spores were not fully mature.
The original lipid pattern can often also be observed in herbarium material, but it requires caution to select those spores which possess enough maturity, and to avoid all those which are overmature or in which the LBs have fused to irregular aggregations. Actually, most reports in the literature concern dead spores that contain irregular rather variably shaped aggregations.
Ascospore reports such as 'often with a large oil globule' (Dennis 1956, p. 93) or 'often with 1-3 oil drops' (Arendholz 1979, p. 80) are due to fusion of the original oil drops during the influence of lethal mountants or the natural death of spores. Surprisingly, White's 1944, fig. 24) drawing of the type of Hy. albidus shows a rather regular spore content with a 'granular', multiguttulate sporoplasm, the largest LBs exceeding not even 1 µm diam. This is astonishing, since oil drops usually fuse in dead material, but never split into small ones. The explanation for this appears to be that White (1943, p. 137) followed a method described by Martin (1934, p. 264), who placed fungal fragments in alcohol before applying KOH and phloxine. We made a test with a specimen of Hy. fraxineus: the large oil drops in the spores indeed disappeared and only the small ones remained visible. Apparently it is the alcohol that dissolves the large oil drops (see Figure 10(h 3 )→(h 4 )).

Ascospore sheath and setulae
Distinct setulae at the spore ends were not observed in any of the species treated here, while a delicate sheath around ejected spores was occasionally seen, though not yet in Hy. honshuanus and Hy. vacini. Because of its fragile nature, it is a rarely reported structure. When studying herbarium material, it remains inevitably undetected since it does not detach from the wall when dead spores are rehydrated, due to the absence of cell turgor. The sheath can only be observed when the spores are ejected from living asci in a water mount: after discharge, they rapidly take up water and swell, and the inelastic sheath bursts and eventually floats besides the spores. Gross, Holdenrieder et al. (2014, fig. 1m-n) reported a thick, hyaline to pale brown mucilage secreted from the spores of Hy. fraxineus, by which they are thought to adhere to the leaf surface after wind dispersal. This mucilage was observed on overmature (1-septate or brown) spores during. In fact, the mature spores are hyaline and aseptate during ejection, and soon dehydrated after dispersal, including their mucilage. Possibly the sheath observed by us covers a thin mucilage that gets more abundant during germination and affixes the spore to the leaf when the appressorium is formed.

Overmature ascospores, pigmentation
Apparently all true members of Hymenoscyphus eject hyaline, non-septate ascospores. However, at an overmature stage of development, the spores frequently get 1(-2)septate and their wall may turn pale to light brown and often also warted. Because brown or septate spores of a Hymenoscyphus are never forcibly ejected, such spores in this genus can only be found inside dead asci or outside asci, but never inside living asci (see also Baral 1992, p. 375).
The presence of pigmented spores has been overestimated as a generic character in the Helotiales. As pointed out by Galán and Baral (1997, p. 61) and Hengstmengel (2009, p. 271), this feature of overmature spores is quite a common character in Hymenoscyphus, though it is seen only inconsistently and sometimes very rarely in a given species. The proportion of brown spores in a preparation depends on the senescence of the apothecia. For some reason, perhaps unfavourable field conditions, some populations produce many such brown spores when senescent, while in others none or only a few can be found.
Spore germination is observed in both hyaline and brown spores. Possibly, the hyaline spores directly infect the living leaves, whereas the brown spores are resting spores being able to survive a longer time. Brown germinating spores were figured by Hosoya et al. (1993) for Hy. fraxineus (as Lam. albida), and by Kowalski and Holdenrieder (2009, fig. 1g). In the present study, brown spores were only two times seen in Hy. fraxineus (Figure 10(f)), once in Hy. honshuanus, and once in Hy. aesculi.

Ectal excipulum
In sections of living specimens of Hy. albidus, the ectal excipulum is a textura prismatica with comparatively thin though slightly gelatinized walls. In dead specimens, or when adding killing agents such as KOH, the cells shrink especially in width, and the gel between the walls becomes more obvious due to imbibition. The shrinking effect was probably the reason why Svrček (1985) found the cells to be only 4-5 µm wide, though he obviously did not see any gelatinization. Likewise, Arendholz (1979), who classified the ectal excipulum of Hy. albidus as a t. intricata, did not see glassy or agglutinated cell walls. White (1944, fig. 22) found the excipular cells in the lectotype 8-10(-15) µm wide, but did not mention or clearly figure a gel between the cell walls.
In Hy. fraxineus strong variation in the width of the excipular cells was noted, which appears to depend mainly on the development stage: in younger apothecia the ectal excipulum forms a t. prismatica-porrecta in both stipe and receptacle, with a cell width of *~6-10 µm, whereas in older and larger apothecia it forms a t. prismatica-angularis with *~10-30 µm wide cells.

Hair-like protrusions
Svrček (1985) mentioned the outside of the receptacle of Hy. albidus to be finely downy, the stipe almost tomentose in the lower part. Also Dennis (1956, p. 94) described and illustrated 'short, slender, obtuse hairs about 10-15 × 2 µm' on the excipular surface in this species. We here describe them in Hy. albidus and Hy. fraxineus (Figure 9(l)-(m)), but saw a pubescent stipe also in Hy. aesculi. These protrusions were not present in each of the studied samples of these species, however.

Vacuolar bodies
In all here treated species of Hymenoscyphus, refractive vacuolar bodies (VBs) occupy most of the lumen of the terminal cells of the living paraphyses and partly also of the cortical cells of the marginal ectal excipulum. In young paraphyses, they are globose or somewhat angular, but soon fuse and finally form very elongate vacuoles (see also Baral 1992, p. 363). In the dead state, they are distorted or entirely disappeared. Also KOH applied to living paraphyses irreversibly makes them disappear (Baral 1992), and that without provoking any colour reaction. Staining agents when added to a water mount, provoke a distinct stain to VBs: CRB slowly stains them turqoise, and IKI light yellowish to bright red-brown. In IKI mounts, many minute red-brown granules may extrude into the surrounding medium, but this effect was only inconsistently obtained in Hy. albidus, Hy. fraxineus, and Hy. vacini, and not in Hy. aesculi.
A colour change of the VBs from hyaline towards olivaceous-brown was repeatedly noted in Hy. aesculi, resulting in partially blackish-brown hymenia in herbarium material. In the other species studied, the VBs turned merely yellowish-cream with age. The phenomenon appears to be due to an oxidative chemical process. Arendholz (1979) observed abundant crystals of 'presumably Calcium oxalate' in the lower 1/4 of the stipe. We have regularly seen these crystals in both Hy. albidus (Figure 3(c)) and Hy. fraxineus (Figures 8(d)-(e), 9(h)-(i), (k)), and Zheng and Zhuang (2014) report them for Hy. fraxineus and Hy. albidoides. Although crystals occur also inside some of the cells of the infected rachises, those found in the dense tissue of the medullary excipulum of the stipe base are undoubtedly formed by the fungus, not by the host.

Crystals
The crystals are hyaline, ±rhomboid, and frequently aggregated in druses. They are KOH-inert and do not stain with CRB or IKI. Under polarized light, they are birefringent by changing the direction of the light for about 90° (Figure 9 (i)). Crystals occur also in the living rachises of Fraxinus, inside cells of the radial rays between the vascular bundles. However, these are not arranged in druses. Crystals were not observed in Hy. aesculi and Hy. vacini, nor in any other species of Hymenoscyphus examined by us.
Anamorph and life cycle Hy. fraxineus is heterothallic and produces a Chalara anamorph in pure culture, whereas the avirulent Hy. albidus is homothallic (Gross, Zaffarano et al. 2012) and not associated with an asexual state ; also it displays slower growth on agar media.
Various authors have shown that the conidia of Ch. fraxinea do not germinate. Their perhaps exclusive function as spermatia was suggested by various authors, including Gross, Zaffarano et al. (2012; see also Gross, Holdenrieder et al. (2014)), whereas distribution seems to be exclusively accomplished by ascospores. The whole life cycle from leaf infection via ascospores to production of apothecia lasts only 1 year (Gross, Zaffarano et al. 2012, fig. 5). Yet, the pseudosclerotia in the fallen leaves may produce apothecia also in the second year (Gross & Holdenrieder 2013). Moreover, seeds of Fraxinus may contribute to dispersal: in 8.3% of the investigated seeds DNA of Hy. fraxineus could be detected (Cleary et al. 2012).
According to Gross, Zaffarano et al. (2012), the formation of spermatia in Hy. fraxineus is linked to the heterothallism of that species. The authors found that Hy. fraxineus needs different mating types in order to produce apothecia, whereas Hy. albidus is homothallic (self-fertile) and able to fruit from a single-spore mycelium (A. Gross personal communication). Because of its homothallism, we assume that Hy. albidus does not require complex mechanisms of nuclear division during the dicaryophase which result in the formation of croziers in Hy. fraxineus.
The situation in Hy. fraxineus/Hy. albidus raises the question whether also in other species of Hymenoscyphus the presence/absence of an anamorph is linked to the presence/absence of croziers. Chalara anamorphs have virtually been unknown in that genus before the connection was established by Kowalski and Holdenrieder (2009). Unpublished observations of a Chalara asexual morph in four species with asci arising consistently from croziers support this correlation. Two of them were observed in ascospore isolates by E. Weber: (1) hyaline phialides in a species close to Phaeohelotium imberbe (H.B. 7469); (2) light olivaceous phialides in Hymenoscyphus infarciens (H.B. 7025a, CBS 122016). The other two concern hyaline to pale brown phialides emerging from the ectal excipulum at the flanks of the receptacle (Baral in preparation): (3) in the lectotype of Hy. subferrugineus; (4) in a specimen identified as Hy. calyculus (H.B. 3128). Hy. linearis, however, a species with simple-septate asci, was found to produce a Chalara asexual state in culture.
Size of conidiophores (13.5-18.5 × 4-5.5 µm) and conidia (2.5-3 × 1.5-2 µm) as given by Hosoya et al. (1993) for Japanese material is at the lower end of the range, except for the inflated basal part of the conidiophores which are more inflated (ventricose).

Remarks on ecological aspects of the two treated species on Fraxinus
Altitude Apothecia of Hy. albidus were recorded in the region of the Alps up to an altitude of 1690 m as indicated by the sample from Giswil (Switzerland, NMLU 2309-80) made in 1980 (Queloz et al. 2011). Before Hy. fraxineus started its invasion, Hy. albidus showed a wide but scattered distribution in lowland regions of Europe, where it became nowadays extinct, except in those regions of Western and Southwestern Europe which have not yet been invaded. In 2009 Hy. albidus was still present in montane regions of Switzerland, according to Queloz et al. (2011): in southern parts of the country at altitudes of 600-1075 m only Hy. albidus could be found, whereas Hy. fraxineus was in that year exclusively present in the northern lowlands at 345-570 m, although some Hy. albidus records were made also there at altitudes of 508-695 m. Kirisits (2011, fig. 5) presented for Hy. fraxineus a map of Eastern Tyrol (Austria), with records of ash dieback by July 2010 at elevations of about 650-720 m in the Drava valley near Lienz, whereas the disease did at that time not yet occur in other parts of Eastern Tyrol, including localities at higher elevations (e.g., in the Gail valley). However, observations made in 2014 indicate that apothecia or dieback symptoms occurred already at 1120 m in the Prättigau (Graubünden, H.O. Baral), at 1130 m in the Hürital (Rosalmig, Zug, U. Graf personal communication), at 1150 m in the Gesäuse National Park (Ennstaler Alpen) and 1300 m in Kärnten (G. Koller personal communication), and at up to 1600 m in the Virgental (Eastern Tyrol) at the uppermost altitudinal limit of F. excelsior (T. Kirisits personal communication). This suggests that Hy. albidus is threatened in the entire distribution range of ash in Central Europe. In Eastern Asia the observed maximum altitudinal records of Hy. fraxineus were 1325 m in Japan, 1100 m in South Korea, and 1310 m in Northeast China (Jilin).

Host range
Within Europe, Hy. fraxineus was recorded, apart from the European temperate species F. excelsior, also on the North American F. nigra (Estonia, Drenkhan and Hanso 2010) and on two submediterranean species from Southern and Southeastern Europea, F. angustifolia and F. ornus. Artificial wound inoculation experiments revealed an equally high or slightly higher susceptibility of F. angustifolia to Hy. fraxineus compared to F. excelsior (Matlakova 2009). When inoculated, the fungus also caused symptoms on F. ornus, though less intensely than on the two other species (Matlakova 2009, Gross, Holdenrieder et al. 2014). However, ash dieback symptoms due to natural infections have so far never been observed on flowering ash, neither on woody parts nor on leaves, and this species may be resistant to the fungus (Gross, Holdenrieder et al. 2014;T. Kirisits personal communication). Gross, Holdenrieder et al. (2014) reported the occurrence of apothecia of Hy. fraxineus on petioles of F. ornus, which may indicate that it occurs endophytically in leaves of this ash species.
F. nigra was badly affected with symptoms, F. pennsylvanica only moderately, and F. americana and F. mandshurica were least affected (Drenkhan & Hanso 2010). In its presently known original area of Eastern Asia, the hosts of Hy. fraxineus are F. mandshurica and F. rhynchophylla (partly treated as a subspecies of F. chinensis).
Hy. albidus is known with certainty only from F. excelsior and sometimes F. angustifolia. For two here mentioned records on Acer, voucher specimens have not been preserved. A preserved specimen on a leaf of 'Pyrus' from Belgium was found to grow in fact on Fraxinus. Misidentifications of either host or fungus explain in some cases a seemingly wide host spectrum.
Similarly narrow host spectra are noted in the genus Rutstroemia. The common Rutstroemia longipes is said to be confined to rachises of Fraxinus in North America (R. P. Korf personal communication), and the closely related Rutstroemia luteovirescens is frequent on petioles of Acer in Europe. However, the independence of these two fungal species and the extent of their host spectra have been questioned in the past (see above).

Fruiting on twigs
As an exception, both Hy. albidus and Hy. fraxineus were found fruiting on twigs of Fraxinus in the present study, though in both cases also petioles were inhabited. The record of Hy. albidus (Echternach, LUX 047701) concerns a corticated 1 mm thick twig with blackened bark, while in that of Hy. fraxineus (Luzern, H.B. 9571) the apothecia grew on blackened wood of an almost entirely decorticated, 5.5 mm thick twig. The rare occurrence of Hy. fraxineus on twigs was repeatedly reported, e.g., by Gross, Holdenrieder et al. (2014, fig . 1d). The rareness of this behaviour is astonishing, since living twigs are regularly invaded by the fungus. Whether also Hy. albidus is able to invade living woody parts is not known.

Phenology
The slightly earlier production of apothecia of Hy. fraxineus in comparison with Hy. albidus is shown in Table 4, based on the specimens listed in our collection data. Literature data suggest that Hy. albidus occurred also in October (e.g., Krieglsteiner 1999, p. 238, 30.X.1997. A fruiting period of Hy. fraxineus during May-October (-November) was demonstrated by Chandelier et al. (2014, fig. 4) using spore traps and real-time polymerase chain reaction (PCR). This earlier start of fruiting plays a role in the dominance of the pathogen over Hy. albidus.

Distribution and replacement
The presently known original distribution of Hy. fraxineus is shown in Figure 24. Whether it covers the entire natural area of its known host trees F. mandshurica and F. rhynchophylla is not known. The original distribution of Hy. albidus appears to comprise all more or less temperate to montane but also oromediterranean regions of Europe, where F. excelsior or F. angustifolia occur. A rather dense distribution at least in Central and Western Europe is evident, based on the presently available data ( Figure 23).
Hy. fraxineus is a neomycete within Europe and threatens other fungi in that ecological niche similarly as introduced plants do with the native vegetation. Our estimation is that particularly host-specific fungi such as Hy. albidus or Cy. fraxinophila are severely threatened by the invasive fungus.
The area-wide extinction of Hy. albidus within Central Europe from ca. 2000 onwards is obvious, due to the mass colonization of the rachises by Hy. fraxineus. As a matter of fact, the species could not be discovered anymore in several countries of Central, Northern, and Eastern Europe, the last known records dating from ca. 1990-2008 (see also below and Figure 23). The species could   so far maintain only in high-montane areas, e.g., in the Alps, and in lowlands outside the present distribution area of Hy. fraxineus, particularly in the southwestern part of France. A history of the spread of Hy. fraxineus along with the extinction of Hy. albidus is given, e.g., by Bakys (2013) and Vasaitis (2013).
Cy. fraxinophila fruited quite abundantly in late autumn (IX-XII) all over Central Europe. A total of 36 records in the first author's database refer to the period 1975-1995, whereas only two concern later years (2006 and 2009, Mecklenburg). Despite a thorough search for the species, it could not be collected in the last years in the area of Tübingen. However, O. Koukol (personal communication) observed Cy. fraxinophila in 2013 at several sites within Czechia, although Hy. albidus disappeared there during the invasion by Hy. fraxineus. It can be assumed that Hy. fraxineus does not infect all living leaves of an ash tree, so there probably remain enough uninfected leaves that fall to the ground during autumn. Infection of these leaves by Cy. fraxinophila is assumed to take place on the ground during the spore discharge period in Sept.-Dec.
The predicted decline of the European ash might also affect lignicolous and other fungi. Jönsson and Thor (2012) discussed the extinction risk for the epiphytic lichen diversity on bark of living ash trees, with 174 species recorded on the island of Gotland (Sweden).

Historical records and present distribution of Hymenoscyphus albidus and Hymenoscyphus fraxineus
When excluding misidentifications, Hy. albidus was predominantly considered as rare in earlier times. The species was sometimes regarded as common, particularly in England and France, but some of these records concern misidentifications on substrates other than Fraxinus. Hy. fraxineus existed originally only in Eastern Asia (Japan, Korea, northeast of China, and far east of Russia). The first records of apothecia within Europe known to us were made in 1999 in Lithuania. With the invasion by Hy. fraxineus, Hy. albidus became extinct in most regions of Europe (see Figure 23).

Japan
Apothecia recorded in Japan on blackened rachises of Fraxinus mandshurica var. japonica were identified as Lam. albida by Hosoya et al. (1993). The very detailed documentation includes a pure culture and an anamorph which fits very well Ch. fraxinea. A sequence (ITS and part of SSU) was deposited at the NBCR (National Biological Resource Center) of the Japanese National Institute of Technology and Evaluation under the number NBRC102368 and was, therefore, overlooked by European researchers. After re-examination of eight specimens collected during 1990-2011, these were found to possess croziers at the ascus base in each of the apothecia tested (see Table 2). Five of these were sequenced and confirmed to belong to Hy. fraxineus (Zhao et al. 2013).
Korea J.G. Han (personal communication) examined 10 records on rachises of Fraxinus mandshurica and F. rhynchophylla morphologically, four of them also genetically. All were found by him to possess croziers, and his sequences refer them to Hy. fraxineus.

China (northeast)
Zheng and Zhuang (2014) recorded Hy. fraxineus on F. mandshurica at four localities in the northeast of China (Jilin province). The authors observed croziers at the ascus base, and their sequences refer these records to Hy. fraxineus.
In a forest with Alnus incana and F. excelsior in Pskov Oblast visited by E. Popov every year since 1996, with Hy. albidus being common, no symptoms of disease were obvious until 2009 when all ash trees showed strong dieback, all dying after 1 or 2 years, also in several similar communities within a radius of 5 km. Other regions (Kaluga and Orel Oblast) did not show the disease so far. Shabunin et al. (2012) recorded Hy. fraxineus southwest of St. Petersburg (Leningrad Oblast) in 2012 by molecular methods.

Poland
Several records of Hy. albidus can be found in Polish publications. For instance, Lisiewska and Bujakiewicz (1976, p. 57) mentioned a specimen (as He. robergei) for the Dębina reserve in Wielkopolska (Central Poland), and Bujakiewicz (2001, p. 116) reported Hy. albidus from the Ostrów Panieński reserve close to Chełmno (Northern Poland) collected during 1981-1984. Molecular data have not been gained from these specimens, but the occurrence on blackened rachises prior to ca. 1990 leaves little doubt about their identity.
Ash dieback symptoms are known from Northern Poland since 1992 (Kowalski 2006), and since 1998 decline of F. excelsior occurred all over the country (Vasaitis 2013). Nevertheless, apothecia were noticed as part of the disease only since 2008 (Kowalski & Holdenrieder 2009) (Kraj et al. 2012;Kowalski et al. 2013).

Lithuania
Six herbarium specimens under the name Hy. albidus, collected on blackened petioles of F. excelsior from the districts of Kėdainiai andVilnius in 1999-2002, were recently reviewed by E. Kutorga (personal communication). He could demonstrate the presence of croziers by microphotos in all of them. The distribution of 'Hy. albidus' in Lithuania based on specimens (not tested for croziers) and literature data was reported by Treigienė et al. (2010) for the districts of Kėdainiai, Ukmergė and Biržai during field trips in 1997-2005, but these observations most likely also concern Hy. fraxineus.
The ash dieback disease was noticed since about 1996/ 97 (Juodvalkis & Vasiliauskas 2002), mainly in Biržai and Panevėžys districts (Northern Lithuania), Kėdainiai, Jonava and Ukmergė districts (Central Lithuania, see also Stepanenkova (2010) and Bakys (2013)). In the following years it spread over the whole country. A DNA sequence that matches Hy. fraxineus (AY787704) was for the first time gained in 2001 by Lygis et al. (2005, as Hymenoscyphus sp. 970, later corrected to Ch. fraxinea in GenBank, V. Lygis/E. Kutorga personal communication) from a stem base of F. excelsior in Biržai district (northeast of Lithuania). Mass production of apothecia was noticed by E. Kutorga from about 1999 onwards including 2013. Rytkönen et al. (2011) isolated Ch. fraxinea from F. excelsior in 2008in NW-and SE-Estonia and in 2009in W-Latvia. Also Drenkhan and Hanso (2010 isolated the fungus in 2009 in SE-Estonia. However, the first indications of ash dieback were as early as 1995 in the northwest of Estonia, where the disease reached most ash stands during 2003, and from where it spread towards the southeast of the country during 2006(Drenkhan et al. 2014. Also in Latvia massive ash dieback occurred since mid-1995s (Vasaitis 2013).

Finland
Karsten (1871, p. 112) stated that collections from his country were unknown to him. The first record of Ch. fraxinea was in 2007 and 2008 in the SW-part of the country (Åland archipelago and mainland), from where the disease spread towards the southeast. The first symptoms of ash disease were in 2000 in the Åland archipelago (Rytkönen et al. 2011).

Sweden
Hy. albidus is known from a single locality near Helsingborg in Skåne (south of Sweden), where it was abundant in 1994 and 1996 (S.Å. Hanson personal communication). The first observation of the disease was in 2001, and in 2004 the entire distribution area of F. excelsior in the south of Sweden was affected . A frequent occurrence of apothecia was first noted in Skåne in 2002 and particularly from 2004 onwards (S.Å. Hanson personal communication).

Norway
Hy. albidus was recorded in 1985 from the southwest of Norway (Bergen) by S. Olsen (K. Homble personal communication), and is today still present along the west coast, according to molecular data by Hietala and Solheim (2011). Possibly it had occurred also in the southeast of Norway before Hy. fraxineus invaded the country, but K. Homble never found apothecia referable to Hy. albidus in the Oslo area before 2009.
The first symptoms were noted in 2006 in the most southern part of the country. The first isolate of Ch. fraxinea was made in SE-Norway in May 2008 in a nursery in the Østfold region, and in 2009 the disease has spread along the entire coastline there (Talgø et al. 2009;Timmermann et al. 2011 Hietala and Solheim (2011) found by molecular methods that Hy. albidus survived along the west coast of Norway. According to Hietala et al. (2013, p. 5), 'The low ascospore number of Hy. albidus detected by real-time PCR in our Norwegian ash stand could relate to such a transition. On the other hand, our data suggest that Hy. albidus is still present in the studied forest.'. Berkeley andBroome (1878, p. 29, No. 1724 27.VII.1946. Clark (1980 recorded Hy. albidus rather often on black patches of Fraxinus petioles (but also on Aesculus) in Warwickshire (West Midlands of England), with the first record in 1967. Ellis and Ellis (1997, p. 137, fig.  600) probably relied on Clark's records when saying that Hy. albidus is common during July-October. A distribution map is shown in Webber and Hendry (2012), which includes also misidentified records, especially on substrates other than Fraxinus.

Great Britain
The British Isles were considered as being free of Hy. fraxineus until January 2012 when ash seedlings were imported from the Netherlands and Germany. At four locations in England they showed the typical signs of the disease (Webber & Hendry 2012). In August 2012 Ch. fraxinea was reported as well from Scotland in a stand of young ashes planted in spring 2009 (Webber & Hendry 2012;Munro 2012). Mainly the eastern part of the country was affected.
An intensive survey carried out since November 2012 confirmed 427 sites with ash trees infected by Ch. fraxinea (status 25.III.2013). Southeastern areas of the British main island are affected the most (http://www.defra.gov.uk/ news/2012/11/09/wms-ash/), but confirmed sites are distributed from Wales to Scotland and Ireland.

Denmark
A couple of collections of Hy. albidus have been made between 1989 and 1999, some of which were DNAchecked (McKinney et al. 2012, T. Laessøe personal communication). According to the molecular analysis, other collections made in 2005-2010 turned out to belong always to Hy. fraxineus, including three sites where Hy. albidus apothecia had been collected previously, and it was concluded that Hy. fraxineus had eliminated Hy. albidus in these regions.
The first records of ash disease were in 2002 near Haderslev (a small port on the Baltic side of the Jutland peninsula), and in 2003 in Zealand and Bornholm (Thomsen et al. 2011;Bakys 2013 Besch, during 1985Besch, during -1990 were reexamined in the present study and found to represent Hy. albidus. According to Roskams and De Haeck (2011), the first record of ash disease was in autumn 2010 in Eastern Flanders (Hainaut), while in Flemish Brabant it was already present as early as 2007. In 2011 the disease was recorded throughout Flanders. For Wallonia (Southern Belgium) the disease was firstly recorded in June 2010

Germany
When Arendholz (1979, p. 79) Baral and Krieglsteiner (1985, p. 121) reported collections from the south of Germany made during 1975-1979 along the rivers Rhein and Donau, also from Franken and Eifel, but were unaware of records from Northern Germany. Beyer (1992, p. 67) described a single collection made in August near Burggaillenreuth (Bayreuth), and believed the species to be rare. G.J. Krieglsteiner (1993, pl. 745) presented a distribution map for W-Germany that shows a scattered occurrence of Hy. albidus within Niedersachsen, Rheinland-Pfalz, Baden-Württemberg, and Bayern, but whether it includes also misidentified samples on other substrates or without pseudosclerotium remains unclear.
L.G. Krieglsteiner (1999Krieglsteiner ( , p. 238, 2004. 601) reported Hy. albidus from 4 localities in the Main area between Schweinfurt and Würzburg (Bayern), and from the Rhön mountains at the edge of Bayern, Hessen and Thüringen with 14 collections. Judging from the phenology data (30.VII.-30.X.) and the occurrence in the years between 1996 and 2004, these records might all concern Hy. albidus. The same is true for nine samples from the Hainich National Park made between 16.VII. and 17.XI. in 2003 made by F. Putzmann, G. Hirsch and A. Gminder (P. Püwert & I. Wagner personal communication). Especially those specimens collected in 2003-2004 should be re-examined for croziers in order to exclude Hy. fraxineus. However, Hy. fraxineus has apparently never been recorded within Thüringen in the years before 2008 (P. Püwert & I. Wagner personal communication); therefore, we strongly suspect that the ash disease was still absent in the Hainich and Rhön region in 2004.
As early as 2002, ash dieback or conspicuous damages on ash were noticed in the northeastern lowlands (Schumacher et al. 2007;Leonhard et al. 2009;Heydeck & Dahms 2012). The first scientific proof of the presence of Ch. fraxinea for Germany was published by Schumacher et al. in 2007 and was detected in material collected from woodlands and nurseries, e.g., from Salzwedel (Altmark, Sachsen-Anhalt). The earliest records of mass fructifications that came to our notice were observed in Sachsen (SN, 2006) and Mecklenburg-Vorpommern (MV, 2007) (see Table 2). In Mecklenburg many ash stands have been cut in order to avoid damage of the timber; therefore, the pathogen was suppressed (T. Richter personal communication).
In Thüringen, the fructifications of Hy. fraxineus became very abundant from 2008 onwards (P. Püwert & I. Wagner personal communication). Although the ascus base was studied by I. Wagner (personal communication) in only a few collections, the absence of abundant fruit bodies in previous years suggests that all these records from the south of Thüringen (around Sonneberg) but also some from the Rhön mountains collected in 2011 belonged to Hy. fraxineus. The presence of Ch. fraxinea was confirmed officially for Thüringen as late as 2009 (Baier et al. 2010;TLWJF 2010).
In Baden-Württemberg ash dieback was first noted in 2006 for a single tree in the eastern part of the country (Gaildorf, Metzler 2010, personal communication). The first detection of Ch. fraxinea dates from 2009 from near Aalen (Schröter et al. 2009, B. Metzler personal communication), and the disease remained almost undetected in SW-Germany before 2009 (Metzler 2009). At four sites in Baden (along the Rhine between Freiburg and Karlsruhe) dieback symptoms could be backdated to 2007 based on repeated yearly infection, followed by proliferation of replacement shoots (Metzler et al. 2012). The first symptoms detected in the region around Stuttgart (Calw, Ludwigsburg, Heilbronn, Schwäbisch Hall, etc.) Table 2).
For Bayern the disease seems to have started in 2008, being recorded at various sites, especially in the north (Main area) and the southeast to east (Leonhard et al. 2009). Apothecia were first recorded in 2009 by P. Karasch in the southeast, and in 2010 in the northeast by H. Ostrow (P. Karasch personal communication).

Austria
Some records of Hy. albidus from Oberösterreich are included in the distribution map of Krieglsteiner (1993, pl. 745). A recent map (Dämon et al. 2009) shows 98 records under the name Hy. albidus, mainly from the north and east of Austria, and merely 7 of Hy. fraxineus. Only a few of the Hy. albidus records date before 2000, whilst the majority was made between 2008 and 2009; therefore, probably most of them concern Hy. fraxineus. For a critical review of these records and a map see Kirisits and Cech (2009). After the invasion of Hy. fraxineus, Hy. albidus has not been observed in Austria anymore .
Ash dieback appears to have invaded Austria from Czechia. The first observations on younger trees date from 2005, while in 2008 symptoms were observed in all Austrian provinces. Ch. fraxinea was first isolated in 2007, and apothecia were first detected in 2009 in great number in various parts of the country (Kirisits & Cech 2011). Breitenbach and Kränzlin (1981) reported Hy. albidus in the cantons Luzern, Obwalden and Nidwalden, and stated the species to be uncommon. Prongué et al. (2004) gave four records for Liechtenstein, and the drawing on our Figure 1 refers to one of them. A recent distribution map of Switzerland (Senn-Irlet 2014) gives four records of Hy. albidus before 1990 and eight between 1991 and 2005, but 26 after 2005. However, the records after 2005 certainly include Hy. fraxineus. Based on molecular identification, Queloz et al. (2011) reported 24 collections of Hy. albidus from the cantons Bern, Glarus, Obwalden, Ticino, Valais, Zug and Zürich, collected mainly in 2009 but also earlier. Most of them originate from montane to subalpine areas in more southern regions of Switzerland (see Figure 23).

Switzerland and Liechtenstein
The epidemic was not evident before 2007 in Switzerland, and diseased trees were noticed in 2008 only in the north and northwest of the country. During 2009During -2011 Hy. fraxineus spread towards the Alps (Engesser & Meier 2012;Pautasso et al. 2013). Apothecia of Hy. fraxineus were recorded by Queloz et al. (2011Queloz et al. ( ) during 2009Queloz et al. ( -2011 In the montane southern cantons Ticino, Glarus and Wallis the disease was still absent during that period.

Italy
Rehm (1893, p. 797) mentioned a collection from Südtirol (leg. G. Bresadola), in which apothecia of obviously Hy. albidus arise from blackened patches on the petioles of Fraxinus (other records in Rehm concern Cy. fraxinophila). A record from Vigevano (Pavia) on rotting petioles of Fraxinus (22.IX.1983) is included in the checklist of fungi from the 'Parco della Valle del Ticino' (Gaggianese et al. 2002).
Symptoms of ash dieback on F. excelsior and isolation of Ch. fraxinea were first reported in May 2008 in NW-Hungary (Szabó 2009).

Slovenia and Croatia
According to N. Matočec and I. Kušan (personal communication), Hy. albidus is not uncommon in montane areas of Croatia. However, Hy. fraxineus was observed to have replaced it since about 2010, though in lower rate at altitudes above 1000 m, while it was abundant also at the few studied colline sites. Because these results are part of a funded project, we refrained from including Croatian records in our map of Hy. albidus.
Ogris et al. (2009)  with the last five nucleotides of the SSU: CATTA]. One exception must be mentioned: though otherwise perfectly matching Hy. fraxineus, a twice sequenced isolate from Czechia (071026.1, FJ429386, GU586921) shows at position 124 the character C which would be typical of Hy. albidus, as well as Chinese, Korean and Japanese Hy. fraxineus.
Apart from the characteristics in the ITS, Husson et al. (2011, table 4) mentioned a difference of four nucleotides in the partial SSU between the two species: positions 56 (C/T), 90 (C/A), 182 (T/C), and 269 (C/G) [the authors counted from the beginning of their sequences that start with CTTGGTCA]. It must be noted that those critical positions which they listed for the ITS bear a few errors: positions 460, 461, 463 should read 460, 462, 463 (bold = gap, position 460 corresponds to pos. 77 in Figures 23-24), and 607 is an error for 507 (=pos. 124 in Figures 23-24). LSU sequences from Hy. albidus were not available in GenBank for comparison with Hy. fraxineus.
Type sequences of Hymenoscyphus albidus Husson et al. (2011) succeeded to isolate ITS rDNA from the two~160 years old duplicates of the type collection of Hy. albidus preserved in the Caen Herbarium. According to their sequences deposited in GenBank, only 67 nucleotides (part of ITS1) were gained from n°1604 (HM193465), but 510 nucleotides (part of SSU, ITS1, and 5.8S) from n°2004 (HM193466). Both sequences fully match previously obtained sequences of Hy. albidus in GenBank, thus confirming the identity of the type material. Five critical positions are located in the region of the 67 nucleotides of n°1604, (77,79,80,87,101;Tables 5 and 6), allowing exclusion of Hy. fraxineus, while the sequence from n°2004 covers all 7 critical positions of the ITS1.

Intraspecific variation in the ITS
Within the many European sequences of Hy. fraxineus variation in the ITS was noted only in on sample (positions 124 and 379, see Table 6). In Hy. albidus three sequences showed variation, which concerns positions 86, 93, and 399. Among the 7 Japanese sequences variation occurred at positions 155, 212, and 443, and among the 6 Chinese sequences only at position 443 (Zheng & Zhuang 2014). The four Korean sequences (KF830850, KF830851, KF830852, KF830853) do not show any variation (J.G. Han personal communication). This infraspecific variation of Hy. albidus and Hy. fraxineus lies in the range of 0-0.5% of the entire ITS1-5.8S-ITS2 region, and the distance between the two species is about 2.2-2.4%.

Does European Hymenoscyphus fraxineus originate from the northeast of Asia?
The phylogenetic analyses by Zhao et al. (2013), J.G. Han (personal communication) and Zheng and Zhuang (2014) have shown that European and Asian Hy. fraxineus cluster together in a clade with rather high conformity and support, sharply separate from Hy. albidus. However, two deviations were noted, which question the asserted invasion of the pathogen from the presently known distribution area in Eastern Asia. According to the results presented by Zhao et al., all seven Japanese ITS sequences fully concur with European Hy. fraxineus, except for two nucleotide positions (80 and 124) at which the Japanese ones show the character of Hy. albidus (80: G instead of a gap, 124: C instead of T; counted when starting with CATTA, see Tables 5-6). The very same peculiarity is shown by all four ITS sequences from South Korea (J.G. Han) and all six from the northeast of China (Zheng & Zhuang 2014). As mentioned above, a sample of Hy. fraxineus from Czechia which was twice sequenced (FJ429386, GU586921) shows at position 124 the character of the Asian samples, while at position 80 it shows the typical gap of European Hy. fraxineus. This sequence further deviates from all sequences of Hy. albidus/Hy. fraxineus at position 379: T vs. C, Table 6).
In the two available sequences of LSU, European Hy. fraxineus (HM145907, Slovenia) deviates from Japanese (NBRC102368) at two positions in the overlapping part. More sequences are needed to clarify whether these two positions are constant markers. In the CAL and EF1-α gene regions, Japanese and European Hy. fraxineus isolates do not consistently deviate from each other, with at best one exception: at position 266 of CAL, Japanese Table 5. Part of ITS1 which includes seven critical positions. At two of them (marked in blue: 80, 124) European Hy. fraxineus deviates from NE-Asian Hy. fraxineus. Positions in red (77,79,87,101,117) concern definition positions against Hy. albidus that apply also for Asian samples (see also Table 6).

Hy. fraxineus (Europe)
GCCCTCT-GGGCGTCGGCCTCGGCTGACTGTGCCTGCTAGAGGACCCTAAATTTTG Hy. fraxineus (NE-Asia) GCCCTCTGGGGCGTCGGCCTCGGCTGACTGTGCCTGCTAGAGGACCCTAAACTTTG Hy. albidus (Europe) GCCCCCCGGGGCGTTGGCCTCGGCTGACCGTGCCTGCTAGAGGATCCTAAACTTTG isolates have A and European G, but one Japanese (TNS-F-12503, AB705208) has also G. Further arguments against an origin from the northeast of Asia were presented by Drenkhan et al. (2014) who showed that the outbreak of the disease in NW-Estonia in 1995 was not associated with the areas where Mandshurian ash had repeatedly been introduced as seeds or seedlings since nearly 150 years.
Variation of Asian Hymenoscyphus fraxineus towards the genome of Hymenoscyphus albidus Various nucleotide positions are presently known at which Asian Hy. fraxineus shows infraspecific variation. These can be divided into two groups: those positions at which Hy. fraxineus does not differ within Europe from Hy. albidus, and those at which the two differ. In the latter group (the definition positions), the Asian sequences show either the character of Hy. fraxineus or that of Hy. albidus, but never a further possibility. In other words: when deviating from European Hy. fraxineus, variation of Asian Hy. fraxineus was always towards the character of Hy. albidus. This strange phenomenon, which reminds of the two consistent deviations in the ITS between European and Asian Hy. fraxineus, might indicate a lineage from the ancient Asian Hy. fraxineus towards the younger Hy. albidus.
Variation in the ITS region within Asian Hy. fraxineus is noted at three positions (155,212,443) which are located outside the definition positions. Variation in the EF1-α gene within Japanese isolates concerns various positions. Among these are 11 definition positions, and all of them show either the character of European Hy. fraxineus or that of Hy. albidus (see Zhao et al. 2013 ,  table 3; gaps not counted, pos. 236 must show a T instead of a gap). Also one position in the CAL gene (244) shows this phenomenon in both Japanese and Chinese specimens, and a further position (49) varies in one Chinese specimen towards Hy. albidus. These variations concern not only transitions but also transversions.

Phylogenetic considerations within the Fraxinusinhabiting species
According to fossil finds, the genus Fraxinus occurred for the first time in the early Eocene(?) in North America. The land bridge between North America and Asia during this epoch allowed it to spread westwards while forming different species. According to molecular data, Asian species such as F. mandshurica are older than European species such as F. excelsior and F. angustifolia (Wallander 2008;Hinsinger 2010). Parallel to this evolutionary progress, different leaf decaying fungi developed on different Fraxinus species, such as R. longipes in North America (see below), Hymenoscyphus fraxineus in Eastern Asia, and Hymenoscyphus albidus in Europe.
The poor genetic variability of both Hy. albidus and Hy. fraxineus in comparison with a rather high variability of Eastern Asian Hy. fraxineus as detected by Zhao et al. (2013) and Gross, Hosoya et al. (2014) might suggest that not only the invasion of Hy. fraxineus but also that of Hy. albidus to Europe started from a small part of the whole Table 6. Variable nucleotide positions of the ITS1 (77-155), 5.8S (212), and ITS2 (355-453) of Hymenoscyphus albidus and Hy. fraxineus from European and Asian samples [position numbers starting with the last five nucleotides of the SSU: 'CATTA']; ***/** = normal case, * = infrequent abnormal case. Definition positions that apply also to Asian samples are marked in red, whereas at the blue markings Asian and European samples of Hy. fraxineus differ, the former matching Hy. albidus. Positions identical to all sequences are omitted, and identical sequences are equally omitted except for a few from Asia. T T -

Hy. albidus
Europe GU586900 infundibuliform collar at the spore base undoubtedly concerns a modification of the setulae which are characteristic of quite a few species of Hymenoscyphus. Due to its very large spores Lam. torquata obviously represents a distinct species, which is here transferred as follows: Hymenoscyphus torquatus (W.Y. Zhuang) I. Kušan & Baral, Basionym: Lambertella torquata W.Y. Zhuang, Mycotaxon 56, p. 41 (Zhuang 1995) Also Lambertella caudatoides Zhuang and Lambertella tengii Zhuang have scutuloid spores and were described in Lambertella because the spores turn light brown when they become median septate (Zhuang 1999(Zhuang , 2002. Undoubtedly they belong in Hymenoscyphus, but their transfer should await careful comparison and redescription, e.g., concerning the ascus base. A further misplacement of stroma-forming discomycetes in the Rutstroemiaceae concerns Dicephalospora Spooner (1987, on twigs) and the two petiole-inhabiting species Lanzia huangshanica Zhuang (1995) and Lanzia aurantiaca (Zhuang) Zhuang in Zhuang and Liu (2007). A collection of D. rufocornea (Berk. & Broome) Spooner from Mauritius was studied (H.B. 6962) and found to have an amyloid apical ring of the Hymenoscyphus-type. The spores are not scutuloid, though slightly heteropolar, and possess round gelatinous appendages at both ends, typical of Dicephalospora. Lan. huangshanica has similar spores but without appendages, but the type of amyloid ring is not discernible on the photos in Zhuang (1995). A pale yellow granular exudate releases a honey-yellow pigment in KOH in Dicephalospora, whereas no such pigment is released in Lan. huangshanica. In a phylogenetic study by Zhuang and Liu (2007), the two taxa cluster in one clade, quite separate from the true Rutstroemiaceae, but also remote from Hymenoscyphus. Possibly Lan. huangshanica will be transferred to Dicephalospora in the future, a view also shared by Zhuang and Liu (2007).
Rutstroemia longipes, a similar though unrelated species on Fraxinus in North America White (1941) emphasized the 'more prominently developed stroma than any other member of the genus' [Rutstroemia]. His precise description of the stroma in section being 'seated directly on the sclerenchyma' matches exactly the situation in Hy. albidus and Hy. fraxineus. However, an acervular conidial state is formed 'more or less associated with the black stromatic lines', with conidiophores and conidia not unlike Ch. fraxinea. Comparable ellipsoid conidia are formed from the germinating ascospores.
The thin apothecial stipe of a very variable length reminds indeed of a Rutstroemiaceae, and the entirely yellow apothecia resemble those of R. luteovirescens (Roberge ex Desm.) W.L. White, a species currently recorded on petioles of Acer but originally described from Tilia and Platanus. Seaver (1951) even synonymized R. longipes with R. luteovirescens. The morphological differences between the two species are actually not very striking, and a genetical comparison has apparently never been done. White (1941) distinguished them by in the fresh state sulphur-yellow vs. greenish-yellow apothecia, which on drying turn dark brown in R. longipes while remaining more or less unchanged in R. luteovirescens, also by slightly larger asci and spores in the latter species.