Additions to Amanita (Amanitaceae, Agaricales) section Arenariae from south-western Australia

Abstract. A recent molecular phylogeny of Amanita recognises three subgenera and 11 sections. Members of subgenus Amanitina are characterised by amyloid spores and a mycorrhizal habit. Section Arenariae falls within this subgenus. Members of this section are known only from southern Australia; they are either sequestrate (secotioid) or agaricoid and lack clamp connections. We describe the following three additional secotioid species: Amanita arenarioides Bougher, E.M.Davison & Giustiniano, A. compacta Bougher, E.M.Davison & Giustiniano and A. pseudoarenaria E.M.Davison, Giustiniano & Bougher, which are separated on macroscopic appearance, spore shape and genetic sequences. We also describe two agaricoid species, namely, A. pupatuju E.M.Davison, Giustiniano, McGurk & E.L.J.Watkin, and A sabulosa E.M.Davison & Giustiniano, which are separated on bulb shape and genetic sequences. We provide expanded descriptions of A. arenaria (O.K.Mill. & E.Horak) Justo and A. griselloides D.A.Reid; we also synonymise A. dumosorum D.A.Reid with A. peltigera D.A.Reid. A revised diagnosis and description of section Arenariae is provided, together with a key to currently recognised member of this section.


Introduction
Amanita Pers. is the largest genus in Amanitaceae, with~1000 described species worldwide; however, only~600 are well defined ). This genus is important ecologically because most species are mycorrhizal with woody plants. It is economically important because some species are comestible, and others are poisonous . The genus is defined by the field character of a membranous or friable universal veil, and by the microscopic and ontogenetic characters of bilateral lamella trama, acrophysalidic stipe trama and schizohymenial development (Bas 1969, pp. 299-333). Like many mushrooms, Amanita spp. are phenotypically variable; consequently, many species cannot be confidently identified using field characters alone; microscopic examination is also essential. Molecular analyses too have been instrumental in both identifying new species and clarifying relationships within the genus Yang et al. 2018).
The family Amanitaceae has been recently revised by Cui et al. (2018), using concatenated datasets of nuLSU (28S nuclear ribosomal large subunit rRNA gene), b-tubulin, ef1-a (elongation factor 1-a), and rpb2 (RNA polymerase II) gene regions. Within Amanita, they recognise three subgenera, namely, Amanita, Amanitina (E.-J.Gilbert) E.-J. Gilbert and Lepidella Beauseign. Most Australian species fall within subgenus Amanitina because they have amyloid spores, a bulb at the base of the stipe, and are likely to be mycorrhizal. Cui et al. (2018)  Phalloideae (Fr.) Quél., section Roanokenses Singer, section Strobiliformes Singer ex Zhu L.Yang, Y.Y. Cui & Q.Cai, and section Validae (Fr.) Quél. These sections can be recognised morphologically by differences in the pileus margin, form of the universal veil, and presence or absence of clamp connections.
Members of section Arenariae are agaricoid or sequestrate, their basidia are thin-or thick-walled and lack clamp connections; agaricoid species have an appendiculate pileus margin and remnants of universal veil on the pileus (Cui et al. 2018, pp. 106, 117). Five species have been described; they are known only from southern Australia. The type is A. arenaria (O.K.Mill. & E.Horak) Justo, a sequestrate (secotioid) species from the south-west of Western Australia (WA). Amanita arenaria is unusual because it has inamyloid spores, whereas they are amyloid in agaricoid members of this section. There are a further five sequestrate species of Amanita known from Australia, and they cluster together in section Amarrendiae (Bougher & Lebel) Zhu L.Yang, Y.Y. Cui, Q.Cai & L.P.Tang, which are in subgenus Amanita, not subgenus Amanitina (Justo et al. 2010;Cui et al. 2018, pp. 19, 44). Miller and Horak (1992) described A. arenaria (as Torrendia arenaria O.K.Mill. & E.Horak) from collections they had made while visiting WA in 1985WA in , 1989WA in and 1991 Miller split two of his collections (OKM 24013, VPI 679;OKM 24018, VPI 412) and deposited duplicates in the Western Australian Herbarium (PERTH). Within T. arenaria, they described T. arenaria f. lutescens O.K.Mill. & E.Horak for material showing yellow stains on the stipe and pileus or slowly turning yellowish on bruising or cutting. Justo et al. (2010) re-examined 10 of Miller's collections of T. arenaria, together with other Torrendia species, and used the nuLSU gene region to determine whether they were closely related to Amanita, as had been suggested by Bas (1975). They showed all Torrendia species nested within Amanita and transferred all species to Amanita. Their results showed that these secotioid species were not monophyletic because they fell into three well-supported clades in different parts of the genus. They also showed that four of the collections identified by Miller and Horak (1992) as T. arenaria were an undescribed species that was morphologically and molecularly distinct, and described it as A. pseudoinculta Justo. One of the collections examined by Justo et al. (2010) was T. arenaria f. lutescens; however, the nuLSU sequence from this collection did not differ from sequences from collections identified as T. arenaria f. arenaria. They also noted that the six A. arenaria collections consistently clustered into two subgroups that differed in spore shape; those from the type locality were ellipsoid to elongate (Q = 1.55-1.72), whereas spores of the others were broadly ellipsoid to ellipsoid (Q = 1.28-1.36). They suggested that this either represented two cryptic species, or there was high genetic divergence among collections from different localities.
While reviewing all A. arenaria collections in PERTH and in private collections, we have noted there are at least three species that are macroscopically similar but can be separated by spore shape and molecular sequences. We describe these in this paper, together with another secotioid species that differs macroscopically from A. arenaria.
There are four agaricoid species that are recognised as being members of section Arenariae; these are A. lesueurii E. M. Davison, A. peltigera D.A.Reid, A. wadjukiorum E.M. Davison and A. wadulawitu McGurk, E.M. Davison & E.L.J. Watkin . In this paper, we describe two additional species. Many more described species may be recognised as being members of this section if widespread sequencing of described Australian amanitas is conducted. Reid (1980) described A. griselloides D.A.Reid as having a layered universal veil on the pileus; however, examination of the type by E. M. Davison found that the superficial layer was formed by hyphomycetes. We provide a revised description in this paper.
The type of A. peltigera was described from South Australia (SA) as having a saccate volva and lacking a partial veil (Reid 1980;. However, examination of collections from SA and the south-west of WA showed that they form a single clade exhibiting variation in both the persistence of the partial veil and remains of the universal veil at the top of the bulb . Reid (1980) also described A. dumosorum D.A.Reid, which he considered to be similar to A. peltigera apart from having a partial veil, lacking a saccate volva, and lacking infrequent large, inflated cells in the universal veil on the pileus. However, examination of the type of A. dumosorum (K(M) 236387) by E. M. Davison found infrequent large, inflated cells in this tissue. The other differences noted by Reid (1980) occurred in collections of A. peltigera ; because these two species cannot be separated by their field characters or micromorphology, they are synonymised in this paper.
The nuclear ribosomal internal transcribed spacer (ITS) region has been used as a barcode marker for species discrimination within the Basidiomycota (Schoch et al. 2012). Hughes et al. (2013) found less than 2% base pair divergence among haplotypes within an individual, and Cai et al. (2014) used ITS sequences to separate Amanita species within section Phalloideae. The ITS region does not appear as useful among Australian species. For example, Davison et al. (2017a) found that ITS sequences fail to separate three species from section Phalloideae, which differ in spore shape and are geographically separated, and within A. peltigera, the base pair divergence among haplotypes from the same individual is between 0.2 and 5.8% . In this paper, we determine the difference between ITS clones from both secotioid and agaricoid members of section Arenariae, to determine whether they show divergence similar to that observed in A. peltigera. This paper has four aims. First, molecular sequences are used to check the placement of the undescribed species and A. griselloides D.A.Reid within section Arenariae, their relationships with one another, and the variation within the ITS region of haplotypes from the same individual and from different collections of the same species. Second, we describe three new secotioid and two new agaricoid species, together with a revised description of A. arenaria and an expanded description of A. griselloides. Third, we synonymise A. dumosorum with A. peltigera. Last, we provide revised diagnosis and description of section Arenariae and a key to currently recognised members of this section.

Taxonomy
The methodology used for describing the macroscopic and microscopic characters largely follows Tulloss (2000). Colour names, including the colour of spores in deposit and other shades of white to cream (designated by the letters A-G), follow Royal Botanic Garden, Edinburgh (1969), whereas colour codes are from Kornerup and Wanscher (1983). In the descriptions of basidiospores (and basidia) the notation [x/y/z] denotes x basidiospores measured from y basidiomes from z collections. Biometric variables for spores follow Tulloss (2000), that is: L = the average spore length computed for one specimen examined and the range of such averages L 0 = average spore length computed for all spores measured W = the average spore width computed for one specimen examined and the range of such averages W 0 = average spore width computed for all spores measured Q = the length:breadth for a single spore and the range of the ratio of length:breadth for all spores measured Q = the average value of Q computed for one specimen examined and the range of such averages Q 0 = the average value of Q computed for all spores measured Author citations follow Index Fungorum (see http://www. indexfungorum.org/Names/Names.asp, accessed 19 February 2021). Herbarium codes follow Index Herbariorum (see http:// sweetgum.nybg.org/science/ih/, accessed 1 February 2021).
Any polymerase chain reaction (PCR) chimeras in ITS sequences were detected using the 'Remove chimeric reads' function in Geneious, using the UCHIME program (ver. 11.0.667, see https://www.drive5.com/usearch/manual/ uchime_algo.html; Edgar et al. 2011). Default settings were used (minimum reporting score 0.3, weight no vote 8, minimum divergence ratio 0.5). The recommended database for chimera detection in fungal ITS sequences was selected as the reference database (Abarenkov et al. 2020). Maximum-likelihood phylogenetic trees were built using MEGA (ver. 5, see https://www.megasoftware.net; Tamura et al. 2011). The best model for each dataset was determined using the Model Function in MEGA. The general timereversible model (Tavaré 1986) with gamma distribution rates was used to determine the placement of the undescribed species in subgenus Amanitina by using the nuLSU gene region. The Kimura 2-parameter model (Kimura 1980) with invariant sites was used for the ef1-a and b-tubulin gene regions. The Kimura 2-parameter model (Kimura 1980) with gamma distribution rates was used for the rpb2 gene regions. The Tamura-Nei model (Tamura and Nei 1993) with invariant sites and gamma distribution rates was used for concatenated nuLSU, rpb2, ef1-a and b-tubulin gene regions. The Tamura 3-parameter model with gamma distribution was used for the ITS gene regions for the secotioid species (Tamura 1992). Bootstrap consensus trees were inferred from 500 replicates.

Placement of undescribed species in subgenus Amanitina
The nuLSU is the only gene region available from GenBank for all type species of sections within subgenus Amanitina. These, together with additional species (Table 1), show that A. arenarioides sp. nov., A. compacta sp. nov., A. griselloides, A. pseudoarenaria sp. nov., A. pupatju sp. nov. and A. sabulosa sp. nov. form clades within section Arenariae with good support (Fig. 1). There are two clades within this section. One well-supported clade contains the secotioid    ), these are considered as species-level clades and are described below (see Taxonomy section). The names of these novel taxa are indicated on the phylogenetic trees, and previous misidentifications are updated.

Additional gene regions
Additional gene regions (ef1-a, rpb2 and b-tubulin) are available for some collections within section Arenariae (Table 2). Phylogenetic analysis of the ef1-a ( Fig  A. sabulosa form a well-supported clade, separate from A. peltigera, A. wadjukiorum and A. wadulawitu.

ITS region
Cloned haplotypes of the ITS region are available for both secotioid and agaricoid species from section Arenariae. No chimeric reads were detected in these sequences.
In the secotioid species, the ITS region is 569-599 base pairs long. Molecular phylogenetic analysis by the maximumlikelihood method shows that GQ925393 and GQ925388 (sequences originally identified as A. arenaria by Justo et al. 2010) are widely separated. All the new species form distinct clades, some with good bootstrap support (Fig. 3). The bootstrap support for A. arenaria is 92% (13 clones from four collections), for A. arenarioides it is 67% (8 clones from two collections), for A. pseudoarenaria it is 77% (20 clones from five collections) and for A. compacta it is 53% (4 clones from a single collection).
As can be seen from Fig. 3, there is considerable variation between the clones; they differ by 0.0-8.1% within an individual, by 0.0-9.6% among collections of the same species, and by 10.3-18.8% among different species ( Table 3).
Examination of the ITS region shows that there is an insertion of up to 20 base pairs (AAAGGAAGCTTCACTTTTGA in MW775323) in ITS1 in all haplotypes of A. arenaria; this is not present in any haplotypes of A. pseudoarenaria. There is a shorter, 15 base pair insertion (AGGGGAATCTTTTGA) in ITS1 in all haplotypes of A. arenarioides. There is a 10 base pair deletion in ITS1 in all haplotypes of A. arenaria and A. arenarioides, which is not present in A. pseudoarenaria and A. compacta.
In the agaricoid species A. griselloides, A. pupatju and A. sabulosa, the ITS region is 582-588 base pairs long. There is considerable variation among the clones: they differ by 0.2-9.9% among haplotypes from the same individual, by 0.7-10.1% among collections of the same species and by 12.2-18.0% among different species (Table 4).
Examination of the ITS region shows that there is a five base pair deletion in ITS1 in all haplotypes of A. griselloides and A. pupatju and a four base pair insertion in ITS1 in all haplotypes of A. pupatju.

Taxonomy Section Arenariae
Cui et al. (2018, pp. 106, 117) stated that members of section Arenariae are agaricoid or sequestrate and lack clamp connections. Since their publication, more species have been recognised as being members of section Arenariae (including those detailed below), enabling us to provide a revised diagnosis and description of this section.

Diagnosis
Basiomata agaricoid or sequestrate; basiodiospores thinwalled or very slightly thickened, smooth, amyloid in agaricoid species, inamyloid in sequestrate species; clamp connections absent in all parts of the basidiome. When basidiomata is agaricoid, pileus with remnants of universal veil, lamellulae attenuate, stipe base variable. When

Description Basiomata agaricoid or sequestrate
When basidiomata agaricoid, pileus plane, margin nonstriate, appendiculate when young, context white or cream, unchanging, or becoming vinaceous-buff when bruised. Universal veil on pileus present. Lamellae adnexed to adnate, white to cream; lamellulae attenuate. Stipe subcylindric, stipe base turbinate and elongating with age or ovoid or conic. Partial veil present. Universal veil at base of stipe indistinct or as soft warts or as small, flimsy free limb. Basidiospores amyloid, thin-walled, smooth. Sub-basidial cell becoming inflated. Clamp connections absent.
When sequestrate, basidiomata hypogeus to emergent, sequestrate and stipitate; pileus subglobose or pulvinate or cylindric, white or cream, yellowing with age; margin inrolled, in contact with stipe. Universal veil on pileus usually absent. Gleba loculate, white to cream. Stipe subcylindric, apex hardly extending into gleba or extending almost to top of gleba. Partial veil absent. Universal veil at stipe base a small cup or absent. Basidiospores statismosporic, inamyloid or very slightly amyloid, wall slightly thickened, smooth. Basidia thin-walled, collapsing early, thick-walled crassobasidia sometimes present. Clamp connections absent.
Known from southern Australia on sandy soil, associated with Agonis, Allocasuarina

Description
Basidiomata sequestrate, secotioid, capitate, very small. Pileus 5-16 mm wide, up to 4 mm thick, white to ivory (B) without surface staining or bruising, globose to pulvinate, surface dry, silky, smooth to lumpy, margin incurved. Universal veil on pileus adnate, thin, membranous flakes, usually missing, white. Gleba white to ivory to cream (B-D), chambers round to elongate. Stipe 3-33 mm Â 3-6 mm, cylindric or narrowing downward, apex domed, white; surface scabrous or smooth; context solid, white to ivory (B), unchanging. Universal veil at base of stipe a small broad membranous cup up to 0.5 mm thick with smooth, dry inner surface, white to cream. Odour not distinctive or of mushrooms or coconut.

Habit, habitat and distribution
Gregarious, initially buried in sand or clayey sand or clayey lateritic gravel or loam, emerging as the stipe elongates. In native vegetation, nearby plants include Agonis linearifolia, A. parviceps, Corymbia calophylla, Eucalyptus marginata, E. patens, Hibbertia sp. and Mirbelia dilatata. Occurs in the Swan Coastal Plain Perth SWA02, Northern Jarrah Forest JAF01, Southern Jarrah Forest JAF02, and Warren WAR01 IBRA subregions (Department of the Environment 2013). Fruiting period is May to August.

Etymology
From the Greek pseudo-meaning 'false', referring to its similarity to A. arenaria.

Notes
This is the cryptic secotioid species, which was discussed by Justo et al. (2010). The nuLSU sequences form a well supported clade, as do ef1-a, rpb2, b-tubulin and ITS sequences that differ from those of A. arenaria (Fig. 1, 2A-C, 3, Table 3). This species is widespread in the southwest of WA. It is, in macroscopic appearance, similar to A. arenaria, but can be distinguished from it microscopically by its subglobose to broadly ellipsoid (Q = 1.11-1.27), not ellipsoid to elongate (Q = 1.54-1.70) spores. It differs from secotioid species in section Amarrendiae, and from A. torrendii Justo because of the absence of clamp connections. On this basis A. pseudoarenaria is described as a new species.
Some collections produce a bright yellow liquid that dries as a yellow stain where they have been damaged by insects (Fig. 6). This is not specific to A. pseudoarenaria because it has been observed on some collections of at least one additional, yet to be described, secotioid species from this section (E. M. Davison, pers. obs.).  Bougher,E.M.Davison & Giustiniano,sp. nov. (Fig. 8,9) Basidiomata sequestrate, secotioid, capitate, very small. Pileus white or ivory white, bruising dull yellow, globose to pulvinate, margin incurved. Gleba chambered, white or ivory. Stipe cylindric, hardly reaching into gleba, solid, white, staining yellow. Universal veil on pileus not seen; at stipe base a small, shallow cup. Spores elongate to cylindric, inamyloid. Clamp connections absent.
Peridium up to 10 mm thick, not continuous, very pale yellow; filamentous hyphae 3-5 mm wide, walls slightly thickened, periclinal orientation, gelatinising; inflated cells not observed; vascular hyphae 6 mm wide, pale yellow. Pileus context consisting of dominant or frequent filamentous hyphae, inflated cells and very infrequent vascular hyphae; filamentous hyphae 4-18 mm wide, with widest hyphae constricted at septa, thin or slightly thickened walls, colourless, gelatinising; inflated cells up to 70 Â 30 mm, thin or slightly thickened walls, ellipsoid or clavate or ovoid or cylindric, terminal or in chains of 2; vascular hyphae 3-5 mm wide, pale yellow. Dissepiments of dominant filamentous hyphae encircling chambers, frequent to infrequent inflated

Etymology
From the Latin -oideus meaning 'like', referring to its similarity to A. arenaria.

Notes
The nuLSU sequences form a well-supported clade, as do ef1-a and ITS sequences that differ from those of the secotioid species A. arenaria and A. pseudoarenaria (Fig. 1, 2A-C, 3, Table 3). This species has been found only north of Perth.

Habit, habitat and distribution
Gregarious, barely emerging from sand. In native vegetation, nearby plants Allocasuarina campestris. Occurs in the Lesueur Sandplain GES02 IBRA subregion (Department of the Environment 2013). Fruiting period is August.

Etymology
From the Latin compactus meaning 'close-packed', referring to the compact, rather stumpy appearance of the basidiomata.

Notes
Only one collection of this distinctive species is available. It differs macroscopically from the secotioid species A. arenaria, A. pseudoarenaria and A. arenarioides because the basidiomes are cylindric, not capitate, and the gleba is thimble-shaped, not globose to pulvinate. Also, the universal veil is absent. Microscopically it differs from A. pseudoarenaria and A. arenarioides because the spores are ellipsoid to elongate (Q = 1.40-1.63), not subglobose to broadly ellipsoid (Q = 1.11-1.27, A. pseudoarenaria) or elongate to cylindric (Q = 1.85-2.13). It differs from secotioid species in section Amarrendiae, and from A. torrendii because of the absence of clamp connections. The nuLSU, ef1-a and ITS sequences differ from those of A. arenaria, A. pseudoarenaria and A. arenarioides (Fig. 1, 2A 3, Table 3). On this basis, A. compacta is described as a new species.

Description
Pileus 22-115 mm wide, up to 6 mm thick, white to very pale clay pink (6A2) with age, without surface staining or bruising, initially convex becoming plane with slightly depressed centre, surface tacky when moist; margin non-striate, appendiculate. Universal veil on pileus adnate, thin, felted to floccose, breaking into small patches over disc, white. Lamellae adnexed to adnate, subcrowded, white, 5-10 mm broad, margin concolorous, fimbriate; lamellulae plentiful to infrequent in several lengths, rounded truncate or truncate or rounded attenuate or attenuate. Stipe 35-75 mm long, 8-40 mm wide, cylindrical or tapering upward or tapering downward, white, surface mealy or floccose or fibrillose. Partial veil superior, arachnoid, white. Bulb 14-35 Â 16-39 mm, conic to tapered, occasionally ovoid when young, white to grey. Remains of universal veil at top of bulb not evident, or as a slight ridge. Pileus and stipe context white, unchanging, stipe solid becoming hollow. Smell none when young, ammoniac when old. Spore deposit white or B (ivory).

Habit, habitat and distribution
Singly or gregarious in sand. In degraded native vegetation, nearby plants include Corymbia calophylla and Eucalyptus marginata. Occurs in the Swan Coastal Plain Perth SWA02 IBRA subregion (Department of the Environment 2013). Fruiting period is April to June.

Etymology
Pupatju means 'to cover the head' in the Nhanda dialect of the Western Australian Aboriginal Noongar language, reflecting the patches of universal veil on the pileus. The epithet is formed as a noun in apposition. Figure 1 shows A. pupatju is a member of section Arenariae.

Notes
There are several species with a white or pale pileus, in which clamp connections are either absent or very infrequent. Amanita pupatju has similar shaped, but longer, spores than does A. preissii (A. pupatju [200/10/8] L = 10.6-13.6 mm, W = 5.2-6.5 mm, Q = 1.80-2.12; A. preissii [340/11/11] L = 8.8-11.5 mm, W = 5.2-6.0 mm, Q = 1.65-2.13), but the pileus, universal veil and partial veil do not age pale cinnamon to saffron (Davison et al. 2017b). It has spores of shape and size similar to those of A. lesueurii ([200/10/5] L = 11.1-12.4 mm, W = 5.5-6.2 mm, Q = 1.90-2.26), but lacks the felted or floccose universal veil on the pileus which becomes vinaceous-buff to grey with age . Amanita pupatju has spores that are wider than those Gilbert & Cleland, two poorly known species from South Australia, by its larger size and different-shaped spores (Reid 1980). The nuLSU sequence, ef1-a, b-tubulin show that it clusters with A. griselloides (Fig. 1, 2A, 2C), but the ITS sequences differ by 12.2-13.8% (Table 4). Amanita pupatju differs from A. griselloides because it has a white, not pale grey or brown, pileus and the inflated cells in the universal veil on the pileus have colourless, not brown, contents (Reid 1980; see below). On this basis, Amanita pupatju is described as a new species.
( Fig. 14, 15) Basidiomata agaricoid, small to medium. Pileus white to ivory, aging pale smoke grey to pale vinaceous-buff to milky coffee to clay buff to drab, paler at margin, margin appendiculate. Universal veil on pileus white to ivory to pale vinaceous-buff, thin, felted, crustose or as flat patches. Lamellae white to ivory. Stipe white. Partial veil superior, white, fugacious. Universal veil at stipe base a cup forming a globose to subglobose marginate bulb. Spores amyloid, elongate to cylindrical.

Description
Pileus 27-90 mm wide, up to 9 mm thick, white to ivory, aging pale cream to pale smoke grey to pale vinaceous-buff to milky coffee to clay buff to drab (5A2-4B2-5B2-6C3-6E3) with pale margin, without surface staining or bruising, initially convex becoming plane with slightly depressed centre, surface tacky when moist; margin non-striate, appendiculate. Universal veil on pileus adnate, thin, felted, crustose or breaking into flat patches over disc, white to ivory to very pale vinaceous-buff (5B2). Lamellae adnate to adnexed to free, close to subcrowded, white or ivory, 2-16 mm broad, margin concolorous, fimbriate; lamellulae plentiful, in several lengths, rounded truncate or truncate or rounded attenuate or attenuate. Stipe 26-60 mm long, 10-20 mm wide, cylindrical, or tapering upward, white, surface smooth or floccose or fibrillose below partial veil. Partial veil superior, descendent, striate above, soft to arachnoid, fugacious, white. Universal veil at stipe base a cup, forming a marginate globose to rounded bulb, 10-18 Â 12-22 mm, sometimes with free limb up to 5 mm high, white. Pileus and stipe context white, unchanging, stipe solid. Smell none. Spore deposit white or B (ivory).

Etymology
From the Latin sabulosus meaning 'sandy', referring to the sandy soil where it occurs. Figure 1 shows A. sabulosa is a member of section Arenariae.

Description
Pileus 33-65 mm wide, up to 10 mm thick, silvery-grey to smoke grey to silvery-beige to cinnamon to pale brown to brown, margin paler, without surface staining or bruising, initially convex, becoming plane with decurved margin and slightly depressed centre, surface slightly tacky when moist shiny when dry; margin non-striate, slightly appendiculate when young. Universal veil on pileus adnate, felted to floccose, as one or several patches in centre of disc, white to light brown to greyish-brown to brown. Lamellae adnexed to adnate, close to subcrowded, white to pale buff (5B2), 6-10 mm broad, margin concolorous, fimbriate; lamellulae frequent to infrequent in several lengths, shortest truncate or rounded truncate, longest attenuate or subattenuate. Stipe 20-48 mm long, 8-27 mm wide, cylindrical or tapering upwards, white or cream, surface at top of stipe powdery, below smooth or minutely floccose or fibrillose. Partial veil superior to median, descendent, striate above, fugacious or not observed, white or pale buff. Bulb 25-30 Â 8-32 mm, ovoid or conic becoming tapered, white or pale buff. Remains of universal veil at top of bulb not apparent or a soft, floccose rim at top of bulb, white. Pileus and stipe context white, unchanging, stipe solid, becoming hollow. Smell not distinctive or unpleasant. Spore deposit white.
Pileipellis poorly developed in young specimens, up to 200 mm thick in old specimens, with a colourless gelatinised suprapellis up to 150 mm thick, and yellow or brown subpellis, consisting of filamentous hyphae and frequent to infrequent vascular hyphae (inflated cells not observed), filamentous hyphae 2-12 mm wide, thick-walled, walls hyaline, gelatinising, contents colourless to yellow to brown, radially orientated; vascular hyphae 2-10 mm wide, occasionally branched, pale yellow or yellowish-brown. Pileus context consisting of filamentous hyphae (dominant or equal), inflated cells and frequent to infrequent vascular hyphae, filamentous hyphae 3-25 mm wide, with widest hyphae constricted at septa, thin-walled, colourless; inflated cells up to 300 Â 35 mm, thin-walled, some gelatinising, clavate or ventricose or ovoid or cylindrical, colourless; vascular hyphae 3-15 mm wide, occasionally branched, pale yellow. Lamella  Notes Reid (1980) described the universal veil on the pileus of A. griselloides as layered, with a grey felty floccose layer of inflated cells with brown contents intermixed with thinwalled, hyaline hyphae, covered by a white arachnoid superficial layer formed entirely of thin-walled, prostrate hyphae. He commented that the structures of the universal veil on the pileus of an additional collection (DAR 32036) were exactly as in the type. A layered universal veil in which there is a superficial layer of mainly filamentous hyphae occurs in some members of subgenus Amanitina (Bas 1969, pp. 307-308). However, in the type K(M) 236386, this superficial layer has phialides, indicating it is a saprophyte growing on the basidiome. This superficial layer was not seen in DAR 32036. It has not been seen in the other collections of A. griselloides.

Discussion
The recent revision of Amanitaceae by Cui et al. (2018), by using concatenated datasets of nuLSU, b-tubulin, ef1-a and rpb2 gene regions, provides a phylogeny for Amanita species with amyloid spores. Our work, using only the nuLSU gene region, shows that section Arenariae is sister to sections Phalloideae, Strobiliformes and Validae, with good bootstrap support (Fig. 1). Within section Arenariae, there is a well-supported clade containing the secotioid species A. arenaria, A. arenarioides, A. compacta and A. pseudoarenaria, and the agaricoid species A. griselloides, A. pupatju and A. sabulosa (Fig. 2B-D). There are insufficient sequences to resolve their relationships with A. lesueurii, A. peltigera, A. wadjukiorum and A. wadulawitu. the other known members of this section.
The multicopy ITS region is the DNA barcode marker for fungi (Schoch et al. 2012). We have found that cloning is necessary to consistently obtain high-quality sequences (McGurk 2013;McGurk et al. 2016;Davison et al. 2017aDavison et al. , 2017b. However, cloning has shown inconsistent utility of this region for delineating species from different sections of Amanita from southern Australia (Davison et al. 2017a. In section Phalloideae, ITS sequences varied by up to 1.42% among collections of the same species, and by 1.42-1.60% between different species which are morphologically distinct and geographically separated (Davison et al. 2017a); however, in section Arenariae, amplicons from the same individual vary from 0.0 to 9.9% (Tables 3, 4; ). There were no changes to the methodology and processing of samples from sections Phalloideae and Arenariae. Examination of the ITS region showed consistent insertions and deletions among different species, and some consistent transitions and transversions among collections of the same species. We have assumed that these represent real differences among clones, not being PCR-induced errors; no chimeras were detected by UCHIME. We were not able to determine whether these differences represent past hybridisation events (Hughes et al. 2015(Hughes et al. , 2018.

Section Arenariae
It is likely that there are many more described agaricoid species in Australia that are unrecognised members of this section. These would be expected to have a slightly appendiculate pileus margin, and inflated sub-basidial cell; possible examples are A. gracilenta, A. albidannulata A.E. Wood, and A. annulalbida A.E.Wood.

Data availability statement
The data that support this study are available in the article.

Conflicts of interest
The authors declare that they have no conflicts of interest.