The two collections of Lactifluus caliendrifer were made during fieldwork by Komsit Wissitrassameewong in 2012 and are deposited in Herbarium Universitatis Gandavensis, Belgium (GENT) and the herbarium of Mae Fah Luang University, Chiang, Thailand (MFLU). For L. russulisporus, fieldwork in 2010 and 2012 by Roy Halling and collaborators resulted in two collections of the species, which are deposited in The William and Lynda Steere Herbarium of the New York Botanical Garden (NY) and the Queensland Herbarium (BRI). We know from earlier research (De Crop et al. 2017; De Crop et al. 2016) that Halling 9398 and Wisitrassameewong 378 belong to L. subg. Gymnocarpi sect. L.teoli. Our dataset contains the ITS sequences used for L. subg. Gymnocarpi in De Crop et al. (2017), supplemented with newly generated sequences. Five species of L. subg. Lactifluus were used as outgroup.

Macroscopic characters are all based on fresh material. Microscopic features were studied from dried material in Congo red in SDS. Possible excretory products were checked for in Cotton blue in lactic acid and Cresyl blue (Clémençon 1997; 2009) Spore ornamentation is described and illustrated as observed in Melzer’s reagent. A total of 40 spores (20 per collection) were measured for each of the two new species. For details on terminology we refer to Verbeken (1998) and Verbeken and Walleyn (2010). Line-drawings were made with the aid of a drawing tube (Zeiss camera lucida on a Zeiss Axioskop 2 microscope equipped with a magnification changer of 2.5× for spores and an Olympus U-DA on an Olympus CX21 microscope for individual elements and pileipellis structures) at original magnifications: 6000× for spores, 1500× for individual elements and sections. Basidia length excludes sterigmata length. Spores were measured in side view, excluding the ornamentation, and measurements are given as (MINa) [AVa-2*SD]–AVa–AVb–[AVb+2*SD] (MAXb), with AVa = lowest mean value for the measured collections and AVb = greatest mean value for the measured collections, SD = standard deviation, MINa = lowest extreme value of collection “a” and MAXb = greatest extreme value of collection “b”. The Q-value (quotient length/width) is given as (MIN Qa) Qa–Qb (MAX Qb), with Qa = lowest mean ratio for the measured collections and Qb = greatest mean ratio for the measured collections, MIN Qa = lowest extreme ratio of collection “a” and MAX Qb = greatest extreme ratio of collection “b”. Other measurements are given as MIN–MAX values. Colour codes refer to Kornerup and Wanscher (1978). Microscopic photographs were taken using a Nikon eclipse NI-U–microscope equipped with a DX-Fi1c camera and Nikon NIS-Elements software including EDF module.

DNA from dried collections was extracted using the protocol described by Nuytinck and Verbeken (2003) with modifications described in Van de Putte et al. (2010), and from fresh material using the CTAB extraction method described in Nuytinck and Verbeken (2003). Protocols for PCR amplification follow Le et al. (2007). The internal transcribed spacer (ITS) was sequenced for a second collection for each new species using the primers ITS1-F and ITS4 (Gardes and Bruns 1993; White et al. 1990). PCR products were sequenced using an automated ABI 3730 XL capillary sequencer (Life Technology) at Macrogen. Forward and reverse sequences were assembled into contigs and edited where needed with SequencherTM v5.0 software (Gene Codes Corporation, Ann Arbor, MI, USA).

Sequences were aligned online using the E-INS-I strategy of the multiple sequence alignment program MAFFT v7 (Katoh and Standley 2013). Trailing ends were trimmed, and where necessary, the alignment was manually edited in MEGA 7 (Kumar et al. 2016). The alignment can be obtained from the first author and TreeBASE (Submission ID S23999). The best partition scheme was selected with PARTITIONFINDER 2 (Lanfear et al. 2016) using standard settings. Aligned sequences were partitioned into 18S (1–56), ITS1 (57–334), 5.8S (335–482), ITS2 (483–820) and 28S (821–868). Maximum likelihood (ML) analyses were conducted with RAxML v8.2.10 (Stamatakis 2014), where a ML analysis was combined with the Rapid Bootstrapping algorithm with 1000 replicates under the GTRCAT option (Stamatakis et al. 2008). All analyses were performed on the CIPRES Science Gateway (Miller et al. 2015).

When it comes to terminology used in the genera Lactarius and Lactifluus, most authors tend to follow Verbeken and Walleyn (2010) and Verbeken (1998). Unfortunately, some confusion seems to exist concerning hymenophoral cells that can be termed either leptocystidia or sterile elements. Even though this type of cell is frequently present in Lactifluus (pers. observations), these cells are only rarely reported in species descriptions (De Crop et al. 2019; Delgat et al. 2017), probably often being dismissed as basidioles and/or of limited taxonomic value. This problem presented itself during the description of the two new species and a consensus between the authors of this paper was pursued.

The term leptocystidium is composed of the Greek leptós, meaning “smooth, thin-walled” and cystidium, meaning “a sterile body, frequently of distinctive shape, occurring at any surface of a basidiome, particularly the hymenium from which it frequently projects”(Ainsworth 2008). In Clémençon (1997), leptocystidia are described in a similar manner, with the addition that they often have an excretory function. For the latter, we could not find evidence in our collections. According to Verbeken and Walleyn (2010), leptocystidia can be regarded as “thin-walled cystidia without remarkable content and thus only deviating by their shape. They are tapering at the top and often have a rostrate apex, which makes them easy to confuse with monosterigmatic basidia. One can consider them to be cystidia if they are regularly observed and if they never bear a spore or spore primordium”. In the two new species, and by extension in most Lactifluus species, thin-walled sterile cells with no remarkable content occur in the hymenium. Furthermore, they do not exhibit a deviating shape, being cylindrical and usually ending blunt. If shape deviation is seen as a vital component for being a cystidium, these cells cannot be named as such. In addition, we dismiss the idea that these cells represent basidioles. Firstly, no intermediate forms between these cells and basidioles were observed. Secondly, in L. russulisporus these cells display a different morphology in both collections. In RH 9674, and by extension in general, they do not protrude from the hymenium and do not exhibit a deviant form, leaving open the possibility that they constitute basidioles or protobasidia (Fig. 7C). However, in RH 9398, they grow out strikingly, protruding clearly from the hymenium (Fig. 7A, B). The same behaviour is seen in the pseudocystidia and marginal cells in this collection. According to Moore (2005), principle nine of fungal developmental biology states that “meiocytes appear to be the only hyphal cells that become committed to their developmental fate. Other highly differentiated cells retain totipotency–the ability to generate vegetative hyphal tips that grow out of the differentiated cell to re-establish a vegetative mycelium.” A possible hypothesis is that some stimulus, perhaps environmental, caused the totipotent cells in the hymenium to grow out, giving rise to the protruding sterile elements, pseudocystidia and marginal cells in RH 9398. This explanation adds to the idea that these cells are not precursor cells of meiocytes (basidia).

As these sterile elements are argued not to be cystidia or basidioles, the question remains as to what they are. Several terms might have been used to indicate the same kind of cells. For example, haplohyphidia refers to unmodified, unbranched or little branched terminal hyphae in the hymenium of (mostly) Aphyllophorales. An intriguing term, paraphyses, is used in the works on the developmental biology of the hymenium done in Coprinopsis cinerea (Horner and Moore 1987; Rosin and Moore 1985a). These cells originate as branches of sub-basidial cells and insert into the basidial layer, later inflating so that they become the main structural component as a pavement from which basidia and cystidia protrude (Horner and Moore 1987; Moore 1985; Rosin and Moore 1985a; b). This description fits well with the sterile elements observed in Lactifluus (Figs 7, 8F). Nevertheless, paraphyses is a term strongly associated with Ascomycota, used for more hair-like (filiform) cells. It cannot be stated with certainty that Ascomycete paraphyses are homologous to the cells we find in Lactifluus.

Given the lack of a distinctive deviating shape in most cases, the improbability of being basidioles and the neutrality of the term, we recommend the use of the term ‘sterile elements’ over the terms ‘leptocystidia’ and ‘paraphyses’ to refer to these cells.

Thereto can be added that marginal cells often bear a striking resemblance to sterile elements (Fig. 8). Furthermore, in Inocybe, little differentiated cystidia are referred to as paracystidia, which also show similar morphology to marginal cells and might constitute the same type of cell (Jacobsson and Larsson 2012; Kuyper 1986). Presently it is difficult to argue whether this is due to homology or homoplasy. Marginal cells are sterile elements on a sterile edge that differ from pleurocystidia and are, in fact, ‘hairs’ sensu Romagnesi (Verbeken and Walleyn 2010). In species where the edge is fertile, sterile elements are also present on the edge. It is possible that, when no differentiated marginal cells are present on an infertile edge, sterile elements are present and consequently reported as being marginal cells. We suggest paying more attention to these sterile elements which occur predominantly in Lactifluus. Given the variation that we observe within L. russulisporus, it is likely that the taxonomic value of this character is rather low, but this needs more observations.