Hebeloma.org - Demystifying *Hebeloma*

• I recently collected several basidiomes that I determined are from genus Hebeloma, but upon further reading I realised that this is a large and rather complex genus.
• Species and even section determination usually requires microscopic examination of spores and cystidia, and there are some difficult and variable species that are not currently resolved by ITS.
• Estimates of around 150 species, in 17 sections, a few with subsections too. Associated with a wide range of trees and shrubs, could be restricted for some species but data is often missing historically or less reliable.
• Worldwide distribution, found in many woodlands, especially in arctic and alpine regions, also common in boreal and temperate regions, and rare in tropical regions.
• Originally described as Agaricus tribus Hebeloma by Fries in 1821 and elevated to genus by Kummer in 1871.
• Typified by Hebeloma mesophaeum since 2013, apparently there was too much controversy around the original type, Hebeloma fastibile (Agaricus fastibilis), based on the original 1801 description of Persoon.
• Currently placed in the family Hymenogastraceae.
• The Latin name translates to ‘Fringe of Youth’, relating to the cortina-like veil found on immature basidiomes of many species in the genus.
• Generally reported as inedible and some as poisonous, common names include ‘Poison Pies’ and ‘Dull Caps’.
• I found the website hebeloma.org, which is an incredible resource, summarising many years of research on the genus, designating types, and describing sections and species.
• All the species descriptions are very detailed, and dynamically parameterised by a database of collections, which contains over 10,000 observations across 120 species. Almost all observations have an associated ITS sequence and many have sequences for other loci too.
• Culmination of decades of work on Hebeloma by H. J. Beker, U. Eberhardt and J. Vesterholt in Europe (2016) and more recently North America.
• Morphology page that explains how to measure, describe, and record morphology (macro and micro) of collections.
• Observations are stored in a dedicated biological database, which also back-ends the website, this allows queries to find certain types of observations and patterns, or to compare the characteristics of a new observation.
• I really like the comprehensive and systematic approach to studying the genus, and the use of morphology, geography, and phylogeny to describe and delimit the sections and species.
• Important features identified for determination of sections and species:
o Geographic location.
o Number of complete lamellae.
o Spore ornamentation, size, and perispore loosening.
o Cheilocystidia shape, length, apical width, and ratios.
• An automatic identifier has been trained on some of the data using machine learning and is available on the website. If some of these features are entered into a form, then suggestions of sections and species with different certainties are generated from the database.
• In testing, the auto-identifier was found to be over 75% accurate for giving the correct species (as determined by expert and molecular analysis) as the top suggestion, and over 95% accurate for giving the correct species in the top 5 suggestions.