Abstract

The fungi kingdom is among the most diverse eukaryotic lineages on Earth with estimates of several million extant species (O’Brien et al., 2005; Blackwell, 2011; Taylor et al., 2014). Fungi play critical roles in carbon andnutrient cycling of terrestrial and aquatic ecosystems, and they are important pathogens and mutualists (Read & Perez-Moreno, 2003; Taylor et al., 2012; Grossart et al., 2016). More than 80% of plant species form symbioses with fungi and these symbioses have been crucial to the colonization of terrestrial ecosystems (Field et al., 2015a; Selosse et al., 2015). Despite their impacts on primary ecosystem functions, assessments of fungal biodiversity estimate that only c. 10% of fungal species have been described (Bass & Richards, 2011; Hibbett et al., 2011). Traditionally, specimen-based taxonomic studies have been the only way to discover new species. Because most fungi have microscopic life-stages and convergent morphological features (Rivas-Plata & Lumbsch, 2011; Wynns, 2015), many fungal groups remain severely undersampled. DNA-barcoding and highthroughput sequencing methods have provided a new framework for studying fungal biodiversity (Fierer et al., 2012; Schoch et al., 2012; Myrold et al., 2014), and diversity estimates based on environmental sequences have increased exponentially. Although these ‘sequence-based classification and identification’ methods are a powerful means to rapidly detect hidden diversity, careful interpretation of these data is needed to make accurate inferences (K~oljalg et al., 2013; Lindahl et al., 2013; Nguyen et al., 2015; Hibbett et al., 2016). In particular, many environmental sequences cannot be associated with a known fungal species or lineage. This remains a major challenge to decipher fungal community composition and understand ecological roles of fungi in leaf litter, soil, or inside plants (Yahr et al., 2016). In some cases, these fungi are truly undescribed and their ecological roles are unknown but in other cases they represent described taxa for which no sequence is available (Nagy et al., 2011; Nilsson et al., 2016). DNA barcoding of herbarium specimens and culture collections is extremely valuable to link unidentified sequences to known taxa (e.g. Brock et al., 2009; Nagy et al., 2011; Osmundson et al., 2013; Garnica et al., 2016).DNA sequences have been generated from fungal type specimens > 200 years old (Larsson & Jacobsson, 2004), but in many cases obtaining sequences from historical material is challenging (Dentinger et al., 2010). Today’s threats to biodiversity from habitat loss and climate change are occurring at an unprecedented scale, and it is possible that many species may become extinct before they have been discovered (Costello et al., 2013; Monastersky, 2014). In the need to describe and protect as many species as possible we addressed the following questions: what are the best methods to rapidly document fungal biodiversity? Are traditional, specimen-based approaches still useful?