**3. Results**

#### *3.1. Fungi from the Laboratory Experiments and Their Prevalence*

Data on 2978 lepidopteran pupae were obtained from the laboratory rearing experiments (Table 2). Fungi were detected from 82 pupae (2.8%) and identified as representing 23 species from 10 families and 6 orders (Table 1). The prevalence of fungal infections differed among the 3 years but remained below 10% (Table 2, note the small sample sizes of *C. pusaria* and *H. punctinalis* in 2016, associated with atypically high prevalence). The majority of the fungi belonged to Cordycipitaceae (59 pupae infected), followed by Aspergillaceae (8), Nectriaceae (7), Mortierellaceae (4), Hypocreaceae (4), Tilachlidiaceae (3), Clavicipitaceae (2), Saccotechiaceae (2), Umbelopsidaceae (1), and Cladosporiaceae (1).


**Table 2.** Incidence of fungal infection in lab reared moths in different years.

On the basis of previous knowledge [12], all of the Cordycipitaceae (six species, overall prevalence 2.2%) were considered to unambiguously represent entomopathogens infecting living hosts. The most abundant of such species, *Cordyceps farinosa*, was found in total on 43 pupae of 3 moth species out of 5. While present each year, it was the most prevalent entomopathogen in 2015 and 2016. Species from other families should be considered potential (opportunistic) pathogens of Lepidoptera as most of them belong to large genera including saprotrophs and pathogens of various hosts, with some representatives occasionally found also on insects [23–25]. There are exceptions from this general scheme, however. In particular, while several species of Clavicipitaceae are known as entomopathogens, the genus *Metapochonia* has been mainly found on nematodes. The genus is represented here by the nematode and rotifer pathogen *Metapochonia bulbillosa* [26], whereas the respective UNITE SH (Table 1) also includes a few sequences obtained from Coleoptera. Moreover, *Tilachlidium brachiatum* has been known to grow only on decaying fungi (K. Põldmaa, personal observation). In addition, our data may include the first records on Lepidoptera/insects for some species from the genera *Mariannaea*, *Mortierella*, *Simplicillium*, *Trichoderma,* and *Umbelopsis*. However, the respective host associations need further investigation as species concepts in these groups are changing as a consequence of advances in molecular systematics.

Lepidopteran species differed in the prevalence of infection—*O. gothica* had the lowest rate (0.4%), while *E. atomaria* and *C. pusaria* had the highest (5% and 3.3%, respectively). There was also a difference between the 3 years of study, showing a trend of increase in the fungal prevalence (Table 2). However, given the somewhat non-systematic character of the data (rearing experiments were not designed to study the prevalence of infections), we refrain here from presenting formal statistical analyses.
