Rove beetles (Coleoptera: Staphylinidae) have proven to be useful indicators of forest disturbance and recovery because they are sensitive to environmental perturbations, diverse in species and trophic roles, easily sampled, and at least in central Europe and Canada, mostly readily identified using a wealth of available taxonomic tools (Boháč 1990, 1999, Pohl et al. 2008). Many staphylinid species show distinct response patterns following forest disturbances (e.g., Pohl et al. 2007, 2008, Klimaszewski et al. 2008, Work et al. 2013). For example, in a recent study of rove beetles following removal of logging residues by whole-tree harvesting in boreal balsam fir forests of Quebec, three Atheta species, Tachinus fumipennis (Say) and Tachinus frigidus Erichson were negatively affected by the removal of forest biomass, while Gabrius brevipennis (Horn), Pseudopsis subulata Herman and Quedius labradorensis Smetana were not and their catch increased (Work et al. 2013). While studies comparing species assemblages can quantify the overall effects of harvest treatments or other forest disturbances, they are often not designed to identify specific underlying mechanisms for individual species’ responses. Study of trophic roles may provide useful insights into these response patterns by assessing factors such as individual predator-prey (or consumer-resource) relationships, the degree of diet specialization, and possible associations of beetles with specific microhabitats that may serve as habitat or substrate for their food resources.

Rove beetles are a diverse group exhibiting a wide variety of trophic relationships and occupying numerous microhabitats in forest ecosystems. Many Aleocharinae and Staphylininae, e.g., species of Aleochara, Philonthus, Platydracus, and Staphylinus, are voraciouspredators of other arthropods such as fly larvae (Klimaszewski 1984, Smetana 1995). At least some species of Scaphidiinae, Osoriinae, Tachyporinae, and Aleocharinae (Gyrophaenina) eat the flesh or spores of fungal sporocarps (Seevers 1951, Ashe 1984, Newton 1984). Oxytelinae are generally detritivores and feed on decaying plant material and algae (Thayer 2005, Makranczy 2006). A few Omaliinae, e.g. Eusphalerum and some other genera, are pollen-feeders (Thayer 1987, 2005). A few species, such as the aleocharine Himalusa thailandensis Pace, Klim. & Cent., feed on live plant tissue (Klimaszewski et al. 2010). Most information on food sources of rove beetles has been obtained through observation of individual beetles in the field or laboratory, or inference from habitat associations of species. For example, some groups (e.g., Aleochara, Philonthus) that are collected in decaying mushrooms are predators of dipteran larvae that co-occur within these fungi (Klimaszewski 1984, Smetana 1995). While direct observations of feeding provide compelling evidence of dietary preferences, inferences based on habitat preferences are not definitive evidence of consumer-resource relationships. Other methods have been used to more definitively establish feeding habits of beetles including microscopical examination of gut contents (Newton 1984, Thayer 1987) and immunological methods (Dennison and Hodkinson 1983). The use of molecular techniques to investigate dietary preferences and trophic links in rove beetles is presented here for the first time, but similar techniques were used in the past to investigate invertebrate predators for multiple prey using DNA targets (Harper et al. 2005). Increasingly large databases of DNA sequences in repositories such as GenBank and MycoBank will make these techniques more and more useful for examining relationships between beetles and cryptic food items such as fungi and bacteria (Crous et al. 2004, Robert et al. 2005). For beetles that feed on organisms with relatively strict habitat requirements, such as fungal species that require lignocellulose, molecular gut analyses may lead to inferences on the importance of habitat elements such as downed deadwood (Suh and Blackwell 2005a, b).

In this study we use both microscopic examination and molecular analysis of gut contents to more precisely characterize the feeding habits and trophic role of 10 rove beetle species common in the boreal forest of Quebec. There are few published data on gut contents, of these species and little is known of their food affiliations, except for some general statements on habitat preferences of Tachinus species (Campbell 1973) and limited observations on hosts and gut contents of some Tachinus and Pseudopsis species (Newton 1984).

Both dissection and molecular analysis of guts strongly suggest that rove beetles in this study may feed primarily on yeasts. Yeasts are ubiquitous (in soil, on decaying plant material including deadwood, and on berries) and they are an important part of the diet of at least some fungivorous beetle species (Suh and Blackwell 2005b). Some species of Candida yeasts have close associations with saproxylic insects and are capable of transforming d-xylose and other important components of lignocellulose to ethanol (Wang et al. 2005). However, yeasts within the Candida mesenterica clade are associated with many insect groups and are likely indicative of habitat associations rather than being highly specific gut symbionts (Suh and Blackwell 2005a). Yeasts in the Candida mesenterica clade, particularly species in its subclade A, are known to be associated with fungal basidiocarps and have previously been isolated from the digestive tracts and body surfaces of six families of basidiocarp-inhabiting beetles, including one unidentified species of Staphylinidae (Suh and Blackwell 2005a).

With the exception of the relatively broad consumption of Candida mesenterica yeasts by most species, finer patterns in feeding preferences among rove beetles were difficult to assess. This was partly a result of the limited number of matched sequences for some species, which in turn probably reflects in part the limits of available reference sequences. The limited numbers of sequences obtained in our study could be related to degradation of DNA as the result of suboptimal preservation medium.

The prevalence of other fungi in addition to yeasts and the presence of spores in rove beetle guts was not unexpected, as many rove beetle species are associated with fungi (Campbell 1973, Newton 1984, Newton et al. 2000, Thayer 2005). Spores of at least seven species of fungi were found in the guts of six of the rove beetle species, although it is not known whether the beetles derive nutrition from spores. Many spores have tough walls that enable them to pass through digestive tracts; however, others are certainly digested and some are cracked by the mouthparts to provide nutrition for beetles specializing in spore feeding (Lawrence 1989, Betz et al. 2003). It cannot be determined from available data whether the ingestion of spores by these rove beetles is incidental or intentional. However, the absence of spores in four species, including all three species of Aleocharinae, is notable and raises the question of whether some species do not ingest spores, either because they are unable to, or because they have difficulty finding them in a particular microhabitat.

Although we isolated bacteria less commonly than fungi, we did find them in six beetle species. The second most commonly detected sequence, in fact, was from the bacterium Serratia marcescens, which is associated with soils; it may sometimes be pathogenic to insects (Flyg et al. 1980). This species of bacteria is probably not an important food source for rove beetles. It may simply be so common in the soil that incidental ingestion is frequent. The other species of bacterium we isolated, Bradyrhizobium japonicum, is a soil-dwelling, nitrogen-fixing species associated with legume plants (Rivas et al. 2009), so seems unlikely to be a food source for rove beetles.

It is notable that no arthropod cuticle or evidence of animal DNA sequences were found in the guts of any of these species despite the fact that predation on arthropods (especially mites, springtails, and smaller insects) is common in the family (Newton et al. 2000, Thayer 2005). Two possible interpretations are that: (1) the adults of these species are entirely fungivorous, or (2) they are predaceous and use preoral digestion, as many staphylinids do (e.g., Evans 1965, Dennison and Hodkinson 1983, Thayer 2005) and specifically as has been hypothesized for Pseudopsis (Pseudopsinae) and the entire group of subfamilies to which it belongs (Grebennikov and Newton 2009). The fact that these 10 species represent seven genera and three subfamilies suggests that analysis of gut contents of many more species is needed to provide a better sampling of rove beetle diets. Identification of the many presently unmatched DNA sequences, which could include animal DNA, and ruling out of preoral digestion are also required before carnivory can be excluded with certainty for the species studied here.

In work conducted to characterize rove beetle responses to removal of logging residues following clearcut harvesting in boreal balsam fir forests of Quebec (Work et al. 2013; Klimaszewski, unpublished data), several response patterns were shown by different species. Seven species (Atheta capsularis, Atheta klagesi, Bryophacis smetanai, Oxypoda grandipennis, Tachinus frigidus, Tachinus fumipennis and Tachinus quebecensis) were found predominantly or exclusively in uncut forest rather than forest subjected to harvesting treatments. Except for Tachinus quebecensis, all of these may feed partly or wholly on basidiocarps, as the predominance of basidiocarp-associated Candida spp. in their guts suggests. These species may not persist well in harvested stands because their drier, disturbed conditions are generally far less favourable to mushroom and other sporocarp production (Langor, personal observation). Tachinus quebecensis was found only in uncut stands, and had Candida sophiae-reginae and Candida mesenterica isolated from guts in the present study. However, this beetle species had the highest diversity of ingested fungal species, five of them unique to it, and possibly one or more of these represent important food sources that are absent (or rare) in harvested stands, although it is also possible that it is predaceous and all fungi are incidental.

Pseudopsis subulata was the most common species to show a strong affinity for disturbed stands, specifically stands subjected to whole tree harvesting, although it was also found in uncut stands and in stands subjected to harvesting with debris left behind (Work et al. 2013). Interestingly, this is the only rove beetle species that did not have Candida spp. in its gut; however, its gut was typically packed with other yeasts that could not be identified. Perhaps these yeasts have a strong association with disturbed and open habitats.

Mycetoporus montanus was not collected during the first year of the study (Work et al. 2013), but it was common during the second year, when it was collected almost exclusively in harvested treatments (Klimaszewski, unpublished data). It appears that this species moved into disturbed stands and multiplied rapidly, taking advantage of food or breeding sites that became more available in such stands. Although Mycetoporus montanus had Candida mesenterica in its gut, it also had a large variety of other organisms that could not be identified, some of which may be the primary food source for this species.

Ischnosoma longicorne was commonly found in both uncut and disturbed stands (Work et al. 2013). This species had a high diversity of species in its gut (six fungal and one bacterium species), which may indicate a broad diet and, therefore, a capacity to succeed in many habitat types.

Feeding associations between rove beetles and yeasts provide some insight into potential mechanisms by which biomass harvesting may impact rove beetles. Our results may suggest that dominant rove beetles are feeding on yeasts and other fungi that may or may not be directly associated with sporocarps growing on deadwood substrates. It is important to understand the complexity of factors linking the studied beetles to biomass removal treatments. The removal of additional forest biomass may be affecting beetles not only via potential food linkages, but also by other non-trophic mechanisms such as changes in physical conditions following the removal of the forest overstory (Work et al. 2013).

In addition to characterizing food sources for some abundant species of rove beetles, many of which are good ecological indicators, our work provides some possible explanations for beetle response patterns in the wake of forest disturbance. The relatively easy application of DNA sequencing to gut contents and the steadily increasing wealth of sequence data available to serve as an identification resource means that these techniques can now be readily applied in disturbance ecology research to investigate species response patterns and habitat preferences. We encourage broader use of this approach to support future work.