Insects Associated With Droppings of Moose, Alces alces (L), in Isle Royale National Park, Michigan

Moose (Alces alces [L]) droppings were sampled in June-August, 2007, in Isle Royale National Park, a remote wilderness island in Lake Superior, to characterize biodiversity of the park’s moose dung fauna. Twelve Diptera and nine Coleoptera species were obtained, for a total of 21 insect taxa. Twenty of the taxa are newly recorded colonists or visitors to moose dung. The Diptera were Psychodidae, Anthomyiidae, Muscidae and representatives of three other families, and Coleoptera consisted of three species in each of Scarabaeidae, Staphylinidae and Histeridae. Species per sample ranged from two in early June to nine in mid-July, and a species accumulation curve indicated a total of six more species remain to be detected. The relatively depauparate nature of the island’s dung insect fauna may be attributed to absence of other large herbivores, to geographic isolation from source populations on adjacent mainland, or to distinct physical or biological properties of moose dung. Among reared specimens, Hylemyza partita (Meigen) (Anthomyiidae) and moose fly, Haematobosca alcis (Snow) (Muscidae) required approximately 3 weeks to complete development from egg to adult. Projections from weather records on the island indicated the two species could have completed as many as five generations between dates of last spring frost in May and first autumn frost in November. ____________________ Adults and immatures of many arthropods use dung of large herbivores for habitat and nutrition. Faunas associated with many large North American herbivores are well characterized, including cattle (Bos primigenius Bojanus), bison (Bison bison L.), sheep (Ovis aires L.) and horses (Equus ferus caballus L.) (see Hammer 1941, Mohr 1943, Blume 1985, Stevenson and Dindal 1987, Cambefort 1991). With notable exceptions, these faunas share many of the same generalist species (Dormont et al. 2007, Tilberg and Floate 2011), and many of the species are naturally Holarctic or have been introduced from Europe through human commerce (Gordon 1983, Hanski and Cambefort 1991). Relatively less is known about the insects associated with droppings of Cervidae, including deer (Odocoileus spp.; although see Brousseau et al. 2010), caribou (Rangifer tarandus [L.]), elk (Cervus canadensis Erxleben), and moose (Alces alces L.). Isle Royale in Lake Superior has a relatively austere community of large herbivorous mammals, largely due to geographic isolation in the lake. Different cervids and cattle have occurred on the island in recent history, but only moose have persisted since the 1930s. Cochrane (1996) reviewed many historic reports and oral histories regarding large mammals on the island. Caribou were common in the 1700-1800s, but the population declined in the late 1800s and was likely extirpated through hunting by 1928. The Island Mine community on the west end of Isle Royale was known to have about 300 cattle for food in 1875. In addition, white-tailed deer (O. virginianus Zimmermann) were released in 1910 and may have persisted as late as the 1930s. Moose apparently were absent through the 1800s, but arrived from a Minnesota or Ontario population 2013 THE GREAT LAKES ENTOMOLOGIST 205 around 1904 (Peterson 1999). The method of arrival is unknown, although the most likely options include swimming or purposeful introduction. In order of decreasing body size, other herbivorous mammals on the island are beaver (Castor canadensis Kuhl), snowshoe hare (Lepus americanus Erxleben), muskrat (Ondatra zibethicus L.), red squirrel (Tamiasciurus hudsonicus regalis Howell) and deer mouse (Peromyscus maniculatus Wagner). The present study was undertaken to document the insect fauna associated with moose droppings on the island, with particular interest in species that may be unique to moose, endemic on the island, or potential pests or vectors of parasites or pathogens among moose and other animal populations. Materials and Methods Study Area. Isle Royale National Park, Michigan, is an archipelago in northwestern Lake Superior, approximately 46 km SE of Thunder Bay, Ontario, 100 km north of Houghton, Michigan, and 57 km east of Grand Portage, Minnesota (Fig. 1). The park consists of one main island surrounded by hundreds of smaller islands, for a total of 544 km2 of land, 99 percent of which is federally designated wilderness. Forest-dominated ecosystems consist of northern hardwoods in western portions of the park, and boreal forest in the east. Access to the park is generally limited to coastal docks and interior foot trails. Figure 1. Map of Isle Royale, showing locations and dates of moose dropping collections, and locations of weather station and Davidson Island where droppings were incubated, 2007. 206 THE GREAT LAKES ENTOMOLOGIST Vol. 46 Nos. 3 4 Samples and Rearing Records. Natural fecal droppings from free ranging moose were collected as encountered at scattered locations in the park between 10 June and 10 August 2007 (Fig. 1). Moose droppings in summer consist of somewhat amorphous piles of varying size (Rice 2010), and samples consisted of about 50-100% of encountered piles, depending on pile size. Most of the droppings were near lakes in mature forest settings, except for one at Francis Point, which was between the Lake Superior shoreline and a semi-flooded meadow. Two defecating moose were observed on 15 June and their dung was collected about 8 h later. In all other cases, droppings from unobserved moose were estimated to be within 1-3 d, as judged from surface moisture and apparent insect activity. Intent was to select droppings that had been exposed long enough to be colonized by flies and beetles, yet retrieved before wandering stage larvae of muscoid flies could have dispersed into the surrounding soil substrate. Each dropping was placed in a plastic bag and transported to a small, unheated outbuilding on Davidson Island (Fig. 1) to rear out insect inhabitants. Each sample was placed on 2.5 cm of sifted sand in a clear plastic tray (30 × 20 × 7.5 cm) and covered with thin linen that allowed ventilation but prevented insect entry and escape. Samples were inspected and intermittently misted with lake water to mimic natural rain and dew. After 2 wk, samples were rechecked almost daily and then disassembled to extract individual insects by hand when most larvae of muscoid flies had pupated in the dung or sand. Mature beetle larvae and fly puparia were transferred to small, perforated Eppendorf tubes and set aside to rear out adults. Free roaming adult beetles were killed directly by freezing, as were representative specimens of small Diptera that were too numerous to extract and count. Free-living mites were noted roaming on the dung and sand, and attached to adult beetles and flies, but none were deliberately collected. Adult beetles and flies with associated puparia were mounted on points or pins, identified using keys in Arnett and Thomas (2001), Bousquet and Leplante (2006), Gordon (1983), Griffiths (2001), McAlpine et al. (1981), and Pratt and Pratt (1980), and compared with determined reference specimens. Although several subgenera of Aphodius (Scarabaeidae) have recently been elevated to genus by Gordon and Skelley (2007); we retain previous names for compatibility with earlier literature on the fauna of mammal dung, but include subgenus names in parentheses. Specimens have been deposited in the University of Minnesota St. Paul Insect Museum. Faunal Diversity. The fauna was characterized initially by the list of species or higher taxonomic units encountered, and then by a species accumulation curve. An accumulation curve shows the cumulative number of species encountered with increasing number of sample units (droppings), and was calculated with EstimateS v. 9beta4, using the Mau Tau estimator. This estimator provides a mean and 95% confidence limits for numbers of taxa actually present, given the number of species encountered in the initial reference sample (Colwell et al. 2012), and has performed well in studies of insect communities elsewhere (e.g., Basset et al. 2012). The difference between number of taxa observed and extrapolated number was interpreted as the number of taxa that remain to be detected. Development Times. Time required to develop from egg to adult was estimated for two species of muscoid Diptera that were frequent and abundant enough for analysis. Median development times (D50) for cohorts of specimens from individual droppings were calculated as the difference between presumed oviposition date and median emergence date, and then examined for seasonal variation. To test the hypothesis that any seasonal pattern was a result of varying temperatures, concurrent records of daily minimum and maximum air temperatures were obtained from a ground-level weather station at Mount Ojibway, 2 km west of Davidson Island (Fig. 1). Mean temperature for each cohort was calculated by averaging daily minima and maxima from afternoon of oviposition date to afternoon of median emergence date. Analysis of covariance 2013 THE GREAT LAKES ENTOMOLOGIST 207 was used to assess differences in median development times among species and dates or temperatures, as D50= a+b1S+b2X+b3S ́X+e, where a was overall mean, S was a dummy variable for species, X was a covariate, either oviposition day of year or mean temperature, S ́X was an interaction term for unequal slopes, and e was for errors, presumed normal. Significance of b3 was first checked with F-test (α = 0.05), and if insignificant, the interaction term was omitted and the simpler model fit to examine b2 and b1 for the covariate and species, respectively. Models were fit with lm in R (R Development Core Team 2012), and F-tests were done with ANOVA in the package car (Fox and Weisberg 2011). To estimate possible numbers of generations per yr on the island, development times were expressed as degree-days (DDs) above arbitrary bases of 0, 5 and 10°C (Pruess 1983). Degree-days were calculated from the weather station’s daily records with the half-day sine wave method (Allen 1976) without adjustment for latitude. Sums from afternoons of oviposition to median emergence were averaged to estimate mean DD requirement for each species, and then the means of cohorts of the two species were compared with Student’s t-test. Number of possible egg-adult intervals was computed by dividing DD accumulations between last to first frost date by mean DDs required.