Bumble Bees (Hymenoptera: Apidae: Bombus spp.) of Interior Alaska: Species Composition, Distribution, Seasonal Biology, and Parasites

Abstract Background Despite the ecological and agricultural significance of bumble bees in Alaska, very little is known and published about this important group at the regional level. The objectives of this study were to provide baseline data on species composition, distribution, seasonal biology, and parasites of the genus Bombus at three major agricultural locations within Alaska: Fairbanks, Delta Junction, and Palmer, to lay the groundwork for future research on bumble bee pollination in Alaska. New information A total of 8,250 bumble bees representing 18 species was collected from agricultural settings near Delta Junction, Fairbanks, and Palmer, Alaska in 2009 and 2010. Of the 8,250 specimens, 51% were queens, 32.7% were workers, and 16.2% were males. The species composition and relative abundances varied among sites and years. Delta Junction had the highest relative abundance of bumble bees, representing 51.6% of the specimens collected; the other two locations, Fairbanks and Palmer represented 26.5% and 21.8% of the overall catch respectively. The species collected were: Bombus bohemicus Seidl 1837 (= B. ashtoni (Cresson 1864)), B. balteatus Dahlbom 1832, B. bifarius Cresson 1878, B. centralis Cresson 1864, B. cryptarum (Fabricius 1775) (=B. moderatus Cresson 1863), B. distinguendus Morawitz 1869, B. flavidus Eversmann 1852 (=B. fernaldae Franklin 1911), B. flavifrons Cresson 1863, B. frigidus Smith 1854, B. insularis (Smith 1861), B. jonellus (Kirby 1802), B. melanopygus Nylander 1848, B. mixtus Cresson 1878, B. neoboreus Sladen 1919, B. occidentalis Greene 1858, B. perplexus Cresson 1863, B. rufocinctus Cresson 1863, and B. sylvicola Kirby 1837. Overall, the most common bumble bees near agricultural lands were B. centralis, B. frigidus, B. jonellus, B. melanopygus, B. mixtus, and B. occidentalis. Species' relative population densities and local diversity were highly variable from year to year. Bombus occidentalis, believed to be in decline in the Pacific Northwest states, represented 10.4% of the overall specimens collected from the three sites studied. Bumble bees were found to be infected by Nosema and nematodes with infection rates up to 2.1% and 16.7% respectively. Of the eight species infected by parasites, B. occidentalis displayed the highest Nosema infection, while B. centralis was the species with the highest infection of nematodes. To our knowledge this represents the first multi-year study on bumble bees from the main agricultural areas of Alaska to provide baseline data on species composition, distribution, seasonal biology, and parasites of the genus Bombus.


Introduction
Bumble bees are considered important pollinators in subarctic Alaska (Washburn 1963, Kevan 1972. Bumble bees will forage during rainy, cool, and windy weather during which honey bee activity is limited (Buchmann 1983). They have even been observed foraging during snowfall, under a full moon (Kearns and Thomson 2001), during the night, above the tree line (Richards 1973, Lundberg 1980, and in temperatures as cold as -3.6°C (Heinrich 1979). Native bees, such as bumble bees, are responsible for the pollination of over $3 billion US dollars worth of fruits and vegetables produced in the US (Losey and Vaughan 2006).
Despite the ecological importance of bumble bees, no published estimates on the value of bumble bee pollination for crops in Alaska are available. Furthermore, there is no consensus on the total number of Bombus species present in Alaska with estimates ranging from 17-24 species (Washburn 1963, Krombein et al. 1979, Bishop and Armbruster 1999, UAM 2013b. These estimates suggest the Bombus fauna of Alaska contains about half of the species known for North America. , the most recent and authoritative publication on North American Bombus, list 23 species from Alaska. Nationwide, honey bees are undergoing extensive die-offs which do not appear to have a single underlying cause; a phenomenon termed Colony Collapse Disorder (CCD) (Bromenshenk et al. 2010). Some predict that native bees will buffer potential declines in agricultural production due to CCD (Buchmann and Nabhan 1996, Kremen and Ostfeld 2005, Winfree et al. 2007), but in many cases, as in Alaska, the native bee fauna is poorly known. There are also concerns about the long-term persistence of bees, some of which are predicted to become extinct, as a result of the planet's changing climate (Rasmont et al. 2015).
Nosema is a genus of obligate microsporidian intracellular parasites that has been known to affect economically important insects such as the silkworm moth, honey bees, and bumble bees (Otti andSchmid-Hempel 2007, Koch andStrange 2012). Nosema bombi infestation has been related to declining bumble bee populations and reduced genetic diversity of North American bumble bees (Flanders et al. 2003, Thorp and Shepherd 2005, Colla et al. 2006, Cameron et al. 2011. Impoverished native bumble bee communities are often associated with the intensification of agriculture and may be insufficient to replace the pollination services currently provided by honey bees (Goulson et al. 2008). Alaskan farms tend to be surrounded by native vegetation and habitat that would benefit native bee populations, but there is little information on bumble bee species composition, geographical distribution, biology, and factors affecting bumble bee species richness associated with agricultural areas in the state. The objectives of this study were to provide baseline data on species composition, distribution, seasonal biology, and parasites of the genus Bombus at three agricultural locations within Alaska: Fairbanks, Delta Junction, and Palmer. Stephen and Rao (2005)The three major agricultural areas of Alaska (Benz et al. 2009) with farms were sampled in 2009 and 2010. These were the University of Alaska Fairbanks experimental farms near Delta Junction (64.04°N, 145.73°W), Fairbanks (64. 85°N, 147.85°W ), and Palmer (61. 60°N,149.13°W,WGS84). Habitat types surrounding field sites ranged from urban areas with mixed boreal forest and a botanical garden near Fairbanks, grasslands and boreal forest near Delta Junction, and large scale commercial agricultural lands near urban areas in Palmer. All three locations grow potatoes, barley, wheat, oats, oilseeds (camelina, canola, and mustard), and rhubarb.

Material and methods
Blue vane Japanese beetle traps (SpingStar Inc; Woodinville, Washington) were placed (five traps per site per year) around agricultural field perimeters and set at a height of one meter from ground level following the methods described by Stephen and Rao (2005). In our study, traps were hung horizontally to prevent rain from entering traps. Both years, the traps were placed along a tree or fence line 200 meters apart in a straight line along the same field edge. Traps had a 6.5 cm piece of Vaportape ® (Hercon Environmental; Emigsville, Pennsylvania) in the bucket to kill captured insects. The vaportape was replaced every 6 weeks. Traps were serviced every seven days; bumble bees were removed, transported to the laboratory, and stored in labeled Ziploc® bags, and frozen until they could be pinned, labeled and identified in the Agricultural Research Service (ARS) laboratory in Fairbanks, Alaska. Sampling dates were May 19 to September 10, 2009 andMarch 27 to September 28, 2010 in Delta Junction;March 27 to September 23, 2009 andMay 3 to September 27, 2010 in Fairbanks;May 4 to September 21, 2009 and May 17 to October 7, 2010 in Palmer.
Initially, a series of Alaskan specimens were identified by Dr. Jamie Strange, United States Department of Agriculture (USDA), Agricultural Research Service (ARS), Pollinating Insects Research Unit, Logan, Utah. Subsequent identifications were made using the keys of Thorp et al. (1983) and Stephen (1957) as well as comparison to the voucher collection identified by Strange. Voucher specimens were deposited in the University of Alaska Museum (UAM) Insect Collection, Fairbanks, Alaska. Records of these specimens are available online via the UAM database (UAM 2013a).
A series of taxonomic changes resulting from recent DNA barcoding work (Cameron et al. 2007, Williams et al. 2012 were made subsequent to the start of, but prior to the completion of, this project. As a result, some species in our study were referenced (on specimen labels, lab notes, data files, etc.) using now invalid names. In this paper we have used the currently valid names from Williams (1988) and Williams et al. (2012).
Two trials were conducted to establish the presence of entomoparasites in bumble bees. From May 26 to September 17, 2010, ten bees per week were hand collected from the University of Alaska Fairbanks, Georgeson Botanical Garden (GBG) and frozen until their abdomens were dissected following the procedure described by Klee et al. (2006) and Plischuk et al. (2009). Bumble bees were collected with the aid of a glass jar. Only bees resting on flowers or structures were collected. During 2011, the sampling technique was modified and bees were captured with blue vane traps as previously described. Ten traps per locality (Fairbanks and Palmer) were setup for 24hr periods once a week from 3 May -16 September 2011. Dissected digestive/reproductive tracts were homogenized in 2 ml of distilled water and the homogenate examined by light microscopy (400x) to determine the presence of microsporidian-like spores of Nosema (Klee et al. 2006, Plischuk et al. 2009

Results
Delta Junction had the highest relative abundance of bumble bees with 4,258 specimens representing 51.6% of the overall catch. Fairbanks and Palmer represented 26.5% and 21.8% of the overall catch respectively. Sixteen of the identified 18 species were collected from Delta Junction, while 14 species were identified from Fairbanks and Palmer. Of the 8,250 specimens examined, 51.0% were queens, 32.7% were workers, and 16.2% were males. Six of the 18 species collected in this study were found at all three locations during both sampling years: B. centralis, B. frigidus, B. jonellus, B. melanopygus, B. mixtus, and B. occidentalis.

Delta Junction
Sixteen species were collected from Delta Junction ( Table 1). The most abundant species both years was B. bifarius representing approximately 46% and 54% of the specimens collected in 2009 and 2010 respectively. In 2009, three species, B. bifarius (46.3%), B. jonellus (17.1%), and B. frigidus (11.0%) represented 74% of the total bumble bees collected. In 2010, a different set of species, B. bifarius, B. occidentalis, and B. jonellus, contributed 76.4% of the specimens that year with percentages of 54.1, 12.4, and 9.9 respectively. Relative abundances were lower in 2010 (n = 2745 specimens) as compared to 2009 (n = 4020); however, the percentage of queens was higher in 2010 (79.1%) as compared to 2009 (47.5%). Flight activity was earlier in 2010 than in 2009 (Fig. 2), but relative abundance was lower than 2009 (

Palmer
Fourteen species were collected from Palmer (  Fig. 3). The fourth species in order of abundance was B. mixtus, but relative densities were low most of the season during both years.
In 2009, B. centralis and B. occidentalis were collected as early as 14 May (Fig. 3)

Parasites
Nematodes were found in nine specimens of two species among 101 Bombus specimens examined. Infection incidence was 16.7% of B. centralis specimens and 6.3% of B. perplexus specimens. The nematodes were identified as belonging to the family Tetradonematidae.
A total of 642 bumble bee specimens, from Fairbanks and Palmer, of seven species were examined for microsporidians (Nosema spp.) (

Discussion
All the species recovered have been previously reported from Alaska in collections, databases, and publications. However, to our knowledge, this represents the first multiyear study focused on seasonality and abundance of Alaskan Bombus species in the major agricultural regions of the state. This also represents the first report on nematodes and the second report on Nosema affecting bumble bees in Alaska.
Although no published reports are available on bumble bee population dynamics in Alaska's major agricultural areas to compare with our results, working with other taxa, Pantoja et al. , Pantoja et al. 2010b, Pantoja et al. 2010a) reported a similar pattern of higher relative densities of leafhoppers (Cicadellidae) and wireworms (Elateridae) in Delta Junction relative to Fairbanks and Palmer. Pantoja et al. , Pantoja et al. 2010a, Pantoja et al. 2010b) suggested that the differences in relative leafhopper and wireworm densities were associated with climatic differences, cropping histories, habitat availability, or agronomic practices. Bumble bee diversity and abundance can be affected by the availability of floral resources and nest sites, climatic conditions, presence of invasive species, habitat fragmentation, parasitic spillover, urbanization, competition, and the use of pesticides (Cane and Tepedino 2001, Roubik 2001, Goulson et al. 2008). The three sites studied have significant climatic differences (Benz et al. 2009) and cropping histories ) that might have affected relative bumble bee densities. Proximity to urban areas might provide another explanation for the differences in species composition and densities between sites. Traps in Delta Junction were located in rural areas, while traps in Fairbanks and Palmer were within two kilometers of major highways and structures. Bumble bee populations respond positively to the presence of unmanaged areas (pastures, meadows, and forests) that provide nesting and forage sites (Williams 1986). Davros et al. (2006) (Kirby)), B. sitkensis Nylander 1848, and B. vagans Smith 1854(Ashmead 1902, Bequaert 1920, Washburn 1963, Milliron 1973, Williams and Batzli 1982, Thorp et al. 1983, Henrich and Vogt 1993  Published literature on parasites of bumble bees in Alaska is scant. Schmid-Hempel and Tognazzo (2010) described a protozoan flagellate, Crithidia bombi Lipa and Triggiani in Alaskan bumble bees. Koch and Strange (2012), discuss the distribution and relative abundance of eight Bombus species in Alaska and the prevalence of Nosema bombi Fantham and Porter 1914, detected in B. occidentalis, B. cryptarum (as B. moderatus), B. bifarius, B. flavifrons, B. jonellus, B. mixtus, and B. sylvicola. To our knowledge, the discovery of nematodes in bumble bees from Fairbanks and Palmer represents the first report of this endoparasite from the state ( Table 4).
The origin of the endoparasites observed is unknown. In Ontario, Canada, higher Nosema prevalence has been associated with commerically raised bumble bees that escaped greenhouses, a phenomenon known as "pathogen spillover" (Colla et al. 2006). However, Plischuk et al. (2009) reported that bumble bees can become infected by Nosema from honey bees. Due to climatic conditions, Alaskan beekeepers have been importing honeybees into Alaska annually since the early days of beekeeping in the state (Washburn 1974) which may have introduced these parasites to the state. It is possible that Nosema occurred naturally in Alaska but without historical data predating beekeeping in Alaska, this is uncertain.
Both parasites, Nosema and nematodes, were identified from bee species collected in high numbers from Fairbanks and Palmer. However, few specimens of the relatively low abundance species were examined for endoparasites. In the nematode study in Fairbanks, we only examined bees hand-collected while resting; this may have skewed the results towards bees in poor health. Research is also needed to study the geographical extent of nematodes and Nosema infecting bumble bees in Alaska. Tetradonematid nematodes are obligate and fairly specific parasites, but are not considered common nematodes of bumble bees (Poinar 1975), stressing the need to further study this group in Alaska.
The western bumble bee, B. occidentalis, once considered to be one of the most common North American west coast bumble bee species, is declining in the Pacific North West (Rao and Stephen 2007, Colla and Ratti 2010, Cameron et al. 2011. In Alaska, this species represented roughly 10.4% of the total specimens collected (Tables 1, 2, 3) suggesting that B. occidentalis is a relatively abundant species in the areas studied. Among all Bombus specimens in the University of Alaska Museum (n=23,001), this species is the fourth most abundant behind (in order of abundance) B. bifarius, B. centralis, and B. jonellus. Koch and Strange (2012) also noted that B. occidentalis appears to be both widely distributed and relatively common species in Alaska. This species comprised 28% of the bumble bees in their survey, which included a total of 15 Bombus species. In addition to the sites studied, this species was collected from the Kenai Peninsula near Soldotna and Wiseman, Alaska (data not presented but specimens vouchered in UAM, and data available online [UAM 2013b]). However, B. occidentalis had the highest Nosema counts among the six species in which we detected this parasite (Table 4). Koch and Strange (2012) (Thorp et al. 1983), all of which occur in Alaska (Washburn 1963) and two species were documented in this survey; emphasizing the need to study the effect of social parasites on bumble bees in the state.
The earliest sampling date recorded was May 6; however, depending on sites and years, flight activity was detected during the first week sampling was initiated, suggesting that flight activity started before the snow melts. In this study we deployed traps as soon as snow melted; future studies should initiate sampling by mid-April, before the snow starts to melt.
The highest counts observed were in Delta Junction during 2009 with a mean number of 11.5 bumble bees per trap per day. No previous reports from Alaska provide comparative data to put these values in context. In Oregon, Stephen and Rao (2005) captured an average of 17.3 bees per day using the same blue vane trapping method.
Counts in Palmer were consistent between years with a difference of 276 bees between the two years ( Table 3). The relatively stable densities between years in Palmer can be explained by trap locations. During both years, traps were hung in close proximity to a patch of Rheum spp. that provided a long and consistent foraging source for bees. Information on the Rheum species available at the site, flowering patterns, and other plant characters were discussed by Pantoja and Kuhl (2009).
Depending on site and year, queens were the most abundant caste collected. The lowest collecting year was 2010 and the location with the fewest queens was Fairbanks where only 57 specimens were collected. Delta Junction displayed the highest overall queen density. It is reasonable to assume that the removal of queens during the previous season (2009) would reduce the overall bumble bee relative density during the following season (2010). However, this was not observed, more queens were captured in Delta Junction during 2010 than 2009 ( Table 2). The first year, sampling started when snow melted. We expect queens to emerge during drier conditions; however, to our surprise, queens were already emerging while snow was still present. In 2010, we set up traps a month earlier than the previous season and had several weeks of no activity before we began to see queens in the traps. The reduction in the relative population density of workers and males recorded for Delta Junction during the 2010 season as compared to the 2009 season cannot be explained by sampling or removal of the queens alone. In Oregon, Stephen and Rao (2005)  This report provides baseline data on species composition, distribution, seasonality, and parasites of the genus Bombus at the main agricultural areas in Alaska: Fairbanks, Delta Junction, and Palmer. Baseline data are needed to help understand reported patterns of bumble bee declines in North America (Cameron et al. 2011). Additional research is needed to better understand the biology, geographical distribution, contribution of bumble bees to Alaska agriculture, and the possible effects of endo-and social parasites on bumble bees in the state.