Changes in Wild Bee Fauna Changes in the Wild Bee Fauna of Rockefeller Prairie

JENNIFER L. HOPWOOD1, 2 University of Kansas, Lawrence, KS 66045

1 Present address: Midwest Pollinator Outreach Coordinator, The Xerces Society for Invertebrate Conservation, PO Box 229, St Louis MO 63166 2 [email protected]

Abstract: Bees are considered to be the most important pollinators in most ecosystems, including tallgrass prairie. While recent research highlights the possibility of a global decline in pollinators, the lack of baseline data, species lists and well-documented collections of pollinators for many localities makes assessing anthropogenic impacts difficult. However, wild bees were sampled in 1978 at Rockefeller Prairie near Lawrence, Kansas, which provides one useful source of baseline data. Management of Rockefeller Prairie, a small but high quality prairie remnant, has been consistent since the 1978 collection. The land surrounding Rockefeller Prairie has undergone minor changes (different agricultural practices, secondary succession). The goal of this study was to revisit Rockefeller Prairie to make comparisons between the past and current bee fauna. I re-sampled the site by collecting bees from flower species previously sampled in 1978, as well as other flower species not previously sampled. I also employed pan traps, a more recent passive method of collection. Pan traps are colored plastic bowls filled with soapy water that attract flower-visiting . I found little overlap between my collection and the bee communities of the past collection. The species composition of my sample was markedly different from past samples, with 24 bee species previously found in Rockefeller Prairie in 1978 that were not recollected in 2004. Most notably missing from the 2004 collection was the most abundant bee in the previous collection, Bombus pensylvanicus, a species that has become less widespread across the Eastern United States in recent years.

Proceedings of the North American Prairie Conference 21:139-145

Keywords: bees, Bombus pensylvanicus, pollinator ecology, prairie forbs, re-sampling.

Introduction 2000). Honey bees (Apis mellifera) contribute pollinators are integral to the life history about $14.6 billion in crop pollination each year in of many flowering plants, providing an essential the United States (Morse and Calderone 2000) and wild bee species contribute $3.07 billion to fruit and ecosystem service in both natural plant communities vegetable pollination (Losey and Vaughan 2006). and agricultural ecosystems. Bees are considered Suitable bee habitat must include a diversity of to be the most important group of pollinators, flowering species and nesting substrates, because possessing structural and behavioral adaptations bees have a wide range of specialized floral and that enable them to pollinate a wide variety of plants nesting requirements. Bees of both sexes may successfully (Delaplane and Mayer 2000, Michener exploit many floral species for nectar, and females of 2000). Bees are important in the maintenance of social species and some solitary species also utilize diversity in wild flowers (Kearns et al. 1998), in the a diversity of forbs for pollen. Female bees of other conservation of threatened plant species (Sipes and solitary species are oligolectic, visiting only one or Tepedino 1995), and in the reproduction of numerous several closely related flowering species for their agricultural crop species (Delaplane and Mayer specialized pollen needs (Wcislo and Cane 1996).

Proceedings of the 21st North American Prairie Conference (2010) 139 Prairie Entomology and Zoology

The nesting preferences of bees are highly more than 200 plant species, including the federally variable, and while many species dig subterranean endangered Mead’s milkweed (Asclepias meadii) nests in their preferred location and soil type, some and Western prairie fringed orchid (Platanthera species nest above ground in plant stems or dead praeclara) (K. Kindscher, pers. comm.). From the wood, or may need specific nest-building materials 1870’s until 1956, Rockefeller Prairie was maintained such as mud, snail shells or resin (Linsley 1958). by annual mowing for hay, and subsequent burning Prairies, which provide a diversity of flowering of the remaining stubble (Kindscher 1992). The species and nesting substrates, can be suitable habitat University of Kansas acquired the land in 1956 and for bees. the prairie was burned every three years until the Globally, researchers have begun to suspect an late 1960’s, when woody plants from the nearby impending pollination crisis in both natural and native oak-hickory forest bordering a portion of the agricultural ecosystems, due to potential declines in prairie began invading. Since that time, the prairie pollinator populations (Buchmann and Nabhan 1996, has been burned in one to two year intervals, and Allen-Wardell et al. 1998). Habitat fragmentation, most recently (1986 onwards) has been burned invasive plant and animal species, and extensive in the spring every other year (Kettle et al. 2000). usage of herbicides and pesticides have been Rockefeller Prairie was burned prior to this study in identified as some of the key threats that pollinators the spring of 2004. As it was in 1978, Rockefeller face (Allen-Wardell et al. 1998, Kearns et al. 1998). Prairie remains bordered by forest, cultivated fields, Losses of wild bees have been documented in some open secondary vegetation and prairie restoration regions of the world (Williams 1982, Biesmeijer et (Kettle et al. 2000), although in the interim between al. 2006), with declines suspected elsewhere (NRC 1978 and 2004, different agricultural practices were 2007). employed, and secondary succession progressed in While the influence of humans on flora and non-cultivated areas. fauna is of considerable interest, it can be difficult to document anthropogenic impacts because detailed Bee sampling historic data from specific localities are rare. The lack During the 1978 study, Laroca sampled bees of baseline data, species lists and well-documented Table 1. List of Prairie Forbs sampled for bees. collections of pollinators for many localities make historic and contemporary comparisons difficult Common name Scientific name (Marlin and LaBerge 2001, NRC 2007). One valuable source of baseline collections was made Lead plant Amorpha canescens near Lawrence, Kansas, U. S. A. by Sebastio Laroca Partridge pea Chamaecrista fasciculata Sumac Rhus glabra* (1983), then a doctoral student at the University of Purple prairie clover Dalea purpurea Kansas. Laroca’s study area included a high quality Pale purple coneflower Echinacea pallida tallgrass prairie remnant, Rockefeller Prairie. I Gray-headed coneflower Ratibida pinnata revisited Laroca’s study in 2004 to re-sample the Mountain mint Pycnanthemum tenuifolium bees to assess changes in abundance and bee species Black-eyed susan Rudbeckia hirta richness after 26 years. A high degree of similarity Compass plant Silphium laciniatum between samples from Rockefeller Prairie in 1978 Indian hemp dogbane Apocynum cannabinum** Butterfly milkweed Asclepias tuberosa** and 2004 was expected, as the management of the Blue sage Salvia azurea** prairie preserve remained consistent. Stiff tickseed Coreopsis palmata** Tall boneset Eupatorium altissimum** Methods Canada goldenrod Solidago canadensis** Study site Rattlesnake master Eryngium yuccifolium** Rockefeller Prairie is a 4.6-hectare native tallgrass * denotes flowers sampled only in 1978; prairie located within the Rockefeller Experimental ** denotes flowers sampled only in 2004 Tract in Jefferson County, Kansas. The prairie has

140 Proceedings of the 21st North American Prairie Conference (2010) Changes in Wild Bee Fauna in Rockefeller Prairie by collecting for one hour on seven sampling dates from June to August in Table 2. Bee species collected at Rockefeller Prairie in 1978 and 2004. 1978 (Laroca 1983). Bees were re-collected at Rockefeller on seven sampling dates from June Family Bee species through September in 2004. I sampled in 2004 using Colletidae * Colletes latitarsus an aerial net to collect bees off blooming flowers for ** Hylaeus affinis one hour on each sample date. Hylaeus modestus Andrenidae Andrena helianthi In both 1978 and 2004, bees were only sampled Perdita ignota on days with warm temperatures (above 70ºF), * Augochlora pura abundant sun (less than 50% cloud cover), and calm Augochloropsis metallica winds (less than 5 m/s). In 1978, Laroca sampled ** Augochlorella aurata bees from nine species of prairie flowers, a subset of ** Halictus ligatus * Halictus parallelus prairie forbs (Table 1). Lasioglossum bruneri In 2004, I swept bees off all prairie flowers in * Lasioglossum rowheri bloom, but made a distinct effort to sweep bees from Lasioglossum obscurum the forb species Laroca had previously sampled * Lasioglossum imitatum (with the exception of sumac, which does not bloom ** Lasioglossum coreopsis * Lasioglossum disparile in the season following a burn). Pan traps were * Lasioglossum nr oblongum also implemented to supplement the collection of * Lasioglossum nr paradmirandum bees. Pan traps are colored plastic bowls filled with Lasioglossum tegulare water and unscented dish soap, and bees attracted Lasioglossum versatum to the color hit the water and drown. Six colors of * Lasioglossum sp. A Nomia bakeri pan traps, light blue, dark blue, fluorescent blue, * Sphecodes sp. A fluorescent yellow, yellow and white, were used in Megachilidae ** Heriades carinata order to attract a variety of bees. Pan traps were left * Megachile apicalis out for six hours, along four 50-m transects, with the Megachile brevis colors randomized and 15 pan traps on each transect. Megachile mendica Megachile petulans Due to the inability to accurately identify most Megachile rugifrons bees on the wing, bees were killed, pinned and labeled Anthophora walshii with flower host or pan trap color. I identified the bees * Bombus auricomis using Michener, McGinley and Danforth (1994), * Bombus bimaculatus Mitchell (1960, 1962) and the reference collection Bombus fraternus ** Bombus griseocollis housed in the Snow Entomological Collection at the * Bombus impatiens University of Kansas. Dr. Charles Michener aided Bombus pensylvanicus in bee identification and Jason Gibbs identified Ceratina calcarata Lasioglossum species. Domesticated European Melissodes bimaculata honey bees (Apis mellifera), although present in Melissodes communis atripes both 1978 and 2004, were not considered in this Svastra obliqua study, as their abundance is primarily determined by Svastra petulca the location of their managed hives. Xenoglossodes albata Xenoglossodes helianthorum Results * Xylocopa virginica In 1978, Laroca collected 219 bees from nine species of prairie flowers and an additional 11 in the * denotes bees found only in 2004, air, for a total of 230 individual bees (Laroca 1983). ** denotes bees found both in 1978 and 2004. Allowing for taxonomic changes, Laroca collected 30 currently valid bee species in 1978. During the

Proceedings of the 21st North American Prairie Conference (2010) 141 Prairie Entomology and Zoology re-collection in 2004, I found 194 individuals (108 2004. Detailed statistical comparisons between the collected from flowers, 84 in pan traps, 2 in the air) 1978 collection and the 2004 re-collection cannot be of 21 species. Of the 21 species collected in 2004, 15 made because of a lack of quantitative data reported were species not found by Laroca, bringing the total by Laroca (1983). However, similarity indices such number of bee species collected so far at Rockefeller as the Jaccard index and the Sorenson index can Prairie to 45 species. Of the species collected by be used to measure the similarity of communities Laroca, only 20% of species were re-collected in based on the presence or absence of species. These 2004 (Table 2). similarity indices measure the amount of similarity The differences between 1978 and 2004 are between bee communities, and if the indices equal particularly striking when just considering the bees one, the collections are completely similar, whereas collected in 2004 from the same prairie plants sampled if the indices equal zero, the collections share no in 1978. Only 10 bee species and 40 individual bees species. For bees found on the 1978 eight study were collected from lead plant, partridge pea, purple plants, the Jaccard index value was 0.09, and the prairie clover, pale purple coneflower, gray-headed Sorenson index was 0.15. When considering all coneflower, black-eyed susan and compass plant in bee species in the 2004 resample, including bees

110 100 1978 90 2004 80 70 60 50 40 Bee abundance abundance Bee (# of individuals) of (# 30 20 10 0

Hylaeus affinis Svastra atripesSvastra obliqua Augochlora pura Megachile brevis Ceratina calcarataHalictus parallelus Megachile mendica AugochlorellaBombus aurata griseocollis Lasioglossum rowheri Xenoglossodes albata Bombus pensylvanicus Lasioglossum disparile Lasioglossum obscurum

Lasioglossum nr oblongum

Figure 1. Comparative bee abundances in 1978 and 2004 of predominant bee species (those species of which four or more individuals were collected).

142 Proceedings of the 21st North American Prairie Conference (2010) Changes in Wild Bee Fauna

Table 3. Number of individual bees that were captured on various plant families at Rockefeller Prairie in 1978 and 2004.

Individual bees (grouped by bee families) No. of Total # plant of bee Plant family Colletidae Andrendidae Halictidae Megachilidae Apidae species individuals visited visiting 1978 Fabaceae 3 1 0 9 4 116 130 Asteraceae 4 3 3 5 4 25 40 Lamiaceae 1 1 0 7 4 11 23 Anacardiaceae 1 5 0 7 4 10 26 Total 9 10 3 28 16 162 219 2004 Fabaceae 3 0 0 19 1 9 29 Asteraceae 7 2 0 13 0 4 19 Lamiaceae 2 0 0 3 0 3 6 Asclepiadaceae 1 0 0 1 0 5 6 Apocynaceae 1 0 0 1 1 0 2 Apiaceae 1 2 0 42 0 2 46 Total 15 4 0 79 2 23 108 collected by pan traps and off an additional seven family in 2004. floral species, the samples were found to be slightly Other interesting shifts include reduced stem- more similar, with a Jaccard index of 0.13 and a nesting bees. While bees that nest in the ground Sorenson index of 0.23. are generally more abundant in North America than Overall, the collections made in 1978 and cavity or stem-nesting bees, only three species seen in 2004 had very little overlap and showed little in 2004 were stem-nesters, down from the eight similarity. Bee abundances were also drastically species observed in 1978. The six species common different between the two collections (Figure 1). to both the 1978 and 2004 collections (Table 2) The most abundant bee in 1978 was the bumble are all considered to be generalist species that visit bee Bombus pensylvanicus, a primitively eusocial a variety of flowers for pollen, and are commonly species that comprised about 44% (100 individuals) found in bee collections. of the collection. Surprisingly, B. pensylvanicus was Discussion entirely absent from the 2004 collection. In 2004, The passing of 26 years was not expected to the most abundant bee was a small semi-social have such a drastic impact on the bee community sweat bee, Lasioglossum rowheri, a species not seen of Rockefeller Prairie as was observed, particularly in 1978. as Marlin and LaBerge (2001) found a surprisingly Bees were primarily collected on plants in the similar bee fauna in the early 1970s after a 75-year bean plant family Fabaceae in 1978, but in 2004 more lapse near Carlinville, IL. Yet 24 bee species found bees visited plants in the family Apiaceae (Table in Rockefeller Prairie in 1978 were not re-collected 3). Rattlesnake master, in the family Apiaceae, is in 2004, and the species composition in 2004 was particularly attractive to both short and long-tongued markedly different from 1978. The bee family bees seeking its nectar and pollen and was relatively Apidae dominated the 1978 sample, in part because abundant in 2004. The predominant bee family in the bumble bee B. pensylvanicus was so incredibly 1978 was Apidae (includes bumble bees, carpenter abundant. A particularly striking find of the 2004 bees, long-horned bees), whereas Halictidae re-sampling was that not a single B. pensylvanicus (includes sweat bees) was the predominant bee individual was observed. While declines in bumble

Proceedings of the 21st North American Prairie Conference (2010) 143 Prairie Entomology and Zoology bee populations have been documented in other rainy or unusually cool days). There was a positive countries (Williams 1982, Biesmeijer et al. 2007), correlation in 2004 between the number of individual declines have only just begun to be observed in bees within a species and the number of floral North America due to lack of baseline data. Declines species upon which the bees were found, indicating of B. pensylvanicus have been noted in some areas that those bees that were generalists visited many across its wide distribution (populations historically flowers and were thus caught most frequently. Thus, extended from the northeast coast west to the Great specialist bees might have been missed if collection Plains and south into Mexico), although healthy was limited on their host plant or if density of that populations are still present in some areas (Colla and plant was low. In addition, Laroca considered 19 of Packer 2008). Several other bumble bee species once the bee species caught in 1978 to be rare (less than abundant across a wide geographical distribution are 2.2% abundance), making their recollection that noted to be in decline in North America (Colla and much more difficult. Generally, many rare species Packer 2008), with habitat loss, pesticide use, and (species with low abundance or not found in multiple pathogen spillover from commercially reared bumble locations) are found in bee collections, and it is these bees used for greenhouse pollination contributing to rare species that usually make collections so variable the decline (NRC 2007). (Williams et al. 2001). For example, separate However, despite purported bee population samples made on creosote bush shared most of the declines, the observed changes in the bee fauna might common species, but only 20% of the rare species be attributed to several additional traits of bees. (Williams et al. 2001). The 45 species recorded in Firstly, the presence of a plant on which to forage Rockefeller prairie is likely an underestimate of the does not necessarily guarantee that the associated actual bee fauna, as is suggested by the numerous bee will be found. The plant must be located within species that are represented by only one individual. the flight range of the bee, and solitary bees have Finally, very little is known of population limited flight ranges. Many bees use partial habitats, dynamics, and it is possible that bee populations nesting in one location and foraging elsewhere fluctuate greatly from year to year, as evidence (Westrich 1996). Bees may be foraging for food in suggests (Williams et al. 2001). Intense sampling the prairie and nesting outside the prairie. This may for more than one season is likely required to learn be particularly true for soil-nesting bees that prefer to more about the overall bee diversity of Rockefeller nest in exposed bare ground. Rockefeller Prairie has Prairie, and to detect long-term population trends. limited patches of bare soil, as the dominant grasses Future sampling is needed, to build on Laroca’s are bunch grasses that are very dense in places. The historical data. Prairie remnants and restorations land surrounding Rockefeller has undergone some are excellent places to sample pollinators, as these changes from 1978 to 2004. Thus, bees that may areas are often protected from development and are previously have nested in nearby hay meadows managed with care. As the landscape surrounding present in 1978 were evicted when the fields were these islands of prairie continues to change, wildlife later tilled for cultivation, and could no longer fly to becomes even more dependent upon the resources the prairie to forage for food. of the prairie. Efforts are needed to continue to The weather and the time of day when collections sample pollinators, to contribute insights towards take place are also important factors that influence understanding the responses of important pollinators bee diversity. Although I took care to collect bees to anthropogenic changes. under similar conditions as Laroca did, 2004 was Acknowledgements an unusually cool and wet summer (NOAA 2004). I thank C. Taylor for his guidance, C. Michener The summer of 1978 was, by contrast, warmer for aiding in bee identifications, J. Gibbs for and drier. The high rainfall in spring and early identifying Lasioglossum species, T. Dickson and two summer of 2004 may have kept certain bee species anonymous reviewers for manuscript comments and from emerging from their nests in the soil, or led M. Ramspott for current and past landscape photos to decreases in population sizes due to the lack of of Rockefeller and its surroundings. This work was foraging opportunities (bees typically do not fly on supported by grants from the Entomology Program

144 Proceedings of the 21st North American Prairie Conference (2010) Changes in Wild Bee Fauna at the University of Kansas and the University of value of ecological services provided by insects. Kansas Field Station and Ecological Reserves. Bioscience 56: 311-323. Thanks to KSR and D. Kettle for permission and Marlin, J. C. and W. E. LaBerge. 2001. The native support to work at Rockefeller Prairie. bees of Carlinville, Ill. Revisited after 75 years: a case for persistence. Conservation Ecology Literature Cited 5(1):9. Allen-Wardell, G., Bernhardt, P., Bitner, R., Burques, Michener, C. D., 2000. The Bees of the World. The A., Buchmann, S., Cane, J., Cox, P. A., Dalton, Johns Hopkins University Press, Baltimore and V., Feinsinger, P., Ingram, M., Inouye, D., Jones, London. C. E., Kennedy, K., Kevan, P., Koopowitz, H., _____, McGinley, R. J. and B. N. Danforth. 1994. Medellin, R., Medillin-Morales, S., Nabham, The Bee Genera of North and Central America. G. P., Pavlik, B., Tepedino, V., Torchio, P., and Washington, D.C. S. Walker. 1998. The potential consequences Mitchell, T. B., 1960. Bees of the Eastern United of pollinator declines on the conservation of States. Vol.. I. North Carolina Agricultural biodiversity and stability of food crop yields. Experiment Station Technical Bulletin 141:1- Conservation Biology 12:8-17. 538. Biesmeijer, J. C., Roberts, S. P. M., Reemer, M., _____. 1962. Bees of the Eastern United States. Ohlemuller, R., Edwards, M., Peeters, T., Vol. II. North Carolina Agricultural Experiment Schaffers, A. P., Potts, S. G., Kleukers, R., Station Technical Bulletin 152:1-557. Thomas, C. D., Settele, J. and W. E. Kunin. 2006. Parallel Declines in Pollinators and Insect- Morse, R. A. and N. W. Calderone. 2000. The value Pollinated Plants in Britain and the Netherlands. of honey bees as pollinators of U.S. crops in Science 313:351-354. 2000. Bee Culture 128 (3):1-15. Buchmann, S. and G. P. Nabhan. 1996. The Forgotten National Climatic Data Center, 2005. Climatological Pollinators. Washington, D.C. Data Annual Summary for Kansas. Colla, S. and L. Packer. 2008. Evidence for National Oceanic and Atmospheric Administration decline in eastern North American bumblebees 119, no. 13. (: Apidae), with special focus National Research Council, 2007. Status of Pollinators on Bombus affinis Cresson. Biodiversity and in North America. The National Academy of Conservation 17:1379-1391. Sciences, Washington D.C. Delaplane, K. S. and D. F. Mayer. 2000. Crop Sipes, S. and V. Tepedino. 1995. Reproductive Pollination by Bees. CAB International, New biology of the rare orchid, Spiranthes diluvialis: York. breeding system, pollination, and implications for conservation. Conservation Biology 9:929- Kearns, C., Inouye, D. W. and N. M. Waser. 1998. 938. Endangered mutualisms: The conservation of plant-pollinator interactions. Annual Review of Wcislo, W. T. and J. Cane. 1996. Floral resource Ecology and Systematics 29:83-112. utilization by solitary bees (Hymenoptera: Apoidea) and exploitation of their stored foods by Kettle, D.W., Rich, P. M, Kindscher, K., Pittman, natural enemies. Annual Review of Entomology G. L. and P. Fu. 2000. Land-Use History in 41:195-224. Ecosystem Restoration: A 40-year Study in the Prairie-Forest Ecotone. Restoration Ecology Westrich, P. 1996. Habitat requirements of central 8:307-317. European bees and the problems of partial habitats. In: Matheson, A., Buchmann, S.L., Kremen, C. and T. Ricketts. 2000. Global perspectives Toole, C., Westrich, P., Williams, P.H. (Eds). on pollination disruptions. Conservation The Conservation of Bees, Linnean Society Biology 14:1226-1228. Symposium Series, vol. 18. Academic Press, Laroca, Sebastio, 1983. Biocoenotics of Wild Bees London. (Hymenoptera, Apoidea) at Three Nearctic Sites, Williams, N. M., R. L. Minckley and F. A. Cilveira. with Comparative Notes on some Neotropical 2001. Variation in Native Bee Faunas and its Assemblages. University of Kansas Doctoral Implications for Detecting Community Changes. Dissertation. Conservation Ecology 5(1):7. Linsley, E.G., 1958. The ecology of solitary bees. Williams, P. H. 1982. The distribution and decline of Hilgardia 27 (19): 543-599. British bumblebees (Bombus Latr.). Journal of Losey, J. E. and M. Vaughan. 2006. The economic Apicultural Research 21 (4):236-245.

Proceedings of the 21st North American Prairie Conference (2010) 145