Australian Journal of Entomology (2011) 50, 174178

Comparison of yellow and white pan traps in surveys of bee fauna in New South Wales, Australia (Hymenoptera: Apoidea: Anthophila)
aen_797 174..178

John R Gollan,* Michael B Ashcroft and Michael Batley

Australian Museum, 6 College Street, Sydney, NSW 2010, Australia.

Abstract

Pan trapping is a standardised and commonly used method for collecting bees, but characteristics of the trap may inuence its effectiveness or bias results. The effect of trap colour on the species and numbers caught has been studied in the Northern Hemisphere, but not in the Australian region. Australia has a unique bee fauna and colour preferences, if any, may differ from those found in other continents. In four separate surveys across a wide area of New South Wales, it was tested whether there was a difference in the abundance or species richness captured by yellow- and white-coloured pan traps. In total, 1267 bees were collected, comprising 66 species, 50 of which are in the family Halictidae. In all surveys, yellow pan traps collected a signicantly larger number and greater diversity of bees. Eight of the thirteen most common species were observed signicantly more frequently (P < 0.01) in yellow pan traps, while the European honey bee (Apis mellifera) was found in larger numbers in white traps. Our results demonstrate that differently coloured traps collect different components of the Australian bee fauna. Therefore, a variety of pan colours should be used when sampling overall bee biodiversity, but specic colours may be more effective when targeting certain groups or species. coloured trap, method, survey, water trap.

Key words

I NTRODUCTIO N
A recent review of the diversity and conservation status of Australias native bee fauna (Apoidea: Anthophila) found that many species remain undescribed and more than half the named taxa were in need of revision (Batley & Hogendoorn 2009). Improving this situation requires both investment in taxonomic expertise and the development of efcient and effective methods for assessing bee distribution and abundance. Large-scale surveys of Australias existing bee fauna are needed as a crucial step in conservation planning (Tuell et al. 2009). The most common standardised collection protocol for bees uses pan traps, which are usually plastic bowls lled with soapy water (Leong & Thorp 1999). Bees approach the bowls, land on the water and drown. Pan trapping is viewed as an attractive choice over methods such as sweeping because it is simple, more time-efcient, not dependent on trained or skilled collectors and can accommodate different activity patterns between the species to be collected (Cane et al. 2000; Wilson et al. 2008). Characteristics of the trap, however, can inuence its effectiveness and the subset of species captured. In particular, colour is considered to have strong effect on the efciency of trapping. Colour is one of the important attractants used by owering plants (Kevan 1972), and bees are likely to respond to colours associated with oral rewards such as pollen and
*john.gollan@austmus.gov.au 2010 The Authors Journal compilation 2010 Australian Entomological Society

nectar (Waser & Price 1981). Studies of the effect of trap colour performed in North America produced mixed results. Some indicated that the majority of bee species had no signicant colour preference when combinations of blue, red, yellow and white were compared (e.g. Toler et al. 2005; Roulston et al. 2007; Wilson et al. 2008; Tuell et al. 2009), while other studies showed quantitative sampling rates were highly sensitive to trap colour (Leong & Thorp 1999; Campbell & Hanula 2007). Any colour preference of Australias bee species remains to be demonstrated. We conducted four surveys of bees throughout New South Wales, Australia, and tested whether there was a difference in the number of specimens trapped or the species richness (the number of different species) between yellow- and whitecoloured pan traps. We also examined differences in the number trapped between these trap colours for the 13 most frequently collected species.

MAT ERIAL S AN D MET H ODS

Surveys of bee fauna were conducted on four occasions: October 2008 (100 sites), November 2008 (94 sites), December 2008 (97 sites) and March 2009 (78 sites). Sample sites were within 40 m of roads, and spaced at distances of approximately 10 km (Fig. 1). Sites were chosen a priori using a Geographic Information System and satellite imagery (Google Earth) so that a wide range of conditions were sampled, including open elds, riverside habitats, forests and roadside verges.
doi:10.1111/j.1440-6055.2010.00797.x

Bees in yellow vs. white pan traps

175

Fig. 1. Map of survey locations.

Two yellow and two white pan traps (plastic bowls 18 cm in diameter and 5 cm deep were placed linearly where possible and approximately 5 m apart at each site. The order of colour along each trap line was random. All traps were half lled with salt water (approximately 3% concentration) and two to three drops of detergent, secured on the ground using wooden foodskewers, and collected after approximately 48 h. All bees were transferred to ethanol or pinned for identication. Specimens were identied to genus or species where possible, and all specimens apart from Apis mellifera were lodged at the Australian Museum, Sydney, Australia.

Analysis
Where there were no bees trapped and/or signs of trap disturbance, the sites were eliminated from analyses. This reduced the number of sites to 55 in October 2008, 72 in November 2008, 51 in December 2008 and 59 in March 2009. Samples from the same colour traps at each site were pooled and a two-tailed paired t-test was used to test for differences in total numbers trapped and species richness between yellow and white traps. Separate tests were carried out for each survey occasion to examine consistency of results. All data were log(x + 1) transformed prior to analyses. For the 13 most abundant species (those that were captured in at least 10 traps), binomial tests were used to test if the species was signicantly more likely to be observed in yellow or white pan traps. We used presence or absence in each trap to avoid results being inuenced by a large abundance in a small number of traps. The null hypothesis was that specimens were equally likely to be found in the different colour pans (n = number of traps where species was captured, P = 0.5).

Sixteen of the 66 species (24%) were collected in white traps but not in yellow, 30 species (45%) were collected in yellow traps but not white and 20 species (30%) were trapped in both colours. Of the total individuals, 957 (76%) were trapped in yellow pan traps (Table 1). The total number trapped and species richness of bees trapped in yellow traps was larger and signicantly different from that of white traps on each survey occasion (Table 2; Fig. 2). Log abundance was higher by between 66% and 161%, while log richness was higher by between 77% and 126%. The two most numerous species trapped were Homalictus sphecodoides and H. urbanus (26% and 22% of the total number trapped, respectively), while the introduced halictid, Halictus smaragdulus, was the third most common species trapped (10% of the total number trapped). The European honey bee (A. mellifera) and the native halictid Lasioglossum cognatum were the fourth and fth most frequently trapped species, respectively (7% and 4% of the total number trapped, respectively). Of the 13 most frequently trapped species, 12 were more commonly captured in yellow pan traps, and 8 signicantly so (P < 0.05; Table 3). The only common species that was more frequently recorded in white pan traps was A. mellifera.

DIS CUS S ION

Standardised sampling protocols should be used when developing large-scale monitoring programmes (Williams et al. 2001; Westphal et al. 2008) in order to increase taxonomic knowledge or establish the conservation status of species. Our surveys of bees across a large region of New South Wales, Australia, demonstrated clear differences in the numbers trapped and species richness caught by yellow and white pan traps. On each survey occasion, yellow pan traps collected
2010 The Authors Journal compilation 2010 Australian Entomological Society

RE SULTS
A total of 1267 individual bees were trapped comprising 66 species, 50 (76%) of which belong to the family Halictidae.

176

J R Gollan et al.

Table 1 Total number of bees trapped in yellow and white pans in four surveys across New South Wales, Australia (see Fig. 1 for dates of surveys and locations of survey sites)
Family Apidae Apidae Apidae Apidae Apidae Colletidae Colletidae Colletidae Colletidae Colletidae Colletidae Colletidae Colletidae Colletidae Colletidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Halictidae Megachilidae Total Genus (subgenus) species Apis mellifera Linnaeus, 1758 Ceratina (Neoceratina) australensis (Perkins, 1912) Exoneurella setosa Houston, 1976 Thyreus caeruleopunctatus (Blanchard, 1840) Trigona (Heterotrigona) carbonaria Smith, 1854 Euhesma (Euhesma) sp. Euryglossa depressa Smith, 1853 Hylaeus (Prosopisteron) asperithorax (Rayment, 1927) Hylaeus (Pseudhylaeus) albocuneatus (Cockerell, 1913) Hyphesma atromicans (Cockerell, 1913) Leioproctus (Leioproctus) advena (Smith, 1862) Leioproctus (Leioproctus) alleynae (Rayment, 1935) Leioproctus (Leioproctus) carinatus (Smith, 1853) Leioproctus (Leioproctus) maculatus (Rayment, 1930) Leioproctus sp. Halictus (Seladonia) smaragdulus Vachal, 1895 Homalictus (Homalictus) caloundrensis (Cockerell, 1914) Homalictus (Homalictus) ctenander Michener, 1965 Homalictus (Homalictus) holochlorus (Cockerell, 1914) Homalictus (Homalictus) punctatus (Smith, 1879) Homalictus (Homalictus) sphecodoides (Smith, 1853) Homalictus (Homalictus) urbanus (Smith, 1879) Lasioglossum (Chilalictus) athrix Walker, 1995 Lasioglossum (Chilalictus) brunnesetum Walker, 1995 Lasioglossum (Chilalictus) cambagei (Cockerell, 1910) Lasioglossum (Chilalictus) cephalochilum Michener, 1965 Lasioglossum (Chilalictus) chapmani (Cockerell, 1910) Lasioglossum (Chilalictus) clelandi (Cockerell, 1910) Lasioglossum (Chilalictus) cognatum (Smith, 1853) Lasioglossum (Chilalictus) convexum (Smith, 1879) Lasioglossum (Chilalictus) ebeneum Walker, 1995 Lasioglossum (Chilalictus) erythrurum (Cockerell, 1914) Lasioglossum (Chilalictus) expansifrons (Cockerell, 1914) Lasioglossum (Chilalictus) gilesi (Cockerell, 1905) Lasioglossum (Chilalictus) globosum (Smith, 1853) Lasioglossum (Chilalictus) gynochilum Michener, 1965 Lasioglossum (Chilalictus) helichrysi (Cockerell, 1914) Lasioglossum (Chilalictus) hemichalceum (Cockerell, 1923) Lasioglossum (Chilalictus) humei (Cockerell, 1905) Lasioglossum (Chilalictus) imitans (Cockerell, 1914) Lasioglossum (Chilalictus) impunctatum Walker, 1999 Lasioglossum (Chilalictus) instabilis (Cockerell, 1914) Lasioglossum (Chilalictus) lanarium (Smith, 1853) Lasioglossum (Chilalictus) macrops (Cockerell, 1916) Lasioglossum (Chilalictus) mundulum (Cockerell, 1916) Lasioglossum (Chilalictus) obscurissimum Michener, 1965 Lasioglossum (Chilalictus) orbatum (Smith, 1853) Lasioglossum (Chilalictus) pachycephalum (Cockerell, 1916) Lasioglossum (Chilalictus) quadratum Walker, 1995 Lasioglossum (Chilalictus) repraesentans (Smith, 1853) Lasioglossum (Chilalictus) sexsetum Walker, 1995 Lasioglossum (Chilalictus) sp. Lasioglossum (Chilalictus) tridens Walker, 1995 Lasioglossum (Chilalictus) victoriellum (Cockerell, 1914) Lasioglossum (Chilalictus) vitripenne (Smith, 1879) Lasioglossum (Chilalictus) willsi (Cockerell, 1906) Lasioglossum (Ctenonomia) sp. Lasioglossum (Parasphecodes) bryotrichum (Cockerell, 1912) Lasioglossum (Parasphecodes) clarigaster (Cockerell, 1918) Lasioglossum (Parasphecodes) olgae (Rayment, 1935) Lasioglossum (Parasphecodes) tribuarium (Rayment, 1935) Lasioglossum (Parasphecodes) turneri (Cockerell, 1914) Lasioglossum (Pseudochilalictus) imitator Michener, 1965 Lipotriches (Austronomia) cf avoviridis (Cockerell, 1905) Lipotriches (Austronomia) phanerura (Cockerell, 1913) Megachile (Eutricharaea) serricauda Cockerell, 1910 Yellow 26 0 0 2 1 0 0 0 0 1 2 1 0 1 11 81 32 1 0 1 296 210 4 3 4 4 18 29 39 11 1 6 11 15 16 1 39 6 0 3 0 1 27 1 1 1 3 2 0 3 2 0 14 9 0 0 1 3 0 0 1 1 1 2 7 1 957 White 65 1 1 0 0 1 1 2 1 0 0 0 1 0 2 43 1 0 2 0 33 71 0 2 0 0 4 1 18 1 0 0 7 0 10 0 10 0 1 1 1 1 4 0 0 0 0 0 1 0 0 1 7 0 2 2 0 0 2 1 0 0 0 7 1 0 310

2010 The Authors Journal compilation 2010 Australian Entomological Society

Bees in yellow vs. white pan traps

177

Table 2 Results of two-sided paired t-tests on four survey occasions to test for differences in numbers trapped and bee species richness between yellow and white pan traps
Survey occasion t October 2008 (n = 55) November 2008 (n = 72) December 2008 (n = 51) March 2009 (n = 59) 3.17 7.19 4.60 11.00 Number of bees trapped P <0.0100 <0.0001 <0.0001 <0.0001 t 3.81 5.78 3.93 9.43 Bee species richness P <0.0010 <0.0001 <0.0010 <0.0001

Table 3 The number of yellow and white pan traps that captured the 13 most frequently observed species
Species Homalictus (Homalictus) sphecodoides Homalictus (Homalictus) urbanus Apis mellifera Halictus (Seladonia) smaragdulus Lasioglossum (Chilalictus) cognatum Lasioglossum (Chilalictus) lanarium Lasioglossum (Chilalictus) helichrysi Homalictus (Homalictus) caloundrensis Lasioglossum (Chilalictus) clelandi Lasioglossum (Chilalictus) globosum Lasioglossum (Chilalictus) chapmani Lasioglossum (Chilalictus) expansifrons Lasioglossum (Chilalictus) gilesi Yellow traps 117 79 22 30 33 22 20 19 17 11 10 9 10 White traps 17 37 49 19 14 4 3 1 1 6 4 5 0 P <0.001 <0.001 <0.010 N.S. <0.010 <0.001 <0.001 <0.001 <0.001 N.S. N.S. N.S. <0.010

The probability, P, that the traps were equally effective was calculated using a binomial test. N.S. indicates non-signicant results.

Fig. 2. Mean ( SE) number of bees and species richness of bees trapped in yellow and white pan traps during four surveys, each of which included between 51 and 72 sites. The counts are the pooled total from two traps of each colour at every site. Data have been transformed (log(x + 1)).

more bees and a larger number of species than white traps. Twelve of the 13 most common species were trapped in higher numbers in yellow traps, but the European honey bee (A. mellifera) was more abundant in white traps. Our result showing that European honey bees were more often captured in white than yellow pan traps is consistent with a similar study of wild bees in North America (Tuell et al. 2009). White traps are therefore likely to be a useful, standardised and passive option for monitoring honey bees. Assuming that other species of Apis display a similar preference for white, white-coloured pan traps would be a useful method for the surveillance of the Asian honey bee (Apis cerana). The

Asian honey bee is considered a major threat to Australias honey industry as well as the feral honey bee colonies that are relied upon for pollinating many types of crops (Lawrence & Anderson 2007). White pan traps offer a simple supplement to the sentinel hives that have been used in major Australian ports for detecting incursions of pests of European Honey Bees. Our surveys only captured a fraction of the over 400 species that have been recorded from the overall region (M Batley unpubl. data 2009; KL Walker pers. comm. 2009) and the sampling appeared to be biased towards the halictids. Apart from site selection, any collection method introduces bias for a number of reasons (Westphal et al. 2008). Sweeping and pan traps, for example, have been considered as complementary methods in North American bee survey protocols because they capture an entirely different species composition (Wilson et al. 2008). Trap nests have been considered complementary to pan traps in studies of European fauna because pan traps failed to detect cavity-nesting bees (Westphal et al. 2008). Nonetheless, a systematic evaluation of six different methods for surveying bee fauna concluded that pan traps were the best indicator of overall bee species richness (Westphal et al. 2008), and therefore it is important to assess the use of pan traps in Australia. Comparison of ground-level yellow and white pan traps is only one of many possible variations. Traps could be placed at different heights above the ground, be a wide variety of colours, contain scented detergents or optical brighteners, or be coated with UV paints (e.g. http://online.sfsu.edu/
2010 The Authors Journal compilation 2010 Australian Entomological Society

178

J R Gollan et al. REF EREN CES

Batley M & Hogendoorn K. 2009. Diversity and conservation status of native Australian bees. Apidologie 40, 347354. Campbell JW & Hanula JL. 2007. Efciency of Malaise traps and colored pan traps for collecting ower visiting insects from three forested ecosystems. Journal of Insect Conservation 11, 399408. Cane JH, Minckley RL & Kervin LJ. 2000. Sampling bees (Hymenoptera: Apiformes) for pollinator community studies: pitfalls of pan-trapping. Journal of the Kansas Entomological Society 73, 225231. Kevan PG. 1972. Floral colors in the high arctic with reference to insect ower relations and pollination. Canadian Journal of Botany 50, 22892316. Lawrence L & Anderson D. 2007. Exotic pests pose pollination threat. Farming Ahead 180. [Accessed 14 Dec 2009.] Available from URL: http://www.csiro.au/les/les/pc11.pdf Leong JM & Thorp RW. 1999. Colour-coded sampling: the pan trap colour preferences of oligolectic and nonoligolectic bees associated with a vernal pool plant. Ecological Entomology 24, 329335. Roulston TH, Smith SA & Brewster AL. 2007. A comparison of pan traps and intensive net sampling techniques for documenting a bee (Hymenoptera: Apiformes) Fauna. Journal of the Kansas Entomological Society 80, 179181. Stephen WP & Rao S. 2007. Sampling native bees in proximity to a highly competitive food resource (Hymenoptera: Apiformes). Journal of the Kansas Entomological Society 80, 369376. Toler TR, Evans EW & Tepedino VJ. 2005. Pan-trapping for bees (Hymenoptera: Apiformes) in Utahs west desert: the importance of color diversity. Pan-Pacic Entomologist 81, 103113. Tuell JK, Ascher JS & Isaacs R. 2009. Wild bees (Hymenoptera: Apoidea: Anthophila) of the Michigan highbush blueberry agroecosystem. Annals of the Entomological Society of America 102, 275287. Waser NM & Price MV. 1981. Pollinator choice and stabilizing selection for ower colour in Delphinium nelsonii. Evolution 35, 376390. Westphal C, Bommarco R, Carr G et al. 2008. Measuring bee diversity in different European habitats and biogeographical regions. Ecological Monographs 78, 653671. Williams N, Minckley RL & Silveira FA. 2001. Variation in native bee faunas and its implication for detecting community changes. Ecology and Society 5 (1), 7. [Accessed 14 Dec 2009.] Available from URL: http://www.ecologyandsociety.org/vol5/iss1/art7/ Wilson JS, Griswold T & Messinger OJ. 2008. Sampling bee communities (Hymenoptera: Apiformes) in a desert landscape: are pan traps sufcient? Journal of the Kansas Entomological Society 81, 288300. Accepted for publication 18 October 2010.

~beeplot/). Other congurations are also worth testing in Australia, and comparisons with blue pan or vane traps may be of particular interest given their success in other countries (Stephen & Rao 2007; Westphal et al. 2008). Our study provides some insights into the sample sizes that might be required to detect differences between these other trap congurations. Two traps of each colour at 369 sites was more than sufcient to detect differences in numbers trapped and species richness. A single survey of 78 sites was enough given the large difference between the two colours. However, even with the full complement of 369 sites, we could only detect differences between trap colours for 9 of 66 species. If only numbers trapped and species richness are of interest, it may be possible to reduce sample sizes, but sample sizes may need to be increased to assess how effective pan traps are at collecting individual species. Results for individual species results may be of particular interest for detecting rare, threatened or introduced species. Pan traps are an efcient sampling technique but white pan traps are less effective for most species and collect fewer species than yellow pan traps. We recommend using a variety of pan colours (potentially included those not testing in our study) when sampling overall bee biodiversity, white pans when sampling European honey bee, or yellow pans when targeting Homalictus or Lasioglossum species or the recently introduced emerald furrow bee (H. smaragdulus).

ACKNOWLE D G E ME NTS
We thank the Winifred Violet Scott Charitable Trust for funding this work. Natalie Sullivan assisted with eld work and laboratory sorting. Scott Ginn assisted with data basing and incorporating specimens into Australian Museums collection. Thanks to Dr KL Walker (Museum Victoria) for comments on an earlier version of the manuscript and assistance with identications.

2010 The Authors Journal compilation 2010 Australian Entomological Society