This post is written by Kit Prendergast, a native bee ecologist and conservation biologist.

Pollinators have been making headlines, with numerous articles in both the academic and popular press reporting that across the globe insect populations are declining. Of all insects, wild bees are vital in terms of their service to the ecosystem — pollination. Beyond that, they are a wonderfully diverse group of organisms with an intrinsic source of value.

There is, therefore, a need to document and monitor wild bees: How are they faring? What habitat types are important for their conservation? How are activities such as mowing, agro-forestry, pesticide application, habitat conversion, pesticides, fragmentation and revegetation — as well as factors like climate change and introduced bee species — impacting wild bees? Are management actions aimed at helping bees, such as agri-environment schemes, having their desired effects in conserving bees and promoting pollinators? To answer any of these questions, it is critical that the right sampling methods are used. Use the wrong methods, and the results will be invalid, and conclusions questionable, if not erroneous and misleading.

A variety of methods are currently in use to survey insects such as bees. Active methods include counting the number of bees visiting flowers and catching bees with an entomological sweep net. Bees can be collected from flowering plants present in the environment, or from phytometers or “mobile gardens” —essentially, standardized arrays of potted plants placed at sites to attract bees. Passive survey methods include putting out colored pan traps (“bee bowls”) of various sizes and colors, as well as vane traps, which are currently available in blue and yellow and consist of two sheets of plastic at right angles that funnel insects into a container. Native bees can also be monitored based on their nesting occupancy.

Which sampling method is best?

I’ve been conducting bee surveys in the urbanized region of southwest Western Australia, a biodiversity hotspot. This region is renowned for its wildflowers, many of which are only found in this part of the world. Yet, the native bee fauna here is poorly known, and no systematic surveys have been conducted. Like much of the world, this region is under threat of ongoing urban development. On top of the threat of urbanization, wild bees may also be impacted by European honey bees, Apis mellifera. This managed species has been introduced across the globe, including in Australia and, like many introduced species, can harm the native fauna. As Australia is free of Varroa mites and their associated diseases, Australia has a thriving population of both domesticated and feral honey bees. I wanted to know the composition of bees in the region, how they responded to urbanization, and if honey bees were impacting the wild indigenous bees. However, when I started planning my studies, I realized there was no consensus about which method was best to accurately and comprehensively survey wild bee communities.

As part of my research, I put different methods to the test. Over four months in the Australian spring and summer, I conducted surveys for native bees at fourteen sites: seven were urban natural vegetation remnants, and seven were residential (home) gardens. During each survey, I spent three hours recording all the bees I saw (observational counts) and collected them with a sweep net. At the same time, I set out bee bowls (20 soufflé cups painted UV-blue and UV-yellow, mounted on stakes and filled with water and colorless, odorless surfactant, and nine yellow party bowls laid on the ground) and four pot plants (“mobile gardens”). In addition, I hung two yellow and two blue vane traps on tree branches and installed eight bee hotels, all of which were checked monthly. The bee hotels (also known as “trap nests” in academic literature) were blocks of wood with fifteen holes (five each with a diameter of 4, 7, and 10 millimeters, and all 10 centimeters deep) providing an artificial nesting habitat for bees that naturally nest in holes made by wood-boring beetles. To minimize inter-observer bias, I performed all the surveys on my own — quite an undertaking, but there’s no better way to spend one’s days than out among flowers looking for bees!

So, what did I find?

The results of the surveys were published in the journal Ecosphere under the title, “The Relative Performance of Sampling Methods for Native Bees: An Empirical Test and Review of the Literature,” co-authored with my supervisors Myles Menz, Ph.D., Kingsley Dixon, Ph.D., and Philip Bateman, Ph.D.

During the survey, observational counts exceeded the collection methods, meaning that I could never catch everything that was present (which is a good thing, as I didn’t want to further deplete populations). However, the relative ratio of observed versus collected varied according to the particular bee taxonomic group: some bees are easier to catch than others.

Of collection methods, sweep netting vastly outperformed all other methods in terms of the number of bees and number of species collected. The majority of native bee species were collected exclusively with a sweep net. Of the most numerous bee species, the superiority of sweep netting over all passive methods held, except for Amegilla chlorocyanea, which was collected most frequently in blue, but not yellow, vane traps. Amegilla are the pandas of the bee world: cute, large, and buzzy. They are a common species in my study region and are generalists and good pollinators that forage on a wide range of plants and perform buzz-pollination (in which the anthers of plants like tomatoes must be vibrated at a particular frequency for the pollen to be released — something Amegilla excel at, while honey bees cannot do).

 

I also found that there were huge disparities in the representation of honey bees and native bees according to sampling method used. Honey bees outnumbered all native bee taxa combined when it came to observations, but this wasn’t the case for the passive sampling methods. This has implications when it comes to investigating the potential for honey bees (often an introduced species) to exert negative impacts on indigenous bees.

The mobile gardens were, to be lucid, a complete flop. Only a few bees visited them, and most of these were honey bees.

The intake of bee hotels was encouraging. They enabled me to measure nesting success and even document new nesting behaviors (including fights over nesting holes; the first bee in the world to use Banksia fuzz in its nest; and an introduced bee and wasp). However, bee hotels can only be used to document a small subset of the total bee community (since the majority of bee species nest in the ground). Furthermore, I found that of all the potential cavity-nesting bees present during my surveys, only a small subset of these nested in the bee hotels. Interestingly, their relative representation in bee hotels differed markedly from their representation based on collections in the field.

To place my results in context, I conducted a literature review of studies that compared the relative effectiveness of two or more methods of surveying bees. My review revealed a lot of variation among studies, which can be related to different sampling durations, study systems (e.g., monocultural fields vs. natural landscapes) and biogeography. However, I found that sweep netting consistently outperformed other methods.

What does this mean?

This study raises concerns about the large number of studies that use only passive methods (typically bee bowls). Moreover, there’s evidence that bee bowls catch more bees in relatively poor habitats (e.g., resource-sparse agricultural habitats). So, if bee bowls are used to look at how native bees are affected by disturbance and to compare habitat types, a completely inaccurate picture could result.

While counting bees provided the highest results in terms of abundance, a major drawback is that it is impossible to identify bees to species level based on observations alone. This means that vital taxonomic information is lost if specimens are not collected. In Australia, the majority of bees are less than a centimeter long, and with so many species that have yet to ever be described, collections are indispensable.

Bee bowls have often been used because they are cheap, cost-effective, and require no skill; just put them out in the field, leave them there for a few hours or a few days, and then collect them. In contrast, sweep netting requires a good eye and quick reflexes, and you need to be in the field (which really isn’t a bad thing at all, especially in the beautiful bushland remnants of southwest Western Australia!). But my research shows that you can’t short-cut a good collection effort, and sweep netting by experienced bee scientists is the best method to accurately document the full suite of bee biodiversity for a study.

You can read the full publication here in the open access journal Ecosphere.