ST19-ABodey-PEN (3)

How bees navigate

Dr Andrew Bodey from Diamond Light Source in the UK discusses research at the facility which has investigated the ocelli of orchid bees.

Orchid bees, insects that live in tropical and subtropical regions, often within the rainforests of Central and South America, are known for travelling great distances in search of specific orchids to collect scents from. Using the bright X-ray beam of the Diamond-Manchester Imaging Beamline I13-2 at the UK’s synchrotron, Diamond Light Source, scientists have now determined how the orchid bee’s simple eyes capture the information it needs to fly through its complex habitat. They have discovered that the bee’s three simple eyes – also known as ocelli – work in ways that had not been previously imagined.

The three ocelli each have a retinal region which receives focused light from a different visual field. Other parts of each retina share a primarily unfocused view of a common region, creating a never-before-seen ‘trinocular’ visual field.

In an interview with Pan European Networks, Dr Andrew Bodey, scientist at Diamond Light Source and co-author of the new paper, discusses the research as well as possible future areas of interest, from bee behavioural studies to possible applications in unmanned aerial vehicle (UAV) technologies.

Could you begin by outlining the main discoveries of the project – the identification of the bee’s trinocular visual field, etc.?

In addition to their large compound eyes, orchid bees have three smaller, single-lens eyes which are situated on top of their heads. These simple eyes, or ocelli, were the focus of this study. We used a number of methods, primarily X-ray microtomography, to study the eyes and discovered that each possesses two distinct fields of view. One is a monocular region which views the horizon and is therefore suitable for flight stabilisation and steering. Using ray tracing simulations (computational simulations of light rays entering and passing through each eye) we demonstrated that these regions receive focused light – they see a sharp image. The other field of view is an upwards-facing trinocular region (the first that has been identified in an insect). Being trinocular, its view is shared by all ocelli – and in contrast to the monocular regions, its image is largely defocused.

Photoreceptors in the ocelli are polarisiation-sensitive: response to light polarised in a given direction depends on the angle of receptors. The photoreceptors in each ocellus are aligned with one another, and the photoreceptors between ocelli are offset by approximately 40°. Polarisation sensitivity has been observed in insects before, but this is the first observation of a system which meets all the requirements for full polarisation analysis; orchid bees have the equipment necessary to accurately measure both the angle and extent (fully polarised, partially polarised, etc.) of polarisation. This raises the question of why orchid bees need to make such accurate measurements, and if this type of system is more widespread amongst insects.

Andrew Bodey

Andrew Bodey

Why did you decide to use X-ray microtomography?

We needed three-dimensional information on the micron scale, we needed to see both internal and external features of the bee, and we required a field of view that was appropriate to the sample. In imaging there is typically a trade-off to be made between the field of view (which determines the size of the sample

that can be studied) and the size of features which can be resolved. We used X-ray microtomography with lower magnification to image the whole head and then increased the magnification so that we could image the ocelli individually at higher resolution. We then aligned the high-resolution ocelli reconstructions into the reconstruction of the head.

What does Diamond Light Source offer that is perhaps unavailable elsewhere?

The ray tracing simulations we performed on the microtomograms were rather ambitious, and so we needed very good quality data. Synchrotrons produce phenomenally bright beams of X-rays, and this enables datasets to be collected relatively quickly, with high signal-to-noise ratios and minimal artefacts.

The majority of the project team1 were based in Sweden, which doesn’t yet have its own third-generation synchrotron – so Diamond Light Source was an obvious choice. Diamond has a very strong capacity for imaging and – with high quality detectors, optics and motorised stages, access to high-performance computing, bright X-rays of appropriate energy and the capacity for phase contrast imaging – I13-2 is well-suited for soft-tissue microtomography.

Over the last couple of years we have developed much expertise in soft-tissue imaging. Biological samples can pose particular challenges relating to beam damage, sample deformation and poor contrast resulting from weak X-ray attenuation. Addressing these challenges can involve the balancing of competing needs, and good imaging relies upon optimisation of various imaging parameters and monitoring of data quality during experiments. The beamline utilises variable ‘in-line phase contrast’ to enhance the visibility of features, based upon differences in their refractive indices. The extent of such contrast is variable, and a balance must be struck between its benefits and the difficulties it can pose for image analysis if an excessive amount is present.

The high quality tomography was made possible by careful sample preparation. The insects were dissected and then preserved via fixation (a chemical process for cross-linking proteins). Samples were given a mild stain to enhance amplitude contrast in imaging, and then dehydrated and embedded in epoxy resin. The process helped to improve mechanical stability, which reduced the extent of artefacts arising from sample deformation during imaging.

We were careful to protect the samples from beam damage. X-ray tomography is a non-destructive method: one can look inside objects while leaving them relatively intact. However, even non-destructive methods can deteriorate soft-tissue samples and cause them to deform during imaging. We therefore modified the energy spectrum of the beam, reduced beam exposure and utilised a low-dose alignment strategy to minimise such problems.

Where will future research efforts lie?

We have observed a visual system which fulfils all the requirements of full polarisation analysis. Behavioural experiments are now required to confirm that the bees use the system for this purpose and to determine why they need to measure polarisation so accurately.

Light in the sky is polarised with respect to the position of the Sun, and bees use this information to navigate. For example, honeybees dance to communicate the angle and distance in which to travel in order to find good food sources; they then use polarisation patterns in the sky to navigate to the food and to return to the nest. Orchid bees operate in a particularly challenging environment: the rainforest is dimly lit and cluttered, and the sky is largely obscured by canopy. It may be that orchid bees have evolved a particularly sophisticated vision system; alternatively, polarisation analysers may turn out to be quite common amongst other bee species. We plan to study two important groups of pollinators – bumblebees and stingless bees – to investigate this.

The focused monocular regions are suitable for stabilisation and steering, but they may also play a role in visualising polarised light reflected from orchid flowers. Orchid bees are picky pollinators and some will exclusively visit a single orchid species. Some plants reflect particular polarisation patterns from their flowers, and orchid bees may use the sharp images produced by their monocular regions to identify orchids; behavioural experiments could be devised to investigate this.

We are still analysing the tomographic data collected at Diamond, and are now performing ray tracing simulations on the compound eyes. With thousands of lenses per eye this poses a particular technical challenge, and joint principal author Dr Gavin Taylor is writing software to deal with this. This should lead to a more complete understanding of the visual information that orchid bees receive from their eyes. Orchid bees are quick and agile flyers, and navigate and search for orchids while avoiding predators and obstacles. Somehow they receive all the visual information they need, and perform image analysis rapidly enough with their tiny brains, to perform these remarkable feats.

The microtomograms collected at Diamond revealed a surprising feature in the retinae. We can’t give away too much at this stage, but the team will be visiting the synchrotron again soon to investigate further.

How do you feel your results can impact applications in areas such as UAV technology?

Our work has focused on fundamental biological questions, but the results do suggest that polarisation analysis could be useful for UAVs. In fact, other researchers have been experimenting with polarisation sensing and analysis in these vehicles for some years now. There is interest in the development of UAVs which can operate in challenging environments such as forests, disaster areas and at high altitudes. Polarisation analysis could serve as part of the sensing mechanisms of vehicles operating in such environments, assisting with navigation, flight stabilisation, steering and object identification.

  1. The project was a collaborative effort involving researchers from Swedish, Australian, South African and UK institutions. Further details can be found in the journal article: Gavin Taylor and Willi Ribi, Martin Bech, Andrew J Bodey, Christoph Rau, Axel Steuwer, Eric J Warrant, Emily Baird: The Dual Function of Orchid Bee Ocelli as Revealed by X-ray Microtomography. Current Biology. 2016. http://dx.doi.org/10.1016/j.cub.2016.03.038.

Image reprinted from Gavin Taylor and Willi Ribi: ‘The Dual Function of Orchid Bee Ocelli as Revealed by X-ray Microtomography’. Current Biology. Copyright 2016, with permission of Elsevier.

 

Dr Andrew Bodey

Diamond Light Source Ltd

http://www.diamond.ac.uk/Home.html

 

This article first appeared in Issue 19 of Pan European Networks: Science and Technology