Bees do not use nearest-neighbour rules for
optimization of multi-location routes
Mathieu Lihoreau1,3, Lars Chittka1, Steven C. Le Comber2, Nigel E. Raine1,4*
1
Queen Mary, University of London, Biological and Experimental Psychology Group,
School of Biological & Chemical Sciences, London E1 4NS, UK.
2
Queen Mary, University of London, School of Biological & Chemical Sciences, London
E1 4NS, UK.
3
Present address: The University of Sydney, School of Biological Sciences, NSW 2006,
Australia.
4
Present address: Royal Holloway University of London, School of Biological Sciences,
Egham, Surrey TW20 OEX, UK.
*Author for correspondence (Nigel.Raine@rhul.ac.uk).
Electronic Supplementary Material
Figure S1. (a) Locations of the artificial flowers (filled circles: 1-6), the nest-box (N) and the landmarks (black bars: a-d) in the flight cage. Black
(unbroken) lines represent the two optimal routes (length: 2462 cm – either turning clockwise or anticlockwise), dotted arrows the unique
nearest-neighbour route (length: 3380 cm). Numbers in parentheses indicate the Cartesian coordinates (x;y, scale in cm) for objects in the flight
cage. (b) Electromagnetic flowers (1-6) were refilled at a distance via a remote control box (a bee is visiting flower 4 in this photo). Artificial
flowers were identical to those used by Lihoreau et al. [1,2]. The landing platform of each flower was a blue plastic disk (Ø = 60 mm; height =
10 mm) with a yellow circular area in the centre (Ø = 20 mm) and a narrow hole (Ø = 1 mm) through which a bee could insert its proboscis to
feed. Underneath the plastic disk sits a reservoir (Ø = 30 mm; height = 70 mm; volume = 50 mL) containing sucrose solution. Floating in the
reservoir is a plastic cylinder (Ø = 15 mm; height = 40 mm), containing a strong neodymium (NIB) magnet, on top of which is a small plastic
feeding cup calibrated to contain a fixed amount of sucrose solution (range: 5-30 µL, depending of the crop capacity of each bee). When the
electromagnet beneath the reservoir is activated, the float is drawn under the surface of the sucrose solution, which acts to fill the feeding cup.
When the electromagnet is switched off, the float bobs up and the cup containing sucrose is guided via a plastic tube just below the hole through
which the bee can feed. The electromagnetic mechanism in all flowers was controlled by the experimenter from a remote control box, reducing
any potential interference with the bee’s foraging behaviour during tests. These flowers can be accessed by a bee equally well when approached
from all angles – so the direction from which they arrive or depart each flower is not constrained. Each flower was placed on a 50 cm height
wooden support in a different location in the flight cage (see panel a).
1
Lihoreau, M., Chittka, L. & Raine, N.E. 2010 Travel optimization by foraging bumblebees through re-adjustments of traplines after
discovery of new feeding locations. Am. Nat. 176, 744-757. (doi:10.1086/657042)
2
Lihoreau, M., Chittka, L. & Raine, N. E. 2011 Trade-off between travel distance and prioritization of high-reward sites in traplining
bumblebees. Func. Ecol. in press. (doi:10.1111/j.1365-2435.2011.01881.x)
Figure S2. Appearance of the geometric patterned posters (a-d) used as landmarks. Posters
were fixed to the walls in each corner of the flight cage (figure 1a).
Figure S3. Average percentage of visits to each flower per foraging bout (mean ± s.e., n = 8
bees). Letter code above bars indicates significant differences between flowers (GLMM,
F5,571 = 12.64, p < 0.01, pairwise t-tests: p < 0.01). Flowers close to the walls of the flight
cage (< 1 m, grey bars) were not visited more often than flowers further away from the walls
(> 1 m, white bars), suggesting that bees did not use a wall-following strategy.
Table S1. Flower visitation sequences (including only the first visit to each flower) for each
of the 8 bees tested. Sequences in each column are sorted in chronological order (bout 1 to
bout 80). Numbers in the table (1-6) refer to the spatial location of each flower in the array
(see figure 1a). Colour codes (39 combinations of font and cell colours) indicate the
sequences observed significantly more often than expected by chance (multinomial test, p <
0.05). * indicates optimal sequences (clockwise 123456 and anticlockwise 654321), † nearest
neighbour sequence (124563), > clockwise sequences, < anticlockwise sequences. Incomplete
sequences where at least one of the six flowers was not visited are indicated by a red font and
white cell background (these sequences were not included in analyses).