I aimed to develop a paradigm that could be used to train bumblebees, Bombus terrestris, to learn that a specific odour-colour combination is associated with a food reward when there is a time delay between the presentation of the odour and colour (cross-modal conditioning). This paradigm would allow the experimenter to measure what odour identity the bees perceive by recording what colour they select after being presented with an odour. Which colour the bee chooses indicates what odour the bee has perceived after training. Bumblebees naturally use vision and odour cues together to optimise their speed and accuracy when foraging. The first chapter is the introduction. The second chapter is on using differential conditioning to train bees to select the rewarded colour over the un-rewarded colour. This was to optimise the paradigm for training bees before moving on to cross-modal conditioning. In Chapter 2 I presented five different experiments, building on each other. I found that the Y maze was better than the previous apparatus used in the lab for learning and that small intertrial intervals caused the bees to learn to avoid the rewarded conditioned stimulus. In the third chapter, I describe an experiment that aimed to train the bumblebees to associate an odour A with a colour A and an odour B with a colour B. This was done using cross-modal conditioning and there were signs of weak learning. The fourth chapter is an overall discussion. Overall, the bees did not successfully learn to differentiate between the two colours but showed signs of learning to weakly associate an odour with a colour.