The neurobiological basis of time estimation are still an unsolved puzzle. A sexy hypothesis proposes that it encloses cognitive processes such as attention, learning, working memory and decision-making. Time estimation would be a specific type of working memory and behaviour would be the result of decisions based on time perception. To understand the ability to estimate time we make use of the interval timing skill of the fly Drosophila melanogaster. We designed an experimental setup that enables the study of a time-referenced memory in the fly based on the proboscis extension response to a sucrose solution. Briefly, we offer a sucrose solution drop to a thirsty/hungry fly at a fixed interval and record its behavioural response. We operationally define time-referenced memory when the fly extends its proboscis anticipating the appearance of the drop. Preliminary results with w1118 strain suggest that flies develop an anticipatory proboscis extension reflex (A-PER) over time. The response gets closer to the interval of the experiment as trials increase. Only 40% of the flies have shown A-PER at the sixth trial while 80% did at the ninth. For an interval of 60 seconds this response gets as close as 55,5s ± 4,6s at the evaluation trial. Here we demonstrate that wild type Canton S stock are able to learn intervals. We conclude that our experimental setup is promising to unveil the mechanisms involved in temporal reference memory and propose a neurogenetic strategy to approach it.