It depends on the variety of food sources. If there are a lot of categories, according to HO hypothesis, you should find a proportionality between available and eaten sources. If some categories are overrepresented (or under-represented) in the diet, there is a choice. I developped this approach in the case of herbivorous grasshoppers. Maybe this paper will be useful :
BENKENANA, N., HARRAT, A. & PETIT, D. 2013. Analysis of the number of sensilla on the labrum and the diet of grasshoppers belonging to the family Pamphagidae. European Journal of Entomology, 110(2): 355–364.
Dear Ana, I would suggest using the Kulback-Leibler divergence between two distributions (http://en.wikipedia.org/wiki/Kullback-Leibler_divergence). The nice thing about it is that it dovetails nicely with the Shannon index of biodiversity. Good luck! Bob U.