In a series of feeding experiments we found that, depending on the larval food plant species or part of food plant ingested, individuals of the blue butterfly Polyommatus icarus (Lycaenidae) exhibit broad variation of wing patterns in the ultraviolet (UV) range of wavelengths which is invisible to humans. Such intraspecific variability in UV wing patterns has been underestimated thus far due to the rather demanding approach needed to study these patterns. We discuss methodological problems with the assessment of butterfly UV wing patterns by UV photography. Given proper standardization, UV photography is a suitable method to qualitatively assess UV wing patterns for possible use in morphology or systematics. Spectrophotometly should preferably be used as quantitative method when consideling UV wing patterns in a communication context. No higher value should be attached to UV wing patterns as compared to human visible wing patterns.