Avian brood parasitism is a well-recognized model for studying coevolution. In this model, hosts adapt the ability to recognize and remove parasitic young to avoid the costs of rearing foreign offspring, while parasites counter-adapt phenotypes to evade detection. A classic example of host-parasite coevolutionary arms race is the great reed warbler and its parasite the common cuckoo (hereafter "warbler" and "cuckoo", respectively). Recent work has found that cuckoo eggs are more likely to be accepted by warblers when those are bluer than their own eggs. Therefore, hosts would select for bluer cuckoo eggs; in turn, this provides directional selection for hosts eggs to remain bluer than cuckoo eggs. Here, we tested whether there were consistent differences in eggshell appearance between cuckoo and warbler eggs using a dataset which provides eggshell coloration of both warblers and cuckoos from the same nest. Despite being a textbook example of mimicry, we found that the cuckoos have significantly browner eggs than the warblers. These findings most likely suggest that this warbler-cuckoo system is experiencing negative frequency-dependent selection, as expected under red queen dynamics. Future research comparing warbler-cuckoo interactions over time would advance our understanding of the coevolution between avian brood parasites and their hosts.