Our paper “Repeatable randomness, invariant properties, and the design of biological signatures of identity” has been published in the journal Evolution. This study, led by Tanmay Dixit and carried out with Kuan-Chi Chen, Cassie Stoddard, L Mahadevan, Christopher Town, and Claire Spottiswoode, was a collaboration between biologists, mathematicians, and computer scientists.
We were inspired by digital signatures in cryptography to study two hallmarks of an optimal signature – consistency (i.e. one’s signature needs to be the same each time it’s deployed) and distinctiveness (i.e. one’s signature needs to be different from everyone else’s). Using a biological system – the highly variable egg signatures of tawny-flanked prinias, which they use to identify their own eggs from those of imposter cuckoo finches – we showed that there is a trade-off between consistency and distinctiveness.
How could hosts get around this trade-off? One way would be to possess a signature trait that is both consistent within their eggs but distinctive from others. Using approaches from computer vision and field data, we showed that the sizes of individual egg markings would be consistent but distinctive, but that prinias do not discriminate between eggs based on these sizes. This suggests that prinia behaviour is not optimally tuned to make the best use of their egg signatures.
Open access link: https://doi.org/10.1093/evolut/qpad134