No support for honest signalling of male quality in zebra finch song

Martin Bulla , Wolfgang Forstmeier

Alam et al.¹ claim to have discovered a song feature, “path length”, that honestly signals male fitness and is therefore preferred by all females. However, their data and analyses provide no statistical support for this claim. (1) The key finding — that long-path songs are difficult to learn (Fig. 4c) — is a statistical artefact: regressing y minus x on x creates an illusory effect where none exists. (2) Their path-length estimates have a measurement error of 45-73%, which undermines their conclusions, including the claim that females prefer long-path songs. (3) This claim is based on playback experiments that use only three artificial stimulus pairs, which, given the measurement error, cannot reliably contrast long and short paths. In sum, there is no evidence that path length functions as an honest fitness indicator. Our re-evaluation highlights the importance of validating new methods and accounting for random noise in small datasets. Finally, we emphasise that in species where females are known to disagree on who is attractive^2-6, searching for a trait that determines male attractiveness is unwarranted.