In a recent paper it was suggested that results published by Kammerer (1911) on the midwife toad could be explained by epigenetics (Vargas et al., 2016). We show that data thought to be fitting are based on untested assumptions about the underlying genetic mechanisms. We cite recent studies on the genetics of life history traits, in particular egg-size and number, to show that these assumptions are not realistic. We review aspects of Kammerer’s experimental results on the midwife toad for which there are no plausible mechanisms, i.e. toads switching from land-breeding to water-breeding in response to an increase in temperature, eggs becoming resistant to moulds within a few generations, the gradual development of nuptial pads, heterochronous changes in the development of water-born larvae, and conclude that Kammerer cannot have obtained the results he claims. We argue that natural selection would not have favoured a change in reproductive mode and the loss of parental care and that an epigenetic master switch, affecting many different traits simultaneously, would have either eroded during more than 15 million years of land-breeding and/or would have disappeared by natural selection against it. Finally, we show that Kammerer's data are remarkably close to the invoked Mendelian ratio and too good to be genuine. We conclude that Kammerer’s data are fictitious and that Vargas et al. (2016) have used non-existing data in search of support for a role of epigenetics in neo-Lamarckian evolution.