Yeap, in case of Tasmanian Devils you are right.

I made a a large swooping statement and had this in mind -> http://www.nature.com/neuro/journal/v17/n1/full/nn.3594.html

When I read it the first time if blew my mind. In principle - traumatic memories developed in presence of a particular scent triggered responses in subsequent generations - without the generations entering into social contact (a case of a female mouse to its offspring). Don't know how well established is statistics in this study though and how well replicated was the study. It suggests a storage of the information beyond central nervous system - in mammalian sperm.

Can you say that for certain in this case? In particular, any time you sequence a population, you see a large number of "novel" SNPs (or collections of SNPs), where the mutations were introduced in the germline of the parent and inherited, but weren't present in the parent's somatic cell line. And that's exactly what they saw in this paper. So I'm not certain you can claim that the DNA didn't adjust (rather, that natural mutations didn't contribute to the survivor's higher reproduction rate).