What you're hinting at here is Lamarckian Inheritance: the notion that an individual organism can acquire a trait during it's lifetime and then pass it on to it's offspring. It's generally considered a dead theory, although slightly related effects are at the core of Epigenetics.
The foundation of modern genetics is built around the notion that changes in DNA occur over time that are not immediately expressed (that is they are often in recessive genes or even more likely in what's colloquially called "junk DNA" which are sequences that appear to be ignored). This is the operation behind Stephen Jay Gould's Punctuated Equilibrium theory that posits that such genes are expressed at evolutionary stress points when environments change significantly.
To get back to your example then, if a "high resistance/Immunity to Neurotoxins" was expressed, by the time it was expressed, the genetic material that allowed it would actually already be widespread in the population, and due to those with the expressed version of the sequence being more likely to survive, subsequent generations would have higher percentages of that useful gene in the population. But that would happen quickly because the sequence would already be in a significant portion of the population, the "expression" simply being a small genetic change to "turn on" the existing code.
With the advent of gene therapy, it's possible that the expansion in populations that were not in the stressed group could be spread more rapidly, but that's not an evolutionary process.
It is not enough to discover and prove a useful truth previously unknown, but that it is necessary also to be able to propagate it and get it recognized,