You seem to think that random mutations are passed on to offspring. Most mutations, as far as I am aware, are somatic mutations. Meaning they are mutations which occur in an organism's non reproductive cells. In order for mutations to be passed on to offspring, they would have to be germline alterations. This is a major sticking point with genetic engineering today, somatic v/s germline alteration. One only affects the organism in question, while the other is passed on to future generations.

If my hypothesis has some truth to it, then it raises some interesting possibilities. One of which would be that evolution is male driven and not female driven. Female eggs are created while in the womb and remain with her throughout her life. They are static and unchanging. Sperm are produced by the hundreds of millions every day. They would be subject to any alterations induced by environmental changes.

I really wish this idea could be tested in a lab, but many generations would have to be raised in order for it to be properly evaluated. Maybe by breeding a few thousand generations of fruitflies in total darkness, we could see if it had any effect on eye development.

BTW: You say that the Ceolocanth of today has different DNA from that of the Carboniferous era, but considering that we have no DNA from specimens from the past, it's hard to come to that conclusion.