I dont know if anyone has adressed this, but during the last Ice Age, the polar ice caps only came down as far south as London, hence nowhere near the modern habitat of crocodiles who have not significantly evolved since the age of the Dinosaurs because there is no mutation that would be favourable.
In other words crocs havent evolved because theyre perfectly adapted to their habitat, which has not significantly changed since the dinosaurs.
Not only do we have scientific evidence of fish breathing air, we have common examples in home tropical aquariums, Anabantoids, or Labyrinth fish. These include Gouramis, Siamese Fighting Fish and Climbing Perch who can crawl out of one body of water, into another, they have even been known to climb trees.
The lung like Labyrinth organ which enables these fish to breathe air would have primarily enabled the fish to survive in waters with low oxygen content, as a consequence some, like the climbing perch developed a land loving habit.
The only random aspect in the theory of evolution is genetic mutation, or a Darwin called it, variation, but observations support that. Natural Selection is also an observable fact, its simply one species causing a decline in another due to more advantageous characteristics.
We also observe the effects of Artificial Selection, where characteristics of domesticated species are promoted through selective breeding.
Goldfish are a typical example Sarasa Comets, Fantails, Orandas and Lionheads can all be traced back to the common goldfish, which in turn is a descendend of a Crucian Carp or similar fish, yet they look very different to their ancestors.
Charles Darwin noted the same kind of thing, but in pigeons at first, and then other domesticated species. Then on Galapagos, he famously noticed the same kind of thing happening to wild finches.
It is argued that a different shaped beak, or modified body shape may be reasonable, but for one type of creature to become an ancestor of another over time is not. The trouble is, on what basis can we draw a line? How can we define a limit in the change of characteristics of species over generations?