Many organisms have been observed to acquire various new functions which they did not have previously (Endler 1986). Bacteria have acquired resistance to viruses (Luria and Delbruck 1943) and to antibiotics (Lederberg and Lederberg 1952). Bacteria have also evolved the ability to synthesize new amino acids and DNA bases (Futuyma 1998, p. 274). Unicellular organisms have evolved the ability to use nylon and pentachlorophenol (which are both unnatural manmade chemicals) as their sole carbon sources (Okada et al. 1983; Orser and Lange 1994). The acquisition of this latter ability entailed the evolution of an entirely novel multienzyme metabolic pathway (Lee et al. 1998). Bacteria have evolved to grow at previously unviable temperatures (Bennett et al. 1992). In E. coli, we have seen the evolution (by artificial selection) of an entirely novel metabolic system including the ability to metabolize a new carbon source, the regulation of this ability by new regulatory genes, and the evolution of the ability to transport this new carbon source across the cell membrane (Hall 1982).
Such evolutionary acquisition of new function is also common in metazoans. We have observed insects become resistant to insecticides (Ffrench-Constant et al. 2000), animals and plants acquire disease resistance (Carpenter and O'Brien 1995; Richter and Ronald 2000), crustaceans evolve new defenses to predators (Hairston 1990), amphibians evolve tolerance to habitat acidification (Andren et al. 1989), and mammals acquire immunity to poisons (Bishop 1981). Recent beneficial mutations are also known in humans, such as the famous apolipoprotein AI Milano mutation that confers lowered risk to cardiovascular disease in its carriers.
Bacteria, plants, and other creatures are incredibly well designed to be able to adapt for survival in different environments, to the extent they can even copy and use external genes in some circumstances. Yet they still always reproduce the same kind. E. coli produce E. coli, plants produce the same kind of plant, animals produce the same kind of animal. There can be a lot of variation within a kind (e.g. all the different dog breeds), but a dog will never produce a non-dog, and a fish will never produce a non-fish. This requirement of microbe-to-man evolution contradicts the fundamental law of heredity.
> but a dog will never produce a non-dog, and a fish will never produce a non-fish. This requirement of microbe-to-man evolution contradicts the fundamental law of heredity.
Agreed! And doing so would of course invalidate our current understanding of how evolution works! We'd have to throw out our theory of evolution by natural selection if we ever observed a dog give birth to a non-dog!
The definition of a "species" isn't something defined by nature. Rather it's a construct that humans invented in our need to categorize things. Genetic dissimilarities one generation to the next are of course very very small. So small that adjacent members will always be of the same "species" (able to breed with each other). But cumulative changes over generations add up. So much so that if we skipped ahead many generations, we'd no longer have compatible breeding. Different "species" as we would say under our admittedly flawed categorization system. At what point in the family tree does the species change? Given that each species can breed with its adjacents? At what point does a color gradient stop being red and start being blue? As you can see, the flaw in this understanding stems only from holding onto the definition of distinct "species" as something real rather than a human creation. A useful one, don't get me wrong! But it's important to understand the fuzziness of this definition. Your argument stems from a fundamental misunderstanding of this idea.
By the way, the example above is not merely hypothetical, since species can be separated in space as well as time. We can observe this phenomenon in action through very fascinating phenomenon like "ring species".
