Sample PagesSelection can only select from variations that are available to be selected from. Objections have long been raised that random mutations--the only source of variation that the theory allows--are incapable of providing a credible raw-material resource. The chances of hitting on anything potentially beneficial are simply too slim. (This was how my own original doubts began.) The stock answer has always been that given enough time, even the improbable becomes likely. Despite the experiences of domestic breeders, there was nothing in principle to prevent the observed adaptations and divergences of species being extrapolated without limit, even if hard evidence that it had happened was lacking. For as long as such arguments revolved around qualitative, higher-level issues like morphology, limited knowledge of the finer workings could be invoked to preserve assumed Darwinian principles. But the detailed expositions of the underlying molecular machinery that have come about in the course of the last twenty years make such a recourse untenable. Every morphological change along the way in the postulated evolutionary changes from a fin to a leg, an air sac to a lung, a light-sensitive spot to an eye, or in the development of a circulatory system or an energy metabolism, requires changes in the relevant biological macromolecules: the proteins required for structure and function, and the controlling genes that reside in the organism's DNA. And at that level, which admits no appeals to further hidden explanatory mechanisms, the sheer, colossal improbability of even one of the many thousands of such molecules found in nature forming by any chance-based process constitutes a very real problem.
All biological proteins, from those forming bacteria to the highest animals and plants, consist of chains built from the same, twenty-strong set of the chemical groups known as amino acids. The structures and roles of proteins are amazingly diverse. In size they range from short polypeptides (not usually classed as protein) just a few amino acids long, such as the endorphins which carry chemical signals in the brain, through hormones like insulin with several tens, oxygen carriers like hemoglobin and cytochrome c, and enzymes with several hundred, to the structural proteins making up every body tissue, culminating in titin, a component of muscle, weighing in at 25,000 to 30,000 amino acids. As well as having a linear structure as represented by its amino acid sequence, a protein also folds itself up to exhibit a very precise and specific 3-D shape that enables it to perform its function, usually in cooperation with other proteins. Building things from proteins is amazingly efficient. It would be like being able to mass-produce parts for, say, a refrigerator, a bicycle, or anything else by means of a universal chain-making machine capable of churning out stiff lengths of links punctuated with springs and hooks in just the right places to make the chain buckle and lock into the exact form required. One of the properties that makes enzymes such superb catalysts, able to speed chemical reactions up by factors of millions or even billions, under far milder conditions than are necessary in laboratories, is that they are tailored like precision machine jigs to bring and hold the reacting molecules together at just the right distance and in the right orientation for their active sites to coincide.
As an example of how unlikely it is that such macromolecules will arise readily, let's look once again at cytochrome c. Cytochrome c is a small protein, usually comprising 104 amino acids, which is found in virtually all cells as a component of the energy metabolism. Its universality leads proponents of the Darwinian view to conclude that it arose early in the history of life, before the various groups of organisms diverged. With 20 amino acids available to choose from at each position in the chain, the number of possible combinations that could be generated is 20104 or 2x10135 -- in other words, 2 followed by 135 zeros.
Nothing in common experience conveys the size of such a number. The number of atoms in the entire observable universe is estimated to be in the order of 1080, which falls short by 55 zeros. The number of ways to construct even a small protein 104 units long equals -- give or take a few -- the number of atoms in 1055 universes! Once the hard numbers are in, there can be no resorting to vague assurances that long spans of time make anything possible simply by being beyond the ability of the human mind to imagine. Even if all the material resources of the universe were applied to generating trial combinations at the fastest rate that physical processes can proceed, the fraction that could have been tried in its entire lifetime is utterly insignificant -- in the order of 1 divided by 1040.
For proteins numbering thousands of amino acids the problem becomes inconceivably greater still. The human body contains somewhere around 20,000 different types of protein. About 2,000 of them are enzymes. Fred Hoyle calculates the improbability of these enzymes alone as a number having 40,000 zeros--40 pages to print in an average book. Every one of these sequences has to be specified (a significant word that we'll return to later) in the DNA code that directs their assembly. Each of the 46 chromosomes that make up a human DNA chain contains many millions of nucleotide base pairs (the units corresponding to amino acids in proteins). The full human genome is estimated to be three billion base pairs long.
It is true that these calculations refer to the odds of producing a specific protein, i.e. of hitting on one of the exact sequences found in nature. The odds against being dealt any particular hand at bridge are also enormous, but that doesn't mean that the combination of cards one is holding amounts to a miracle, since one of them was bound to happen. What would be miraculous is the ability to specify them consistently in advance. In the same kind of way, it is sometimes contended that if other variations of a protein are capable of doing the job, the precise form that happens to have evolved in nature isn't essential, and so the constraints can be relaxed. Examples cited to show that this is in fact the case include the phenomenon of polymorphism, or variants existing within the same species such as the light and dark forms of Peppered Moth, which result from slightly different versions of the same protein, and equivalent but not identical proteins performing essentially the same role across a wide range of species--cytochrome c is a good instance. And coupled with that is the redundancy of the genetic code, by which different DNA sequences can produce the same protein, which introduces more latitude for variation at that level--a bit like holding multiple tickets in a raffle: Whichever one comes up, you still get the prize. In summary, the claim is that unique amino acid sequences are not necessary for protein function; many of the sequences that work can be arrived at by multiple coding paths, and the general effect is to mitigate the prohibitive improbabilities involved.
However, on closer examination this kind of optimism turns out to be not very well founded. In proteins that do exhibit some wiggle room for variation, the variations all occur in relatively unimportant positions, such as on the outside of the folded protein structure, where the function is more to contain the functional inner parts, and what actually does the containing doesn't make a lot of difference. But where it matters, in places where the crucial folding operations take place, or the active regions inside must come together in precisely the right way for the right things to happen, the sequences are highly specific and strongly conserved. So while you can choose things like the color and seat upholstery when ordering your new car, the way the engine is put together is not something you have options on.
As an example, aptly named ubiquitin plays a key role in regulating protein degradation and is found almost everywhere in eukaryotic organisms (ones consisting of cells with nuclei, which means everything above bacteria). It has just 76 amino acids, 69 of which are totally invariant. Only three differences exist between the sequences in yeast and in humans. Actin, a structural protein found in all eukaryotic cells, has a sequence of 375 amino acids, 80% of which is identical in all animals from amoebas to humans. The core portion of the plant protein rubisco has 476 amino acids, 105 of which are absolutely constant, and in a further 110 positions only one substitution is possible. Hence, the essential, highly specified sequences can, and as a rule do, greatly exceed the figure of a hundred or so upon which the previous calculations were based.
The other common response to the problem of improbabilities is that the mountain doesn't have to be scaled in one leap. Macromolecules arose from simpler ones in the same way as the organisms that express them did. Just as species evolved through natural selection of advantageous combinations of genes, so crude precursors of proteins (and the genes responsible for them) evolved through selection of advantageous combinations of amino acids (and nucleotide base sequences) into the forms we see today. As methods for determining amino acid sequences were developed through the 1950s and 1960s, clear similarities were found across species, and the closer that species were morphologically, the closer their protein sequences matched. This was received as strong supporting evidence for an evolutionary process, and work followed in earnest to construct molecular phylogenetic trees showing how the sequences observed today could have branched from common ancestral ones. And, indeed, the lines of descent inferred in this way bore a good resemblance to the phylogenies already deduced from morphology and the fossil record. Molecular chemistry, it was therefore confidently expected, would provide the conclusive proof that neo-Darwinism had been seeking.
But once again, the optimism of the early days was clouded by the later findings. The variations measured and graphed to construct the trees occurred overwhelmingly in peripheral locations of minor importance, where their effect on the protein's activity was practically neutral--hence providing nothing of significance for selection to work on. The core regions carrying the functionally critical sequences were highly conserved, if not invariant, across wide ranges of species. What this showed was the fine-tuning about a basic theme that was consistent with adaptation inside limits, where the effectiveness of selection had never been disputed. But the highly conserved core sequences meant that getting from one basic theme to another was as big a jump as ever, and the absence of intermediates provided no evidence of it's having happened. No amount of juggling with colors, upholstery, and other accessories will turn a Pinto into a Chevy van. And it turns out repeatedly that some minimum combination of amino acids is required for the molecule to do anything useful at all, far above any number that could plausibly arise by chance. This would seem to preclude progressive development from simple precursors.
What the groupings display is the hierarchical structure of a typological order that we met before--islands of related variants scattered across an ocean of non-viability that produces no bridges from one to another. But neo-Darwinism has become so entrenched that the early molecular matching results are insisted to be the result of evolution nevertheless and interpreted from an evolutionary perspective accordingly.
Once even a moderate number of sequences becomes available, the number of possible trees by which they might be related rises too rapidly for even a small fraction of them to be assessed. Adding to the difficulty is that different proteins yield different phylogenetic implications. The cytochrome c of birds was found early on to be more like that of reptiles than of mammals, which was advertised as a clear indicator of reptilian ancestry. But then avian hemoglobins turned out to be more like that of mammals, so this was explained by birds and mammals being both warm-blooded. The myoglobin of seals is similar to that of whales and dolphins, which was considered unsurprising because of their similar habitat and life style, even thought they were said to have diverged 50 million years ago; but whale cytochrome c is identical to that of the camel.
What happens, then, is that given the need for some starting point, the conventionally accepted morphological tree becomes the guide for constructing the molecular trees. This is reasonable as far as it goes, but it negates the suggestion of the results being independent evidence of an evolutionary picture. As ad hoc explanations contrived to account for the discrepancies accumulate, the case becomes less convincing.