I seldom read a book that I think has little if anything to recommend it, but I just read one. It is Jeffrey H. Schwartz' Sudden Origins (Wiley, 1999). Subtitled "Fossils, Genes, and the Emergence of Species," it should more accurately have been called "an anti-Darwinist paleoanthropologist reviews the history of evolutionary thought." Billed as a book which explains the discovery of homeobox genes' role in evolution and presents a new theory of the origin of species, that description only accurately fits the last 30 pages of an almost 400-page book.

For the "fossils" portion of the subtitle, 99% of the discussion of fossils is of those of hominids. There is only a passing reference to any other group. The "genes" reference is primarily a review of various researchers' work since publication of the Origin of Species, a rehashing of Mayr's classic The Growth of Biological Thought. I had to wait until p. 367 before being exposed to anything new.

Before I go into my complaints about the theories expounded, let me say that throughout I was constantly annoyed by grammatical errors and inaccuracies. Every few pages or so I came across a sentence in which the verb did not agree with the subject. In a discussion of hominid tools on p. 73, he states that these were generally made of "obsidian, flint, or chert, which are volcanic rocks" -- the last two of which are actually micro- or noncrystalline varieties of quartz of sedimentary origin. On p. 128 it is said that an organism reaches sexual maturity concurrent with a cessation of physical growth. While true in many species, the author is obviously unaware of the biology of fish and reptiles (as well as trees), which grow continually throughout their lifetimes, long past reaching the age of reproduction. When I am faced with a tome which presents obvious factual and linguistic errors, indicating sloppy editing, I can't help but wonder how much of the rest of the material presented is inaccurate as well.

With vehemence the author attacks Darwin's theory of the origin of species by means of gradual selection as an outgrowth of the flawed principle ("a Victorian mindset") of geological uniformitarianism. True, Darwin believed in "blending inheritance" in which the offspring's phenotype is intermediate between those of the parents. His lack of understanding of the nature of particulate inheritance was a direct consequence of the fact that the Origin was first published 7 years prior to Mendel's work on the genetics of peas, and fully 41 years before the "rediscovery" of Mendel's theories by the general scientific community (due to the fact that they had been published in a small, obscure journal). The Darwin-bashing which pervades the book is predicated on this, and it seems to me hardly fair to fault Darwin for being unaware of scientific discoveries which had yet to be made. An example from physics is relevant here. When Sir Isaac Newton formulated his laws of mechanics and motion, Einstein would not be born for another 200 years. Do we therefore throw out all of Newton's formulae because relativity has since been discovered? Of course not. Newton's laws are fully applicable to everyday situations when the systems in question nowhere near approach the speed of light. Calculations based on relativity yield the same results, to a level of accuracy which is practical in everyday use. Newton's laws are still an integral part of any physical science or engineering curriculum, being as they are merely a special case of the phenomena articulated by Einstein. Newton merely was ignorant of the "complete picture", not inherently in error.

The author, however, throws out Darwin's entire theory of evolution by means of natural selection based on similar logic. OK, genes do not "blend", but are transmitted intact from generation to generation. Thus the phenotypic traits they control are minutely discontinuous as well. On a nit-picking scale, you could say that Darwin was wrong. But this is akin to haggling over the accuracy of calculus in describing the area under a curve. The basic premise of such a calculation is that the area under a curve can be thought of as the sum of a series of rectangular areas (easily calculated) which fill and approximate the entire area under the curve as described in x-y space. If the x-dimension of those rectangles is large (figure 1), the result will be at best a crude approximation. As the x-dimension (called dx in calculus) of the rectangles becomes smaller (figure 2), the approximation of the area will become better, only yielding the precise value when the value of dx becomes exactly zero. But for all practical purposes (and this is the method used by computers to do calculations on complex equations), the area yielded when dx is very small will be a very good and quite adequate approximation of its exact value. By pertinent analogy, although genes and their phenotypic expressions are discrete, there are so many of them in the genome of a complex organism that they effectively create continuous phenotypic variation within a population, as Darwin said.

Thus, from my viewpoint, taking issue with the uniform nature of phenotypic variation is much the same as saying that a mountain does not erode in a continuous fashion just because sandstones yield up one grain at a time, and other rocks dissolve chemically one atom or molecule at a time.

But Schwartz' assertion that variation in a population is not technically, at the most minute level, continuous leads him to reject Darwin's entire theory of natural selection, a gross error in logical extrapolation as it seems to me.

The "Evolutionary Synthesis" of the 1940s melded an understanding of the nature of genetics with Darwinian natural selection (engendering the theory which is called "neo-Darwinism"). The prevailing viewpoint became that novelty was introduced by random genetic mutation (not from blending of existing characteristics as Darwin had thought), and that these novelties were subsequently subject to the pressures of natural selection with regard to whether they would spread through and persist in the gene pool or be eliminated. Such mutations were thought at the time to probably produce minute variations in the phenotype of organisms, a premise which is again akin to a comparison of Newton's laws to Einstein's -- yes, but not under all conditions.

Genes, and their resulting proteins and hormones, orchestrate the production of every component of the body. However, the products of some genes have more profound effects than others on the phenotype of the organism -- hardly surprising if you give it much thought -- and this is what is "new" about the theory presented in this book (although it was actually presaged by Gould and Eldredge back in 1979, albeit without explicit proof).

Some genes code for the proteins that make up the substance of cells. Others produce hormones and enzymes that control cellular processes and the activation and deactivation of sections of DNA (remember that each cell of your body contains the full set of your chromosomes and therefore DNA, but different portions of that DNA are activated to cause one cell to become muscle and another to become bone). The class of genes dubbed "regulatory genes" (of which homeobox genes are a subset) controls embryological and developmental processes. They determine the unfolding of the genetic program. As such, a mutation in one of these genes can alter the development of an organism and produce a larger phenotypic effect than can one which is in a gene which codes for a structural protein and controls, say, hair color. I take no issue with this. But does that mean that mutation and selection of regulatory genes is an entirely different process from mutation of structural genes and their products -- and therefore that the processes of speciation and adaptation are completely decoupled? This stretches the imagination.

This, however, is exactly what Schwartz expounds. He gives examples of where a mutation in a homeobox gene can determine whether a human has five fingers, stunted fingers, or six or more. This is related to the number of repeating alanine molecules (in the product of one regulatory gene) which the organism possesses. But it is not a matter, demonstrated experimentally on mice and birds, of a singular mutation in one DNA base determining the number of fingers formed or whether the number will be reduced into the wing of a bird. The repeating alanine segments are long, requiring a correspondingly long DNA segment to be repeated. Merely because there are discontinuities in the number of alanine repeat codons present in the pertinent genes of existing mice and birds, must we assume that intermediate mutations cannot and must not have ever existed? I do not mean to imply that a mutation did not perhaps consist of a multi-alanine segment having been duplicated in a single step. But must all mutations in regulatory genes occur in such a fashion?

I doubt it, and here I give an example from my own family. It runs in our family to lack lateral incisors (the two flat teeth flanking the central incisors and between them and the canines). My grandmother didn't have them, and neither does my dad or do I. But intermediates clearly demonstrate that this is not the effect of an either/or mutation. My mom has lateral incisors, and although both my older sister and I lack them, my younger sister has them and they are dwarfed. My husband has them, and our union resulted in one daughter who is like him and another daughter who possesses one lateral incisor but not the one on the opposite side. The fact that I am phenotypically different from my husband in this respect did not cause him to consider me a different species and thus unsuitable for a mate. It is merely an example of my having a "progressive" adaptation (fewer teeth which better fit the human trend towards smaller mouths, allowing my wisdom teeth to erupt without any problem) than in his case where his teeth are crowded and crooked. Clearly the phenomenon is one of incomplete dominance of one gene over the other, and here I take issue with Schwartz again.

Anyone who has had an introductory course in genetics knows that the terms "dominant" and "recessive" when applied to genes provide an incomplete picture. Mendel was lucky in that the peas he worked with expressed dominant and recessive traits without complication. But most traits are not like that. A gene can be "incompletely dominant", that is, dominant over some alternate alleles but not others; additionally, few traits are controlled by a single gene -- multiple genes contributing to a feature are the rule. Schwartz adheres to a terminology of dominant/recessive as if they are the only two states to be found, and nowhere discusses the phenomenon of multiple genes contributing to a given trait. This is the high-school simplification whereby we learned that there are genes responsible for blue eyes and brown eyes, and that brown is dominant and blue recessive (such that a heterozygous individual will have brown eyes, and only someone homozygous for the blue allele will express blue). So how come my eyes are hazel? There is clearly more to the picture than our high-school example.

When discussing the conundrum of how a mutation in a regulatory gene which causes a macroscopically discrete phenotypic difference can become established in the gene pool of a large population, Schwartz continues his either/or dominant/recessive distinction in providing an example (fish going from toothless to an entire set of teeth). He states that most mutations arise in the recessive state. Admitting that most mutations are likely to be deleterious to the organism, he overlooks the likelihood that recessive mutations are probably found to be more numerous in living organisms simply because dominant (expressed) mutations are much more likely to be quickly weeded out of the gene pool via immediate death of the organism. OK, then his hypothetical fish with the recessive mutation for toothiness is not subject to natural selection (he rejects that natural selection works in evolution, while invoking its influence in his example), and therefore the mutation is free to spread slowly throughout the gene pool without being expressed. The population subsequently, some generations later, reaches the point at which sufficient heterozygotes for this trait exist and breed to result in enough homozygote progeny with the expressed trait that a viable breeding population of fully-toothed fish arises within the course of a single generation, and a new species is born.

Again, this stretches the imagination. Gould and Eldredge allow in their theory of punctuated equilibrium that the "punctuations" may reflect evolution which is instantaneous on a geologic timescale, but fully reconcilable with gradualism on an ecological one. Carrying this to the extreme of a new species arising within the space of a single generation, required if the individuals expressing the new trait are to find identical mates, seems excessive. I do not profess that it's impossible, and may have occurred occasionally, but to use it as an explanation for the origin of species in general seems a bit much.