by Bill Bryson
Darwin never crossed the digital bridge. If he had, he would have had a ready answer to Fleeming (pronounced Fleming) Jenkin, the Scottish engineer who – independently of his colleague Lord Kelvin (with whom he collaborated on the trans-Atlantic cable) – gave Darwin a hard time over matters of theory.11 Jenkin pointed out that, on the current non-digital, blending view of heredity, variation would be swamped by successive sexual crossings, and after a few generations would disappear. There’d be no hereditary variation for natural selection to work on. Blending inheritance would be like mixing black and white paint: you get grey, and no amount of subsequent mixing of grey with grey will give you back the original black and white.
As a matter of fact, any fool could have seen that Jenkin’s premise must be wrong. Variation does not dissolve away as the generations go by. We are not more uniform than our grandparents were, and our grandchildren will retain the same level of variation as we possess. Jenkin thought he was doubting Darwin. Actually he was doubting observable facts. Nevertheless, his criticism worried Darwin.
Enlightened by Mendel’s nineteenth-century peas and building on Hardy and Weinberg’s elementary algebra, the twentieth-century founders of population genetics, R.A. Fisher, J.B.S. Haldane and Sewall Wright, buried Fleeming Jenkin. If genes are countable, digital entities that don’t blend, their frequencies have no inherent tendency to change. If they do change, that is evolution, and it happens for a reason. The most interesting reason is non-random selection, but random drift also occurs – to an extent disputed among the founding fathers but now widely admitted among molecular geneticists. Even those three founding fathers never knew quite how digital genetics really is. In the light of the Watson/Crick revolution, we now see the very genes themselves as digitally coded messages, digital in exactly the same sense – and in the same way to an astonishing level of detail – as computer information is digital.
Of the three founding fathers of population genetics, it was Fisher who, in his great book of 1930, The Genetical Theory of Natural Selection, most clearly expressed the evolutionary significance of blending inheritance and its Mendelian antithesis.12 If genes did indeed blend, the variance available for selection would be halved in every generation. It’s the grey paint over again, but Fisher proved it mathematically. Mutation rates would have to be colossal – utterly unrealistic – to maintain the variation. Fisher quotes a letter from Darwin to Huxley, tentatively dated to 1857, before The Origin, which shows how tantalisingly close Darwin himself came to Mendelism:
… I have lately been inclined to speculate, very crudely and indistinctly, that propagation by true fertilisation will turn out to be a sort of mixture, and not true fusion, of two distinct individuals, or rather of innumerable individuals, as each parent has its parents and ancestors. I can understand on no other view the way in which crossed forms go back to so large an extent to ancestral forms. But all this, of course, is infinitely crude.
Even Fisher didn’t know how breathtakingly near Darwin really was to discovering Mendelian genetics, even working on sweetpeas! In 1867, he wrote a letter to Wallace that began as follows:
My Dear Wallace
I do not think you understand what I mean by the non-blending of certain varieties. It does not refer to fertility, an instance will explain. I crossed the painted lady and purple sweetpeas which are very different coloured varieties, and got, even out of the same pod, both varieties, perfect but non-intermediate. Something of this kind I should think, must occur with your butterflies … Though these cases are in appearance so wonderful, I do not know that they are really more so than every female in the world producing distinct male and female offspring.
That last sentence is a beautiful example of the power of reason, and the importance of seeing through the obvious. When a male mates with a female, you do not get a hermaphrodite. You get either a male or a female, with approximately equal probability. In a way, Mendel never needed to go into his monastery garden. All he had to do was take the inheritance of sex itself, and generalise it to all other cases of inheritance. Digital heredity was staring us in the face, in the most obvious way you could imagine. The trouble was, it was too obvious to be noticed. Darwin noticed it, and he came close to making the connection. But, just as Patrick Matthew didn’t quite cross the bridge that Darwin and Wallace crossed, so Darwin didn’t quite manage to cross the Mendel/Fisher Bridge – at least not decisively enough to answer Fleeming Jenkin.
I distinguished Bridge One from Bridge Two as ‘stabilising selection’ versus ‘directional selection’. But there’s more to it than that – or perhaps the distinction I am about to make really separates Matthew’s Bridge Two from Darwin and Wallace’s Bridge Three. I am talking about the distinction between selection as a negative force and selection as a positive, constructive force that puts together complex new ‘designs’. My own preferred way – the ‘selfish gene’ way – of explaining this is again to deploy ‘digital genes’, so perhaps we really have to cross Bridge Five in order to paint the full picture.
In modern genetic terms, not Darwin’s own, natural selection may be defined as the non-random survival of randomly varying coded instructions for how to survive. We see – and admire – the products, the phenotypes, of the successful instructions. The instructions are DNA and their most visible products are bodies that survive by doing something impressive such as flying, swimming, running, digging or climbing – all in the service of reproduction, which means they also tend to be good at attracting a mate and warding off rivals. An important part of the environment that each gene must exploit, if it is to ensure its survival in the form of copies of itself, is the other genes it encounters in the genomes of a succession of bodies – which, because of sexual recombination, means the other genes in the gene pool of the species. As a result of this, cartels of mutually supportive genes cooperate to build bodies that specialise in some particular method of surviving, such as grazing or hunting. Different cartels are the gene pools of different species, bound together by the remarkable phenomenon of sexual recombination – and separated from all other cartels, for it is part of the definition of species that they can’t interbreed. Occasionally, often through accidents of geography, gene pools find themselves subdivided for long enough to become sexually incompatible, and the subdivisions are then free to go their separate evolutionary ways as distinct species. Eventually, ‘separate ways’ can mean ‘very separate indeed’, for animals as different as vertebrates and molluscs originally split apart as members of the same species. Successive branchings of this kind have given rise to hundreds of millions of species, over thousands of millions of years.
At least in sexually reproducing species, evolution consists of changes in gene frequencies in gene pools. I stipulate sexual reproduction, because without it we have no clear idea what ‘gene pool’ even means. Where there is sexual reproduction, the gene pool is the set of available alleles from which the individual members of a species draw their genomes – ‘draw’ as in a lottery, the lottery of sex. Each individual genome is like a shuffled pack of cards. The available cards to be shuffled are sampled from the gene pool. The statistical frequencies of these available cards change as the generations go by, and that is evolution. We can monitor evolution by measuring a sample of the phenotypes – the anatomy and physiology of typical members of the population. As the average phenotype changes – as legs get shorter, horns longer, coats shaggier, or whatever happens to be evolving at the time – it is tempting to see natural selection as a sculptor’s chisel, carving the bones and flesh of the animals themselves.
But if we want to talk chisels, a sharper representation of evolution sees them as working not on the bodies of animals but on the statistical structure of gene pools. As crests get longer, or eyes rounder, or tails gaudier, what is really being carved by natural selection is the gene pool. As mutation and sexual recombination enrich the gene pool, the chisels of natural selection carve it into shape. We observe the results in the form of change
s in the average phenotype, and it is phenotypes that serve as the proxies for genes. As the external and visible manifestations of genes, they determine whether those genes are eliminated, or whether they persist in the gene pool.
Natural selection carves and whittles gene pools into shape, working away through geological time. It is an image that might have seemed strange to Darwin. But I think he would have come to love it.
1 W. Paley, Natural Theology (Oxford, Vincent, 1802); R. Dawkins, The Blind Watchmaker (London, Longman, 1986).
2 R. Dawkins, Climbing Mount Improbable (London, Viking, 1996).
3 D. Dennett, Darwin’s Dangerous Idea: Evolution and the meaning of life (New York, Simon & Schuster, 1995).
4 D. Hume, Dialogues Concerning Natural Religion (1779).
5 Patrick Matthew, Naval Timber and Arboriculture (Edinburgh, 1831).
6 W.J. Dempster, Evolutionary Concepts in the Nineteenth Century (Edinburgh, Pentland Press, 1996).
7 Unlike Patrick Matthew or Edward Blyth, Wallace was a Fellow of the Royal Society, although elected rather late – about thirty-five years after his landmark paper on evolution by natural selection. Darwin was elected in 1839, when still not yet thirty. Both Wallace and Darwin were honoured with the Society’s Royal Medal and Copley Medal.
8 The distinguished physicist Freeman Dyson has read it in exactly this sense, to buttress his own partiality for group selection.
9 The one exception – a rare exception in Darwin’s thinking – is his treatment of the evolution of human cooperation and kindness through a kind of group selection among rival tribes.
10 S.J. Gould, The Structure of Evolutionary Theory (Cambridge, Mass., Harvard University Press, 2002).
11 Kelvin’s attack centred on his (entirely erroneous) ‘demonstration’ that the Sun and Earth were too young to allow enough time for evolution. His calculations were based on the assumption that the Sun’s energy came from some kind of combustion. Pleasingly, it fell to Sir George Darwin FRS, Charles’ second son, to redo the calculations on the assumption that the Sun was a nuclear furnace and thereby vindicate his father.
12 I like to think that Ronald Fisher, arguably Charles Darwin’s greatest intellectual descendant, was also his intellectual grandson through his mentor, Major Leonard Darwin, the dedicatee of Fisher’s great book. Leonard, Charles’ fourth son, lived into my own lifetime and died on my second birthday, 26 March 1943.
10 HENRY PETROSKI
IMAGES OF PROGRESS:
CONFERENCES OF ENGINEERS
Henry Petroski, the Aleksandar S. Vesic Professor of Civil Engineering and a Professor of History at Duke University, is the author of more than a dozen books on engineering and design, including To Engineer Is Human: The Role of Failure in Successful Design and Engineers of Dreams: Great Bridge Builders and the Spanning of America. His newest book is The Essential Engineer: Why Science Alone Will Not Solve Our Global Problems. He is a Distinguished Member of the American Society of Civil Engineers; a Fellow of the American Society of Mechanical Engineers, the Institution of Engineers of Ireland, and the American Academy of Arts and Sciences; and is a member of the American Philosophical Society and the US National Academy of Engineering.
AS CELEBRATED IN THEIR DAY AS THE STATESMEN OF SCIENCE WERE THE GREAT ENGINEERS OF THE NINETEENTH CENTURY. HENRY PETROSKI EXPLAINS HOW THEY BUILT THESE AWESOME STRUCTURES AND WHY THEY ATTRACTED SUCH ACCLAIM.
One of the great engineering achievements of the nineteenth century was the expansion of the railways into an ever-widening network. Extending the right of way across major bodies of water naturally presented especially difficult problems for engineers, and so early railways often relied upon ferries at these locations. But this solution was not in keeping with the developing image of a fast and uninterrupted journey in a string of carriages pulled by a steam locomotive, and so bridges were built whenever possible. The most daring of these bridges, symbolic of the creativity, resolve, and integrity of the engineers that designed and built them, proved to be great engineering achievements in their own right, especially when the body of water to be crossed presented unique challenges, as it did at the Menai Strait.
This strategic strait, which separates the isle of Anglesey from the mainland of north-west Wales, was controlled by the Royal Navy, and so the Admiralty required that any bridge that was to cross it had to provide a clearance of at least 400 feet horizontally and 100 feet vertically so that tall-masted sailing ships of the day could pass between its piers and beneath its roadway without hindrance. Furthermore, because of the importance of the strait, temporary supports were not allowed in the water during construction. This virtually ruled out the choice of an arch bridge, which traditionally required the use of an elaborate system of falsework upon which the arch was assembled until it was self supporting.
Thomas Telford had already been presented with this problem when he was charged with completing the highway that connected London and Dublin and thereby providing a reliable route for the delivery of, among other things, the royal mail. The Irish Sea could only be crossed by ferry. The ideal location for a terminal was at Holyhead, which is on the west side of the island of Anglesey.
To carry the road from London to Holyhead meant bridging the Menai Strait. Telford initially wanted a cast-iron arch, which in 1811 he proposed to support by cables from above and thereby not obstruct ship traffic during construction. This untried method would have worked, as would be proven a half-century later, but it was not to be tried first at Menai. Instead, Telford designed the only other then-known bridge type that could span the distance and provide enough headroom: a suspension bridge.
The Menai Strait Suspension Bridge was completed in 1826 and remains an aesthetic paragon of what can be achieved with the form. Telford’s early experience as a mason enabled him to design graceful viaducts and towers bracketing the main span, which was a record-shattering 580 feet. He employed wrought-iron chains that were tested before installation, and the completed bridge was a structural marvel of its time. Unfortunately, the wooden roadway of the bridge proved not to be as substantial as its stone towers and viaducts and iron chains. When the wind was especially unfavourable, the roadway was susceptible to being tossed about, and on occasion it was destroyed.
When the Chester & Holyhead Railway was being laid out, routing its tracks across the Menai Bridge seemed the natural thing to do. However, as the wind had demonstrated, the structure’s roadway was light and flexible, and this would not serve the purpose of the contemporary railway. As well as the possibility of the road being destroyed in a storm, there was also the problem of a heavy steam locomotive causing the roadway of the bridge’s main span to deflect so much that the engine would have had to climb out of a valley of its own creation. The engineer George Stephenson suggested decoupling the train of carriages from its locomotive and using horses to pull the carriages to the other side of the bridge, where they could be coupled to another locomotive for the continuation of the journey. This was not what engineers would call an elegant solution.
Stephenson’s son, Robert, had a different idea. It involved designing a bridge that relied on neither the arch nor the suspension principle. Stephenson identified a site about a mile south of Telford’s Menai Suspension Bridge, where a large rock formation divided the strait into two wide navigation channels. Since this natural formation, known as Britannia Rock, was a recognised and accepted obstacle to shipping, there could be no reasonable objection to constructing a tall stone tower upon it. Similarly tall towers could also be erected outside the navigation channels on either side of the rock. Massive wrought-iron girders could then be installed at a sufficient height between these towers so that the vertical clearance was equal to that beneath the suspension bridge.
Robert Stephenson’s scheme was acceptable to both the railway company and the government, and so the detailed design and construction of the bridge was begun in the mid-1840s. Since no such structure had ever been designed, let alone built, it fell t
o Stephenson to organise what would today be called a research-and-development project. In order to keep the weight of deep girders exceeding 450 feet in length within acceptable bounds, it was decided early on that they should be hollow. At the time there existed no structural theory sufficiently advanced whereby the design of such girders could proceed by calculation alone. An experimental programme was thus embarked upon.
The experimentalist-engineer William Fairbairn, who had established a shipyard and had tested cast-iron beams years earlier, was responsible for conducting scale-model strength tests to establish the preferred shape and detailed design of the wrought-iron tubes. He began with small-scale models to compare the relative strengths of different shapes and arrived at the conclusion that a rectangular cross-section was the best. The model tubes were tested by hanging from their centre weights that represented the load of a heavy locomotive. Weights were added until the tube failed, which revealed the weakness of the structure and thereby provided guidance for how to modify it in the next model. By progressively increasing the scale of his models, Fairbairn was able to establish trends of behaviour, and from the experimental data the theorist Eaton Hodgkinson established an empirical formula by means of which he could extrapolate to the requirements for the full-size tube.