No, Dr. Mayr did not dislike the beach. That's just how all the kids are saying, "You ain't down with the shit I'm doing." Anyway, read this post attacking the physicists for calling us stamp collectors on a blog with a very cumbersome name. MissPrism puts a throw-away link at the end of the post to Roger Lewin's 1982 interview with Mayr (available here if you have a subscription to JSTOR). I actually like MissPrism's post better than Lewin's article, so you're not missing much if you can't get access to the Mayr interview. I have some quotes from Mayr, and my responses, below the fold.
Both Mayr and MissPrism point out that physical scientists don't appreciate biology because they don't understand evolution. MissPrism goes so far as to suggest that physicists are "jealous because biologists got our Grand Unified Theory in 1859." Very nice.
Mayr presents an example that illustrates how some non-biologists don't understand the biology and, therefore, fail to grasp evolutionary theory:
"Few physical scientists understand the uniqueness of individuals," suggests Mayr. "I was at a conference at the Wistar Institute in 1967 where a group of evolutionists and physical scientists, including mathematicians, met to ask, was 4.5 billion years long enough for the evolution of diversity and adaptation seen in the world? The physical scientists said no, we can calculate that the time is nowhere near sufficient. And we evolutionists argued that it was indeed sufficient."
The two sides disputed for 3 days, and then dispersed with nothing resolved. Mayr says that it wasn't until about 9 months later that he realized what "mistake" the physicists and their companions had made. "They assumed that all the individuals in a species are identical, just as all atoms of sodium are identical, for example. For them, a mutation has to spread through all the individuals in the population, and this must be followed by another mutation, and so on. If one were to adopt such a process of tandem evolution, no amount of time would be sufficient to account for the diversity we see now."
This sounds a lot like the modern day creationists (calling their form of bullshit "Intelligent Design") making erroneous assumptions in a feeble attempt to disprove evolution. It sure is easy to argue against something when you distort reality. But it appears that Mayr also thinks that theoretical evolutionary biologists are disconnected with reality:
"Two years ago I saw a paper in the Proceedings of the National Academy of Sciences, and the author wrote, 'Let's assume the gene has a constant selective value; let's assume there is no gene flow from any other population.' He made about five such assumptions, each of which was equally unrealistic, and then he went on to prove something very beautiful mathematically, but it was meaningless."
Did Mayr not understand the purpose of models? In modeling any natural phenomenon (be it evolutionary, ecological, or physical), we simplify the real world. It's impossible to take every variable into account, so we hold many of them constant while we examine others. Modelers do not claim to be recreating the real world. One problem with increasing the amount of variables in your model is that you increase variance. There is a trade-off between being realistic and minimizing noise. Maybe Mayr was just a stamp collector...
Mayr also points out the differences between evolution in astrophysics and evolution in biology:
"In physical phenomena, rates are rather constant, whereas in biological evolution they are incredibly variable. Look at the Isthmus of Panama, which, when it was formed 3.5 million years ago, divided into two what previously was a single ocean. Some species on either side of the barrier have remained indistinguishable whereas others have evolved so far as to become different genera.
"There are lagoons around Lake Victoria in Africa, one of which was separated from the lake about 4000 years ago. Of the six species of Lake Victoria fish in the lagoon, five have become new species. At the other extreme there is Limulus (the horseshoe crab) and Nautilus, neither of which has appreciably changed for at least 100 million years and perhaps as long as 250 million. You can see, the rates of biological evolution are extraordinarily variable, differing by three to five orders of magnitude. This would be extremely unusual in a physical process."
Another point of departure is in the mode of change. "In cosmic evolution changes are continuous," suggests Mayr. "Every once in a while there might be the origin of a nova or something, but generally it is a continuous process. In biological evolution, when each new generation arises, the formation of a new individual results in a new assembly of genetic material, and the process of selection starts anew."
We'll ignore that fact that genera are arbitrary classifications. But does Mayr have a point? Is cosmic evolution a continuous process? Does biological evolution proceed at variable rates?
The contrast Mayr creates is actually a product of the examples he chose. Phenotypic evolution does not proceed at a constant rate. Certain attributes may change rapidly whereas others remain constant for long periods of time. Evolution on the molecular level, on the other hand, does have a very constant rate. Now, Lewin's article was published at the very infancy of the molecular revolution, and the neutral theory was less than 15 years old. The mounds of data available to us today reveal that much of the genome is evolving at a regular rate. This process is analogous to the constant cosmic (and geological) processes, whereas rapid phenotypic changes can be thought of as the catastrophic cosmic events.
In the end, everything is governed by physical laws. Biology just has so many variables that we cannot take the same reductionist approach used by physicists. Although this approach is often very practical, you do need to step back and look at the big picture to ensure you have not gone too far down the rabbit hole. All disciplines of science use both hypothesis driven and discovery driven approaches. They all use empirical and theoretical methods to understand the natural world. Where many scientists stumble when talking about a research area outside their own is in understanding the wealth of knowledge in other disciplines. The physicists who did not understand how alleles are transmitted through a population made this mistake in arguing that evolution was mathematically impossible.
- Log in to post comments
The issue many physicists are running into now is that you can't even take the reductionist approach when trying to solve what should be bread-and-butter physics problems!
As for the 'grand unified theory' of evolution by natural selection, to a physicist it isn't a very satisfying theory because it is extremely tricky to predict anything specific. It is a wonderful coincidence that many systems that we physicists have studied in the past (say reasonably dilute gases, or magnetic spins on a lattice) have "easily" predictable outcomes after random perturbations of the system.
That's not to say there's anything _wrong_ with the theory of evolution, it is just extremely foreign to those of us brought up on ideal gases and celestial mechanics. Give us time.
I don't know if saying "molecular evolution occurs at a constant rate" is a good idea. Certainly there are chemical rates involved that are constant, but given the varieties of ways that molecular evolution can occur (HGT, recombination,retroviral integration, retrotransposons, etc) and how little we know about the effects of these elements on populations, I think this is a stretch. It is fairer to say "we have molecular clocks that we can use to calculate evolutionary rates on geological/evolutionary timescales; these appear to change at constant rates." But those clocks are the exception, it seems, rather than the rule. IMHO, of course :).
From the little Mayr I've read, I suspect he was not that fond of molecular phylogenetic techniques. I'd probably chalk it up to him being a human being who was quite advanced in age and career when molecular techniques became available.
The mutations you mention, Paul, are the exception, not the norm. The most common mutation is a single nucleotide substitution, and these are the prefered tool of molecular clockers. The other types are important (hell, I study these other types), but they are far to rare to be used for clocking. The subsitution rates for nucleotide to nucleotide (or amino acid to amino acid) changes work quite well.
Yeah, models are great, and I think Mayr understood their value. There is a point at which, however, any model falls apart on faulty assumptions. Mayr was perhaps too cautious in his trust of models, but skepticism is fundamental to any model's success, so Mayr may have also served his role well.