From: Subject: =?iso-8859-1?Q?Stephen_E._Jon?= =?iso-8859-1?Q?es_:_Creation/?= =?iso-8859-1?Q?Evolution_Arti?= =?iso-8859-1?Q?cles:_Colin_Pa?= =?iso-8859-1?Q?tterson:_addre?= =?iso-8859-1?Q?ss_to_the_Amer?= =?iso-8859-1?Q?ican_Museum_of?= =?iso-8859-1?Q?_Natural_Histo?= =?iso-8859-1?Q?ry=2C_New_York_C?= =?iso-8859-1?Q?ity=2C_1981?= Date: Mon, 1 Mar 2004 22:34:41 -0500 MIME-Version: 1.0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Location: =?iso-8859-1?Q?http://members?= =?iso-8859-1?Q?.iinet.net.au/?= =?iso-8859-1?Q?~sejones/patta?= =?iso-8859-1?Q?m06.html?= X-MimeOLE: Produced By Microsoft MimeOLE V5.00.3314.1001 Stephen E. Jones : Creation/Evolution Articles: Colin = Patterson: address to the American Museum of Natural History, New York = City, 1981

Stephen E. Jones

Creation/Evolution Articles

Colin Patterson's address to the American Museum of Natural History, = New=20 York City, 1981, pages 6-10:

[Home = page] [Site map] = [Update = log] [Articles]=

Pages [1-5] [6] [7] [8] [9] [10] [11-14]

page 6

SECOND PARABLE

Now I'd like to go on to the second parable and return again to this = same=20 diagram but I'd like you now to think of the way in which it was = previously used=20 by Ernst Mayr in his 1969 book with his 1974 paper on "Systematics." At = that=20 time it looked like this:

15%
10%
70%

Now, that's how Mayr tried to put some precision into evolutionary=20 systematics. Then he said, "Let A be the common ancestor of BCD and = suppose the=20 genome of B is diverged from the ancestral genome by l5%, that of C = diverged by=20 10%, and D has moved off rapidly into some new adaptive zone diverged by = 70%."=20 Then he said we should be quite wrong to classify C with D which appears = towards=20 its nearest relative by common ancestry because relationship means = inferred=20 amount of shared genotype not the inferred recency of common ancestry. = Here is=20 the very simple sum:

B and C share 75% of the ancestral genotype
C and D share only = 30%

..... show the kind of mistakes that might be made, Mayr said that=20 systematists might group crocodiles with birds rather than other = reptiles or=20 might group African ape with man rather than with the orangutan. So Mayr = in his=20 knowledge of evolution is making predictions about the genotype of = crocodiles=20 and African apes.

In the first example crocodiles, birds, and reptiles, he predicts = that the=20 proportion of genotypes shared by C, the crocodile and B, another = reptile be=20 greater than proportion shared by C, the crocodile and D, the bird. He = predicts=20 that in some shared genotypes BC will be greater than CD.

Now next is Mayr ..... demonstrating the explanatory power here of a=20 hypothesis of common ancestry, something that I've [said] scarcely had = zero=20 explanatory power. He is also demonstrating a knowledge of evolution, = writes a=20 theory, makes a prediction and is going to test it.

Last month in Ann Arbor, a student of Morris Goodman gave me the = amino acid=20 sequences for alpha and beta hemoglobin of 3 crocodilians: a caiman, a = Nile=20 crocodile the Mississippi alligator. We already knew the alpha and beta=20 hemoglobin sequences of 2 birds, a chicken and a goose but the problem = is=20 finding another reptile. As far as I know the only other reptile = available at=20 the moment is the alpha hemoglobin sequence of a snake, the viper. Alpha = hemoglobin is 143 amino acids long so there 3 times that or 429 = nucleotides=20 long. That is a very small part of the genotype but at least it is worth = checking. The prediction is that the amino acids common to B the viper, = C the=20 crocodile and D the chicken, that BC would be greater than CD. And here = of=20 course are his findings:

BC =3D	8 out of 143	5.6%
CD =3D	25 out of 143	17.5%
BD =3D	15 out of 143	10.5%

(Patterson C., "Evolutionism and Creationism," Transcript of Address = at the=20 American Museum of Natural History, New York NY, November 5, 1981, p.6) = [Top of=20 page]

page 7

Here we are. The theory makes a prediction, we've tested it and the=20 prediction is falsified precisely. CD far outweighs BC so something is = wrong=20 with the prediction. Something is wrong with the theory. But, of course, = we know=20 that falsification is never absolute, for you're never sure what it is = you have=20 falsified.

And here we see only three possibilities. First, we've falsified the = data;=20 there is something wrong with alpha hemoglobin. Second, we've falsified = the=20 diagram; there is something wrong about the physical relationships of = the snake,=20 bird and crocodile. And third, we've falsified Mayr's knowledge of = evolution,=20 either the particular stuff about rates and adaptive zones or something = more=20 general.

We can check one of those pieces of data by taking another sample of = the=20 genome: a crocodile, bird including chicken and two other reptiles a = lizard and=20 turtle. Myoglobin is 153 amino acids long.

BC	16 out of 153	10.5%	(B is lizard)
CD	13 out of 153	8.5%	(C is crocodile)
BD	16 out of 153	10.5%	(D is chicken)

This time the BC proportion of the genotype is larger than the CD, = slightly=20 larger, as Mayr predicted. What happened here? What fits BD? It ought to = be by=20 far the smallest proportion of the genome, yet it is exactly the same as = the BC.=20 Something has gone wrong again. Again there are 3 possibilities: data, = diagram=20 or the theory

In fact, when you think about the diagram and this data but yet a = different=20 diagram BC same as BD, and possibility of dichotomy, like that. Perhaps = the=20 diagram is wrong.

Let's check again with the turtle, turtle myoglobin. B is the turtle=20 (terrapin as I remember), C is crocodile and D is chicken.

BC	11.8%	(B is turtle)
CD	5.2%	(C is crocodile)
BD	5.9%	(D is chicken)

Now we're getting closer to what Mayr expected. BC is much larger = than BD=20 here. Again there is another problem. The BD portion (turtle-bird) = should be=20 tiny [but] is larger than the CD. Again, something is wrong. This data = suggests=20 a different diagram like this where BC go together, D is separate.

So it seems with these three examples, either we get the right = diagram as we=20 did with the first sample .... with the viper and alpha hemoglobin but = we had=20 the wrong proportion, or we get the right proportions as here but the = wrong=20 diagram or we get a set of data that is just agnostic that gives us a=20 trichotomy. So what is it we are falsifying? There aren't any more = lizard=20 sequences available, no more reptile sequences.

Two things we might do: One is to accept Mayr's assumption that = reptiles are=20 a group and to sum up the data and take averages. You add together 3 = kinds of=20 beings: the turtle, viper or the lizard; 3 kinds of crocodiles; 3 kinds = of=20 chickens - when you do that, the 3 summations are for BC you get 27.9, = for CD=20 you get 30.9 and BD 26.9. They are virtually identical...

(Patterson C., "Evolutionism and Creationism," Transcript of Address = at the=20 American Museum of Natural History, New York NY, November 5, 1981, p.7) = [Top of=20 page]

page 8

The creationist makes one assumption - that there are some groups in = a set of=20 data. The evolutionist, I think, has to make another assumption - that = there=20 some groups in there and that the groupings tell us something about the = history=20 of groups.

Someone from audience says: "Which creationist?"

Patterson: "Alright, me."

Audience member: "You don't mean Duane Gish?'

Patterson: "No, I don't. I mean a creationist systematist."

Audience member: "Duane Gish doesn't make any assumptions."

Other member: "Oh, yes he does."

Patterson continues lecture

I think I am able to treat that DNA .... data, taking all of it as = saying=20 something, but I only find two signals in it. In order to get a tree out = of it,=20 you have to infer that some of it is good data - the stuff that gives = you that=20 (at board) and some of at is bad data - the stuff that gives you that = (at=20 board). Sounds like [in] this case the stuff that gives you BC and AB is = more=20 than twice as numerous as the stuff that gives you BC is inferred. Now = what is=20 it that is being paired here? These are all identities at the nucleotide = level,=20 the ones that gives you AB are identities. The ones that give you .... - = they=20 are all identities at the nucleotide level. Yet somehow you'll have to = say that=20 some of those identities are ...... the same. Some of them are not the = same,=20 because the theory demands that. It seems to me somehow nonsensical.

There is one more problem with homology at the DNA level and that is = this=20 business of plurality, or inferred duplication. There is a similar = problem=20 coming in beside ...A. A couple of weeks ago Roger Lewin had a piece in=20 Science (Oct. 23, 1981) on globin genes and in it he touched on a = model=20 of DNA that is now factual among molecular geneticists. He called it the = "Vesuvian" model and a simple description of it by Roger Lewin and = others is=20 that every gene is constantly bombarding the rest of the genome with = pseudogenes=20 which are more or less perfect copies of itself. Now if this is so, and = the=20 model does have empirical support, then the problem of plurality of = duplication=20 of DNA is even more pressing.

In order to do DNA sequencing, you mix up the genome and clothe = [clone?] a=20 bit that you think is right that you can get out. If there really are = all these=20 bombarding pseudogenes lying, around I see no possible way of knowing = whether=20 you've got the right one or not.

Well, I'm sorry to have on so long about that and the point of it is = that I=20 think it has something to say about evolution because of all the = discussion in=20 the last few years whether evolution is tastable, and by evolution here = I mean=20 the general theory, the descent with modification, that species are = mutable and=20 related by descent rather than a specific theory about mechanisms.

If the general theory of evolution is testable, it must have some = function=20 ...... that can be confronted with reality. In other words, it must = make

(Patterson C., "Evolutionism and Creationism," Transcript of Address = at the=20 American Museum of Natural History, New York NY, November 5, 1981, p.8) = [Top of=20 page]

page 9

some predictions. As far as I know only one sensible prediction has = been=20 offered. Niles Eldredge put it like this in a letter to = Science.

"If evolution is descent with modification, the hierarchical array of = organisms defined by nested sets of evolutionary novelties must result. = This is=20 evolution's grand prediction."

Then Niles went on to say that whatever organism you look at, = whatever aspect=20 of it you study, you find the same hierarchy. I've heard this same point = made=20 repeatedly, at meetings. There really is a hierarchy and there can be no = hierarchy without history and therefore the prediction of evolution is = met.

The first thing that strikes me about this is that it seeps [steps?] = inside=20 of evolution as a deductive inference from the systematic hierarchy. The = people=20 like Cuvier, Linneus, Hooker and a thousand other pre- Darwinians were = merely=20 fore- thinkers but they failed to see the necessary...

The second thing concerns the prediction of whatever aspect of = organisms you=20 look at, you find the same hierarchy. Not everyone wants to agree with = that.=20 Here's Ernst Mayr again in Science. Different types of = characters:=20 morphological characters, chromosomal differences, enzyme genes, = regulating=20 genes, and DNA matching made these a rather different groupings. = Different=20 stages in life cycles result in different groupings.

Here's Arnold ...... with the conclusion of his cladistic study of = apes and=20 men. His study denotes a clear lack of congruence between molecular and = other=20 more traditional kinds of data. Notice they are both saying the same. = Mayr is=20 saying that molecular data or whatever level you look at it, it doesn't = seem to=20 matter, morphology and Arnold are saying the same thing.

Now, the prediction. of evolution, according to Niles is that = hierarchy and=20 congruent hierarchy is what we'll find no matter what aspect of the = species you=20 look at. ...... the experience of Mayr and Arnold that there is no such = thing as=20 congruence. In particular, molecular data are incongruent with = morphology.

Well, is that so? I'm not sure but I think there are other signs of = it among=20 protein sequences. For example, that cladogram that I was building up = earlier is=20 one of the congruents with everything that we know about morphology. Now = there=20 are published cladograms that have bird with sister group of mammals, = Nile=20 crocodiles, snakes and sister groups of all other amniotes. Never mind = the=20 details of it.

Let me put the question at the most basic level. How is it you = recognize a=20 hierarchy? At the level of the phenotype I think you have no real = problem. We=20 have a fairly rational concept of homology and I agree with Gary that = the=20 organizing principle is ontogeny and Von Baer's* law in particular. And = as Gary=20 suggested we can analyze phenotypification that way and get a hierarchy = by a=20 method that has no evolutionary implication at all. There is a history = ahead but=20 the history is on the genetics. The history is a what gives you the = direction in=20 Von Baer's law. We don't have to infer anything more about the = geological=20 [genealogical?] history, for example.

So what about this molecular level or the level of protein and DNA = sequences?=20 How do we recognize a hierarchy there? First of all, the concept of = homology=20 is

* Note: the original has "Conbear's"

(Patterson C., "Evolutionism and Creationism," Transcript of Address = at the=20 American Museum of Natural History, New York NY, November 5, 1981, p.9) = [Top of=20 page]

page 10

estimates of the old autapomorphies we derived. The gibbon has 66=20 autapomorphies, 53 orang, 21 gorilla, chimp 19 and man has only 14. = Notice that=20 is the exact reverse of Mayr's prediction... So we get no useful = grouping out of=20 that.

The next set of characters is the one that picks out 3 of the 5.

ABC	30
ABD	11
ACD	10
CDE	10
ADE	9
ABE	7
BCD	7
BDE	4
ACD	3
BCE	3

One signal here - ABC form a group. Since there are 10 ways of = picking 3=20 species out of 5, the probability of a repetition by chance in that sort = of data=20 is 1 in 10. This group has 19 repetitions over its nearest competitor, = so far as=20 I understand it; you can't be getting a signal like that. The = probability of=20 getting a signal like that purely by chance is 10^-18.

The last set of characters pick out 2 from the 5. Again there are 10 = of them.=20 I won't list them all. There is only one that gives a signal .... E. The = rest of=20 them are grouped very much like this with as far as I can see no = distinguishable=20 differences between them and here the probability according to my = calculation is=20 10^{- 17}.

So as I understand this data, the information that is in there, is = there are=20 2 groups ABC ... DE; together they form a larger group and that's all = the data=20 says. It also gives an estimate of the number of autapomorphies to be=20 grouped...

So what about the tree here and the numbers on the branches? As Steve = said,=20 it is produced by a program. Those numbers don't pop out of the data in = any way,=20 so I suppose those come from massaging the data with evolutionary = theory. It is=20 a program that assumes evolution to be true and tells the computer to = find a=20 tree. So my question will be: What is the tree telling us about? Is it = telling=20 us something about nature or something about evolutionary theory?

One last thing, at this level of DNA, the level of DNA, we also have = a=20 problem of homology. What does homology mean in terms of DNA? The = alignment=20 procedure is the same with protein sequences, its a purely statistical = business=20 but because in DNA we only have 4 possible nucleotides in any one = position, we=20 expect a 25% match by chance alone. Amongst these 5 very closely related = species=20 there is only a 7% match, that leaves a 45% variation to accommodate all = other=20 eukaryotes. I think that the problem with aligning DNA...will be=20 extremely...

Then in the level of individual nucleotides we have to assume that a = match of=20 one position, say an adenine in the same position of all these, is = homology. A=20 mismatch is a non-homology. So bearing that notion in mind, let's look = again at=20 the tree. The tree tells us that - no, I'm going to forget all that. I = was going=20 to talk about the effects of putting data through a tree generation = program.

(Patterson C., "Evolutionism and Creationism," Transcript of Address = at the=20 American Museum of Natural History, New York NY, November 5, 1981, p.10) =

[Top of=20 page]

Copyright =A9 1999-2002, by Stephen E. Jones. All rights reserved. = This page=20 and its contents may be used for non-commercial purposes only. If used = on the=20 Internet, a link back to my home page at http://members.iinet.net.au= /~sejones=20 would be appreciated. Created: 2 December, 1999. Updated: 23 January,=20 2002.