Part II:
Arlin: I was looking over some of Senapathy's writing, and I have to admit that it makes me angry to see him claim that he has demonstrated in some way that unsplit genes are "astronomically unlikely" and split genes are "very likely." He has done nothing of the sort! He has never made a valid side-by-side comparison of two models.
Senapathy says that genomes (and this really only applies to mammalian genomes -- genomes of plants, fungi, protists, etc., have fewer & smaller introns, if they have introns at all) show a "near-perfect" exponential distribution of reading frame lengths. This is false, but I won't quibble with it right now. For the sake of argument, let us say that Senapathy is right and that the mammalian genome reflects its origin as a random nucleotide sequence.
How common are split genes in the mammalian genome? The length of a mammalian gene is typically 5 to 100 kb, with an average of roughly 20 kb. Including upstream and downstream flanking regions and inter-genic spacers, the length of DNA per gene is more like 30-40 kb. You can check to see that this figure makes sense by considering that we have about 50,000 genes (some estimates go as high as 80,000) and 2,000,000,000 bp. That is, we have 2,000,000,000/50,000 = 40,000 bp per gene.
How does this compare to random UNSPLIT genes? That is, how much random DNA is there per 200-codon ORF (open reading frame)? This is easy to calculate. If the chance of a non-stop, or "sense," codon is s, then the probability of a randomly generated ORF (open reading frame) with a length of at least L codons is simply
s^L
Thus, the chance that an ORF will be at least 200 codons long is s^200, which is 0.0000674 (given s = 61/64 sense codons, i.e., 3 nonsense codons as in the canonical genetic code). Now we just need to calculate how many ORFs there are in a random DNA sequence to get the density of 200-codon ORFs.
We need to make some simple assumptions to calculate this. Let us assume, as Senapathy does in his 1986 paper, that an ORF is any string of sense codons (i.e., it doesn't have to start with a methionine codon, as do most modern genes). And let us assume that an ORF may occur in any of the 6 different reading frames in a double-stranded DNA (dsDNA) molecule. That is, there are 6 ways to render a dsDNA sequence in triplet codons (3 forward phases, and 3 backward phases), and we will assume that all are allowable, as is the case in real genomes. This means that in a nucleotide sequence of N bp, there are 6 different reading frames, each with N/3 codons, for a total searchable codon- string-length of 6 * N/3 = 2 * N. Each nonsense codon can be counted as the end of an ORF, thus there are 2 * N * (1 - s) ORFs in a sequence of length N. Thus, the probability that an ORF of length L will occur in a DNA sequence of length N is:
P = 2 * N * (1 - s) * (s^L)
Or, we can set P = 1 and solve for the average amount of DNA needed per ORF of length L:
N = 1 / ( 2 * (1 - s) * (s^L) )
Now all we need to do is solve for s = 61/64, and L = 200, to get the average length of DNA needed per 200-codon ORF. The answer is
N = 157,782
Now our comparison is complete.
DNA per split gene in a mammalian genome: 40 kb. DNA per 200-codon ORF in a random sequence: 158 kb.
I readily admit that the comparison isn't exactly fair. I am comparing random 200-codon ORFs to real genes, a large proportion of which have several times more than 200 codons. I would LOVE to make a completely fair comparison, by calculating the DNA needed per random split gene, but Senapathy never provides the details that would allow such a calculation, because he never specifies what is the splice site target.
The point is that Senapathy has not, and cannot, rule out randomly occurring unsplit ORFs on the basis of probabilities. Unlikely though they may be, they are not "astronomically unlikely." At best, I suspect that if Senapathy were to actually provide details of a model, you would find that the probabilities are equal. At worst, if the above comparison is accurate, then we would just need a slightly bigger "primordial pool" of DNA to allow unsplit genes rather than split ones.
Whew! just had to get that off my chest. Now I need to get to work. Keep in touch.
Jeff: Could you explain a little about what you mean by that ("splice site target")? Is this information that only he has?
Arlin: I would guess that he has not considered it. But you can consider it with a gedanken experiment. Imagine a ticker-tape machine connected to your computer. Have it print a random string of letters. Now take a scissors and cut the tape at an arbitrarily chosen position in a pre-determined string of characters. For example, cut at the 2-character string ";I", just after the ";". Now tape together every second piece of ticker-tape or every third piece or whatever you like. It doesn't matter. You will find that none of the resulting ticker tapes will make any more sense than the original. You will also find no difference in statistical properties of the sequences. Even the interval or "wait length" (Senapathy's term) between target strings or substrings (e.g., ";") will be the same. In fact, I cannot think of any global property of the original sequence of characters that will be altered by this procedure [perhaps you can think of something].
Obviously, in order to claim that splicing would somehow improve the chance of finding an open reading frame, Senapathy has to come up with something different than just a simple target sequence. Suppose that (to continue the gedanken experiment) we are set with the task analogous to Senapathy's task. We must design a set of rules to increase the chance that the strings ligated from cut pieces of ticker-tape will be intelligible sentences. Here is a start: three things that a sentence cannot have in it are ".", "?" and "!", otherwise it would not be one sentence, but more than one. If we just use random strings of 128 different characters, one of these sentence-terminators would occur every 43 characters. Sentences in standard written English in a newspaper or magazine are typically 100 to 300 characters long; some of my longer sentences may run to 500; unless you read Joyce, you will rarely see a sentence as long as 1000 characters, which is roughly the length of the 5 sentences in this paragraph.
So, let us cut the random string of characters just before each ".", "?" or "!"
Oops! This doesn't help! If we cut before each of these, then each resulting piece (except the first) will begin with ".", "?" or "!"! That is, if we ligate the 3rd piece to the 1st, the initial sentence will be no longer, since the 3rd piece (and every other piece) will by definition begin with ".", "?" or "!". As in the previous example, the distribution of characters will be unchanged!
So, if we pick a random cut site, we can't change the length of sentences. Even if we pick a cut site that is non-random with regard to sentence-beginnings and sentence-ends, we still cannot change the resulting length of sentences.
In order to increase the sentence length, it is absolutely essential that we distinguish upstream and downstream splice sites. That is, we must have one type of cut at the end of a fragment-to-be- saved/beginning of a fragment-to-be-discarded, and another type of cut at the beginning of a fragment-to-be-saved/end of a fragment-to-be- discarded.
Likewise, to rationalize Senapathy's model, we must have a different exon/intron and intron/exon splice site. An even stickier problem involves phases. That is, a ".", "!" or "?" is always a sentence- terminator, but a stop codon is only a stop codon when it is in the correct phase (e.g., the "TGA" in AAT GAA is different from that in CAA TGA, though they are biochemically identical). How do we determine whether a "TGA" sequence must be cut off of an exon and discarded? -- how do we know we are not screwing up a gene by cutting out a "TGA" that is actually supposed to mean "AAT GAA"?
However, even if we rise to the challenge of the ticker-tape experiment, we will only have succeeded in lowering the frequency of sentence- terminators in the resulting sequence of characters. This increased sentence length does not ensure that the sentence will have meaning.
Another fundamental problem is empirical -- the data suggest that upstream and downstream splice sites both have a sequence roughly like AG^GT, where "^" is the cut site. As mentioned above, if upstream and downstream cut sites are identical, it is theoretically impossible to change the information content of the resulting spliced sequences.
Jeff (posted in s.b.e): ...the method driving new organisms. With evolution, it's natural selection and descent with modification. ...
Arlin: Natural selection isn't necessarily in the driver's seat with respect to speciation, though some educated people favor this view. Darwin specifically proposed that natural selection is the most common general cause of adaptation. That is, if something has a feature that "fits", appears to benefit the organism, ensure its survival, etc., this is likely (so goes the theory) to be attributable to the cumulative effects of natural selection. Natural selection is not necessarily the answer to those other big questions of evolution: "why are there so many species?" or "why does species X exist."
Jeff (continuing on s.b.e): With Senapathy, it's a new assembly of DNA and then a simple pass/fail test of viability.
Arlin: Maybe the test that will distinguish Senapathy's ideas from Darwin's in your mind would be a test that would show that organisms have been subjected to multiple successive pass/fail tests -- with different criteria each time! -- instead of a single test.
Here is an anecdote illustrating what I mean. Some of my details may be wrong here, but you could talk to an anatomist to clarify them. There is a nerve going to your larynx that emerges from the medulla with the vagus nerve, at roughly the level of the hairline on the back of your neck. You would think that this nerve would make only a short journey to the larynx, which is that cartilaginous tube (with vocal chords inside) that makes the "Adam's apple" in males. However, this nerve, called the "recurrent laryngeal nerve", takes a torturous course. The left branch loops under the systemic aortic arch (the aorta), which is roughly at the level of the nipples, then travels back up to the larynx. Why?
The assumption that there was a single pass/fail test in which all the nerves and blood vessels must fit together would not help us to understand the path of the recurrent laryngeal nerve.
However, an explanation is possible if, instead, one starts with the assumption (some would say it is more of a conclusion than an assumption, based on other grounds, but that would be to beg the question) that primates and other mammals evolved ultimately from animals that were more like fish or lampreys. Comparing such organisms, one would conclude that, in the evolution of mammals, the aorta has dropped down into the chest, while the enhanced "encephalization" of mammals has squeezed alot of nerves and neurologically important organs up into the head and neck region, leaving the circulatory and breathing apparatus in the chest. In our postulated fishy ancestors, these things were spread around in the same region, along with sets of aortic arches to correspond to the gill arches. Throughout evolution (one surmises), as the aortic arches dropped back, this path became more and more torturous.
There are many other examples in which it is possible to explain twisted plumbing in our anatomy by proposing this sort of historical explanation. For instance, the ureter, which (if I recall correctly) in females emerges from the kidneys and travels up over the fallopian tubes before descending again to the bladder, and in males is tangled with the vas deferens. The vas deferens itself is a wild case in point. The vas deferens carries sperm from the testes, passing by the urethra on its way up over the ventral side of the pubic bone back over the bladder and ureter and through the prostate gland, where it joins the urethra. If all of these parts had been subject to a single pass/fail test (or had been designed intelligently), the vas deferens would simply emerge from the testis and join the urethra immediately without ever leaving the scrotum.
Make sense?
JM: I suppose such a test could be used, but I don't see how you could say with certainty that such things were not from the pond. However, I have a problem with your examples. You wrote:
Arlin: The assumption that there was a single pass/fail test in which all the nerves and blood vessels must fit together would not help us to understand the path of the recurrent pharyngeal nerve.
JM: I spent a lot of time thinking about this and other interesting and unique organs and behaviors of organisms. Senapathy mentions quite a few examples (Chapter 10, page 480-487) and Michael Denton does, too (Evolution: A Theory in Crisis). I agree that there are many amazing characteristics in life, and they are hard to contemplate without thinking: how could this happen "randomly"? However, at the genetic level, all these complexities disappear. There is no relationship between the complexity of a genome and the complexity of the organism it specifies (Senapathy, page 301-306). (I find it amazing that only 30,000 genes or so are needed to make a human.) So, I find it plausible that given a random genome, we'd find many random, unique, and curious attributes. So, why is there still so much order in the "random" result? Well, organism have to be viable, so that eliminates 99.999% of all seed cell trials and likely many interesting characteristics (perhaps viability is inversely related to having some of those interesting features -- too weird to work). Also, order tends to "accumulate" in the pond because new seed cells have reused material from previously successful (and generally "boring") organisms.
Furthermore, just because we see so many interesting and useful characteristics (organs and behaviors taken together) does not mean they came about through selection. Why aren't there more such things? We can imagine other useful characteristics that we do not observe, and we can realize there are multitudes of things NOT seen. So, perhaps only 0.00001% of all possible nifty things actually exist, and so it no longer seems so fantastic to me.
You wrote: However, an explanation is possible if, instead, one starts with the assumption (... conclusion...) that humans evolved ultimately from animals that were more like fish or lampreys.
I understand that at some point a human embryo has "gills." There are developmental genetic pathways in the human genome that are not fully expressed. Some pathways are partially "on" during the embryonic stage, but later those pathways are turned "off." Here's my interpretation of the new theory on that. Rather than being left-overs inherited from a distant ancestor, these unused pathways are random inserts or tag-alongs from the pond. Perhaps there's a pathway in the human genome for growing feathers, but the genes are never turned on, or only part of the necessary feather genes are there. Maybe the genes and proteins for which we can find no function are remnants of these unused pathways. This is my own extension of Senapathy's theory -- I don't recall that he got into this very much.
So, to me your examples seem to fit the new theory -- at least they don't refute it.
Arlin: I have to say that the response to the recurrent laryngeal nerve was disappointing. I claim that there exists a class of observations that demand a historical explanation, that such a class is predicted if one assumes common ancestry/descent- with-modification, and that no such class is predicted or explained by Senapathy's theory. Your response to this is, to paraphrase, 'well, it isn't predicted or explained, but who is to say it couldn't have happened by chance?'
Yes, Senapathy's theory can allow for weird and poorly-designed part to arise by chance. However, Senapathy's theory is inconsistent with the observed recurrence of poorly designed parts. Furthermore, the descent with modification theory can make predictions about what the weirdness will look like, whereas Senapathy's theory cannot. That is, you have gone beyond Senapathy's logic in suggesting that perhaps mammals may have extra pathways for feathers. This is not consistent with Senapathy's theory. They might have an extra enzyme or two by chance, but an entire pathway does not get assembled unless it has been selected to be part of a viable seed cell, in Senapathy's theory. To suggest otherwise is a bit like waiting for the wind to blow through a junkyard and assemble a 747 (which is possible, metaphorically speaking, under Senapathy's theory), then expecting that the 747 will have the parts for a Yamaha 650 inside, because, hey, who is to say that it might not have some extra pathway for transportation in it, just by chance?
The theory of descent with modification allows that cases of weirdness may have a historical explanation discernible in the features of close relatives. If we wonder why we have a useless appendix, we may look at other tetrapods and find an explanation superior to "just by chance". If we want to know why whales have hips, we may look at other mammals and find an explanation that is better than "just by chance". [BTW, not all weird things can be explained this way. For instance, if we want to know why humans have protruding noses and fleshy buttocks, we're out of luck, because these weird features are not seen in close relatives].
The pattern that will emerge is: weird organ (whale hips) with no apparent function --> similar organ with not-so-mysterious function (e.g., hips with legs attached) in a relative. I see no reason why such a pattern should be seen under Senapathy's theory. If we find the same organ (e.g., an appendix) appearing in different organisms, this means that it was independently selected -- EVIDENCE of its usefulness under Senapathy's theory. That is, the hips of sperm whales, right whales, humpbacks, harbor porpoises, orcas, etc., are useful organs selected at the seed-cell stage. they are not remnants of the hips of a terrestrial ancestor. Likewise, under Senapathy's theory, the reduced tail of chimpanzees, gorillas, and (yes) humans cannot be a glitch, a random mistake, because (under Senapathy's theory) it occurred three times independently ("independent" as in not related as in having nothing to do with each other as in Independent Birth of Organisms). Therefore, having a reduced tail must have contributed to the survival of seed cells. You cannot have it both ways. Historical explanations are the exclusive property of the descent-with-modification theory, which is denied by Senapathy.
JM: Now it's my turn to be disappointed. Just as I thought we were making good progress teaching each other about the two theories, I find that you have missed some important aspects of Senapathy's theory. I think you just don't see these things because of your acceptance of the prevailing theory -- a presupposition of major proportions, if you will. I will try to explain...
You wrote: I claim that there exists a class of observations that demand a historical explanation, that such a class is predicted if one assumes common ancestry/descent-with-modification, and that no such class is predicted or explained by Senapathy's theory.
JM: I believe such a class is predicted by Senapathy's theory. But a better word would be "allowed," in that if the class did not exist, it would not refute independent births. I don't know if the absence of the class would refute evolution (you can tell me) -- probably not, so "allowed" is a better word there, too.
You wrote: Yes, Senapathy's theory can allow for weird and poorly-designed part to arise by chance.
JM: OK, I think we just agreed on "allowed." I hope you would also agree that finding such a class is not "required" by Senapathy.
You wrote: However, Senapathy's theory is inconsistent with the observed recurrence of poorly designed parts. Furthermore, the descent with modification theory can make predictions about what the weirdness will look like, whereas Senapathy's theory cannot.
JM: There is where you tossed in your presupposition of evolution. Descent with modification will predict those things. But proving that they came about via that method is another story. Evolutionists have constructed a tree based on those characteristics and characteristics that are correlated with those things, and also on your belief of evolution. Poorly designed organs is part of the assumed ancestor-descendant structure, and that data has been used to place creatures on the tree. Then you say that evolution predicts something that was used to describe evolution. Circular.
I can hear you screaming: "No! -- the tree is based on molecular data." Since when? There have been trees around since before any molecular data or even DNA was known, and they have not fundamentally changed a whole lot since. Now you are saying: "But the new molecular data agrees with the old trees!" Yes, and where that is true, it could be because the molecular compositions of the creatures we see today are the result of in-pond reuse of genomes. In that case, there will be the appearance of an ancestor-descendant tree, but not at the living level. The in-pond tree (if one could be constructed) would be the same as your evolutionary tree, but more accurate. (I previously described such a tree as looking more like a web.) This is what you have missed in the independent birth theory.
Please look at figure 8.8 on page 323, and the text starting on page 321.
You wrote: They might have an extra enzyme or two by chance, but an entire pathway does not get assembled unless it has been selected to be part of a viable seed cell, in Senapathy's theory.
JM: [Note that as point "1" for reference below]
But, it could. An entire unused feather pathway is possible. This would be less likely to occur in two seed cells that were constructed at different places or times, but what if a bird fell and died in the soup and one of its cells got tangled up in the construction process for some other creature? You might get a bird-like-thing, but not a bird. I just used the bird/feather concept to illustrate my point -- please don't get hung up on feathers. Maybe the human appendix would be a better example. It could be all or part of some pathway for some organ that was perhaps useful to another creature, now extinct. Maybe it's a misplaced pathway or a bad copy of some other organ. Maybe it's just a random mistake. This will eventually -- long time away -- be figured out by looking at more molecular data.
You wrote: The theory of descent with modification allows that cases of weirdness may have a historical explanation discernible in the features of close relatives. If we wonder why we have a useless appendix, we may look at other tetrapods and find an explanation superior to "just by chance". If we want to know why >whales have hips, we may look at other mammals and find an explanation that is better than "just by chance". [BTW, not all weird things can be explained this way. For instance, if we want to know why humans have protruding noses and fleshy buttocks, we're out of luck, because these weird features are not seen in close relatives].
JM: Sometimes I wonder whose side you're on. :-) I like those contrary examples. But you are not being completely honest, because you should add this inside your brackets: "and THIS IS predicted by Senapathy." Because you have been fair in your messages, I believe this oversight is evidence that you do not fully understand the new theory, or are unable to release yourself from the grips of evolution to contemplate all of the new theory. There's more:
You wrote: The pattern that will emerge is: weird organ (whale hips) with no apparent function --> similar organ with not-so-mysterious function (e.g., hips with legs attached) in a relative. I see no reason why such a pattern should be seen under Senapathy's theory. If we find the same organ (e.g., an appendix) appearing in different organisms, this means that it was independently selected -- EVIDENCE of its usefulness under Senapathy's theory.
JM: When you used "selected" way up above (at reference point "1"), I thought it was just a neutral word being used, meaning "chosen," not having the special connotation of natural selection. But, you just used that word again in the same sentence with "Senapathy" and "usefulness." And now, a third time...
You wrote: That is, the hips of sperm whales, right whales, humpbacks, harbor porpoises, orcas, etc., are useful organs selected at the seed-cell stage. They are not remnants of the hips of a terrestrial ancestor. Likewise, under Senapathy's theory, the reduced tail of chimpanzees, gorillas, and (yes) humans cannot be a glitch, a random mistake, because (under Senapathy's theory) it occurred three times independently ("independent" as in not related as in having nothing to do with each other as in Independent Birth of Organisms). Therefore, having a reduced tail must have contributed to the survival of seed cells.
JM: There is no reason it could not have occurred three times. In fact, it probably occurred 1377 times (just my estimate :-), but most of those creatures never made it. There are plenty of characteristics we see reused more often than three times, and a few that are reused EVERY time (necessary for viability). Just because a characteristic appears three times (or more) does not mean it must be useful to the organism. Why don't we have four or six fingers (including the thumb)? We could survive just as well either way. Is five so much more useful that you can say we have only five because it helped us survive? My point is: we see some weird organs and we also have five fingers -- these are not much of a surprise to me, it just happened that way.
But your use of "selection" in the same sentence with "Senapathy" and "usefulness" is a surprise, and probably not a random occurrence :-), so I must get back to that. There is no natural selection operating in the pond. The only way you could get there was from viable organisms contributing their DNA back to the pond. Senapathy calls this "random perfection," and you can find more about that via his index. You wrote: You cannot have it both ways. Historical explanations are the exclusive property of the descent-with-modification theory, which is denied by Senapathy.
JM: They are only historical at the living level. But there is a history in the pond too, and if I had been there at the time, I could have documented it. I'd see the genomes of previously-viable organisms being picked apart and reassembled in the pond in all kinds of interesting and random ways. Most of those random ways would produce failures; some would not. Because viability is required in order to have creatures to study in the first place, there are going to be things about that creature that appear to be related to other creatures, and along with that will be traces of the in-pond history, or a "pond-tree" (my new word). "Pond-web?"
Arlin: I did not simply talk about "weirdness" in general, but instead tried to explain that a particular subset of weirdness would, to a greater-than-random (greater-than-Senapatheistic) degree, follow a historical pattern. I am using "weird" as a casual term to mean anything that seems out of place if we assume (as Senapathy and Darwin do) that organismal features have been subject to some degree of selection. More objectively, we can define "weird" as something that is at odds with a priori design specifications. For instance, chimpanzees have a tail. How would we build a tail to fulfill a function? Tails in other animals are used for balance (e.g., cats and many others), grasping (opossums, some monkeys), defense (some lizards) and warmth (foxes and squirrels). Perhaps you can think of some other use for a tail. The chimpanzees tail is suited for nothing that I can think of. We could experiment surgically, cutting off the tails, to see if there is any effect. You accuse me of making special assumptions to use "weird", and that this is a historical designation, but the term can easily be made objective, as described. Let us call them "useless or radically non-optimal features" if this makes you more comfortable.
The tail of the chimp can be explained by invoking a tailed ancestor, under the assumption of evolution (before objecting to this statement, please read about 'assumptions' below), whereas no such explanation is available under Senapathy's theory. However, not all weird things will have this kind of historical explanation. I explained this in my previous letter, giving a specific example of how a weird thing does not necessarily have a historical explanation.
When I use "recurrence" I am instead presupposing Senapathy's theory, since in evolution the tail of vertebrates (for instance) does not "recur", but only arose once in a common ancestor. In Senapathy's theory, tails have arisen thousands and thousands of times.
Also, be aware that there is such a thing as reductio ad absurdum, where one "presupposes" or "assumes" certain conditions, then sees whether or not they lead to falsehood.
I cannot prove to you that descent with modification explains everything. We are engaged in this conversation because I have claimed that Senapathy's theory is false, and you have differed. Darwin's theory may also be false, but
(Quoting JM): I can hear you screaming: "No! -- the tree is based on molecular data."
Arlin: No, this is not what I am screaming. I am screaming at myself for not having clearly demonstrated the argument that is most relevant here, the argument about congruence that I presented two months ago. I agree that a single tree, taken by itself, cannot show common ancestry, since it is based on the assumption of common ancestry. Not only do I agree with you on this point, I was the person who explained this to you, with the examples of manuscripts and so forth (remember?). I never claimed that a single tree was evidence of anything. ONE TREE proves nothing, disproves nothing. Two trees, three trees, many trees, IF THEY ARE CONGRUENT WITH EACH OTHER, DISPROVE INDEPENDENT BIRTH, whether it's creationism or Senapatheism. It is an observable FACT that trees are congruent. Therefore, Q.E.D., Senapathy's theory is false. The point is that I don't use the trees to "prove" evolution (impossible), I use them to disprove creationism and Senapatheism (possible, and actual).
There are two other specific tests based on the distribution of features of organisms that can disprove Senapatheism and creationism. One of them is the test for correlated characters. Another one is the test for hierarchical structure. The null hypothesis in the first case is complete independence (as in Independent Birth of Organisms). If correlated features are found, this means that the features arose non-independently. Such a result is consistent with common ancestry, but does not prove it; such a result is inconsistent with creationism and Senapatheism. In fact, characters are correlated -- the null hypothesis of independence can be excluded.
However, Senapatheists might argue that correlation could arise through 1-step selection. This is where the hierarchical structure comes in. In this case, the null hypothesis is that characters may have arisen non-independently, but they arose only by a unitary process, not by way of common ancestors. For example, suppose that I characterize people's gardens by what things are planted, how they are arranged, etc. I would definitely find correlated characters, since some gardeners plant flowers and others plant vegetables; also, Nova Scotia gardeners plant different things than Carolina gardeners, given the different growing seasons. This is analogous to your modified Senapathy view, in which there is non-independent re-sampling in the pond. The null hypothesis of independent characters would be falsified. However, the null hypothesis of separate, non-hierarchical structure test would not be falsified.
If the new theory is correct, five-fingered limbs have appeared thousands of times independently (as in "Independent Birth of Organisms"). It would also be the case that eyes have occurred thousands of times independently. We can justify why eyes would be used over and over again (actually, we can't, because seed cells can't see, and even if they could, all they would see is DNA -- but that's yet another wretchedly illogical stupid nonsensical aspect of Senapathy's theory that is not relevant to this point), because they are useful, but why would five-fingered hands occur over and over again? You cannot just brush it off by saying "parts can get re-used". Which parts? Why? Is there a predictive science of parts re-use, or do you just invoke re-use whenever it is convenient?
One of the major problems to be explained in biology is the apparent purposefulness of organisms. Creationists invoke a designer to solve this problem. Darwin invoked "natural selection". Senapathy has to invoke some sort of selection or sorting or whatever, some process whereby things-that-work are separated from things-that-don't-work. His procedure is consummately stupid, since the "seed cells" cannot have arms, eyes, fur, etc., yet he expects us to believe that only functionally-armed, -eyed and -furred organisms would arise. The survival of genomes that can be replicated and can form functional seed cells is a naturally selective process acting on genomes. The continued survival of seed cells, and their division (if they divide) are also naturally selective processes. If you want to have successful seed cells contributing their dead DNA to the pond, to be re-used, that is also a naturally selecting procedure. If you want to have only some seed-cell-generated organisms survive, while other perish, that is yet another round of natural selection. These are all natural selective procedures, and they will all contribute to changing the constitution of the pond. You cannot pretend that natural selection and evolution do not exist. These are principles, not hypotheses. Natural selection will happen under all sorts of circumstances, the only question is whether it is negligible, or will influence your system to some unknown degree.
Back to [Part I]
Continued in [Part III]