Macroevolution: What Were the Evolutionary Ancestors of Whales?

ID supporters often assert that, although there is abundant empirical evidence for microevolution, there is no such evidence for macroevolution. One of the macroevolutionary transitions that they often cite is the evolution of whales from a land-dwelling ancestor. Which leads me to ask the following question:

What empirical evidence would verify (i.e. support) or falsify (i.e. undermine) the hypothesis that whales had evolved from a land-dwelling mammal? Please note that this is a hypothesis about macroevolution, not microevolution.

A basic principle of hypothesis validation in the natural sciences is that if one can find multiple lines of evidence, all of which support the hypothesis, then such evidence is much stronger than if there were only a single line of evidence. This is especially the case if the different lines of evidence are from very widely separated fields.

Until recently the main line of evidence for the evolution of whales (i.e. members of the mammalian order Cetacea) from even-toed ungulates (i.e. members of the mammalian order Artiodactyla) was anatomical. This anatomical evidence was derived from two sources:

1) similarities between the anatomy (especially skeletal anatomy) of living (i.e. "extant") Artiodactyls and Cetacea, and

2) an evolutionary phylogeny of the transition from terrestrial Artiodactyls to aquatic Cetacea, based on fossils.

Rather than summarize this comparative anatomical evidence here, I recommend that interested readers follow this link. What you will find is a fairly detailed summary of the evidence from comparative anatomy, all of it pointing to the conclusion that whales (i.e. Cetaceans) evolved from even-toed ungulates (i.e. Artiodactyls). It is this evidence that most evolutionary biologists have until recently cited as support for the macroevolutionary derivation of Cetaceans from Artiodactyl ancestors.

However, one can also ask the question Does the comparative genomics of Artiodactyls and Cetaceans support the same hypothesis? That is, are there observable DNA sequence similarities and differences that are similar in both scope and timing to the similarities and differences in the fossil record (and as reflected in the comparative anatomy of Artiodactyls and Cetaceans)?

This is an easily falsified hypothesis: If the genomic evidence does not support the Artiodactyl into Cetacean hypothesis — e.g. that Cetaceans evolved from some other clade, or that they had not evolved at all, but rather sprang into existence fully-formed and without genetic evidence of a macroevolutionary transition — then this evidence would not support the evidence from comparative anatomy and the macroevolutionary hypothesis based on comparative anatomy would be called into question.

So, what does the comparative genomic evidence indicate about the macroevolutionary relationships between the Artiodactyla and the Cetacea? Here’s a summary of the findings from comparative genomics:

The idea that whales evolved from within the Artiodactyla was based on analysis of DNA sequences. In the initial molecular analyses, whales were shown to be more closely related to ruminants (such as cattle and deer) than ruminants are to pigs. In order for the order name to reflect a real evolutionary unit, the term Cetartiodactyla was coined.

Later molecular analyses included a wider sampling of artiodactyls and produced a more complete tale. Hippos were determined to be the closest relative of whales, ruminants were related to a whale/hippo clade, and pigs were more distant. In addition to producing the controversial whale/hippo clade, these analyses debunked the idea that hippos and pigs are closely related. This had been a popular taxonomic hypothesis (Suiformes) based on similarities in morphological (physical) characteristics.

In addition to DNA and protein sequences, researchers tracked the movement of transposons called SINEs in the genome. A transposon is a DNA sequence that will occasionally make a copy of itself and insert that copy into another part of the genome. It is considered highly unlikely that SINEs will insert themselves into the exact same part of a genome by chance. The data indicate that several transposons inserted themselves at the same point in the genomes of whales, ruminants and hippos (sometimes referred to as "pseudoruminants" because although they have four-chambered stomachs like true ruminants, they do not chew the cud). This insertion point is not shared with camels and pigs.

This hypothesis has been tested with DNA sequences from a host of genes: the complete mitochondrial genome (as well as several of its genes independently), beta-casein, kappa-casein, von Willebrand factor, breast cancer 1, recombination activating genes 1 and 2, cannabinoid receptor 1, and several others. These sequence data and the transposons converge on the same conclusion that hippos and whales are more closely related to one another than either is to other artiodactyls.

Sequences analyzed in combined analyses with morphological characters have also produced the same results as sequences alone. Some have argued that the sheer number of characters (one for each nucleotide) in sequences swamps out the effects of morphology. There have been a few morphology-based studies that have suggested (weakly) the same results as the molecular results, but overall most morphological studies have conflicted with the whale/hippo hypothesis of Cetartiodactyla.

An important exception is a recent conducted by Boisserie et al. (2005). They examined 80 hard morphological characters of fossil and extant cetartiodactylan taxa. Their results suggest that hippopotamids evolved from within a clade of anthracotheres. That anthracothere/hippopotamid clade appears to be sister to the Cetacea and supports the molecular results.

[sources: http://en.wikipedia.org/wiki/Cetartiodactyla (summary article), where you can find links to many primary references]

Note that much of the genomic data (especially from transposon sequences) supporting the macroevolutionary hypothesis is based on non-adaptive DNA sequences. That is, DNA sequences that do not code for adaptive characteristics, and in many cases that do not code for anything at all. This is like figuring out which students have been copying the answers to test questions from other students by comparing their wrong answers. The right answers are the same for everybody, but wrong answers vary from student to student in virtually random ways. If two students have the same wrong answers, you can be reasonably confident that one of them copied the wrong answers from the other. You can then test this hypothesis by looking at seating charts, past test performance (cheaters are often identified by sudden increases in test scores without apparent increases in effort), and – often the last resort – asking them if they copied answers.

Conclusion: The empirical evidence from comparative genomics closely matches the empirical evidence from (both extant and fossil) comparative anatomy.

Is that all, or is there yet another line of evidence that might be pursued to verify or falsify the Artiodactyl into Cetacean hypothesis? Yes, there is. Consider the observable fact that whales reproduce much more slowly than even-toed ungulates (such as deer and hippos). Indeed, there is a general principle in zoology that the larger the members of a species are (on the average) the fewer offspring they have, the more widely spaced those offspring are in time, the fewer offspring they can have over their lifetime, and the longer the average lifespan of individuals.

For example, deer can have offspring every year, and under good conditions can sometimes have twins or even triplets in one reproductive cycle. By comparison, baleen whales can only have offspring every few years (it can take up to two years for one pregnancy in large baleen whales), they virtually never have more than one calf at a time, they have only a few reproductive life cycles per lifetime, and they have much longer lifespans than deer.

This means that, if Cetaceans evolved from Artiodactyls, one might be able to find empirical evidence that the rate of the macroevolutionary transition from Artiodactyl ancestors into Cetacean descendants had slowed down as the result of the increase in size, decrease in number of offspring per reproductive cycle, decrease in total number of offspring per lifetime, and increase in average lifespan. In brief, there might be evidence that the macroevolutionary “clock” slowed down as Cetaceans evolved larger and larger size.

Here’s the latest genomic evidence vis-a-vis this hypothesis:

Big and slow: phylogenetic estimates of molecular evolution in baleen whales (suborder mysticeti).
Molecular Biology and Evolution. 2009 Nov;26(11):2427-40. Epub 2009 Jul 31.
Jackson JA, Baker CS, Vant M, Steel DJ, Medrano-González L, Palumbi SR.
Marine Mammal Institute, Hatfield Marine Science Center, Oregon State University, OR, USA.

ABSTRACT: Baleen whales are the largest animals that have ever lived. To develop an improved estimation of substitution rate for nuclear and mitochondrial DNA for this taxon, we implemented a relaxed-clock phylogenetic approach using three fossil calibration dates: the divergence between odontocetes and mysticetes approximately 34 million years ago (Ma), between the balaenids and balaenopterids approximately 28 Ma, and the time to most recent common ancestor within the Balaenopteridae approximately 12 Ma. We examined seven mitochondrial genomes, a large number of mitochondrial control region sequences (219 haplotypes for 465 bp) and nine nuclear introns representing five species of whales, within which multiple species-specific alleles were sequenced to account for within-species diversity (1-15 for each locus). The total data set represents >1.65 Mbp of mitogenome and nuclear genomic sequence. The estimated substitution rate for the humpback whale control region (3.9%/million years, My) was higher than previous estimates for baleen whales but slow relative to other mammal species with similar generation times (e.g., human-chimp mean rate > 20%/My). The mitogenomic third codon position rate was also slow relative to other mammals (mean estimate 1%/My compared with a mammalian average of 9.8%/My for the cytochrome b gene). The mean nuclear genomic substitution rate (0.05%/My) was substantially slower than average synonymous estimates for other mammals (0.21-0.37%/My across a range of studies).

CONCLUSION: The nuclear and mitogenome rate estimates for baleen whales were thus roughly consistent with an 8- to 10-fold slowing due to a combination of large body size and long generation times. Surprisingly, despite the large data set of nuclear intron sequences, there was only weak and conflicting support for alternate hypotheses about the phylogeny of balaenopterid whales, suggesting that interspecies introgressions or a rapid radiation has obscured species relationships in the nuclear genome. [emphasis added]

So, there are indeed empirically falsifiable hypotheses for the macroevolution of whales from land-dwelling ancestors. If whales (Cetacea) evolved from even-toed ungulates (Artiodactyla), then the following predictions should be supported by the observable data:
• that there should be anatomical similarities between extant Artiodactyls and Cetaceans,
• that there should also be anatomical similarities between fossil Artiodactyls and Cetaceans,
• that there should be shared similarities and differences between the genomes of extant clades of Artiodactyls and Cetaceans, and that the overwhelming majority of these similarities and differences would mirror the comparative anatomical evidence for the macroevolutionary origin of the various clades of the Cetartiodactyla, and
• that the inferred slowing of macroevolutionary change during the transition from Artiodactyl ancestors to Cetacean descendants should also be consistent with the hypothesis that the rate of this transition would have slowed as the result of increasing body size, increasing reproductive spacing, decreasing numbers of offspring per life cycle, and increasing longevity.

And it is.

Clearly, an ID supporter might then ask for specific empirical evidence on how the various transitions occurred at the genetic and developmental level, and if these details could unambiguously distinguish between natural and supernatural causes for such genetic mechanisms. Evolutionary developmental biologists are currently working on answers to the first part, but I personally cannot imagine how one could empirically test the second part. Furthermore, it seems to me that invoking a supernatural cause for the macroevolutionary transition from Artiodactyls to Cetaceans would be unnecessary, and would add nothing whatsoever to our understanding of the mechanisms by which this transition occurred.

Ergo, if I were doing this research and publishing my results I wouldn’t mention it, as it would be completely unnecessary for a scientific explanation of this phenomenon.

Just out of curiosity, ask yourself how one might use any of the foregoing as positive or negative empirical evidence for the existence of God. I mention this because some evolutionary biologists believe they can use the data of evolutionary biology to disprove the existence of God, and some ID supporters believe they can use the data of evolutionary biology to prove the existence of God. Personally, I believe both attempts are misguided, pointless, and ultimately futile. That’s why I don’t make such attempts, and wonder why anyone would.

************************************************

As always, comments, criticisms, and suggestions are warmly welcomed!

--Allen