A really fun family outing in San Diego is to visit Sea World and see the many fascinating and exciting marine exhibits. But the unquestioned main attraction is Shamu, the killer whale. If you are a real bona-fide thrill-seeker, you sit in the first few rows next to the tank, virtually guaranteeing that when the sleek but massive animal breaches the water and then falls back, you will be inundated by a huge wave and soaked to the skin! How did such marvelous creatures arise in the first place? It has taken many years of patient work by scientists operating in very different specialties, but we are now at the point where we can relate the “Whales’ Tale.” It is a story of evolution over a critical period of about ten million years, which is supported by three main types of evidence. We will consider the first two types of evidence in this essay: molecular and fossil.
If evolution is true, then modern whales and other mammals should be related to previously living ancestral species, through a process of “descent with modification.” It should therefore be true that the living organisms and ancestral ones (now extinct) should form a sort of “family tree.” If you have taken an interest in your family genealogy, then you know right away what this means. You, your siblings, parents, aunts, uncles, grandparents, and so forth, can be arranged in a diagram that passes from one generation to the next. If we visualize this going deep into the past, we can use the “tree analogy” even further – the most recent generation of members of the family can be said to lie at the tips of the branches, while very early generations of the family would lie deeper in the tree, at branching points.
The metaphor of using a tree to represent ancestry comes in other varieties too—not just families. Consider, for example the growth and diversification of the historic Christian church – from its roots in ancient history to the tips of its branches—the various denominations still in existence today. As shown below, the Christian traditions which are especially closely related to each other are located near one another at the branch tips. The more distant the relationship, the further away they are in the tree of Christian traditions.
So how can one derive the family tree for organisms like whales—how can one determine the tale of the whale? Cetaceans, after all, have such a dramatically different body plan compared to all other mammals; deciphering their family tree presents a fascinating challenge. If evolution is true though, there is one group of organisms to which whales are more closely related than any other. Furthermore, if evolution is true, independent ways of deriving tree structure ought to produce very similar results.
In today’s essay we will show two methods that have enabled biologists to trace the lineage of the whale family: two somewhat independent methods that allow us to explore the structure of the whale’s family tree. In our next post, we will examine a third.
The instructions on how to build an organism are contained within the four letter DNA code: A, G, C, and T. Each gene is a short stretch of this code and the specific order of the 4 letter code is called its “sequence.” The cells of the organism read the code, gene-by-gene, working in concert with one another in constructing the body. Because it is very different than that of other living mammals, understanding the origin of the whale body presents an interesting challenge. Whales are mammals though, so if evolution is true they must have a family tree which shows how they are connected to other groups of mammals.
One useful source of information in whale family tree construction is the sequence of the DNA code-letters (bases) in a particular gene in whales compared to the sequence of that same gene in other mammals. Why would this information help us? Both whales and their related mammalian “sibling and cousin” species will each possess a version of whatever gene we look at that was inherited from their common ancestor. Random mutation will have changed each version of the gene slightly, so that the descendant organisms will generally each have a distinct sequence. More closely related species will have a more recent common ancestor, and will, therefore, have more similar sequences. This means they will tend to lie closer together in our reconstructed family tree.
We can put this DNA gene sequence information from whales and comparison mammals into a tree-building computer program. The living organisms form the tips of the branches and the interior branch points represent extinct predicted ancestral organisms. It turns out that whales sit closest in the tree to a set of hoofed mammals including cows, sheep, pigs, camels, and hippopotamuses.1 This entire group of hoofed mammals is technically called the “Artiodactyla” (Greek for “even toed”). If evolution is true, this means that whales and these even-toed hoofed mammals share a common ancestral species that existed much more recently than the ancient common ancestral species that gave rise to all mammals. Indeed, even before that there would have been a common ancestral species that gave rise to all mammals and all reptiles. All of this can be represented on the metaphorical tree of life.
There are other independent ways in which DNA analysis can be used to test whether we have correctly positioned whales on the tree of life. Scientists are always eager to obtain different sorts of data. If all independent methods lead to the same conclusion, if “all roads lead to Rome” to use the analogy introduced in an earlier essay, then we can become increasingly convinced that our model is correct. So what is another DNA feature that can be used to determine the whale family history? There are certain chunks of DNA which, on rare occasions in the history of life, move to a new location in a chromosome. These mobile chunks of DNA are sometimes called “jumping genes” although it should be emphasized that they don’t “jump” very often. The location at which a jumping gene inserts itself into a chromosome is quite random. When such an element inserts itself into a particular place in the chromosome, it will reside at that location for many generations. Indeed since “jumping” is so rare, it generally stays at the same location for millions of years.2 Since the insertion process is almost random, and the element almost never moves out once it is in a chromosome at a particular position, the chance that a “jumping gene” will be in precisely the same place in the chromosome of unrelated organisms is vanishingly small—(essentially zero). In other words, the “jumping gene” makes an ideal “marker” to trace the ancestry of living species. If you examine a set of such “jumping genes,” each inserted into a particular place in the chromosome, only related organisms will share a particular insertion, since they inherited it from their common ancestor. If one of a pair of organisms lacks this insertion at this site, it supports the conclusion that those two organisms do not share a recent common ancestor.
The figure below shows a set of chromosomes, and then enlarges one part of one chromosome to show the DNA molecule. Imagine a “jumping gene” moving in precisely between two of the millions of units of DNA in a chromosome. Since DNA replicates each generation, the chromosome with its inserted “jumping gene” gets passed on faithfully through millions of years. Once a piece of DNA has moved into a chromosome between two bases, it is a great marker to identify species that descend from a common ancestor.
One of the very nice things about this type of DNA information is that it can be tabulated, and is simple enough that you can do a little head scratching and puzzle out the relationships of the organisms involved. The data either consists of a particular “jumping gene” being present (call that a “1”), or if it is absent (call that a “0”). In practice we need a third category, and that is “we don’t know if the “jumping gene” was there or not” (call that a “?”). This third category is necessary because sometimes a random genetic event will result in the loss (deletion) of the entire region which might have contained the jumping gene insertion. Now with this background, take a look at the following figure.3 For this somewhat simplified example, we show 20 “jumping genes.” If two species share a “jumping gene” at exactly the same position, this means those species are derived from the same ancestral species. This tree confirms the prediction made based on DNA sequence data previously, that is, that whales should be closely related to the group of even-toed hoofed mammals. For example, whales share “jumping genes” 10,12, and 18 with a broad assortment these animals. This means that they all share a common ancestor with insertions in these exact same positions. No other living organisms will share this group of common insertions, or this common ancestor. In addition, these data show that whales are most closely related to hippos (note that they each share “jumping genes” 4, 5, 6, and 7). (In fact, DNA gene sequence studies also support such a relationship, so this is not an aspect of using “jumping gene” data alone).4
Now we come to the bottom line: so far we have two roads (DNA sequence data and “jumping gene” data), both of which lead to “Rome.” Both point to exactly the same conclusion. Whales, despite their highly specialized body form, can now be confidently predicted to lie within the group of even-toed hoofed mammals. Furthermore, of that group of living mammals, hippos are predicted to be the most closely related to whales. There is agreement between two types of DNA data, and more confidence in our result.
Therefore, if evolution is true, we would expect that living whales and living hoofed mammals should share extinct common ancestors, from which they descended with modification. Or, put another way, we should be able to find “transitional fossil forms” which we can identify by their structural features as being ancestral to both living hoofed mammals and also whales. But about how long ago would we expect such extinct forms to have been alive? It turns out that application of DNA data once again can give us a time estimate with which to start.
We mentioned above that random mutational changes to DNA in an ancestor are passed on to descendant organisms. It turns out that for a particular gene, this sort of change acts as a sort of “molecular clock.” That is, for a particular gene, the rate of change over time is approximately constant. If we can “calibrate” how fast a particular molecular clock for a particular gene is ticking, then we can use it to determine how long ago in the past two species last shared a common ancestor. For example, we know from the fossil record (which has been dated by radioactive isotope clocks, as discussed in a previous essay), that cows and pigs last shared a common ancestor about 55-60 million years ago. We can measure the total number of changes in the DNA of a particular gene in cows and pigs, divide that by the age of a fossil from an ancient species believed to be ancestral to both of them, and determine an average rate of DNA change. Our molecular clock for this gene is now calibrated. If we want to determine when whales last shared a common ancestor with cows, and then pigs, we can measure the total DNA change in our clock gene between whales and cows, and between whales and pigs. We can then divide by the rate of “ticking” of the clock, and determine when in the past these ancestors should have lived. When we do this, it turns out that such common ancestors should have lived about 45 to 50 million years ago.1 So if evolution is true, we should expect to find fossil “transitional forms” showing evidence of common ancestry of hoofed mammals and whales, dating from about this period.
Previously, we learned that a tree summarizing species relationships can be built using DNA information, and how we can use DNA as a “molecular clock” to date ancient events. Both of these methods have made specific predictions about the origin of whales. If evolution is true: whales are related to the even-toed hoofed mammals and should share common ancestors with them; transitional fossil forms dating from about 45 to 50 million years ago should be found which can be shown to be related to both the even-toed hoofed mammals and modern whales; whales are most closely related to modern hippos, and should share a common ancestor with them.
What other types of information might we be able to use to construct a phylogenetic tree (i.e. a family tree) of species relationships? It turns out that characteristics of body structure also can be used—for example, the presence or absence of certain bones, or the specific shapes of those bones. An advantage of using bony features is that they can be recovered from fossils, whereas DNA (with only certain limited exceptions) must come from living organisms.
We can also derive functional information from an examination of bony features. The various protrusions, bumps and knobs found on bones usually have important implications. For example, smooth rounded areas at the ends of bones allow them to fit together and move easily. The shapes of such surfaces determine which bone motions are “allowed” or “disallowed.” Consider, for example, the motion of the forearm against the upper arm at the elbow. This is a “hinge” joint, whose normal motion is defined by the shapes of the upper arm bone and one of the forearm bones, where they meet each other. You might normally exercise the action of this hinge joint when you pick up a cup of coffee, bring it to your mouth, then set it back down again. Let’s try to imagine another motion. For this exercise we first need to get our arm into the proper starting position. Place your arm at your side, bent at the elbow at a ninety degree angle, with your palm up. Now, while keeping your palm up, let’s attempt to move your arm only at the elbow (no shoulder motion – that’s cheating!). Now swing your forearm out to the side and attempt to end up with your fingers pointed directly away from your side. Most of you will not be able to do this. If you can, it’s because your shoulder is rotating in spite of yourself. This motion at the elbow is normally not allowed. Hence a careful analysis of bone shapes can allow us to infer how the bones were used. This in turn can assist us in the task of phylogenetic (evolutionary) classification of organisms. That is, we will have more confidence in the grouping together of animals in our tree diagram if corresponding bones are used functionally in the same way.
Therefore we would expect that we could use various bony features to help us examine the predictions generated by our previous look at different types of DNA data. Are there any bony features that are particularly relevant to the even-toed hoofed mammals? Well, it turns out that there are. These are mainly running animals, and there are several features of their ankle bones, which taken together define the “allowed” motions which make them efficient runners. If one takes the various ankle bones of a large group of mammals, examines them carefully to note their shapes, scores that information into a table, then uses a computer program to build a phylogenetic tree, it turns out that all the even-toed hoofed mammals are placed together. So far, so good. But what about whales? Well, now we have an obvious problem. Modern whales are very specialized—they have flippers which correspond to the forelimbs, and they have almost no hind limbs! I say “almost” because they do have small pelvic bones, which are not attached to the rest of their skeletons. But they certainly have no ankles. This is where the fossils should come in—if evolution is true, we should expect to be able to identify transitional fossils which are ancestral to whales which contain the characteristic ankle bony features of the even-toed hoofed mammals.
Now let’s look at bony features from the whale perspective. We have already mentioned the almost complete loss of hind limbs, and the presence of forelimbs modified into flippers. In addition, as air breathers, whales have a blowhole at the top of their skull. And as powerful swimmers, which use a large tail fluke in vertical motions, whales have enormous sets of muscles which attach to enlarged projections from their vertebral column. So if evolution is true, we should begin to see fossil forms which manifest changes in bony features which correspond to the gradual accumulation of these whale-like characteristics. However, we still need more, because these various bony features all would be expected to occur in largely or exclusively aquatic forms. We might expect this to correspond to the later stages of a transition from terrestrial even-toed hoofed mammals. But what about the earlier stages? It would be very helpful if we had some “defining” characteristic of whales, similar to the ankle structure of even-toed hoofed mammals.
It turns out that the structure of the bones of the skull and ear apparatus of whales are highly modified to allow efficient hearing underwater. The mechanical aspects of efficiently receiving sound through water are somewhat different than receiving sound traveling through air. If evolution is true, we should expect to be able to find key transitional fossil forms with a progressive series of modifications of the skull and ear bones, features which would not be found in any other mammals.
Now that we know what we should expect to see, if evolution is true, let’s look at what has actually been found in the fossil record. Over the last fifteen years or so, a series of fossils, many of them discovered in the Indian subcontinent, have fulfilled nearly all of our predictions.1,2 The entire fossil progression illustrated occurs from a little over 50 million years ago to about 40 million years ago. So a remarkable alteration in general body form occurred in a little over 10 million years. This time frame agrees well with the previous prediction from the DNA “clock” that we discussed earlier. Second, the general change in body shape corresponds to what we predicted in our discussion of whale bony features above. That is, there is a gradual elongation and streamlining, there is a modification of the forelimb into flippers and progressive reduction of the hindlimb, the nostrils for breathing move toward the top of the skull to form a blowhole, and the vertebrae develop enlarged projections to support the attachment of swimming muscles.
There is probably little question that the fossil species Dorodon is well on the way to becoming a modern whale. However, it might be argued by a skeptic that the earlier species (like Rhodocetus, Ambulocetus, or Pakicetus), despite the “cetus” (whale) part of their names, are not so obviously “whale-like” that they deserve to be considered as fossil whale ancestors. However, remember the characteristic whale skull modifications for hearing? It has been shown very clearly that throughout this series of fossil species, the various bony changes necessary to support efficient hearing in water were being acquired in a stepwise fashion. Organisms earlier in the sequence had skeletal characteristics consistent with them being able to hear well in both air (using the “classic” mammalian hearing apparatus), and newly acquired changes to also allow better hearing in water. Later organisms in the sequence become increasingly specialized for hearing in water only.4
What of the earliest fossil shown in this diagram—Pakicetus? Careful examination shows that it has the features we would predict for an early whale ancestor. It has the ankle bone characteristics of the even-toed hoofed mammals (in fact these features are also found in several of the later fossil forms as well, ensuring their continuity). This confirms one of the predictions made by the DNA evidence we discussed earlier. Furthermore, it has some of the modifications of the skull bones necessary for more efficient underwater hearing, which were previously documented only for modern whales and their later (more obvious) ancestors.4 These features are also shared with the “whale cousin” Indohyus.5 Preservation of more of the skeleton of this latter species has allowed detailed analysis indicating characteristics likely shared with whale ancestors. Indohyus was probably a wading animal, which spent much of its time in the water. It appears to have fed mostly on land, so it is suggested that resort to the water was made to escape predators.5
Finally, we need to look back at the last prediction from our previous DNA evidence, namely that modern whales are most closely related to hippos. If evolution is true, we should expect to find fossil forms linking these two modern groups. This has proven to be a tougher nut to crack, mainly because the ancestral whales first appear about 50 million years ago in what is now south Asia, and the hippo family first appears about 15 million years ago, in Africa. The most recent tree diagram, produced by using a combination of skeletal features and DNA data, still supports this family connection, as shown by the following figure (Figure 2).6
The blue lines in the diagram show species in which the skeleton is specially thickened, and the bone structure more dense. This is an adaptation which allows wading animals (like modern hippos and the fossil Indohyus) to be good “bottom-walkers” (it prevents them from floating due to lighter body tissues), and allows fully marine organisms (like modern whales) to have “neutral buoyancy” (so they don’t always tend to pop up to the water surface, like a cork). There has also been progress in clarifying the relationships between fossil ancestors of hippos and those of modern whales. A recent study of hippo evolution, based only on skeletal characteristics, has conclusively shown that the hippo family are descended from an extinct group of fossil Artiodactyls, known to go back more than 40 million years, and whose fossils are from southern Asia. Furthermore, this study produced a phylogenetic tree predicting that this extinct hippo ancestor group also shared a common ancestor with the fossil whales.7 Thus the investigation of hippo origins is independently leading us back toward the origin of whales. However, in this study the statistical support for predicted common ancestor of the ancient hippo group and the ancient whale group is not as strong as scientists would like to consider this “case closed.” What is necessary is more fossils, of the appropriate age in order to complete the story of hippo evolution. We still need that to fill in the details of the predicted relationship of hippos to modern whales.
Thus the “Whales’ Tale” is not yet complete. It is a story of scientific discovery in progress. As we finish, let’s briefly summarize what we have found out. Different types of DNA evidence agree that modern whales are most closely related to the even-toed hoofed mammals, despite the obvious great changes in limb anatomy of the modern whales. This prediction has been amply confirmed by the fossil record. The DNA sequence evidence predicted a time frame during which critical early events in evolution of whale ancestors should occur. This prediction has also been amply confirmed. Finally, DNA evidence predicts that modern whales are most closely related to hippos. There is some fossil evidence supporting a predicted common ancestor, but more data is needed. A final caution to possible sceptics—this state of “unfinished business” is precisely how the scientific process works. There is no “crisis.” There is no indication that evolution is not true. There is simply the ongoing work of mapping out of various lines of evidence. A scientific conclusion is considered well supported if “all roads lead to Rome.” In the case of whale evolution it might be prudent to say that the evidence has not quite converged in Rome yet, but that we are now in the suburbs. That is precisely what makes science interesting and fun. Stay tuned!