Tracking the Course of Evolution


by Sandra J. Carlson

NOTE: This is page 3 of a five-page document.

If we agree that life originated only once, then the evolutionary relationships among all living things can, in principle, be expressed on a single, huge clade. Then we must decide on a method or a variety of methods for reconstructing the pattern of phylogenetic relationships among organisms. To attempt to reconstruct, in any detail, the pattern of relationships among all living beings is an impossibly large task (although the beginnings of such a massive undertaking can be viewed at the University of Arizona "Tree of Life" website: Fortunately, because of the property that clades nest within other clades (Figure 4), we can focus on one small clade nested within the enormous clade of "all life."
Briefly, phylogenetic analysis involves: (1) choosing the taxa whose evolutionary relationships are of interest, (2) selecting the characters to be included and compared, (3) deciding on a criterion (or point of reference) for determining the direction of character change in evolution, and then (4) choosing a method for selecting among the many cladograms (hypotheses of relationship) that can result from a computer-generated phylogenetic analysis.

1. Choosing the Taxa
What organisms are we interested in studying the ancestry of? Dinosaurs! I have constructed an example that includes several familiar kinds of vertebrates, namely frogs, humans, whales, crocodiles, and birds (Table 1). I hope to accomplish three tasks in this example: determine the phylogenetic relationships among these five groups, decide how dinosaurs are related to them, and then represent their phylogenetic relationships in a cladogram. This group of six taxa is referred to as the ingroup of the analysis.

2. Selecting Characters
We can make observations about the five living taxa in the ingroup, and then make a list of attributes or features of each (Table 1). These features are often anatomical or morphological, but they may also be genetic, developmental, physiological, or behavioral. Any attributes that can be inherited from a common ancestor and are observed to vary among the ingroup taxa are acceptable. This list will most likely include both homologous and homoplastic characters. Ideally, only the homologous characters would be used to reconstruct relationships because clades are defined by shared (homologous) and derived characters. If there were a foolproof way to determine which characters were homoplastic before doing an analysis, they could be eliminated right away. However, it is not always easy to distinguish the two types of characters before a preliminary analysis is completed. The homoplastic characters are almost always "exposed" as such after completing an analysis and comparing the distribution of each character to all the other characters. In this example, characters of both skeletal (hard) and non-skeletal (soft) anatomy are included, as well as a character involving a biological process or mechanism, namely physiology.

3. Deciding a Point of Reference
We need to compare features possessed by each member of the ingroup with some point of reference outside the group of interest. This reference point will be used to "root" the cladogram. As mentioned earlier, we could use as a point of reference the stratigraphic record of these taxa or their patterns of development. I have chosen a third method, the outgroup method, which is widely used by biologists today. With this method, a group presumed to be closely related to the entire ingroup serves as a reference for determining the relative ancestral and derived state of characters among the ingroup. All five ingroup taxa are tetrapods (with two arms and two legs), so I chose non-tetrapod animals as the outgroup, or point of reference. Sharks and tuna are both members of the outgroup in this analysis; I chose them because they are vertebrates, and share with the tetrapods many characters of skeletal structure and embryological development.

4. Selecting the Method
The method used most commonly by systematists today is that of parsimony — given the data currently in hand, the simplest explanation to account for those data is to be preferred over a more complex explanation. A complex explanation that requires many ad hoc assumptions carries a higher burden beyond that suggested by the immediate data alone. For example, let's say that you find the door to your house unlocked after returning from work one day. Nothing else about your house is out of the ordinary. The door might be unlocked because: burglars carefully picked the lock to break into your house but were scared off before they could steal anything; aliens slipped down the chimney and unlocked the door from the inside; or you forgot to lock the door when you left the house in the morning. The last explanation is the most parsimonious because it requires the fewest additional assumptions, although some people would argue that either of the other two explanations is at least possible. In systematics, "simplest" is usually interpreted as the minimum number of character state changes required to account simultaneously for the distribution of all character states among all taxa. For example, it is simplest to suggest that the common ancestor of birds evolved feathers once and that all descendants from the common ancestor retained that feature, rather than suggesting that each separate clade of birds evolved feathers independently.

5. Phylogenetic Analysis
The information on character variability compiled in Table 1 can be used to reconstruct relationships among the ingroup. Which groups of taxa share which character states? Let's begin by considering each character, in order, and comparing only the five living groups to one another. We will construct the simplest (shortest) cladogram using the characters we have listed, and that cladogram will suggest a hypothesis of relationships among the vertebrates in the ingroup. After a hypothesis of relationship has been diagrammed, we can ask where the dinosaurs would be located on it, and examine the evidence that would help us make that decision.

Character 1. Vertebral column. As their name reveals, vertebrates all have a vertebral column as the central component of their skeleton. All members of the ingroup are vertebrates, and the two outgroup taxa chosen are also from the vertebrate clade. The fact that all the ingroup and outgroup taxa share a vertebral column indicates that they share more recent common ancestry with each other than any one taxon does with a non-vertebrate animal. Thus, sharks and tuna are more appropriate outgroup taxa for a group that includes frogs, humans, whales, crocodiles, and birds than are, for example, clams or insects, which both lack a vertebral column and are more distantly related to vertebrates.

Character 2. Bony internal skeleton. Tuna share with the ingroup taxa an internal mineralized skeleton composed of calcium phosphate (bone), while sharks have an internal skeleton composed entirely of non-mineralized cartilage, not bone. Because bone develops in the same way, has the same chemical composition, and is located in the same region of the body in all bony vertebrates, bone is considered to be a homologous character. The distribution of bone among the ingroup and outgroup taxa indicates that the five ingroup taxa share more recent common ancestry with the tuna than with the shark.

Character 3. Four limbs. Frogs, humans, crocodiles, and birds have two forelimbs and two hindlimbs; they are all tetrapods. Birds have wings instead of "arms" as forelimbs, but they are still tetrapods. Whales are somewhat confusing, because they have become fully aquatic secondarily, a habitat that their fully terrestrial ancestors successfully invaded and adapted to millions of years ago. Their forelimbs have been modified to flippers, and hindlimbs have been lost entirely (at least in living whales). Since we can't observe two hindlimbs in living whales, I have indicated with a question mark the state of this character in Table 1. Before doing an analysis, we can't be sure if they lost their hindlimbs secondarily, or never had hindlimbs at all. Sharks and tuna both have fins that are fundamentally different from limbs, even those present in whales, revealing fins as the ancestral state of this character. Having four limbs is the derived state of Character 3, and the four tetrapod taxa that share the derived state are thus united in a clade.

Character 4. Lower temporal fenestra. The number and position of openings in the skull (fenestra) within and between bones are diagnostic characters, particularly among the tetrapods. From a side view of the skull, mammals (humans and whales) possess a lower temporal fenestra — a single opening below and behind the eye socket (Figure 6). Even extinct relatives of mammals, including Dimetrodon, possess this skull opening. Crocodiles and birds possess an upper temporal fenestra, in addition to a lower temporal fenestra (see below). Sharks, tuna, and frogs have skulls, but do not possess openings in the same positions on the skull that other tetrapods do; they lack a lower temporal fenestra, which is the ancestral state of this character. Characters 5 and 6. Upper temporal fenestra and antorbital fenestra. As the names imply, the upper temporal fenestra is located above the lower temporal fenestra (Figure 6), and the antorbital fenestra is located in front of the eye socket (orbit). Of all the taxa in this analysis, only crocodiles and birds possess these two skull openings (in addition to the lower temporal fenestra discussed above). These additional two skull openings are shared derived features that unite crocodiles and birds in a clade nested within the tetrapods.

Character 7. Amniotic egg. Of our five ingroup taxa, only humans, whales, crocodiles and birds have a protective membrane, called an amnion, surrounding their eggs. The amnion allows the egg to survive outside of an aquatic environment. In both humans and whales, the egg develops inside the mother's body, but still possesses an amnion. Frog (and shark and tuna) eggs are different from human, whale, crocodile, and bird eggs; they lack this membrane, which is the ancestral state of the character, and must lay their eggs in water if they are to survive.

Character 8. Mammary glands. All living mammals, including humans and whales, possess mammary glands to nurse their young. No other living vertebrates share this derived character, which is clearly related to the mammalian mode of reproduction.
Using the distribution of shared states of Characters 1-8, with sharks and tuna as our point of reference, we have been able to construct a hypothesis of the relationships among the five ingroup taxa (Figure 7). Any other characters present and variable among the ingroup can be, and should be, used to test this hypothesis of relationships. Try to think of several other characters that support this cladogram, and others that appear to conflict with it. Can you determine if each character is shared and derived, or shared and ancestral, or homoplastic and not shared through common ancestry at all?


Character 9. Endothermy. Endothermy is the ability to regulate body temperature internally, rather than rely on the external environment to warm up or cool down body temperature passively (a condition referred to as ectothermy). Humans, whales, and birds are all endothermic and their body temperature doesn't vary by more than a degree or two throughout the day. Sharks, tuna, and frogs are all ectothermic. This is the ancestral state of this character, as determined by its pattern of distribution relative to Characters 1 through 4. Because crocodiles also share the ectothermic condition, we infer that they retain it from their ancestors. Therefore, we infer that endothermy evolved two times independently, once in mammals (humans and whales) and once in birds. In this analysis, endothermy is a homoplastic character, and by itself might falsely lead us to conclude that mammals and birds are more closely related to each other than either is to crocodiles. Based on the distribution of a larger suite of characters that are in conflict with this interpretation (including Characters 5 and 6), we can detect endothermy as a homoplastic rather than homologous character, at this level of analysis.
Now let's try to place dinosaurs on this cladogram. We know that dinosaurs are egg-laying tetrapod vertebrates, so they belong somewhere in the tetrapod clade. Are they more closely related to mammals, crocodiles, or birds? Representatives of each of the two major groups of dinosaurs, the saurischians (like Apatosaurus and Tyrannosaurus) and the ornithischians (like Triceratops and Stegosaurus) are included in this example.

Characters 10 and 11. Reduced fourth and fifth digits on hands; fully upright posture. A shared derived character uniting tetrapods (in addition to two arms and two legs), is the presence of five fingers and five toes. In frogs, humans, and crocodiles, all five fingers are of more or less similar length. In birds and both saurischian and ornithischian dinosaurs, the fourth and fifth digits (ring finger and pinky finger) are reduced considerably relative to the other fingers. Birds and both groups of dinosaurs also share a fully upright posture, with the hindlimbs approximately parallel to the vertebral column. Even though they are extinct, we can infer posture in terrestrial dinosaurs from footprints and from the shapes of the joints between limb bones, particularly the hip, knee, and ankle joints. Crocodiles have a semi-erect posture, and frogs have a squat, sprawling posture, with their limbs emerging at an angle from the vertebral column. Humans and other mammals (with hind limbs, so not including whales) have evolved an upright posture independently from that in birds and dinosaurs, similar to the evolution of endothermy in both mammals and birds.

Characters 12 and 13. Long S-shaped neck; long hands. Birds and the saurischian dinosaurs both have relatively long legs, arms, and hands, and long S-shaped necks, while crocodiles have much shorter legs, arms, hands, and necks relative to body length. This is considered to be the ancestral condition, since crocodiles lie outside the clade containing birds and dinosaurs, based on the distribution of Characters 10 and 11 (and many others).
These four additional characters give strong support to the hypothesis that dinosaurs are more closely related to birds than to any other taxon in the ingroup (Figure 8). Birds are nested within the clade that contains the common ancestor of all dinosaurs, and are more closely related to saurischians than ornithischians. For this reason, we now consider birds to be dinosaurs, albeit highly derived dinosaurs whose skeletons are modified for flight.

As is true for each of the nodes in these cladograms (Figures 4, 7, 8), numerous other characters can be listed to provide support, but are not discussed for lack of space. However, since the saurischian and ornithischian dinosaurs are known most commonly as the "lizard-hipped" and "bird-hipped" dinosaurs, it is worth mentioning the distribution of pelvic structure on this cladogram (Figure 8), because it appears to be in conflict.

Character 14. "Bird-hipped" pelvis. Saurischian dinosaurs share the ancestral "lizard hipped" condition with crocodiles (and, of course, lizards). Birds and ornithischian dinosaurs have each evolved a "bird-hipped" pelvic structure independently from this ancestral "lizard-hipped" condition, a conclusion that is supported by many other additional skeletal features. The character is thus homoplastic, in this example.

Taking all the information on character distribution in the ingroup and comparing it to the states present in the outgroup, we can construct the cladograms illustrated in Figures 7 and 8. The shared derived characters that support each node (and, consequently, each clade nested within each more inclusive clade) are listed beside them. On the basis of this phylogenetic hypothesis, we can make the evolutionary statement that birds and other dinosaurs share more recent common ancestry than either group does with crocodiles, or with mammals, or sharks. This is a powerful evolutionary statement, and a very useful one, if we agree that our goal in systematics is to discover the pattern of evolution. Simply stating that birds are in the Class Aves, dinosaurs in the Orders Ornithischia and Saurischia, and crocodiles in the Order Crocodylia emphasizes each group as a unit separate from the others, rather than as elements in a clade that are related by common ancestry.

BACK to page 2
CONTINUE to page 4

Return to top of page   |   TRACKING HOME   |   PRESENTATIONS