To test phylogenetic hypotheses, scientists must be able to find out which similarities indicate a close relationship between species and which do not. The key to this process is determining the evolutionary origins of the similar features. Only similarities inherited from the species' common ancestor can provide evidence of phylogenetic relationship, because they are evidence of a genetic continuity from the common ancestor. Such a similarity, inherited in common form from a single common ancestor, is called homology.
The importance of homology in determining relationships is easily illustrated. For example, whales and humans share many homologies that sharks do not have. Lungs, warm-bloodedness, lactation, three middle ear bones and a single jaw bone are all features that humans and whales share because of their common history of descent. These features did not all evolve in the most recent common ancestor of humans and whales--in fact, none of them did. Lungs are common to amphibians, reptiles, mammals, and birds, while lactation and the evolution of three former bones of the jaw into three middle ear bones are shared only by mammals. But all these homologies arose more recently than the common ancestor of humans and whales with sharks, and therefore provide evidence of the close relationship of humans and whales relative to sharks.
Different organisms can evolve in similar ways even if they are not similar by descent from a common ancestor. For example, although humans and whales share many features that sharks lack, sharks and whales share many features also. Like whales, sharks have a streamlined body shape, fins, and an aquatic habitat. Some whales have single-pointed teeth like sharks, and some sharks have live births, like whales. None of these features were inherited by both these species from their most recent common ancestor. Instead, such non-homologous similarities are examples of homoplasy.
Which features are homologous?
Testing whether similarities are homologies or not involves many comparisons. In some cases, we can look directly at the genetic basis of the traits. For example, a number of genes contribute to the hemoglobin protein, which carries oxygen in red blood cells. Mammals and other vertebrates have more of these genes than non-vertebrates, because gene duplications occurred in our distant ancestors. After these ancient duplications, several genes acquired more specialized functions, which are shared among diverse groups of animals. The odds that this complex series of events occurred more than once is so low that the hypothesis of homology among species that share these genes is virtually certain.
In other cases, the physiological basis of the trait compellingly favors homology. For example, the mechanisms of live birth in marsupials and placental mammals are different--marsupial offspring emerge from the mother and enter an external pouch very early in life, where they develop and grow, while placentals remain inside the mother for an equivalent length of time. However, the mechanism of lactation is the same. In both these groups, milk is produced by similar structures, the mammary glands, is transmitted in the same way, by suckling, has a very similar composition, and serves the same purpose. The detailed similarities of this physiological process would be unlikely to arise by chance--after all, there are many possible ways of providing nutrition directly to offspring. Therefore, we can argue quite strongly that lactation is a homology.
In some cases, the fossil record is the ultimate judge of homology. If paleontologists can find traces of the common ancestor of two groups, and can show that a feature existed in this ancestor, then the presence of the feature in the descendants is probably by homology. For example, our closest living relatives, the chimpanzees and gorillas, are both knuckle-walkers. Based on the weight of other evidence, chimpanzees are more closely related to humans than to gorillas, and humans are not knuckle-walkers. Knuckle-walking could have evolved in parallel in chimpanzees and gorillas, or it could be a homology--if humans actually descended from a knuckle-walking ancestor as well as chimpanzees and gorillas. Present-day evidence from anatomy does not strongly support either view. But fossil evidence from some of the earliest human relatives shows features in the wrist that may be signs that humans also evolved from a knuckle-walking ancestor (Richmond and Strait 2000). If these features are marks of a knuckle-walking ancestry, these fossils support the hypothesis of homology for knuckle-walking.
Patterns of homoplasy
Homoplasy can occur by convergence or by parallelism. Convergence describes similarities between two species that evolved independently from different features in their common ancestor. For example, the wings of birds and the wings of bats are similar in function, but bat wings involve the bones that in humans make up the hands, while bird wings lack many of these bones entirely, and instead include only the bones that in humans make up the arms. Both structures support flight, but because the two lineages had a long evolutionary separation before they independently became fliers, the wings are different in structure, development, and genetics.
Parallelism occurs when two groups independently develop similarities from the same structures. For example, gorillas, orangutans, and some fossil relatives of humans can have bony crests along the top of their skulls, called sagittal crests. These crests, which provide attachment points for massive jaw muscles, do not occur in the earliest human ancestors, and they may not have occurred in the common ancestor of gorillas and orangutans. However, sagittal crests perform the same function and develop from the same anatomical structures for the same reasons in all these animals. Many mammals with large jaw muscles have sagittal crests, including some bears and pigs. Thus, the feature is a case of parallelism.
Because much of the evolution of species is caused by selection, which can affect gene frequencies in different populations in the same ways, homoplasy has been very common in evolutionary history. Convergence can occur whenever different organisms adapt to the same environment. For example, flight has evolved at least four times in the history of life, in birds, bats, pterosaurs, and insects, each time involving different underlying structures. Even very complex structures, like eyes that can focus light, have evolved many times in different groups of organisms. Parallel evolution is also common, because it is a likely result whenever similar species are subjected to the same series of events or sequence of environmental pressures. Closely related species tend to share many genes, because of their common ancestry, and when exposed to the same selective factors they tend to adapt in the same way. Parallelism especially creates challenges for paleontologists attempting to study relationships, because a feature that evolves in parallel in two closely related species is very difficult to distinguish from a homologous feature inherited from their common ancestor. Some groups of species have undergone substantial parallel evolution, and for this reason it can be extremely difficult to sort out their phylogenetic relationships.
Richmond BG, Strait DS. 2000. Evidence that humans evolved from a knuckle-walking ancestor. Nature 404:382-385.