Biology › Classification, biodiversity and conservation › Classification and phylogeny: sorting by ancestry, not by looks

Classification and phylogeny: sorting by ancestry, not by looks

Sorting organisms by how similar they look is easy and sometimes wrong: a dolphin is not a fish and a bat is not a bird. Modern classification tries to group organisms by shared ancestry instead, and the evidence that settles the arguments comes out of a sequencing machine rather than off a dissecting board.

Before this Natural selection and adaptation · DNA base sequence, the genetic code and protein synthesis · Antigens and the antibody response

What you should be able to do

• Write the eight taxonomic ranks in order and name a group at each rank for a named organism.
• Write a binomial correctly and say what each part of it means.
• State the biological species concept and give two situations it cannot handle.
• Distinguish a phenetic from a phylogenetic classification, and explain why convergent evolution defeats the first.
• Describe the evidence from ribosomal RNA that split the prokaryotes into two domains.
• Explain how DNA sequences, amino acid sequences and immunological comparisons are used to estimate how closely two species are related.
• Read a cladogram: say what it asserts, and say what it does not.

Names, and why they have to be dull

A robin in Surrey is a small brown bird that will follow a spade around a garden. A robin in Ohio is a thrush the size of a blackbird. A buzzard in Britain is a hawk; a buzzard in Arizona is a vulture. Common names are local, they change, and two of them can collide, so biology uses a name agreed internationally that belongs to one species only.

The system is binomial nomenclature, and it is old — Linnaeus set it out in the eighteenth century, long before anybody knew why organisms fall into groups at all. Every species gets two words. The first is the genus, and it is shared with the species' closest relatives; the second is the specific name, and it is unique within that genus. Homo sapiens. Panthera leo. Quercus robur. The genus takes a capital letter and the specific name never does, both are printed in italics, and when you are writing by hand you underline them instead. After the first mention you may abbreviate the genus to its initial: H. sapiens.

Those conventions look like fussiness and they are worth a mark. Writing homo Sapiens in an exam tells the examiner you have learned the name and not the system.

A species does not sit on its own. It is placed in a genus, the genus in a family, and so on up through eight ranks: domain, kingdom, phylum, class, order, family, genus, species. Botanists usually say division where zoologists say phylum, which is worth knowing so that a plant question does not look unfamiliar.

Two features of the hierarchy do real work. It is nested: every mammal is a chordate, every chordate is an animal, and there is no organism that belongs half to one class and half to another. And it has no overlap at a rank: nothing is in two orders. That is not an administrative convenience. It is what you get if groups are built by descent, because a lineage that split cannot rejoin.

Taxonomy

The naming and classifying of organisms into a hierarchy of groups.

Taxon

Any one group in that hierarchy — Mammalia and Primates and Homo are all taxa.

Binomial

The two-part international name of a species: genus then specific name, italicised, genus capitalised.

Phylogeny

The evolutionary history of a group: which lineages split from which, and in what order.

What counts as one species

The rank at the bottom of the hierarchy is the only one that people have tried to define from first principles, and it is still argued over. The definition every board wants first is the biological species concept: a group of organisms with similar characteristics whose members can breed with each other to produce fertile offspring.

The word fertile is doing the work. A horse and a donkey will breed and produce a mule, and the mule is almost always sterile, so horses and donkeys stay two species. Members of one species also share the same ecological niche and, in animals, usually a specific courtship behaviour — a display, a call, a scent, a dance — that lets a male and a female recognise a member of their own species, of the opposite sex, and in breeding condition. Courtship is how many species keep themselves separate in the first place, and AQA asks about it directly.

Now the awkward part, which good answers acknowledge. The biological species concept cannot be applied to anything that does not reproduce sexually, which rules out every bacterium and archaean and a great many protoctists and plants; prokaryote species are defined instead by how similar their genomes are, using an agreed threshold. It cannot be applied to fossils, because nobody can breed them. It struggles where populations form a chain around a barrier and neighbours interbreed but the two ends do not. And it is embarrassed by the number of species that hybridise successfully in the wild — polar and brown bears do, and so do a large fraction of flowering plants, where whole new species have arisen from hybrids doubling their chromosomes.

None of this means the concept is useless. It means a mark scheme that asks for a limitation is asking for one of the situations above, and that classification at the species level is an argument rather than a measurement.

Two ways to sort, and only one of them survives a dolphin

Suppose you know nothing about ancestry and simply group organisms by how much they resemble one another — count the features they share, put the most similar together. That is a phenetic classification, and it is what Linnaeus was doing. It has the great advantage of needing no theory, and one fatal weakness.

A dolphin is streamlined, has a dorsal fin, flippers, and lives its whole life in the sea, exactly like a shark. A bat has wings and flies, exactly like a swift. A cactus and a euphorbia are both fat, spiny, green-stemmed desert plants that look almost identical, and they belong to families that are not closely related. In each case the resemblance is convergent evolution: unrelated lineages under the same selective pressure arriving at the same solution. Group by resemblance and you put the dolphin with the shark.

Phenetic classification

Grouping organisms by observable similarity, without regard to how they are related.

Phylogenetic classification

Grouping organisms so that each group contains a common ancestor and all of its descendants.

Homologous structures

Structures with the same underlying plan, inherited from a common ancestor, whatever they are now used for.

Analogous structures

Structures with the same function and different origins, produced by convergent evolution.

The way out is to ask not how similar but similar how. A dolphin's flipper opens up to show one upper arm bone, two forearm bones, a set of wrist bones and five digits — the pentadactyl limb, the same plan as a bat's wing, a horse's leg and your hand. A shark's fin has nothing of the kind. The flipper and the wing are homologous; the flipper and the shark fin are analogous. Homology is evidence of ancestry. Analogy is evidence of a shared problem.

A phylogenetic classification uses homology and throws analogy out. Each group it recognises is a piece of the evolutionary tree cut off whole: a common ancestor together with every descendant of that ancestor, and nothing else. That is why 'reptiles' is an uncomfortable group — the ancestor of crocodiles and lizards also gave rise to birds, so a group that holds the reptiles and excludes the birds has cut the tree across a branch rather than at one.

Three domains, found in a molecule

For most of the twentieth century the top-level split was between prokaryotes and eukaryotes: cells without a nucleus, and cells with one. The five-kingdom scheme that grew out of Whittaker's 1969 proposal — Prokaryotae (his Monera), Protoctista, Fungi, Plantae, Animalia — put every organism without a nucleus into a single kingdom, on the reasonable ground that they all looked alike under a microscope.

Carl Woese did not look at them under a microscope. From the 1970s he compared the base sequence of the RNA in the small subunit of the ribosome — 16S rRNA in prokaryotes, 18S in eukaryotes — across as many organisms as he could get hold of. What came back split the prokaryotes into two groups whose rRNA sequences were about as different from each other as either was from ours. In 1990 Woese, Kandler and Wheelis proposed a rank above kingdom, the domain, and three of them: Bacteria, Archaea and Eukarya.

Ribosomal RNA was not chosen at random, and a question about why it was used has a specific answer. Every organism that has ever been sequenced has ribosomes, so the same molecule can be compared across all of life. Its job — holding the ribosome together and positioning the tRNAs — is the same everywhere, so equivalent positions can be lined up against each other. Most of its sequence changes very slowly, because almost any change breaks a working ribosome, so it still carries a readable signal across billions of years. And it is long enough that the similarities are not coincidence.

The cell biology then turned out to support the split. Archaea are not simply odd bacteria; on several counts they resemble us more than they resemble Bacteria.

Which scheme a board asks for differs, so know both. AQA teaches the three domains directly. OCR A teaches the five kingdoms and then the three-domain revision, and can ask you to compare them. CAIE 9700 teaches three domains with the eukaryotic kingdoms beneath them. All three want the same underlying point: a classification is a hypothesis about ancestry, and new evidence can overturn it.

Viruses sit outside the whole arrangement. They are not cells, they have no ribosomes and therefore no ribosomal RNA to compare, and they carry out no metabolism of their own, so the molecule that built the tree cannot place them on it. They are classified by their own scheme, based on the kind of nucleic acid they carry and how they replicate it.

Reading relatedness out of molecules

Anatomy runs out quickly: two bacteria look the same, and homology can be argued about. Molecules give a sequence of discrete units that can be counted, and counting is what turns an argument into a measurement.

DNA base sequence. The direct comparison. Line up the same gene from two species and count the positions where the bases differ. Fewer differences means less time has passed since the two lineages separated, because mutations accumulate steadily once two populations stop exchanging genes. Whole-genome comparison is now routine and has rearranged several parts of the tree.

DNA hybridisation. The older technique, still examined. Heat DNA from two species so the strands separate, mix them, and let hybrid double strands form. The more similar the sequences, the more hydrogen bonds hold the hybrid together, and the higher the temperature needed to separate it again.

Amino acid sequence. Compare the same protein — cytochrome c and the haemoglobin chains are the standard examples — and count the positions that differ. This is a slightly blunter tool than the DNA, because the genetic code is degenerate: a base can change without changing the amino acid, so some differences are invisible at the protein level.

Immunological comparison. Inject a rabbit with human blood serum. The rabbit makes antibodies against the human proteins, and the serum containing those antibodies is collected. Now add that antiserum to serum from another species: the antibodies bind wherever the other species' proteins are similar enough to human ones, and the amount of precipitate formed measures how similar they are. More precipitate means a closer relative. The numbers below are idealised to show the pattern rather than taken from a particular study.

Molecular clocks. If substitutions accumulate at a roughly steady rate in a given molecule, then the number of differences between two species is a measure of the time since they diverged. Calibrate the rate against a split whose date is known from the fossil record, and the same molecule will date splits that left no useful fossils.

Cladograms: what the lines claim

A cladogram is a diagram of branching order. Each line is a lineage, each fork is a point where one ancestral population split into two that stopped interbreeding, and each tip is a taxon. A clade is any fork together with everything above it: an ancestor and all of its descendants.

The tree is built from shared derived characters: features that are new in a group and inherited from that group's common ancestor. Jaws are a shared derived character of everything except the lamprey in that animation; four limbs are a shared derived character of the tetrapods; the amniotic egg, with its membranes that let an egg be laid on dry land, is shared by lizard and mouse. Characters that are merely similar — the streamlining of a dolphin and a shark — are not used, which is exactly the discipline the phenetic approach lacked.

Here is what a cladogram asserts, and it is less than most readers assume. It asserts the order of branching: which lineages separated before which. It asserts common ancestry at every fork. It asserts a set of nested groups.

Here is what it does not assert. It does not say how long each branch is in years unless the diagram is explicitly drawn to a time scale. It does not say how much change happened along a branch — a lineage may have changed enormously or hardly at all. It does not say that any tip is descended from any other tip; all five animals in that animation are alive now, and none of them is the ancestor of the others. And it carries no direction of improvement whatsoever. Swap the two branches at any fork and you have redrawn the same tree, which is worth trying on paper once, because it kills the idea that the bottom of the diagram is the destination.

That last point is where the marks go. 'Humans are more evolved than fish' is not a claim a phylogeny can support; every living lineage has been evolving for precisely the same length of time, since life began. 'Fish are our ancestors' is wrong in the same way as the chimpanzee sentence at the top of this lesson: living fish are cousins, and the shared ancestor was neither a modern fish nor a modern human.

In the exam

• Get the binomial conventions right without being asked: capital for the genus, lower case for the specific name, italics or underlining for both. It costs nothing and questions do award it.
• When asked why classification changed, name the evidence. 'New technology' earns little; 'comparison of ribosomal RNA base sequences showed archaea and bacteria differ as much from each other as from eukaryotes' earns the marks.
• For a limitation of the biological species concept, pick one situation and explain it — asexual reproduction, fossils, or fertile hybrids. Listing all three quickly and explaining none scores as one.
• Molecular clock calculations want the assumptions as well as the number. Constant rate, and one visible difference per substitution, are the two that appear on mark schemes.
• Never write 'more evolved', 'higher' or 'more advanced' about a living organism. Examiners read those as evidence that the tree has been misunderstood, and there is always a precise phrase available instead: 'branched off earlier', 'shares a more recent common ancestor with'.
• A cladogram question that asks what the diagram 'shows' is usually testing restraint. Answer with branching order and common ancestry, and refuse the invitation to talk about time or progress.

Questions

Written to the command words the boards use. Try them on paper before opening a scheme: the marks go to points made, not to length.

← Genomes, screening and gene therapy: what a sequence is good for · Measuring biodiversity: richness, evenness and the index that catches both →