How to make a cladogram

A cladogram shows how a group of organisms are related by descent. You build one from shared derived characters, traits that appeared in a common ancestor and were passed on. Here is the whole process, from a blank page to a finished diagram.

What you need before you start

Two things, and nothing else: a list of taxa (the organisms or groups you're comparing) and a list of characters (features that are either present or absent in each one). Everything in a cladogram is derived from those two lists.

A cladogram is a hypothesis about branching order, who split off from whom, and in what sequence. It's not a measure of time, and it's not a measure of how much any lineage has changed. That distinction matters when you interpret the finished diagram.

Step 1, Choose your taxa and an outgroup

Pick the organisms you want to compare. These are your ingroup. Then add one more taxon that you already know branched off before all of them: the outgroup.

The outgroup is what makes the diagram readable. It tells you which state of a character is ancestral (original) and which is derived (new). Without one, you can work out that some taxa cluster together, but you can't tell which end of the diagram is the root.

A good outgroup is closely related to your ingroup, but clearly outside it. If you're comparing fish, amphibians, reptiles, birds and mammals, a lancelet or a lamprey works well. If you're comparing primates, a non-primate mammal is a reasonable choice.

Step 2, Choose characters that are actually informative

A character is any feature you can score consistently: a vertebral column, four limbs, hair, amniotic eggs, flowers, jointed legs. For each taxon you record whether it's present (1) or absent (0).

Not every character helps. A useful character is:

• Shared by more than one taxon, a trait unique to a single taxon (an autapomorphy) tells you nothing about grouping.
• Absent in the outgroup, that's what marks it as derived rather than ancestral.
• Heritable, a trait passed down, not one acquired during an organism's lifetime.

Be careful with traits that evolved more than once. Wings in birds and wings in bats look similar but arose independently, that's convergent evolution, and treating them as one shared character will group unrelated taxa together. Traits that are similar because of shared ancestry are called homologous; those similar for other reasons are analogous.

Step 3, Build a character matrix

Put taxa in rows and characters in columns, and fill each cell with 1 (present) or 0 (absent). This table is the real input to a cladogram, everything else is bookkeeping.

Notice the staircase pattern. Each character is present in one more taxon than the last, which is exactly what a nested set of groups looks like when written as a table.

Step 4, Group taxa by shared derived characters

Count how many taxa share each derived character. The character shared by the most taxa marks the earliest split; the one shared by the fewest marks the most recent.

Reading the matrix above:

• Vertebral column, 4 taxa. The largest group, so it branches first.
• Four limbs, 3 taxa, nested inside the previous group.
• Amniotic egg, 2 taxa, nested again.
• Hair, 1 taxon. It marks mammals but doesn't group them with anything.

Each group you find is a clade: an ancestor plus all of its descendants. Properly nested clades are the whole point of the diagram.

Step 5, Draw the branching diagram

Start with the outgroup on one side and work inward. Every derived character becomes a branch point, and every branch point is a node, the common ancestor of everything beyond it.

1. Draw a line for the outgroup splitting off first.
2. At the next node, split off the taxon that shares only the most widespread character.
3. Repeat, working toward the taxon that shares the most characters.
4. Mark each character on the branch where it first appears.

For the matrix above you get a ladder: lancelet, then fish, then amphibian, then reptile, then mammal, with vertebral column, four limbs, amniotic egg and hair marked in order along the spine of the diagram.

Step 6, Check it with parsimony

With real data your characters will conflict, one suggests grouping A with B, another suggests A with C. The standard tie-breaker is maximum parsimony: prefer the tree that requires the fewest evolutionary changes overall.

Count the changes each competing arrangement needs. If one requires a trait to evolve twice independently and another explains it with a single origin, the second is more parsimonious. It's not a guarantee of truth, evolution isn't obliged to be economical, but it's the convention, and it gives you a defensible answer.

Common mistakes

• No outgroup. Without one the diagram has no root and no direction.
• Using shared ancestral traits to group. Fish and amphibians both lack hair, but "no hair" doesn't make them a clade, it's the ancestral state.
• Reading tip order as relatedness. Two taxa printed next to each other aren't necessarily close relatives; only the nodes tell you that.
• Treating branch length as time. In a cladogram, branch lengths are drawn for legibility and carry no meaning.
• Grouping on convergent traits. Wings, streamlined bodies and eyes have evolved repeatedly.

Frequently asked questions