Group species by derived traits the outgroup lacks. Read relatedness from nodes, not tip order.
A phylogenetic tree is a hypothesis about which lineages share the most recent common ancestor. Read it from the branch points, not from which tips sit side by side. Build it from shared derived traits.
Key terms
- Node — a branch point. It represents the most recent common ancestor of all the species that descend from it.
- Clade — an ancestor plus all of its descendants. Snip one branch; everything that falls off is a clade.
- Outgroup — a related species outside your group that branched off earlier. It shows which trait states are ancestral.
- Shared derived character — a trait that arose in a group's common ancestor, so members have it and the outgroup lacks it.
- Horizontal gene transfer — a gene moving sideways between organisms, not from parent to offspring. Bacteria do it often.
Build and read a tree
- Start with the outgroup. Treat its state for each trait, present or absent, as ancestral. The other state is derived.
- Count how many species share each derived trait. In a table where traits nest inside each other, the most widely shared one usually arose earliest, so it goes nearest the root.
- Mark each derived trait on the branch leading to the species that have it. Every species beyond the mark inherits it.
- Read relatedness from nodes. Two species are more closely related when their most recent common ancestor lived more recently.
- Treat the tree as a hypothesis. More data, especially DNA sequences, can move branches or resolve unclear nodes.
Why it matters
Trees are the picture of common descent, so they pull the whole unit together. Expect to build one from a trait table or read relatedness off one. For DNA or protein data, match the molecule to the timescale. Slow-changing genes, like ribosomal RNA genes, resolve ancient splits. Fast-changing ones separate close relatives, whose slow genes may differ at only a site or two.
Exam traps
| Trap | Actually |
|---|
| Tips drawn side by side are the closest relatives. | Tip order carries no information; nodes can spin like a mobile. Compare how recent the shared nodes are. |
| Any trait two species share can group them. | Only a derived one. A state the outgroup shares, having the trait or lacking it, is ancestral and can't define a clade. |
| Three lineages meeting at one node split at the same moment. | That's a polytomy. It usually means the data can't yet tell which two share the most recent ancestor. |
| Two gene trees disagree, so one must be wrong. | Bacteria swap genes by horizontal transfer. A transferred gene's tree tracks the gene's travels, not the species' ancestry. |
Worked example
Biologists compare the same 400-base gene in four salamander species, W, X, Y, and Z, plus a frog as the outgroup. W and X differ at 4 sites; Y and Z differ at 15. Every other salamander pair differs at 30 to 33 sites. The frog differs from each salamander at 60 to 64. Which salamanders form clades, and which pair shares the most recent common ancestor?
- The frog differs most from all four, so its branch splits off first. It roots the tree.
- Among the salamanders, two counts are small: W–X (4) and Y–Z (15). Every cross-pair is 30 or more, so W with X is one clade and Y with Z is another.
- Differences pile up with time since a split. W and X, at only 4, share the most recent ancestor. Y and Z split earlier, and the two pairs earlier still.
W with X is a clade, Y with Z is a clade, and all four salamanders form a clade. W and X share the most recent common ancestor.
Fewer sequence differences means less time for mutations to build up since two lineages parted, so the youngest node joins the least-different pair. The near-miss is Y and Z, because they also pair up. They are a clade, but 15 differences put their shared ancestor further back. Where the letters sit on a drawing changes none of this.