What Is an Outgroup on a Cladogram?

An outgroup on a cladogram is a species or group of organisms that sits outside the cluster of species you are actually studying. It serves as a reference point, something known to have branched off before the organisms of interest shared their last common ancestor. By comparing the traits or DNA sequences of the organisms being studied (called the ingroup) to this external reference, researchers can figure out which characteristics are ancestral and which evolved more recently. The outgroup is what gives a cladogram its direction, turning a web of relationships into a tree with a clear starting point.

Why a Cladogram Needs a Starting Point

A cladogram is a branching diagram that shows how organisms are related by common ancestry. Without an outgroup, a cladogram is “unrooted,” meaning it shows which species are more closely related to each other but gives no indication of where the whole group began. Think of it like a family photo that shows siblings standing together but does not tell you which direction the parents are. The outgroup provides that orientation.

When an outgroup is added to an analysis, the root of the tree is placed along the branch connecting the outgroup to everything else. The most recent common ancestor of the ingroup and the outgroup becomes the root, representing the oldest divergence on the tree.¹ Once you know where the root is, you can read the cladogram from past to present, tracing the sequence in which new traits appeared along different lineages. Without that anchor, you cannot distinguish an old trait shared by many species from a newer one that evolved in just a few.

How Researchers Choose an Outgroup

The ideal outgroup is a close relative that diverged from the ingroup just before the ingroup’s own common ancestor existed. If you are building a cladogram of cats, dogs and bears, you might use a seal as the outgroup, because seals are related to the group but branched off early enough to serve as a useful comparison. The key is balance: the outgroup needs to be far enough outside the group to genuinely be an outsider, but not so far away that meaningful comparisons become impossible.

Choosing well matters because the outgroup directly determines which traits are treated as “original” versus “derived.” A trait shared by both the outgroup and some ingroup members is interpreted as ancestral, something inherited from a deeper common ancestor. A trait found only within part of the ingroup is interpreted as a newer evolutionary innovation. If the outgroup is poorly chosen, those assignments can flip, and the entire tree topology can shift. Research on large genomic datasets in seed plants showed that limited outgroup selection can lead to incorrect phylogenies even when the dataset itself contains hundreds of genes.²

Another consideration is that adding an outgroup to a dataset does not just root the tree; it can actually change the relationships among the ingroup members themselves. Simulations and empirical analyses have shown that including even a single new taxon can substantially rearrange the resulting tree topology for species already in the analysis.¹ This means outgroup selection is not a minor bookkeeping step tacked on at the end. It is a decision woven into the analysis from the start.

What Goes Wrong When the Outgroup Is Too Distant

One of the most common pitfalls in building a cladogram is using an outgroup that diverged from the ingroup so long ago that the two groups barely share informative traits. When sequences have been evolving independently for hundreds of millions of years, the same nucleotide positions can mutate to the same state by sheer coincidence rather than by shared ancestry. This creates a well-known artifact called long-branch attraction, where two distantly related lineages with long evolutionary branches get incorrectly grouped together on the tree simply because they have accumulated similar-looking random changes.

Long-branch attraction is especially misleading when the dataset has sparse taxon sampling or the outgroup is very distant from the ingroup.³ The artifact is not just a theoretical curiosity; it has led to high-profile mistakes in animal and plant phylogenetics that took years to untangle. One way researchers reduce the problem is by using more sophisticated statistical models of sequence evolution. Another is by choosing outgroups that sit closer to the ingroup, shortening the offending branch and reducing the window for coincidental similarities to pile up.

When No Good Outgroup Exists

For some groups of organisms, picking an outgroup is genuinely difficult. Certain bacterial pathogens, for instance, have diverged so sharply from their nearest relatives that few genes are still similar enough to align confidently. When an outgroup does not exist, is unknown, or is so distant that almost no informative characters are shared, the standard approach of rooting with an external species breaks down.⁴

In these cases, researchers sometimes turn to duplicated genes within the ingroup itself. If a gene duplicated before the last common ancestor of the ingroup, the two resulting copies can serve as each other’s outgroup. One copy acts as the reference to root the tree of the other copy. This is a clever workaround, though it introduces its own complications, including the need to correctly identify which gene copies descended from which duplication event.⁴

Alternative rooting methods that do not use an outgroup at all have also been developed. Midpoint rooting places the root at the midpoint of the longest path on the tree, and minimal ancestor deviation (MAD) rooting tries to find the root that minimizes variation in evolutionary rates across the tree. Both methods rely on the assumption that evolution proceeds at roughly similar rates across all branches, and when that assumption is violated, the results suffer. A study testing both methods under challenging conditions found that each had average success rates below 50 percent, though MAD slightly outperformed midpoint rooting.⁵ These outgroup-free approaches are useful fallbacks, but the success rates make clear why external outgroups remain the preferred strategy whenever a suitable one is available.

Using More Than One Outgroup

A single outgroup can be misleading if it happens to have an unusual evolutionary history, like an accelerated rate of mutation or a reversal of some trait to an ancestral-looking state. For this reason, many analyses now include multiple outgroups. The logic is straightforward: if several independent external lineages all agree on which ingroup traits are ancestral and which are derived, the resulting root placement is much more robust than if it rested on a single species.

A multiple-outgroup approach has been used, for example, to resolve deep relationships among bacterial divisions. By incorporating several distantly related bacterial lineages rather than relying on just one, researchers were able to test whether proposed bacterial groups were genuinely monophyletic, meaning they included all and only the descendants of a single ancestor.⁶ Including additional outgroups helps buffer against the quirks of any single reference species, though it also increases the complexity of the dataset and can introduce its own alignment and computation challenges.

Outgroup Choice Affects More Than the Root

The outgroup does not just tell you where the tree starts. It ripples through every downstream analysis that depends on the tree, including estimates of when lineages diverged, how fast traits evolved, and what the ancestors of living species looked like. A study on the evolution of feathers in stem-birds found that ancestral state estimates were strongly influenced by outgroup selection, alongside factors like how the tree was calibrated in time and which evolutionary model was used.⁷ In practical terms, changing which species served as the external reference changed the researchers’ conclusions about what kind of skin covering the ancestors of modern birds likely had.

This sensitivity is worth keeping in mind for anyone reading cladograms in textbooks or news articles. Two research groups studying the same set of organisms can arrive at different tree topologies and different evolutionary stories simply because they chose different outgroups. The tree itself is not a photograph of evolutionary history; it is an inference, and the outgroup is one of the most powerful assumptions baked into that inference.

A Real Example That Stumped Botanists for Years

One of the best illustrations of outgroup headaches comes from the search for the root of all living flowering plants. Flowering plants, or angiosperms, are an enormous group with over 300,000 species, and figuring out which living lineage diverged first from the rest is a question that hinges entirely on where the root of the angiosperm cladogram falls. For years, a small shrub from the Pacific island of New Caledonia, Amborella trichopoda, has been strongly supported as the single living species of the sister lineage to all other extant flowering plants.⁸

But that placement was not always uncontested. Earlier analyses sometimes placed water lilies, or a group including both Amborella and water lilies, at the base of the tree instead. Researchers traced part of the discrepancy to problems with paralogy (accidentally comparing the wrong gene copies), suboptimal outgroup selection, long-branch taxa, and sensitivity to the method used.⁹ The outgroup used to root the angiosperm tree was typically a gymnosperm like a pine or a cycad, but because gymnosperms diverged from flowering plants over a hundred million years ago, the long branch connecting them to angiosperms made the analysis vulnerable to exactly the kind of artifacts described earlier. As datasets grew and methods improved, the placement of Amborella at the root stabilized, but the wobble in earlier results was a vivid demonstration of how outgroup distance and quality can swing the answer to a major biological question.

Gene Trees and Species Trees Can Tell Different Stories

Modern phylogenetics often builds trees from hundreds or thousands of genes simultaneously, and not all of those genes tell the same story. Individual gene trees can disagree with the species tree because of events like hybridization, incomplete sorting of ancestral genetic variation, or horizontal gene transfer. When researchers root a species tree using an outgroup, the outgroup is being compared against a consensus of these gene trees, and high levels of gene tree discordance in the ingroup can make the outgroup’s job harder.

A genomic study of the plant family Amaranthaceae found strong concordance among gene trees for the monophyly of the family as a whole, with over 99 percent of informative gene trees agreeing, but much weaker agreement deeper in the tree, where one key group was supported by only about a third of informative gene trees.¹⁰ In situations like these, the outgroup can confidently root the broad family tree but cannot fully resolve the tangled relationships among closely related subgroups that experienced rapid diversification or ancient hybridization. The outgroup is doing its best, but it cannot fix ambiguity that originates inside the ingroup itself.

How to Read the Outgroup on a Published Cladogram

If you are looking at a cladogram in a textbook, a museum display, or a research paper, the outgroup is almost always the lineage that branches off first at the very base of the tree. It sits on its own branch on one side while the rest of the organisms cluster together on the other. Sometimes the outgroup is labeled explicitly; other times you need to recognize it by its position. The branch connecting the outgroup to the rest of the tree represents the root, the deepest split in the diagram.

A few things are worth remembering when you encounter one. The outgroup is not necessarily the “most primitive” organism on the tree. It is simply the one that diverged from the ingroup earliest. It has had just as much time to evolve since that split as the ingroup species have. Calling the outgroup primitive is a common misreading that conflates being an early branch with being unchanged, and those are very different things. A crocodile used as an outgroup for a cladogram of birds is not a living fossil frozen in time; crocodilians have their own long evolutionary history since their lineage split from the one leading to birds and other dinosaurs.

It is also important to realize that outgroup status is not a permanent property of a species. The same organism can be an outgroup in one analysis and an ingroup member in another, depending on the question. A lizard might be the outgroup in a study of snake relationships, but an ingroup member in a broader study of all reptiles, with a frog serving as the outgroup instead. The role is entirely relative to the question being asked.

Practical Lessons from Outgroup Disputes

The debates over outgroups in the scientific literature carry a few lessons for anyone trying to interpret evolutionary trees. One is that the root of a tree is often the least stable part of the analysis. Tips of the tree, where closely related species sit, tend to be well supported by abundant shared data. The root, where the outgroup anchors the whole structure, is the place where the data have to stretch the furthest, and confidence often drops accordingly. If two published trees for the same group of organisms differ mainly in where the root falls, outgroup choice is a likely culprit.

Another lesson is that “more data” does not automatically solve outgroup problems. Throwing thousands of genes at a phylogenetic question helps in many ways, but if the outgroup is too distant, more data can actually amplify systematic errors rather than dilute them. The seed-plant study mentioned earlier found that inadequate taxon sampling and limited outgroup selection led to spurious results even at phylogenomic scale.² Good outgroup choice and broad taxon sampling remain as important in the age of genomics as they were when researchers were comparing a handful of morphological characters by hand.

Finally, when you see a cladogram presented as a settled fact, it is worth asking whether the outgroup was stated and whether it was a reasonable choice. In many popular depictions of evolutionary trees, the outgroup is silently assumed rather than discussed, which can give the false impression that the tree’s orientation is beyond question. In reality, the root is often the part of the tree most open to revision as new species are sampled and new analytical methods are developed.