The problem with "Basal"

Amblyopone australis: a primitive ant?

Earlier I chastised Christian Peeters and Mathieu Molet for misinterpreting the term "basal" in a phylogenetic context.  What was that about?

The issue relates to the classic fallacy of viewing evolution as a linear progression from primitive to advanced. Popular conceptions of evolution aside, the process is not linear like a ladder so much as branching like a bush.

I don't know what quirk of human psychology so strongly predisposes us to frame ideas in linear narratives, but the fact that we do so makes evolution an unfortunately difficult concept to grasp. We imagine a tidy parade of forms from simple to complex, even though evolution in the real world is entirely different. Real evolution runs in duplicating strands in many directions at once, some gathering complexity, other losing it.

Ant science has its own version of the monkey-to-man mythology, whereby brutish prehistoric hunting ants evolved into sophisticated superorganisms. This narrative is apparent in the Peeters & Molet study:

In higher ants, polyphenism can be extreme: high queenâworker size dimorphism and worker polymorphism allow for better success at colony founding by lone queens, better division of tasks, and ultimately access to new ecological niches. However in basal ants, polyphenism is much more limited, restricting them to narrow niches.

While it is undoubtedly true that the largest and most intricate ant societies emerged from less refined ones, it would be wrong to extrapolate that progression as a general rule into "higher" and "lower" ants. Many modern "higher" ants live in small colonies (Strumigenys, for instance) and evolved from relatively complex ancestors. And, as Peeters & Molet document in their Amblyopone data, some "lower" ants have morphological complexity in their worker caste. Ant evolution is not a straight line. Overarching patterns are obscured by numerous reversals and exceptions, so that reducing evolution to a story not only oversimplifies the process but actively misrepresents it.

Here is a common view of the phylogenetic position of Amblyopone australis, the subject of Peeters & Molet's study:

They justify the placement of Amblyopone as a "lower" ant by pointing to the basal position in this tree and noting that it is close "to solitary vespoid wasps."

But basal does not mean primitive. Nor does it imply evolutionary proximity to external groups like vespids.   To see why, let's look at a different tree.  Here I have included a different set of species, retaining only A. australis and S. invicta.  This tree is just as accurate as the preceeding one- I've taken it from a study by Corie Moreau- it just shows the relationship of a different set of taxa:

Among these taxa, Solenopsis invicta is older and basal and Amblyopone australis is recent and apical. Exactly the reverse of the first tree. Is Solenopsis invicta closer to wasps than Amblyopone australis?

Well, no. These diagrams are only contradictory if we assume that modern species serve as stand-ins for ancestors.  Solenopsis invicta and Amblyopone australis cannot both be ancestral to other ants.  And indeed they are not.  Both are equally distant from their wasp progenitors, as both are modern species with the same amount of elapsed time since the Ur-ant ancestor.   Here is a combined tree:

It should be clear that a basal phylogenetic position has an awful lot to do with the taxon sample, and not a whole lot to do with the amount of evolution.

Many researchers have taken to using the term "basal" instead of "primitive" as more phylogenetically acceptable. But the word never was the problem. The word was a symptom. The problem was the underlying misconception of evolution as a ladder of progress, and just swapping the terminology won't cut it.   Many biologists still can't think clearly about phylogeny.

"Primitive", "advanced", "lower" and "higher" are dangerous words. They come laden with baggage, imbuing values on species that bear little resemblance to the species' actual biology. Worse, these words make us think we know things that, in fact, we don't.  They allow us to substitute prejudice for knowledge.  If we assume that Amblypone australis is primitive, then we are surprised when it turns out to have some derived features.  We shouldn't be. It's standing on just as much evolutionary time as the other ants in the tree.

Finally, here's our own species added to the mix:

Draw your own conclusion.

More like this

The chaotic evolution of colony size in ants.  (Tree re-analyzed from Brady et al 2006, colony data taken from Hoelldobler & Wilson 1990 and other sources) This tree depicts how colony size evolves in ants.  The purple/blue colors represent small colonies with only a few to a few dozen ants…
Martialis heureka Rabeling & Verhaagh 2008 drawing by the inimitable Barrett Klein for PNAS Most scientific discoveries these days emerge through carefully planned and controlled research programs.  Every now and again, though, something unexpected just pops up in a distant tropical jungle. …
Although these two ants in northern Argentina look like they're ignoring each other, they are in fact doing just the opposite. This end-to-end confrontation is an intense chemical duel. What's particularly interesting about the image is the juxtaposition of two different defense systems. At left…
Figure 1. For the 32 most-studied ant species, the percentage of publications 1984-2008 in various contexts. In thinking about where the myrmecological community ought to devote resources in the age of genomics, it occcured to me that putting some numbers on where researchers have previously…

Very nice.

That sounds like some pattern cladism right there! :P

With all seriousness, the above is why I try to avoid basal, derived, ancestral, descendant, and all of those words when I talk about systematics. All we have are the special similarities, which form a pattern. The hypothesis of process comes later.

Thanks for writing this.

I actually prefer "primitive" to "basal", with "ancestral" coming in a clear first. However, these words are applicable for characters only, and should not be used on taxa.

That is also one of my pet peeves, and I agree that it is highly entrenched in ant biology, particularly with the higher and lower Attines. This also involves a misconception of how evolution works. People tend to view lineages nested within a whole bunch of other lineages as being more derived, whereas lineages depicted with few off-shoots as more primitive. That clearly implies two assumptions: that evolution is punctuated (concentrated in cladogenetic events) and that the tree is well-sampled. Both are most often not the case. Moreover, if evolution is continuous (as in a Brownian motion model), all extant lineages have experienced the same amount of evolution.

I prefer reserving those terms for traits and not for lineages.

Let me write shortly* in defense of Peeters and Molet here.

As Kai rightly jokes, unless you are a hardcore pattern cladist, and by that I mean one that maintains that neither taxa nor characters should ever be treated as ancestral, you would concede that phylogenies allow us to analyze the evolution of traits as steps that come historically one after the other.

The misuse of terms like basal, derive, high and low as tree descriptors aside, the point of their paper is that suddenly one finds the evolution of a trait (i.e., worker polymorphism) inside a clade where you would not expect it given the "simple" type of social structure and low queen-worker dimorphism shared among the rest of the species in that group (Amblyoponinae), that is, the ancestral or basal condition for the clade. Appeals to non-ants is necessary to clarify that such condition is not only basal to amblyoponines, but to ants in general.

Take the simplified cladogram above in which Solenopsis invicta is the sole sister group to the amblyoponines and everything still holds of course: black is low queen-worker dimorphism, blue is high queen-worker dimorphism and red is the appearance of worker polymorphism nested within black (as opposed to blue, where it commonly occurs). Yes, both modern taxa are equally distant in time from the most recent common ancestor of ants, but for the traits that are considered to be relevant for the evolution of worker polymorphism, Amblyopone australis, and indeed amblyoponines, seems to have changed little from the most parsimonious/likely/probable condition found in the Ur-ant.

Their paper is about character evolution, and in that regard basal is a perfectly reasonable metaphor: a trait seems to have evolved from a set of initial (basal) conditions not expected by the theory.

That is not to say that I don't agree with you that many people have problems interpreting cladograms properly. In the case of Peeters and Molet, I think their misuse of terms is rather harmless.

This is an interesting discussion and almost makes me want to learn something about ant evolution (which apparently isn't very parsimonious).

In normal English, even the word 'evolve' comes with baggage - usually a gradual change is implied (perhaps that is why Victorian English liked 'Natural' [another bugger of a word] Evolution). Myrmecos is quite right to be annoyed at the biases inherent in the dichotomies above â although this does remind me of one of the first posts I read here â about how cute ant pictures were more highly valued than uncute ones.

Also, I think Roberto has a point, and if I understand his point correctly, we do need to tell a story about our results. Metaphor, analogy, and the rest of the tricks of the storyteller help us to do so. A scientist who canât make their results interesting is unlikely to contribute much to their discipline. I suspect that the very nature of phylogenetic inference would cause perfectly objective presentation of results to be pretty boring (except to specialists who are secretly using the results to tell themselves stories about evolution in the privacy of their own heads).

I feel that I understand the main points of this discussion fairly well, but still have some of questions, as follows:

Okay, so granted there are no primitive/basal ants in existence. The only truly primitive/basal ant was the first eusocial vespoid with a metapleural gland, let us say.

Could it be said perhaps that there are (potentially) primitive, basal ants in the sense of ants currently existing that are the predecessors of the ant taxa/groups that will arise in the future?

I get that all existing ant species have an equally long evolutionary history, in being ants, as do all others? But certainly as species, they do not.

Could someone discuss here the application of (or lack/inappropriateness of application) to situations such as these?:

-- Hybrid fire ants arising from the miscegenation of Solenopsis invicta and S. richteri, which two still exist as thriving species in their own right.

-- Lasius neglectus that appears to have arisen in quite recent history out of L. turcicus.

-- etc.

Marcio said "Moreover, if evolution is continuous (as in a Brownian motion model), all extant lineages have experienced the same amount of evolution."

A very unlikely possibility. Selective pressure is much more likely to be discontinuous with massive evolutionary radiation following some release from biological/environmental continuity THEN followed by gradual 'brownian motion' type evolution - years of boredom followed by minutes of sheer terror if you like.

I personally do not have any problem with the use of "primitive" to describe a phenotype - some morphological characteristics/whatever are bound to have a history of less selective pressure than others.

myrmecos, the proper cautionary should be "imbuing values on species that MAY bear little resemblance to the speciesâ actual biology." since they MAY well be accurate.

James C. Trager said "Okay, so granted there are no primitive/basal ants in existence."

Actually not accurate - there are quite a few fossilized ants represented in amber and inferences about these specimens have been incorporated into the known evolutionary history of ants. Presumably it would be accurate to ascribe clusters of phenotypical characteristics found in both fossilized ants and extant species as "primitive" or "basal" without too much fear of criticisms other than the normal noise concerning convergence etc. to which all phenotypical characteristics are susceptible.

but the concept of "primitiveness" can be useful. considering the last common ancestor of chimps and humans, it's generally thought that ancestor was chimp-like. it lived in the jungle just as chimps still do. the adaptation to the savannah is what turned the hominid line into walking and tool using and eventually sapient apes. certainly the chimp lineage and human lineage have spent an equal amount of time undergoing mutations since divergence, but the selective circumstances, and therefore amount of mutations that were preserved, are completely different. that's why humans are something new and unusual in the world, and chimps aren't.

Actually jon, neither the underlying philosophy nor the empirical data bear you out on this. The recent discovery of Ardipithecus shows the opposite: that chimps actually did undergo a significant amount of post-divergence evolution, and that our preconceptions of "primitive" actively misled people who looked to chimps as a model for the common ancestor.

Though I may be alone in this, I often treat 'basal' and 'derived' as having the meanings of closer to the node for which whatever follows those terms applies, but I don't see them as explicitly having meanings that indicate a taxon is 'more primitive', 'less specialized', 'very advanced', etc. I suppose you could use 'higher' and 'lower' the same way.

For instance, one of my research interests is in crocodyliform phylogeny, where tree topologies are not that stable and not a lot of nomenclature exists to cover different nodes (thank god, compared to Coelurosauria, which has so many freaking different names for its internal clades), it gets a little bit frustrating referring to taxa which are more derived than dyrosaurid-pholidosaurid neosuchians but yet basal to eusuchians.

An example on the other side of the branches for more derived crocodyliforms: most of the taxa which we've come to think of when we talk about notosuchians (i.e. Simosuchus, Mariliasuchus, etc.) are basal members of that clade compared to taxa like Sebecus and allies (Sebecosuchians), but obviously they aren't any less derived in their own weird ways.

The biggest problem I see is that these terms ('basal' and 'derived') are relative to whatever it is you're talking about.

I don't know if I'm part of the problem or not, but I don't associate the baggage with these terms that others do. But if we want to move away from this type of baggage, then I don't know what terms we should in fact use to refer to these taxa or to explain their relative position to other taxa within our trees.