I got interviewed: Palaeontology as a science

A friend interviewed me for his assignment in his journalism course (specialising in science journalism). May he succeed. He transcribed the whole thing and is graciously allowing me to post the meaty part of the interview. Minimal editing on my part to make it web-friendly.

You wrote a post about taxonomic bias in the fossil record and came to the conclusion that it’s not too big a deal. Which of the fossil record biases do you count as the most worrisome?

Just to make it clear, it’s not that I consider taxonomic biases insignificant, it’s that even with them, we can still make many reliable inferences from our datasets. We just need to acknowledge their existence and treat our analyses accordingly.

Let’s consider that there are four major biases in the fossil record:

  • Taxonomic: Absence of certain types of taxa, e.g. unimeralised animals.
  • Ecological: Absence of fossils from certain ecologies, e.g. animals from jungles.
  • Stratigraphic: Absence of fossiliferous sediments from the geological record, e.g. complete erosion of river beds.
  • Biogeographical: Absence of fossils from full distributional range of a species.

Of these, the biogeographical bias is the most annoying, because it will still creep in even when all other biases are avoided. Say you want to study an insect species from the Eocene. You find specimens of it trapped in amber. This is already a lucky avoidance of taxonomic and ecological biases, since these individuals were hanging around the resin-producing trees at just the right time. Also luckily, the amber didn’t get washed away while it was still soft. But even then, you can’t know the full range of this species, because all the specimens come from this one amber-producing forest, in a single window of time. You won’t know if the species lived outside the forest, if it had a distribution that spanned the continents or if it was an endemic of a single clearing in that forest.

The only place where the biogeographical bias is successfully avoided is in the case of ancient lakes. See as an example the Steinheim Basin snails. While we can’t be sure of the original Gyraulus kleini founder population’s distribution, all the subsequent species that formed from it are definitely confined to the Steinheim Basin. But the only way we know this is because we have the complete layering of the lake, with no stratigraphic gaps. The snails lived there, died, sank at the bottom, and remained undisturbed until we uncovered them.

On land, this doesn’t occur. Bacteria eat away at remains in soil. Winds blow away anything light like leaves. Rivers transport small remains and erode big carcasses. Ocean currents ensure that anything light drifts away rather than sink straight to the bottom.

So, in the light of all these hazards that will affect even the most fossilisable animals in the most fossilisation-friendly locations, I deem biogeographical biases to be the most annoying. Of course, most research doesn’t need such tightly-known biogeographies of fossil species, but when they are known, they allow for very good macroevolutionary analyses to be done.

From our private conversation and the last bit of your answer to the last question, I got hints that your interest in palaeontology isn’t just about cool extinct animals, it’s also about evolution as a whole. Is this correct?

Absolutely. Historically, palaeontology has been very underappreciated as a science, referred to as “glorified stamp-collecting” or as a tool for stratigraphers and geologists, but having little intrinsic scientific value. Even now, more often than not, neozoologists and neobotanists will struggle to tell you about the fossil history of their taxa. And don’t get me started about the blatant disregard of the contribution of fossils to reconstructing phylogenies, which is too often ignored in the glut of flimsy molecular data (this is a personal pet peeve).

Palaeontology isn’t just about freaky Cambrian animals and giant dinosaurs. Fossils are data that need to be interpreted. Evolution is partly a historical science, especially when we study phylogenetics, and fossils are the only solid historical data we have (note: I am considering deep time here, not museum archives and ancient DNA which can’t go more than a few thousand years). Even if you ignore this, there are ecological questions that can be addressed using fossils, for example on the effects of climate changes. We need to look at fossils and palaeontology not just as a source of cultural enjoyment and pop sci zaniness, but as a biological discipline with both empirical and theoretical contributions to make to organismal biology, evolutionary biology, evolutionary theory, ecology, among other subdisciplines.

So would you say that palaeontology is one of the more important fields that contribute to evolutionary biology?

Interdisciplinary dick waving isn’t really useful. Palaeontology, zoology, botany, genetics, molecular biology, mathematics, geology, they and more all contribute to evolution, often not singularly but by combining knowledge.

For the working scientist, what’s important is to know which information to use. If you’re studying the evolution of Golgi bodies, you don’t need palaeontology at all. If you’re studying the evolution of C4 photosynthesis, then knowledge of when the first C4 plants evolved certainly helps, although most of your work will be in the lab. If you’re studying the evolutionary history of birds, then not knowing palaeontology will result in an epic fail.

I’d like to go back to your previous answer. Can you think of an example where palaeontology can contribute to population genetics?

I mentioned such an example before: the Steinheim Basin snails! You cannot ask for a more perfect situation with an isolated founder population in a bottleneck. One of the fundamental questions in population genetics is how speciation in these scenarios occurs: does fitness decrease in small populations due to less genetic variability and inbreeding? Or does a bottleneck serve as a particular driver for increasing the fitness of the founder population?

Population geneticists will craft complex theoretical models to try to solve these questions, but as precise as models are, they are useless without an empirical foundation and/or empirical testing.

The Steinheim Basin snails are one way to test them, since it provides a complete run through a bottlenecked founder population that then speciates due to environmental conditions. Sure, we don’t have the genetic information of the Gyraulus species involved, but we do have the entirety of the macroevolution that occurred, and that is one big part of the answer to these questions.

This is fun. Developmental biology?

Sure. Homeobox genes. In 1984, McGinnis et al. discovered that there is a cluster of genes, the HOX genes, that regulate body patterning in all animals. At a very basic level, HOX genes control where each body part will grow. By comparing them across lineages, we can see why insects have a three-segmented body with only three pairs of legs, while crustaceans have limbs all over, and vertebrates only have two. To get an idea of how fundamental HOX genes are, we can modify HOX gene expression in the lab to get a fly with antennae instead of legs.

How does palaeontology creep into this? Simple. We can check how HOX gene expression changed through time by checking up on basal fossils. Cambrian freaks are great for this. We can find out how body plans evolved, and the palaeontological and developmental biology data work in synchrony to provide us with a more complete picture.

So does that mean palaeontologists subscribe to the concept of “hopeful monsters”?

Otto Schindewolf was a palaeontologist, but that doesn’t mean we will all blindly follow him, especially when it’s been known even in Schindewolf’s time that individuals showing such drastic mutations don’t really survive so well (it was R.A. Fisher who showed this). Regardless, the caustic reaction to Goldschmidt’s idea was unwarranted – he got a lot of personal attacks at the time and his legacy is very tarnished despite his considerable influence on evolutionary theory, contributions which fell outside of the hopeful monsters idea. It’s a real shame. But anyway, no, you’ll be hard-pressed to find anyone subscribing to the hopeful monster concept.

Shifting back to the main topic, you also mentioned that palaeontology is important for phylogenetics. Why?

Every fossil is a remnant of the macroevolutionary process during a lineage’s history. The consequence of this process is a classification of all of life into monophyla – groups of organisms that share certain unique characteristics which are hypothesised to have evolved independently within the monophyla.

Palaeontology is useful because it allows us to check whether these hypotheses are correct, since every fossil is a snapshot from the macroevolutionary process. So, say you hypothesise that whales are related to some group of land mammal. You uncover some fossils that share many traits found only in modern whales… but belonging to an animal that could only have been on land. So the fossil animal might have been either a land-capable whale, or it convergently evolved the traits that it shares with modern whales.

You can never be 100% sure whether you’re dealing with convergence or whether you really did uncover a basal fossil, but if you then discover more fossils from many ages so that you can build a rough timeline, you will be able to test which of the two scenarios is more plausible. It’s all about identifying and interpreting fossils correctly (with an open enough mind!), and then analysing everything by taking into account all possible scenarios. In the history of systematics and phylogenetics, there have been cases of very surprising convergences, cases of very surprising non-convergences, and all degrees in between.

Speaking of timelines, you also mentioned in our conversation that you dislike molecular clock studies. I would have thought they are right up your alley. Is your dislike shared by palaeontologists?

You’ll have to ask every single palaeontologist for that :) All I can say for myself is to repeat what lead molecular clock researcher Lindell Bromham said in a 2006 Palaeo Electronica article: “I find the more I learn about molecular clocks, the more I despair that they can ever be trusted.” Bromham then continues by saying that it would be great if they could work reliably.

This is exactly my stance, and I go through it in some detail in a series of posts on molecular clocks: 1, 2, 3, 4.

If I had to generalise for all palaeontologists, disgruntled or pleased, I would say that the molecular clock, even if it worked perfectly, is at best a complement to the fossil record, not a replacement of it. As useful as the molecular clock could be, it still only provides part of the picture: a timeline. It doesn’t give details on what actually occurred at the macroevolutionary level, which is what palaeontology is here for.

In any case, a perfect molecular clock will never exist without the input of palaeontologists, so that should stop those people from getting too ahead of themselves!

Leave a Reply