Fossil record biases strongly distort scientific conclusions about how mammal body size and diet evolved over time, according to a 2026 simulation study in Nature Communications. Researchers found that biological and taxonomic filters, which determine which animals are more likely to become fossils, significantly skew evolutionary interpretations, while geographic location has minimal impact on bias.
Scientists studying how animals changed over millions of years rely on fossils, but the fossil record is incomplete and biased. A new study simulated what future fossils of today’s mammals might look like to understand how these gaps distort our understanding of evolution. Researchers found that certain factors, like which animals are more likely to be preserved as fossils, can seriously skew what we think we know about how body size and diet evolved in mammals. This research helps scientists understand which evolutionary conclusions from fossils are trustworthy and which ones might be misleading.
Key Statistics
A 2026 simulation study published in Nature Communications found that biological and taxonomic filters strongly skew inferences about mammal evolution from fossil records, while spatial filters do not significantly distort these conclusions.
Research simulating a future fossil record of present-day land mammals showed that fossil record incompleteness can substantially bias interpretations of deep-time biodiversity change, with effects varying depending on preservation rates.
According to the study, which animals are more likely to fossilize based on their size and species type creates stronger evolutionary biases than geographic sampling differences in the fossil record.
The Quick Take
- What they studied: How incomplete fossil records mislead scientists about how mammal traits like body size and diet changed over time
- Who participated: Computer simulations of present-day land mammals, filtered through realistic fossil preservation conditions
- Key finding: Biological and taxonomic filters (which animals are more likely to fossilize) strongly distort evolutionary conclusions, while geographic location has minimal impact
- What it means for you: When you read about how ancient animals evolved, remember that fossil conclusions may be incomplete. Scientists need to account for these biases when interpreting deep-time evolutionary patterns.
The Research Details
Instead of digging up actual fossils, researchers used computer simulations to predict what a future fossil record of today’s mammals might look like. They started with a complete family tree of modern land mammals and then systematically removed species based on realistic factors that affect fossilization. These factors included how much sediment was available in different regions, how large or small the animals were, how wide their geographic range was, and how much scientific effort went into finding fossils. The researchers tested three different preservation rates, meaning they simulated scenarios where fossils were preserved at different frequencies. After removing species according to these realistic biases, they analyzed what remained using the same statistical methods that paleontologists use to study actual fossils. This allowed them to see how much the evolutionary conclusions changed compared to the complete, unbiased starting data.
This approach is important because it lets scientists test their methods before applying them to real fossils. By knowing exactly which species were removed and why, researchers could measure exactly how much the conclusions changed. This is impossible with real fossils because we never know what we’re missing. The study essentially creates a controlled experiment that shows which biases matter most when interpreting fossil data.
This is a simulation study published in Nature Communications, a highly respected scientific journal. The strength of this approach is that it uses realistic parameters based on actual fossil data and modern mammal biology. The limitation is that simulations, by definition, cannot capture every complexity of real fossilization processes. The study provides first-order estimates, meaning it gives us a starting point for understanding bias rather than a complete picture.
What the Results Show
The research revealed that some biases distort evolutionary conclusions much more than others. Biological and taxonomic filters, which determine which types of animals are more likely to become fossils, strongly skewed the results. For example, if large animals are more likely to fossilize than small animals, scientists studying fossil records might incorrectly conclude that mammals evolved to become larger over time, even if that did not actually happen. Similarly, if certain groups of animals are better represented in the fossil record than others, conclusions about how diet evolved can be seriously misleading. In contrast, spatial filters, which relate to geographic location and where fossils are found, did not significantly distort the evolutionary conclusions. This suggests that uneven geographic sampling of fossils is less problematic than uneven sampling of different animal types.
The study found that the degree of distortion depends on the preservation rate, meaning how frequently fossils are preserved. Under lower preservation rates, where fewer fossils survive, the biases become even more pronounced. This makes sense: when you have very few fossils to work with, the ones you do have become more likely to be unrepresentative of the true diversity. The research also showed that these biases affect different evolutionary questions differently, suggesting that some conclusions from fossil records are more reliable than others.
According to Gram Research analysis, this study provides the first systematic quantification of how fossil record incompleteness distorts specific evolutionary conclusions. While paleontologists have long recognized that the fossil record is incomplete, this research moves beyond general awareness to measure the actual magnitude of bias. Previous work acknowledged these problems conceptually, but this study uses simulations to show exactly how much evolutionary conclusions can change.
The study uses simulations rather than actual fossils, which means it cannot capture every real-world complexity of fossilization. The researchers made assumptions about preservation rates and filtering mechanisms based on current knowledge, but actual ancient conditions may have differed. Additionally, the study focuses on land mammals, so the results may not apply equally to marine animals, plants, or other organisms with different fossilization patterns. The sample size for the simulation is not specified in the abstract, which limits our ability to assess statistical power.
The Bottom Line
When reading about mammal evolution based on fossil evidence, consider that conclusions about body size and diet changes may be biased by which animals were more likely to fossilize. Scientists should explicitly acknowledge these biases when presenting fossil-based evolutionary conclusions. This research suggests that geographic sampling bias is less concerning than biological sampling bias, so conclusions about where animals lived may be more trustworthy than conclusions about how their traits changed.
Paleontologists, evolutionary biologists, and science communicators should use these findings when interpreting fossil records. Students and the general public should understand that fossil-based conclusions have built-in limitations. Museum exhibits and textbooks discussing mammal evolution should acknowledge these biases. This is less relevant to people studying living animals or recent evolutionary history.
This research does not propose changes to current practices but rather provides tools for evaluating existing conclusions. Scientists may begin incorporating these bias estimates into their analyses immediately, but changing textbooks and public understanding will take years.
Frequently Asked Questions
How do gaps in the fossil record affect what we know about animal evolution?
Fossil gaps create biases that can make scientists draw wrong conclusions about how animals changed over time. A 2026 study found that certain types of animals are more likely to fossilize than others, which distorts our understanding of evolutionary trends in body size and diet.
Why are some animals more likely to become fossils than others?
Large animals, animals with hard body parts, and animals living in areas with lots of sediment are more likely to fossilize. These preservation biases mean the fossil record overrepresents certain animal types and underrepresents others, skewing evolutionary conclusions.
Can scientists fix the bias problem in fossil records?
Scientists cannot fix the incomplete fossil record itself, but they can account for known biases when interpreting data. This 2026 research provides estimates of how much different biases distort conclusions, helping scientists evaluate which fossil-based claims are most reliable.
Does where fossils are found matter as much as which animals fossilize?
Geographic location matters less than animal type. The study found that spatial sampling bias does not significantly distort evolutionary conclusions, while biological filters that favor certain animal types create much stronger distortions.
Should I trust what textbooks say about extinct mammal evolution?
Textbook conclusions about mammal evolution are based on real evidence, but they have built-in limitations from fossil record bias. Look for sources that acknowledge these limitations rather than presenting fossil conclusions as complete certainty.
Want to Apply This Research?
- Track which evolutionary claims about extinct mammals you encounter and note whether the source acknowledges fossil record bias or preservation limitations
- When learning about extinct animal evolution, seek out sources that explicitly discuss fossil record limitations and biases rather than presenting conclusions as definitive facts
- Over time, notice how scientific communication about fossil-based evolution becomes more nuanced and includes caveats about preservation bias
This research is a simulation study that estimates how fossil record biases affect evolutionary conclusions. It does not represent actual fossil data but rather models realistic scenarios. Conclusions about specific extinct mammals should be evaluated by consulting primary paleontological research and considering multiple lines of evidence beyond fossils alone. This study provides methodological insights for scientists rather than definitive answers about any particular animal’s evolutionary history.
This research translation is published by Gram Research, the science division of Gram, an AI-powered nutrition tracking app.