Biological Diversity: Fact or Artifact?

A sandstone fossil of a trilobite, which probably lived 500 to 600 million years ago.

It’s one of the biggest questions of biology and one of the most intractable. How did biodiversity change after the ‘Cambrian explosion’ 540 million years ago? For many years, paleontologists have thought that this rapid diversification of multicellular life started a long-term increase in biodiversity, as measured by the number of genera, during the Phanerozoic era (from the Cambrian to today). Although periodic mass extinctions punctuated this increase in biodiversity, the overall trend was toward more genera (groups of closely related species).

A study published last year of a fast-growing database on Phanerozoic marine diversity questions this conventional wisdom. The study was reported in the May 22, 2001, issue of the Proceedings of the National Academy of Sciences (PNAS). John Alroy, of the National Center for Ecological Analysis and Synthesis in Santa Barbara, which houses the database, and colleagues used statistics to try to correct biases in previous analyses. Their study compared two intervals: 460 million years ago (Ma) to 300 Ma and 164 Ma to 24 Ma.

Estimating biodiversity by looking at fossils is akin to looking for keys under the lamppost because that’s where it’s light. Bias can result from the tendency to collect fossils in accessible geologic formations, from researchers who focus on certain eras, and from the decomposition of older specimens.

The study looked at marine invertebrates, Alroy says, because there was no terrestrial life until the Ordovician period, between 510 Ma and 440 Ma; because marine invertebrates comprise a huge part of overall diversity; and because these animals fossilized well.

Juggling Fossil Numbers

Nonetheless, the marine invertebrate record does suffer from systematic biases. Optimally," says Alroy, ‘you’d want to sample an equal amount at every time and space, in every kind of environment.’ Instead, fossils from places that were extensively collected tend to predominate in fossil collections. Rather than fund a new wave of more systematic field paleontology and then wait decades for the results, Alroy and company tested statistical techniques to standardize" remove sampling biases from the existing data which the new database has assembled.

One method for standardizing is extrapolating multiplying diversity in under-sampled eras to compensate for under-sampling. But that, Alroy says, tends to exaggerate errors. Instead, his group chose to interpolate to statistically improve" the data so it would seem that collectors had expended equal effort in each geologic era.

Alroy and his colleagues tried several different approaches. A statistical technique dubbed the "unweighted" manipulation examined a restricted number of collections. If, say, 100 collections were available from one era, but 600 from another, they might sample each period by looking at 50 randomly chosen collections from that period. The resulting statistics would tell how well any group of 50 collections represent true biodiversity from that era.

Brachiopods were found in the Ordovician, Silurian, Devonian, Mississippian, and Pennsylvanian.
Credit: Dept. of Earth & Atmospheric Sciences, Cornell University

In the so-called "occurrence-weighted" manipulation, the researchers looked not at the number of lists from a particular time period, but at the occurrences of genera in various lists. If a particular genus is present as one specimen in collection A, three specimens in collection B, and 100 specimens in collection C, the researchers counted just three occurrences, Alroy explains. Instead of setting a quota of lists, occurrence weighting sets a quota of occurrences. Occurrence weighting, Alroy says, is a new mathematical technique to standardize lists. Rather than having a firm justification in reality, it is an "operational" technique that can be used for comparisons.

The resulting curves from four separate statistical manipulations of marine invertebrate diversity were confusing, as even some study authors admit. ‘Given the ambiguity of the curves we presented, it’s not obvious that there’s any immediate conclusion you can draw,’ says Charles Marshall, an evolutionary biologist and invertebrate paleontologist at Harvard, another author of the database paper. Still, the overall result was relatively flat curves that, Alroy says, raised questions about earlier analyses, which showed a long-term increase in diversity. ‘The empirical conclusion is that previous estimates may be inflated toward [recent times],’ he says.

The paper’s many authors listed numerous biases that could affect their results, such as the emphasis in the database and the underlying fossil collections on North America and Europe, where funding for paleontology is relatively abundant.

‘Inconclusive at the moment,’ is the judgment of Mark Newman, an applied mathematician at the Santa Fe Institute who commented on the Alroy study in the same issue of PNAS in which the study appeared. ‘It’s not clear to what extent they have really fully corrected for biases. It’s a difficult thing to do and the data are incomplete, so it’s possible that the results are still influenced by the same or different biases.’ However, Newman says the big news may not be the paper’s conclusion so much as its announcement of the large, growing database, which will eventually expand beyond marine invertebrates and thus will be available for addressing many questions of paleontology, ecology and evolution.

Flip Side

But does the database provide enough trustworthy data to reach even tentative conclusions on Phanerozoic diversity? No, according to Jeremy Jackson, of the Scripps Institution of Oceanography, and Kenneth Johnson of the Natural History Museum of Los Angeles County. In the September 28, 2001, issue of the journal Science they wrote that, Ever more complete and detailed compilations of old data collected for different purposes, by different methods, with different taxonomic standards, and without abundance data cannot resolve the problem of inadequate and inconsistent sampling." These authors proposed to resolve the issue by spending about $100 million on 20 intensive paleontological projects around the world.

‘Our data are mostly new, not old,’ Alroy responds. ‘Well over half of our references were published after 1980, and more than 90 percent were published after 1960.’

In the ideal world, we’d love to have the type of data" that Jackson prefers, comments Marshall. ‘But we don’t and won’t for a long time, and rather than shoot ourselves in the foot, we’re trying to see to what extent we can use a database" to explore the reality behind the existing collections. ‘Jackson is essentially saying, ‘We’ve got two centuries of field collections, and they’re not really worth anything.’ Get real."

Rather than detracting attention from field paleontology, Marshall says the database should have the opposite impact. We will have done the discipline an enormous service if we can show that the data [for a specific geologic period] are insufficient." Such evidence should help requests for support for research into some of paleontology’s dark holes.

Otherworldly Significance?

A Cambrian eurypterid chases trilobites. Beneath, the legged worm-like animal is Aysheaia pedunculata, an onycophoran.
Credit: Univ. of Michigan Exhitbit Museum of Natural History

In terms of astrobiology, the study helps flesh out our understanding of how fast, and under what conditions, life expands to fill available niches. Previously, scientists had thought that biodiversity increased more or less steadily since the Cambrian rebounding after each mass extinction to reach the present high level. But that may not be true at least when diversity is measured by number of genera.

‘Few thing we do are more relevant to astrobiology,’ says Alroy. If this paper holds up, it suggests that .. even for multicellular, complex organisms, diversification can reach its limit in a few hundred million years. That strongly predicts that if we find multi-cellular life on other planets, it should be close to the limit of diversity. When you step on that planet, you are likely to see the carrying capacity of that planet in terms of biodiversity."

What’s Next

Now that a single database contains data from fossil collections from around the world, Marshall says existing research should be easier, and that new types of research become possible. The database should, for example, help determine, as some suspect, taxonomists, who focus on identifying and naming new organisms, find more species than ecologists, who tend to concentrate on organisms with larger ecological impacts.

The new database, Marshall continues, has ‘the potential to revolutionize the rigor, depth, intricacy, and subtlety of these studies. More factors can be analyzed simultaneously if all the information is sitting there. The PNAS paper was as much about announcing this new sophistication, which is allowed by information technology and large-scale cooperation at the raw data level of the paleontology community.’

The new database is already being extended to terrestrial animals. Overall, Marshall says, the technology is, ‘a little akin to having a new telescope with 10 times the resolving power.’