Wednesday, April 8, 2015

Another Unsettled Science: The Case of the Brontosaurus

It was; then it wasn't; now it is again.
Brontosaurus Is Back: New Study Says the Dino Is Real After All

"The problems start with the original Brontosaurus holotype. That's what they call the fossil that gets named, and to which every other fossil is compared. The Brontosaurus holotype was found in Montana in 1877 during a period of fierce, unwieldy fossil hunting known as the Bone Wars—where the prestige of naming a new dinosaur often trumped scientific scruples.

The dino's discoverer was notable rapscallion Othniel Charles Marsh, the leading paleontologist at Yale. He named the remarkably complete fossil of Brontosaurus excelsus after declaring it to be a new species. But in 1903, paleontologists decided that this naming had been hasty. Marsh's finding, they said, seemed to be little more than a smaller version of another closely related dinosaur called Apatosaurus. They declared the name invalid and that's been generally agreed upon up until today (even if the public loves the name Brontosaurus).

Later in the 20th century, the case against the thunder lizard got even stronger. Remember what we said about scientific scruples? Paleontologists who went back to the original find in the 1970s uncovered this eyebrow-raising fact: Marsh's original finding was actually a mishmash of two completely different dinosaurs from two entirely different quarries, Poole tells PM. The skull was taken from one (now known to be the skull of Camarasaurus) and just plopped onto the skeleton from another.

Return of the thunder lizard

With that kind of skullduggery around the supposed skull and bones of Brontosaurus, the name seemed an unlikely candidate for a scientific renaissance. Yet that's where Tschopp ended up.

First, he and his colleagues Octavio Mateus and Roger Benson collected a major trove of fossil data on almost all the known Diplodocid fossils. This group includes apatasaurus and most of the other long-necked, long-tailed dinosaurs, though not creatures like Brachiosaurus that have larger front legs than back or massive titanosaurs like Dreadnoughtus. The scientists then ran the data through statistical programs which grouped the dino fossils based on their various bone peculiarities. For the most part, the computer groups matched how paleontologists currently view the evolutionary tree of these dinosaurs—this convinced the team they were on the right track, says Tschopp. But it led them to a few surprises, including Brontosaurus.

According to Tschopp, there are seven specific bone differences that make the body of the original Brontosaurus its own species and genus, not just some other big dino that's been mislabeled. Most of are rather subtle, including facts like this: The tail vertebrae in dinosaurs related to Brontosaurus have spiny prominences called "neural spines," he says, "and for most of these dinosaurs these spines project kind of backwards, but in Brontosaurus they're more straight up." Brontosaurus's hips are unusual, with two bones (the ilium and pubis) meeting a curious junction. And its lower leg fibula meets its ankle bones in an equally unusual manner. Like we said, we're talking about subtle differences. But these are the differences that make a species."
One major problem with science is: scientists.

1 comment:

Robert Coble said...

The science is settled. The debate is over.

Until. . . it's not.

Perhaps we should take this with several grains of salt. (See previous blog posts for the relevance.)

"The scientists then ran the data through statistical programs which grouped the dino fossils based on their various bone peculiarities. For the most part, the computer groups matched how paleontologists currently view the evolutionary tree of these dinosaurs—this convinced the team they were on the right track, says Tschopp. But it led them to a few surprises, including Brontosaurus."

Perhaps these statistical programs are closely related to the statistical programs used to "establish" global climate warming, er, cooling, er CHANGE in the absence of correlating data.

Formulation of a "theory" dependent solely or even primarily on statistical (perhaps Bayesian?) programs seem somewhat fraught with significant potential for major error, especially in areas for which historical empirical data is not only absent, but cannot be created even in theory. Perhaps the "gold standard" should be a continuous reminder that Garbage In = Garbage Out when using any and all computer programs.