Ancient grazers: find adds grass to dinosaur menu.

Analyses of fossilized dinosaur feces in India reveal the remains of at least five types of grasses. The finding not only provides the first evidence of grass-eating dinosaurs but also shows that grasses evolved diverse forms much earlier than scientists had previously recognized.

Bits of silica called phytoliths indicate the grasses' presence. The tiny crystals, which form within cells of many plants, are especially plentiful in grasses, according to Caroline A.E. Stromberg, a paleobotanist at the Swedish Museum of Natural History in Stockholm. Because each type of grass produces distinctly shaped phytoliths, scientists use the readily preserved grit to identify the mix that once grew in an area.

Stromberg and her colleagues examined spherical coprolites, or fossilized feces, that measured up to 10 centimeters across. The material was probably created by titanosaurs, the most common type of dinosaur represented in the rock layer holding the coprolites, says Stromberg.

The coprolites contained phytoliths and other remnants from a variety of plants, including palm trees, conifers, and cycads. The researchers identified some of the 65-million-year-old phytoliths as coming from a variety of grasses. Modern relatives of these ancient species of dino fodder include rice and bamboo, says Stromberg. She and her colleagues report their findings in the Nov. 18 Science.

From other research, the oldest fossils that are indisputably grasses are about 56 million years old, says Stromberg. However, some sediments from at least 70 million years ago contain pollen grains that may have come from that group of plants. Also, recent genetic analyses of modern grasses hint that some major forms may have appeared as early as 83 million years ago.

The team's new results are "very important," says Dolores R. Piperno, a botanist at the Smithsonian's National Museum of Natural History in Washington, D.C. Because phytoliths, particularly those of grasses, provide definite identification of a plant type, "it's conclusive that grasses were there," she adds.

"This is really pretty exciting," says Elizabeth A. Kellogg, an evolutionary biologist at the University of Missouri--St. Louis. If the research holds up, "it would completely revise what we've thought about the origin of grasses," she notes. "This isn't that much older than the oldest previous grass fossils, but to find such diversity at that time is surprising."

The presence of grasses at least 65 million years ago may give clues to a mystery regarding ancient mammals called gondwanatheres. They appeared in the waning days of the dinosaurs, could be as large as groundhogs, and had extremely long teeth with a flat chewing surface, according to Greg Wilson, a paleontologist at the Denver Museum of Nature & Science.

Those features are characteristic of modern grazers such as horses, whose teeth must stand up to constant abrasion by phytoliths in grass. Before the recent finding, such teeth were inexplicable in the mouths of gondwanatheres because there didn't seem to have been grasses for them to eat, Wilson notes.

TOUGH CHEW

This phytolith, which was extracted from fossilized dinosaur dung unearthed in India, indicates that the reptiles dined on grasses.

[ILLUSTRATION OMITTED]