Animals may have sprung from the tree of life more than 300 million years later than a recent estimate suggests. According to a new analysis  in the current issue of the Proceedings of the National Academy of Sciences, primordial animals arose 670 million years ago.
Although the oldest animal fossils date back 544 million years, most evolutionists agree that complex, multicellular creatures probably appeared at least 150 million years earlier than that. In 1996, Gregory Wray of the State University of New York, Stony Brook, and his colleagues shook up these assumptions when they published a paper in Science (25 October 1996, p. 568 ) that--by averaging the mutation rates of eight genes shared by animals and nonanimals--estimated that animals arose about 1 billion to 1.2 billion years ago.
Startled by such an early date, Francisco Ayala, a molecular evolutionist at Pennsylvania State University in University Park, took a closer look. After discovering "a host of statistical problems," Ayala teamed up with his eponymous father, an evolutionary biologist at the University of California, Irvine, and statistical analyst Andrey Rzhetsky of Columbia University to rework Wray's data.
Like Wray, the scientists estimated the genes' rate of evolution by comparing their DNA sequences in two different species, a technique calibrated to the date when the two species are known to have diverged. They calculated evolution rates for 12 additional genes, and--most importantly, they say--eliminated from their pool both slowly and rapidly evolving species that might have skewed their average. They then came up with revised estimates of the branching dates for organisms with no fossil record, including the first animals.
Wray stands behind his estimate, contending that it's wrong to ignore certain species. "Most statisticians wouldn't trust an analysis that throws away half to two-thirds of the data," he says. But Charles Marshall, an evolutionary biologist at the University of California, Los Angeles, disagrees. Species that don't adhere to normal evolution rates, he says, are like yardsticks that don't measure 36 inches.