MEXICO CITY (AP) -- It started with a burst of gas through the drilling well. Workers scrambled to close the safety valves but within moments, the platform caught fire and collapsed. Tens of millions of gallons of oil gushed into the Gulf of Mexico. Numerous attempts to stanch the spill failed.
Three decades later, the 1979 Ixtoc disaster remains the Gulf's - and the world's - worst oil spill.
The parallels between that disaster and the current BP oil spill offer sobering lessons. There were no quick fixes for Ixtoc: It took 10 months to stop the leak, with Mexico's state-owned oil company, Pemex, trying methods similar to those that BP has attempted at its Deepwater Horizon rig.
Pemex managed to slow the spill a little using several methods including forcing metal spheres into the well. But it couldn't stop the leak until two relief wells were drilled - and even that didn't work right away: the oil kept gushing for another three months after the first well was completed.
In the end, Ixtoc spewed a record 140 million gallons of oil. Massive slicks reached the northern Mexican Gulf coast and Texas, where it would eventually coat almost 170 miles (275 kilometers) of U.S. beaches.
By comparison, Deepwater Horizon has spilled an estimated 21 million to 45 million gallons of oil. But if the Pemex disaster serves as a precedent, the BP spill could continue even after the two relief wells are expected to be finished in August.
By then, it could surpass Ixtoc as the worst oil spill in history, said Tad Patzek, chair of the Petroleum and Geosystems Engineering Department at the University of Texas-Austin.
"We are looking at an August time frame for stopping it," Patzek said. "If for one reason or another that stopping is left imperfect or it takes another drill or what have you, we're looking at another four months, at that time this spill would look like the Ixtoc spill."
The BP spill, at a depth of nearly 5,000 feet, is proving more complicated to choke off than Ixtoc, a shallow-water rig about 150 feet deep. The Ixtoc well could be accessed directly, while the Deepwater spill must be combated remotely, using robots to wield clamps, saws and other tools while monitoring the action by video feed.
"They could fight Ixtoc from and at the surface," said Robert Bea, a professor of engineering at University of California at Berkeley who has studied offshore drilling for 55 years and worked for Pemex for a number of years. For Deepwater, "they must fight from the sea floor remotely, from the sea surface miles above."
Most recently, BP tried to stop the gusher by pumping in heavy drilling mud and cement. The tactic, called a "top kill," had never been tried 5,000 feet underwater. It didn't work.
The good news is the Ixtoc experience suggests the Gulf of Mexico has natural properties that help it cope with massive oil spills, scientists say. Warm waters and sunlight helped break down the oil faster than many expected. Weathering reduced much of the oil into tar balls by the time it reached Texas.
Two decades after the Ixtoc disaster, marine biologist Wes Tunnell sank his diving knife into an area where he had spotted a tar patch just after the spill. The blade came out black and tarry but the hardened surface of the patch was under sand, shells and algae that had completely covered it.
"No one else would know that it was anything other than a rock ledge," said Tunnell of Harte Research Institute for Gulf of Mexico Studies at Texas A&M University. "I think that the Gulf of Mexico is hugely resilient, or at least it was 30 years ago. We've insulted it a lot since then in various ways."
The Gulf has also long dealt with oil that naturally seeps from the seafloor. Some experts estimate that tens of millions of gallons seep into the Gulf from natural up-wellings each year, fostering large populations of oil-eating bacteria and microorganisms.
However, it is unclear how much any of that will help this time around.
The Deepwater spill is closer to sensitive coastlines than Ixtoc was. And it is affecting Louisiana marshlands that are more sensitive than the more sparsely populated Texan and Mexican coastlines Ixtoc reached.
"Obviously there were some helping factors - nature, climate, current - that in the end helped people (with Ixtoc) so that's good news," said Patzek. "However ... the Ixtoc well seemed to have been a little farther out from sensitive places."
The depth of the BP spill could also complicate the Gulf's ability to cope.
The oil-eating bacterial populations are located mainly on the surface or near shore, where the Ixtoc oil appeared. BP has tried to break up the oil deep underwater, pumping chemical dispersants directly into the damaged well.
That could be a mistake, said Larry McKinney, the director of the Harte Research Institute. While chemically dispersing the oil keeps the spill less visible and ugly than Ixtoc, it prevents the oil from floating up to the surface where wind, waves, bacteria and sunlight could help break it up, he said. And some environmentalists question the safety of the dispersant itself.
"I know, out of sight, out of mind," McKinney said. "But also, out of sight is what can kill you, like a cancer, and that may be the bigger problem."
Pemex estimates that about half the Ixtoc oil may have burned away in the rig fire that lasted months. About a quarter dispersed and the rest was either recovered or evaporated.
Ixtoc threatened coral, sea turtles, shrimp and commercially valuable fish. Some species have rebounded while others have drastically declined.
Jack Woody, a retired officer with the U.S. Fish and Wildlife Service, was in northern Mexico a month after the Ixtoc spill, leading the U.S.-Mexico effort to save the Kemp's ridley sea turtles.
He watched as the oil slick closed in and tar balls began washing up at the Rancho Nuevo beach, the only significant nesting ground for the female turtles, which had dwindled to a population of about 300. Hatchlings were just emerging, and helicopters desperately tried to ferry the baby turtles to open ocean beyond the slick.
"I thought we were too late," Woody said.
The turtle population continued to dwindle until 1988 but has since rebounded to up to 12,000, due in part to programs to relocate hatchlings to safer beaches in Texas and cut down on turtle deaths in shrimp nets.
"I was wrong. This is the one time I like to be wrong," Woody said.
Scientists say it is difficult to know how much long-term damage Ixtoc caused because it is hard to separate the effects of the spill from overfishing, sediment, runoff and other pollution.
Tunnell said his students saw coral islands ringed by oil "like black doughnuts" after Ixtoc. Many of those islands have since lost most of their coral cover but Tunnell could not say whether the spill was to blame. Other causes, he said, could include overfishing of coral-friendly fish, coral collecting and sewage.
Perhaps the most tragic thing about Ixtoc is that it is still a mystery.
The tar mats on Texas beaches largely disappeared some five years after the Ixtoc spill, probably broken up and swept seaward. People stopped paying attention once the beaches looked better and funding for research largely dried up.
"Unfortunately, from a science standpoint, not much was followed up on, to learn from," says McKinney. "That was a mistake that was made that hopefully won't be repeated."