nvestigators have traced the events that led to Thursday's blackout to the failure of several high-voltage transmission lines near Cleveland, officials said yesterday, but it remained unclear how the problem spread from the Midwest to the Atlantic seaboard.

Power experts and companies that use large amounts of electricity had warned for years that the nation's transmission grid was seriously strained, especially in the Midwest. Some of those warnings focused on Ohio.

For years, industry groups have complained of a rapid rise in the number of times that transmission bottlenecks have led regulators to restrict the flow of power around the region. A report by a group called Industrial Energy Users-Ohio cited a nearly fivefold rise in such incidents between 1999 and 2000, and called for changes to be made in the system.

Reports by the North American Electric Reliability Council, which is investigating the blackout, show that the sharp growth of such problems has been limited almost entirely to the Midwest.

Yesterday, the council gave its first detailed timeline of Thursday's events leading up to the huge power failures. The first breakdown, said Michehl R. Gent, the group's president, occurred at 3:06 p.m., when a 345-kilovolt transmission line west of Cleveland shut down for reasons that are still unknown.

Ordinarily, that would have no effect on service because the load would be transferred to other lines nearby. But adding power to a line makes it heat up, expand and sag. Twenty-six minutes after the first line failed, Mr. Gent said, a second 345-kilovolt line in the same area, probably one helping to carry the load from the first failed line, sagged into a tree, causing it to shut down. Trees near transmission lines are supposed to be kept cut back to prevent such accidents.

With the remaining lines in the Cleveland area becoming more strained, three more failed over the next 34 minutes, the last at 4:06 p.m. Over the next few minutes, systems throughout the Eastern United States and Canada began to see huge swings in voltage and in the direction of power flow. More lines went down, and power plants shut themselves off. At 4:11, the blackout began.

The first lines that failed are owned by the FirstEnergy Corporation, a Cleveland-area utility. Last night, FirstEnergy released a statement acknowledging the breakdown and suggesting that an alarm in its control system had failed. The company did not respond to a reporter's request for more details.

The transmission lines in the Cleveland area are part of what the power industry calls the Lake Erie Loop, a ring of lines in the United States and Canada that have become a sort of electricity interstate, moving vast amounts of power from one state to another. On Thursday, wild fluctuations along the lines between Michigan and New York through Ontario caused power plants to shut off, and the blackout began.

In assessments of the reliability of the national grid, Mr. Gent told reporters Friday, that loop "has always been a big, big problem" in part because it is so heavily used.

Readings taken by a company that monitors electric flows for its commercial customers suggest that the problems in Ohio might have started the day before the blackout. Sensors used by SoftSwitching Technologies, a Wisconsin company, show that its customers in Ohio experienced frequent swings in voltage starting after 3 p.m. on Wednesday, continuing into Thursday, according to a database accessible on the company's Web site.

The readings show power lines dropping below their usual operating voltage for a few seconds to several minutes, while the flow to consumers in other parts of the country remained far more stable. Small voltage swings are not uncommon, but they are usually very mild and brief. These were sharper, more prolonged and much more numerous than usual, and the database shows that some customers lost power completely on Wednesday night.

"Instability in the voltage is something you really worry about," said Karl E. Stahlkopf, senior vice president of the Hawaiian Electric Company and a former vice president of the Electric Power Research Institute in Palo Alto, Calif. "In the '96 outages in California, we had very large voltage swings, which caused relays to trip, which caused plants to go offline, which caused blackouts."

The company's database also shows Ohio consumers repeatedly experiencing a much rarer condition: voltage running higher than normal for a minute or more, from mid-June through late July.

Several experts said that such "overvoltage" can indicate systemic trouble in the local grid, and that if it is severe enough, it can damage power systems. But they also cautioned that it is not clear whether any of these patterns were more widespread, or had anything to do with the blackout.

Blackouts are often caused by summer heat, as peak demand strains the system and equipment overheats. Mr. Gent and others have ruled that out in this case, noting that neither temperatures nor demand were terribly high on Thursday.

But a report by the Electric Power Research Institute proposes that warm weather could have played an important role, though an indirect one. When a problem causes voltage to drop, a lamp or a television might dim, but when the voltage rises again, the appliance returns to normal without a notable increase in power consumption.

Air-conditioners, though, work differently, and they account for a huge proportion of summer electricity use. Air-conditioners have electric motors that slow down when voltage drops. The motor's response is to draw significantly more power, trying to resume its usual speed.

The research institute's report proposes that that can lead to damaging swings in voltage, and turn a slight drop into a much steeper one that causes equipment to fail.