PASADENA, CALIF., NOV. 24 -- An unusual earthquake swarm that struck California's Imperial Valley Monday night and today, culminating in a 6.3 magnitude shock early this morning, was blamed for two deaths and caused at least 94 injuries in the United States and Mexico.

Scientists at the California Institute of Technology, a major center for seismological research, first reported a strong earthquake registering about 6.0 on the Richter scale at 5:53 p.m. Monday, but they anticipated more activity because of the tendency of Imperial Valley temblors to be preceded by foreshocks.

Lucile Jones, a seismologist with the U.S. Geological Survey here, said she returned to the laboratory here at 2 a.m. today in case of even bigger shocks. The 6.3 earthquake struck just over three hours later, at 5:16 a.m.

Seismologist Tom Heaton, also with the U.S. Geological Survey, said the swarm of quakes in the slowly widening, seismically complex desert valley may yield clues to connections between fault systems that can aid earthquake prediction in more densely inhabited parts of the state.

Teresa Romero, a communications supervisor with the Imperial County sheriff's department, said she was roughly shaken awake at her El Centro home by this morning's earthquake. "It sure felt a lot stronger than the one last night," she said.

In addition to some broken windows, fallen masonry and displaced trailers, she said an earthquake-related fire was reported in Mexicali, a Mexican city near the border.

She said one of the injuries involved a pregnant woman who fell down stairs and broke an ankle. Fifty persons were reported injured in Mexico, and 44 were treated in California.

There were other stress-related injuries, including some mild heart attacks, according to police.

Carmen Garcia, 30, of Mexicali, and her son, Esteban Garcia Jr., 4, died when the Monday quake jolted a highway 50 miles east of Mexicali, causing two cars to collide, Ignacio Aguirre, spokesman for the government of the Mexican state of Baja California, told the Associated Press.

Seismographs recorded two significant foreshocks before Monday's major quake, and many aftershocks into the evening, including two in the 5.0 range.

Several aftershocks followed this morning's 6.3 magnitude quake.

Seismologists call such a series of earth movements a swarm.

Monday's and today's quakes were centered within six miles of each other, the first about 11 miles northwest of Westmorland, along an unmapped fault southeast of the brackish Salton Sea, and the second about 14 miles west of that same 1,500-population desert town, near the Superstition Hills fault.

The quake caused a ground shift of four to six inches 10 miles from the epicenter -- typical of a California earthquake of that magnitude.

Heaton said the many faults in that area are difficult to detect on the surface because of centuries of sediment laid over the valley by the meandering Colorado River.

Scientists have laced the Imperial Valley with seismographs in order to record its unusual patterns of earthquakes, growing out of its position at the northern end of a subterranean rift that is pushing apart two large continental plates. Known as the East Pacific Rise, this zone of volcanoes and other marks of thinness in the Earth's crust has over the last 10 million years separated Baja California from the rest of Mexico and still pushes it slowly out to sea. If it were not filled with river sediment, the Imperial Valley would now be part of the Gulf of California, Heaton said.

Major earthquakes in such spreading zones tend to signal their arrival, apparently because of the same underground seismic forces that now allow scientists to predict some major volcanic eruptions at places like Mount St. Helens in Washington State. Jones said 90 percent of the major quakes in the Imperial Valley are preceded by foreshocks. A 6.6 magnitude Imperial Valley earthquake near El Centro in 1979 caused $30 million in damage and about 70 injuries.