He was undergoing treatment at St. Anthony's Medical Center in St. Louis, but had returned to his in-laws' home in St. Louis, where he died. He was 67.

McEvilly, who for 11 years directed the Earth Science Division at Lawrence Berkeley National Laboratory, is best known for his 15-year project with U.S. Geological Survey scientists to monitor movement of the San Andreas fault at Parkfield, Calif., in hopes of finding changes in the fault that precede, and thus could be used to predict, large quakes. The spot is hit by magnitude 6 earthquakes on a regular basis, originally thought to be every 20-25 years.

As one of the principle investigators for the Parkfield Prediction experiment, he and colleagues in 1986 began installing bore-hole seismometers, strain gauges and creep meters to monitor fault movement. McEvilly's most important contribution was installation of state-of-the-art deep borehole sensors and recording systems for the High Resolution Seismic Network.

"Tom was very technologically innovative, going straight into earthquake faults with the latest instruments to examine how they work," said William Dietrich, professor and chair of UC Berkeley's Department of Earth and Planetary Science.

Since the last major quake on the fault took place in 1966, the Parkfield team expected to record one by the late 1980s or early 1990s. The expected quake never came - the period is now thought to be about 35 years - but McEvilly learned an enormous amount about the day-to-day movement of the fault at Parkfield. In particular, the deep bore hole recordings allowed very sensitive detection of microquakes, which seem to repeat every one to two years.

With Robert Nadeau, assistant research geophysicist at UC Berkeley's Seismological Laboratory, McEvilly reported in 1999 that these repeating clusters of microquakes could be used to predict the rate at which the fault is slipping deep underground.

The Parkfield site is so well-studied that it has been chosen by the USGS as the site of a 2 1/2-mile deep bore hole to be drilled into the San Andreas fault to record fault properties much deeper than ever before. Called the San Andreas Fault Observatory at Depth, it is in line for funding from the National Science Foundation as part of a larger effort called EarthScope.

"Tom got a lot of data that really helped characterize the physics of earthquake faults on the San Andreas system," Nadeau said. "He was working with the USGS and the National Science Foundation to improve instruments on the fault so he could get data that would let him correlate surface observations with the physical properties of the fault at depth."

At the time of his death, McEvilly was working with other seismologists to determine where to drill the hole and the types of rock expected at different depths.

McEvilly also helped set up a high-frequency borehole seismic network along the Hayward Fault near San Francisco, which is judged the most dangerous fault in the state. Monitoring of the fault helped determine that the northern Hayward Fault slips deep underground to relieve strain and is thus much less likely to generate severe quakes. These studies were partly responsible for the decision by the USGS in 1999 to reduce the likelihood of a major quake along the fault. He also studied plate-boundary dynamics of the Alpine fault system in South Island, New Zealand, an active zone similar to the San Andreas.

In addition, during the height of the Cold War in the 1960s and '70s, he and UC Berkeley colleague Lane Johnson, professor of earth and planetary science, were heavily involved in developing seismic methods to monitor nuclear test ban treaties. They traveled frequently to the Nevada Test Site to monitor nuclear explosions and obtained much information that went into monitoring explosions around the world.

While still a professor, he took on the duties of associate director and then director of the Earth Science Division at Lawrence Berkeley Laboratory from 1982 to 1993, leading it through a period of significant changes in research direction. He served also as geology department chair from 1976 to 1980, and was assistant director of UC Berkeley's Seismographic Station (now the Berkeley Seismological Laboratory) from 1968 to 1989.

In 1984, he helped found and was first president and board chairman of the highly influential Incorporated Research Institutions for Seismology, an NSF-sponsored consortium of nearly 100 U.S. universities and affiliates devoted to development, deployment and support of modern digital seismic instrumentation. It currently operates a global seismographic network of 126 stations, a field program with 250 portable seismic instruments, a data management system that supplies seismological data to scientists around the earth, and an outreach program that promotes science education at levels from K-12 through college.

McEvilly was born on Sept. 2, 1934, in East St. Louis, Ill., and grew up in Belleville, Ill., across the Mississippi River from St. Louis. He earned his BS in geophysical engineering in 1956 from St. Louis University, then worked as an exploration seismologist with the California Company in Louisiana and Mississippi.

In 1960, he returned to St. Louis University as a research associate and graduate student in geophysics. During this period, he also worked part time for the Sprengnether Instrument Company and assisted in the installation of seismographic stations expanding the university's seismic network. After receiving his PhD in geophysics summa cum laude in 1964, he continued with Sprengnether until 1968, eventually becoming engineering vice president.

McEvilly came to UC Berkeley in 1964, where he cemented his reputation as a world leader in earthquake seismology and exploration seismology, specializing in measurements and instrumentation. He retired in 1994, but continued to supervise graduate students.

"He always had a twinkle in his eye and a smile," Dietrich said. "He was a very positive presence in the department whom we will really miss."

His retirement wasn't all work. He loved the Lake of the Ozarks area of southeast Missouri, where he and his wife owned a home and spent half the year. One of his great joys, his wife said, was fishing daily on the lake.

He published more than 200 scientific papers, mentored many students, participated in a variety of professional activities and organizations and served on numerous government committees and panels. He also was editor from 1976 to 1985 of the Bulletin of the Seismological Society of America,

He was a fellow of the American Geophysical Union and the American Association for the Advancement of Science, an honorary member of the Seismological Society of America, and a member of the Society of Exploration Geophysicists, the Royal Astronomical Society and Phi Beta Kappa.

He is survived by his wife of 31 years, Dottie Hopfinger McEvilly; children Mary Scott, Susan McEvilly, Ann Dodson, Joe McEvilly and Adrian McEvilly of the San Francisco Bay Area; son Steve McEvilly of Orlando, Fla.; brothers Robert McEvilly of San Diego and Mike McEvilly of Berkeley; a sister, Dorothy Hennessy of Conn.; and nine grandchildren.

Burial will be in Mt. Carmel Cemetery, Belleville, Ill., following a mass of Christian burial celebrated by McEvilly's friend and cousin, Fr. Jack McEvilly. Mass is scheduled for noon on Saturday, Mar. 2, at Our Lady Queen of Peace Catholic Church in Belleville. Visitation is at 11 a.m.

A memorial service will be held in Berkeley at a later date.