SAN FRANCISCO (AP) — An old memo just revealed by Pacific Gas & Electric Co. to California pipeline regulators indicates the company had noted weld defects on the same line that exploded in San Bruno in 2010 years before the fatal blast occurred.

The San Francisco Chronicle reported on Sunday (http://bit.ly/JlxtiQ ) that a March 1989 memo summarized an investigation into a leak about 9 miles south of where the 2010 explosion occurred, killing 8 people and destroying 38 homes.

Both state and federal investigations into the tragic explosion have been completed, but PG&E said it only recently found the memo and turned it over to regulators last month.

The memo said an internal PG&E investigation had determined the likely cause of the October 1988 leak was a defective seam weld: the same problem that contributed to the San Bruno blast.

If the company's investigation had found such a defect, state law would have required it to test the entire 51-mile-long pipeline years before the San Bruno incident.

The memo said an X-ray of the excavated segment of pipe where a leak was thought to be "showed the weld to be of low quality." That segment was located near the Crystal Springs Reservoir, just off Interstate 280.

Federal investigators have determined that the segment of the line that exploded in 2010 had a defective weld that dated back to 1956, when it was installed. If the company had done pipeline-wide testing after the 1988 leak, it may have discovered this defect.

PG&E said the findings of the 1989 leak investigation were not definitive, and therefore did not require it to test the entire pipeline.

Company spokesman David Eisenhauer told the newspaper that PG&E's investigators never found the actual leak in 1988, leading to a conclusion that a pinhole leak had occurred on a so-called "girth weld." This finding would not have required the company to do the pipeline-wide testing.

Pipeline safety experts said the fact that the company just turned over this document to the state raises questions about PG&E's record keeping and safety practices.

"If you have seam weld risk of this magnitude, you have to assess it — that means a hydro test," Richard Kuprewicz, a pipeline safety consultant, told the newspaper.

"But to me, this illustrates they were not following federal regulations that are very clear, that if you don't have certain records, you have to assume you have the (seam weld failure) threat and assess for it."