Weight of Opinion: More times than most of us would like to admit, a doctor's opinion is taken at face value merely on the strength of his reputation. Such seems to have been the case of a Canadian forensic pathologist whose evidence - although somewhat sloppy by pathology standards - was taken as gospel at trial. Some cases that are now in doubt:

William Mullins-Johnson of Sault Ste. Marie has been in prison 12 years for the murder of his four-year-old niece Valin. It was recently discovered that Dr. Smith lost tissue samples that could exonerate him.

Eleven-month-old Nicolas Gagnon of Sudbury died in 1995 after apparently bumping his head on a table. Dr. Smith disagreed with the findings of the first pathologist, and concluded after a second autopsy that the cause of the baby's death was non-accidental, blunt-force trauma. The parents were never charged, but while they were under suspicion authorities seized their second child.

In Kingston, Louise Reynolds spent two years in jail facing a murder charge in the 1997 death of her seven-year-old daughter Sharon. Dr. Smith believed that 80 cuts on the girl's body were caused by scissors. He then changed his view and said that the injuries could have been from a pit-bull attack. The Crown withdrew the charge.

Three-year-old Tyrell Salmon of Toronto died in 1998. Based on Dr. Smith's conclusions, his father's girlfriend, Maureen Laidley, was charged with murder. The charge was withdrawn on the eve of trial after three other pathologists concluded that the bump on Tyrell's head was likely caused by falling on a coffee table.

"Expert opinion" doesn't mean infallible, or apparently, even accurate.