Kimbrough Sherman was watching the evening news this week when he learned that six manufacturers had voluntarily agreed to stop producing accordin-style "baby gates" because they pose a possible safety hazard to the small children they were intended to protect.

"I thought at first that they were doing this because the product had caused hundreds of deaths," said Sherman, a business professor at Loyola University and a member of the Maryland Consumer Council. "But then the news said there had been eight deaths and 23 near deaths , which is a rather small number, particularly when you consider the lives that the gates may have saved."

The gates are used to keep children from falling down stairs or getting into rooms where there are hazards.

The point raised by Sherman strikes at the heart of safety regulation, namely: How many deaths and injuries, if any, should occur before a product is removed from the market?

The answer varies, depending on which regulator you ask.

Terrence M. Scanlon, chairman of the Consumer Product Safety Commission, the agency that worked out the agreement to halt production of the baby gates, said, "One death, to me, for any reason is bad."

However, he said, if the agency had tried to force a halt in the production of baby gates by taking the companies to court, "eight deaths would not have been enough for us to prove that there was a significant hazard, because there are 15 million baby gates out there."

But some other regulators say death statistics don't adequately measure the emotional reaction to the loss of a baby who was strangled after its head became trapped in the V-shaped and diamond-shaped openings of the accordion-style baby gate.

"The parents are so afflicted with guilt and self-blame in these cases that they are reluctant even to talk about the incident; they blame themselves for not having provided adequate supervision," said CPSC Commissioner Stuart M. Statler, "and it appears that they are even reluctant to report what happened."

Statler said that in the most recent reported death, which occurred in Idaho in November, "the father, after he found his child dead, kicked the gate out into the yard and it ended up behind the bushes. He didn't even want to see it again because it reminded him of what had happened."

Even more tragic, Statler said, is the fact that there are safer gates available to consumers for comparable prices. They typically have mesh-style material, slats or grids with openings too small to entrap a baby's head.

Statler said the fact that baby gates may have prevented some injuries and deaths isn't the issue.

"That is what the industry said -- look at how many children have been saved by this product," he said. "I told them, 'poppycock.' Because whenever you find a product that has killed kids, you should recall it from the market, because its potential for further death and tragedy is all too apparent."

Statler said the accordion- style baby gate is a product that parents buy in order to protect their children. As a result, they may be lulled into a false sense of security, he said, just as someone who buys a smoke detector that is defective thinks he is protected against fire.

"Both the gate and the smoke detector are supposed to protect, but if they are hazardous, you want to get them off the market and out of people's homes," he said. "Particularly when there are safe substitutes available at comparable prices."