The large stuffed dog was not in its usual place. The toy, still wrapped in a thin plastic bag, ordinarily sat on stacked boxes and watched over the twins' crib. The twins' mother awoke and found her 8-month-old son curled asleep in a corner of the crib. The baby girl lay face up beneath the toy dog. She was dead.

At the Kings County Hospital in New York, the baby's death was attributed to sudden infant death syndrome, SIDS, in part because there were no obvious signs of injury or abuse.

SIDS is defined as the sudden and unexpected death of an infant when no other cause can be found. It is the most common diagnosis in infants who die between one month and one year of age. An estimated 7,000 to 10,000 babies are believed to die of SIDS each year in the United States.

But according to a study published in The New England Journal of Medicine last week, those estimates may be too high.

After careful examination of the scene of the baby girl's death, the authors of the study concluded that she was asphyxiated by the fallen toy.

In fact, concluded the authors, of 26 consecutive cases of sudden unexpected death in infants brought to the emergency room of Kings County Hospital 24 were actually due to suffocation, heat stroke or other accidents caused by a parent's or babysitter's poor judgment.

Twenty-three of the 26 cases had been diagnosed as SIDS by the medical examiner.

Across the county, many of the deaths attributed yearly to SIDS, also known as crib death, are probably due to preventable accidents, said Dr. Millard Bass, a forensic pathologist at the State University of New York and an author of the controversial study, which took place in an economically disadvantaged area of Brooklyn.

"This is not a study peculiar to Brooklyn's inner city, but it's representative of all major cities in the U.S. where problems can occur with toys, overwrapping infants, and tired mothers sleeping with babies," said Bass. "The diagnosis of SIDS is an easy way out. It's like a waste paper basket term."

The study by Bass and associates already has provoked a heated debate among physicians and an emotional backlash from parents of SIDS victims.

"In the study, there are implications of parental misunderstanding of dangers to sleeping infants -- that's blaming parents," said Dr. Frederick Mandell, who, as vice chairman of the National SIDS Foundation, a support group, has received a number of calls and telegrams from parents reacting to the findings. "Parents are very angry and say this study can't be true. They feel needlessly blamed when their child, in fact, died of natural causes."

Some experts question whether any far- reaching conclusions can be drawn from a study that focuses on a small group of low-income families, many of them living in crowded tenements.

"It's a very unique study population," said Dr. Charlotte S. Catz, chief of the pregnancy and perinatology branch at the National Institute of Child Health and Human Development. "Are the results applicable to other areas? Does it apply to suburban Bethesda? To some areas, it might apply, but not universally."

Other physicians, including Mandell, question the methods used to conduct the study.

"The methods are crude. Going into a home following a death and retrospectively attempting to recreate what happened is a crude way of determining facts," said Mandell, who is also an associate clinical professor of pediatrics at Harvard Medical School. "Parents may be confused about what happened. They may react to suggestions on the part of the interviewer. It's a very confusing time."

Mandell and members of the National SIDS Foundation dispute Bass' claim that more thorough investigation would reveal causes of death other than SIDS. They say most babies who die of the syndrome show no signs of parental negligence and that the causes of deaths cited in the study are unusual.

"The majority of SIDS cases are not due to accidental deaths, asphyxiation or trauma," said Dr. Frederick Mandell, vice chairman of the National SIDS Foundation and associate clinical professor of pediatrics at Harvard Medical School. "Those are very unusual cases."

In the Brooklyn neighborhood served by Kings County Hospital, the diagnosis of SIDS is not unusual. This inner-city area has more than four SIDS deaths for every 1,000 births, a rate 10 times higher than in more affluent neighborhoods 20 minutes away by car. It is also a mortality rate similar to that of poor areas in Washington and cities across the nation, Bass said.

For the study, researchers visited the family's home, sometimes repeatedly, within one week of the infant's death. They interviewed the families and babysitters; examined rooms, beds, bedding and sleepwear; and checked environmental conditions for temperature, ventilation and unusual odors. To reproduce the death scene, researchers provided a doll that was the size of the dead infant. The family was asked to position this doll where the infant was found at the time of death.

After those in-depth investigations, Bass concluded six of the deaths probably resulted from accidents and the other 18 cases were probably not the result of SIDS.

Many of the deaths were attributed to asphyxiation and hyperthermia, or heat stroke. In one case, a 4-month-old boy, who had a cold, was dressed in a polyester undershirt and jumpsuit, and put to bed next to a radiator on a chilly night. The baby, covered with three blankets, was found dead in the morning. Later study showed the temperature in the crib rose to 105 degrees Fahrenheit when the heat went on at dawn.

"The infant was literally cooked," said Bass. "But when the radiator is off and the house is cold, who would think his death was due to heat?"

In six cases, Bass suspects a sleeping parent inadvertently smothered the baby. Such deaths, known as overlying, were first mentioned in the Bible and thought to be common during the 19th century. In more recent years, however, most experts have concluded that death by overlying was improbable, believing the parent would wake before the infant died.

But overlying has not been widely studied. Bass and associates say overlying, though difficult to prove, occurs and should be considered among possible causes of sudden infant death. The authors' conclusion that some of the deaths were caused by overlying were based on interviews with the mothers.

"Experts say overlying doesn't happen," said Bass. "Baloney it doesn't happen. That's a myth."

Causes of death like overlying, hyperthermia and asphyxiation are difficult to detect and get lumped into the SIDS category, said Bass. "The medical examiner, overworked and understaffed, doesn't give the same attention to SIDS as he gives to a homicide or an auto accident -- that little infant won't go to court. If the medical examiner can't find anything anyway, he says 'SIDS' and it's over -- that makes parents happy. But it also means you get statistics on SIDS that are meaningless."

Dr. Richard E. Kravath, an author of the study, is quick to point out that overuse of the SIDS diagnosis is not a reflection of the quality of New York medical examiners.

"New York City medical examiners are as good as any in the nation. It's not the medical examiners who are at fault -- we've all not been seeing what is going on," said Kravath, director of pediatric inpatient services at Kings County Hospital. "If you give a name, you've achieved something -- rather than saying you really don't know what caused the death, which creates a lot of anxiety. It's the desire we all have to comfort the family."

Kravath also believes the families' economic status was not solely responsible for the infants' deaths. "Some characteristics related to being impoverished, others to poor judgment, and others to a lack of knowledge that we all have," he said. "Insufficient information is all over -- it's a problem for all of us. You can't blame poor people or rich people."

Nor does Kravath think the families in the study made unique mistakes. "Some things may be unusual, but in most cases, even people who have taken parenting courses would make these mistakes. For instance, what's the risk of sleeping with a baby? Nobody knows."

Kravath and Bass believe the true causes of many deaths ascribed to SIDS can be found through careful investigation at the scene of the infant's death. Autopsies, generally conducted for SIDS cases, do not detect asphyxiation or certain other causes of death. But an investigation, they say, would detect these types of causes and save the lives of other children.

"With more in-depth investigations, I suspect the SIDS rate would go down and the accident rate would go up," said Bass. "We'll never prevent deaths if we act like ostriches putting our heads in the sand and never looking. To sign these deaths off as SIDS is unfair to other children whose lives may be endangered by the same problems."

Kravath says an investigation can be conducted in a way that would not add to parents' grief or induce feelings of guilt. "I'm talking about investigation, not accusation," he said. "We should avoid the one extreme of not investigating and avoid the other extreme of assuming these are homicidal murders."

According to Kravath and Bass, the families in the study were relieved to learn the findings of the investigation. "There's no question that the families were very saddened, but they were also relieved to know what had happened," said Bass. "Then things could be changed in the household and those families could be shown how to prevent that accident from happening again." Some physicians and SIDS Foundation members condemn the study's call for death scene investigation. They say it would only contribute to the agony of grieving families, who are also victims of the SIDS tragedy.

"To unnecessarily traumatize a family with an investigation is particularly harsh," said the SIDS Foundation's Mandell. "It's the nature of investigation that it's difficult to do without causing some guilt or innuendoes of guilt."

Kathy D'Elosua, of Springfield, also opposes in-depth death scene investigations.

Last July, D'Elosua had checked on her 2-month-old napping daughter, Melissa. The baby was sound asleep and breathing easily. D'Elosua left the room and brought back her husband to photograph Melissa, who had looked so cute sleeping in a sun dress. When the husband and wife entered the room, Melissa was blue and lifeless. At the hospital, the diagnosis was SIDS.

"When a thriving infant dies, you blame yourself, you just do. I will die with the guilt," said D'Elosua, speaking haltingly. "I would rather see more money spent on research and less on badgering parents. I believe in facts, but why investigate? Nothing can be proved and it just means more pain. An investigation should consist of no more than six questions asked in a gentle, compassionate way."

Nora Zamichow is a New York-based writer who frequently contributes to Health.