One of the most common events in human reproduction is poorly understood and only sketchily studied. Its name is "spontaneous abortion," a process by which -- for reasons largely unknown -- a developing human being is expelled from its mother's body long before it has a chance of independent life.

No one knows how prevalent spontaneous abortion is, but it certainly accounts for more than half the human ova that are fertilized. Counting misadventures later in pregnancy plus stillbirths, perhaps as many as 80 percent of conceptions fail.

Most people think of an unsuccessful pregnancy as a miscarriage, which a standard dictionary defines as "expulsion of a fetus before it is viable, especially between the third and seventh month of pregnancy." Far more numerous are fetal losses in the first three months.

A genetic defect in the fertilized egg is regarded as the leading cause of failed pregnancy. Others include the woman's age -- older women are more likely to miscarry -- and physiological problems with the woman's ability to have a pregnancy. None of those causes can be controlled.

But there are others, such as hormonal deficiencies and infections, which can often be dealt with medically. And there are external factors -- such as environmental pollutants -- over which the woman may or may not have control, including tobacco and alcohol.

The first time a woman becomes pregnant, there is no way to know in advance whether she is miscarriage prone. Across the board, the failure rate is regarded one in five or one in six for a first pregnancy.

Once a woman has lost a fetus, however, her risk of losing another rises. Drs. Richard I. Feldbloom and Betty Y. Forman, in the medical text "Pregnancy, Birth and the Early Months," cite these odds:

"The chance in a first pregnancy is about 20 percent. After having a miscarriage in her first pregnancy, a woman runs a 28.2 percent risk of one in a second pregnancy. After two miscarriages, her risk rises to 37.9 percent; and after three, to 50 percent."

"Spontaneous abortion occurs randomly in most cases and there's nothing that can be done about it," says Dr. John Graham, deputy director of the Washington-based American College of Obstetricians and Gynecologists.

A novel experiment still under way in North Carolina has given new insight into very early spontaneous abortion. Dr. Allan Wilcox at the National Insitute of Environmental Health Sciences, in a recent interview, explained:

More than 200 women of childbearing age who had recently discontinued birth control in order to become pregnant were studied for six months or until they became pregnant.

Urine samples were analyzed for human chorionic gonadotrophin (HCG), the most reliable early marker for pregnancy. The researchers found, Wilcox said, "that out of all the chemically detected pregnancies we found . . . about 22 percent had ended without the woman being aware of the loss."

This 22 percent loss is in addition to the 20 percent considered normal for miscarriage of known pregancies, suggesting that more than two-fifths of detectable pregnancies end in this very early stage.

The HCG test, Wilcox points out, is not reliable before about seven days, so "there's probably more pregnancy loss that takes place earlier."

Other researchers are working on ways to improve the odds for miscarriage-prone women. Drs. Jerome Check and Khosrow Nowroozi at Thomas Jefferson University Hospital in Philadelphia are pursuing research on a steroid hormone, progesterone, which the female reproductive organs produce to help prepare the uterus for implantation of a fertilized egg.

Last January at a conference on sterility and fertility in Puerto Rico, Check reported that when 34 women with two or more spontaneous abortions were treated with progesterone, 22 carried both of their next two pregancies to term, and six others were successful in one of their next two pregnancies.

A parallel -- if contrary -- development was reported late last year in The New England Journal of Medicine: a drug called RU 486 that induces miscarriage by nullifying progesterone's effects. In an editorial comment on the drug, the phrase "medical abortion" was used, suggesting that denying progesterone to the reproductive system at a crucial time may become an alternative to surgical abortion. When spontaneous abortion occurs, it is often found that the fetus actually died some time earlier. It usually happens around the 12th or 13th week of pregnancy but can occur later. Dr. Heinz Berendes of the National Institute of Child Health and Human Development says that about half the fetuses expelled at 12 or 13 weeks are found to have died four weeks or more before.

The risk period for spontaneous abortion is usually considered over by 20 weeks, when the fetus is about six inches long and weighs just about a pound -- well on its way to full development but still unable to breathe independently.

But at about 24 weeks, and having nothing to do with the condition of the fetus, there is another potential pitfall. This is called "incompetent cervix," in which the mouth of the womb gives way with little or no warning.

Many doctors believe damage done in previous childbirth or surgical abortion can cause this condition, since is rarely noted in women pregnant for the first time. One textbook calls physical damage "the most common cause" of incompetent cervix.

Women who have lost one fetus due to incompetent cervix are regarded as extremely likely to lose others. Two procedures, called the McDonald cerclage and the Shirodkar technique -- in which the neck of the womb is fitted with a pursestring-like suture -- can save future pregnancies in 85 to 90 percent of cases.

The procedure must be performed between the 14th and 24th weeks of pregnancy, and the closure must be cut when labor begins. A woman with this type of repair surgery must seek prompt help if labor begins prematurely. Of all controllable causes of fetal loss, smoking is recognzied as No. 1, with alcohol consumption probably No. 2. Doctors also believe that most fetuses lost as a result of smoking had nothing fundamentally wrong with them.

"If I were a woman planning to get pregnant," says Dr. James Mills of the National Institute of Child Health and Human Development, "I would stop smoking beforehand; I wouldn't wait until I was pregnant to stop."

Two alcohol studies showed that women who drank twice a week had a 60 percent greater risk of miscarriage than those who did not drink. And those who drank daily had twice the normal risk of miscarriage.

Evidence is less clear cut on other environmental hazards. Early this year a study of women in the computer-chip industry showed spontaneous abortion rates of 29 and 39 percent in two groups of production workers, compared with 18 percent in non-production employes. The theory is that chemicals used in the manufacture of these electronic components may harm the fetus, but not all scientists are satisfied with the study's reliability.

The National Institute of Occupational Safety and Health, an arm of the Labor Department, is supervising a study of 4,000 female employes of the Bell South telephone company to determine whether the emanations from computer screens and the stress of working in front of them has any bearing on pregnancy outcome.

Environmental lead has long been suspected as a promoter of miscarriage, but some scientists are dubious. Even though lead is much less widespread today than it was in the 1960s due to the advent of lead-free gasoline and disappearance of lead-based paints, miscarriage rates nationally have not changed significantly in the last 20 years.

William Hines is a Washington free-lance writer.