IS FALLOUT from nuclear reactors exposing people to dangerous levels of radiation -- even without an accident? The Nuclear Regulatory Commission claims that normal emissions from reactors are safe. But a report which the commission has been sitting on since early this year shows that these government safety claims are based on fraudulent research.

Performed by Atomic Energy Commission scientists 20 years ago, the experiments demonstrate the Eisenhower-Lewis Strauss policy of making fallout look harmless. Thus today, the new report says, fallout from normally operating nuclear reactors is exposing people to radiation in excess of the authorized limits.

Although the NRC admits the charge about the fraudulent experiments, it denies that fallout from nuclear reactors is a health hazard. To find the truth in these conflicting claims, it helps to begin with what is known.

It is known that each of this country's 72 reactors releases some radioactive waste. This happens by design and because of human and mechanical flaws. Uranium fuel rods crack, pipes leak, filters fail and so do workers when they open the wrong valve. Thus the NRC limits how much atomic waste each reactor may discharge into the water and air.

To measure what this is doing to people, the commission makes a dose calculation. This tracks the radioactivity from the reactor through the food chain into human bone and tissue. The resulting "dose estimate" is the core of the claim that a normally operating nuclear reactor is safe.

But the NRC now acknowledges that this safety calculation is based in part on the dubious experiments of the 1950s. The problem begins with the conviction that atmospheric nuclear testing was essential to national security.

To quiet the Nervous Nellies who wanted a test ban, the old AEC took steps to show it was keeping on top of fallout. One such step was a program of experiments to find whether food crops would take up dangerous levels of fallout from the soil. More than just looking for results that would make fallout appear safe, most of the scientists rigged their experiments to produce the desired reassurance.

In measuring how much fallout plants would pick up from the soil:

The AEC scientists made preliminary tests on a variety of soils. They chose for their experiments those soils which absorbed the least amount of fallout.

It was known that plants have difficulty assimilating many fallout ingredients until they are acted upon by soil bacteria. To prevent this, the scientists cooked their soil in ovens and killed its bacteria.

Then they added the radiotoxic substances to the soil shortly before the plants were harvested. This avoided the conditions of reality, where the plants would grow from seeds in the contaminated soil.

Not surprisingly, these experiments showed hardly any fallout was getting into crop plants.

The report which reveals this information breaks new ground. It's the first time independent scientists have dug into the NRC's safety assurances to expose their foundations. Written by a team of 14 West German scientists -- agricultural biologists, physicists, chemists, a mathematician, a physician and a veterinarian -- from the University of Heidelberg, it applies directly to this country. Not only does Germany build its reactors from American designs but it proves their safety with the same set of calculations the Nuclear Regulatory Commission uses for that purpose here.

As performed by the NRC or the utilities, these calculations show a reactor giving little radiation to a person living within 10 miles -- a fraction of a millirem to less than 5 millirems yearly. (Current estimates peg a chest X-ray between 15 and 30 millirems.)

But the German scientists say that in measuring fallout's journey from reactor to residence in the body, the NRC figures are "either at the lower end of the range given in the literature or far below the values that may be regarded as realistic. It follows that the results of these assessments are unrealistically low."

They say, for example, that NRC judgements on how much plutonium, cesium and strontium crops pick up from the soil are "between 10 and 1,000 times too low."

The Heidelberg group reached its conclusion after digging through 25 years of scientific journals to find what experiments had been done on how much fallout was getting to people. They then compared the results of these other experiments with the NRCfigures -- eight other experiments in the case of cesium, 11 others in the case of strontium.

After examining the NRC safety estimates, the report scores another first. It calculates the dose from a nuclear reactor, using figures chosen by independent scientists. What the German scientists did was feed their figures into the NRC computer model.

They found that a pressurized water reactor planned for the town of Whyl on the Rhine could be expected to expose people to a yearly dose of 1,071 millirems of wholebody radiation. The major part of this dose would come from radioactive substances taken into the body with food and drink. The reactor's exhaust air is the principal source of this contamination, with its waste water playing a significant but lesser role. t

The exposure limit in the United States is 170 millirems of wholebody radiation yearly from all nuclear facilities. On Dec. 1, the Environmental Protection Agency will reduce this limit to 25 millirems.

But it would be inaccurate to transfer the 1,071-millirem result from the Whyl study to a reactor in this country or anywhere else. Conditions specific to each power plant's site, such as wind patterns, nearness of farmland, size and type of reactor could bring these figures up or down.

Still, the study strongly suggests that realistic safety calculations would show each of this country's 72 reactors burdening people with more radiation than the new 25-millirem limit. Since the NRC uses these dose estimates at its basic yardstick for licensing and regulation, it would have cause to act against all the country's nuclear plants in more than a cosmetic way.

But this is not likely to happen, for more than economic reasons. Just as the Heidelberg group calls the NRC figures too low, the commission replies that the German scientists' figures are too high. "Their literature search was not comprehensive," an NRC spokesman comments. "They looked for experiments that would support their conclusions."

More importantly, Dr. Frank Congel, leader of the NRC's radiological impact section, says that "real measurements" by the utilities show that radiation emmited by neclear plants is well within current safety limits.

Can the utilities be trusted to monitor how much radiation their own plants are giving off? Asked whether the NRC has assigned the fox to guard the chickencoop, Dr. Congel responded that the NRC "reviews the records and procedures of each plant on an average of twice a year." The commission also "spotchecks seven or eight plants each year."

But these "real measurements" by the utilities fail to answer the Heidelberg report. The German scientistss are attacking not the utilities' emission figures but the set of caluclations used to estimate what dose these releases give to human beings. To verify the Heidelberg report wold require monitoring food from farms and dairies near rectors, a step which the NRC calls unnecessary and expensive.

Yet both the Environmental Protection Agency and the states monitor mile for radioactivity. Although milk from dairies near reactors often shows high levels of strontium 90, both the utilities and the NRC claim this contamination comes from atmospheric fallout. They blame this partly on the residue left in the atmosphere from old U.S. and Soviet tests, but mostly on the more recent Chinese bomb testing.

Since strontium 90 lacks a label of origin, no one can know for sure. But state and EPA monitoring shows a pattern. After the U.S.-Soviet test ban in 1963, strontium 90 milk levels remained high across the country for two or three years. Then the average dropped from roughtly 25 picocuries per liter to 4 picocuries, where it remains. After a Chinese test, these levels jump briefly, then go back to what is now normal.

But in certain areas near nuclear plants the level of strontium 90 in milk has at times risen higher than it was at the height of atmospheric nuclear testing. Such readings were found in milk from dairies in Waterford, Conn., near the Millstone nuclear plant, when that facility was experiencing high radiation releases in 1976.

In Wisconsin, South Carolina and other states, rises in strontium 90 milk levels from dairies near reactors have also been recorded after releases of radiation from thoses facilities. Still, NRC and utility officials insist the strontium 90 comes from Chinese tests. This presents a pictures of an America covered by a huge umbrella in the sky with holes punched over certain reactors.

This is less ridiculous than it sounds, since out technical institutes teach fledgling physicists that reactors do not give off strontium 90. Dr. Bernard L. Cohen, director of the Scaife Nuclear Laboratories at the University of Pittsburgh, writes in "nuclear Science and Society" that "strontium 90, which has received wide publicity for its importance in bomb fallout, is removed in the chemical purification and hence is of little consequence here." (That is, it is removed by a filter within the reactor.)

Unforunately, what's written in textbooks does not always happen in reality. Not only can filters be less than perfect, but the largest portion of radioactivity released from reactors, which is in the form of gases, presents special problems. These gases decay into radioactive particiles, including strontium and cesium. Nuclear plants are designed to hold these gases long enough so that the decay takes place within the plant, where the particulates can be filtered out.

But if the gases leave the plant before they have had time to mellow, rather than floating away harmlessly, they deposit strontium, cesium and other isotopes in the environment. These gases can be released prematurely due to plant emergencies. Or, according to the Heidelberg report, significant quantities of the gases often leak out of the plant before they reach its filtering system.

Further, a Wisconsin investigation based on 14 years of milk monitoring by state officials supports the validity of the Heidelberg report, giving evidence that reactors are indeed releaseing strontium 90.

America's dairyland state is bordered on three sides by a Big Dipper-shaped chain of 14 nuclear reactors. Wary of safety assurances, a Wisconsin evironmental foundation asked the State Radiation Section to prepare a dose estimate based on the official milk sampling program, started in 1963. The state refused. So this concerned group of middle-aged, middle-class professionals -- Land Educational Associations Foundation Inc. -- took on the project with the assistance of a University of Minnesota boilogy professor. To insure credibility, they chose to use data only from state monitoring records.

The chain of reactos around Wisonsin grew from two to 14 between 1970 and 1976. State records show milk strontium 90 levels jumping in 1973 from just below to more than twice the national average, and staying there at least three years. (The study had to close with 1976, since state officials were more than two years late in supplying records which are supposedly public.)

The largest increses of strontium 90 came in a Wisconsin area 50 miles downwind, and downriver, from Minnesota's Monticello reactor, and in the Green Bay area around the Point Beach nuclear plant. Both high readings of strontium came and persisted during the year that Monticello was leading the nation in gaseous releases of radioactivity, and Point Beach was tripling its allowed emissions.

Showing instance after instance where nuclear plant releases were followed by high radioactivity in milk, the study points to reactor fallout as the major source ot contamination.

In calculating its dose, the Heidelberg report judges the impact of a large reactor on people living in a small area -- roughly within a 10-mile radius, making a circle around the plant of 314 square miles. But the Wisconsin study judges the impact of 14 smaller reactors on people spread across a territory of 54,000 square miles. Thus exposure for each person will be smaller.

Since they based their study on just the three radioactive poisons monitored by the state -- strontium 90, cesium 137 and iodine 131 -- the Wisconsin investigators label their "Total Fission Dose" findings as "not the whole dose."

Using government formulas, they find the average yearly dose to Wisconsin citizens from nuclear waste in food and the environment is 33 millirems of whole body radiation for an adult, and 67 millirems for growing child. The yearly dose to the bones is 76 millirems for an adult and 174 millirems for a growing child. The study says this "extra" radiation has more than doubled the risk of blood cancer for Wisconsin 14-year-olds.

They caution that this radiaiton comes from more than milk. Although ideal for monitoring, milk is one of the least radiotoxic foods. It contains only 10 percent of the radioactivity found in the grass eaten by the cow, which filters out the rest. As foods high in fallout radiation, the investigators note potatoes, whole wheat, leafy vegatables, soybeans, berries, venison, nuts, cabbage and cheese, which multiplies milk dose times six.

The NRC rejects the Wisconsin study, even though it is based on verified state milk records and calculated with EPA and NRC formulas. The NRC's Dr. Congel states the Wisconsin study "showed extreme bias in its data and its presentation when we review it."

At the same time that the EPA is establishing the new 25 millirem exposure limit, the NRC is stopping the monotoring of strontium 90 at nuclear plants. The reason given is that the utilities, assigned by the NRC to monitor themselves, aren't finding much. But like most manufacturers, utility executives want responsiblity to end at the front gate. The public can't see the strontium 90 leaking out of the plant. Why should the utitlity? More importantly, why should nuclear plant owners be responsible for monitoring themselves?

The agency responsible for protecting the environment has given some indication of what its new 25 millrem limit means. One of its radiaiton officals explains," the EPA does not have any regulatory requirement to monitoring the environment around nuclear power plants; this monitoring is required by the NrcY of their licensees."

When told of high strontium 90 levels being monitored in milk and fish from the area near the three Oconee reactors in South Carolina, another EPA radiaiton official responded that the 25 millirem limit applies to "planned discharges of radioactive materials." It does not apply, he said, "to background or fallout radiation in the vicinity of nuclear power plants." Chinese fallout again, and a large loophole for the many unplanned releases which reactors experience.

Both EPA and NRC plan to enforce the new limit by letting the utilities tell them what radiation is being released. From this they will estimate the dose to the public, using that same set of calculations which the Heidelberg report examined.

Both agencies have copies of the Heidelberg report. The NRC translated and printed it last spring (NRC translation 520: "Radioecological assessment of the Whyl Nuclear Power Plant.") Its distribution has been delayed while the German authors made some revisions. But Bernd Franke, one of the Heidelberg scientists, told this writer they had changed nothing of substance.

The significance of the report is that it indicates there is a great range of uncertainty in what is known about the impact of nuclear plants on human beings. The only way to determine whether the official dose estimates have validity is by a rigorous and continuing program of food monitoring performed by an independent agency.