For the last three decades, deaths caused by all cancers for white women have been going down slightly -- by 1.8 percent every five years -- even though deaths caused by lung cancer among women have been on going up 35.7 percent every five years for the last three decades.

For white men, deaths from all kinds of cancer have been going up 3.8 percent every five years.

The data, published last week by the National Cancer Institute in its "Atlas of U.S. Cancer Mortality Among Whites: 1950-1980," provide a kind of a snapshot of the distribution of one disease -- cancer -- among white Americans.

A separate report on the distribution of cancer deaths among blacks will be published later.

The calculations are based on death certificates reviewed by the National Center for Health Statistics. This is the first year

the atlas showed cancer trends for various geographic regions.

The cancer institute does the study in order to identify trends and cancer hot spots, and to stimulate additional research in the areas where cancer seems to be more common.The last survey of a decade ago, for example, discovered that oral cancer killed more white women in the rural South than in any other part of the country.

A follow-up study showed that they used more snuff than did the rest of the population. The study provided the first concrete evidence that smokeless tobacco could be dangerous.

The current study helps identify cancer death hot spots, though it does not explain why there are differences in the distribution of cancers.

Compared with the 50 states, the District of Columbia had the nation's highest cancer death rate for men in the 1950s. During the 1960s, the District's cancer death rate dropped to third, and during the '70s dropped to 16th.

For women, the District was the ninth most dangerous place to live in terms of dying of cancer and rose to seventh in the '60s and third in the '70s.

Maryland had the seventh-highest cancer death rate for men during the '50s, climbed to fifth during the '60s, and to second dur- ing the 1970s. For women, Maryland ranked 10th in cancer deaths in the '50s, climbing to eighth in the '60s and fifth in the '70s.

Virginia has a lower cancer rate for both men and women than the other jurisdictions in the metropolitan area.

In the 1950s, Virginia ranked 30th in terms of cancer deaths for men, but began to climb to 27th in the '60s and '70s. For women, Virginia was 32nd in the '50s, climbed to '27th in the '60s and then fell to 29th in the '70s.

Despite relative rankings, actual differences in the cancer death rates are not great from the highest to the lowest states.

If anything, Americans are becoming more alike in the cancers they contract. The North-South differences in colon cancer rates, for example, have diminished over the last three decades, though colon cancer remains more common in the North.

The distribution of lung cancer deaths among women, however, showed an interesting and as yet unexplained pattern. Lung cancer deaths ringed the country, with elevated rates in areas of Florida, along the Mid-Atlantic and West coasts.

"The lung cancer pattern for women is strikingly different from that for men," said Dr. Joseph F. Fraumeni Jr., director of the NCI's epidemiology and biostatistics program. "The reason for the new pattern is not certain, but cigarette smoking may be the main factor."The largest increases in cancer deaths were caused by lung cancer, malignant melanoma, multiple myeloma and connective tissue tumors in both sexes and by laryngeal cancer in women.

The greatest decreases in death rates were seen for cancers of the stomach, lip, bone, rectum and non-melanoma skin cancer in both sexes; and for cancers of the liver, cervix and endometrium, nasal cavity and thyroid in women; and for Hodgkin's disease in men.

Said Fraumeni: "It is not so much where you live but how you live."