US puts crop data on computer to help plant breeders. New network may spur development of wider variety of crucial US crops

For some time, scientists have worried about the lack of genetic diversity in the country's major crops. For instance, of the 150 or so distinct racial groups of corn available in the Western Hemisphere, no more than three have been used in the United States. And scientists are concerned by the lack of variety in other major crops as well.

In the past, a narrow genetic base in crops has proved disastrous. The classic example is Ireland's potato famine in the mid-1800s, in which the varieties were so closely related that it allowed an obscure blight to virtually wipe out the crop for several years. A century later, US farmers were hit by the Southern corn-leaf blight, which reduced the overall corn crop by more than 10 percent in 1970.

Now, with the help of computers, government scientists here at the Beltsville Agricultural Research Center are taking a small but important step to improve the situation.

For many years, the United States has collected plant material, or germplasm, from all over the world. Germplasm - the part of a cell that contains the hereditary information of a plant - has been used to breed new and better varieties of food crops. But lack of information on the government's germplasm collections has made it much harder for plant breeders to use them, scientists say. Instead, breeders traditionally have reshuffled the few varieties already known to be successful.

And most of the government's germplasm collections have not been evaluated. Scientists say they need more-detailed knowledge on plants' resistance to disease and insects.

A new computer network has been created here that will gather together the available information on the government's germplasm collections. Dubbed GRIN - Germplasm Resources Information Network - it will allow plant breeders and researchers to plug into the databanks by phone and quickly find the known information, says Leonard L. Jansen, the network's program analyst.

The new network will be become more valuable to scientists as more evaluation work is done, says Quentin Jones, national coordinator for the National Germplasm System of the Agricultural Research Service, one of two research arms of the United States Department of Agriculture.

The system will help fill in some gaps of knowledge, since scientists experimenting with germplasm will be able to report their findings into the system.

Still, Dr. Jones concedes, much more evaluation is needed. One estimate is that less than 10 percent of the germplasm collections in the country has been evaluated, he says.

Earlier this year, speakers at a plant-breeding research forum in Washington pointed out several problems in storing, maintaining, and enhancing the germplasm collections so they would be readily usable by plant breeders.

Their No. 1 area of concern, however, was the lack of evaluation.

''In the past 30 years we've put no emphasis on it,'' says Major M. Goodman, professor in the crop science department at North Carolina State University. ''We've had committee after committee after committee. If we had as many people in the fields as in committee meetings, we'd have no problem.''

Is this of immediate concern? For most crops, the genetic base ''is broader than it was 10 or 15 years ago,'' says William L. Brown, former board chairman and president of Pioneer Hi-Bred International Inc., a major seed-corn producer based in Des Moines. ''(But) it has a long way to go before it gets to the point where it should be.''

The main obstacle appears to be funding.

Jones points out that the federal germplasm system received only about $13 million this year - even though ARS estimated a need for $21 million this year and $40 million by 1988.

In this budget-cutting era, Jones isn't particularly optimistic either about more funding. ''You don't hold your breath,'' he says.

But ARS has placed more emphasis on evaluation work. In the past two years, funding for it has quadrupled, Jones says, although ''it's still low in relation to the total need.'' GRIN costs roughly $500,000 a year, he adds.

It will be near the end of this year or early 1985 before the system has enough information to be useful, says Jim Mowder, the network's database manager.