The first of the many wonders of Karl H. Norris was the machine that replaced the old candling techniques and lifted egg quality control into the 20th century.

After that came the machine to separate the brown eggs from the white eggs, and then there was no stopping Karl Noris. Inventions for measuring food and grain quality have tumbled out of his laboratory to such a degree that it's probably safe to call him the Bell or the Edison of his field.

He came up with a commonly used device to measure protein in wheat. He developed a tool for measuring oil, protein and moisture in soybeans. He invented a machine to measure the fat in a package of hamburger. Another one parses the "water core" in apples -- a deterrent to long-term storage.

Norris, a friendly, easygoing man of 60, is one of those government researchers whose creative strokes have altered the visage of American agriculture. His work has led to four patents, from which, because he's a federal employe, he receives no royalties. In return for his brainpower, Uncle Sam pays him $50,112 a year and honors him with a GS-15 grade.

Much of that creative scientific work goes on at the Department of Agriculture's big research complex in Beltsville, Md., where Norris hangs his apron, and at the USDA-supported laboratories at the land grant colleges.

They're into everything: improving plant species, finding new uses for farm products, increasing crop yields, determining how to combat pests and disease, improving animal productivity and nutrition.

Karl Norris fits into this picture as a researcher on instrumentation, looking for ways to measure more quickly and accurately the food content of grains and meats.

His most important development was 10 years ago, when he introduced infrared reflectance spectroscopy (IRS), a simple and inexpensive technique for grading the chemical composition of seeds and other plant materials. Before IRS, traditional chemical measurement could take weeks before results were ready.

Norris' big breakthrough was called the greatest advance in grain analysis since the 1880s, and was significant enough to win him the 1978 Alexander von Humboldt Award for agricultural research.

The newest by-product of his research was on display the other day at Beltsville -- a computer-operated system mounted in a standard van that in two minutes can give a farmer an analysis of the nutritive content of the hay, silage or grain he is feeding his livestock. With such information, a farmer's profits could suddenly soar.

The unit was developed, with USDA moeny for the most part, by Dr. John S. Shenk, a research agronomist and professor at Penn State Funiversity, who has taken Norris' ideas about infrared technology and tried to put them to work for the farmer.

The result is the high-tech van that Shenk believes could be a boon to farmers who need to know the nutritional content of the feed they give their livestock. With the van estimated to cost $70,000, an individual farmer probably couldn't justify buying it. But a cooperative or an entrepreneur could, providing the on-the-farm testing service to subscribers.

The feed, be it grass, alfalfa or a grain, is quickly dried in a microwave oven. The dried material is ground to a dust, slipped into the IRS unit and analyzed. A computer printout gives a reading of protein content, digestible nutrients and moisture. Beyond that, the computer will tell the farmer how he should alter his herd's diet to get the best results.

Norris has been working in USDA research for 30 years, pushing from one frontier to another. A few of his inventions -- the hamburger fat analyzer and the egg color separator -- didn't really catch on. Brown eggs became passe, and supermarkets weren't enamored of the meat tester.

No matter. Norris keeps on in his austere lab at Beltsville, working now on ways to measure the nutritional content in grapes, information essential for marketing and harvesting purposes.

But that doesn't mean private enterprise will rush the work of Norris and his fellow researchers into production.

The classic example might be the hand-held unit Norris developed a Year or so ago to measure "water core" in apples as they come from the tree. If an apple has too much moisture, it can't be stored for long periods of time. Ordinarily, Norris explained, an orchardman will cut open sample apples to make this kind of test.

"They can determine water core by observation, but they don't like to do it," Norris said. "There really is a need for an instrument like this."

But no one seems thrilled with the idea, and that's the trouble with being out there on the front lines of invention. Sometimes you have to lead a horse to water.