The outbreak of World War II transformed Caltech's research priorities. Almost overnight armed guards appeared, and a flood of federal funding became available for a range of new projects investigating weapons and defense. Old projects were dropped. New ones, from building guided missiles to making better gas masks, were started.

Pauling, an ardent anti-fascist, was fingerprinted, given a security check, and invited to war-research planning meetings in Washington, D.C. He quickly shifted his laboratory's work toward the development of rocket propellants and explosives, and he personally worked on better instrumentation for submarines and weapons sighting, and designed (and patented) an armor-piercing shell. Eventually he was invited to participate in the Manhattan Project to build the atomic bomb (an invitation from Robert Oppenheimer himself, which Pauling turned down because, he said, he did not want to relocate his family).

All of Pauling's prewar protein projects were pushed into the background, with one exception: antibodies. Funding agencies were interested in wartime medical breakthroughs as well as military advances. Pauling capitalized on that interest, garnering a government grant to develop an artificial, gelatin-based substitute for blood plasma, and getting other funding to expand his research into the workings of the immune system.

Before Pearl Harbor he had already been at work proving, as his theory of antibody formation proposed, that each antibody molecule had two and only two binding sites. To do that, he and his coworkers developed ingenious methods using precisely defined synthetic antigens with one, two, or three binding sites. By reacting these antigens with antibodies and carefully measuring the resulting complexes, Pauling's work provided strong evidence that antibody molecules are indeed bivalent, an important step forward in the field.

He also claimed to have found evidence that a single antibody molecule could bind to two different antigens, a result also predicted by his theory of antibody formation. Later it was found that this observation was wrong; in fact, while individual antibody molecules are indeed bivalent, both "arms" bind to a single type of antigen.

This was not the only time that Pauling's meticulous, often important work with antibodies was tarnished by his enchantment with his own theories. He seemed always to find results that matched his preconceived ideas. For instance, Pauling's basic idea about antibody formation, in which the two ends of a "vanilla" long-chain protein molecule molded themselves to an antigen's molecular shape, also predicted something that might turn out to be a gold mine: according to the theory, it should be possible to manufacture antibodies in the laboratory. If a more or less generic protein, say something common like beef globulin, was carefully denatured, then allowed to renature in the presence of antigen, it should mold itself to the antigen to create a specific antibody, a molecule specifically built to bind that antigen. The result would be artificial antibodies, made to order.

The possibilities were breathtaking. Antibodies were among the most powerful medicines known; they were the end result of a million years of evolution that allowed the body to fight off infections and invaders. If they could be made by the quart in a laboratory, they would represent a breakthrough in medical care. The rights could mean millions to the discoverer. Pauling's theory predicted that they were possible. Early experiments in his lab started to indicate that they were. His excitement grew.

In late 1941, Pauling applied for a patent for a method of making artificial antibodies.