At the end of winter 1948 Pauling came down with a severe cold. The damp English winter worsened it into a sinus infection, and he soon found himself confined to his London flat, congested, feverish, and at his wits end. "The first day I read detective stories and just tried to keep from feeling miserable, and the second day, too," he remembered. "But I got bored with that, so I thought, 'Why don't I think about the structure of proteins?'"

He grabbed some paper and a pencil, and began sketching possible structures for keratin - a somewhat off-hand attempt to do what he had first tried in 1937, to build a model based on what he knew about the amino acid building blocks, the planarity of the peptide bond, and what was known about the x-ray patterns, including Astbury's 510-picometer repeat. Almost all of the basics were the same. Corey's work with amino acids had, for the most part, simply confirmed what Pauling had thought about their structure more than a decade earlier; the peptide bond appeared to have the properties he thought it had; and Astbury's work had shown little new about keratin.

Only one significant idea was new: the helix. Pauling started his sketches by drawing the molecular backbone, a chain of amino acids linked end-to-end. He used a heavy line to indicate the peptide bonds between them; these were spots that would have to stay relatively flat on the page to represent the planarity of the bond. With the chain roughed out more or less to size – that is, the positions of the backbone atoms relative to one another were about what they should be - he started playing with the paper, making his flat sketch into a sort of squared-off tube, folding it where carbon-carbon bonds in the backbone allowed the amino acids to bend, keeping the peptide bonds as flat as possible. He tried to keep the fold angles roughly the same as a tetrahedron, the basic bond angle for chain-linked carbons in amino acids. He worked the paper around, folding and refolding, telescoping the tube in and out to align the atoms he had sketched, trying to line them up so that as many hydrogen bonds as possible could be formed between the turns of the helix.

It was not long before Pauling came up with a structure that looked surprisingly good. The bond angles and lengths were about right. A reasonable number of hydrogen bonds appeared to form naturally, pinning the structure into place. "Well," he remembered, "I forgot all about having a cold then, I was so pleased."

But there was a problem. The model Pauling put together did not seem to match a critical measurement, Astbury's 5.1 picometer repeat, which would theoretically measure the distance between each full turn of the helix. Pauling knew that much more careful model-building would have to take place before it was confirmed one way or another, but for the moment he held off spreading the news. He did not write Corey about it. No record indicates that he mentioned it to any of his British hosts. He did not even save his original sketch. Instead, he filed the idea away. There would be time to work on the details when he got back to California.