Most of what is written about Gregg Toland (1904-1948) treats Citizen Kane (1941) as the year zero of deep focus cinematography. This is convenient and almost entirely wrong. By 1941, Toland had been working out the technique for the better part of a decade, on films that have nothing to do with Welles. The reason Kane gets the credit is that Welles was generous enough to put Toland's name on the same title card as his own, and the film is more famous than the others. The credit is fair. The history is not.
A short list of pre-Kane Toland deep focus work: The Long Voyage Home (1940, dir. John Ford); The Grapes of Wrath (1940, Ford); Wuthering Heights (1939, William Wyler), which won Toland his only Oscar; Dead End (1937, Wyler); These Three (1936, Wyler). The technique on display in Kane arrives there fully developed because Toland had already developed it.
What deep focus actually requires is straightforward in principle and difficult in practice. To hold sharpness from foreground to background you need a small aperture, which means you need a lot of light, which means either bright daylight or large arc lamps. You also need a fast film stock to avoid burning out the highlights, and a lens whose optical design holds sharpness across the frame at small apertures. None of these conditions were standard in the late 1930s. Toland built his career around assembling them.
He worked at the Goldwyn studio under a contract that gave him unusual freedom to experiment with his own equipment. He shot tests with Eastman Plus-X, which became his preferred negative stock once it could be pushed past its rated speed. He used coated lenses, then a recent innovation, to reduce internal reflection that would have killed the contrast at f/8 and below. He documented his methods in American Cinematographer in two articles, one in 1939 on Wuthering Heights and a longer one in February 1941 on Kane itself, the latter titled "Realism for Citizen Kane."
The "accident" of my title is a small one. Toland told colleagues that he had stumbled on the wide-and-deep look while trying to solve a different problem, namely the difficulty of using the relatively static cameras of the period to follow performances across a room. If you can keep everything in focus, you do not have to pull focus during a take, and you can let the actors block the scene without being chained to a focus mark. The aesthetic that we now read as a deliberate stylistic statement about depth, simultaneity, and the ambiguity of the image was, on Toland's account, a workaround.
This does not diminish what Citizen Kane does with the technique. Welles and Toland use deep focus there for purposes Wyler and Ford did not. The shot of the young Charles Foster Kane through the window, his mother in the foreground signing him away, the father at middle distance arguing, is a scene that depends on the simultaneity. These shots were possible because Toland had spent five years working out how to make them.
David Bordwell's On the History of Film Style (1997) is the best single account. The American Cinematographer archive has Toland's own articles, in their original 1939 and 1941 issues, available now in PDF.