Nikkor 24mm f/2.8

Nikon spent almost ten years between the 2.8cm f/3.5 of March 1960 and a 24mm lens, and Kouichi Ohshita’s account of why is the most useful thing written about this design. A retrofocus lens shorter than 28mm was, in his words, uncharted territory; the optical design report from those years is thick with trial data because there was no fast computer to trace rays through a wide field, and the arrangement that would hold performance across that field was still being found by hand.

The obstacle that mattered was not resolution at infinity but what happened at three feet. Retrofocus wide-angles generate astigmatism as they focus close, and the corners go soft. Nikon’s answer on the 2.8cm f/3.5 had been to forbid the problem — the lens stops focusing at 60cm. Competitors’ 28mm lenses focused nearer, customers noticed, and Wakimoto Zenji would not fix it the easy way. Racking the whole lens further would shorten the minimum distance and ruin the picture taken at it, and every super-wide Nikon had yet to design would inherit the fault. What he proposed instead was to split the lens between the sixth and seventh elements and let that gap close as the barrel focuses. The separation is computed so that changing it barely touches spherical aberration and moves astigmatism a great deal; the astigmatism generated by focusing close is cancelled by the astigmatism generated by the gap. Shimizu Yoshiyuki did the basic design, Wakimoto supplied the mechanism, and the NIKKOR-N Auto 24mm f/2.8 reached the market in June 1967 as the first lens anywhere to carry close-range correction. The figures engraved on the focus ring — 0.3m, 1 foot — are the whole point of the lens.

The versions divide into two computations rather than the usual one. Everything from 1967 to 1977 is nine elements in seven groups: the Nippon Kogaku barrel, then the Nikon barrel of 1971, then the multicoated N·C of 1972, then the K of 1975 with seven blades and f/22 in place of six and f/16. That is 227,289 lenses on the original computation, by Vink’s arithmetic. Then in January 1977 the Ai opened a new serial block at 525001 with nine elements in nine groups — both cemented pairs separated — and nothing on the outside of the barrel says so. Ohshita, writing about the older lens, mentions in passing that his own Ai sample has better field flatness and that its elements are the same ones in the AF Nikkor 24mm f/2.8D. That is the only confirmation Nikon has published that the two designs differ, and it appears inside an essay about a different version.

After 1977 the story is mounts, not glass. The Ai-S of August 1981 halves the focus throw to 80°, moves the depth-of-field marks to a chrome ring, and picks up Super Integrated Coating partway through its run. The AF of September 1986 puts a screwdriver coupling and a plastic focus ring on the same computation; the AF New of 1991 swaps the twist aperture lock for a slide; the AF-D of April 1994 adds distance reporting. The Ai-S and the AF-D both ran until 2020, thirty-nine and twenty-six years respectively, and Vink’s serial spans put the whole family at about 814,300 lenses.

What is missing is the factory record. Nikon has published no production figure for any version, no cross-section of the 1977 redesign, no dated price for the 1967 lens — the ¥70,000 on Nikon’s own page belongs to the Ai-S — and no statement of why the computation changed when it did. Rørslett, who rates the lens highly and warns that its lateral colour becomes expensive on a high-resolution sensor, dates close-range correction to 1968; Nikon and Vink both say 1967. The lens that solved close focus for wide-angle photography has left almost no paperwork behind it.