Super-35 is not only 3-perf (even the Wiki link that David posted shows that Super-35 is either 4-perf or 3-perf) -- it's been 4-perf for most of its existence as a format! People only started shooting 3-perf full aperture for theatrical movies in the early 2000's, though it was being used for TV before that. Super-35 means shooting full aperture, meaning exposing picture information out to each perforation row, not leaving room for an optical soundtrack.
Super-35 started out with the movie "Greystoke" (1984) and was 4-perf. Back then it was called "Super Techniscope" because John Alcott felt it was an improvement over 2-perf Techniscope, allowing him to shoot "Greystoke" with spherical lenses for a blow-up to 4-perf 35mm anamorphic and 5-perf 65mm.
After "Greystoke", you had "Top Gun" and "Silverado", all made using the Super Techniscope moniker. But since Techniscope was a system developed by Technicolor Labs, the more generic term "Super-35" replaced it by the mid 1980's.
Again, these were all 4-perf Super-35 movies. Even most of Roger Deakin's Super-35 movies have been 4-perf.
As far as the negative aperture, some early Super-35 movies used a hard matte in the camera gate (something easy to do with Panaflexes for example) so that they weren't exposing a full 1.33 : 1 negative, feeling that this was too hard to keep clear of equipment if composing for a crop to 2.40. So movies like "Greystoke" and "Terminator 2" have a hard matted negative, matted to something near 1.50 : 1 or so, then composed for cropping to 2.40.
But even this practice soon was dropped as being unnecessary, and many visual effects people wanted the cinematographer to shoot and expose the whole negative area in case they could use that information.
Super-35 was not really a new idea though in 1984. In the 1950's, there was a process called SuperScope where a 2:1 image was composed within the 1.37 Academy area of the 4-perf 35mm negative, the idea being that 2:1 was about the most you could get away with cropping Academy. The 2:1 area was blown-up and squeezed to 4-perf 35mm anamorphic (CinemaScope) with a hard matte on the sides to keep the image 2:1 instead of 2.35 : 1. In the last few years of SuperScope's use in the late 1950's, they were experimenting with shooting 4-perf 35mm full aperture, not Academy, and composing for cropping to 2.35 : 1. Some people called this process "SuperScope-235" but it's unclear if any actual features were shot in this process. By this time, low-budget producers started just cropping Academy to 1.85 and not spending the money to blow-up SuperScope to CinemaScope. And then by the early 1960's, 2-perf Techniscope was invented.
This is some history of 3-perf that I gave to Mark Mueller at Arri that they used on their website:
For more information about 3 perf and a historical perspective, here are some articles (Thanks to David Mullen, a Cinematographer from LA, for sharing his research):
* In the Sept. 1973 issue of the SMPTE Journal, pg. 742, is an article titled "A Universal Format for Film Production" by Bernstein, Wysotsky, and Konoplev. It is a translated reprint of an article in Tekhnika i Kino Televideniya of January1973. It recommends shooting in 3-perf full-aperture as a way of generating either 1.85, 2.35 anamorphic, or 4:3 TV versions from the same negative.
* In the March 1975 American Cinematographer magazine is an article by Kenn Davis proposing a new lens system with a 25% anamorphic squeeze, in order to get a 1.85:1 image onto a regular 16mm frame, and a 2.35:1 image onto a 3 perf 35mm frame. He then goes on to explain other advantages of a 3 perf system, including for 1.85 production.
* In the June 1976 American Cinematographer magazine is an article by Miklos Lente, proposing a 3 perf format called "Trilent-35". It used the Academy width but the 3-perf height.
* In the July 1986 American Cinematographer magazine is an article by Rune Ericson titled "3-Perf in the Future".
* In the February 1998 issue of "International Photographer", Vittorio Storaro proposed his "Univision" 3 perf Super35 format.
You can see from that list of articles above that even though 4-perf Super-35 started being used in 1984, 3-perf was mainly just in the proposal stage at this point.
3-perf started being used for TV as a way of saving stock, and the negative has a native 1.78 : 1 (16x9) aspect ratio, although the start of 3-perf for TV predated the transition to HDTV and most shows were composed for 4x3 inside of 16x9. I think "Max Headroom" (1987) might have been one of the first 3-perf shows, and one of the first U.S. shows to shoot film at 30 fps (3-perf was being used to compensate for the increased frame rate in terms of stock usage). TV shows could shoot 3-perf because it was all for telecine transfer only; features avoided 3-perf at the time because it not only required an optical printer blow-up to 4-perf for release printing (as did 4-perf Super-35) but negative cutters had a hard time with the edge code system in matching to 3-perf EDL's. It wasn't until D.I.'s became popular in the mid 2000's that most of the reasons to avoid 3-perf for Super-35 went away.