About

Glass Bodies is a variant-level encyclopedia of vintage lenses and cameras. Not one page per model — one page per family, going all the way down: every variant, its serial ranges, how to tell one from another in a blurry auction photo, what reviewers across languages actually thought, the history of how it came to exist, and where one is for sale right now. This site grew out of a love of finding new rabbit holes — discovering lenses and systems I didn't know existed, and following each one further down than any single source would take me.

Version one covers the Nikon rangefinder system in full depth: 46 S-mount lens families from the 5cm f/1.1 to the 50cm f/5, and the bodies from the Nikon I of 1948 to the 2005 reproductions — every family known to exist in the mount, and a few the reference books miss.

How it's built

Every claim traces to a source a reader can check, and the page says so when it can't: figures that sources dispute are shown as disputes, numbers I computed myself are labelled as derived, thin evidence carries a flag rather than a confident sentence, and a handful of findings this project worked out itself are marked as exactly that. Original prices are shown as listed at the time, in the currency they were listed in, alongside what that money is worth today. History sections are AI-authored and say so; the identification data they sit beside is checked against published sources and photographed examples.

Live Listings are read from the public catalogs of the shops and auction houses I track — independent dealers on four continents, not just the big marketplaces.

Our own findings

A few claims here are not repeated from anywhere — they came out of comparing sources against each other, or against photographed examples, and finding something nobody had written down. Those carry an our own finding tag on the page.

The bar is deliberately narrow. Noticing that two published facts connect is ordinary work and belongs in the prose; the tag is for a claim that had to be inferred from the evidence, and that a better source could overturn tomorrow. There are 69 of them so far.

• AF DC-Nikkor 105mm f/2D

  Nikon credits the 135mm f/2 DC's improved longitudinal chromatic aberration to the cemented front group in its seven-element, six-group layout. This lens has six elements in six groups, which by definition means nothing is cemented, and it is the lens on which reviewers report strong longitudinal chromatic aberration at full aperture. Connecting the two is this entry's inference from the published construction and the published reviews; no source states it.

• Bessa R2S / R2C

  The flange-depth account of why Nikon and Contax lenses cross-mount but do not hold focus cannot be the whole explanation, because a fixed 0.31mm offset at an identical thread pitch would put focus out by a constant amount at every distance, where every published description of these two bodies has them correct at infinity and drifting progressively worse the closer you focus — which is the signature of a difference in helicoid travel, not of a difference in flange depth

• Biotar 7.5cm f/1.5

  Three readings here are this page's own. The Leitz Auction lens no. 3170182 is read against the 1945–49 serial band as a pre-batch test lens; the auction house says "possibly a lens for testing purposes" and no more. The first postwar batch is closed at 3467850 by adding 100 to the published opening number, and the two known lenses inside it (3467776 and 3467786) are offered as confirmation of the block; no source prints the closing number. And the conclusion that the Contax lens kept the slim barrel into 1954 rests on one lens Bischoff describes, set against Morrison's version boundaries.

• Contax IIa

  That Camerapedia's black-dial and colour-dial figures sum to 69,025 for the IIa and IIIa together, within twenty-five of Ghisetti's separate figure for the IIIa alone — an arithmetic coincidence noticed here, recorded in the production dispute above as a caution and not as a finding about either source.

• Contax TVS III

  The proposal that the widely printed 2000 introduction date is an artifact of two things being sold at once — a camera launched in Japan in 1999 and a Year 2000 commemorative version of it — is this project's own reading of the Japanese dealer sources. No source consulted makes that argument.

• ELCAN 66mm f/2

  Three findings are made here rather than taken from a source. First, the arithmetic against the production estimate: every serial that has surfaced in any catalogue, dealer listing or first-hand account falls between 0003 and 0028, and no source sets that span beside the ~200 figure it repeats. Second, the provenance of the nine-element figure — traced through the catalogues to a single WestLicht sentence of March 2018 citing Lager II page 228, reprinted verbatim by two later catalogues and reused by the same house under its later name, with the two handlers outside that chain giving no count at all. Third, that the 1972-versus-c.1970 dating conflict is internal to one auction house, which catalogued the same lens, 283-0028, both ways six years apart.

• F-Distagon 30mm f/3.5

  That the 9-element count some sources give for the C and CF versions is Zeiss's 8-element figure plus the mandatory filter glass. All three Zeiss data sheets print 8 elements in 7 groups without change, and all three say the filter "is part of the optical system" and must always be fitted, so a count that includes it arrives at 9 — no source consulted states this reconciliation, and it is offered as the likeliest reading of the discrepancy rather than as a fact.

• Fisheye-Nikkor 6mm f/2.8

  The introduction date given as 1970 in most English sources is the date of the prototype, not of the lens. Cavina documents serial 261001 at Photokina in October 1970 and the first production serial, 628001, in March 1972; Vink's serial dating agrees. The two dates have been collapsed into one in general circulation.

• GN Auto Nikkor 45mm f/2.8

  The reconciliation of the two launch dates is this project's, not a source's. Nikon dates the release to March 1969 and the end of production to 1977, and calls that "a sales life of approximately nine years" — a span that fits Vink's August 1968 production start better than it fits Nikon's own March 1969 release date. That is an argument for 1968 being when the lenses were built and 1969 when they went on sale, not proof of it.

• Komura 105mm f/2.8 · Komura 135mm f/3.5 · Komura 135mm f/2.8

  The bezel engraving dates a Komura — earlier lenses read the bare focal length over Sankyo Kōki Tokyo, later ones f= over Sankyo Kōki Japan

• Leica 72

  The Leica Wiki's own two statements about this camera's dates cannot both be true — 1957–1963 in its production line, 1948–1950 in its serial table — and the serial numbers themselves, in the 357xxx range, match the earlier dating in the same wiki's other model tables. This site records 1948–1950 as the better-supported reading and leaves the model page's dates as the disagreement.

• Leica Ic

  The Leica Wiki's Ic page notes that camera 59478 is a black-lacquer Ic. That serial belongs to 1930 — the Leica Wiki's own tables place it among Model A and Model C production — and cannot be a camera introduced in 1949, whose blocks begin at 455001. The note is a transcription error, and this site records it as one rather than repeating it.

• Leica Ig

  The Leica Wiki's Ig serial list double-counts: its two 34-camera Siemens blocks fall inside the 1,412-camera block at 924589–926000 that the same table already counts in full. Deduplicating brings the list from 6,368 to 6,300 — still 332 above the model page's 5,968, but the smaller discrepancy is the real one.

• Leica IIf

  A two-screw accessory shoe means different things on different models. On a IIIc it is evidence of a converted IIc, because the factory IIIc shoe has four screws; on the IIf the two-screw spring-blade shoe is what the factory fitted. The same observation supports opposite conclusions depending on which body it is made about.

• Leica IIIc

  The Leica Wiki's IIIc page carries no K at all, though it lists the grey wartime bodies that occupy the very serial block the K cameras come from, 388927–397650. What the letter abbreviates is also unsettled: camera-wiki reads it as Kugellager, ball bearing; Leitz Photographica as Kältefest, cold-resistant; Leica Historica gives both.

• Leica IIId

  Separating what the Leica Wiki's IIId serial list documents (102 cameras) from what it assumes (a further 325, marked as such in its own table) — the source publishes both but only ever quotes the combined total of 427.

• Leica M4-P

  The Mount Everest edition is given as 200 cameras with serials 1586371–1586565. That range is 195 numbers inclusive, so either five cameras fall outside it or the edition is 195. The source publishes both figures without noticing they disagree.

• Leica MD

  The MD's own page carries a caution that older production counts are wrong: cameras once attributed to the MD are now assigned to the M3, following Pierre Jeandrain's research on the Postkameras. Any MD figure published before that reassignment includes M3 bodies.

• Leica MDa

  The MDa serial list prints the block 1185291–1185300, ten Postkamera 24×27 cameras of 1968, twice in identical form. Its stated total of 14,892 therefore counts those ten cameras twice; 14,882 is the deduplicated figure.

• Light Lens Lab 15mm f/8 ("Hologon")

  Two conclusions here are this project's rather than any source's. First, that the lens never shipped: no source states it, and the finding rests on a negative sweep — no product page, no entry in the maker's lens collection, no dealer listing, and no rumour-site coverage after October 2021 — set against the maker's own on-hold statement. Second, that the fault the prototype was criticised for is the original's fault: the cyan shift the reviewer reported is the same rear-element-to-sensor geometry that makes the 1972 Zeiss Hologon unusable on a digital M. The most faithful thing about the copy was its defect.

• Light Lens Lab 28mm f/2.8 "Nine Element"

  Where the copy stops. The element count is the original's and the rear geometry is not: nine in six on both lenses, but a rear group recessed by 1 to 2 mm, which is what lets the recreation meter on the digital and metered film M bodies the original cannot be used on. The shared M5 and CL exclusion is the shared limit of a swinging meter arm rather than a shared prescription. No source sets the two compatibility lists against each other, and reading the overlap as proof of a faithful rear group — which is what it looks like at a glance — gets the lens backwards.

• Light Lens Lab 50mm f/1.5 "Z21"

  The decode of the name. Angénieux's own type code for the original is S21; this lens is the Z21, and Z is the initial of Light Lens Lab's founder, Mr. Zhou. No source consulted states that substitution — the maker never writes the word Angénieux on the product page at all, and the press coverage that identifies the original does not comment on the letter. The reading is this project's, offered as inference rather than fact.

• Medical-Nikkor 120mm f/4 IF

  Vink's blocks give 23,220 of the 120mm against at least 19,217 of the 200mm f/5.6 it replaced — one of the six blocks Vink lists for the older lens carries no quantity, so its true total is higher and the two may be closer than the figures suggest. Either way the second generation matched or beat the first over a comparable seventeen years, with one unaltered version against the earlier lens's six blocks and two revisions. Vink publishes the blocks, not the comparison.

• Micro-Nikkor 105mm f/2.8

  The production totals — 94,722 for the manual-focus run, 212,416 for the AF-D, and 443,368 for the three versions together — are sums of Roland Vink's per-block quantities performed here. No published total exists for any of them, and each is a floor: Vink derives quantities from observed serial spans.

• Micro-Nikkor 200mm f/4

  The production totals — 40,143 for the two manual-focus versions together, 30,573 for the autofocus version, and 70,716 for the family — are sums of Roland Vink's per-block quantities performed here. No published total exists for any of them, and each is a floor: Vink derives quantities from observed serial spans. The observation that the Ai-S and AF-D serial blocks overlap between 200001 and 212588 is likewise read off his table rather than stated by him.

• Micro-Nikkor 55mm f/2.8

  The totals — 489,201 for the manual-focus run and 529,780 for the family including the sixteen NASA lenses and the autofocus version — are sums of Roland Vink's per-block quantities performed here. No published total exists for either, and both are floors: Vink derives quantities from observed serial spans.

• Micro-Nikkor 55mm f/3.5

  The production totals — 72,903 for the compensating Auto, 119,284 for the Micro-P, and 517,719 for the family as a whole — are sums of Roland Vink's per-block quantities performed here. No published total exists for any of them, and each is a floor rather than a count: Vink derives quantities from observed serial spans.

• Nikkor 105mm f/2.5

  The production totals — ~131,258 for the Sonnar versions, ~517,088 for the Xenotar versions, ~648,346 for the family — are sums of Roland Vink's per-block quantities performed here. No published total exists for any of them. The "about 1,500" for the first Xenotar block is likewise a serial-span calculation (408816 minus 407301), not a source's figure.

• Nikkor 135mm f/2

  Adding Vink's per-version counts, the autofocus DC lenses — 13,258 plus 32,987 — outnumber the entire manual-focus run of 44,596. The 135mm f/2 sold more copies as a lens whose optics Nikon had thrown away and replaced than as the design the family is named for. Neither source states this; it follows from the two sets of serial blocks.

• Nikkor 135mm f/2.8

  Two arithmetic points that follow from Vink's blocks and are not stated by him. The four-element K and the five-element compact K were in production simultaneously for two to three months in early 1976, in separate serial series under the same name. And the Ai block's upper end, 916935, sits above the Ai-S block's start, 900001, so the 146,935 and 76,038 figures overlap by roughly 17,000 numbers and cannot simply be added.

• Nikkor 15mm f/5.6

  The QD·C engraving is read here as an element count rather than a model code. Nikon's Latin prefixes ran U for one to D for ten and were compounded above ten: the 20mm f/3.5 UD has eleven elements, this lens marked QD has fourteen, and the 13mm prototype marked HD·C has sixteen. All three counts match Vink's optical specifications exactly, which is the evidence for reading Q plus D as four plus ten rather than as an arbitrary designation.

• Nikkor 180mm f/2.8 ED

  Nikon says the Nikkor Auto 180mm f/2.8 went to the press for the 1970 Sapporo pre-Olympics and to the public in 1971. Vink's first serial block, 312011–313379, holds 1,369 lenses and is dated June 1970 to 1971, after which a separate block opens at 320001. The size and the timing of that first block are consistent with a press-only run preceding general sale — this entry draws that connection; neither source states it.

• Nikkor 200mm f/2 IF-ED

  Two claims here are not in any source. The manual-focus total of 3,753 lenses is a sum of Vink's four per-block quantities, performed here. And the absence of an AF or AF-D 200mm f/2 is read off the gap in his table rather than stated by anyone: no source consulted remarks that the lens skipped the screw-drive autofocus era entirely.

• Nikkor 20mm f/2.8

  Summing Vink's per-block quantities gives roughly 317,500 across all four production versions — 72,472 manual and 245,039 autofocus. Vink publishes the blocks, not the total; the addition is this site's, and it inherits every uncertainty in serial-span estimates.

• Nikkor 24mm f/2

  Adding Vink's two blocks gives about 65,000 lenses across the whole run, roughly seventy-three per cent of them Ai-S. Vink publishes the blocks, not the sum; the addition is this site's.

• Nikkor 24mm f/2.8

  Vink's blocks add to about 814,300 lenses across fifty-three years — roughly 552,000 manual focus and 262,000 autofocus, with the Ai-S the single largest version at about 165,000. Vink publishes the blocks, not the sums; the addition is this site's.

• Nikkor 28mm f/2

  Adding Vink's blocks gives roughly 116,700 civilian lenses across the whole run — about 29,300 pre-Ai, 12,800 K, 30,000 Ai and 44,700 Ai-S. Vink publishes the blocks, not the sum; the addition is this site's.

• Nikkor 300mm f/2.8 ED

  The correspondence between Nikon's verbal figure for the preset lens and Vink's serial counts is drawn here, not by either source: Nikon says production reached one hundred and several tens, Vink's two blocks come to 72 and 78. The per-version sums for the AF New (12,818) and the G VR (10,884), each spanning two restarted serial series, are likewise arithmetic performed here.

• Nikkor 35mm f/1.4

  The family total of roughly 134,000 manual-focus lenses is a sum of Roland Vink's per-block quantities carried out here, not a figure any source publishes. It is only as good as his serial counts, and it excludes the three prototypes.

• Nikkor 50mm f/1.2

  The 241,752 family total and the comparison with the Noct-Nikkor's 11,413 are sums of Roland Vink's per-block quantities carried out here, not figures any source publishes.

• Nikkor 50mm f/1.4 (F mount)

  The family total of roughly 3.88 million is this project's sum of Roland Vink's per-block quantities for every version from the 1962 Nikkor-S Auto through the AF 50mm f/1.4D, excluding the optically unrelated AF-S G. Vink publishes the blocks and the per-block counts but no total; van Dijk's independent "more than 4 million" is the nearest published figure and agrees.

• Nikkor 50mm f/2 (F mount)

  The family total of roughly 1.34 million is this project's sum of Roland Vink's per-block quantities across all six versions. It is reported here alongside van Dijk's much larger published estimate rather than in place of it; the two cannot both be right and no factory figure exists to settle them.

• Nikkor 55mm f/1.2

  The total of about 178,700 is this project's sum of Roland Vink's per-block quantities. It is offered as a cross-check on van Dijk's independently derived "around 175,600," not as a replacement for it; the two figures agree to within two per cent, which is the strongest evidence available that either is roughly right.

• Nikkor 85mm f/1.4

  The 103,405 figure for the AF-D is a sum of Roland Vink's three per-block quantities performed here; no source publishes a total for that lens.

• Nikkor 85mm f/1.8

  The production totals — ~74,789 for the pre-AI blocks, ~120,642 for the whole manual-focus run, ~244,014 for the AF-D and ~355,133 for the autofocus lens in total — are sums of Roland Vink's per-block quantities performed here. No source publishes any of them.

• Nikkormat FTn

  The FTn's published serial span cannot be used as a production count. FT 3500001 to FT 4700001 is 1,200,000 numbers against published totals of 850,000 and "over one million", so the block is either unfilled or the lower figure is understated; the arithmetic is set out here because collector sources quote the span and the totals interchangeably.

• Nikon F

  That the million-plus production figures for the Nikon F are serial spans with a gap left in them. Nico van Dijk's serial matrix runs 640xxxx to 745xxxx — about 1,050,000 numbers, which is Casual Photophile's figure to the rounding — and marks 661xxxx to 669xxxx as never issued, that block having been held for the Nikon S3M. Netting out those 90,000 numbers leaves about 960,000. The subtraction is this project's; van Dijk publishes the table and the exclusion but draws no total from them.

• Nikon F4

  Nikon's history gives 230,000 F4 bodies sold by the end of 1991; van Dijk records serial 2230000 as the number reached at that same date, from a start of 2000001. The two agree to the unit, which indicates F4 serials were issued consecutively with no reserved blocks — and makes the 26xxxxx range of 1996 readable as about 600,000 bodies, which is independently what van Dijk reports.

• Nikon F55

  That the Japanese-market Nikon Us corresponds to the F55D rather than the plain F55: Nikon's specification for the Us lists the quartz date in the main table, not as an option, and gives 129 × 92 × 67.5mm and about 360g — the depth and weight English sources attach to the dated body, against 129 × 92 × 65mm and 350g for the plain one. The same pattern holds for the Nikon U and the Nikon U2.

• Nikon F65

  That the Japanese-market Nikon U corresponds to the F65D rather than the F65: Nikon's specification for the U lists the quartz date in the main table, not as an option, and gives 139.5 × 92.5 × 68mm and about 400g — the depth and weight English sources give for the dated body, against 395g for the plain one. Nikon's 2002 agnès b. release then flags the absence of date imprinting on that model as a departure from the ordinary U, which only makes sense if the ordinary U had it.

• Nikon F75

  That the Japanese-market Nikon U2 corresponds to the F75 QD rather than the plain F75: Nikon's specification for the U2 lists the quartz date in the main table, not as an option, and its 385g is five grams above the 380g English sources give for the F75. The same pattern holds for the Nikon U and the Nikon Us. The inference has since been confirmed from the other end — Nikon's own N75 manual gives 380g for the plain body and 385g for the N75 QD — so the five grams are the date back, and the U2's figure is a dated body's.

• Nikon M

  That the four published production figures for the Nikon M are not in conflict but count different things — about 1,600 for the unsynchronised camera alone and about 3,200 for unsynchronised and synchronised together — and that Mike Eckman's 1,649 is therefore an unsynchronised figure presented as a total. No source states the reconciliation; the serial blocks support it, spanning some 1,703 numbers to the end of the unsynchronised range and about 1,538 more to the end of the M prefix.

• Nikon S

  The 1,619-camera gap between Pacific Rim Camera's 35,127 and the 36,746 Nippon Kogaku's own book gives is within 24 of the 1,643 Nikon M bodies ja.wikipedia records, which points at M-prefixed cameras being the difference. Suggestive only: the Nikon M total counts unsynchronised bodies, not the synchronised MS subset, and no source states the relationship.

• Nikon S2

  That 11,715 is probably not an independent count of black-dial cameras. It differs from the 56,715 total by exactly 45,000 — too round a number to be coincidence. It looks as though someone assumed 45,000 chrome-dial cameras, worked out the black-dial figure by subtracting, and everyone since has repeated it as though it had been counted. If that is what happened, the chrome-dial and black-dial numbers are not two findings that agree; they are one number and its remainder.

• Nikon S3

  Nikon's stated intention for the life-size finder is only half met on the published figures. Effective base length — the distance between the two rangefinder windows multiplied by finder magnification — is what sets how precisely a rangefinder can be focused, and the S3's 60mm base at 1.0× gives 60mm, the same figure the SP already had. The Leica M3's 69.25mm base at 0.92× gives 63.7mm, so the magnification half of Nikon's claim holds while Leica's longer base absorbs it. Against the M-series finders that came later the S3 does win: the common 0.72× M6 manages 49.9mm on the same 69.25mm base. No source performs any of these multiplications.

• Nikon S3M

  Where the S-72's serial numbers sit. No source publishes a serial range or a production total for the motor, but the two numbers on record — 95025 with the chrome 6600156 and 95155 with the black Leitz camera — fall at and just above the top of the published S-36 range of 94001 to about 950xx, which is consistent with the S-72 continuing the same motor numbering rather than opening a block of its own. Nobody publishes that connection.

• Nikon S4

  The published serial block, 6500000 to 6505900, spans 5,901 numbers — the start number plus its author's own rounded production figure of about 5,900. The ceiling therefore looks calculated rather than observed, which would explain why the same author's S4 page reports two bodies above it, #6505922 and the 6505928 last reported to the Nikon Historical Society. Nobody performs that subtraction.

• Nikon SP

  The black-paint total in circulation does not survive its own evidence — Japanese Wikipedia gives 1,500–1,600 cameras and then lists twelve serial blocks that sum to 1,271, a shortfall of 229 to 329

• Noct-Nikkor 58mm f/1.2

  The total of about 11,415 is this project's sum of Roland Vink's per-block quantities for the prototype, Ai and Ai-S blocks. It is set out beside van Dijk's "some 10,000" rather than in place of it, since the two published figures differ by around fifteen per cent and neither is sourced to Nikon.

• PC-Nikkor 28mm f/3.5

  Vink's blocks put 35,058 of this lens against 7,457 of the 28mm f/4 it replaced — the successor outsold its predecessor by roughly four and a half to one, on a run five times as long. Vink publishes the blocks, not the comparison.

• PC-Nikkor 35mm f/2.8

  Vink's blocks add to 63,462 lenses across thirty-one years, of which 26,671 — about forty-two per cent — are the 1980 barrel alone. Vink publishes the blocks, not the sum. Separately, all four of Nikon's stated release dates for this family fall four to six months after the serial block Vink opens for the same version. That is consistent with production preceding sale, but no source states it.

• Reflex-Nikkor 1000mm f/11

  The family total of 16,760, and the 12,539 for the Ai version across its two serial series, are sums of Vink's per-block quantities performed here; no published total exists for either. Setting those against the 1000mm f/6.3's 56 and 52 is likewise a comparison drawn here from two rows of his table, not a claim any source makes.

• Reflex-Nikkor 500mm f/8

  Three claims here are not in any source. The reading of Vink's 55,631 as a total for both ·C blocks rather than the second alone, from the arithmetic 585631 − 530001. The family total of 104,156, and its split into 72,094 old-design and 32,062 new-design lenses. And the serial overlap at 183201–183318 between the last Reflex-Nikkor 500mm f/5 and the first 500mm f/8 New, which follows from putting two rows of Vink's table side by side and which he does not remark on.

• Rollei 35

  camera-wiki's Singapore figure of 215,000 for the original Rollei 35 is English Wikipedia's '185,000 plus 30,000 with S-Xenar' added up: the Schneider-lensed cameras are inside the Singapore total, not beside it, which puts the model at 527,000. Neither source says so, and neither prints 527,000 — nor, contrary to what circulates in English write-ups, 497,000.

• Summar 5cm f/2

  The Leica Wiki's Summar serial table prints the block 186001–187000 as "Summar 5 cm 1:3.5". No Summar was made at f/3.5 — that is the Elmar's aperture — and the same table renders f/2 as "1:20" in the rows either side, so this is a typing error rather than a lost variant.

• Summarit 5cm f/1.5

  The Summarit's serial list opens with 103 lenses dated 1939 at 491898–492000 — a decade before the model existed, and inside the block the Xenon's own page claims for its final 1939 run. Both pages count the same hundred numbers. The likeliest reading is that these are Xenons, or transitional lenses, rather than Summarits built ten years early.

• Tele-Tessar 350mm f/4 IF

  That the shutterless Tele-Tessar 4/350 was introduced between January 1980 and 1985: it is absent from the five-item 'F Lenses w/o Shutter for 2000FC' section of Victor Hasselblad Inc.'s dealer price list effective 1 January 1980, and present as product number 20486 in the 1985–86 Hasselblad product catalogue. No source consulted states an introduction year at all.

• Tele-Tessar 350mm f/5.6

  That the 1982 CF rebarrelling left the optics untouched: the C and CF data sheets print identical entrance-pupil position and diameter (325.5 mm / 59.4 mm), identical exit pupil (13.2 mm / 24.2 mm), identical principal planes (H 173.6 mm, H' 47.5 mm) and an identical 168.1 mm between the first and last lens vertices. Neither sheet says the design was carried over; the numbers only make sense if it was.

• Xenon 5cm f/1.5

  The Xenon's last serial block, 490001–492000 in 1939, contains the Summarit's first: 103 lenses at 491898–492000 appear on the Summarit's own list. The two lenses are claimed to share the same hundred numbers, which neither page notices.

What's coming

More systems, in the same depth:

• Medium format beyond Hasselblad — 120 and its frame sizes, 6×4.5 through 6×9: Rollei, Mamiya, Pentax, Bronica
• More SLRs — the manual-focus systems beside the Nikon F: Canon FD, Pentax, Olympus OM, Minolta, Contax/Yashica
• More rangefinders — the Zeiss Ikon ZM and the Rollei 35 RF, the two Cosina-built bodies not yet here

And more site around them:

• Marketplace
• Store
• Feature requests — a proper channel for telling me what to build

Spotted an error, or own an example that proves a page wrong? The corrections invitation at the bottom of every family page is the way in — the catalog gets better one contradicted sentence at a time. [email protected]