Findings

Every checkable claim this directory makes, numbered, dated and statused, with what would overturn each one, and the analysis drawn from the register

The material on this page is drawn from the Global Cat Health Survey 2026, which is where its sources and its declarations of interest are set out in full.

The recommendation this page's evidence produces, and the five others on where to spend a test, are on Recommendations.

The findings

A finding is a claim about the evidence that names what would overturn it.

Where the register stands, in one paragraph. It holds 1044 cats that carry a dominant eye code or are recorded blue, odd or latent. 219 of them descend by documented parent links from one of 6 founders, and 825 do not yet connect to any. 29 animals have been genotyped. The lines collide in 15 animals, across 4 distinct pairings. Every one of those figures is a floor: the register is bounded by the documents supplied to it, and a registry code is not a carrier test.

19 open, meaning the question behind the finding is still live, and 1 settled by a document or a test. What the two words mean is set out on the method page.

F35. Every reported outcome of a cross between two dominant blue eye determinants in this register was deafness, and neither case is confirmed by a measurement OPEN

Entered 30 August 2026

Two crossings of two determinants in this register have a reported outcome, and both produced deaf kittens. The first is published: a latent dam carrying DBE^ALT bred to a DBE^CEL sire, and the two kittens genotyped as compound heterozygous were white and deaf. The second was reported to the survey on 30 August 2026: Seymour (DBE-ALT) x Igor (DBE, variant not identified), made deliberately by a breeder who held both lines, and both daughters of it were born deaf.

The two have the same shape, which is why the second is recorded rather than filed as an anecdote. In both, the dam showed no blue eye and carried; in both, the sire was of a second line; in both, the affected kittens were the first animals in which the two determinants met. The 2026 dam stands six generations below Nikita'l Seymour by unbroken parent links on stamped certificates.

A single copy of one variant does not do this. Nine genotyped heterozygous carriers of DBE^ALT were BAER tested at Glasgow and all nine heard normally in both ears. The reported harm in the survey is confined to animals standing below two determinants, and 114 records in this register stand below two, of which 27 have one determinant identified and one that nobody has sequenced. For that group no panel now sold can predict the mating, because there is no test for the second determinant to require.

And the mechanism carries an inference about the Igor determinant that is worth more than the deafness itself. A compound heterozygote is two different alleles AT ONE LOCUS. The harm the literature describes in these cats is the PAX3 compound: the two white deaf kittens of the published litter carried DBE^CEL and DBE^ALT, both PAX3 variants, one on each chromosome. A cross to a second line that produces the same outcome is therefore evidence that the second determinant is also at PAX3, because a variant at some other locus would give a double heterozygote at two loci and not a compound at one. All six identified determinants are PAX3 variants, so the prior is already high and this cross is consistent with it.

What settles it is smaller and cheaper than the survey has been asking for. The recommendation has been to sequence the Igor line, which means a research group and a whole genome. If the determinant is at PAX3 then the first step is a PAX3 screen on one Igor-line cat, which is a targeted read of a single gene and not a discovery project, and which a variant panel does not perform: a panel reports whether a cat carries one of the variants already named, and returns clear on a seventh. The informative animals are named on this record: the sire, the dam and either daughter of the 2026 cross. A PAX3 read on the sire alone would say whether the Igor eye is a seventh PAX3 variant, and one on a daughter would say whether she is a compound.

And the daughters look like what the literature describes. In the published litter the two kittens genotyped compound heterozygous for DBE^CEL and DBE^ALT were white and deaf, and this register already records that a double dose at PAX3 gives an animal white or nearly white and at risk of deafness. Solomiya Josephine is white, blue eyed and deaf, and so was her sister. Her certificate codes her 02, harlequin, which is the nearest white-spotting figure a registrar has for a white kitten, but a harlequin carries colour on head and tail and she has none; neither parent is white and neither is deaf, and two low-grade white-spotted parents throwing the same extreme result twice, with deafness in both, is not what white spotting predicts. The compound predicts it, and predicts it in every kitten that inherits both.

What would overturn it. A PAX3 screen on an Igor-line cat returning no variant in that gene; a BAER recording on either 2026 daughter showing normal hearing; or a reported two-determinant cross whose kittens hear. What is missing is the genotype and nothing else: no laboratory result exists for either daughter or either parent, and the Igor determinant has never been sequenced at all.

The record: Solomiya Josephine, the second daughter of the two-line cross, not named by the source, King_Pride Linda, Kittystar Kyzya

F34. Every Maine Coon in this material carrying either variant descends from the one 2016 Ukrainian mating, and outside it the two variants meet in no Maine Coon at all OPEN

Entered 27 August 2026

Counted straight down from the founders, the two variants look almost like one population. 103 cats descend from Nikita'l Seymour, the DBE-ALT founder, and 131 from Roxi, the DBE-CEL founder. 84 carry both labels, and 58 of them are Maine Coons.

A descent count cannot separate two lines that have already met. Nikita'l Markiss is the animal in which the two crossed, so every cat below him inherits both labels whatever else is true of it. The walk is therefore run a second time refusing to pass through him, and what survives is the part of each line reached by a route that never crossed.

Run that way, the Maine Coon vanishes from both lines completely. DBE-ALT falls from 103 cats to 45 and DBE-CEL from 131 to 73, and the Maine Coon count in both falls from 58 to 0. Not one Maine Coon in this register carrying either variant reaches it by any route other than through Markiss. The whole Maine Coon involvement in both variants is a consequence of a single mating made in Ukraine in March 2016.

What each line did on its own is what is left. 26 cats still carry both labels without passing through the crossing, and they sit in these breed groups: Celeste (13), German Longhair (7), Siberian (6). Those are the descendants of the second crossing, Kaena over Lazuli in France four months later, which reaches no Maine Coon.

What would overturn it. One Maine Coon carrying DBE-ALT or DBE-CEL whose descent does not pass through Markiss. The figures are a floor: the register carries records of one name entered from more than one chart, which a descent count reads as more than one animal until the two are shown to be the same.

The record: Nikita'l Markiss [Cherkassy, UA], Nikita'l Seymour, Roxi

F33. Reading the pedigrees upward as well as downward finds a variant or line standing above cats that carry none on their own record OPEN

Entered 26 August 2026

The Aerlin case generalises. Where a cattery's cats descend from animals of a named variant or line, and none of the cattery's own cats is recorded as carrying it, the variant may be in that programme unrecorded. It may equally have been lost, or never have been there at all. The register can only say where the question arises. Walking the recorded parent links upward from every animal, it arises for 421 cats, in 77 cattery prefixes and in 79 cattery and breed group pairings.

The breed group is counted separately, and this is a rule and not a refinement. Where a cattery holds two breeds that do not interbreed it is holding two populations, and a variant in one of them is not standing over the other. Belle Ayr holds British Shorthairs carrying DBE^ALT and Maine Coons descending from DBE^RE, and no mating in that cattery could bring the two together, so the two breeds are counted apart and only the Maine Coon side appears below.

The largest cases, by how many attributions stand above them.

64 further cases, each with fewer attributions standing above it.

Descent puts an allele within reach; it does not put it in an animal.

• The variant is in the cattery and was never written down. A pedigree records ancestors, not the alleles they carried. A cattery can hold a variant through many generations while every certificate it issues stays silent, because no document has a field for it and no cat was tested.
• The variant was in the line and has since been lost. A heterozygote passes the allele to half its offspring, so a few generations of breeding away from it can carry it out of a line altogether. The ancestor is still on the pedigree; the allele is gone.
• The attribution above the cattery is itself untested. Where the ancestor was given its variant by line tracing rather than by a laboratory, the row rests on an inference, and the ancestor may never have carried what it is credited with.

What would settle a case: a PAX3 panel on one cat of the prefix and breed, and, where the line above it has never been sequenced, a genotype on any animal of that line first, since otherwise there is nothing to match against.

F32. Every cat sorted into five states of knowledge about its variant, and fewest cats fall into the state where nobody can name one OPEN

Entered 26 August 2026

The register was built because a large body of blue-eyed cats sits under names nobody has sequenced. Every cat carrying a variant, a founder line or a descent falls into one of five states, and only the last is a cat about whose variant nothing can be said at all.

That last group is 163 cats in 17 lines. Against the 1392 cats this classification covers, that is the part of the record on which the survey can name nothing at all, and it is the smallest of the five groups rather than the largest.

What the narrowing is made of, and it is not new testing. The reduction comes from three things the register does with paper alone: reading a line's variant off descendants who were tested, so a founder who will never be sampled is still identified; separating a line named in the published work from a label the survey raised itself; and refusing to count the same variant twice where a line and a variant are two names for it.

What would change these figures. A genotype in any of those unnamed lines that matches none of the six named variants. That would make the line a seventh variant and take it out of the group, and testing is the only route by which the group gets smaller. A genotype matching one of the six would take it out too, by folding the line into a variant already named.

It is 70 cats showing a dominant blue or odd eye whose variant nobody can name, and a small group is not a settled one.

F31. Onyx Magic Smile: one name that looked like two ancestries, and the dam link it was blocking SETTLED

Entered 26 August 2026

The name stood on four records and on two apparently different sets of parents, so a dam link that three documents call for was left unmade. The FBU pedigree certificate for Kittystar Brilliant names Onyx Magic Smile as the dam of BG*Amur Dream (Igor line, unsequenced), whose sire on the same document is Igor Azure Dream.

Three of the four were one cat entered from three charts, and they are now one record. The fourth ancestry is not a second cat either. The certificate gives her out of Maximilian von der Jochhoeh by Amore Mio vom Grutholz (Igor line, unsequenced). Both animals are in this register, and both stand exactly one generation above the parents the record already carried:

• Maximilian von der Jochhoeh is the sire of Lord Creme Jonas vom Wernerwald (Igor line, unsequenced), who is her sire.
• Amore Mio vom Grutholz (Igor line, unsequenced) is the sire of Ice Lady v Peo's Cat (Igor line, unsequenced), who is her dam.

One grandsire on each side. That is what the grandparent row of a four-generation chart looks like when it is read one row out of place. The certificate agrees with the register rather than contradicting it.

The dam link is made on that reading, and BG*Amur Dream gains his dam and the whole Lord Creme Jonas ancestry above her. What it rests on: the two certificate names matching the two grandsires, one on each side, and nothing else. What would overturn it: a registration number for Onyx Magic Smile that does not match the cat on the Azul chart. None of the sources carries one, which is why this is settled on the pedigree and not on the paper.

The record: Onyx Magic Smile, BG*Amur Dream, Igor, given on the PawPeds pedigree as Igor Azure Dream and on the breeder charts simply as Igor

F30. No registry holds a carrier count for the variant a breed is recognised on OPEN

Entered 18 August 2026

The number of registered cats carrying DBE-CEL worldwide does not exist. It is not withheld. No body computes it.

No registry outside the GCCF records a genetic variant on a registration record. FIFe requires a test for the PAX3 gene rather than for a named variant, and its rule 3.5.1 places the duty to register the result on the member, which produces a paper record in each member country and no central total. TICA, CFA and the WCF record nothing genetic beyond parentage. LOOF holds no genotype for the breed it recognised.

The laboratories necessarily hold the number and none publishes it. Three currently offer a DBE-CEL test. A fourth withdrew its test on the ground that false positives could not be excluded. That such counts can be published is demonstrated elsewhere: one commercial laboratory publishes per-breed carrier counts across more than 11,000 cats, and PAX3 is not among the variants it covers.

The nearest thing to a count is a count of the breed, and it is not a count of carriers. LOOF publishes pedigrees issued per breed per year, read from its own annual figures on 18 August 2026.

103 Celeste pedigrees in three years, against 4,151 British Shorthair and 24,197 Maine Coon in 2025 alone. The breed was recognised on 1 January 2023 and issued nothing in 2022. The shorthair has trebled while the longhair has stayed flat: 10, 18, 32 against 14, 15, 14.

A registered Celeste is not necessarily a carrier. The standard forbids Celeste to Celeste and requires an outcross, so a carrier bred to a non-carrier gives about half carriers and half not, and the whole litter is registered Celeste. One is named: Uskya d'Ysorian (no variant or line recorded), born 26 June 2023 and registered RIEX, is described by her breeder as non porteuse du gene DBE-CEL. The 103 is therefore an upper bound on registered carriers in France and not a floor.

Nor can the non-carriers be subtracted by looking. DBE-CEL is latent in a proportion of carriers, so a Celeste with no blue eye may carry it and a Celeste described as clear may be clear on a laboratory result or on an eye. Where the source does not say which, this register grades it reported.

It counts pedigrees issued in a year, so it is a flow and not a population, and a cat registered in one year is not counted again in the next. It covers France only, and reaches no carrier outside the breed at all, because a cat carrying the variant without showing it is registered under whatever breed it was bred as.

What this register can say instead. It holds 24 records attributed to DBE-CEL and 1 with a laboratory result. The published genotyping remains the only verifiable floor for carriers anywhere, and the 65 carriers in one paper exceed the 46 pedigrees LOOF issued in 2025.

F29. A breed defined by a variant cannot close its stud book OPEN

Entered 18 August 2026

The Celeste is constituted by carriage of DBE-CEL. The LOOF requirement of 2019 is that every Celeste descend from Lazuli, the one animal in this register with a laboratory result for that variant.

The variant is reported lethal in the homozygote, so the breed may not be bred to itself. LOOF prohibits Celeste to Celeste mating and permits only outcrossing, to the British Shorthair, the British Longhair or a domestic cat.

Those two rules together have a consequence. Every generation must take in cats from outside the breed, and because the required outcross is to a non-carrier, every generation is by construction about half non-carrier. A population that must outcross in every generation cannot become homogeneous, and a stud book that cannot become homogeneous cannot be closed.

It is a structural result, not a criticism of the breed. It follows from defining a breed by a heterozygous variant and holds whatever the variant is called and whoever breeds it.

Three years after recognition the cats are not in the main register at all. LOOF separates the register a cat is entered in from the Livre Principal. RIEX is an auxiliary section of it, for the offspring of cats of known origin whose ancestry is not perfectly conforming, and the suffix numbers the generation holding the non-conforming mating. RF, the registre de filiation, is one LOOF states expressly does not confer the status of a pedigree cat. The founding cattery describes its own cats as encore au stade RIEX donc experimental, and of the five Celeste-programme cats this register holds a LOOF number for, three are RIEX or RIEX4 and one is RF.

The registration figures show it happening. LOOF issued 24 Celeste pedigrees in 2023, 33 in 2024 and 46 in 2025, 103 in the three years since recognition. Every one of those cats is the product of an outcross, so the breed cannot accumulate recorded ancestry of its own. Three years in, it is among the smallest breeds LOOF registers: 4,151 British Shorthair and 24,197 Maine Coon were registered in 2025 alone.

What the register can check. It holds 24 records attributed to DBE-CEL across 10 breeds, which is the other half of the same mechanism: a required outcross in every generation carries the variant out of the breed it defines.

The record: Lazuli

F28. Roxi Darlin and Roxi of the Topaz cannot be one animal on the dates recorded for them OPEN

Entered 18 August 2026

The later date is fixed by his own breeding. Roxi, the founder of the Topaz at the Nikita'l cattery in Ukraine, sired Lazuli (DBE^CEL, genotyped), born 2 April 2015 and so conceived at the end of that January. On a breeding life of eighteen months to eight years, Roxi of the Topaz was born between 2007 and 2013.

The earlier date comes from an account. A cat named Roxy Darlin is placed among the Darlin cattery's early animals at Ust-Kamenogorsk in 1997. A cat already breeding then would have been eighteen years old at that conception.

So one of exactly two things is true: either they are two animals, or the 1997 date is wrong. Neither animal holds a registration carrying a date of birth. Both dates come from breeder accounts, so choosing would mean preferring one account over another with no document behind either.

The two readings do not cost the same. If they are two animals, Roxy Darlin carries no established variant and drops out of the history of DBE-CEL entirely, and the variant has no documented origin before about 2007. If the 1997 date is wrong, they are one animal born between 2007 and 2013 and the line to that cattery survives with its date moved forward by about a decade.

One claim in the wider history turns on it. That both Kazakh variants were in one population before either breed existed rests on Fyodor and Roxy Darlin being at the Darlin cattery together in 1997. Fyodor's presence is not in question. Roxy Darlin's is exactly what is in question. The survey therefore does not date the origin of DBE-CEL earlier than about 2007.

The spellings cannot settle it. Roxi, Roksi and Roxy are transliterations of one Russian name, and the presence or absence of the Darlin suffix does not distinguish two animals: this register documents repeatedly that a cat moving between programmes is re-registered under a different breed code and appears with and without a prefix depending on the document.

What would settle whether the 1997 cat and the Topaz founder are one animal is a registration or certificate carrying a date of birth for either animal, a photograph of the 1997 cat, or the founding cattery's own records.

The record: Lazuli, Roxy Darlin, also written Roxi and Roksi

F27. The Maine Coon route to DBE-CEL is a dated chain from Roxi that does not pass through the original British Shorthair Celeste from the d'Aerlin cattery OPEN

Entered 18 August 2026

DBE-CEL reported in a Maine Coon is commonly read as evidence that Maine Coons have been outcrossed to British cats, because the Celeste is mated to British Shorthairs and British Longhairs under the terms of its own standard. The dated record does not support that reading.

The route runs from Roxi through the Topaz, in Ukraine. Roxi and Nikita'l Seymour both stood at the Nikita'l cattery, so a daughter of the one could be put to the other. Nikita'l Pheer-Ca (Blue Eyed Maine Coons Markiss line, untested), an odd-eyed daughter of Roxi, was mated to Nikita'l Seymour (DBE^ALT, by descent), giving Nikita'l Markiss on 26 March 2016, registered an experimental longhair. Markiss was mated to Celebrity (Blue Eyed Maine Coons Markiss line, untested), a Maine Coon born 20 November 2016 and not blue-eyed, and their son Blue Ridge Ynjir (Blue Eyed Maine Coons Markiss line, untested), born 8 January 2018, is the first Maine Coon of the line.

The dates settle the direction on their own. Blue-eyed Maine Coons of this line were on the ground in 2018. LOOF recognised the Celeste as a preliminary breed in 2022 and as a full breed on 1 January 2023. A Maine Coon of this line cannot have taken DBE-CEL from a breed that did not yet exist. Both took it from Roxi, by separate routes, and the naming records the later of the two.

The two branches leave Roxi by two different offspring. The British Celeste descends from Roxi through his son Lazuli (DBE^CEL, genotyped), a Topaz. The Maine Coon line descends from Roxi through his daughter Nikita'l Pheer-Ca (Blue Eyed Maine Coons Markiss line, untested). Neither route passes through the other.

The mating that produced Markiss held both variants, one parent of each, and the published genotyping settles which one passed. Ten Topaz Maine Coons returned wild type for DBE-ALT and nine returned heterozygous for DBE-CEL. That is the only route in this register where a two-variant mating has a laboratory result on the generation below it.

The risk it leaves is on the individual animal, not on the line. Nikita'l Markiss is the son of a Seymour-line sire and a Roxi-line dam, so he may carry one copy of each variant, and 57 records in this register descend from him. A cat of that descent may carry DBE-CEL, or DBE-ALT, or one of each, and one of each is a compound heterozygote. No genotype for DBE-ALT in a Maine Coon has been published, and the question can be answered for any individual by a panel test.

Source: the Blue Eyed Maine Coons genealogy, breeder-compiled, read 13 August 2026, graded reported because no cat on it carries a genotype; the PawPeds records for the four named animals; and the two 2024 papers for the genotyping. The published route, the Maine Coon lineage founded by a Topaz cat, names no individual animals.

The record: Nikita'l Seymour, Nikita'l Pheer-Ca, Nikita'l Markiss [Cherkassy, UA], Blue Ridge Ynjir

F26. Nadeya (Ermine Trace) and the Azul chart material share ancestors by two independent routes: two unsequenced lines and a DBE-ALT claim, and none of the three is tested OPEN

Entered 18 August 2026

• Nadeya, the Ermine Trace line. Published as a dominant blue eye line in its own right, with the search for its variant ongoing. Unsequenced.
• Igor (Igor line, unsequenced), the Igor Azure Dream line. Published on the same terms. Unsequenced.
• DBE^ALT, a sequenced variant, which chart C13 claims for the Ermine Trace material. This register has not adopted it for any cat on that line, because no cat on it has been genotyped.

Where this finding says the Azul chart it means the population drawn on chart C15, not the Igor founder and not DBE^ALT.

The Ermine Trace line is the population chart C13 heads NANOTIGR BLUE-EYED BRITISH SHORTHAIRS, DBE-ALT and its right-hand panel LINK TO ERMINE TRACE BLUE-EYED BRITISH. This register holds 241 records on it.

It shares 129 ancestors with the population standing on the Azul and Igor chart. The shared ancestors reach the Igor chart at two animals, Barby vom Wuerzburgerstern (the Igor line, unsequenced), BSH ny 12, and her daughter Karina vom Wuerzburgerstern (the Igor line, unsequenced), BSH ny 11 61. Both are blue-eyed. Neither carries white, white spotting or the colourpoint gene.

Barby's sire and her dam are each shared with the Ermine Trace line, and the two routes are independent of one another.

Sire side: Silver Angel's Lucjan of Lemurcat (no variant or line recorded), BSH ny 11, 6 September 2012, to Glamour MiracleBri, to Lemurcat Fleur, to Lemurcat Valentin, to Al' Zart Golden Leader, to Barby. The same animal runs the other way to Lemurcat Xantares, to Lusy-Mur Dionis, to Ermine Trace Lazurit and Ermine Trace Jeana, and so to Ermine Trace Barkhat and the cats coded 68 and 69 below him.

Dam side: Muslin Cats Patrick Peach (no variant or line recorded), BSH ny 11 33, 6 June 2012, to Julius GoldenBri, to Happy Anthill's Vittorio, to Wilma vom Wuerzburgerstern, to Barby. The same animal runs the other way to Quest GoldenBri, to Wanessa Emerald Flower, to Paula Emerald Flower, and so to the four Zojka cats the chart labels DBE-ALT.

Neither shared ancestor has been tested, and neither carries an eye code. Every step above is a recorded parent on a PawPeds pedigree.

Both directions are excluded. No cat of the Ermine Trace line is an ancestor of any cat on the Azul and Igor chart, and no cat of that chart is an ancestor of any cat of the Ermine Trace line. The relationship is collateral and carries no direction.

Barby and Karina have no recorded ancestor named Igor. They stand on the chart headed for that founder and their pedigrees hold 175 and 311 ancestors respectively, none of them Igor Azure Dream.

What would decide which variants these two populations carry. A PAX3 panel on Karina vom Wuerzburgerstern or on any cat of the Ermine Trace line. Three outcomes are possible and the finding does not choose between them: the two populations carry one variant, they carry two different unsequenced variants, or one of them carries DBE^ALT as chart C13 claims.

The record: Barby vom Wuerzburgerstern, Karina vom Wuerzburgerstern, Silver Angel's Lucjan of Lemurcat, Muslin Cats Patrick Peach, Ermine Trace Lazurit, Paula Emerald Flower

F25. Tracing the variants rather than the catteries: what each line actually reaches OPEN

Entered 18 August 2026

A cattery name is not a unit of inheritance. Each variant is traced through recorded parents and nothing else.

How a dominant allele travels. It comes down from an ancestor, so the animals that could carry it are the descendants of any ancestor of a known carrier, and the trace has three tiers: below a known carrier, where a cat showing the eye must have had it from that carrier; above one, where some ancestor transmitted it; and collateral, the other descendants of those ancestors, who could carry it and mostly have never been tested. Where a variant has a laboratory result the trace is anchored on the genotyped animals only.

One variant of the 5 reaches the Igor chart, and it is DBE-ALT. It does so through 118 shared ancestors, two of which are the halves of a single mating: Zinedin of Mystery's Dream (no variant or line recorded) over Clarinda Zolotoi Zamok (no variant or line recorded). No other identified variant connects to that chart. Finding 23 sets the chain out step by step, and finding 26 reports a second join, from a line whose variant is claimed rather than genotyped.

DBE-CEL traces upward through a single parent link. Every chart draws Lazuli, its one genotyped animal, out of an unnamed female by Roxi. The variant holds 2 ancestors and 59 collateral relatives, all of them reached through that one link, and Roxi is computed lineage genotyped DBE-CEL: his son carries it, a dominant allele must come from a parent, and the dam is drawn as a question mark.

A variant with no recorded parent on its carriers returns no ancestor, no collateral relative and no contact with any other line.

DBE-RE has the deepest ancestry and reaches nothing. 30 records, 173 ancestors above its genotyped animals and 187 collateral, and not one of them touches the Igor chart or any other variant line. Depth of pedigree is not what decides whether two lines meet.

No two identified variants share a pedigree. 5 of the six identified PAX3 variants have a carrier in this register: DBE-AGO, DBE-ALT, DBE-CEL, DBE-RE, DBE-xHW. Tested as 10 pairs, no two of their pedigree webs meet at a single animal. Every meeting the survey has found between blue-eyed populations is a meeting with the Igor variant, which is unidentified and is named after the cat Igor Azure Dream, and not a meeting between two identified variants.

The record: Zinedin of Mystery's Dream, Clarinda Zolotoi Zamok, Lazuli, Natalyland's Monaco

F24. The Aerlin holds cats of the CEL line and cats of the ALT line, and the ALT line reaches Igor OPEN

Entered 18 August 2026

The four Nanotigr cats this finding turns on were transcribed from close-up images of genealogy C13 on 13 August 2026. The holder of that chart does not recognise them, queried 18 August 2026. Until the names are checked against the chart, the finding is a reading and not a fact, and if the reading is wrong the finding falls entirely.

Three propositions, each true on its own.

1. The Aerlin holds cats attributed to the Celeste line, carrying DBE-CEL.
2. The Aerlin holds cats attributed to the Nanotigr line, carrying DBE-ALT.
3. Cats of the Nanotigr line and one cat drawn on the Igor chart, Natalyland's Monaco (the Igor line, unsequenced), have a pair of ancestors in common. The two are joined by eleven recorded parent links in a row, each one written on a pedigree the survey holds, with no gap and no step assumed. That makes them collateral relatives, cousins rather than ancestors of one another. The chain is set out step by step at finding 23 on this page.

The three together do not make three variants in this prefix. The Aerlin holds two attributions, DBE-CEL and DBE-ALT. The third proposition is a shared ancestor with a cat that a breeder chart groups with the Igor material, and a shared ancestor is not a variant: nothing in this register puts an Igor-line variant in an Aerlin cat, and Natalyland's Monaco does not carry one that anybody has identified either.

Natalyland's Monaco stands on the Igor chart and does not descend from Igor by any recorded parent. The chain therefore joins the DBE-ALT population to one chart-mate of the Igor cats, not to the Igor line. What would confirm or exclude a link: a pedigree for Natalyland's Monaco naming Igor Azure Dream in his ancestry, or a PAX3 panel on him.

It is the only prefix in the register of which the first two are both true.

The register holds 546 cattery prefixes. 30 of them hold a cat carrying a named variant on its own record. Exactly 1 holds cats carrying two different named variants, and it is the Aerlin. The cats are Muscari of Aerlin (DBE^CEL, by descent), Mirage of Aerlin (DBE^CEL, by descent), Pimprenelle D'Aerlin (DBE^CEL, by descent), Rhum Coda D'Aerlin (DBE^CEL, by descent) and Nautilus D'Aerlin on one side, and Plumette d'Aerlin (DBE^ALT, by descent) with Urist d'Aerlin (DBE^ALT, by descent) on the other. None of the seven has a laboratory result of its own, so both attributions rest on descent and on the charts the animals were entered from. What would confirm them: a PAX3 panel on one cat from each group.

Its 18 records arrive in two bodies of material that share no cat. Three stand on the Celeste chart, out of Lazuli: Muscari of Aerlin and Mirage of Aerlin, both DBE-CEL, and Mavi of Aerlin, whose variant this register withholds because her dam is labelled DBE-ALT and her sire DBE-CEL. Four stand on genealogy C13, Blue Eyed British Shorthair Line (Nanotigr line): Pamina d'Aerlin, Ti'Punch d'Aerlin, and Plumette d'Aerlin and Urist d'Aerlin, both DBE-ALT.

The chart the second group stands on also holds genotyped animals. The Aerlin cats are not among them. The Nanotigr line runs to 67 records and holds 11 animals with a laboratory result, all DBE-ALT, including Cotomodda Hubba-Bubba and the Belle Ayr cats. Not one of them is an Aerlin cat, and no Aerlin cat is joined to one by a recorded parent.

The Igor line does stand in this prefix's own ancestry, and that is worth pursuing. Four Aerlin cats, Nuances D'Automne, Okashi, Pimprenelle and Rhum Coda, each carry six recorded ancestors on the Igor line, among them Arischa from Sham (Igor line, unsequenced) and Mirabella from Sham (Igor line, unsequenced). Two more, Miyuki and Mavi (no variant or line recorded), descend from Kaena (Altai, early population line, unsequenced) of the early Altai population. Every step is a recorded parent on a pedigree the survey holds.

What that cannot show is whether the Igor variant is carried. No cat of the Igor line has been sequenced, so there is nothing for a laboratory to match an Aerlin cat against, and an ancestor on a line is not a variant in a descendant: a heterozygote passes it to half its offspring. The question is open and answerable. Sequencing any Igor-line animal would give the variant a name, and a panel on an Aerlin cat would then say whether this prefix holds a third.

No Aerlin cat stands on the chain reported at finding 23, which runs from the genotyped DBE-ALT population through Zinedin of Mystery's Dream and Clarinda Zolotoi Zamok to Natalyland's Monaco.

Searched through recorded parentage, no path runs from any Aerlin cat to a DNA-genotyped DBE-ALT cat, and none to the Igor group. The reason is flat: 8 of the 18 Aerlin records carry no parent at all. Plumette, Ti'Punch and Pamina are drawn on a chart and are not attached to a pedigree, and neither is Crocmou Kotomodda (DBE^ALT, by descent), the DBE-ALT cat of the Nanotigr line that Pamina d'Aerlin is recorded mated to. That mating is the only documented tie between the Aerlin and the Cotomodda programme, and it is a mating and not a descent.

A cattery is not an animal. Two variants under one prefix is a statement about a breeding programme, not about a gene.

Two spellings, one cattery. The English-language Celeste chart writes of Aerlin and the French Nanotigr chart writes d'Aerlin, and both are the Chatterie d'Aerlin. They are one cattery.

What would settle whether the two Aerlin groups are joined by descent rather than by a shared prefix. Four sheets: Plumette d'Aerlin, Ti'Punch d'Aerlin, Pamina d'Aerlin and Crocmou Kotomodda. Any one of them that reaches the recorded Nanotigr pedigree turns a cattery-level coincidence into a descent, and by finding 23 that descent already runs on to the Igor chart.

The record: Muscari of Aerlin, Mavi of Aerlin, Mirage of Aerlin, Plumette d'Aerlin, Urist d'Aerlin, Pamina d'Aerlin, Crocmou Kotomodda

F23. A shared ancestor between the genotyped DBE-ALT cats and one cat of the Igor chart, and what it is not OPEN

Entered 18 August 2026

This is a join to one cat that a breeder chart groups with the Igor material, and not to the Igor line. The animal at the far end of the chain, Natalyland's Monaco, stands on the Igor chart and does not descend from Igor by any recorded parent: none of his twenty-eight recorded ancestors is the founder, and Igor and Zinedin of Mystery's Dream are unrelated in both directions. Chart membership is not descent.

Finding 22 reports that no pedigree connects the Celeste cats to the Igor lines. That result does not carry over to DBE-ALT. Searched the same way, the DNA-genotyped DBE-ALT animals and the Igor chart animals do share ancestors, and the connection is a chain of recorded parent links rather than a shared cattery name.

The join is one mating. Zinedin of Mystery's Dream (no variant or line recorded), BSH ny 25, over Clarinda Zolotoi Zamok (no variant or line recorded), BSH ny 25, produced two recorded offspring, and they are full siblings:

• Alen Delon Zolotoi Zamok (no variant or line recorded) to Felicita of Coterie to Diana's Dream Union to the Belle Ayr cats that carry a laboratory result for DBE-ALT. Four generations.
• Elizabeth Zolotoy Zamok (no variant or line recorded) to Favorit-S Opium to Natalyland's Golden Oskar to Natalyland's Monaco (the Igor line, unsequenced), and on to his blue-eyed daughter Barbariska Elisa Vetsky. Four generations. Monaco is drawn on the chart headed line descended from Altai gene Igor and does not descend from Igor; that is a grouping by the chart's compiler and not a parentage.

No single document asserts this. The two descents are read off three separate PawPeds charts, for Belleayr Roxy, Belleayr Raine and Natalyland's Monaco. Remove any one of them and the connection still stands.

The arithmetic checks against the registry's own. PawPeds prints 0.781 per cent for Natalyland's Monaco. Computed here from the recorded parentage, he returns 0.7812 per cent.

The relationship is close enough to measure. Kinship between the genotyped cat Rain and Natalyland's Monaco is 0.00391. A hypothetical mating of those two would give a kitten 0.391 per cent inbred.

Not one animal on the connecting path carries white or colourpoint. Across all fifteen, from the genotyped Belle Ayr cats up through Zinedin and Clarinda and down to Barbariska Elisa Vetsky, there is no w, no 01, 02, 03 or 09, and no 33. Every code is a golden or silver coat marker: ny 25, ny 11, ny 12. This is the test of finding 19 holding along the whole of the join, so neither KIT nor the pointed gene is available as an explanation at any step.

It does not run in either direction, and that is a result rather than a gap. Both directions were tested. No Igor-chart cat is an ancestor of a genotyped DBE-ALT cat (0), and no genotyped DBE-ALT cat is an ancestor of an Igor-chart cat (0). The join is collateral: two branches off one mating, four generations down each side. The pedigrees held exclude both readings.

And the shared pair is too young to be the origin of DBE-ALT. Zinedin of Mystery's Dream was born 7 August 2011. DBE-ALT is named for the Altai and traces to Fyodor, in Kazakhstan, in 1995 or 1996. A British Shorthair born in 2011 cannot be the source of a variant already segregating fifteen years earlier.

What this does not establish. That Zinedin and Clarinda carried the variant. Two readings fit: either they transmitted it into both branches, or the blue eye in the Natalyland branch entered through one of that branch's other parents. No variant is adopted on the pair or on any cat of the path, and Natalyland's Monaco has no eye recorded anywhere: he is BSH ny 12 with no code and no icon. The blue-eyed animal is his daughter.

What would settle what the blue eye in that branch is. A genotype on Barbariska Elisa Vetsky (the Igor line, unsequenced), or on any cat of the Natalyland or Elisa Vetsky descent. She is blue-eyed, she carries no white, and she is four generations below the shared pair. One test on her says whether the blue eye in that branch is DBE-ALT. It does not answer the Igor question, because she is not shown to descend from Igor; for that, a cat with a recorded descent from the founder must be tested, and Chonky Bear's Commander Wolffe or Heavenly Eyes Cinderella from Sham would serve.

Scale. The Igor group is 436 records and the component it sits in, once these links are counted, is 2284. 16 cats in the register carry a laboratory result for DBE-ALT.

The record: Zinedin of Mystery's Dream, Clarinda Zolotoi Zamok, Alen Delon Zolotoi Zamok, Elizabeth Zolotoy Zamok, Natalyland's Monaco, Barbariska Elisa Vetsky

F22. Is there a line trace from the Celeste cats to the Igor lines? Not yet OPEN

Entered 18 August 2026

The two populations look alike and the question is whether they are related. They are not related on any pedigree this register holds. The 91 records of the Celeste group and the 436 of the Igor group sit in separate components of the parent graph: there is no path of parents and offspring between them, in either direction, at any depth, and they share no ancestor.

The same search does not come back empty on the other variant. Run against the DNA-genotyped DBE-ALT animals instead, it returns a shared ancestral pair four generations from each side. Finding 23 sets it out.

Resemblance is not descent, and this register has already had to say so once. Seimur and Roksi are photographed with closely similar silver tabby markings and the coats rule out their being father and son: neither Roksi nor Marta carries the silver marker, silver is dominant, and two non-silver parents cannot produce a silver kitten. Two blue-eyed populations looking alike is what a dominant allele at one locus produces whether the allele is shared or not, and the six known variants sit at the same gene.

The one place the two do meet is a cattery, and a cattery is not an animal. The name from Sham stands in both bodies of material. 131 records in the register carry it. On the Celeste side it is a founder: Willow, given as Asahi from Sham, a red British Longhair. On the Igor side it is the whole registered pedigree: Heavenly Eyes Ultra Black from Sham and Heavenly Eyes Cinderella from Sham are both by Jimmy Azure Dream, whose sire is Igor Azure Dream. No document held here joins the two.

There is a documented route between the programmes and it is not a Celeste route. messybeast.com records Nikita'l Seymour bred to a British Blue Shorthair, introducing the variant to the SuperQueen and from Sham catteries. Seymour is the DBE-ALT founder. Roxi, the cat DBE-CEL is traced to, stood at the same cattery and no such transfer is recorded for him.

The contemporary record says the opposite of what the resemblance suggests. Jimmy Azure Dream is drawn on the Celeste dispersal material, and LOOF refused to recognise his progeny as Celeste on the express ground that Igor was unrelated to the Altai cats that produced Lazuli. That is a registry, at the time, ruling on this exact question.

What the two new charts did settle. The Igor line now has a registered pedigree rather than a name. Igor Azure Dream is BSH n 02 61; his son Jimmy Azure Dream is BSH d 03 63; and Jimmy's two offspring here are coded 68. Only 44 records descend from him by a recorded parent link, against the 436 the register groups on his line: the line is a name on a chart for most of its length and a pedigree for five animals. All four ancestor positions above him read No information in the database on both charts, so the registered record of this line begins at him and stops.

The arithmetic checks out on the way. PawPeds prints a coefficient of inbreeding of 0.781 per cent for Heavenly Eyes Cinderella from Sham. Computed here from the recorded parentage, she returns 0.8133 per cent, and the single repeated ancestor the figure implies is identifiable: Jim Bob van Steffina's Garden (no variant or line recorded), four generations above her on each side.

What would settle whether the Celeste and Igor populations are joined. One document: the pedigree of Willow, Asahi from Sham. He is the only named animal standing on the Celeste chart and carrying the prefix through which the Igor line runs. If his ancestry meets Igor Azure Dream's, the two populations are joined and this finding is overturned. If it does not, the resemblance is convergence at one gene and nothing more. Failing that, a genotype on any Igor-line cat would name the variant directly and end the question without a pedigree at all.

The record: Igor, given on the PawPeds pedigree as Igor Azure Dream and on the breeder charts simply as Igor, Jimmy Azure Dream, Asahi Red from Sham, called Willow on the Celeste charts, Nikita'l Daisy, Heavenly Eyes Ultra Black from Sham, Heavenly Eyes Cinderella from Sham

F21. Two pedigrees for DNA-genotyped cats, and a check on this site's arithmetic OPEN

Entered 17 August 2026

Belle Ayr Olympas and Ermine Trace Nadeya both carry a laboratory result and until today neither held a single recorded parent. Their PawPeds charts add 139 ancestors and take the register from 1549 records with a recorded parent to a larger figure across 4225.

The arithmetic on this site was checked against the registry's own. PawPeds prints a coefficient of inbreeding on each chart. It gives 0.0 per cent for Belle Ayr Olympas and 1.56 per cent for Ermine Trace Nadeya. Computed independently here from the recorded parentage by Wright's coefficient, the same two cats return 0.0 per cent and 1.5625 per cent. THE FIGURES AGREE, which is the first external check the survey has been able to run on its own inbreeding calculation.

Where Nadeya's figure comes from. Two generations above her, Schoenweg Mona Lisa returns 25 per cent: her sire Schoenweg Dorian and her dam Passimilla's Upper Class are son and mother. That one mating is the whole of the inbreeding in the pedigree, and it sits directly above a DNA-genotyped animal.

A genotyped cat sired by a genotyped cat. The sire of Belle Ayr Olympas is Cotomodda Hubba-Bubba, himself DNA genotyped DBE-ALT. Above him the sire line runs Roddiffer Afina at n 69, Vanguard Camelia My Goddess at n 69, Cotomodda A'Effi Farrow at o 61 and Hubba-Bubba at n 25 61. Five generations of coded eyes ending in two laboratory results, with the 69 codes standing two generations above the 61 codes in one family.

Not one of the thirty ancestors of Belle Ayr Olympas carries white. No w and no 01, 02, 03 or 09 appears anywhere on that chart. This is the test of finding 19 holding across a whole pedigree rather than a single animal.

The 68 code stops at Ermine Trace Nadeya. All thirty of her recorded ancestors are coded 11, 12, 21 or 33, and not one carries an eye code of any kind. She also carries 09 white spotting that no ancestor carries. Four generations of pedigree contain no coded eye, so this chart does not push the Nadeya line further back; it shows how far back the record is silent.

What her chart does not rest on. The only laboratory result for her is a colourpoint test at the C locus, which is not a variant test. Her grade is reported, and the whole of this ancestry, thirty-one records, now rests on no genotype at all.

What would overturn it. Nothing in the arithmetic; both charts state four complete generations of four. Deeper generations would raise both coefficients, because every figure here is a floor computed at the depth the register holds.

The record: Belle Ayr Olympas, Ermine Trace Nadeya, Cotomodda Hubba-Bubba, Schoenweg Mona Lisa

F20. Lineage genotyped: a grade between a laboratory result and a traced descent OPEN

Entered 17 August 2026

A DNA genotype in a first-degree relative can settle which variant an untested animal carries.

Upward is certain. A dominant allele has to come from a parent. Where an offspring is genotyped and only one recorded parent shows the eye, that parent transmitted it. No probability is involved. 5 cats are graded this way.

Downward is not certain by descent alone. A heterozygote passes the variant to half its offspring, so descent from a tested parent proves nothing on its own. It is admitted only where the cat itself shows a blue or odd eye and is not dominant white: the eye establishes that a variant is present and the parent's genotype names which one. 3 cats are graded this way.

8 cats in total, against 29 carrying a laboratory result of their own. The grade is first degree only, so nothing is inferred across an unrecorded generation, and it is withheld wherever two variants sit in the ancestry. That withholding is not a technicality: it is exactly the Fletcherfield and Graf Reinhard case in finding 3.

What it unlocked, which is the larger result. Fixing a variant low in a pedigree lets the name travel back up the path the deduction came down. Of the 1083 cats the pedigree forces to carry or allows to carry, 349 now have the variant named rather than left as an unnamed carriage: DBE-ALT, DBE-RE.

Naming the variant does not upgrade the carriage. A cat on the path above an obligate pair still may carry and may not. What has changed is that a breeder now knows which test to order, and a laboratory result on any one of them settles a whole branch.

A further 56 cats sit more than one generation below a genotyped animal. They are not graded. Their pages name the variant available to them and say plainly that it is not established.

What would overturn it. A second variant found in any of these ancestries, which is why finding 3 and finding 6 stay open.

F19. Cats coded blue or odd eyed that carry no white OPEN

Entered 17 August 2026

PawPeds codes the eye separately from the coat: 61 blue eyed, 63 odd eyed, 67 pointed (Siamese) eye colour, 68 dominant blue eyed, 69 dominant odd eyed. A pointed eye has its own code, so a cat written 61 or 63 has not been coded as pointed.

That makes a test possible. Where the EMS code contains no w, no 01, 02, 03 or 09 white spotting and no 33 colourpoint, the blue or odd eye is not attributable to dominant white, to white spotting or to the pointed gene. 120 cats in this register meet it.

7 of the 120 have been genotyped, among them Cotomodda Hubba-Bubba. The other 113 carry no result of any kind, and they are the cats a test would settle.

The test says what the eye is not caused by. It does not name a variant. A cat carrying no white and coded blue or odd eyed is a candidate for genotyping, not an attribution.

The converse is the larger result. Counted across the register, of the 219 coded Siberian and Neva Masquerade records, 10 are dominant white and a further 80 carry a white spotting code: 41 per cent of the coded Siberian population is white or white-spotted, which is the direction finding 10 and finding 11 asked about, on the records this register holds rather than on a census.

What would close it. A genotype on any of the 113, and the same test applied to the remaining breeds.

F18. The 68 and 69 codes on PawPeds are a recent and partial re-coding OPEN

Entered 17 August 2026

PawPeds states directly that codes 68 and 69 are used for the dominant blue eye, and points to the breed and colour list linked from its database pages. That is the first statement the survey holds from a registry-derived database saying which coding system it uses for this trait.

The codes were introduced there within roughly the last year. Cats entered before the change remain under the older 61 and 63 codes unless they have been re-coded since.

The dates of birth in this register bear that out. Of the 575 Maine Coons here carrying 68 or 69, 513 were born in 2019 or later and 26 before it, with a single record from 2008 and a single one from 2013. A trait present in the breed since at least 2008 does not produce that distribution; a recent re-coding does.

What follows for every count on this site. Absence from the 68 and 69 listings is not evidence that a cat lacks the trait. It may be evidence only that its record has not been revisited. The same animals may appear under 61 or 63 in the older listings, and this register already holds cats coded both ways by two sources.

What would close it. The 61 and 63 listings for the same breeds, matched by name and date of birth against the 68 and 69 listings and against this register.

F14. Cattery prefixes holding more than one population OPEN

Entered 16 August 2026

Nikita'l, Aerlin, Hermes, annealtindan, Katrin Style and SuperQueen. In four the second population is another identified variant, which is a testing problem with a solution: both can be bought. In two it is a line with no variant sequenced at all, Katrin Style with the Nahal line and SuperQueen with the Igor line, for which no test exists at any laboratory at any price.

Qualified by finding 2. Where both populations turn out to descend from one source this is a naming problem rather than two variants, and only genotyping tells them apart.

F13. Latent carriers: derivation and counts OPEN

Entered 17 August 2026

A dominant allele cannot skip a generation, so who must have carried unseen is derivable. Recomputed at every build: 3 obligate carriers, 74 cats in an obligate pair, 1006 on a path above a pair, and 363 possible carriers by descent, each with the variant or line it could be carrying named.

Dominant white cats are excluded from the deduction. White masks the eye whatever the cat carries at PAX3, so there is nothing for a descent to explain and counting them would manufacture carriers out of a coat colour.

The limit is parentage, not logic. Only 1549 of 4225 records carry a recorded parent, so the true count is larger and rises with every sheet entered.

the full lists, on Latent lines

What is unresolved, and what would close it

In the order effort would best be spent on them.

Observations, outcomes and analysis

Observations

1. 4225 cats are held. 100 carry an identified variant and 29 of those have a laboratory result identifying it. 1518 sit in 43 named founder lines in which no variant has been sequenced.
2. The evidence is overwhelmingly assertion. 4019 reported, 165 line-traced, 29 genotyped, 12 tested and carrying none of the known variants.
3. Four registries code this trait and no two of them code it the same way, and a fifth, one of the largest, TICA, does not record eye colour at all. A cat registered there carries no eye information on its certificate, by design and not by omission, so the trait is invisible in those records however carefully a breeder observes it.
4. Every founder entered pedigree breeding without a cattery prefix, as a found cat, a shelter cat, a village cat or a household pet, and in several cases with no recorded ancestry whatever.
5. 43 cats have no name in any source, and on one chart sixteen kittens exist only as the counts "4 kittens non-blue-eyed" and "6 kittens copper-eyed".

Outcomes

Each observation has a consequence that does not depend on anyone's intentions. Because the variant is attested by descent rather than by measurement, a line name is doing the work a test should do, and a line name is only as good as the paperwork behind it. Because four coding systems are in use, a cat crossing a border arrives with a code its new registry either reads differently or cannot read. Because the founders were unregistered, the origin of most of these variants is unrecoverable: there is nothing behind the founder to trace. And because unnamed animals cannot be looked up, traced forward or asked for by query, the descendants of the very cats most likely to carry the trait unseen are the ones no record can reach.

Survey analysis: what this register can and cannot support

The register is not a sample. Nobody drew it at random, no denominator exists, and its contents are whatever has been charted and supplied. It cannot support prevalence, incidence or rate. Coverage is uneven by construction: 1184 of 4225 cats have an eye status that can be scored at all, 3859 carry an EMS code, and 546 cattery prefixes appear in the cattery table of which 305 have a single named animal recorded. The bias runs one way throughout, because a cat enters the record by being interesting, so every count here is a floor and none is an estimate.

What the register can and cannot say

It can say where a line has reached. Every descent on these pages is a chain of parent links.

It cannot say whether two lines carry the same allele. Twenty-nine animals are genotyped and they fall in two of the founder lines. Every collision recorded here is a place where a test would answer that question, and no test has been done.

The counts are floors. Registry eye codes miss carriers, and the register holds direct evidence of it.

• BelleAyr O'Blue Sea is blue-eyed and genotyped DBE^RE, and her EMS code carries no eye code at all. Her two full sisters from the same litter are coded 68 and 69.
• BelleAyr Pluto is odd-eyed and genotyped DBE^RE, and is coded MCO ns: no eye code, no white code, no point. Four generations of his dam line are coded 68.
• Ermine Trace Xrustal carries no eye code and was placed as not carrying the gene. He has three coded offspring: 64, 68 and 69.

A code is a registry's opinion. One sheet writes the same descending trait as 61, then 60, then 68, then 68, then 69 across five generations and three registries, and a fourth registry writes the same animals with no eye code at all.