Registries that do not require health screening commonly give cost and access as the reason. This case study takes the two measures to which that reason cannot apply. Blood group typing is a single line on a laboratory panel most breeders already buy. The coefficient of inbreeding costs nothing at all, because it is arithmetic performed on a pedigree the registry itself issued. Both are within reach of every organisation in this survey, and neither is required by more than a handful.
A queen of blood type B mated to a tom of type A or AB can produce kittens of type A or AB. Her first milk carries strong antibodies against type A blood. Kittens that are born healthy, suckle, and then fade and die in the first days of life are dying of neonatal isoerythrolysis, in which the mother's own antibodies destroy the kitten's red blood cells, and the whole litter can be lost. It is preventable in its entirety by knowing the type of both parents before the mating: the pairing is either avoided, or the kittens are withheld from the queen for the first day and hand-reared.
Type B is not rare, and it is concentrated in particular breeds: the British group, meaning the British Shorthair and the British Longhair, together with the Devon Rex, the Cornish Rex, the Turkish breeds and the Birman. The test is on the standard menu of nineteen laboratories in this directory and is included as standard in every large multiplex panel, so most breeders using a panel are already paying for it. It is needed once in a cat's life and never repeated.
No organisation in this survey requires blood group typing of the cats it registers. Eleven name it as a recommendation. The remaining sixty-nine say nothing about it at all.
The eleven are FIFe at federation level, SVERAK, NRR, the Finnish Cat Association, KKÍ Iceland, Mundikat, ANFI, ASFE, SFDH, Felis Britannica and FASA. An organisation is counted here only where its own published document names blood group or AB antigen typing. FIFe's Appendix I recommendation formally reaches every FIFe member, so a reading that counted the federation's recommendation as binding on all thirty-six would give a much larger figure. The narrower reading is used throughout this survey.
There is no cost argument available here, and no access argument either. The result is a single line on a panel a breeder is already buying, and it prevents a kind of loss that is total when it happens. This is the clearest single omission the survey found.
The veterinary profession does not share the registries' silence on the point. The Canadian Veterinary Medical Association's Code of Practice for the Care of Domestic Cats (second edition, 2025) states at section 2.3.2.2 that part of responsible breeding is ensuring that all cats and kittens receive preventive veterinary care, including testing for feline leukaemia and feline immunodeficiency virus and blood typing as appropriate. Three Canadian provinces incorporate an edition of that code into their own regulations. So a national veterinary body names the test, and the registries that actually decide which cats may be bred from do not.
COI costs nothing at all. It is arithmetic performed on the pedigree the registry itself issued. Free tools exist that breeders already use, PawPeds among them, and several registries hold the pedigree data that would let them compute it for every planned mating without asking the breeder for anything.
Positions across the eighty organisations divide as follows. Fourteen impose a binding limit of some kind and three impose one in defined circumstances only. Three state a recommendation. Thirty-seven have no provision at all, and twenty-three publish nothing from which a position could be established. Put on the common scale used in the inbreeding section above, where a first cousin equivalent is 6.25 per cent, eleven registries already meet that standard, usually without ever mentioning a coefficient, because they forbid the close matings that would exceed it. Twenty-five leave a full sibling mating available simply by saying nothing about it.
Almost none of those rules is a coefficient. Only three registries in the world state a COI figure at all, and a figure is meaningless without the number of generations it was calculated over, because the same pair of cats returns a higher number the deeper the pedigree goes. Two of the three state that depth. Every other rule counted above works by naming forbidden matings or by counting distinct ancestors, so no depth is required of it. Which instruments state a figure, which state a depth, and why the two differ, is set out in the section on generation depth.
No registry anywhere in the survey controls the rate at which inbreeding accumulates, which is what the quantitative genetics literature actually recommends and what determines whether a breed remains viable. Every instrument that exists controls the level of a single mating, and a population can climb steadily towards any fixed ceiling while every individual mating remains compliant.
The numbers are not missing for want of anyone writing them down. The European Union's own Platform on Animal Welfare endorsed a Guide to Responsible Cat Breeding on 3 November 2020. It sets four figures: no breeding from a cat above 20 per cent; a coefficient ideally under 10 per cent; no selective breeding where the pair's relationship is unknown over at least five generations; and no stud siring more than 5 per cent of his breed's kittens in five years. The last of those is the only published attempt anywhere in this survey to limit the popular sire effect by a number. None of it binds anyone, and no registry has adopted any of it.
Blood group typing is mostly held as a recommendation and COI mostly as a limit on a single mating. In both cases the question of what happens when a breeder does not comply is one this survey can pose but not answer.
Thirty-two organisations gate compliance at registration, so their requirements can be verified from the outside, and a further seventeen do so for some of their requirements but not all. Twenty-four publish nothing from which any enforcement mechanism can be established, six state that they have none, and one relies on the breeder's own declaration.
Nowhere in this survey does an organisation publish how many of its members follow a recommendation, how many results it has actually received, or what happens when one is not produced. A recommendation with no verification and no reporting is, from the outside, indistinguishable from silence. The enforcement discount applied in the scoring here is a crude proxy for that, and it is a proxy precisely because no registry publishes the figures that would let it be measured properly. This is a reasonable thing to ask each of them for.
Counts as at the census date, across the eighty organisations surveyed. Blood group and inbreeding positions are coded from the same primary rulebooks as the rest of the survey and can be traced through the requirement records and the source directory.