A pedigree is a sales document until you recompute it. The inbreeding coefficient, F, is the probability that the 2 copies of a gene at a random locus are identical by descent, and it is only as good as the 5 generations behind it. Recompute F from the ancestor slots rather than the printed figure, hold the working line under 6.25 percent, treat 12.5 percent as a warning that needs a reason, and check 3 type risks that appear long before any blood test would flag them.
What a five-generation pedigree actually holds
The form most registries issue has 30 ancestor slots: 2 parents, 4 grandparents, 8 great-grandparents, and 16 great-great-grandparents. That is 4 generations of ancestors plus the cat, which is what the trade calls a 5-generation pedigree. A repeated name in those 30 slots is the only thing that raises F, and the position of the repeat sets the size of the contribution. One name appearing twice in the 4th generation matters 16 times less than the same name appearing as a grandparent on both sides, which is why a pedigree can look crowded and still compute under 3 percent.
Recomputing F by hand in 4 steps
One, list every ancestor that appears more than once and note the generation on each side. Two, for each one, count the number of steps up from the sire (n1) and from the dam (n2). Three, apply 0.5 to the power of (n1 + n2 + 1); that is the contribution of that ancestor when it is not itself inbred. Four, add the contributions. A grandsire shared by both parents gives n1 = 1 and n2 = 1, so 0.5 to the power of 3, or 12.5 percent. A grandparent shared at the same depth on both sides gives 0.5 to the power of 5, or 3.125 percent, and a full pair of shared grandparents doubles that to 6.25 percent. A shared great-grandparent at the same depth on both sides drops to 0.5 to the power of 7, or 0.78 percent.
The two lines: 6.25 percent and 12.5 percent
| Shared ancestor position | F contribution | 2 such ancestors | Reading |
|---|---|---|---|
| Both parents' sire (n1 = 1, n2 = 1) | 12.5% | 25% | Warning; needs a documented reason and health data |
| Grandparent on each side (n = 2, 2) | 3.125% | 6.25% | At the working ceiling |
| Great-grandparent on each side (n = 3, 3) | 0.78% | 1.56% | Comfortably inside the line |
| Repeat at 4th generation (n = 4, 4) | 0.195% | 0.39% | Negligible on its own |
Under 6.25 percent, a line stays usable across generations without stacking the same recessives twice per litter. At 12.5 percent, roughly an eighth of the genome is expected to be homozygous by descent, and that is where smaller litters, weaker neonates, and the type faults below start to concentrate. A full-sibling pair sits at 25 percent, outside any working range. A printed F of 5 percent that recomputes to 14 percent is not a rounding error, it is a different pedigree.
Three type risks that show up first
One, face and airway: a shorter muzzle and narrower nares arrive before anything else, and a cat that breathes through its mouth at rest or snores at 1 m has crossed from a type preference into a welfare problem. Two, jaw and teeth: an underbite or overbite of 2 mm or more, crowded incisors, and retained deciduous teeth at 6 months are the classic trio in a tightened line, and all 3 are visible in a 10-second look at a closed mouth. Three, coat and skin: a sparse or greasy coat, seborrhea along the spine, and a tail kink at the last 3 vertebrae are the faults breeders most often crop out of photographs. Each of these is a reason to read the 3-generation photograph set rather than 1 hero shot, and to ask for the same view of both grandparents.
Cross-checking the document itself
Four checks catch most errors, and 1 of them is the paperwork trail described in the registration walkthrough. One, registration numbers: the same cat name with 2 different numbers is either a rename or a fabrication, and both change the count. Two, dates: a queen listed with 2 litters 5 months apart is possible, 2 litters 8 weeks apart is not. Three, titles and health results: a champion title at 8 months is unusual enough to verify. Four, the coefficient printed on the form: ask which generation depth it was computed to, because a 3-generation figure is often 30 to 50 percent lower than the 5-generation figure for the same cat. A DNA panel for the main breed mutations commonly runs 60 to 200 dollars and an echo 150 to 300, which is cheap next to the cost of a litter that has to be rehomed, as the cost breakdown shows line by line.
A 2-week review, 20 minutes at a time
| Day | Minutes | Task | Output |
|---|---|---|---|
| 1 | 20 | Transcribe 30 ancestor slots into a spreadsheet | Duplicate name list |
| 3 | 20 | Compute F for each duplicate | Own F figure |
| 5 | 30 | Request 3-generation photo sets and health results | Written reply |
| 8 | 20 | Compare printed F with your figure | Gap in percent |
| 11 | 30 | Check jaw, nares, coat, tail on both sides | Pass or fail per item |
| 14 | 20 | Decide: proceed, ask, or stop | One-page note |
Two weeks is enough because nothing here needs a laboratory. The output is a 1-page note with 3 numbers: your F, the gap to the printed F, and the count of type faults you could confirm in photographs. If the gap is above 5 percentage points or any type fault appears on both sides, the honest move is to stop and look at an unrelated line.
What this review is not
It is not a green light. Most cats should not be bred at all: food, testing, and vet cover for 1 litter commonly pass 1,500 dollars before a single kitten is placed, and the market for kittens is not a reason to produce them. This page exists for someone already committed to a litter and holding a pedigree they were handed, so they can check 2 numbers before a queen is bred. Nothing here replaces the pre-breeding genetic and health panel, a cardiology screen, or the breed club's own code of ethics, and a coefficient under 6.25 percent does not cancel a single one of those.
Five steps before you accept the number
- Count the slots: 30 ancestors, and every duplicate flagged with its generation.
- Compute F yourself and write the gap to the printed figure in percentage points.
- Hold the line at 6.25 percent and demand a documented reason for anything at 12.5.
- Ask for 3-generation photographs of jaw, nares, coat, and tail on both sides.
- Put the pre-breeding panel and the echo in writing before, not after, the decision.
Terms mentioned in this article
Frequently asked
Why recompute a coefficient that the registry already printed?
Because the printed number depends on how many generations it was computed to, and a 3-generation figure is often 30 to 50 percent lower than the 5-generation figure for the same cat. Registries also accept corrected pedigrees, and a renamed ancestor can drop out of the count entirely. Recomputing 30 slots takes about 20 minutes and is the only way to know whether the 5 percent on the form is real.
Is 6.25 percent a safe number in every breed?
It is a working ceiling, not a guarantee. A line with a known heart or fertility problem should stay nearer 3 percent, while some closed breeds have no unrelated line left and live above 6.25 by necessity. The number also says nothing about which recessives are shared, so a 6.25 percent cat from a tested line can be safer than a 3 percent cat from one with no cardiac screening at all.
Do type faults really appear before health problems?
They are easier to see, not necessarily earlier in the body. Jaw alignment, nares width, and tail kinks are visible in a 10-second look and in photographs, which is why they are the practical first screen. Airway narrowing matters most because it affects breathing at rest, and a cat that snores at 1 m or mouth-breathes while sitting still needs a clinical assessment, not a pedigree discussion.