Our post on endogamy explains why these populations show the signal. This one is the country-by-country audit — and the three are not equally evidenced.
Colombia — the strongest genetic case
- Chacón-Duque et al. 2018: Colombia had the highest Sephardic component of the five CANDELA countries, ~10% on average, measured haplotype-wise and reported over and above Iberian reference individuals.
- Antioquia is the endogamous sub-population everyone means: a small founder group, generations of cousin marriage, the paisa surnames recurring.
- The famous 2000 figure — ~14% of Antioquian Y-haplotypes "could have a Jewish ancestry" — is six-STR resolution from 2000, leans on the Cohen Modal Haplotype (common among non-Jewish Arab populations), and carries the authors' own warning that it needs far more Mediterranean sampling. Quote the 2018 autosomal figure; mention the 2000 Y figure only to explain why it is weak.
Mexico — the strongest documentary case
Mexico's converso history is not inferred, it is tried in court.
- Luis de Carvajal y de la Cueva, governor of the Nuevo Reino de León (1579), settled the north with kin and associates. He and much of his family were prosecuted by the Mexican Inquisition; his nephew Luis de Carvajal "el Mozo" was burned in 1596.
- That is why Nuevo León / Monterrey is the region named in every serious discussion of colonial crypto-Judaism in Mexico, and why New Mexico — the far end of the same settlement current, at maximum distance from Mexico City — carries the same tradition.
- Genetically, Mexico sits at ~4% Sephardic in the 2018 study — below Colombia. Note the asymmetry out loud: the country with the best documents does not have the highest percentage. Documents and percentages measure different things, and that is the whole lesson.
Costa Rica — the strongest genealogy, the weakest genetics
- The Central Valley is a textbook founder population: small colonial founding group, long endogamy, repeatedly studied as a possible genetic isolate, and exceptionally well documented in Sanabria Martínez, Genealogías de Cartago hasta 1850, which lets a Costa Rican trace real lines to the founding families rather than guess.
- But Costa Rica is not in CANDELA. The 2018 study did not sample it, and we have found no peer-reviewed converso-ancestry estimate for it using modern haplotype methods.
- Colombia's 10% does not transfer. What Costa Rica has is strong genealogy, real endogamy, and no published genetic estimate. That gap is a research opportunity, not a finding.
At a glance
| Documentary evidence | Modern genetic estimate | Endogamy | |
|---|---|---|---|
| Colombia (Antioquia) | moderate | ~10% Sephardic (2018, haplotype, vs Iberian controls) | very high |
| Mexico (Nuevo León) | strong — Inquisition trials, named individuals | ~4% | high in places |
| Costa Rica (Central Valley) | strong genealogically | none published | very high |
The point
A reader from any of the three deserves the same answer in different order: Colombia, start with the genetics and go looking for documents. Mexico, the documents already exist — start there. Costa Rica, you have the best family trees in Latin America and nobody has done the genetics yet.