The Genetic Atlas of European Immigration: How Italian, Scandinavian, French and German Migration Painted the DNA Map of the United States and Canada

White Americans are not a uniform population. Project the average G25 coordinates of each state onto a PCA of European reference populations and a coherent map emerges: Minnesota plots beside Denmark and Norway, Connecticut and New Jersey drift toward Bergamo and Tuscany, Chicago and Manitoba lean east toward Warsaw and Lviv, Louisiana sits almost on top of northern France, Maine, Vermont and New Hampshire lean unmistakably toward Brittany, and the Deep South, Mountain West and Appalachian states cluster tightly around Welsh, English and Cornish averages. These are not random signals. They are the genetic shadow of three centuries of European immigration, port by port, decade by decade. The largest-ancestry county map of the 2016 US Census and the 2021 Canadian Census ancestry map both anticipate these patterns at the cultural level. G25 PCA simply confirms them at the genome level.

1. From Census Categories to Genetic Coordinates

The 2016 American Community Survey and the 2021 Canadian Census let respondents declare their ancestry in their own words. The resulting county-level largest-ancestry maps tell an immediately recognisable story. German is the single largest reported ancestry across the entire Midwest and Great Plains, from western Pennsylvania to eastern Washington. Mexican dominates the Southwest. Irish is the largest reported ancestry across most of the Northeast and Atlantic coast. African American is the modal category through most of the Deep South. "American" appears as the dominant self-identification through Appalachia, Eastern Kentucky, Tennessee and the Ozarks, a marker of deep colonial-era ancestry that respondents no longer trace to a specific European country.

The PCA above is the single most informative figure in this analysis. Read from left to right and top to bottom, it reproduces in two dimensions exactly the regional immigration history written in the census maps: the Anglo-Celtic core in the top-right; the French-Belgian Atlantic populations sliding leftward; the Italian-shifted tri-state populations sinking downward toward the Dutch and Belgian midpoint; the Scandinavian-shifted Upper Midwest drifting right toward Sweden and Denmark. None of this alignment is hand-fitted. The G25 coordinates were calculated independently from millions of SNPs, the European references were positioned by Davidski years before this analysis, and yet the immigration record falls into place state by state, province by province.

Self-reported ancestry, however, is not the same thing as genetic ancestry. A respondent in Iowa who calls themselves "German" may carry one German great-grandparent on a substrate of British, Irish and Scandinavian descent. A New Jersey respondent who reports "Italian" may be 75% Italian or 25% Italian. The advantage of G25 PCA is that it sees through the labels and shows the actual genetic position of each regional average. When we project the white-population G25 coordinates of every US state and Canadian province onto the same European PCA space used to plot Welsh, Norwegian, Italian_Bergamo or French_Brittany, we get a coordinate-level map of immigration that the census alone cannot produce.

2. The Anglo-Celtic Foundation: Old Stock America

The single largest genetic component of white America is the British Isles substrate carried by colonists who arrived between 1607 and 1775. In the G25 PCA, this substrate is captured almost perfectly by the Welsh, English, English_Cornwall and Scottish averages, which themselves cluster within a tight 0.005-radius region of PC space. Modern Bell Beaker descendants from Britain, in other words, look genetically like one another regardless of which corner of Britain they come from. The same is true of their American descendants. From Alabama to Utah, from Tennessee to Idaho, from West Virginia to Texas, the closest European reference is always Welsh or English_Cornwall, and the second-closest is always the other.

This is why the southern, mountain and Appalachian states share an almost identical PCA position. Alabama whites (n=74), Tennessee whites (n=95), Kentucky whites (n=130), Georgia whites (n=96) and Texas whites (n=98) are all within 0.025 G25 units of the Welsh average and within 0.030 of English_Cornwall. Utah whites (n=29) sit at the far end of this baseline, the most British-baseline of any large US state, reflecting the predominantly English, Welsh and Scottish founder pool of early Mormon migration. Iowa whites (n=34) and Newfoundland whites (n=60) are similarly British-baseline.

3. The Tri-State Italian Cluster

Between 1880 and 1924, more than 4 million Italians immigrated to the United States, of whom roughly 80% came from the Mezzogiorno (Sicily, Calabria, Campania, Basilicata, Apulia, Abruzzo) and 20% from the industrial cities of the North. Some 97% of them passed through Ellis Island, and the largest concentrations settled within fifty miles of New York Harbor: Manhattan's Little Italy, Brooklyn, the Bronx, Newark, Jersey City, Paterson, Trenton, Bridgeport, New Haven and Providence. Together with simultaneous Greek, Polish, Russian Jewish and Eastern European Catholic migration, the demographic composition of the entire tri-state region was permanently restructured between 1880 and 1924.

The G25 PCA records this restructuring with surgical precision. On Figure 3, New_Jersey_White is the lowest US state on PC2: it sits below the entire French-Belgian cluster, in the empty space between Belgian and German, in the direction of Italian_Bergamo and Italian_Tuscany. Connecticut_White sits one notch above New Jersey, also clearly south of the British and French-Belgian centre. New_York_White falls in between, dragged down by Italian and Eastern European immigration but still partially anchored by its substantial Irish and German Old Stock substrate. By contrast, the Old Stock cluster of Vermont, Newfoundland, Utah and the Western Australian populations sits a full 0.005 to 0.010 G25 units higher on PC2, with no Italian or Eastern European pull at all.

The closest single European reference for New York is no longer Welsh: it is French_Nord (a population that itself reflects substantial Belgian and German Beaker ancestry). For New Jersey, the closest is French_Alsace. For Connecticut, BelgianA. These are not random results: French_Nord, French_Alsace and BelgianA all sit between Britain and central Europe on PC2, exactly where one would expect a state mixing British, Irish, Italian and Polish ancestry to land.

The Italian signal is real, demonstrably so, and tracks the historical immigration record. New Jersey carries the highest North-Italian-like fraction of any US state (32%), with secondary contributions from Polish and Greek references. New York follows at 26%, with a much larger French_Brittany component (which here functions as a midpoint absorber for Irish-and-Italian-and-Eastern-European mixture rather than indicating direct French ancestry). Connecticut is the most Slav-shifted of the New England states (24% Polish), reflecting late-19th-century industrial Polish, Lithuanian and Slovak immigration to Hartford, Bridgeport and Waterbury. Massachusetts and Rhode Island are slightly more diluted but still show 19% Italian-like. Pennsylvania, despite its much larger size and German agricultural core, still carries 17% Italian and 13% Slavic-like ancestry due to Pittsburgh, Philadelphia and the anthracite coal towns of the northeast.

4. The Upper Midwest Scandinavian Belt

Between 1850 and 1920, approximately 1.3 million Norwegians, 1.3 million Swedes, 350,000 Danes and 250,000 Finns crossed the Atlantic. Most were small farmers and rural labourers attracted by the Homestead Act of 1862, which offered 160 acres of public land in exchange for five years of cultivation. The geography of available homestead land funnelled these migrants overwhelmingly into the same few states: Minnesota, Wisconsin, North Dakota, South Dakota, Iowa and the eastern halves of Montana and Nebraska. The 1900 US Census recorded that more than 16% of Minnesota's population was Norwegian-born or had two Norwegian-born parents, and another 12% Swedish. North Dakota was even more concentrated, with Norwegian-Americans forming the absolute plurality of the state in the 1900 and 1910 censuses.

The G25 signature is unmistakable. Minnesota whites (n=28) are the only US state whose three closest European references are all Scandinavian: Danish at distance 0.0271, Norwegian at 0.0294, Icelandic at 0.0314. The Welsh distance for Minnesota is 0.0414, by far the largest of any state in this study. South Dakota whites (n=6, small sample) sit even further north, closest to Norwegian at 0.0409 and Swedish at 0.0468. North Dakota whites (n=10) show the most Polish-shifted profile of any state (Polish at 0.0387, Czech 0.0355), reflecting the substantial Volga German, Polish and Bohemian Catholic farming colonies that settled the western prairie alongside the Lutheran Scandinavians.

5. The Polish and Ukrainian Belt: From Chicago Polonia to the Manitoba Parkland

Between 1880 and 1924, approximately 2.5 million Poles emigrated to the United States, and another 170,000 Ukrainians arrived between 1891 and 1914 in two parallel streams: one to the industrial cities of the American Midwest and Northeast, the other to the prairie homesteads of western Canada. Both groups left a clean, measurable G25 footprint that NNLS captures cleanly when an explicit Polish and Ukrainian source is included in the model. Without those sources, the signal collapses into "Lithuanian_PA" and "French_Brittany" proxies, which is what often confuses commercial-test interpretations of "Eastern European" or "Baltic" assignments in Polish-American customers.

Three observations are worth highlighting. First, the Polish signal in Connecticut (24% of total ancestry) is the single highest Slavic share of any US state in this analysis, reflecting the dense Polonia of Hartford, New Britain, Bridgeport and Waterbury. Second, the North Dakota signal (35% Polish + 17% Lithuanian) confirms the well-documented but often-overlooked Volga German, Polish and Ukrainian Catholic farming communities of Stark, Pierce and McIntosh counties, on top of the Norwegian Lutheran prairie substrate. Third, Manitoba shows the cleanest Ukrainian-bloc settlement signature in North America: NNLS distributes the non-Anglo, non-French ancestry across Lithuanian_PA (17%) and German (17%), with the unmistakable rural-Ukrainian fingerprint of Dauphin and Interlake.

6. French Atlantic Inheritance: Louisiana, Maine, Vermont, New Hampshire

French settlement in North America followed two corridors. The first was the Saint Lawrence Valley, where Samuel de Champlain founded Quebec City in 1608 and where roughly 8,500 French settlers (the so-called filles du roi and their husbands) became the founder population of essentially all 7 million modern French-Canadians. The second was the Acadian peninsula and the Gulf of Saint Lawrence, where 12,000 to 15,000 French Catholics settled between 1604 and 1755 before being deported by the British during le Grand Derangement. About 3,000 of those deportees eventually reached Louisiana, where their descendants became the modern Cajuns. A smaller contingent stayed in Maritime Canada, where their descendants form the modern Acadian populations of New Brunswick, Nova Scotia and Prince Edward Island.

Three of these populations remain genetically French in 2026, and one US state was permanently shaped by them.

The Louisiana case is particularly clean. Of all the US states in this analysis, Louisiana is the only one whose closest single European population is French (specifically French_Pas-de-Calais at distance 0.0185). The second-closest is BelgianC at 0.0190, the third French_Nord at 0.0195. All four French-and-Belgian references sit within 0.005 G25 units of Louisiana whites. Compare this to Alabama, where the closest French reference (French_Brittany) sits at 0.0250 and the next French reference at 0.0483. The Louisiana profile is genuinely shifted toward continental northwestern France, not just slightly tinted by it.

On Figure 3 the visual confirmation is immediate. White_Cajun (n=20) is the leftmost American population on the entire plot, sitting almost exactly on top of French_Auvergne and French_Nord with French_Alsace and BelgianC as its near neighbours. White_Acadian (n=36), Louisiana_White (n=96) and New_Brunswick_White (n=46) form a tight diagonal a fraction of a unit to the right of the Cajun position, all of them embedded in the French-Belgian reference cluster rather than the British one. Bahamas_White (n=5) plots in the same region, reflecting the substantial Loyalist French and Acadian-derived population of the eighteenth-century Bahamian settlements. No other US state ventures into this part of the European PCA. The northern French shift of Louisiana, the Cajun community and the Acadian Maritimes is not just statistical: it is geographically immediate when the PCA is plotted.

7. The Full Picture: State Distance to Closest European Population

Figure 3 above shows the position of each state on the European PCA. The bar chart below summarises the same information differently: for each US state and each Canadian province we plot the closest European reference population and the weighted G25 distance to it. The three views (the PCA, the regional NNLS sections in 3, 4, 5 and 6, and the bar chart below) are complementary. The PCA shows direction (north-south on PC2 is the British-Italian axis; west-east on PC1 is the French-Scandinavian-Slavic axis), while the bar chart shows the absolute closeness of each state to its best European match.

8. G25 Coordinates: Selected State Averages

The following are weighted G25 averages of consumer-DNA samples from each state's white population. Use them as targets in Vahaduo (Piotr Kapuscinski) or as input to Karl Hogstrom's K248, the Standard G25 Calculator by Davidski or any of the Migration Era Calculator on ExploreYourDNA to model your own ancestry against the regional baseline of any US state or Canadian province.

9. Summary: State by State

10. Common Misconceptions

11. The Bigger Picture

What the G25 PCA reveals is that the United States is not a melting pot in the literal genetic sense. It is closer to a regionally stratified mosaic in which post-1880 immigration left durable, geographically bounded fingerprints on top of a colonial Anglo-Celtic substrate. Three centuries of intermarriage have not erased these signals; they have softened them, and the softening is itself informative. The closer a state's white population sits to the Welsh-English_Cornwall reference, the older its colonial substrate and the smaller its post-1850 immigration share. The further a state drifts toward Italian, Scandinavian, French or Slavic references, the more recent and concentrated its immigration history.

For the genetic genealogist, this has a practical consequence: when commercial DNA tests assign you "Italian" or "Scandinavian" or "French" ancestry, they are not picking up some abstract continental signal. They are picking up the genetic memory of a specific 19th- or 20th-century migration into a specific port or homestead corridor, mediated by the marriage pool of the descendants who stayed behind in those specific neighbourhoods. Your ancestors are not a statistical abstraction; they are individuals who passed through Castle Garden or Ellis Island or Boston Harbor or New Orleans, who farmed wheat in Otter Tail County or rolled cigars in Tampa or made shoes in Brockton. The state-level G25 averages above are the average of all those individual stories.