Show MeYes, third cousins are blood related, but only barely. Third cousins share a pair of great-great-great-grandparents and, on average, about 0.781 percent of their DNA, which works out to roughly 53 centimorgans across about 73 shared segments. That is real genetic overlap, not zero, but it is small enough that two third cousins usually share no detectable DNA at all on a single consumer test. The relationship is genuine in genealogy and negligible in medical genetics. Most people who ask this question are really asking a second one: does that sliver of shared DNA mean anything for health, inheritance, or marriage.
The Mechanism: How Third Cousins Inherit the Same DNA
Every person inherits half their autosomal DNA from each parent, and that halving repeats with every generation. The math that matters here is not the 50 percent halving, though. It is recombination, the shuffling that happens when eggs and sperm form. Each parent passes down a reshuffled deck, not a clean half of each grandparent. That shuffling is why relatives share segments of DNA rather than neat fractions.
Trace the path from you to a third cousin. You and that cousin are both descended from one ancestral couple sitting four generations above each of you: your great-great-great-grandparents. The shared ancestor is a great-great-great-grandparent to both of you, which places you both at the same generational depth, making this a "same-generation" cousin relationship rather than a removed one.
The count of shared ancestors doubles as you go back. You have 2 parents, 4 grandparents, 8 great-grandparents, 16 great-great-grandparents, and 32 great-great-great-grandparents. A third cousin descends from one of those 32 ancestral couples. You have 16 such couples at that level, so any single third cousin represents a tiny slice of your ancestral pool.
In plain terms: you and a third cousin both hold a few shuffled fragments of the same original deck, and most of the deck has been reshuffled beyond recognition.
The Numbers: Shared DNA by Cousin Degree
The percentages below are averages. Real shared DNA between any two cousins varies widely around those averages because recombination is random. A sibling pair can share anywhere from about 37 to 62 percent, not exactly 50. The same spread applies, proportionally, all the way down the cousin ladder.
| Relationship | Shared ancestors | Average shared DNA | Approx. shared cM | Typical range |
|---|---|---|---|---|
| First cousin | Grandparents | 12.5% | ~866 cM | 396–1,397 cM |
| First cousin once removed | Grandparents | 6.25% | ~433 cM | 102–980 cM |
| Second cousin | Great-grandparents | 3.125% | ~229 cM | 41–592 cM |
| Second cousin once removed | Great-grandparents | 1.56% | ~122 cM | 14–353 cM |
| Third cousin | Great-great-great-grandparents | 0.781% | ~53 cM | 0–173 cM |
| Third cousin once removed | Great-great-great-grandparents | 0.39% | ~27 cM | 0–98 cM |
| Fourth cousin | Great-great-great-great-grandparents | 0.195% | ~13 cM | 0–74 cM |
The centimorgan (cM) is the unit geneticists use to measure shared DNA. One centimorgan corresponds to roughly a 1 percent chance that a segment gets split by recombination in a single generation. The ~53 cM average for third cousins comes from the expected genetic contribution of a shared great-great-great-grandparent couple, halved four times and corrected for the fact that inheritance is probabilistic, not guaranteed.
The zero in the third-cousin range is the counterintuitive part. Because recombination breaks DNA into ever-smaller pieces, some third cousins inherit none of the ancestral couple's segments. Studies of real cousin pairs find that a meaningful share of third cousins share no detectable autosomal DNA at all on a standard consumer array. The relationship exists in the family tree, but the DNA evidence can be invisible.
What DNA Tests Actually Detect at This Distance
Consumer DNA tests look at roughly 700,000 markers out of the 3 billion base pairs in the genome. That is a sampling strategy, not a full read. A third cousin who shares 53 cM in reality may show up as 40 cM, or 70 cM, or 0 cM, depending on which markers the array happens to test and which segments the two of you inherited.
Companies label matches by centimorgan thresholds, not by family-tree labels. AncestryDNA and 23andMe both flag matches above roughly 7 to 8 cM as "distant," and both warn that small segments are unreliable because identical-by-chance segments can appear between unrelated people. At the third-cousin level, a 53 cM average sits comfortably above the noise floor, but individual segments of 6 to 10 cM are common and often false positives.
The practical consequence: if you and a known third cousin both test and do not match, that does not disprove the relationship. It means the sampled segments missed each other. Absence of a match at this distance is weak evidence. Presence of a match is stronger, but still needs triangulation with a third relative before you treat it as proof.
Where the Relationship Breaks: Half, Removed, and Double Cousins
The clean third-cousin label assumes full siblings all the way down the line and exactly four generations of separation. Real pedigrees rarely cooperate.
Half third cousins share one great-great-great-grandparent instead of two, which roughly halves the expected shared DNA to about 0.39 percent, or 27 cM. They can appear in the same 0–98 cM range as a third cousin once removed, so labels blur.
Third cousins once removed sit one generation apart: your third cousin's child, or your parent's third cousin. The expected share drops to about 0.39 percent, roughly 27 cM.
Double third cousins arise when two lines loop back on themselves, such as when two siblings from one family marry two siblings from another and both branches have children who later marry. The shared DNA roughly doubles to about 1.56 percent, which puts a double third cousin near a second cousin on a test. Isolated communities with limited founder populations produce these cases more often than outsiders expect.
Consanguinity distorts the numbers further. In populations where marriage between relatives is common, cousins share more DNA than the textbook average because their ancestors overlap in multiple places. This is why genetic counseling for cousin marriage weighs the actual pedigree, not the generic third-cousin figure.
Does the Shared DNA Carry Health Risk?
For third cousins, the inherited-disease risk is close to that of unrelated people. The reason is the coefficient of relationship. Third cousins have a coefficient of 0.0078, against 0.5 for a parent-child pair and 0.25 for full siblings. That number is the expected fraction of genes shared identical by descent, and it scales the risk of recessive disease.
A recessive disorder appears when a child inherits two defective copies of the same gene. The risk rises when both parents carry the same defective variant inherited from a common ancestor. First cousins have a coefficient of 0.0625, and the classic estimate is that their children face roughly a 4 to 6 percent absolute risk of a significant birth defect, against about 2 to 3 percent in the general population. Third cousins, with a coefficient eight times smaller than first cousins, carry a risk barely distinguishable from background.
A 2020 study of more than 400,000 UK Biobank participants quantified this directly. The researchers found that the offspring of first cousins showed measurable but modest increases in some disease risks, while the offspring of second and third cousins showed effects close to zero. The signal diminished sharply with each step down the cousin ladder.
The professional caveat is that "barely distinguishable" is not "identical." If a rare recessive variant is concentrated in a family, even distant cousins can both carry it. Genetic counselors treat the coefficient as a starting point and check the specific family history, carrier status, and ancestry before giving a number.
Correcting the Common Misconceptions
Several myths persist about third cousins and blood relation.
Misconception one: third cousins are not really "blood" because the shared DNA is under 1 percent. Wrong framing. Genetic relatedness is a continuum, and there is no sharp line below which relatedness stops. Third cousins share measurable, traceable DNA; fourth cousins share less; tenth cousins share an amount indistinguishable from chance. "Blood related" is a legal and genealogical label, not a genetic cutoff.
Misconception two: a DNA test that shows no match means the family tree is wrong. At third-cousin distance, a negative result is uninformative for the reason above. The test samples a fraction of the genome and the segments may simply not overlap.
Misconception three: third-cousin marriage is dangerous. The disease-risk data does not support meaningful danger. The absolute risk elevation for a third-cousin couple's children is so small that it falls within the noise of population baseline. This is the relationship where the coefficient of relationship genuinely approaches zero in practical terms.
Misconception four: relatedness is a single number. It is not. Two third cousins can share 0 cM or 170 cM and both be accurately described as third cousins. The label describes the pedigree, and the centimorgan count describes the realized inheritance.
A Timeline: How Shared DNA Fades Across Generations
The fading is not linear in a way people intuit. It is geometric, and the tail flattens into noise.
| Generation back | Relationship to you | Expected shared % | Approx. cM | Detectable on a test? |
|---|---|---|---|---|
| 1 | Parent or child | 50% | ~3,400 cM | Always |
| 2 | Sibling | 50% (avg) | ~3,400 cM | Always |
| 3 | First cousin | 12.5% | ~866 cM | Almost always |
| 4 | Second cousin | 3.125% | ~229 cM | Usually |
| 5 | Third cousin | 0.781% | ~53 cM | Often, not always |
| 6 | Fourth cousin | 0.195% | ~13 cM | Sometimes |
| 7 | Fifth cousin | 0.049% | ~3 cM | Rarely, near noise |
| 10+ | Tenth cousin or beyond | <0.01% | <1 cM | Indistinguishable from chance |
The generational count above tracks the line from you to the shared ancestor, not the number of people in between. The table shows why genetic genealogy becomes speculative past the fifth-cousin level: the shared segments shrink below the reliability threshold of consumer arrays, and the only remaining evidence is documentary.
What to Actually Do With This Information
If you are researching a third cousin, use these concrete actions.
Run the DNA test at a company with the largest possible database. Match rates at the third-cousin level depend heavily on database size, and a 20-million-person database finds far more distant matches than a small one.
Look for multiple shared segments above 10 cM rather than a single small one. One 7 cM segment is weak evidence; three segments of 15 cM each, spread across different chromosomes, is strong evidence.
Triangulate with a third relative. If you and a suspected third cousin share a segment, test a known second cousin of one of you and check whether all three share the same segment region. Triangulation confirms the segment descends from the common ancestor rather than from chance.
Upload your raw data to a second service with a different matching algorithm. A segment missed by one company's array may appear in another's.
Do not use a third-cousin match to make health decisions. The shared DNA at 0.781 percent is far too small to infer carrier status for a recessive condition or to predict disease risk. For that, order direct carrier screening.
If you are planning a child with a third cousin, the standard advice is to mention the relationship to your clinician so it appears in the family history. The risk is close to baseline, but accurate records help the counselor rule out concentrated variants in your specific family.
Limits of the Evidence and Where Science Stops
The centimorgan averages are population estimates, derived from large studies of real pedigrees. They describe what happens on average, not what happened in your family. An individual third-cousin pair can be an outlier in either direction, and the only way to know their actual shared DNA is to test them.
Consumer arrays are also not sequencing. They infer relatedness from sampled markers and can misjudge small segments, which is why companies caution against treating distant matches as certain. Whole-genome sequencing would be more precise but is rarely used for routine relatedness testing because the cost is not justified.
Finally, the risk figures for cousin marriage come largely from historical and population studies with their own confounders. The genetic counselor's rule that third-cousin risk "approaches baseline" is well supported, but the exact number for any family depends on variants the studies could not see. Established science says the shared-DNA fraction and the disease-risk elevation both shrink with each cousin step. It does not claim a guaranteed zero for any specific family, and it should not be presented as if it does.