# T6 and T7 verdicts (30 August 2026, 20:45 ACST)

Under FROZEN.md (commit 8f79f68) with DEVIATIONS.md entries 0 to 3. Order of
record in git: freeze 20:02 (8f79f68), T6 dump enumeration and deviation 1
20:06 (0ed385a), T6 FTDNA checks and results 20:13 (faff48f), T7 results
20:35 (c824bf5), this verdict.

## T6, the fork-splitter search: CLEAN NULL

No H1 (no node interposed between J-Z1865 and J-Z1853 in either tree; the
Z1853 block's variants sit under one node in both), no H2 (FTDNA own-level
testers at J-Z1865: 0, among 13,014 testers under Z1865), no H3 (own-level 0
at J-Z1853, the Z643 block and J-CTS9721; P58* sensitivity also 0). The
dump arm was uninformative for occupancy because its `*` nodes are
structural (deviation 1). The "private off-tree variants" clause was
unapplied, as pre-stated: no public source exposes them.

Reading R1, verbatim: "empty bins = clean null, gift-line unsampled or
extinct, consistent with the standing form, no probability shift claimed".

J-Y519675: INDETERMINATE under the frozen rule (AR1, AR2 met, AR3 in part,
AR4 untestable; OB2 met, OB1 not, OB3 untestable). Its 14 leaves are not
counted in any bin. The rule's OB2 was written too broadly to separate "the
descendants are real" from "the grouping is real"; logged, not corrected.

Resolution limit restated: nothing public can show whether any of the
13,000 testers carries a partial Z1853 block and was filed at a child
regardless. Only a lab's naming of an intermediate node would show it, and
none has.

## T7, the side-branch geography test: STRAINED

Seven of eight branches scored (V6422 insufficient, three known-origin
carriers). Corridor-consistent: L860 (0.979), ZS6178/M11522 (0.957), FGC4745
(0.907), and Y3081/ZS241 under the sub-lineage D2 ruling (0.960; 0.085 on
the primary row). Mixed: Z27682 (0.600). Non-corridor: Y15152/Z18297 (0.224)
and YP4763/ZS2518 (0.292). Two or more non-corridor after the diaspora rule
puts the result in the frozen category "strained".

Frozen wording (B6): the geography of the confirmed stem side-branches
strains the identification of the L858 kin-family as the elected house;
weight shifts against that identification; the standing form itself is
untouched.

Asymmetry (B5, frozen before scoring, verbatim): "J1 as a whole is Middle
Eastern, so corridor-consistent geography is partly baseline expectation and
supports the election only weakly, whereas clearly non-corridor geography
would genuinely strain it; the test's power is asymmetric and the verdict
must say so". The five corridor-classed branches therefore add little; the
two non-corridor branches are the result.

What the two strains are, distinctly:
- Z18297's Class C (163 of 210 known) has a disclosed Jewish-diaspora
  component (a Jewish-heritage source names Levant-origin sub-branches under
  it, without counts), which the frozen no-guessing rule could not reassign;
  the rule's conservatism may overstate this strain, by an amount no source
  quantifies.
- ZS2518's Class C is Sardinia 25 and Italy 9 of 46: not diaspora in any
  source. Sardinian J1 is attested from the Nuragic period in the literature.
  This is the genuine strain: a confirmed stem side-branch whose surviving
  carriers are mostly west Mediterranean.
Both are exactly the two branches Sahakyan et al. 2021 names as Z1853's
"minor branches [that] contain samples from Europe". What this test cannot
separate: an early westward departure of a corridor family from a stem that
was never in the corridor world. Sensitivity (Class C removed) is 1.00 for
every branch, the ceiling of any diaspora explanation, and says nothing.

Branch 8's D2 ruling (deviation 3) was made after the numbers were seen and
moves that one branch from non-corridor to corridor-consistent; both rows
are reported and the category is "strained" either way.

## Consequence for the standing form

Recorded whatever it is. T6 finds nothing and shifts nothing. T7 lands
against interest on the only question it had power to answer: two of the
seven scorable stem side-branches live outside the corridor world, and the
identification of the L858 kin-family as the elected house carries that
cost. The standing form (election among then-living corridor chromosomes)
does not depend on which family was elected. Revival condition unchanged:
corridor aDNA on or near the Z1865 to Z1853 stem.

No website change. Editorial decisions are Mike's.