Priority (resolved against the primary source): t(3)=t(4)=t(5)=1 is IMPLICIT in EHT91 Theorem 3 — (3r-3,1)->_r ceil(r/5) for all r>=3 gives upper bound 1 for r<=5 — but is nowhere stated in the paper, which records no exact small-r values. The contribution of this finding is the explicit statement, independent self-contained proofs by a different method (inclusion-minimal empty-intersection analysis), and exhaustive machine certificates. The erdosproblems.com background mis-transcribes both EHT91 bounds (drops the floor and the ceiling); an erratum is filed separately.