Large discrepancies have been reported between the Gamow-Teller strengths determined from Ca P-delayed proton decay, and the isospin-analog strengths deduced from "Cl(p, n) studies. Some of these discrepancies could be removed if the Ca daughter level at 3.24 MeV in K had I'~forty times bigger than its I'", and not negligibly small as assumed in the initial analysis of the Ca P decay. We studied proton and p-ray branching ratios of low-lying K levels populated via the Ca(p, o) reaction. We find that the 3.24 MeV level has I~/1"")8, and that the 2.75 and 3.08 MeV levels have I'~/ I'"= 0.52 + 0.23 and I'~/I'"= 0.41 6 0.014, respectively. In addition we identified new K levels at E = 3.272 + 0.002 and 2.967 + 0.002 MeV. PACS number(s): 23.20.Lv, 21.10.Hw, 27.30.+t The reduced Gamow-Teller transition strength B(GT) for neutrino capture on Cl into a given state in Ar should be equal to the corresponding B(GT) for Ca P decay to the analog level in K if isospin is a valid symmetry. Several surprising difFerences [1] have been observed in the Cai K* B(GT)'s measured by Pdelayed proton emission [2] and the Cl~A r* B(GT)'s deduced from intermediate-energy zero-degree Cl(p, n) cross sections [3]. In particular, the Ca -+ K transition to the second 5/2+ excited state at E =3.24 MeV was reported to be forty times weaker than the strength assigned to the Ar analog level at E =3.17 MeV in