After making the ‘Langer transformation’, r = ex, ψ(r) = ex/2u(x), Langer found the first-order JWKB hydrogen radial wavefunction to be as if the centrifugal potential were ℏ2(l + 1/2)2/(2r2), thereby ‘justifying’ the substitution suggested by Kramers and known to get, in first order, the correct rl + 1 behavior at the origin, the correct phase shift and the exact energy levels. There have been many extensions of the Kramers–Langer substitution: to get the exact origin behavior at any pre-specified higher order; to show that no substitution is necessary at infinite order; to replace ℏ2l(l + 1) by L2 + ℏL, with L set equal to lℏ at the end. Recently, it was discovered that Langer's JWKB solution in x was exactly equivalent to a JWKB solution in r for r−1/2ψ(r): namely the Langer transformation was irrelevant. How can there be many seemingly incompatible JWKB expansions to solve one equation? The key is the ambiguous treatment of ℏ: in the radial kinetic energy, ℏ is the expansion parameter; in the centrifugal potential, ℏ is implicit, passive and not expanded. By designating the implicit ℏi by its own symbol, one sees immediately how the different JWKB expansions correspond to different partitions of the centrifugal potential between expansion ℏ and implicit ℏi and therefore solve different equations. The different expansions represent the same physical solution only when ℏi = ℏ. Moreover, in the two-ℏ notation, ‘the generalization’ of the Kramers–Langer substitution is made transparently simple: That is, the implicit ℏ2i/4 that completes the square is compensated by the expansion −ℏ2/4 that modifies the second-order JWKB wavefunction directly and higher orders indirectly.