The standard picture of CV secular evolution predicts a spike in the CV distribution near the observed short-period cutoff P_0 ~ 78 min, which is not observed. We show that an intrinsic spread in minimum (`bounce) periods P_b resulting from a genuine difference in some parameter controlling the evolution can remove the spike without smearing the sharpness of the cutoff. The most probable second parameter is different admixtures of magnetic stellar wind braking (at up to 5 times the GR rate) in a small tail of systems, perhaps implying that the donor magnetic field strength at formation is a second parameter specifying CV evolution. We suggest that magnetic braking resumes below the gap with a wide range, being well below the GR rate in most CVs, but significantly above it in a small tail.