We present measurements of the component masses in 15 Cataclysmic Variables (CVs) - 6 new estimates and 9 improved estimates. We provide new calibrations of the relationship between superhump period excess and mass ratio, and use this relation to estimate donor star masses for 225 superhumping CVs. With an increased sample of donor masses we revisit the implications for CV evolution. We confirm the high mass of white dwarfs in CVs, but find no trend in white dwarf mass with orbital period. We argue for a revision in the location of the orbital period minimum of CVs to 79.6+/-0.2min, significantly shorter than previous estimates. We find that CV donors below the gap have an intrinsic scatter of only 0.005R_{sun} around a common evolutionary track, implying a correspondingly small variation in angular momentum loss (AML) rates. In contrast to prior studies, we find that standard CV evolutionary tracks - without additional AML - are a reasonable fit to the donor masses just below the period gap, but that they do not reproduce the observed period minimum, or fit the donor radii below 0.1M{sun}_.

Cone search capability for table J/MNRAS/486/5535/table1 (Ephemerides for the CVs modelled in this paper)

Cone search capability for table J/MNRAS/486/5535/table2 (System parameters for the 15 eclipsing systems analysed in this paper)

Cone search capability for table J/MNRAS/486/5535/table3 (Orbital (Porb) and superhump (Psh) periods of the systems used to calibrate the {epsilon}(q) relation)

Cone search capability for table J/MNRAS/486/5535/tablea1 (System parameters for supplementary systems)

Cone search capability for table J/MNRAS/486/5535/tablec1 (Journal of observations)