We extend our initial study of the connection between the UV colour of galaxies and both the inferred stellar mass-to-light ratio, {Upsilon}*, and a mass-to-light ratio referenced to Salpeter initial mass function (IMF) models of the same age and metallicity, {Upsilon}*/{Upsilon}_Sal_, using new UV magnitude measurements for a much larger sample of early-type galaxies, ETGs, with dynamically determined mass-to-light ratios. We confirm the principal empirical finding of our first study, a strong correlation between the GALEX FUV-NUV colour and {Upsilon}*. We show that this finding is not the result of spectral distortions limited to a single passband (e.g. metallicity-dependent line-blanketing in the NUV band), or of the analysis methodology used to measure {Upsilon}*, or of the inclusion or exclusion of the correction for stellar population effects as accounted for using {Upsilon}*/{Upsilon}_Sal_. The sense of the correlation is that galaxies with larger {Upsilon}*, or larger {Upsilon}*/{Upsilon}_Sal_, are bluer in the UV. We conjecture that differences in the low-mass end of the stellar IMF are related to the nature of the extreme horizontal branch stars generally responsible for the UV flux in ETGs. If so, then UV colour can be used to identify ETGs with particular IMF properties and to estimate {Upsilon}*. We also demonstrate that UV colour can be used to decrease the scatter about the Fundamental Plane and Manifold, and to select peculiar galaxies for follow-up with which to further explore the cause of variations in {Upsilon}* and UV colour.