We present mass and radius derivations for a sample of very young , mid- to late M , low-mass stellar and substellar objects in Upper Scorpius and Taurus .
In a previous paper , we determined effective temperatures and surface gravities for these targets , from an analysis of their high-resolution optical spectra and comparisons to the latest synthetic spectra .
We now derive extinctions , radii , masses and luminosities by combining our previous results with observed photometry , surface fluxes from the synthetic spectra and the known cluster distances .
These are the first mass and radius estimates for young , very low mass bodies that are independent of theoretical evolutionary models ( though our estimates do depend on spectral modeling ) .
We find that for most of our sample , our derived mass-radius and mass-luminosity relationships are in very good agreement with the theoretical predictions .
However , our results diverge from the evolutionary model values for the coolest , lowest-mass targets : our inferred radii and luminosities are significantly larger than predicted for these objects at the likely cluster ages , causing them to appear much younger than expected .
We suggest that uncertainties in the evolutionary models - e.g. , in the choice of initial conditions and/or treatment of interior convection - may be responsible for this discrepancy .
Finally , two of our late-M objects ( USco 128 and 130 ) appear to have masses close to the deuterium-fusion boundary ( \sim 9–14 Jupiters , within a factor of 2 ) .
This conclusion is primarily a consequence of their considerable faintness compared to other targets with similar extinction , spectral type , and temperature ( difference of \sim 1 mag ) .
Our result suggests that the faintest young late-M or cooler objects may be significantly lower in mass than current theoretical tracks indicate .