We present the results of a 3 year monitoring program of a sample of very low mass ( VLM ) field binaries using both astrometric and spectroscopic data obtained in conjunction with the laser guide star adaptive optics system on the W.M .
Keck II 10 m telescope .
Among the 24 systems studied , fifteen have undergone sufficient orbital motion , allowing us to derive their relative orbital parameters and hence their total system mass .
These measurements triple the number of mass measurements for VLM objects , and include the most precise mass measurement to date ( < 2 \% ) .
Among the 11 systems with both astrometric and spectroscopic measurements , six have sufficient radial velocity variations to allow us to obtain individual component masses .
This is the first derivation of the component masses for five of these systems .
Altogether , the orbital solutions of these low mass systems show a correlation between eccentricity and orbital period , consistent with their higher mass counterparts .
In our primary analysis , we find that there are systematic discrepancies between our dynamical mass measurements and the predictions of theoretical evolutionary models ( TUCSON and LYON ) with both models either underpredicting or overpredicting the most precisely determined dynamical masses .
These discrepancies are a function of spectral type , with late M through mid L systems tending to have their masses underpredicted , while one T type system has its mass overpredicted .
These discrepancies imply that either the temperatures predicted by evolutionary and atmosphere models are inconsistent for an object of a given mass , or the mass-radius relationship or cooling timescales predicted by the evolutionary models are incorrect .
If these spectral type trends are correct and hold into the planetary mass regime , the implication is that the masses of directly imaged extrasolar planets are overpredicted by the evolutionary models .