If there are three kinds of mass
(1) resistance (classical inertial mass, rest mass)
(2) sensitivity to gravity (weight)
(3) creation of gravity,
the Eotvos experiment shows the proportionality between (1) and (2).
Can we think the Cavendish experiment and the slowing of binary solar stars show the proportionality between (1) and (3) implicitly?
Or I suspect the three definitions have some flaw.