The undeformability of the primary mirror is a main requirement for the optical quality of a telescope. While thermal deformation effects are prevented by the use of quasi-zero thermal expansion materials such as the Xerodur glass, the effect of static (the change of gravity orientation with respect to the mirror) and dynamic loads (wind and inertia loads) used to be minimized by making the primary mirrors quite thick in proportion to their diameter.
The recent development of active mirror support systems, generally called active optics, has changed this situation. These systems are mainly aimed at maintaining the shape of the mirror through the change of gravity orientation during tracking and can also correct slight alignment and focusing errors. As the shape of the mirror can be adjusted very finely by active actuators that apply controlled forces, the mirror itself does not need to be very rigid. Moreover a too high stiffness would require higher control forces.
Adding to this, there are presently technological limitations in the manufacturing of 8-m glass mirror that prevent them from having a thickness of more than 20 cm. Therefore the primary mirrors of the new 8-m telescopes of the latest generation have a relatively low stiffness which results in a larger sensitivity of their shape to externally applied forces. For instance the mirrors of the VLT 8-m telescopes will be intrinsically 37 times less rigid than the one of the 3.5-m NTT (the first telescope with active optics), which is already four times less stiff than the mirrors of predecessor telescopes in the 4-m class.
The lower stiffness of the new "thin" mirrors makes their shape much more sensitive to a varying wind loading. In fact, while the effects of constant or quasi constant pressures can be corrected in closed loop with the active optics system, this has a bandwidth limited by a number of reasons to a period of the order of 1 minute. Pressure variations over shorter times cannot be compensated and will result in dynamic deformations of the mirror and consequent optical aberrations. These are quantified in two ways.
One figure of merit is the rms
of
the surface displacements, hence of the
deformation 
of the wavefront, which is
generally expressed in nanometers.
The other one is the rms
of the slope of the wavefront 
, expressed in arcsec,
which is
also equal to the rms size of the image PSF and then directly comparable
to the FWHM used for the quantification of seeing.
If the main modal components of the deformation are known, the two
quantities can be related by analytical or numerical computations.