hi
as my memory serves me. by resonate frequency you have by definition created a standing wave form in the object or device. eg that bridge over the canyon you forever see on TV. It stands to reason that If you remove the stimulus (and have no other source of energy) the wave form will decay. Now how much damage is done getting to a standing wave and decaying back to stasis is in the lap of the gods.

If I'm wrong would some of our residdent gurus clarify.

regards Don

Comment: It depends. There are different resonances, e.g. subharmonic resonance, super (or higher) harmonic resonance, interharmonic resonance, intermittent resonance, etc.
Theoretically and practically, the resonance may cause deviation from its center resonating (or vibrating) position due to other factors, e.g. mechanical support, material fatigue, up to a complete breakdown.

I used to work for a loudspeaker manufacturer some years ago. The resonance we used to refer to was the mechanical resonance of the speaker cone/coil combination, where the cone provided the energy storing "spring" and the coil/cone combo was the "mass" suspended and oscillating on the spring. The air provided most of the "friction" which removed the energy from the system as soundwaves, with the cone and coil each taking a bit away as heat energy.

We measured resonance by putting a breif pulse ("pop" sound) on the speaker coil, then having a microphone pick up the resulting (rapidly decaying) tone, which fed a triggered audio frequency counter which read out the resonant frequency. All this was in an unsealed padded (baffled) enclosure to eliminate factory noise and to be accoustically non-resonant, but a quiet room would also work.

At the specified resonance, a minimum amount of energy is needed to make sound with the speaker - very important to the tiny amps and speakers used in battery powered talking toys, themometers, etc. The small size of speaker required is a main reason why female or child voices are preferred for these applications.

Alternatively, for relatively loud stereo speakers, tuned enclosures and multiple speakers are used to flatten out the resonance somewhat so that a wide spectrum of frequencies will be heard at about the same sound level. The speaker enclosure designer needs to know the resonant frequencies in order to design a speaker system with no holes or peaks in its spectrum.

Kevin VanZuilen "KevinVZ"

well, I don't have much to add after what's said, about this subject you can't fill two lines in witch there sais it all...
what I can tell you is that the speaker itself is as you probably know a electrical->mechanical->acoustical transducer so it has different loads:electrical(impedance),mechanical(membrane,bobin...)and acoustical(air). they all add together resulting such lower efficiency of the speaker. anyway, the fact is that at one point, loads have a minimum: the impedance and the air, (since the mass of the mechanical stuff is the same all the time), the impedance given by the electrical factors and the air-load by a sum of "stuff"(in wich: the area of the membrane in proportion with the length of the frequency wave, its mass,...,...); at this point(read frequency), the equivalent electrical load is minimum so physically, the membrane gets a pretty high amplitude of vibrations in compare to the response at other frequencies, and the result is the speaker goes nuts and, if you're lucky(we're talking at a considerable amount of power), either it will descentrate or will go boom!. so this chapter is a don't-try-this-at-home-folks!.
to avoid this search for an impedance  correction(series/parallel Notch filter, Zobel network) documentation. it shold be easy to find.