OK MACCA350 - you are right!
But that has to do with loudpseaker efficiency and a whole lot of other factors. I should perhaps have mentioned more clearly that I was talking most basically and theoretically, just to get the matter of amplitude vs. input power right - one thing at a time. I hope this did not also lead No 5 astray - if so my apologies. I understood his question to be about basics.
What I meant was that IF you put constant power into an ideal driver, as would be the case e.g. when testing amplifier amplitude response vs. frequency into a fixed load resistor, the basic inverted ratio of loudspeaker mechanical amplitude vs frequency followed. (The power is constant; the resistor would get as hot from 10KHz, as it would from 50 Hz, with constant input signal.) But because of efficiency, hearing characteristics, propagation, etc. your ears would fall off in a manner of speaking if you have to listen to that. It is true that in practice tweeters are generally more efficient than low frequency units - and please allow for "in practice". I do not want to make a simple sweeping statement where so many factors play a role! Thus when one starts to introduce the actual loudness perceived or measured, the picture can change dramatically. E.g. the impedance of h.f. units often rises with frequency, also causing a decrease in current drawn, etc., despite which they still sound louder. Then also, for "balance", orchestral levels (at least in classic music) at h.f. is much lower that at mid and low frequencies etc. etc. - so please don't get me wrong there!
Regards.