minor key signatures misleading

You don't have to think about it too hard to realize that key signatures are really major-key oriented affairs. They do in fact and 'to a T' accurately describe the makeup of a major key: C major has no sharps or flats, F-sharp major has six sharps (all but B) and A-flat major has four flats and so on.

But when it comes to the minor keys it's a whole different story. Well, if we restrict ourselves to the natural minor (aeolian) keys then it isn't a different story. But historically, even in Renaissance music where they did use the 'Church modes', sharps and flats entered into cadences. So much so that G# really did become a part of A minor. And melodically so did F# (hence the name melodic minor).

So, in effect, if we're talking about the melodic minor the key signature of its relative major is likely to be 'off' by 2 accidentals: A minor's key signature has no sharps or flats, but A melodic minor has 2 sharps; C minor's key signature has 3 flats but C melodic minor has only one flat (Eb).

Now, none of this may be an issue if you're playing Bach, etc (though in the older Baroque music the key signatures really hadn't been worked out, exactly because of the issue we're discussing), but it is an issue if you think key signatures when you consider modes from an improvisational/compositional point of view. For example: C melodic minor and E minor might seem quite distant by key signatures (3 flats and 1 sharp respectively), but in fact there's much more overlap: the only accidentals are Eb and F#.

Consider also the difference between a melodic minor key and its parallel major. A melodic minor, for example, only differs from A major by 1 tone: the 3rd (in this case C). Obviously if you think of scales/modes/keys from a numeric skeletal point of view then this won't ever be a problem. But we're generally taught the key signature approach at first, so in a sense there's a bit of unlearning to do.