Here's a description from Bill Fitzmaurice I picked up which may be easier to understand:
Two reasons for vertical stacking. One, you can hear them better. Two, the audience can hear them better. That has to do with two other facts. First, the all important midrange frequencies travel in relatively straight lines. The further the speakers are below your ear level the more difficult it is to hear them. Second, the angles of dispersion are inversely proportional to the width and height of the source. A low wide source has narrow dispersion on the horizontal plane, wide dispersion on the vertical plane, and that's the precise opposite of what's desired. Wide horizontal dispersion, for the benefit of the audience, and narrow vertical dispersion, so as not to waste power scerenading spiders on the ceiling and mites on the floor, is what you want, and you get that from a source that is narrow and high.