Melody is not "a sequential production of tones."  An Alberti bass is not a melody.  There are literally an infinite number of examples that can illustrate the problems of this definition.

Harmony is not two tones sounded simultaneously.  There are two distinct problems with this definition.  First, the number of tones -- you really do need three, at least if you're talking about functional harmony.  You need that third tone to give you your triad, which as we all learned in music theory is the basis of functional harmony.  In context, two tones can suffice but you have to have the right context.  Second, an arpeggio is not sounded simultaneously but you get your harmony anyway.

Two melodies playing together do not necessarily make counterpoint -- they will (if they are nonparallel melodies) make some form of polyphony, though.  

Bach did not employ five melodies simultaneously, but he did write the occasional five-voice fugue.  There is the quodlibet but that's only four melodies.  :-)

I will mercifully stop now.  I hope the following negates any pain I may have caused:

<object width="425" height="355"><param name="movie" value="http://www.youtube.com/v/fmoEaiV0b3A&hl=en"></param><param name="wmode" value="transparent"></param><embed src="http://www.youtube.com/v/fmoEaiV0b3A&hl=en" type="application/x-shockwave-flash" wmode="transparent" width="425" height="355"></embed></object>

here's the score in PDF for anyone who wants to follow along:  http://www.icking-music-archive.org/scores/g.gabrieli/gabcan2part.pdf