The PolyToMono node mixes a polyphonic input (denoted by a thick wire) to a monophonic output (thin wire). Each voice is mixed equally.
Typically, you'll want to place linear effects (reverb, delay, EQ) after the PolyToMono, since it will sound the same as placing them before but only a single voice needs to be processed. On the other hand, nonlinear effects, such as Distortion will have quite a different effect if placed before the PolyToMono versus after.