Computing rumours!

Prithee, listen well ;
I heard a bustling rumour, like a fray,
And the wind brings it from Capitol.
Shakespeare (Julius Ceaser)

American scientist runs a regular computing science column: in the May-June 2005 issue, Brian Hayes , in his essay titled "Rumours and Errours" discusses the modelling of rumour propagation and the close relationship between rumour models and models of epidemic diseases. Apparently, if there is one spreader initially, the fraction of population that would fail to hear the rumour is 20% (No wonder Lucius heard nothing when Portia heard 'bustling rumour, like a fray'). On the other hand, if the number of spreaders are more, more people fail to hear it (the upper limit being 36.8%). Why is this so? To know the answer to that question, or to write a code that would help you solve the mystery yourself, turn to 'Rumours and Errours'. And, forget not to spread the news!