A possible mechanism for the transition between bubbly and slug flow is proposed and tested in a simulated slug flow system. No sudden collapse of slug flow with increasing velocity is found and it is concluded that: a. Slug flow is generally stable at voids greater than 35%. b. Bubbly flow at voids higher than this is a result of entrance conditions. c. Visual observations of bubbly flow in unheated systems at higher voids are most likely faulty. No simple asymptote limit or criterion that would predict the location of the bubbly-slug transition was found.