In class today we looked at the breakup of a drop in a viscous fluid in creeping flow in the absence of surface tension. Theory suggests that while a spherical drop should be neutrally stable at zero Re, any imperfections are swept around to the back and cause it to transition into a toroidal shape. Inertia then causes it to grow in radius, and this larger torus becomes unstable, resulting in two drops forming from the ring. These drops are unstable in turn, leading to a cascade of ever smaller drops. The theory behind this process is described in the paper by Kojima and Acrivos given here.
It actually turns out to be a little tricky to determine the best fluid combinations for a nice demonstration of the effect. When the first demo failed (breakup occurred about the time the drop hit the bottom of the cylinder) I first turned to dimensional analysis to get an idea of the important relationships and then re-analyzed the data of Kojima (and later measurements by Pignatel et al., for clouds of particles) to determine a combination that actually would work. The dimensional analysis is given here and the data analysis is given here.