A capillary droplet is a liquid/gas interface with a spherical-cap shape as determined due to a dominant surface tension force. Predicting how surface-tension redistributes liquid volume is important for a number of small-scale applications. The application of our interest – “reversible super-adhesion” – involves manipulating a large system of coupled droplets to make and break “bonds” with a substrate. We are thus motivated to study the dynamics of a system of n identical droplets coupled by pressure. No matter what the topology of the coupling network, one droplet will eventually take the all the volume from the rest of the community; this is the winner-takes-all scavenging phenomenon. Predicting which droplet wins is our focus. If, at the initial instant for a fully-connected network, the volume is equi-partitioned between the n-droplets, then each droplet is equally likely to win. If, on the other hand, the initial distribution of volume is slightly biased, then one droplet becomes favored and a winner emerges. What may be surprising is the way that the identity of the winner under this initial condition can change depending on the topology of the coupling network. We illustrate with computational observations and explain the observed behavior by computing the boundaries between competing domains-of-attraction for a variety of network topologies (small n).