Hexagons are an ideal shape for clustering sensor networks, because clustered areas can be seamlessly divided by the hexagons. In addition, hexagons are the largest regular polygon (in terms of the number of sides) that has this property. In this paper, we propose a novel scheme for subdividing a hexagonal cluster where sensors are densely populated and distributed uniformly throughout the cluster. We provide an analytical estimate of power saving resulting from the subdivision, and show that our proposal gives rise to a substantial reduction in power consumption. We also perform subdivisions at various scales, and analyze the corresponding power saving patterns. Our results show that the proposed scheme will yield significant saving in overall power consumption of a cluster, and the deeper the subdivision, the less the power consumption.