This note studies the consensus problem of a class of multi-agent systems (MASs) with linear dynamics, where each agent is able to use the specified measurement output (e.g., position) relative to its neighbors at certain sampling instants. It is assumed that the network has switching but jointly-connected topologies in the sense that the union of the associated graphs over a certain time interval always has a spanning tree. Under this assumption, a linear sampled measurement output feedback controller is proposed to achieve full state consensus for a sufficiently small sampling period.