This paper presents an experimental study on subjective quality of 3D video under packet loss conditions. A priority network is assumed, such that the base view of a stereoscopic video stream is delivered through a guaranteed channel and the auxiliary view is subject to packet losses over a low priority channel. Due to the packetisation scheme, one frame is lost whenever a packet is lost. Three frame concealment methods are proposed to deal with frame losses and seven different 3DTV sequences are used. The subjective evaluation study was carried out for different packet loss ratios and concealment methods, using shutter glasses with synchronized display. The experimental results show that sequence disparity, type of loss concealment and packet loss pattern should be jointly considered as relevant quality factors in 3D video over packet networks.