A communication device such as a video telephone is connected to a telephone line together with a telephone set at both parties. In response to an operation for removing a handset from a hook of the telephone set at this party, the video telephone at this party is initiated. Each of the video telephones includes a mode selecting switch and DTMF receiver which receives a DTMF signal from one of the telephone sets and decodes a received DTMF signal to output data representative of the mode of the video telephone. When a flag "1" is set, a mode of the video telephone at this party can be controlled in response to the DTMF signal which is received by the DTMF receiver from the telephone set at this party and, when a flat "2" is set a mode of the video telephone at this party can be controlled in response to the DTMF signal which is received by the DTMF receiver from the telephone set at the other party.