An account is given of experiments on the aphistransmission of the composite virus disease of tobacco known as rosette. It is shown that the component viruses—the vein-distorting and mottle viruses—persist for long periods in the body of the aphis, and that as many as twenty consecutive tobacco plants can be infected in 24 hr. serial transfers without the insect having recourse to a fresh source of virus.

The aphis Myzus persicae is the most efficient vector, the aphis M. convolvuli being less effective. The aphides M. circumflexus and Macrosiphum gei failed to transmit the virus.

It is shown that the mottle virus can only be transmitted by the aphis when it is accompanied in the plant by the other component of the rosette disease, the vein-distorting virus. The latter virus is easily transmitted by the aphis, either alone or in combination with the mottle virus.

The writer's best thanks are due to Miss Margaret Short for her assistance with the large numbers of aphis-transmissions, and to Dr Douglas Lea and Dr Roy Markham for suggestions in the course of the work.