Four cassava mosaic geminivirus (CMG) isolates; African cassava mosaic virus from Namulonge, Uganda (ACMV-[Nam]), East African cassava mosaic virus from Mtwara, Tanzania (EACMV-[Mtw]), EACMV-Uganda from Namulonge (EACMV-UG[Nam]) and Indian cassava mosaic virus from Trivandrum, India (ICMV-[Tri]) were compared for their ability to be transmitted by four colonies of cassava whitefly, Bemisia tabaci (Gennadius), collected from Namulonge (NAM), Mtwara (MTW), Kumasi (KUM) and Trivandrum (TRI). With 20 adult whiteflies per test plant, the CMGs from Africa were transmitted by all three of the African B. tabaci populations to 60-79% of the target plants. Indian cassava B. tabaci transmitted ICMV-[Tri] to 89% of the target plants. In contrast, Indian cassava B. tabaci transmitted EACMV-[Mtw] and Tanzanian cassava B. tabaci transmitted ICMV-[Tri] less than a tenth as often, even when using 50 adults per plant and with increased acquisition and inoculation access periods. The complete coat protein genes of the CMGs had sequences typical of their source viruses, the major differences occurring between those originating from India and Africa. Symptom severity of the CMGs was quantified precisely by both visual assessment and image analysis techniques. EACMV-[Mtw] and ACMV-[Nam] were the most and least damaging isolates with 15.4 and 10.0% of the leaf area diseased, respectively. In these tests the transmission frequency was not linked to symptom severity in the source plants. These data support the hypothesis that virus-vector co-adaptation exists in the cassava mosaic disease (CMD) pathosystem and the results are discussed in relation to CMD epidemiology.
