Blackman (1987) used life cycle and morphology to separate Myzus nicotianae Blackman, a tobacco-feeding species of aphid, from Myzus persicae (Sulzer). In the present study, the first objective was to investigate the influence of temperature and host plant on the morphology of M. nicotianae and M. persicae. The second objective was to assess Blackman's 1987 key to Myzus for separating tobacco and non-tobacco originating morphs under different environmental conditions. Four host plants were used: tobacco, turnip, pepper, and okra, and three temperatures, 15°C, 20°C, and 25°C. The intraclonal plasticity of two tobacco collected morphs and one turnip collected morph was investigated in relation to these combinations of host and temperature in a 4 x 3 x 3 factorial experimental design. Fifth generation mature apterous aphids were mounted on slides and 10 different morphological structures utilized in morphometric analysis were measured.

Data support a morphologically distinct, host-adapted tobacco race but not a separate tobacco-feeding species of M. persicae. The key developed by Blackman (1987) did not discriminate between the tobacco and non-tobacco originating clones but the canonical variates generated from the analysis successfully separated the tobacco and non-tobacco groups. Other studies have used many different clones to investigate the possible distinctions between M. persicae and M. nicotianae; the objective here was to see how much morphological perturbation may be induced within a clone by rearing at different temperatures and on different host plants.

Temperature and host plant had substantial influences on the morphology of these aphids. The physiological interactions of temperature-host plant-aphid morphology are very complex yet controlling only for temperature and host plant was sufficient to group specimens according to these independent variables with remarkable accuracy using the linear discriminant functions generated with these data. Percent of aphids in which rearing temperature was correctly identified using linear discriminant functions generated for temperature classes was 87%, 63%, and 64% for 15°C, 20°C, and 25°C, respectively. Random designations would be 33%. Correct identification of host plant was 65%, 45%, 47%, and 48% successful for tobacco, turnip, pepper, and okra, respectively. Random designations for host plant would be 25%.

Canonical variates produced clusters by host, temperature, morph, and combinations of these independent variables with varying degrees of discreteness. CV1 by CV2 for host plants gave a very distinct cluster for tobacco and also separate groupings for aphids reared on turnip and pepper. Aphids from the host plant okra were scattered quite widely across the CV1 by CV2 graph. CV1 by CV2 for temperature conditions showed a tight cluster for aphids from 15°C and still distinct though less closely grouped clusters for both 20°C and 25°C rearing temperatures. CV1 by CV2 for the three morphs gave substantial overlap for the two tobacco originating morphs and a more separate cluster for the morph originally collected from turnip.