The primary mode of action of petroleum spray oils (PSOs) on pest insects is through direct contact. Indirect effects are, however, also possible, and deposits of the oils may influence pest populations by killing insects and/or by influencing their behaviour. The indirect effects of deposits of a new nC24 oil against the cotton aphid, Aphis gossypii Glover, were therefore determined. The effects of oil deposits on the acceptance of cotton as host plant by the aphids were assessed, as well as aphid mortality rates and their success in the establishment of colonies. The efficacy of deposits of a heavier oil (nC27) was also evaluated. Deposits of PSO were toxic to A. gossypii and remained effective until 8 days after spraying. Mortality decreased with time, so that, the older the deposit, the lower was the mortality. Significantly higher aphid mortalities were achieved on younger leaves than on mature ones. Thus, leaf age proved a significant factor in the efficacy of the deposits. Consecutive prophylactic applications (at 9 day intervals) did not have a cumulative effect, and their killing power proved to be independent of one another. Thus, applying the oil prior to aphid infestations would confer only minimal protection. The mortality inflicted by the deposits was not improved by increasing the molecular mass of the oil used (nC27 oil), but the toxic life of the oil deposit was increased. Oil deposits did not deter alates from landing on oil-sprayed plants. Oil deposits did, however, affect subsequent alate and nymphal survival, and thus the establishment of aphid colonies. The impact that the oils could have on the longer-term development of aphid populations in the field was thus demonstrated. First- and second-instar nymphs were the most susceptible life stages, with > 50% mortality compared with < 10% for the other stages. These nymphs did not show the typical signs of oil-induced mortality observed in aphids killed by direct oil applications, which suggests an alternative mode of action to that of the directly applied oil. Anoxia does not seem to be involved in either process, and alternative modes of action of the oil deposits are discussed. The implications of these findings for cotton aphid control are also considered, primarily in relation to the timing and frequency of oil application.