Most aphids are cyclic parthenogens, so are ideal models in studies of the mechanisms and consequences of sex and recombination. However, owing to a shortage of physical and genetic markers, there have been few studies of the most fundamental genetic processes in these organisms. For example, it is not known whether autosomal segregation during male spermatogenesis is in Mendelian proportions: we address that question here. The aphid Myzus persicae has a typical karyotype of 2n = 12 in females (XX), while males are XO (2n = 11). During male meiosis, only the spermatocytes with an X chromosome are viable. We hypothesized that assortment of autosomes might be non-random because chromosomal imprinting leading to elimination of the paternal autosomes is seen in the closely related coccoids. In other aphid models, we have observed segregation distortions at single microsatellite loci (Wilson, 2000). Such distortions may have nothing to do with 'selfish' behaviour, but may be caused by mutation accumulation causing fitness differentials. Thus single-locus distortions might be predicted to be more likely to be detected via the male lines of clones that have lost the ability to reproduce sexually (male-producing obligate parthenogenesis (androcyclic)). Using microsatellites we show that genetic imprinting or selfish autosome behaviour does not occur in male M. persicae. Generally, loci segregated in Mendelian proportions in both sexes of cyclically parthenogenetic (holocyclic) clones. However, in androcyclic clones, segregation distortions consistently involved the same two autosomes. This is consistent with linkage of markers to deleterious mutations associated with a loss of sexual reproduction.