In previous work (L. Karlsson et al., Journal of Applied Polymer Science, 2002, 2003, Vol. 90, pp. 905-915; J. M. Stubbs et al., Colloids and Surfaces A: Physiochemical and Engineering Aspects, 1999, Vol. 153, pp. 255-270) we have explained the development of composite particle morphologies produced by seeded emulsion polymerization in terms of the ability of second stage polymer radicals to diffuse into, or "penetrate," the seed particles. This has been quantified (Stubbs et al., ibid) by calculating so-called "fractional penetration" values for the second-stage radicals. In this article the effect of the second-stage initiator type, specifically nonionic vs ionic initiators, on particle morphology is investigated. The question to be answered is whether charged (ionic) end groups (from the initiator) on second stage polymer chains "anchor" to the particle surface, making it more likely to form core-shell morphologies. This is investigated by using a poly(methyl acrylate-co-methyl methacrylate) [P(MA-co-MMA)] seed latex and polymerizing styrene in the second stage in a semibatch manner using various feed rates of styrene. At each feed rate one reaction was conducted using potassium persulfate as the initiator, which produces charged end groups, and another using VA-086 initiator, which produces uncharged end groups. The morphologies of the resulting particles were then observed by transmission electron microscopy. It is shown that under some conditions ionic initiators do make it more likely to obtain core-shell morphologies, but that this effect is not dominant under most conditions. The resulting morphologies agree quite well with the predictions of the fractional penetration calculations. (C) 2003 Wiley Periodicals, Inc. J Appl Polym Sci 91: 1538-1551, 2004.