Communicated by P. M. A. Tigerstedt J. Whitton ( ) · D. E. Wolf · L. H. Rieseberg Department of Biology, Indiana University, Bloomington, Indiana 47405, USA D. M. Arias Centro de Investigacio´n en Biotecnologi´a, Universidad Auto´noma del Estado de Morelos, Cuernavaca, Morelos, Mexico 62210 A. A. Snow Department of Plant Biology, The Ohio State University, 1735 Neil Avenue, Columbus, Ohio 43210-1293, USA Theor Appl Genet (1997) 95 : 33— 40  Springer-Verlag 1997 J. Whitton · D. E. Wolf · D. M. Arias A. A. Snow · L. H. Rieseberg The persistence of cultivar alleles in wild populations of sunflowers five generations after hybridization Received: 1 November 1996 / Accepted: 17 January 1997 Abstract The development of transgenic plants has heightened concern about the possible escape of geneti- cally engineered material into the wild. Hybridization between crops and their wild relatives provides a mech- anism by which this could occur. While hybridization has been documented between several crops and wild or weedy relatives, little is known about the persistence of cultivar genes in wild populations in the generations following hybridization. Wild and weedy sunflowers occur sympatrically with cultivated sunflowers throughout much of the cultivation range, and hybrid- ization is known to occur. We surveyed two cultivar- specific RAPD markers in 2700 progeny in a naturally occurring population of wild Helianthus annuus over five generations following a single generation of hybrid- ization with the cultivar. Moderate levels of gene flow were detected in the first generation (42% hybrids at the crop margin) and cultivar allele frequencies did not significantly decline over four subsequent generations. These results indicate that gene flow from cultivated into wild populations of sunflowers can result in the long-term establishment of cultivar alleles in wild populations. Furthermore, we conclude that neutral or favorable transgenes have the potential to escape and persist in wild sunflower populations. Key words Gene flow · Hybridization · Sunflowers · Transgenes Introduction Over the last decade, a dramatic increase in the number of requests for field trial permits for transgenic plants has been accompanied by increasing concern about the potential negative effects of the wide-scale commercial release of genetically engineered crops (Kareiva 1993; The Gene Exchange 1993; Seidler and Levin 1994). Potential negative consequences include the possibility that transgenic crops will become feral, or that the transgenes will themselves escape via hybridization with wild or weedy populations of crop relatives (Keeler and Turner 1990; Rees et al. 1991; Raybould and Gray 1994). Escape by either means may have serious agricultural and/or ecological consequences, the severity of which will depend primarily on the nature of the engineered trait and its ability to persist and spread in the environment. The act of introducing genetically engineered mater- ial, in and of itself, is unlikely to result in major changes in life-history traits of crops (Miller and Gunary 1993). Thus, the risk of a transgenic crop escaping cultivation is likely to be highest in crops where non-transgenic varieties have weedy tendencies, or where the engineer- ed trait is likely to improve invasive tendencies (Keeler 1989; Keeler and Turner 1990). This is supported by recent work in which Crawley et al. (1993) found no significant increase in the invasiveness of transgenic oilseed rape (Brassica napus) under a wide variety of environmental conditions. It should, however, be noted that the genetically engineered traits carried by these cultivars (resistance to the antibiotic kanamycin and tolerance of the herbicide glufosinate) were not ex- pected to be favorably selected under the conditions of the experiment. This limitation in the design of this and other experiments has been noted by Miller and Gunary (1993) and others (e.g., Kareiva 1993), who state that in order to assess the potential deleterious consequences of escaped transgenic plants, experi- mental conditions should mimic a plausible scenario