1. ¹Department of Plant and Microbial Sciences, University of Canterbury, Christchurch, New Zealand
2. ²Institute of Botany, Academy of Sciences of the Czech Republic, CZ-252 43 Pruhonice, Czech Republic

Correspondence: S Trewick, Department of Plant and Microbial Sciences, University of Canterbury, Christchurch, New Zealand. E-mail: s.trewick@massey.ac.nz

³Current address: Allan Wilson Centre for Molecular Ecology and Evolution, Massey University, Private Bag 11-222, Palmerston North, New Zealand.

Received 20 January 2003; Accepted 4 March 2004; Published online 12 May 2004.

Abstract

Hieracium pilosella (Asteraceae) was accidentally introduced to New Zealand about 100 years ago. Since then it has become an aggressive weed, and an unexpected degree of genetic and genome size variation has been detected; features that might result from interspecies hybridization. We investigated the possibility that H. pilosella has hybridized with related taxa. Of the four other subgenus Pilosella species introduced to New Zealand, H. praealtum is the most abundant and, on morphological and distributional evidence, most likely to be the other parent. Flow cytometry was used to estimate relative genome size for 156 Hieracium plants collected from the wild. Plants assigned to either parental or hybrid morphotypes were found to comprise tetraploid and pentaploid individuals using genome size measurements, and this was confirmed with direct mitotic chromosome counts for a subset of plants. The haploid DNA content of H. praealtum was approximately 22% larger than that of H. pilosella. Putative hybrids that were tetraploid had mean genome sizes equivalent to two H. pilosella and two H. praealtum haploid chromosome sets, implying they were hybrids arising from the fertilization of two reduced gametes. Similar results were obtained from tetraploid hybrids produced by controlled pollination. However, the majority of field hybrids were pentaploid with a genome size equivalent to four H. pilosella and one H. praealtum haploid chromosome sets. We infer that these are not first-generation hybrids but represent successful backcrossing with H. pilosella and/or hybrid–hybrid crossing, and that sexual tetraploid hybrids have been the parents. We note that populations putatively of H. pilosella often comprise apomictic pentaploid hybrids. Significantly, our data indicate the emergence of sexual hybrids that provide further opportunity for gene flow among taxa in this complex.