Hybrids are common, being found in about 25% of all plant species, but the isolating barriers which preserve species integrity are poorly studied. I investigated this question in the large New Zealand genus Celmisia Cass. (Asteraceae), which hybridises readily in cultivation, but wild hybrids are relatively rare. My study quantitatively tests four potential reproductive isolating barriers in 12 sympatric species of Celmisia found in the Craigieburn Range, inland Canterbury, New Zealand. I examined two potential prezygotic reproductive isolating barriers (flowering phenology and pollinator specialisation), and two potential postzygotic barriers (pre-dispersal seed predation and hybrid seed germination). I used null models to test whether Celmisia species had temporally segregated flowering times, and found that some Celmisia are temporally segregated and thus less likely to form hybrids. I used experimental pair-wise flowering arrays to observe insect visitation to six different Celmisia species pairs. While I found no difference in the overall pollinator community, several insect families showed preferences for some Celmisia species. Furthermore, I found that subtle floral character differences were driving these insect preferences. In particular, I found scape height to be positively associated with insect visitation with taller Celmisia being favoured over shorter species. Insect preferences did not translate into strong floral constancy, therefore indicating that Celmisia flower visitors are likely to be a weak barrier to hybridisation. I reared a range of insect seed predators from field-collected capitula of the hybrid C. x pseudolyallii and both parent species (C. lyallii and C. spectabilis). There was no overall difference in the number of seed-predators per capitulum between hybrid and parent Celmisia taxa. I collected and sowed seeds from three Celmisia hybrids and their parent species in order to test whether hybrids were less fertile than their parent species. I found no evidence to suggest that the seeds of hybrids had lower germination success than those of their parents. Overall I found evidence for only weak prezygotic reproductive isolation and no evidence for postzygotic isolation in the four barriers I examined in Celmisia.