Post-Fire Recruitment of Four Co-Occurring Banksia Species

(1) The factors responsible for genet mortality and survival after autumn and spring burns were studied for Banksia attenuate R. Br., B. leptophylla A.S. George, B. menziesii R. Br. and B. prionotes Lindley, in a 15-year-old stand of fire-prone scrub-heath in southwestern Australia. (2) Late summer to winter seed release from unburnt cones was recorded for B. menziesii (30%) and B. prionotes (3%0,) only. One hundred and forty days after both burns, more than 950%0 of seeds of all species were released, except for B. leptophylla which retained 17% of seeds in cones throughout the summer after the spring burn. The rate of seed release from burnt cones was significantly slower for all species after the spring than the autumn burn. (3) Post-dispersal seed predation after the autumn burn was negligible whereas it ranged from 90%/0 (B. prionotes) to 98%/0 (B. menziesii) after the spring burn. (4) Seed of all species germinated optimally at 15-20 'C. After the autumn burn, field germination inside enclosures ranged from 9% (B. menziesii) to 69% (B. leptophylla). More seeds of all species germinated in the unburnt than in the autumn burn site. No seed sown after the spring burn had germinated by the end of the following summer and their viability was severely reduced. (5) Most (54-97%/,) seedlings which established after the autumn burn had died, largely as a result of predation, by the end of the summer. Predation intensity was probably unnaturally high, due to the small size (c. 2 ha) of the resource-rich, autumn-burnt patch relative to the vast tracts of adjacent unburnt scrub-heath. Significantly fewer seedlings died inside enclosures after the autumn burn (32-520%0) than in the unburnt site (71-80%/,), indicating the importance of drought-induced mortality in maintaining even-aged populations in mature scrub-heath. (6) By the end of the first winter, the number of seedlings recruited per parent of all species was more than twice as high after the autumn than after the spring burn. More than 2 years after both burns, seedling replacement levels were higher than pre-burn adult densities for all species except B. menziesii. There was no recruitment in the unburnt area over the same period. (7) In order to maintain the populations of the Banksia species, extensive tracts of scrub-heath should be burnt in autumn, prior to the onset of winter rain. Conflicts between property protection and conservation cannot be resolved without information on the fireregime effects on scrub-heath populations.