There is limited understanding of how fire-related cues such as heat shock and smoke can combine to affect the germination response of seeds from fire-prone vegetation because combinations of multiple levels of both cues have rarely been investigated. Germination response surfaces were determined for the combination of heat shock and smoke by applying factorial combinations of temperature (up to 100ºC) and aerosol smoke (0-20 min) to 16 species that form soil seed banks in the Sydney region of south-eastern Australia. Duplicate populations of three species were also examined to assess the constancy of a species response surface. Of the 19 populations examined, 16 showed a germination response to both the fire cues, which combined interactively in 14 populations, and independently in two. No population responded only to a single cue; however, seeds of 11 populations responded to heat in the absence of smoke, and nine responded to smoke in the absence of heat. Heat applied in the absence of smoke negatively affected germination in seven populations, either progressively as temperature increased, or above a set temperature. Negative germination responses over part of the temperature range were fully reversed at higher temperatures for unsmoked seeds of four populations (curvilinear heat response). Smoke effects were most frequently positive over all or part of the range of durations used, and when combined with heat frequently fully or partially reversed negative heat effects. Three populations required the obligatory combination of smoke and heat. A novel response to the cues was observed for three species, with smoke reversing negative heat effects at 75ºC, being supplanted by a positive heat response of unsmoked seed at 100ºC. The response surface for duplicate populations of two of the three species examined was variable. Heat shock and smoke frequently combined to affect germination, in both positive and negative ways. Consequently, to gain an accurate assessment of the response of seeds to fires, an experimental design that samples within the potential response zones of germination cues is essential. © 2007 Ecological Society of Australia.