We show that long–period dwarf novae offer a promising route for making Type Ia supernovae .
For typical dwarf nova duty cycles d \sim 0.1 - 0.01 , mass is accreted by the white dwarf mainly during dwarf nova outbursts at rates allowing steady nuclear burning of most of the accreted matter .
Mass gains up to \sim 0.4 M _ { \odot } are possible in this way .
Although these are too small to allow a 0.7 M _ { \odot } WD to reach the Chandrasekhar mass , they are sufficient if the WD grew to \ga 1 M _ { \odot } in a previous episode of thermal–timescale mass transfer , i.e .
for those long–period dwarf novae which descend from supersoft binaries .
A further advantage of this picture is that the supernova always occurs in a binary of small secondary/primary mass ratio , with the secondary having very little remaining hydrogen .
Both features greatly reduce the possibility of hydrogen contamination of the supernova ejecta .