What stars become supernovae

A variety of empirical lines of evidence is assembled on the masses and stellar population types of stars that trigger supernova (SN) explosions. The main theoretical motivations are to determine whether type I supernovae (SN I) can have massive precursors, and whether there is an interval of stellar mass, between the masses of precursors of pulsars and white dwarfs, that is disrupted by carbon detonation. Statistical and other uncertainties in the empirical arguments are given particular attention, and are found to be more important than generally realized. Relatively secure conclusions include the following. Statistics of stellar birthrates, SN, pulsars, and SN remnants in the Galaxy show that SN II (or all SN) could arise from stars with masses greater than M/sub s/ where M/sub s/ approximately 49 to 12 M solar mass; the precursor mass range cannot be more closely defined from present data; nor can it be said whether all SN leave pulsars and/or extended radio remnants. Several methods of estimating the masses of stars that become white dwarfs are consistent with a lower limit, M/sub s/ greater than or equal to 5 M solar mass, so carbon detonation may indeed be avoided, although this conclusion is not secure. Studies of the properties of galaxies in which SN occur, and their distributions within galaxies, support the usual views that SN I have low-mass precursors (less than or equal to 5 M solar mass and typically less than or equal to 1 M solar mass) and SN II have massive precursors (greater than or equal to 5 M solar mass); the restriction of known SN II to Sc and Sb galaxies, to date, is shown to be consistent, statistically, with massive stars in other galaxies also dying as SN II. Possible implications of the peculiarities of some SN-producing galaxies are discussed. Suggestions are made for observational and theoretical studies that would help answer important remaining questions on the nature of SN precursors