Observed fractions of core-collapse supernova types and initial masses of their single and binary progenitor stars

We analyse observed fractions of core-collapse SN types from the Lick\nObservatory SN Search, and we discuss corresponding implications for massive\nstar evolution. For a standard IMF, observed fractions of SN types cannot be\nreconciled with expectations of single-star evolution. The mass range of WR\nstars that shed their H envelopes via their own mass loss accounts for less\nthan half the observed fraction of SNeIbc. Progenitors of SNeIbc must extend to\na much lower range of initial masses than classical WR stars, and we argue that\nmost SNIbc and SNIIb progenitors must arise from binary Roche-lobe overflow.\nSNeIc still trace higher mass and metallicity, because line-driven winds in the\nWR stage remove the He layer and propel the transition from SNIb to Ic. Less\nmassive progenitors of SNeIb and IIb may not be classical WR stars; they may be\nunderluminous with weak winds, possibly hidden by overluminous mass-gainer\ncompanions that appear as B[e] supergiants or related objects having aspherical\ncircumstellar material. The remaining SN types (II-P, II-L, and IIn) are\nredistributed across the full range of initial mass. We consider direct\ncollapse to black holes without visible SNe, but find this problematic. Major\nareas of remaining uncertainty are (1) the influence of binary separation,\nrotation, and metallicity, (2) mass differences in progenitors of SNeIIn\ncompared to SNeII-L and II-P, and (3) SNeIc arising from single stars with\neruptive mass loss, its dependence on metallicity, and how it relates to\ndiversity within the SNIc subclass. (abridged)\n