THE HUBBLE DIAGRAM FOR SUPERNOVAE OF TYPE IA .2. THE EFFECT ON THE HUBBLE CONSTANT OF A CORRELATION BETWEEN ABSOLUTE MAGNITUDE AND LIGHT DECAY-RATE

New Hubble diagrams in B and V are derived for supernovae of type I ba sed on light curves from the archive literature plus 13 new light curv es with superior modern photometry observed in the Cerro Tololo/Univer sity of Chile program (Hamuy et al. 1995). The sample is restricted to SNe Ia whose light curves are defined by photometry beginning 5 days or less after maximum light and with (B-V)(max) < 0.5 mag. Supernovae of known type Ib or Ic are also excluded. The resulting Hubble diagram s, extending to redshifts of 30,000 km s(-1), have dispersions in abso lute magnitude of 0.34 mag in B and 0.33 mag in V, confirming that spe ctroscopically ''normal'' (Branch et al. 1993) SNe Ia are among the be st standard candles known. A solution for the slope of the Hubble diag ram gives n(B) = 0.977 +/- 0.025 and n(V) = 1.020 +/- 0.024 for the ex ponent in upsilon similar to D, proving linearity of the expansion fi eld to a high level. The residuals in magnitude from the ridge line of the Hubble diagram are compared with the light decay rate during the first 15 days to test the correlation between the two suggested by Psk ovskii and by Phillips. The strongest possible correlation using the e xtant data has a slope 3 times smaller than that derived by Phillips, and 2 times smaller than suggested by Hamuy et al., leading to a decre ase of less than 10% in the distance scale based on the present (1995) SNe Ia calibration by means of three supernovae whose distances are k nown from Cepheids in their parent galaxies. Applying the maximum poss ible correction to M(max) for a Pskovskii-Phillips effect would give H ubble constants of H-0(B) less than or equal to 54 +/- 4 km s(-1) Mpc( -1), and H-0(V) less than or equal to 59 +/- 4 km s(-1) Mpc(-1), where the errors are internal. It is argued that the absence of measurable bias effects in the Hubble diagrams shows that the three local (neares t) SNe Ia presently calibrated via Cepheid distances cannot all be ove rluminous relative to the average of more distant SNe. If they are und erluminous, which must be the case by the statistics of the Malmquist effect if the large dispersion in M(max) for SNe Ia claimed by Hamuy e t al. applies to the calibrators, then the value of H-0 = 52 km s(-1) Mpc(-1) given by Saha et al. is an upper limit to the Hubble constant.