21 centimeter line width as an extragalactic distance indicator. II. Does the Tully-Fisher relation depend on Hubble type

The proposal that the slope b of the Tully-Fisher (T-F) relation, -M/sub T/0 = a+b(log V/sub M/-2.2), is the same (b = 10) in the blue (B) band and in the infrared (H) band and that its zero point a depends on morphological type T is tested by several methods. The evidence from small samples at T = const. or in the Virgo cluster is not conclusive, but with the previously adopted value (b = 5) the zero point a is constant and independent of T and of the luminosity index Λ/sub c/ within statistical errors (m.e.roughly-equal0.1 mag) for our sample of approx.150 ''best-observed'' galaxies in two velocity ranges (V0greater-than-or-less-than2000 km s-1). Solutions of the form -M/sub T/0 = a0+a1(T-4)+b1(log V/sub M/-2.2) for all types (-2< or =T< or =10) and for spirals only (2< or =T< or =7) give a small, nonsignificant coefficient a1 and a good fit (sigmaroughly-equal0.6 mag) when b1 = 5, but a large coefficient a1 (0.21--0.38) and a poor fit (sigmaroughly-equal0.9 mag) when b1 = 10. These tests confirm the validity of the T-F relation previously adopted with b = 5 for the B band magnitudes