We present the results for the galaxy luminosity function as estimated from the Durham/UKST Galaxy Redshift Survey .
This survey is magnitude limited to b _ { J } \sim 17 , contains \sim 2500 galaxies sampled at a rate of one in three and surveys a \sim 4 \times 10 ^ { 6 } ( h ^ { -1 } Mpc ) ^ { 3 } volume of space .
The maximum likelihood parameters for a standard Schechter luminosity function are estimated to be M ^ { * } _ { b _ { J } } = -19.72 \pm 0.09 , \alpha = -1.14 \pm 0.08 and \phi ^ { * } = ( 1.2 \pm 0.2 ) \times 10 ^ { -2 } ( h ^ { 3 } Mpc ^ { -3 } ) .
Attempting to correct for the scatter in the observed magnitudes leads to a flatter faint end slope , \alpha = -1.04 \pm 0.08 , which , combined with the different luminosity function shape , causes a higher normalisation to be estimated , \phi ^ { * } = ( 1.7 \pm 0.3 ) \times 10 ^ { -2 } ( h ^ { 3 } Mpc ^ { -3 } ) .
Neither of these parametric functions provides a good formal fit to the non-parametric estimate of the luminosity function .
A comparison with galaxy luminosity functions from other redshift surveys shows good agreement and the shape of the luminosity function now appears well-defined down to M _ { b _ { J } } \simeq - 17 .
There are some discrepancies between the different surveys for galaxies fainter than this absolute magnitude .
However , our estimate agrees well with that from the APM-Stromlo Galaxy Redshift Survey and we measure a fairly flat faint end slope .