We have measured the clustering of 0.30 < z < 0.90 red galaxies and constrained models of the evolution of large-scale structure using the initial 1.2 \Box ^ { \circ } data release of the NOAO Deep Wide-Field Survey ( NDWFS ) .
The area and B _ { W } RI passbands of the NDWFS allow samples of \gtrsim 10 ^ { 3 } galaxies to be selected as a function of spectral type , absolute magnitude , and photometric redshift .
Spectral synthesis models can be used to predict the colors and luminosities of a galaxy population as a function of redshift .
We have used PEGASE2 models , with exponentially declining star formation rates , to estimate the observed colors and luminosity evolution of galaxies and to connect , as an evolutionary sequence , related populations of galaxies at different redshifts .
A red galaxy sample , with present-day rest-frame Vega colors of B _ { W } - R > 1.44 , was chosen to allow comparisons with the 2dF Galaxy Redshift Survey and Sloan Digital Sky Survey .
We find the spatial clustering of red galaxies to be a strong function of luminosity , with r _ { 0 } increasing from 4.4 \pm 0.4 h ^ { -1 } { Mpc } at M _ { R } -5 { log } h \approx - 20.0 to 11.2 \pm 1.0 h ^ { -1 } { Mpc } at M _ { R } -5 { log } h \approx - 22.0 .
Clustering evolution measurements using samples where the rest-frame selection criteria vary with redshift , including all deep single-band magnitude limited samples , are biased due to the correlation of clustering with rest-frame color and luminosity .
The clustering of -21.5 < M _ { R } -5 { log } h < -20.5 , B _ { W } - R > 1.44 galaxies exhibits no significant evolution over the redshift range observed with r _ { 0 } = 6.3 \pm 0.5 h ^ { -1 } { Mpc } in comoving coordinates .
This is consistent with recent \Lambda CDM models where the bias of L ^ { * } galaxies undergoes rapid evolution and r _ { 0 } evolves very slowly at z < 2 .