We measure spectral indices for 1823 galaxies in the CNOC1 sample of fifteen X–ray luminous clusters at 0.18 < z < 0.55 , to investigate the mechanisms responsible for differential evolution between the galaxy cluster and field environments .
The radial trends of D4000 , W _ { \circ } ( H \delta ) and W _ { \circ } ( OII ) are all consistent with an age sequence , in the sense that the last episode of star formation occurred more recently in galaxies farthest from the cluster center .
Throughout the cluster environment , galaxies show evidence for older stellar populations than field galaxies ; they have weaker W _ { \circ } ( OII ) and W _ { \circ } ( H \delta ) lines , and stronger D4000 indices .

From our primary sample of 1413 galaxies , statistically corrected for incompleteness and selection effects , we identify a sample of K+A galaxies , which have strong H \delta absorption lines ( W _ { \circ } ( H \delta ) > 5 Å ) but no [ O ii ] emission ( W _ { \circ } ( OII ) < 5 Å ) , perhaps indicative of recently terminated star formation .
The observed fraction of 4.4 \pm 0.7 % in the cluster sample is an overestimate due to a systematic effect which results from the large uncertainties on individual spectral index measurements .
Corrected for this bias , we estimate that K+A galaxies make up only 2.1 \pm 0.7 % of the cluster sample , and 0.1 \pm 0.7 % of the field .
From the subsample of galaxies more luminous than M _ { r } = -18.8 + 5 \log { h } , which is statistically representative of a complete sample to this limit , the corrected fraction of K+A galaxies is 1.5 \pm 0.8 % in the cluster , and 1.2 \pm 0.8 % in the field .
Compared with the z \approx 0.1 fraction of 0.30 % ( ( ( Zabludoff et al .
1996 ) ) ) , the fraction of K+A galaxies in the CNOC1 field sample is greater by perhaps a factor of four , but with only 1 \sigma significance ; no further evolution of this fraction is detectable over our redshift range .

We compare our data with the results of PEGASE and GISSEL96 spectrophotometric models and conclude , from the relative fractions of red and blue galaxies with no [ OII ] \lambda 3727 emission and strong H \delta absorption , that up to 1.9 \pm 0.8 % of the cluster population may have had its star formation recently truncated without a starburst .
However , this is still not significantly greater than the fraction of such galaxies in the field , 3.1 \pm 1.0 % .
Furthermore , we do not detect an excess of cluster galaxies that have unambiguously undergone a starburst within the last 1 Gyr .
In fact , at 6.3 \pm 2.1 % , the A+em galaxies that Poggianti et al .
( 1999 ) have recently suggested are dusty starbursts are twice as common in the field as in the cluster environment .

Our results imply that these cluster environments are not responsible for inducing starbursts ; thus , the increase in cluster blue galaxy fraction with redshift may not be a strictly cluster–specific phenomenon .
We suggest that the truncation of star formation in clusters may largely be a gradual process , perhaps due to the exhaustion of gas in the galactic disk over fairly long timescales ; in this case differential evolution may result because field galaxies can refuel their disk with gas from an extended halo , thus regenerating star formation , while cluster galaxies may not have such a halo and so continue to evolve passively .