Predicted and Observed Evolution in the Mean Properties of Type Ia Supernovae with Redshift

Recent studies indicate that Type Ia supernovae (SNe Ia) consist of twogroups - a "prompt" component whose rates are proportional to the host galaxystar formation rate, whose members have broader lightcurves and areintrinsically more luminous, and a "delayed" component whose members takeseveral Gyr to explode, have narrower lightcurves, and are intrinsicallyfainter. As cosmic star formation density increases with redshift, the promptcomponent should begin to dominate. We use a two-component model to predictthat the average lightcurve width should increase by 6% from z=0-1.5. Usingdata from various searches we find an 8.1% +/- 2.7% increase in averagelightcurve width for non-subluminous SNe Ia from z=0.03 - 1.12, correspondingto an increase in the average intrinsic luminosity of 12%. To test whetherthere is any bias after supernovae are corrected for lightcurve shape we usepublished data to mimic the effect of population evolution and find nosignificant difference in the measured dark energy equation of state parameter,w. However, future measurements of time-variable w will require standardizationof SN Ia magnitudes to 2% up to z=1.7, and it is not yet possible to assesswhether lightcurve shape correction works at this level of precision. Anotherconcern at z=1.5 is the expected order of magnitude increase in the number ofSNe Ia that cannot be calibrated by current methods.Comment: 5 pages, 3 figures, accepted to ApJ Letters, addressed referee's comments, table adde