So far, I have shown carbon monoxide data with older cars (on average) higher emitters than new cars. This result is observed everywhere and is independent of speed/load. The same age effects are observed for HC ("On-Road Carbon Monoxide and Hydrocarbon Remote Sensing in the Chicago Area in 1992," (ILENR/RE-AQ-91/15 and "On-Road Hydrocarbon Remote Sensing in the Denver Area," Zhang et al., Env. Sci. Tech, vol. 27, 1885-91, 1993), and NO ("Enhancement of Remote Sensing for Mobile Source Nitric Oxide," Zhang, et al, J.Air, Waste Mgmt. Assoc, 1995), but in both cases, the averages depend on vehicle speed and load.

Figure 4. In this graph, "91" symbols show 1991 HC versus CO data from several sites in Mexico City, and "94" symbols show data mostly from the same sites in 1994. The letter D shows the current Denver average.

In all cases, the average readings tend to obscure as much as they illuminate. When we use our large on-road data bases to divide up each model year into five groups (quintiles) from lowest to highest emitting, a more startling result appears. The observed effect of increasing average age on increasing average emissions is overshadowed by the dramatic differences between