Detecting trends in short data sets of stratospheric molecules is difficult because of variability due to dynamical fluctuations. We suggest that one way around this difficulty is using the measurements of one molecule to remove dynamical variability from the measurements of another molecule. We illustrate this using Aura MLS measurements of N₂O to help us sort out issues in the determination of trends in HCl. This shows that HCl is decreasing throughout the middle stratosphere as expected.