The trends in age-adjusted death rates we expected because we've been seeing an increase in life expectancy for some years now.

Older folks who might otherwise have left us decided to hang on to life a little longer.

We were surprised. We were scratching our heads. Something of this magnitude is really out of the ordinary.

We were surprised by the sharpness of the decrease. It's kind of historical.

The most striking aspect of the data this year was the intensity or volume of the decreases. This is the largest single-year decrease in the raw numbers of deaths that we've seen since the 1940s.

We really don't know -- we can't pin it down.