(ProQuest: ... denotes formulae omitted.)1. BackgroundHuman mortality has undergone remarkable declines over the years. The increase in life expectancy is probably the best expression for the dramatic mortality decline in the last 170 years (Oeppen and Vaupel 2002). Improvements in living conditions, nutrition and medicine are among the main reasons for this development (Riley 2001; Oeppen and Vaupel 2002). These changes in economic, social, and sanitary conditions first triggered an important decline in infant, child, and early adult mortality, which contributed to the reduction in lifespan disparities (Wilmoth and Horiuchi 1999; Edwards and Tuljapurkar 2005; Vaupel, Zhang, and van Raalte 2011). As individuals became more homogeneous in their ages at death, a of the distribution of deaths in a more narrow age-interval was observed in many low-mortality countries in the first half of the twentieth century (Fries 1980; Wilmoth and Horiuchi 1999; Kannisto 2000, 2001; Cheung et al. 2009). Fries (1980) hypothesized that this dynamic can be interpreted as a of deaths against the upper limit of the human lifespan. Assuming a nearly negligible role for premature mortality, he stated the limit of the average age at death as approxi-mately 85 years, with 95% of all deaths occurring in an age range of 4 years deviation (Fries 1980). The compression of mortality hypothesis motivated a rich discussion on the occurrence and interpretation of this development. Several studies provided evidence for a compression, but emphasized that the achieved mortality levels differ substantially from Fries' predictions (Nusselder and Mackenbach 1996; Wilmoth and Horiuchi 1999; Cheung et al. 2005).After the period of strong compression, low-mortality countries entered a new era of change. Since the second half of the twentieth century, the main contributions to the increase in average age at death shifted from infant and early adult ages to old and very old-ages (Christensen et al. 2009). This generated changes in the mechanisms behind the increase in life expectancy (Wilmoth and Horiuchi 1999; Edwards and Tuljapurkar 2005; Smits and Monden 2009). The new mechanism behind improvement in life expectancy is best illustrated by a shift in the distribution of death toward older ages with a shape remaining nearly constant (Yashin et al. 2001; Bongaarts 2005; Cheung et al. 2005; Cheung and Robine 2007; Canudas-Romo 2008). Vaupel (1986), Vaupel and Gowan (1986) and Bongaarts (2005) were among the first to articulate the idea of shifting mortality. Canudas-Romo (2008) deepens this idea by studying the variability around and the change of the modal age at death. He finds that over time mortality shifts to higher ages, with approximately constant variability in age at death. He concludes that the shifting mortality pattern might be the new dynamic behind mortality improvements, subsequent to the process.The ages at which mortality reductions occur tend to determine the dominating mortality dynamic: or shift. Compression is more pronounced when mortality reductions occur at very young and adult ages (Nusselder and Mackenbach 1996; Wilmoth and Horiuchi 1999; Kannisto 2000; Cheung et al. 2005). On the other hand, shifting mortality requires changes at old and very old-ages (Canudas-Romo 2008). Vaupel, Zhang, and van Raalte (2011) report relatively stable variability patterns for survivors beyond age 50 in the last 100 years. Engelman, Caswell, and Agree (2014) and Engelman, Canudas-Romo, and Agree (2010), however, provide evidence for a modest expansion of lifespan variability for survivors at older ages, resulting from mortality improvement at these same ages.The measurement of and shifting mortality is an important issue, as both dynamics translate differently into survival, mortality density and hazard distributions (Wilmoth and Horiuchi 1999). Alterations are, however, visible in all three functions due to their interrelation. …