The gestational age pattern of human mortality

In order to check hypotheses about the cause for "ontogenescense" -- the phenomenon of a declining force of mortality prior to maturity -- I analyse data on human mortality by gestational age. Based on extensive microdata on births, fetal- and infant deaths in the US 2009 I calculate a joint fetal-infant lifetable by gestational age spanning week 23 until week 100 after the last menstrual period of the mother. This joint lifetable shows a remarkable regularity in the gestational age profile of fetal- and infant mortality: Mortality rates are declining over the whole observed age range with the exception of a "birth hump" peaking week 38. The absolute rate of decline slows down over age. The observed gestational age pattern of the force of mortality is consistent with three hypotheses concerning the causes for ontogenescense: 1) Adaptation: as the organism growths it becomes more resilient towards death, 2) transitional timing: the transition of birth is a stressful event and momentarily increases the force of mortality, 3) mortality selection: The frailest die first, resulting in the mean force of mortality to decline with age. In order to quantify the relative importance of these three processes I fit a three component mortality model against the observed force of mortality. The model describes the data with high accuracy, suggesting that the phenomenon of ontogenescense in humans is fully explained by the three hypotheses.