The two foci of MM analysis

Analyses of historical MM have focused on two issues: (1) estimating the level, and (2) searching for correlates- Since most determinations of MM are based on elapsed time since birth, the accuracy and completeness of baptismal registration are essential, as noted- Some crosschecks are available from other data sources- In the reconstitution data, there were 417 maternal deaths within sixty days of 56,546 parturitions (see above)- In 350 burial records, there are notations suggesting childbirth issues-²⁰ Of these only 311 are the deaths of adults; neonates also die in childbirth-²¹ The overall rate of 0-0074 maternal deaths per live birth approaches the high end of the range calculated by Schofield (1986: 238) for Sweden 1756—1860 (maximum 0-011), is higher than any of the analogous rates from thirteen English parishes except in the period 1650—99 (maximum 0-041), approaches the peak German rates given by Imhof (1986: 125) for 1810—99 (maximum 0-010) and the German rates reported by Knodel 1825—99 (0-010—0-012) (1986)- The Slavonian peak rates, as noted, are even higher-

In the census of 1910 data given for all of Croatia 1906—10 show MM per live birth closely fluctuating around 0-005 (Croatia 1905: table III, 56)- The census distinguishes babinje (Wochenbett, childbed) at 0-004, bolesti rodila (Geburtskrankheiten, diseases of birthing) at 0-0007, and babinja grognica (Kindbettfieber, childbed fever) at 0-0002- The first category accounts for 81% of the sum of these, the second for 14%, and the last for 5%- These categories scale well to our expectations of the timing of MM of different kinds: childbed deaths

from direct obstetrical causes occurring rapidly, indirect deaths more slowly, and deaths from infection perhaps slowest.

5.3