Sometime in the next few weeks we will all fill out our federal census forms (skimpy as they are this year). The United States has the longest tradition of government censuses, made necessary by our method of determining representation in the House of Representatives. While we have had a federal census since 1790, they have changed greatly over the years. Until the mid-20th century a census taker with a big spread sheet showed up at your door to record all inhabitants at the residence. From 1850 to 1880, the census taker asked if anyone died at the residence in the last year and recorded any on a second spreadsheet called the Federal Mortality Census. If you think they asked a lot of questions about the living, it was less than about the dead.
In theory it should be a snapshot of mortality in every county of the United States for the previous year. Keep in mind that the United States was a growing nation in the 19th century, so not all modern states are represented in the mortality censuses and county boundaries have changed in many states. They asked interesting data. The mortality census covers name, age, color (white, black, mulatto), marital status, state of birth, month of death, occupation, cause of death and length of illness. These schedules are the forerunners of death certificates. The Federal Mortality Schedules can be searched here and local mortality schedules may be available at local libraries. If your local library does not have them, then look for the library that hosts the county genealogical society. Better genealogy society libraries will have the data for most or all of the state.
For epidemiological purposes these schedules have limited usefulness because of under-reporting and misdiagnosis. Cause of death is usually reported by the family, though doctors in the earlier years often didn’t do much better. Many early doctors in the first half of the 19th century had little or no official medical training and/or used a limited set of diagnoses. Nineteenth century medicine had a series of, well, just plain odd thoughts. One of my personal favorites was the notion that children could not contract malaria! Not only can they contract malaria, but they have a higher mortality rate than adults. They also listed causes of death like “teething” in children. Some diagnoses like cholera morbus no longer exist. Cholera morbus was a collection of gastrointestinal diseases that caused diarrhea. Most cases of cholera morbus were probably related to food contamination (food poisoning).
Under-reporting is apparent if you just look at the schedules. I looked at the Illinois schedules and it is obvious that not all census takers were equally vigorous in collecting mortality data. Even in areas where the census taker asked, he was reliant upon surviving family members or boarding house keepers for data. Single people could easily be missed because there was no one left to report. The same could be said for travelers who died in a district (and in some trading or market areas there were a lot transients). When one or both parents died, families were often broken up and scattered. This could lead to some deaths being overlooked. During the worst epidemics, sometimes entire families died. With no one left to report on them, they were overlooked. In 1850 Illinois census takers complain that they know they are missing many of the dead from the cholera epidemic that year because entire families have died with no one left to report. Even where there are people left to report, it seems likely there is still under-reporting. The number of children reported does not match physician’s estimates of childhood mortality. There are very few infants reported when we know that without premature infant care, C-sections, or antibiotics, infant mortality must have been relatively high.
So with all these problems, is there anything that can be taken from the federal mortality census? Yes, there is within reason. It is naturally useful to genealogists interested in particular family histories. With the life expectancies of the 19th century and incomplete recording, it is not uncommon for people to be listed on only one census, dead or alive. They are also of some use for the study of public health. The federal mortality censuses can show seasonality and a rough estimate of mortality. Trends can be observed and epidemics identified. Diagnoses can tell us a lot about paradigms of 19th century disease and life-stage.