Nineteenth-Century Death Records: How Dependable Are They?
By William B. Saxbe Jr., M.D., M.P.H., CG, FASG
The past is a foreign country; they do things differently there.1
Sarah Black, a forty-five-year-old "colored" widow, died 22 May 1880 in Champaign County, Ohio. While the discovery of this information may gratify any descendants who seek it, they are likely to be puzzled by poor Mrs. Black’s cause of death, as given by the county’s death registrations: “change of life.”2 No surviving obituary explains this fatal menopause, the only other cause of death recorded for Sarah is even stranger: “poisoned face.”3 This is an extreme example of the difficulties researchers may have with nineteenth-century death records, but less-dramatic challenges are frequent in dealing with such documents. A closer look at these records, and at why and how they were created, will sharpen our awareness of their virtues and defects.
Genealogists have always used death records. In the past, perhaps only the date and place of death were considered important. Modern family historians, however, want a fuller picture of their forebears’ lives; and family-health historians seek causes of death to identify hereditary patterns of disease.
Before the mid-nineteenth century, information about deaths usually came randomly from estate records, from gravestones and Bible records, from town and church registers, and from newspapers, With the first of these, the record was created to facilitate the transfer of property to the heirs; for the second, to memorialize a family member; and for the last two, to record the passage of a town’s citizen or a church’s member. Causes of death, when known, were irrelevant to these purposes.
By mid-century, however, the growing science of public health demonstrated that many diseases were preventable by clean drinking water, decent living conditions, healthy diets, proper sanitation, vaccination, and vermin control. For example, the transmission of communicable diseases could be interrupted by separating those infected from those susceptible. Since public-health priorities had to be guided by reliable data on the frequency and location of disease, the first step in the collection of such data was the death register, which listed the cause of death as well as its date and place. Initially kept at the city or county level in most jurisdictions, these registers generally became state responsibilities in the twentieth century. In Ohio, for example, the registration of births and deaths began at the county level in 1867 and at the state level in November 1908.4
Like censuses, the registers were not created to preserve information on individuals for the benefit of future genealogists; the intent was to analyze their data in the aggregate in order to devise public-health interventions. Despite their considerable value, these records present a number of challenges to both epidemiologists and genealogists, especially in the areas of completeness and accuracy. The following study was undertaken to gauge the extent of these problems in a typical Midwestern county, with particular attention to the recorded causes of death.
The study focused upon Champaign County, Ohio—in the rural, west-central part of the state—and covered the one-year period from 1 June 1879 to 31 May 1880.5 The county’s population that year was 27,701.6 A county-level death register exists; but from this source alone, it is impossible to tell how fully deaths in the county were properly reported and recorded. As external control, two other contemporary types of death records were used for comparison; obituaries in county newspapers and census mortality schedules.
The death register collected information from reports submitted by officials in each civil division (twelve townships; the villages of Mechanicsburg, St. Paris, and North Lewisburg; and the city of Urbana). The following information was gathered:
|Name in full||Place of birth|
|Date of death (year/mont/day)||Place of death|
|Marital status (married/single/widowed)||Place of residence|
|Age (years/months/day)||Color (white/colored)|
|Cause of death (direct/indirect)||Occupation|
|Names of parents (when an infant without name)||Informant|
For example: Thomas Stone, a seventy-four-year-old white farmer, died 19 November 1879 in Wayne Township, his place of residence. He had been born in Adams County, Virginia. His cause of death: “cold.” Benjamin McCauley, a white sixteen-year-old, died 20 May 1880 in Urbana, his native city, of “inflammation of the bowels.” Whether cause-of-death information originated from families or from physicians was not indicated in these records.
Census takers gathered this set of data by interviewing families door-to-door, seeking the following information on those who had died in the preceding twelve months:
|Name||Age (in years)|
|Month of death||Place of birth|
|Marital status (single/married/widowed/divorced)||Places of parents' births|
|Color (white/black/mulatto/Chinese/Indian)||Place of death|
|Length of residency in county||Disease or cause of death|
|Place where disease was contracted, if not in the same county||Name of attending physician|
As sample reports: Margaret Olden—a twenty-seven-year-old, single, white housekeeper—died in Adams Township of dropsy in August 1879. She and both of her parents had been born in Ohio, and she had lived in Champaign County for twenty-five years. There was no attending physician. George B. Dotson—a thirty-six-year old married white minister who had lived in Mechanicsburg for six years—died there in March 1880 of consumption, which he had contracted “in the Army.” He and his father had been born in Pennsylvania; his mother had been born in England. “Dr. Jones,” his attending physician, was said to have “gone away.”
The Census Bureau instructed the enumerators, via the back of each census sheet, to have the reported causes of death reviewed by the attending physicians. Those physicians, then were to enter their agreement with (or corrections to) the data and add their signatures.
The effort to gather mortality statistics produced varying degrees of success—partially because of the poorly recompensed burden it placed upon the enumerations. The Androscoggin County, Maine, census taker of 1870 vocalized the problem plainly:
I noticed that instructions in relation to this schedule are very imperative that asst. marshals heretofore have been indifferent on the subject considering the compensation for the service, as not proportionate to the time consumed ... I have been very careful not to let a death go unrecorded in my division. Total Nu[umber] 42 for which I get 84 cents for my pains.7
Conditions had not improved by 1880, as noted even more tersely by the Concordia Parish, Louisiana, enumerator, who grumbled: “This [three-page] list caused me to work harder [than] I did on the census”8 and by his colleague in Panola County Mississippi, who echoed, “It gave me more trouble to get the exact requirements for the death Schedule than any thing else in this[s] business.”9
Finally, this study examined obituaries that appeared in the county seat’s newspapers; the Urbana Union-Democrat, the Champaign Democrat, and the Urbana Citizen and Gazette for the same period.10 These ranged from very brief notes concerning deaths to more-formal obituaries containing considerable data about the decedent’s life.
As examples: a man named Epps who lived near Springhill, in Harrison Township, was killed in a buggy accident there in June 1879; the news item supplied no given name, no age, no specific date of death and not additional details.11 On the other hand, when Father Martin Walsh—the dynamic and popular twenty-seven-year-old Irish native who had pastored St. Mary’s Roman Catholic Church in Urbana—died suddenly on 31 March 1880 of “heart disease,” the local paper thoroughly covered his death, his funeral, and his grieving parish.12
Collation and comparison of the three sets of data—for completeness and consistency of data as well as for causes of death—revealed drawbacks in each record type. The problems of ambiguous diagnoses and obsolete medical terms, which existed for all the record types, are discussed later in this article. Situations more or less peculiar to each type of record are discussed below.
The legibility of county records (which report 208 deaths) was excellent, but 26 (12.5 percent) failed to specify a cause of death. Furthermore, the records are usually entered in groups, evidently from returns forwarded to the probate court whenever a sufficient number of entries had been gathered at the local level. Some were not registered until many months after the deaths, so that—in a few instances—there is reasonable doubt that the recorded year of death was correct. Comparisons with the mortality schedules and obituaries helped to exclude obvious errors of year.
Legibility was a major problem with the census mortality schedules—some of which were written in sepia, red or purple ink that had faded badly. The microfilmed records were difficult to read, but even the original manuscript schedules at the Ohio Historical Society were frequently a challenge. Of the 310 death records, 6 had unreadable names, 7 had unreadable ages, and 14 had unreadable statements of cause. Adding to the legibility problem, many of the causes of death were overwritten with additions and corrections. In general, this study used the first-written diagnoses, because they were more legible. As with population returns, the mortality schedules have all of the usual census problems: their accuracy depends on the diligence of the take, the reliability of the informant, and the stability of the community. An additional problem with mortality schedules is their reporting window of exactly one year; if an informant were wrong about a decedent’s month of death, a name might be inappropriately added to or omitted from the record.
None of the newspapers had a regular obituary column. Deaths were mentioned in varying entries scattered throughout the pages, often in news from correspondents in communities further out in the county. Of 128 citizens whose deaths earned a notice in papers, only 64 had ages given and causes of death appear for only 46.
|Deaths in Champaign County, Ohio|
|1 June 1879 to 31 May 1880|
|Type of||Deaths||Missing or||Causes|
|County death registrations||208||26||12.5|
|Census mortality schedules||310||14||4.5|
|* This number is not the sum of the column above, since most deaths appear on more than only one type of record|
† This is the number of deaths for which no legible cause was found in any record.
Completeness of Coverage
Omitting stillbirths, premature births, and neonatal death, 371 documented deaths occurred in the county in the one-year period. (see table 1.) The county-level death registrations captured 208 (56 percent) of these; the mortality census had 310 (84 percent).13 Only 46 percent of all deaths appeared on both these lists. Obituaries or death notices appeared in the newspapers for 35 percent of known cases—with more-complete coverage for those who died in Urbana, the county seat. Only 16 percent of the total appear in all three sources. A surprising 42 percent appear in only one: 29 percent in the census mortality schedules only, 7 percent in the county death records only, and 6 percent in obituaries only. Probably, a number of death escaped recording by any of the three methods. The records varied widely in stating a cause of death: 96 percent of the census mortality schedules, 88 percent of county death registrations, and only 36 percent of the obituaries named a specific cause. For 8 percent of deaths, there is no reported cause in any record.
|Identified Causes of Death|
|(with typical stated causes in brackets)|
|Infections (see table 3)||221||65.1|
|Old age & heart disease* [neuralgia of heart, dropsy]||35||10.3|
|Cerebrovascular disease [apoplexy, paralysis]||25||7.3|
|Malignancy [cancer, tumor]||15||4.4|
|Maternal deaths form childbirth||3||.8|
|* Although those who died of cerebrovascular disease tended to be elderly, a number of the deaths from heart disease were in people aged 20-40. Probably, these people had congenital or valvular heart disease (correctible by surgery in modern times), infections such as viral myocarditis, or bacterial endocarditis, or inflammatory diseases such as rheumatic fever.|
Agreement as to Cause
“Confirmed” diagnosis—those whose sources agree as to cause of death—exist in only 35 percent of the total number. Of the 153 cases in which two or more sources report an individual’s cause of death, the diagnosis generally agree 84 percent of the time. The most-frequent explanation, by far, is infection—representing 65 percent of the cases for which a diagnosis appears.14
Table 2 groups all causes of death, and table 3 breaks down the types of fatal infections. The year 1 June 1879-31 May 1880 saw no significant epidemics in Champaign County; had any occurred, they would have increased both the absolute and the relative contribution of infectious disease to the total number of deaths. The infections listed in table 3 were endemic—always present in the population. Over the next several decades their toll sharply declined, more from the enaction of public-health measures than from improvements in individual medical treatment.
|Probable modern equivalents are in brackets|
|Brain fever/brain congestion [meningitis, encephalitis]||23|
|Cholera infantum [diarrhea disease of infants]||15|
Comparison with Modern Causes of Death
Modern-day medical diagnoses and the assignment of mortality causes are based on three sources of information; the history of the disease, as related by the patient or family, the physical examination, as performed by the physician, and the laboratory and imaging studies, assisted by a burgeoning array of high-tech machines and specialists to operate them. The current medical dependency on technology is great and growing; but nineteenth-century physicians had very little of it. None of these stated causes of death were substantiated by X rays, blood tests, biopsies, or bacterial cultures. None of the obituaries suggest that any deaths were preceded by diagnostic or therapeutic surgery, and only 1 of the 371 deaths had a documented postmortem examination (autopsy).
All of theses stated causes of death were based on clinical findings; signs and symptoms only. For all their experience and expertise, nineteenth-century physicians frequently erred, because many diseases shared similar symptoms and physical findings with other conditions. (Of course, little harm is done by diagnostic error when no effective treatment exists.) The attached list of common causes of death in these records gives the probable modern-equivalent diagnoses—or the difficulties in assigning modern equivalents.
It is unwise, nonetheless, to equate the stated nineteenth-century causes with their most-likely present-day counterparts. From the current vantage point, physicians of that era were guessing at their diagnoses, and genealogists would be second-guessing to think they know what those patients had. Humility and the acceptance of uncertainty are required here; no researchers should assume those diagnoses were right or that one can now correctly interpret what was meant. For instance, it is unlikely anyone can now learn what was intended by such stated causes of death as asthenia, debility, disability, infirmity, or insufficiency—all judgments rendered in the records covered by this study. All of these imply profound weakness and probably were used as synonym for wasting—the common outcome of many malignancies or consumption—or for old age. “Head affection,” the stated cause of death of one citizen, seems destined to remain a permanent mystery.
In many cases, family-health historians may be disappointed by nineteenth-century death records. Very few of the fatalities in Champaign County in 1879-80 are likely to have been manifestations of inherited or genetic diseases. Medical science has demonstrated that most malignancies are sporadic rather than familial, that infections and trauma strike randomly, and that the degenerative diseases of old age eventually catch everyone who has escaped the scythe of an earlier death.
A Short Glossary of Ninetheeth-Century Diseases
Most medial dictionaries, particularly old ones, adequately explain obsolete medical terms. One work currently in print approaches the subject from a genealogical standpoint.15 The list below defines some terms from the Champaign County records that may be unfamiliar to modern researchers.
Ague. See malaria.
Apoplexy: Equivalent to stroke, or cerebrovascular accident. A common consequence of uncontrolled high blood pressure.
Bilious attack: Possibly gallbladder disease, although hepatitis could give similar symptoms.
Bone erysipelas: Chronic bone infection, or osteomyelitis.
Brain fever: Either meningitis (itself either bacterial or viral, and spread by the respiratory route,
or viral encephalitis (transmitted by mosquitoes). The cerebral manifestations of acute polio infection could look the same.
Cancrum oris: A mixed bacterial infection of the tissues around the mouth, similar to the modern “flesh-eating” bacteria.
Cerebral congestion. See Congestion of the brain.
Cholera infantum: Diarrheal disease of infants, caused by numerous vital, bacterial, and protozoal agents. It killed though dehydration. Like true cholera, it was transmitted by contaminated water.
Congestion of the brain (also cerebral congestion): Probably the same as brain fever.
Congestion of the liver: Possibly hepatitus (usually viral), but gallbladder disease, leptospirosis (bacterial), cirrhosis, alcoholic and other poisonings, and amebic infections could earn this name.
Consumption: Pulmonary tuberculosis, so called because it “consumed” the body with generalized wasting and a racking, productive cough. Some other lung conditions or malignancies might produce a similar clinical picture.
Convulsions: Most seizures in children result from fevers, epilepsy and meningitis also cause convulsions.
Disease of the liver. See congestion of the liver.
Dropsy: An older term for edema, or retained fluid in the tissues. This is generally the consequence of heart disease, either valvular or coronary. Some kidney and liver diseases could produce a similar appearance.
Dyspepsia: Literally, “bad digestion.” As a cause of death, it might represent ulcer disease. Cancer of the stomach or pancreas could not be excluded as possibilities without an autopsy. Even today, the pain of a heart attack is often misinterpreted as indigestion.
Erysipelas: Cellulitis, or infection of the skin and soft tissues, usually streptococcal. Potentially fatal before antibiotics. See also bone erysipelas.
Hemiplegia: Paralysis of one side of the body by a stroke.
Inflammation of the bowels: In a young person, this might be appendicitis. In an older person, it might be diverticulitis.
Inflammation of the heart: Probably rheumatic fever; possibly viral myocarditis, or bacterial endocarditis.
Lung fever: Pneumonia, usually bacterial. Typhoid fever and tuberculosis can cause pneumonias as well.
Liver complaint. See congestion of the liver.
Malaria: Also typhomalaria or “malarial poison.” The clinical picture of malaria is one of the recurrent fevers alternating with normal temperatures. Many infections besides real malaria (which is transmitted by mosquitoes) could give the same picture, typhoid being the most common. There was real malaria in Ohio in the nineteenth century, but it was usually called ague, and it was ordinarily nonfatal.
Neuralgia of the heart: Neuralgia means “nerve pain”—an elliptical description, because pain is felt only through nerves. Probably the condition meant here is angina pectoris, the pain felt when the heart muscle is starved for oxygen (because the coronary arteries are blocked by arteriosclerosis). The pain of an actual myocardial infarction or heart attack would be similar.
Neuralgia of the stomach: Pain in the upper abdomen, probably from ulcer disease or pancreatitis.
Palsy: Perhaps Parkinson’s disease. Possibly paralytic polio.
Paralysis: Inability to move, as a consequence of a stroke.
Paralysis of the bowels: Possibly intestinal obstruction, such as from a tumor, a strangulation hernia, or a twisting of the intestine.
Paralysis of the heart: Possibly heart block, where disease produces a slow and weak heartbeat.
Phthisis: A synonym for consumption. Pronounced “ti-sis.”
Pleuropneumonia: Pneumonia that has led to an empyema, or abscess in the space surrounding the lung.
Pocks/pox: Probably secondary syphilis, which produces a characteristic rash.
Puerperal fever: Also childbed fever. Infection of the lining of the uterus, a common cause of maternal death following childbirth.
Scrofula: Tuberculosis of the lymph nodes of the neck, seen mostly in children. In Latin, scrofula means “little sow” the multiple draining points on either side of the neck suggested the rows of teats on a sow’s belly.
Spasms: Presumably convulsions, although tetanus would be a likely alternative.
Spinal fever: Spinal meningitis, an infection of the fluid and tissues surrounding the spinal cord. Bacterial and tuberculous meningitis were usually fatal.
Summer complaint: The same as cholera infantum; the disease was most common in the summer.
Typhoid fever: An infection of the intestinal wall, which could produce perforation and peritonitis. The responsible bacterium, usually transmitted by contaminated water, could also cause pneumonia and meningitis.
Typhomalaria. See malaria.
Worms: The death of one five-year-old child in this study was blamed on worms. Perhaps they caused intestinal obstruction. Worms will also migrate through a hole in the intestine caused by another disease, such as typhoid fever.
Genealogists are typically interested in specific records—not average or cumulative ones. Yet knowledge about record survival and reliability is useful to guide research and to keep expectations reasonable. This one-year sample in nineteenth-century death records in one Ohio county reveals that no more than 35 percent of the known deaths produced obituaries, only 56 percent appeared in the county death registrations, and only 84 percent were picked up by that year’s mortality census. These figures will give genealogists some idea of the likelihood of locating a death records in this period, and where one might be found.
When it comes to the completeness and accuracy of documented causes of death, diagnoses were frequently omitted from these records and some stated causes are difficult to interpret in modern terms. By any measure, however, the causes of death 120 years ago differ markedly from those of today. In this county, tow-thirds were the result of infections, with consumption still—as John Bunyan styled it in 1680—“the captain of all these men of death.”16 Consumption and other respiratory diseases made up more than half of the deaths from infections, and such waterborne intestinal diseases are typhoid fever and cholera infantum were responsible for another quarter. Relatively few deaths could be attributed to potentially hereditary conditions.
William B. Saxbe Jr., M.D., M.P.H., CG, FASG; 346 Reamer Place; Oberlin, OH 44074-1408. Dr. Saxbe has degrees in both medicine and public health, and is board-certified in surgery as well as genealogy. He is an Affiliate Scholar of Berlin College and a member of the NGS Family Health Historians Committee. ProGenealogists® thanks Dr. Saxbe who has graciously given his permission for the online reprinting of this article, which first appeared in the National Genealogical Society Quarterly in March 1999 (vol. 87, pages 43-54).
3. 1880 U.S. Census (Mortality Schedule), Urbana Township, Champaign County, Ohio, p. 1. The original schedules are preserved in hard copy at the Ohio Historical Society, Columbus. They are also available as National Archives microfilm T-1159, roll 102. The “poisoned face” is explained on the reverse of the mortality schedule for Urbana Township: Sara was bleeding from the socket of an extracted tooth, and the remedy applied to stop the bleeding proved fatal. [top]
4. An abortive attempt to record deaths at the county level in Ohio in the 1850’s was abandoned because of local resistance and noncompliance. Few of those records survive, See Kip Sperry, Genealogical Research in Ohio (Baltimore: Genealogical Publishing Co., 1997), 41. [top]
6. 1880 U.S. Federal Census (Population Schedule), Champaign County, Ohio; National Archives microfilm T-9, roll 998. The county’s demographics (the age and sex distribution of its people) were unquestionable skewed by the death of several hundred of its young men in the Civil War, but the same was true of the rest of the country. [top]
7. 1870 census, Androscoggin County, Maine, mortality schedule, city of Auburn (enumerator’s comments), FHL microfilm 0,009,741; original records in the Division of Vital Statistics, Augusta, Maine. The quotations in nn. 7-9 and similar comments from mortality schedules are reported in Elizabeth Shown Mill’s “Walk-abouts and Chicken Men: Tales of the U. S. Federal Census Takers,” luncheon address, 1993 National Genealogical Society Conference, Baltimore; available on audiocassette as Baltimore F-121 (Hobart, Indiana: Repeat Performance, 1994). [top]
10. The Ohio Historical Society has these newspapers but does not have a complete run of any of the three. Issues used were the Urbana Union-Democrat (on microfilm), for June-December 1879; its successor, the Champaign Democrat (on microfilm), for January-March 1880; and the Urbana Citizen and Gazette (in bound volumes), for April-May 1880, Newspapers were published in Mechanicsburg and St. Paris in this period, but no complete runs are known to survive. [top]
12. Champaign Democrat, 1 April 1880, p. 3; and 8 April 1880, p. 3. The obituaries consistently identify the priest as Father M. W. Walsh, the mortality census cites him as Michael Walsh, but a newspaper article praising his preaching, two weeks before his death, call him Father Martin Walsh; see Champaign Democrat, 18 March 1880, p. 2. He does not appear in the county death registrations. According to Sadliers’ Catholic Directory ... 1879 (New York: D. and J. Sadlier, 1879), 81, 495 he was Rev. Martin Walsh. [top]
13. By comparison, the mortality schedules for 1850, 1860, and 1870 are said to be only 60 percent complete. See Val D. Greenwood, The Research’s Guide to American Genealogy, 2d ed. (Baltimore: Genealogical Publishing Co., 1990), 229. Greenwood attributes these estimates to the U. S. National Office of Vital Statistics, Vital Statistics of United States, 1950 (Washington: U.S. Public Health Service, 1954), vol 1: 7. [top]
14. Another study, this one of Texas deaths between 1 June 1849 and 31 May 1850, attributes only 43 percent of mortalities for which a cause is reported to infection. See James Byars Carter, M.D., “Disease and Death in the Nineteenth Century: A Genealogical Perspective,” National Genealogical Society Quarterly 76 (December 1988): 290. [top]
15. Jeanette L. Jerget, A Medical Miscellany for Genealogists (Bowie, Maryland: Heritage books, 1995). Carter, “Disease and Death in the Nineteenth Century,” 294-300, also offers a short glossary that includes some of the terms given above (with varying perspectives), omits some others, and adds additional terms. [top]