My recent scanning of thousands of cemetery records at the Mesa City Cemetery graphically showed me the high incidence of the death of infants and young children. As genealogists, records of infant deaths are some of the most difficult to find and document. I suggest that cemetery records are mostly unknown and very little examined, but contain records of the death of infants that may appear in no other place.
Day after day as I scanned burial records, and I was devastated to find that a huge percentage of the early burials in the cemetery were infants or children. The cause of death was listed in many of earlier records and I saw huge numbers of deaths from scarlet fever, typhoid, intestinal disturbances, whopping cough, accidents, pneumonia, and influenza. Remember, antibiotics and vaccinations are a modern developments and these people had no access to proper medical care. An ear infection, common among children today and cured with a few doses of antibiotic, could be fatal before 1930 or so. Antibiotics were discovered in antiquity, but only became medically available after the discovery of penicillin in about 1928 and only generally available after World War II. It wasn't until well into the 1900s that the germ theory of the origin of disease was commonly accepted and understood.
Especially in smaller towns and agricultural based communities, the norm at any time before about 1920, was to have larger families. Any time there is a gap in the ages of the known children of more than two years, you can suspect that a child was born and died. In the 1800s and before, in Europe and where Europeans settled, it was not uncommon to give the name of a subsequent child of the same gender, the name of a deceased child and it is not a mistake when you see two children or even more recorded in a family with the same name but with different birth dates.
From my personal observation of the burial records, I can only assume that many of the infants that died and were buried in the cemetery, likely were not recorded in any other fashion. I found it highly unlikely that death certificates were issued for these children because the deaths were not otherwise reported and the issuance of death certificates were reserved to "official" deaths when a doctor was in attendance or later called. For that reason, I suspect that historical real infant mortality rates are vastly understated. Current infant mortality rates in underdeveloped countries exceed 10% but I suspect that those figures are understated. Finding historical rates is very difficult, but I assume that for many areas there are no figures.
The primary source for records of early infant deaths must come from the personal records of the family. Although losing an infant was more historically more common, it was in no way less traumatic or emotionally difficult that it is today. Family records such as Bibles and other records may show children that did not reach maturity. One source, as I have mentioned, is cemetery records. By examining only some of the cemetery records and ignoring others as repetitious, you may be failing to find the lost children. For example, in the Mesa Cemetery, many of the infants were buried with no grave marker at all. So looking at headstones only will fail to show many burials. Sometimes, as was the case with the Mesa City Cemetery, there are records that are normally not consulted by the cemetery workers that show more information than is in the readily accessible records. These include Permits for Burial and Funeral Records.
If you find a gap in the ages of children in a family or one where a marriage occurs and no children are born for more than two years, you should always suspect that a child was born and died and your investigation should begin.