We have attempted first, to show that raw, mass statistics on erythroblastosis, however valuable they may be in some respects, can lead to serious error. Results will vary with the mode of collection, and with the care with which every family included in the statistics is studied. The inclusion of mild, often subclinical, cases will make the prognosis appear better. The inclusion of families in which unrecorded abortions or transfusions have taken place, will make the prognosis appear worse. Abortion and transfusion are most active agents in producing disease in later born children.
We have further set out our observations on certain blood group relationships and their possible effect on the development of Rh-erythroblastosis. These observations we do not regard as more than indications for further study.
Finally we have given, so far as we have them, the pregnancy summaries of 13 families in which anomalous results occurred, namely 9 in which normal children succeeded diseased children; 4 in which normal children were born to mothers with anti-Rh antibodies in their blood, of whom 2 had received Rh-incompatible transfusions, one had antibodies at the sixth week of a first pregnancy, and was therefore probably presensitized, and 1 was sensitized by multiple pregnancies only. While it is possible that the blood groups may be a factor in the last 4, they do not seem to play a part in the first 9. To us, this is a group of great importance, suggesting as it does, that other factors than blood incompatibility play an important, at times a deciding, role in the development or non-development of erythroblastosis.