SOME ASPECTS OF ABORTION. 341 



of these small masses, one feels quite certain that some of them belong in the 

 category of rare, full-term specimens such as that reported by Siemens. Few of 

 these probably survive till term on account of an insufficient blood-supply, and 

 the occasional survival of one in association with a normal, full-term fetus is an 

 extremely interesting occurrence. 



One need only to regard the question of abortion historically and recall 

 the practices of primitive races to be reminded of the fact that in such a matter as 

 this human custom, conduct, and frailty play a very large role. Nor need attention 

 be directed to the practices of primitive or uncivilized peoples alone, even if Rob- 

 inson's (1919) estimate that 1,000,000 criminal abortions are performed annually 

 in the United States can not be accepted without question. Since the number of 

 annual births in the United States, estimated on the basis of the registration area, 

 is only about 2,400,000, this would imply that 1 criminal abortion is performed 

 for every 2.4 births. Or, to put it another way, if somewhat less than 40 per cent 

 of all pregnancies terminate prematurely, as Pearson (1897) estimated, then, 

 according to Robinson, 1 out of every 4 pregnancies is terminated criminally. 

 But the highest estimates of the ratio of abortions to pregnancies are those of 

 Taussig (1910) and myself. Those of Taussig were based upon the experience 

 at a St. Louis gynecological clinic, and mine upon about 700 selected histories 

 from the Carnegie Embryological Collection. From these data it seems that 

 there is 1 abortion to about 1.7 to 2.3 pregnancies. Hence, if we accept Rob- 

 inson's estimate of the incidence of criminal abortion in the United States as 

 1 to every 2.4 births, it would follow that in the women considered by Taussig or 

 myself about 70 per cent of all pregnancies which terminated prematurely were 

 terminated criminally! Furthermore, upon the basis of Pearson's estimate of 

 prenatal mortality, the criminally induced actually would exceed the abortions 

 due to all other causes by over one-half, while upon the basis of Mall's earlier 

 estimate of a prenatal mortality of about 20 per cent, the criminal abortions in 

 the United States, as estimated by Robinson, actually would exceed the grand 

 total of abortions from all causes by 200,000 cases annually! 



It is regrettable that we are left partly to surmise regarding the exact incidence 

 of prenatal death. Ahlfeld (1898) estimated that there is 1 abortion for every 4 

 or 5 normal births. This would be 1 to every 5 or 6 pregnancies. A. Hegar (1863) 

 estimated that 1 out of 8 or 10 pregnancies ends prematurely. Michailoff (1897), 

 as reported by Chazan (1904), gave a frequency of 10.18 per cent, and Keyssner 

 (1895) a frequency of 15.1 per cent, or 1 abortion to every 5 or 6 births, an 

 estimate confirmed also by Williams (1917). According to Lechler (1883) and 

 Chazan (1904), this was also the figure reached by Busch and Moser (1840), upon 

 theoretical grounds alone. 



Stumpf (1892) found 1 abortion for every 3.56 pregnancies, a mortality of 

 28 per cent, and Keyssner still less, or 1 abortion in 9 pregnancies, or a mortality 

 of but 11 per cent. 



Since the causes responsible for postnatal mortality differ so widely from those 

 which operate before birth, it is wholly unlikely that a curve of postnatal mortality, 



