Have you seen this table
or the chart it's derived from it?
At several points in life and on the web I've encountered them, cited as evidence that the rate of age-based infertility in the US has risen over the past 100 years. Sometimes an attribution is given, to an legitimate article by researcher George Maroulis, and in fact that is the chart's source. But context means a lot, and this presentation leaves out key information. The table, which also cites Maroulis, seems to draw on the chart, but does so very selectively and inaccurately (among other faults, it conflates and misstates the pre-20th century lines [there is no 18th or 19th-century data in the chart], omits the Iranian and Hutterite lines and misrepresents "modern USA" in several ways).
In the essay this chart is taken from, it's clear that what's being compared here are pre-birth-control ("natural") populations and data from the US in which birth control of various kinds is employed--non-hormonal (1955 and 1981) as well as hormonal (1981). It's the effect of the use of birth control that this chart documents--not a decline in ability to bear. The latest data in the chart comes from very early in the new later motherhood trend (1981), so the representations of births to women in their forties in the chart tell us little about dynamics today.
As I note in Ready, while the table includes a footnoted comment that the "older data is likely to include substantial inaccuracies," no note appears on the huge inaccuracy created by the presentation of the modern data.
This kind of erroneously presented material amounts to a form of statistical fakery, and contributes to the high level of contemporary fertility anxiety.
Here's Maroulis's discussion of the charts:
"Data from the United States do not reflect natural fertility rates since the populations are, as mentioned earlier, practicing birth control. However, they are of interest since they are the product of a combination of biological and social inferences. . . . the fertility rates observed in natural cycles of historical populations may have a bias in that older women who have already conceived previously may not be as anxious to get pregnant. So it may be more appropriate to review data from populations in whom women purposely delay childbearing and try to get pregnant at older ages. . . . A considerable amount of data from such natural populations that delay childbearing and did not practice contraception existed from the late 1700s and early 1800s in Belgium, England, France, Germany and Scandinavia, where the mean age of marriage was over 27 and even close to 30. . . . Results [show] that women over 40 years are not, as often portrayed, hopelessly infertile but indeed can, in up to 48% of cases, achieve a pregnancy." (George B. Maroulis, "Effect of Aging on Fertility and Pregnancy," Seminars in Reproductive Endocrinology, 9, no. 3 [August 1991]: 168.)
Note that the "over 40" category he invokes at the end of the quote is not defined in terms of upper age limit, and so could involve a number of women over 43, who skew the information on women 40-42, who current data indicates are substantially more likely to be fertile than women 43 and over.