Does the Flynn Effect Drop to Zero at a Very High IQ?

The idea is that the IQ of the average person has risen, but the IQ's of our topmost minds haven't.         Their IQ's are no higher than those of the greatest minds of the past. Since the Flynn Effect has been observed only at the mean IQ of 100, perhaps it doesn't apply at higher IQ's.  This is an attractive Idea, since it would explain why there could be such sup
Unfortunately, it leads to logical absurdities.
All that is required to measure the Flynn Effect is to administer IQ tests from the 1916-1920 era to a set of subjects whose ratio IQ's have also been measured on a current IQ test. We would expect that a set of today's 6-year-olds with IQ's in the neighborhood of 100 and mental ages in the neighborhood of 6 would score about 133 on the 1916-1920 test(s), corresponding to a 1916 Stanford Binet mental age of 8.

Proposition 1. The Range of IQ's Above 100 Over Which the Flynn Effect Is Effective Must Be At Least As Great As the Magnitude of the Flynn Effect
It is evident that the Flynn Effect must obtain at least over a ratio-IQ range equal to the magnitude of the Flynn Effect. Otherwise, if the Flynn Effect is 33 points of IQ, as measured on the 1916 Stanford Binet Test, and the Flynn Effect went to zero at a present-day IQ of 130, then anyone with a 1916 S-B IQ between 130 and 133 would have the same 103-to-133 IQ in 2002. And that would mean that someone with an IQ of 100 today could be expected to earn a 133 IQ score on the 1916 S-B, while someone who earned a Mensa-entry-level score of 132 on an IQ test today would also score a 132 on the 1916 S-B. In other words, they would score one point below someone who had a present-day IQ of 100! Obviously, that can't be.

Proposition 2: The Flynn Effect Can't Decline With Increasing IQ
1.  Assume That The Flynn Effect Decreases to Zero at IQ 133:
Let's start with the assumption that the maximum present-day IQ to which the Flynn Effect applies at all is 133, and that the Flynn Effect tapers off linearly as the present-day IQ goes from 100 to 133. In that case, everyone who had a present-day between 100 and 133 would earn the same IQ score of 133 on the 1916 Stanford Binet! In other words, the range of present-day IQ's between 100 and 133 would map into a single huge spike on the 1916 S-B. What we would see would be a present-day distribution that is log-normal, and a 1916 distribution that stretched from 100 to 133 for today's scorers in, perhaps, the 75-IQ to 100-IQ range), soared around 133, and then matched the rest of today's ratio-IQ curve for ratio IQ's above IQ = 133.

2.  Assume That The Flynn Effect Decreases to Zero at IQ 150:
Consider the case of an average 6-year-old child with an IQ of 100. That child would be expected to score 133 on the 1916 S-B, and to have a mental age of 8, as measured by that test.
Next, consider a gifted 6-year-old with a present-day ratio IQ of 150 and a present-day mental age of 9. Since we're assuming that the Flynn Effect hasn't acted upon individuals with IQ's at or above 150, that child will be expected to make a 150 on the 1916 S-B, also, and to have a mental age of 9 on that test, as well. But consider what that means! Our gifted 6-year-old with the 150 ratio IQ will make the same mental-age score of 9 on our present-day IQ tests as our average, present-day 9-year-olds. And inasmuch as we have assumed that there is no Flynn Effect for IQ's of 150 or above, he will be expected to earn the same mental age of 9 if he takes the 1916 S-B IQ test. But our average present-day 9-year-olds may be expected to score a mental age of 12 on a 1916 S-B. That means that the average 9-year-old today will be expected to earn a mental age of 12 on the 1916 S-B, while the gifted 6-year-old, with the same mental age (9) as average present-day nine-year-olds will earn only a mental age of 9 on the 1916 Stanford-Binet. Our gifted 6-year-old has the same mental age as today's average 9-year-old on today's tests, but has a considerably lower mental age (9 vs. 12) if the 1916 S-B is given to both of them. But that's absurd. implying that our initial proposition that the Flynn Effect drops to zero when the IQ reaches 150 must be wrong... a reductio ad absurdum

3.  Assume That The Flynn Effect IQ Cutoff Is 200
Although this changes the age at which Flynn Effect disappears to a mental age of 12 instead of 9, the end result is the same. The Flynn Effect must disappear at a mental age of 12. But the choice of age was arbitrary. We could have picked 5.

Conclusion: There can be no reasonable IQ for which the Flynn Effect drops to zero.

UPDATE:
There is another, and perhaps clearer way to see into what's going on.
Today's average young adults should score an average IQ of 133 (mental age of 21.67) on the 1916 Stanford Binet IQ Test. Presumably, they will score that same IQ on the 1916 S-B throughout childhood. That means that at age 6, they will have a mental age of 8 on the 1916 S-B. At age 9, they will have a mental age of 12 when measured by the 1916 S-B. When chronologically 12, they will earn a mental age of 16 on the 1916 S-B. Then today's 6-year-old with a ratio IQ of 150 on today's IQ tests (e. g., next year's Fifth Revision of the Stanford Binet) will have a present-day mental age of 9 by definition. But a present-day mental age of 9 corresponds to a mental age of 12 on the 1916 S-B, or a ratio IQ of 200 on the 1916 S-B. Presumably, he or she will register a mental age of 12 if she or he actually takes the 1916 S-B
The Bottom Line:  It seems to me to be a contradiction in terms to suppose that the Flynn Effect is anything but proportional to both the calendar and the ratio IQ. Anything else would lead to a distortion in the IQ distribution curve and a violation of the definition of ratio IQ. (I'm assuming that the Flynn Effect is independent of age from some early age through adulthood.)

It should be straightforward to test this assertion. All we have to do is administer the 1916 S-B or a contemporary IQ test, such as the Thorndike CAVD, the Army Alpha, or the National Intelligence Test  to children with known IQ's. We need to carry this out over a range of ages and IQ's.