One almost has to remind oneself that it's the physics, or biology, that we're trying to understand, not the math model. 
 
Greg and Henry apply the Fisher-Kolmogorov equation in their new book. Basically, they point out that strong selection, even if uniform in sign over large geographic areas, will not necessarily result in fixation of the favored allele in a given time interval despite the elementary equations saying that fixation will happen. That's because the elementary equations only track the frequency of an allele over time and ignore its distribution over space. 
 
So if the assumptions underlying Fisher's wave of advance (e.g., lad mates with a lass from the next village over) are accepted as a baseline, then an exceptionally rapid fixation of an allele over a large geographic extent requires special dispersive mechanisms (war, migration, etc.).

Especially for two dimensional systems, you can quickly get an intuition for the dynamics of the system, corresponding to paths in phase space. 
 
http://www.math.psu.edu/melvin/phase/newphase.html

Something to do with Jason and the Argonauts?

This is why GTNS is a classic. As your understanding of evolution improves, with each re-reading you notice something else that Fisher stated first.

David, 
 
You can find a good non-mathematical explanation in this book: Measuring Intelligence.

David Rowe and his colleagues objected strenuously to this chain model of causation. They argued that IQ and income should load on a common factor. 
 
As sometimes occurs in path models with latent variables, the semantic interpretation is not entirely clear. 
 
Regardless, the results "feel" different. 35 and 12 percent, respectively, of the variances in IQ and income are accounted for by genetic influences affecting both variables. 30 percent of the variance in income is accounted for by genetic influences that do not affect either IQ or educational attainment. Interestingly, 49 percent of the variance in income is accounted for by nonshared environmental factors that do not not affect either IQ or education.

Bravo, David!

The confidence interval in Schoenemann et al. does not preclude a substantial within-family correlation.

I don't get the "Height is 90% genetic" thing at all. 
 
Take each individual's height, subtract it from the mean height in the population, square the difference, and take the average of the squared differences over the population. This quantity is called the "variance" and can be apportioned to different sources. 
 
If everyone in the population were genetically identical with respect to the loci affecting height, roughly ninety percent of the variance in height would disappear. 
 
Secular increases in height (and IQ) seem to affect the population uniformly. That is, everyone in each successive generation experiences a boost of similar magnitude. The variance in height (and the proportion attributable to genetic variation) does not change under these circumstances (because adding a constant to the mean does not affect the squared deviation from the mean).

David, 
 
If the product of the population size and the selection coefficient (favored type leaves 101 offspring for every 100 left by disfavored type -> selection coefficient = 0.01) is much larger than one, then the allele frequency will change more or less deterministically. If the product is much smaller than one, then random drift predominates. If the product is between these extremes ... then it is more complicated.

Does anyone have an opinion about Fitzgerald's translations?

have you read it? 
 
No. On my Amazon shopping list ...

Thanks, p-ter. Have you checked this out yet?

I've left some comments. They are "in moderation" and so haven't been posted yet ...

If you want to preserve the 2s(N_e/N), just think of s as the selective advantage of heterozygotes over the ancestral homozygote. It is the fate of the mutant in heterozygotes that matters to its chances of fixation anyway. Then fitnesses are essentially parameterized as 1:1+s:1+s+t. If fitnesses are additive, which is an okay assumption for mutants of small effect, this becomes 1:1+s:1+2s.

A.W.F Edwards severely criticizes Wright's views about evolution, particularly with respect to the "adaptive landscape," in his 1994 Biological Reviews article about the Fundamental Theorem of Natural Selection. The crux of the criticism, IIRC, (the article is not available online and I do not have my photocopy with me), is that Wright's intuitions regarding biological complexity and multidimensional spaces do not survive rigorous mathematical scrutiny. However, Edwards credits Wright for making a greater effort to understand the Fundamental Theorem than did other biologists at the time and most biologists since then.

Model 2 includes "only people under the age of 40 from mid-century." This partition is useful because of the ambiguity as to whether the large B-W gap "among persons born in the 1920s and earlier reflect a widening B-W difference among those tested at older ages rather than a difference that was greater in the early 20th century."

i assume this is still broad sense? 
 
Rather complicated. It was assumed that all genetic effects were additive. The presence of dominance deflates the DZ correlation relative to the MZ, while assortative mating (which appears to be substantial in this sample) inflates the correlation. To some extent these factors cancel each other, which allows for good fit to complete additivity tenable in the twin design.  
 
The twin method is not that great at distinguishing between heritabilities in the narrow and broad senses. To get at the former, we need parents and offspring. The Texas Adoption Project is the best parent-offspring study of IQ that I know of. Remarkably, the narrow-sense heritability was estimated in this study to be 0.78. Probably a tad high. But it really does seem to be the case that the additive genetic variance makes up a large proportion of the total genetic variance in g.

the rank-order of individuals on the tasks is invariant, doesn't this mean that there is more than one g factor? 
 
No, it means that there is more than one common factor. Imagine that you administered seven different vocabulary tests. All would load on g, verbal ability, and factors specific to word knowledge. But this does not mean that verbal ability and the specificity are elevated to the status of "g factors"--at least, not to that of the hypothetical g factor, as opposed to the one that happens to emerge from the factor analysis of your particular battery.

Yes, the correlations of these tasks with Raven's Progressive Matrices in this sample of UC Berkeley students are only -0.30, so their g loadings are not as high as those of typical psychometric tests. In the common-factor model this means that they must load not only on g but other factors.

David, what it means that the individuals' rank orders of reaction times on the two tasks, after correction for attenuation, are perfectly correlated (r=1). This is of interest because the brain clearly computes these two tasks in different ways; many variables such as mean reaction time differ across the two tasks. This tends to support the argument (contra Thomson's sampling theory) that two different cognitive processes can be correlated, not necessarily because of overlapping components, but because they are similarly influenced by widely distributed properties of the neural substrate. 
 
I can put the paper in GNXP Backchannel if you like.

which most of the time doesn't resonate with audiences 
 
Shakespeare? Cervantes? Bach? Mozart? Beethoven? Dickens? Tolstoy? Dostoevsky? Each of these artists has been appreciated by millions. 
 
Do you really think Michaelangelo had an IQ less than 115? What about Picasso?  
 
Perhaps this is a difference of emphasis. When I hear the word "artist," I think composers of literature and music. There can be simply be no question that distinguished practitioners of these arts must be of well-above-average intellectual ability. It seems that when you use the word "artist," you mean dancers, musical performers, actors, and others whose work is generally not preserved for posterity. But even here I don't concede. Notice the enrichment for talented classical musicians among East Asians and Ashkenazi Jews. And I suspect that the mean IQ of rappers with albums promoted by respectable labels is well above the African American mean. Jay-Z anyone?

My own experience is that arts talent has zero to do with G, at least once past basic competence.  
 
Michael, this cannot be true of all artistic endeavors. One cohort of the Study of Mathematically Precocious Youth was selected on the basis of being among the top 1 in 10,000 in IQ (as assessed by SAT scores at age 12). This accomplishments of this group include: 7 publications of poetry or prose, 6 noteworthy pieces of art or music (including an apparently well-known rock opera), 2 Fulbright awards, a solo violin debut at age 13 with the Cincinnati Symphony Orchestra, a Mellon Fellow in the Humanities, a Presidential Scholar for Creative Writing, a Hopwood writing award, a Creative Anachronisms Award of Arms, and a first-place prize in midreal-medieval poetry. (This list of accomplishments is rather vague in order to protect the anonymity of SMPY participants.) 
 
Is there any other measurement that you could have taken at age 12 to select as accomplished a cohort? I doubt it.

Thomson showed, quite early on, that g need not be attributable to any *single* common factor, but could be explained by a random overlapping of different elements. 
 
See pages 117-121 for counterarguments to this view.