I think there might be several places to read a description of these experiments, but they are discussed extensively in a book by historian-philosopher Lindley Darden, entitled "Theory Change in Science: Strategies from Mendelian Genetics", parts of which are available online. See p. 112 of Darden's book for references to other accounts of these experiments by historians E.A. Carlson (1966) and W. Provine (1988).
Everyone knows what Castle did but his aims have been obscured by "Synthesis Historiography" (i.e. making things turn out right for Ernst Mayr and his homies).
What Castle did was to breed rats in the laboratory and implement selection on coat patterns, with such success that he has able to get a large seemingly continuous range of forms. Imagine breeding oreos, with one line of selection of darker forms, and one of lighter, and achieving the success of getting both pure-black or pure-white oreos, and then lining up all the oreos from black to white to show the continuous range of forms.
The story that we hear from people like Wright and Provine is that this proved that selection can work on continuous variation, an idea that the "mutationists" allegedly rejected, paving the way for acceptance of Darwinism.
Yet, while the experiments may have helped one version of Darwinism, they also represented the dying gasps of another form of Darwinism-- the form that Darwin proposed and Castle was (partially) defending.
You can get an inkling of this from p. 143 of Provine's book, where you will find a curious confession, made in passing. Provine is explaining why Pearson did not want to publish Fisher's famous 1918 paper. He writes "Pearson claimed, and Darwin would probably have agreed, that the continuous variations in a pure line were heritable and that continued selection in a pure line should be effective."
Why would Pearson believe-- 17 years after the rediscovery of Mendelian inheritance-- that the environmental variations in Johannsen's pure lines were suitable material for selection? The reason is that this is what Darwin believed, and Pearson was a follower of Darwin. Darwin's "indefinite variability", the fuel for modification in his theory of "natural selection", described in Ch. 1 of the OOS, is clearly a description of environmental variation-- it is always present and emerges anew every generation in response to "conditions of life" (see Winther, 2000). Darwin knew about "definite" single variations or "sports", which could be inherited perfectly (whereas indefinite variations were inherited by blending), but he didn't think they were important for evolution. Darwin bet on the wrong horse, as Johannsen showed in 1903.
Pearson, Castle, and others defended Darwin's original view well into the 20th century. They were not satisfied with the idea that selection merely sorted out stable Mendelian factors that undergo rare mutations. Instead, they held out hope that there was some other form of heredity, or that the Mendelian units were squishy and underwent continuous shifts in potency under conditions of selection. Thus Castle and Phillips in 1914 argued that "the unit character for hooded pattern is itself variable"
Muller, Sturtevant and others from Morgan's lab disagreed, arguing instead that there were simply various Mendelian modifying factors in the background affecting coat color. There was a long-running dispute over the proper genetic interpretation of Castle's results. If you want a very clear statement of the dispute, go to a 1916 article by Jennings, p. 287:
"Castle finds that in rats he can, by selection, gradually increase
of decrease the amount of colour in the coat, passing by continuous
stages from one extreme to the other. As to this, he holds two main
The change is an actual change in the hereditary characteristics of the stock; not a mere result of the recombination of Mendelian
factors. This is the general and fundamental point at issue.
More specifically, he holds it to be an actual change in a single unit factor; this single factor changes its grade in a continuous and
On the other side, the critics of these views
maintain that the changes shown are not actual alterations in the
hereditary constitution at all, but are mere results of recombinations
of Mendelian factors. And specifically, they find a complete
explanation of such results as those of Castle in the hypothesis of
multiple modifying factors."
A few years later in 1919, Castle recanted his earlier view and accepted the multifactorial theory (the development of which is described by Kyung-Man Kim, cited below).
And finally, the idea that the "mutationists" rejected selection on quantitative variation is mistaken. What the mutationists rejected was selection on non-heritable environmental variations. Johannsen's 1903 results were revolutionary because scientists at the time understood that they refuted Darwin's theory. We don't "get" this today because "Darwinism" has been redefined to disentangle "natural selection" from Darwin's errant views of heredity-- see Jean Gayon's "Darwinism's Struggle for Survival" for more explanation.
Kim K-M. 1994. Explaining scientific consensus : the case of Mendelian genetics New York: Guilford Press. xxiv, 239 p. p.
Gayon J. 1998. Darwinism's Struggle for Survival: Heredity andthe Hypothesis of Natural Selection Cambridge, UK: Cambridge University Press.
Winther RG. 2000. Darwin on Variation and Heredity. Journal of the History of Biology 33:425-55.