Once again, I'm looking through source material for a very different reason, but ran across an interesting piece of history. J. Barnard Davis was a British physician and anatomist who, as a private collector, amassed an immense collection of nearly 1800 skulls. His studies on the cranial capacity of these skulls, including comparisons of skulls of different races, were cited by Darwin in The Descent of Man (which is what brought me to Davis' work).
He published this work on human variation in the critical period between the 1859 publication of Darwin's Origin of Species and the 1871 publication of Descent of Man. During this period, Darwin did not publish on the subject of human evolution, but was engaged in a great deal of reading and correspondence on the subject. In his stead, publications and lectures by Huxley, Wallace, and others began to apply Darwinian principles to human variation.
I may write more about Davis, but I wanted to make a note of a passage in the introduction to the 1867 catalog of his collection, the Thesaurus Craniorum [1]. He addresses the importance of a large collection of skulls, which among other things allows an assessment of the breadth of variation within populations. One consequence of our additional sampling of human variation in the 20th century was the recognition that variation among human populations was clinal -- with characteristics forming a gradient across geographic space.
The extent of a collection is of much moment; for, besides affording more reliable averages of measurements, a large one is far more sure to illustrate the types of each race fully, and to contain its aberrant forms. The statement made by Prof. Theodor Waitz, that only small collections of race-skulls exhibit different forms of skulls strikingly whilst rich collections fill up the apparent intervening gaps and show a continual transition from every one form to every other, is only very partially correct, and is an assertion much more characteristic of a Professor of Philosophy than a Professor of Anatomy, essentially a science of observation. Although large collections, philosophically considered, must of necessity, by containing skulls that have intermediate forms, tend to lessen distinctions, they, at the same time, serve to develope [sic] race-characters more fully, and to define the play of diversities round these race-characters with more precision.
The citation to Waitz is to Anthropologie der Naturvölker, 1859.
From a certain point of view, Davis was correct: An increase of sample size will increase the range of the sample, but not greatly increase the proportion of overlap between samples drawn from two populations with different means. Large samples might increase a statistical precision in the description of races. Yet, Waitz' point is also correct. Large samples destroy the typological description of races by showing that no character uniquely typifies any human race. The effect of large samples on the range is often the key evidence that populations share common biology.
Statistics is fundamental to population biology -- so much so that population geneticists like R. A. Fisher and Sewall Wright invented many statistical concepts. Until they understand statistical concepts, people seem inevitably drawn toward essentialism. Essentialism in the history of biology led to typological concepts of species, race, characters, and developmental stages, all of which explained variation in terms of deviation from the ideal type. We now appreciate that populations transform according to statistical rules, not typological rules.
With his immense collection of skulls, Davis showed the extensive variation within populations. He argued so forcefully for the importance of variation that he predicted that the original Neandertal skull would be soon matched within the cranial diversity of living populations. Davis maintained that the skull's elongated shape and browridges could be explained as a result of craniosynostosis, premature closure of the cranial sutures. He looked to human pathology for anatomical intermediates with the Neandertals, arguing that the variation attributable to craniosynostosis would be found to grade continuously right up to the extreme found in the Neandertal skull. In other words, he argued against typology when it came to pathology.
He turned out to be wrong about Neandertal, it was an overreach of his assumptions about variation in development. At the same time, he believed that "race-characters" were stable and that they reflected a long history of separation of human races. He provides an interesting case of how a nineteenth-century anatomist could toss a typological salad.