Darwin was the first to envisage global common descent for species. Who was the first to make the analagous conceptual leap for genes? In other words, who was the first to propose that all genes originate by duplication of other genes?
I've not yet done a proper historical analysis, but I have made a little progress on these questions. A good place to start with a historical analysis of gene duplication is with a review article that I'm a fan of (Taylor and Raes, 2004). They point out that ideas about the copying of genetic material date back to the early 1900's, especially with studies on chromosome duplication and ploidy changes.
But this is a different question. The idea that chromosomes or even genes can be duplicated is not tantamount to a hypothesis of global common descent for those genes. I've argued, in a paper with Michael Rose, published in Biology Direct, that even if early evolutionists and geneticists recognized gene duplication, they did not often recognize the possibility of the relatedness of all genes. Instead, genes were thought to be molded very quickly by natural selection, such that their history is quickly erased. Here is a graphical representation of this idea:
There are two critical differences in figures A and B above. First, in A, genes have only short histories. Whatever their origins (and often their origins were not contemplated, as genes represented very abstract entities), their history was short. Natural selection should quickly mold gene function to current utility and erase history. Second, in B, genes have a branching history of their own, semi-independent of the branching history of species.
In my initial attempts at understanding the history global common ancestry for genes, I've focused on finding the origin of the idea of separate histories for genes and species. One of the best candidates is a paper by V Ingram in Nature, 1961. (Here is a biography of Ingram, a protein biochemist). Ingram had obtained the first amino acid sequences of hemoglobins. He noticed the similarities of different human hemoglobins (e.g. fetal, adult), and constructed a gene tree of hemoglobins as the only figure in the article (everyone knows the only figure published in Origin of Species, right?). Ingram concludes with the sentence "Such a scheme involves an increase in the number of haemoglobin genes from one to five by repeated gene duplication and translocations; the scheme may thus illustrate a general phenomenon in gene evolution."
Ingram was close. But "general phenomenon" does not equate with an idea that gene duplication is the primary or even sole mechanism for the origin of new genes. Perhaps Ohno (1970) or Zuckerkandl (1975) make the bold leap to common ancestry of all genes. I will have to check into that when I get a chance.