I have previously noted that the first known phylogenetic network concerned dog breeds, in 1755 (The first phylogenetic network) — a network is needed because many dog breeds are hybrids between other breeds. I have also noted the inappropriate recent tendency to use phylogenetic trees for these breeds, instead (Why do we still use trees for the dog genealogy?). I have also provided a sampling of known phylogenetic networks from the early 20th century (Phylogenetic networks 1900-1990). This post combines all of these themes.
Max Emil Friedrich von Stephanitz was a cavalry captain (rittmeister), but his enduring legacy was as a dog breeder and historian of German dog breeds. His best known book is (available here):
Der Deutsche Schäferhund in Wort und Bild (1921) Ant. Kämpfe, Jena.This was translated into English as:
The German Shepherd Dog in Word and Picture (1923), translated by J. Schwabacher.In this book von Stephanitz argued that the German Shepherd was a specific type of shepherd or herdsman's dog. Indeed, over the previous two decades he had tried to standardize the German Shepherd breed as a working dog, rather than as a show dog in the British tradition of the 19th century.
As part of his argument, he presented a stammbaum of related breeds.
The second picture shows the genealogy from the English translation.
Note that the German Shepherd is not involved in a recent reticulate history, as are all of the herdsman's breeds. This is part of von Stephanitz's argument for the importance of preserving the German Shepherd's identity.
You can read a bit more about the German Shepherd and its ancestor the Hoffwart, the farm guard dog, at König the Hovawart founder revisited: the myth of the Hoffwart.