He argues his point still more vividly: If a Scotsman uses a split-bamboo trout rod, a waterproof silkline, and a barbed hook, it is not necessarily a case of diffusion if a man in Saskatchewan is found to be fishing with a willow twig, a piece of string, and a bent-pin, since each item is dependent upon the others. But if in addition to their fishing tackle, the Scotsman and the man in Saskatchewan are found to possess a shot-gun, a flask, a brier-pipe and bagpipes, then it would look like a case of diffusion since no one item of the assemblage is dependent upon any other.[9] Dispersion as Well as Diffusion The difficulty of this second step in the diffusionist’s argument lies in the fact that it is hard indeed to find a complex of traits in _one_ American locality that resembles exactly a complex of traits in a _single_ Old World one. If the traits are all together in Peru, some may come from one place in the Old World and some from another. Or, if we take a group of traits from a single Old World locale, we find them spread out widely and separately in the Americas. An excellent example of this may be drawn from Oceania and South America. Dixon writes of the diffusionists: When in South America, they say, you find not only coca-chewing, plank canoes, and tie-dyeing, but also terraced irrigation, Panpipes, and the blow gun—all traits widespread in the western Pacific and southeastern Asia—how can you deny that their occurrence is due to diffusion, or believe for a moment that so many similar and parallel inventions could take place? The challenge is a formidable one. Is there anything that can be said in reply? Dixon points out that these Oceanic traits are not found _together_ in the New World. The plank canoe is confined to the Santa Barbara Islands and southern Chile; tie-dyeing, to the arid coasts of northern Peru; coca-chewing, originally to the Andean highlands and the tropical forests along its eastern border; terraced irrigation, to the Andes of Peru and Bolivia; the blow gun, to the upper Amazon and Orinoco forests, the Antilles, and the eastern United States; the Panpipe, to the Amazon-Orinoco drainage and southward through Bolivia to northern Chile and the Peruvian coast, and to one or two isolated spots in Ecuador and Colombia. “With one exception the only area where the distribution of any two of these traits is found to overlap lies in the Andean highlands and the tropical forest area to the eastward. Only tie-dyeing and the Panpipe are found together on the coast.” Further, the two traits we find on the coast are separated in the Old World. Tie-dyeing is found specifically in Indonesia “and known in Melanesia only in degenerate form in one small area, whereas the Panpipe is primarily Melanesian and almost unknown in Indonesia.”[10] He seems to be ignorant of double Panpipes connected by a cord which are found in the hinterland of Burma and also in Panama and South America.[11]