THE UNDULATION THEORY. 65 does the e of the European Languages Prove for the Existence of an Original European Language ? ( Was heweist das e der Europdischen Sprachen fiir die Annahme einer einheitlic/ien Eurojmischen Grund- sprache ? K. Z., xxiii. p. 373.) If a picture is needed of the way in ■which these partial coincidences were distributed over the Indu- Germanic region, the following diagrana may be found useful: — r. a-e II. k=s,c,sz lll.bh=m IV. r in the Passive V. o gen. fern. Fig. 6. In words, however, the diagram amounts to this : — Just as it is impossible in the diagram to pass out of any one of the spaces en- closed by the five lines drawn therein, without at once falling into a space surrounded by another of those lines, so in the area of the Indo-Germanic languages it is impossible to refer a particular group to an original language peculiar to that group, and so detacli it from the whole, because this would necessarily break the threads uniting and allying that group with all parts of the linguistic area. If we wished to refer the Slavo-Lithuanian, along with the Teutonic, languages to a special group, we should have to ignore the points of relationship (line II.) which bind it to the Hindu-Persian languages. If we wished to get out of this difficulty by ranking the whole of the North European languages nearer to Hindu-Persian, we should have to break the bond (line I.) which embraces all the European languages (and the Armenian), and so on. As, then, according to J. Schmidt, the whole linguistic area of the Indo-Europeans was originally connected together by a chain of "continuous varieties," he has now still to answer the question: How comes it that this state of things no longer exists at the present day, how comes it that instead of the gradual transitions between linguistic regions, such as the Slavonic and Teutonic, the Celtic and Italian, &c., there are now sliarp delimitations of language; that " the unbroken slope from Sanskrit to Celtic " has now become "a flight of steps" {Verwandtschaftsverh., p. 28)? J. Schmidt explains jthis by the dying-out of certain intermediate varieties. Supposing that two dialects, A and X, in the linguistic area were connected together by the varieties B, C, D, Sec, in a continuous chain, it might easily happen that one family or tribe, which spoke