blossoms. Of course we can express the idea in half a dozen words, but the poetic effect is lost.
As a final major difficulty, there is the fact that the overtones of words are not the same. Sometimes this results from the fact that the thing itself is different; thus, the frog is celebrated in Japan for the beauty of its cries, which are not at all like the croaking with which we are familiar here. But more often it is the poetic tradition which is different. Japanese are never tired of writing about the autumn grasses—all of which have disagreeable Latin names when one attempts to translate—but they seem utterly indifferent, say, to the rose, although familiar with it. This list might be prolonged to cover almost all the most frequent images of both languages.
It must be clear from the above that to appreciate Japanese poetry fully it must be read in the original, but I think that it is possible to communicate some of its qualities by describing the developments in one branch of the poetry. I have chosen the renga, or linked-verse, together with the related haiku. These in some ways are the most Japanese of verse-forms, and suitable therefore as illustrations.
The linked-verse, in its simplest form, consisted of one tanka composed by two people; that is, one person wrote the first three lines, and the other the last two lines, to make one normal poem. An example of this type of linked-verse may be found even in the Record of Ancient Matters (Kojiki) of 712 A.D., the oldest surviving Japanese book, but it was not until the eleventh and twelfth centuries that the form became popular. How it happened that one poem came to be divided in two may be seen by comparing the early anthologies with the New Collection of Ancient and Modern Poetry (Shin Kokinshū) of 1205 A.D. In the earlier poetry there was no fixed place for the break in the verse, which would come at the end of the second, third, or
