Many of the fragments of ancient writers have come down to us through rather curious channels. In the later age of Rome it became the fashion to make anthologies and compilations, and it is through such collections that the majority of classical authors are known. One of the most curious of these anthologies is that made by Athenæus about the beginning of the third century A.D. This author called his work Deipnosophistæ, or the Feast of the Learned. He attempted to give it a somewhat artistic form, making it ostensibly a dialogue in which the sayings of a company of diners were related to a friend who was not present at the banquet. The diners were supposed to have introduced quotations from the classical writers, so that the book is chiefly made up of such quotations. The work has not come down to us quite in its entirety, but, even so, no fewer than eight hundred authors and twenty-five hundred different works are represented in the anthology. Of these authors about seven hundred are known exclusively through the excerpts of Athenæus.
Two or three centuries later another Greek named Stobæus compiled a set of extracts from the Greek writers of all accessible periods prior to his own. The number of authors quoted in this anthology is more than five hundred, and here again the major part of them are quite unknown to us except through this single source. Yet another collection of excerpts was made in the latter part of the ninth century by Photius, patriarch of Constantinople, who made excerpts from about 280 authors with whose works he had familiarised himself through miscellaneous reading. In addition to these works of individual compilers there were two or three anthologies compiled in the Byzantine period, including an important collection of fragments of the Greek poets which is still extant under the title of The Greek Anthology, and the elaborate set of encyclopædias made under the direction of Constantine Porphyrogenitus. But for such collections as these, supplemented by the biographical notices of such workers as Suidas, and by fragments that have come to us through a few other channels, it would scarcely have been conceived that so many authors had written in the entire period of Grecian activity, since only a fraction of this number are represented by complete works that have come down to us. Such facts as these give an inkling as to the mental activity of the old-time author, while pointing a useful lesson as to the perishability of human works. In this age of easy multiplying of books through printing, one is prone to forget how precarious must have been the existence of a manuscript of the elder day. It was a long, laborious task to produce an edition of a single copy of any extended work, and each successive duplication was precisely as slow and as difficult as the first. Under these circumstances no doubt a very considerable proportion of books were never duplicated at all, and the circulation of a very large additional number most likely was limited to two or three copies. It was only works which were early recognised as having an unusual intrinsic interest or value that stood any reasonable chance of being copied often enough to insure preservation through many succeeding generations.