this, and to give you some idea of the sadness of their religious experience before the advent of an age of philosophy, of theological syncretism, and of the worship of the rulers of the state.

Let us now turn for a few minutes to the special difficulties of our subject. These are serious enough; but they have been wonderfully and happily reduced since I began to be interested in the Roman religion some twenty-five years ago. There were then only two really valuable books which dealt with the whole subject. Though I could avail myself of many treatises, good and bad, on particular aspects of it, some few of which still survive, the only two comprehensive and illuminating books were Preller's Römische Mythologie, and Marquardt's volume on the cult in his Staatsverwaltung. Both of these were then already many years old, but they had just been reedited by two eminent scholars thoroughly well equipped for the task—Preller's work by H. Jordan, and Marquardt's by Georg Wissowa. They were written from different points of view; Preller dealt with the deities and the ideas about them rather than with the cults and the priests concerned with them; while Marquardt treated the subject as a part of the administration of government, dealing with the worship and the ius divinum, and claiming that this was the only safe and true way of arriving at the ideas underlying that law and worship.10 Both books are still indispensable for the student; but Marquardt's is the safer guide, as dealing with facts to the exclusion of fancies. The two taken together had collected and sifted the evidence so far as it was then available.

The Corpus Inscriptionum had not at that time got very far, but its first volume, edited by Mommsen, contained the ancient Fasti, which supply us with the religious calendar of early Rome, and with other matter throwing light upon it. This first volume was an invaluable help, and formed the basis (in a second edition) of the book I was eventually able to write on the Roman Festivals of the Period of the Republic. At that time, too, in the 'eighties, Roscher's Lexicon of Greek and Roman Mythology began to appear, which aimed at summing up all that was then known about the deities of both peoples; this is not even yet completed, and many of the earlier articles seem now almost antiquated, as propounding theories which have not met with general acceptance. All these earlier articles are now being superseded by those in the new edition of Pauly's Real-Encyclopädie, edited by Wissowa. Lastly, Wissowa himself in 1902 published a large volume entitled Die Religion und Kultus der Römer, which will probably be for many years the best and safest guide for all students of our subject. Thoroughly trained in the methods of dealing with evidence both literary and archaeological, Wissowa produced a work which, though it has certain limitations, has the great merit of not being likely to lead anyone astray. More skilfully and successfully than any of his predecessors, he avoided the chief danger and difficulty that beset all who meddle with Roman religious antiquities, and invariably lead the unwary to their destruction; he declined to accept as evidence what in nine cases out of ten is no true evidence at all—the statements of ancient authors influenced by Greek ideas and Greek fancy. He holds in the main to the principle laid down by Marquardt, that we may use, as evidence for their religious ideas, what we are told that the Romans did in practising their worship, but must regard with suspicion, and subject to severe criticism, what either they themselves or the Greeks wrote about those religious ideas—that is, about divine beings and their doings.

It is indeed true that the one great difficulty of our subject lies in the nature of the evidence; and it is one which we can never hope entirely to overcome. We have always to bear in mind that the Romans produced no literature till the third century B.C.; and the documentary evidence that survives from an earlier age in the form of inscriptions, or fragments of hymns or of ancient law (such as the calendar of which I spoke just now), is of the most meagre character, and usually most difficult to interpret. Thus the Roman religion stands alone among the religions of ancient civilisations in that we are almost entirely without surviving texts of its forms of prayer, of its hymns or its legends;11 even in Greece the Homeric poems, with all the earliest Greek literature and art, make up to some extent for the want of that documentary evidence which throws a flood of light on the religions of Babylon, Egypt, the Hindus, and the Jewish people. We know in fact as little about the religion of the old Italian populations as we do about that of our own Teutonic ancestors, less perhaps than we do about that of the Celtic peoples. The Romans were a rude and warlike folk, and meddled neither with literature nor philosophy until they came into immediate contact with the Greeks; thus it was that, unfortunately for our purposes, the literary spirit, when at last it was born in Italy, was rather Greek than Roman. When that birth took place Rome had spread her influence over Italy,—perhaps the greatest work she ever accomplished; and thus the latest historian of Latin literature can venture to write that "the greatest time in Roman history was already past when real historical evidence becomes available."12

We have thus to face two formidable facts: (1) that the period covered by my earlier lectures must in honesty be called prehistoric; and (2) that when the Romans themselves began to write about it they did so under the overwhelming influence of Greek culture. With few exceptions, all that we can learn of the early Roman religion from Roman or Greek writers comes to us, not in a pure Roman form, clearly conceived as all things truly Roman were, but seen dimly through the mist of the Hellenistic age. The Roman gods, for example, are made the sport of fancy and the subject of Hellenistic love-stories, by Greek poets and their Roman imitators,13 or are more seriously treated by Graeco-Roman philosophy after a fashion which would have been absolutely incomprehensible to the primitive men in whose minds they first had their being. The process of disentangling the Roman element from the Greek in the literary evidence is one which can never be satisfactorily accomplished; and on the whole it is better, with Wissowa and Marquardt, to hold fast by the facts of the cult, where the distinction between the two is usually obvious, than to flounder about in a slough of what I can only call pseudo-evidence. If all that English people knew about their Anglo-Saxon forefathers were derived from Norman-French chroniclers, how much should we really know about government or religion in the centuries before the Conquest! And yet this comparison gives but a faint idea of the treacherous nature of the literary evidence I am speaking of. It is true indeed that in the last age of the Republic a few Romans began to take something like a scientific interest in