they are careful of their scientific materials; (6) That the peculiar nature of these materials as the only revealers of the growth of the culture and civilisation of the nation, their extreme rarity and special liability to destruction and loss, require that any such effort must be special and speedy, as well as national, in order to be effectual; and (7) That we owe it as a duty to our country to transmit to posterity all its monuments and relics of national interest, of which our position as the present possessors has constituted us the guardians in trust for all future generations.

LECTURE II.

(17TH OCTOBER 1879.)

STRUCTURAL REMAINS OF THE EARLY CELTIC CHURCH.

THE characteristics of ecclesiastical architecture in Scotland in the first half of the twelfth century are well known. But previous to that time there are the ecclesiastical constructions of five centuries to be accounted for. The object of this Lecture is to inquire whether it may be possible to determine the types of the earlier constructions of Christian character and use, and to demonstrate the sequence of these types by examination and comparison of the existing remains.

The earliest of the existing twelfth-century churches1

1 The great conventual and cathedral churches erected, or begun to be erected, chiefly in the reign of King David I.—Jedburgh, Inchcolm, Kelso, Holyrood, Melrose, Dunblane, Cambuskenneth, Dunfermline, Dryburgh, etc.-do not now exhibit much of the work of that early time. The nave of Dunfermline, as it was rebuilt by King David, is partially preserved, but in all the others the earliest work is only to be recognised in fragmentary portions of the existing structures. Examples of the smaller churches, however, are more numerous, and of these Mr. Muir has given a pretty complete list in the first chapter of his Characteristics of Old Church Architecture in Scotland. It is curious that we should have to look to the distant Orkneys, and to the work of an alien people, for the best preserved example of the Romanesque in Scotland. The Cathedral of St. Magnus, designed by the Norwegian Kol, and commenced by Earl Rognvald in 1137, contains "the greatest amount of Norman work of any building in Scotland," and in its internal aspect, according to Mr. Muir, is "nowhere equalled by any interior in Scotland." It is equally curious that we must look to Orkney for the only specimen in Scotland of a circular church-that at Orphir, now only a mere fragment. This interesting ruin has been adduced as an example of the development of a church from the early drybuilt circular or beehive

present certain features of form and construction which are of importance for this inquiry as enabling us to separate the known from the unknown. They give us a typical form consisting of nave and chancel, and the typical features of construction are round arches with radiating joints over doors and windows having perpendicular jambs or sides.

Having thus obtained a fixed starting-point in time, and a known type of structure for comparison, I proceed to deal with the unascertained types on the principles of archæological classification.

Supposing that all the ecclesiastical structures which are known by their characteristics of construction or decoration to be twelfth century or later, are swept away, there are left a considerable number, of which the most that can be said is, that they possess no architectural features in the common acceptation of the term-no moulding, ornament, or distinctive feature to be found in the architectural books. It is obvious that in this residue there must be some that are earlier than others. It is possible that there may be some that are even later than the twelfth century, because the earlier type may have survived longer in some places than in others. It is even possible, at least it is conceivable, that there may be places where the earlier type was never superseded by a later type at all. This, we must remember, produces an element of uncertainty in dealing with individual specimens. But in dealing with types or classes it has no disturbing influence, because, if the sequence of the types can be established, it is of no moment whether the specimens may be early or may be late examples of the type.

dwellings of the native inhabitants; but it is on record that the Norwegian Earl Hakon, who had his residence at Orphir, made the pilgrimage to Jerusalem in expiation of the murder of St. Magnus, and as the church is plainly one of the well-known twelfth-century imitations of the Church of the Holy Sepulchre, it was more likely to have been erected by him than by any one previous to his time.

In point of fact, it is not necessary that we should be able to say of the specimens to be used for the determination of the succession of types, that they severally belong to the periods when their types were prevalent, for the succession of types may be determined, although no specimen of the period proper to each type may exist. Let us take a case in illustration, which we can verify from actual experience. We know that the succession of the typical forms of the instruments used in spinning was as follows:-(1) The whorl and spindle; (2) The spinning-wheel; (3) The machine driving many spindles by water or steam. Now, although no spindle of the time when spindles only were used were extant, although no wheel of the time previous to the invention of the machine driven by water or steam were preserved, we should still be able to demonstrate the complete succession from specimens of the two earlier types made and used in the days when the third type alone was prevalent. So, it is not a necessary step in my argument to furnish proof that any of the churches from which I deduce the types in existence previous to the twelfth century, are themselves earlier than that time. The age of the particular specimens is not the object of the investigation. It is a term in the equation which must always be expressed by an unknown quantity, but the fact of its being unknown, or even unknowable, does not render the solution of the problem impossible.

Proceeding then to the examination of the residue which remains after all the churches of known twelfth-century date or later are swept away, we shall find that on grounds of structural form alone they are separable into two classeschurches with nave and chancel, and churches that are simple oblongs of a single chamber only. Of these two typical forms of structure, the more complex and refined is certainly the later. We shall find it passing from great rudeness of style and construction up to the decorated style and elegant con

struction of the Norman manner, and thus linking itself on to the current architecture of the twelfth century. Even its greatest rudeness stops short of reaching the rudeness of a primitive style. We have chancelled churches built of unhewn stones, but we have none built without mortar. On the other hand, the single-chambered type of church, which never rises to the dignity of the Norman edifice, can be traced backwards by a series of gradations of style and construction into a type which is truly primitive, and corresponds in all its characteristics with the type of the earliest churches in Ireland, from which our Christian institutions, as ultimately established, were derived. The conclusion is, therefore, that by thus subdividing the residue which is left, after all churches of known twelfth-century date and later are swept away, we obtain two types, of which the type with nave and chancel is the later, and the single-chambered type the earlier. This does not imply that any church which may be found to be constructed of a single chamber is on that account necessarily earlier than one constructed with a chancel. The conclusion relates to the typical form alone, and does not apply necessarily to all the specimens of that form. Whether any of them which now exist may be of such an early date requires a separate demonstration and special evidence. The thing now determined is merely that the single-chambered form goes farther back than the chancelled form. What is the precise date of any specimens of either of them is a question which is undeterminable by any scientific method. No better illustration of this could be adduced than the example which I shall presently cite.

The chancelled churches on the mainland of Scotland are mostly Norman in style. This fact removes them from among those with which I have to deal. An exception may be made in the case of one which is unique in its features. This exception is the Church of St. Regulus (Fig. 1), whose