[From Manx Crosses, 1907]

CLASS II.-SCANDINAVIAN

Our Scandinavian monuments are in all respects a continuation of the earlier series. They consist of upright headstones, generally rectangular or slightly tapering below, sometimes round-headed, in one or two instances with the spaces between the limbs pierced, and in two cases cruciform in outline. The Celtic type of cross which had now become the established form was that adopted, and in all cases figured by the Scandinavian sculptors, with slight variations of form. Though we cannot allow Gaut's claim that he carved " all the crosses in Man," we must recognise by it that he was the first Scandinavian sculptor,1 and the two pieces which bear his name, as well as the four or five others which appear to show his handiwork, have the Celtic cross and connecting circle of the form figured by Mr. Romilly Allen, 102A.2 Of these we may regard Andreas 73A as the type. The head measures from the centre of the limbs thirteen inches by seven inches wide; the shaft fifty-four inches by eight and a quarter inches to eight and a half inches at the foot; and the arms eight and a half inches by seven and three-quarter inches. The connecting circle is two and a half inches wide and about seventeen inches in outer diameter (fig. 25).

A form distinct, differing from this in the narrower limbs and proportionately longer shaft, and having no circle, occurs also at Andreas, and not elsewhere; of this I take the Santulf cross 103A as the type. The head measures seven inches by three and three- quarter inches ; arms, six and three-quarter inches bythree and a half inches; and shaft, thirty- eight inches by three and three- quarter inches, tapering to the bottom by five-eighths of an inch.

FIG. 25.-FORMS OF CROSSES, SCANDINAVIAN.
(1) Andreas 73A ; (2) Andreas 103B ; (3) Maughold 7213; (4) Maughold 91A; (5) Bride 97A.

A third form is that at Maughold on the Roolwer cross 72, which approaches the Irish form with semi-circular expansion to the limbs, as on the Chobthac (Coffey) cross, and as regards the very long upper limb and short arms, several other early slabs at Clonmacnois,³ but none show quite the same proportions. On face " B " the measurements are as follows :- Head, sixteen and a half inches by two and a half inches to two and three-quarter inches, expanding at the top to five and a half inches ; shaft, eighteen and three-quarter inches by two and three-quarter inches to three inches, with truncated pedestal four and three-quarter inches wide; arms, seven inches by three inches, expanding to four and a quarter inches. The expansions at the ends of the limbs are lightened by central hollow recesses. The circle connecting the limbs is a true one, and two inches wide, with an outside diameter of eleven inches. In only four of this class do we find the cross with approximately equal limbs; in two of these the surrounding circle is a true one. The first-Maughold 91A-is of the Celtic form, with square ends overlapping the circle in such way as to leave the recessed spaces with thistle-head outline, the limbs being four and three-quarter inches long by two and a half inches to three and a quarter inches wide, and the square ends two and a half inches deep. A shaft has, however, been added below, eight inches long by three and three-quarter inches to four and three-quarter inches wide. Bride 97 has the limbs expanding to the point where they are cut off by the surrounding circle. They are seven inches long by three and three- quarters to five inches wide. Braddan 108, and Conchan 113, are slight variations of these. These Scandinavian pieces occur in twelve out of our seventeen parishes in the following numbers: -Nine, seven (three times), four, three (twice), and one (five times) : thirty-four of them in the Northern and eleven in the southern parishes.

We meet with as little symbolism in this Class as in the former-figures which may have been symbolic in their origin appear to be treated rather as decorative designs; for example, the two birds above the arms of the cross. The use of the triquetra would seem to be always decorative. The pastoral staff on the Maughold slab 72 was probably symbolic of the episcopal office (see p. 30), as the mitre at a later period. There remains the cross itself, with the circle, as in Class I.-the great symbol of the Christian faith.

As our earliest Celtic pieces are in the vernacular, and incised in the peculiar characters known as Ogams, so are our Scandinavian inscriptions-not in the Latin language and Roman characters, but in the Norse vernacular, and incised in Runes. Out of the total of forty-five, we find that twenty-six bear runic inscriptions,' eighteen in the Northern and eight in the Southern District. Probably many of the remaining nineteen also bore inscriptions, but if so these are now broken off and lost.

The Christian symbol appears in the form of the Celtic cross on all, the two late inscriptions at Maughold alone excepted, and we find also that the decorative treatment and the designs themselves are of Celtic type and origin. But they are by no means mere copies: the artists have entered into the spirit of the earlier work, have developed simple designs, and evolved from them new patterns. It was not from any monuments in the Isle of Man that they got their inspiration. Such a characteristic Celtic design as the step-pattern which occurs several times on our Scandinavian pieces2 is not met with once on our earlier monuments; the triple divergent spiral which occurs on the Roolwer cross, Maughold 7=', and the chevron on the Bride Cross 97, are unknown to our Celtic pieces. In figure drawing we notice the frequent use of the curious little spiral at the junction of the limbs, which, though we find it occasionally on Scandinavian carving of the Viking period, as on the Ramsund and Goek stones, and on heathen stones from Klistad and Langthora, and Overselo, Sõdermanland, Sweden,' appears to be of Celtic origin. So, too, we find a peculiar Nimbus, with fringed edges, which is not found outside the Isle of Man, though the three crosslets with which it is decorated might have been suggested by the Book of Kells or some other illuminated MS4 .4

Gaut, the first and greatest of our Scandinavian sculptors, introduced a peculiar pattern- the Ring-chain-which became such a favourite with his successors that it is met with in all eighteen times, besides the form taken in the terminations to interlaced bands and other developments or applications of it, and may be considered a characteristic of Manx art, as elsewhere in the British Isles it is not found more than five or six times.' It was Gaut, also, who developed from a simple Twist of two bands a. series of knots and designs culminating in the graceful Tendril pattern which is seen in perfection on the Ballaugh Cross 77. This, too, becomes a characteristic of the Manx Crosses, occurring twelve times.

Finally, it was Gaut who invented the peculiar Manx decorative treatment of the head of the cross, which became general in the Isle of Man, and, as it is not to be found elsewhere, would serve to identify a Manx cross if ever removed from the Island.

In considering the art of our earlier Celtic pieces we have taken them in the order of their development from simple boulders, pillars, and slabs which have no ornament at all, to

FIG. 26.-DECORATIVE TREATMENT OF THE HEAD OF THE MANX SCANDINAVIAN CROSSES.
(I) Michael 74A; (2, 3) Ballaugh 77A, s; (4, 5) Jurby 99A, B.

those pieces sculptured in relief with elaborate geometrical patterns and pictorial representa- tions, this arrangement falling in with their probable chronological sequence.

With the Scandinavian pieces it will be more convenient to follow here the same plan, and to treat them in the order of decorative development. In doing so, it will be necessary to alter the arrangement somewhat from that in the descriptive portion of our work, where they are arranged, so far as possible, in what appears to be their probable order in time; so much so, that the few pieces which show absolutely no ornament, no sign of forming the outline of the stone or dressing the surface, and so are here first referred to, belong chronologically to the very end of the series.

FIG. 27.-FORMS OF LINK-TWIST.
(I) Andreas 73A; (2) Andreas 83B; (2a) Andreas 84A; (3) Michael 74B; (4) Jurby 99B I. ; (5) Malew 94A.

Such are two unhewn and unshaped slabs of the local clay- slate from Maughold, one, 114 found near an old Keeil at the head of Corna Valley, and the other, I 15, from the Parish Church. Both of these, to judge from their inscriptions, are as late as the end of the twelfth or beginning of the thirteenth century. Another, from Conchan 113, has incised on each face a rudely- drawn Celtic cross and circle. Except in the form of the cross itself, which is bordered by a fine line, it exhibits no sort of artistic treatment. Both faces are inscribed in late runes.

A broken slab from Marown 111 cannot be placed, as what now remains shows no trace of sculpture or design. The use of the words " this cross " shows that there has been some carving on the portion now lost. So in Peel Cathedral, German, another slab, 112, of which only the edge is visible, bears the words "this cross," and has, no doubt, been carved on one or both faces.

With these few exceptions, our Scandinavian pieces are handsomely decorated on both faces, and sometimes on the edges. Several exhibit geometrical designs only. First among these is that which we are told in the inscription was carved by Gaut of Cooilley, Andreas 73. Each face bears a cross of our first and most general type, with the limbs connected by a circle. The treatment of the head is peculiar to the Isle of Man (fig. 26). A pair of bands carried up from the plait-work or other design on the shaft interlaces in a central ring with another pair crossing the arms at right angles to the first.2 The extremities loop backwards, and connect in a V-shaped tongue to buckle with an interlacing ring, thus presenting the appearance of the ring-chain pattern, by which possibly this may have been suggested. Some-times the bands split to form diagonal oval rings, interlacing with a square or diamond-shaped one. This appears to have been the case in the upper limb on both faces of this cross. The shaft on each face bears a plain plait-of-five, the spaces to the right and left of it being occupied on the one face by a key-fret (No. 899 of the Analysis) and twist-and-ring (No. 574); this latter design is met with in Ireland only at Castle Dermot, and three times in Wales, but in Scotland on twenty-four stones and in England nineteen. The other face shows a plain twist and a development of this simple design which I term a link-twist, forming one of the series leading up to the tendril pattern (fig. 27).

The form here taken, i, has an off-shoot or link at each point of crossing thrown out from the inside of the right-hand band to pass under and over, linking on to the other band above the point of crossing. Except in this one instance, the link is always thrown off from the outside of the right-hand band, passing horizontally instead of vertically as, 2, in Andreas, where it passes over and under, to unite below the point of crossing, with the left band. In Michael 75 the edge of the stone is unfortunately broken, but I believe the design took this form. Again, 3, the direction may be reversed, the link passing below instead of above, as in Michael 74. Cumming' figures a development, 4, which is practically the same as the second, but carried further, so as to return to the band of origin below, thus forming a complete loop. I have, however, examined this stone, Jurby 99, with the utmost care, and though it is very badly worn, I think that Cumming was mistaken, and that the lower end of the off-shoot terminates in a loose coil below the two bands. There is an interesting variation on the Malew slab, 5, where the link is thrown from the inside of the right-hand band, passing downwards and round, so as to loop over itself and link on to the inside of the left-hand band. Lastly, 6, we have the fully-developed tendril pattern, which is seen in beautiful perfection on Ballaugh 77, and is of frequent occurrence in the work of Gaut and his successors.2

FIG. 28.-TENDRIL PATTERN AND DESIGNS OF LINKED BANDS.
(1) Ballaugh 77B ; (2) Bride 92B ; (3) Jurby 99B ; (4) Ballaugh 77A.

Still further patterns are developed from these designs, as 7, on Bride 92, where two bands not twisted, but converging and diverging at regular intervals-it might be regarded as a twist with vertical breaks-throw off from the outside two links, one passing above, the other below, the spaces being occupied by pellets. From this we get 8, the knot on the stone at Jurby 99, which takes a form as of a double triquetra back to back. And upon Ballaugh 77 we have a design, 9, similar to the last, but the bands connected by a single loop passing downwards, and expanding into the V-shaped tongue to buckle with an inter- lacing ring, being, in fact, the same idea as in Gaut's treatment of the extremities of the limbs (fig. 28).

The other cross slab, Michael 74, which bears Gaut's name, with the bold assertion that he made it, "and all in Man," shows also pure geometrical decoration. The shaft on one face exhibits our first instance of the ring-chain pattern (see fig. 29). This has been called "vertebral pattern," and more aptly, I think, by Cumming, "chain-cable." On our Manx monuments it invariably takes the form of a single line or chain, which upon analysis is found to be composed of three distinct sets of linked rings interlaced, the lower part of each ring expanding into a Y-shaped tongue, where it overlaps the next below.

It appears to be derived from a design such as we meet with around the lower part of the shaft of the beautiful Gosforth cross, near Whitehaven, and was thought by the late Mr. Calver- ley to have originated in a conventional representation of the interlacing foliage of the sacred tree, Yggdrasil !' This view is supported by the design on one face of the Dearham cross,2 on which we see the stem of the tree from which the interlacements arise. In the form of a double row, placed horizontally, it occurs on a granite font from Gdllstead, West Gottland, figured by Du Chaillu in The Viking Age, i., 186.

The next step was to apply the design to fill in a long panel by reducing it to a single row, which becomes a chain of rings; this we see on the west and east faces of the Gosforth cross, where it is formed by a double bead, and again on the south side. A wooden Spoon found about 1845 in a ship twelve to fourteen feet below the level of the streets at Upsala, and described by Prof. Carl Save, bears this design on the haft, which ends in a dragon-head like those on Rune-stones (O.N.R. Hon., ii., 98). In a single row, also, we meet with it at Muncaster, Cumberland,3 and at Burnsall, Yorkshire. The only other examples in Britain are at Penmon Priory, Anglesey,4 and Cardynham, Bodmin, Cornwall.5 In the two latter the position of the tongue or buckle is reversed, turning downwards instead of upwards, and it is in this form that it is invariably applied in the Isle of Man, where it occurs no less than eighteen times, almost always as a decoration on the shaft of the cross.

The space to the right of the shaft on this face of the stone contains the fully-developed tendril pattern. On the left is a plain plait-of-four; the spaces between, not being cut away, are occupied by rectangular pellets.

On the other face, the shaft has plait-of-five, the space to the left twist-and-ring (with almond-shaped pellets), and that to the right the variety 3 of link twist, in which the off-shoot is thrown, downwards and horizontally, to lap over and under, and link on to the other band. Above the circle we find the design of two diagonal rings interlaced.

In Michael 75 we have the ring-chain design on the shaft, to the right a plain twist with little oblong pellets between the bands, and to the left form 2 of link-twist. The spaces between the bands are left, forming long flat pellets, which are ornamented by little incised chevrons. The other face has twist-and-ring on the shaft, which is very exceptional;

FIG 29. - EVOLUTION OF THE RING CHAIN DESIGN.1

(i) Cross-Canonhy ; (2, §) Gosforth ; (3) Dearham ; (4) Panel on Font, Gi llstead ; (6) Cardynham. (7) Michael 749, shaded to show how the Manx form consists of three chains of rings interlaced.

the bands and rings are unusually broad, and the spaces between occupied by truncated pellets. It is finished below by the use of incised spirals, such as may be seen on many Scandinavian blocks early and late. To the right is a plain key-fret, the inscription occupying the space to the left.

In the Port St. Mary slab, Rushen 76, we meet again with the ring-chain and the tendril pattern, also plaits-of-five and of four and twist-and-ring (twice).

In Ballaugh 77 the outline of the stone has been shaped to form a wheel-head and expanding base. The cross on one face lacks the surrounding circle; the shaft and limbs have a slight taper to their extremities, which is exceptional. The treatment of the head is after Gaut's manner, but a central boss takes the place of a ring, and the loop terminations in the limbs of one face take the form of triquetra knots. The shaft of this face shows the tendril pattern decorated by pellets and double beads; on the left is a double-beaded key-fret, and on the right the ring-chain finished below in spiral flourishes, and terminating at the top in a small Celtic cross. Above the head, a plait-of-three is continued to end at either side in two diagonal oval rings, double-beaded and flourished at the ends. The other face has on the shaft the development of link design, before described (p. 43, fig. 28, 4), showing converging bands connected by a tongue-shaped loop buckling into an interlacing ring, and decorated with pellets. The space to the right has a plait-of-four ending in elaborate flourishes; that to the left bears the inscription.

The Nappin cross, Jurby 78, has on one face, for the first time, the circle ornamented, showing a plait-of-four within plain bead borders. What remains of one arm bears Gaut's treatment with the ring pelleted ; and the space to the right of the shaft has the tendril pattern, which is curiously decorated with little curved scores. On the other face the shaft has plait-of-five, and to the right ring-chain; the space to the left has had interlacing of broad and narrow bands now almost entirely worn away.

The two fragments from Jurby 79, 8o, also seem to show Gaut's treatment of the head; the bands are pelleted.

In the St. John's slab 81 we have the name of another carver, Asruth. Unfortunately, only a portion of the shaft remains; it shows the ring-chain pattern, plain. One edge has an irregular design of diagonal scores from right to left, and from left to right; the other bears the inscription. This stone is exceptional in having only one face carved. A broken slab from Maughold 82 is very similar, but the ring-chain is decorated by border lines. The other face also is carved, having the double twist and diamond-shaped ring design, double- beaded, with the rings bordered by fine lines, the spaces where the plait gets looser in the lower part being filled in with pellets. The shaft is the full width of the stone, and the arms must have been unusually short, even if the stone was wider at the top, which is now broken off.

A fragment from Andreas 83 shows ring-chain, key, and form 2 of link-twist as on Michael 75.

The bind-rune stone, Andreas 84, shows tendril, key, twist-and-ring, and link-twist, all double-beaded, the latter of the form 2.

The altered stone, Michael 85, turned round and converted, in 1699, into a tombstone with skull and cross-bones, has plait-of-five (probably),,tendril pattern, double-beaded, and twist-and-ring. The large slab, Braddan 86, displays a new treatment of the head, the bands carried from the shaft, turning to the right and left, instead of passing through the centre, and at the ends forming two loops linked by a ring. One face, very much worn, shows traces of plain key-fret and of some interlacing, the other ring-chain and twist-and-ring, with the inscription at the right of the shaft (fig. 30, 5).

Andreas, as before stated (p. 39), gives us a new variety in the form of the cross, the shaft narrow in proportion to its length, as well as to the stretch of arms, and the circle absent. We meet with this form in Andreas 87, which shows faint traces of ring-chain pattern and of plait-of-four or it may have been a link-twist. The cross on the Roskitil stone, Braddan no, is of the ordinary type, but has no circle; only a portion of the head remains. The arm shows two interlaced diagonal rings, double-beaded; the inscription occupies the shaft and the space above the left arm. The other face shows two pelleted bands crossing and interlacing with two others, but without central ring; these terminate in loops with curious foliations, and two circular rings interlaced. The remaining space between the limbs is occupied by diagonal rings interlaced, above which a narrow diagonal fillet.

FIG. 30.-DECORATIVE TREATMENT OF HEAD.
(1,, 2) Braddan 108A, B ; (3, 4) Michael 89A, B ; (5) Braddan 86A.

A fragment from Peel Castle, German 88, is the first of the Scandinavian pieces to have the spaces between the limbs pierced. The ornament is a loose plait, too fragmentary to show the character.

1 With the possible exception, however, of the Maughold slab 72, certainly carved by another hand. But, if this was earlier than Gaut's work, it must have been unknown to him; it is more likely to have been later.

2 The Early Christian Monuments of Scotland, part ii., p. 52.

3 Petrie, Christian inscriptions, pl. xiii., 32 ; also pls. x. to xvi.

1 One of them, Michael 104, has two inscriptions in Runes and two in Ogams.

2 We find this design on the heathen block at Eke, Upland, Sweden (O.N.R. Mon., ii., 672) ; also on the Ramsund rock. 3 O.N.R. Mon., ii., 733, 746, 769. Sigurds-Ristningarna å Ramsunds-Berget och Goks-stenen.-Carl Save, 188.

4 The use of the crosslets seems to be a refinement on the Cruciferous Nimbus of Byzantine Art. In sculpture, we meet with this in Norman work. See article by Mr. Romilly Allen on " Sculptured Norman Capitals at Southwell Minster" (The Illustrated Archæologist, vol. i., 1894).

5 At Gosforth and Muncaster, Cumberland, and Burnsall, Yorkshire. One in Anglesey and one in Cornwall alone show it in the form taken in the Isle of Man. Owen Jones, in his Grammar of Ornament, Chapter xv., considers this to have been " probably borrowed from the Roman tesselated pavements on which it is occasionally found." But his examples, pl. xxii., figs. 22, 27, have only a superficial and accidental resemblance-consisting of two parallel rows of bold cable pattern, which, at each point of crossing, throw inwards and upwards an offshoot to connect with the other, so as to leave a V-shaped space between the cables. -

6 See page 4^, fig. 28, 1.

i According to Cumming (fig. jib). The stone is much worn, but I think the design was as in fig. 28, 1. I give Cumming's figure here, however, as being a likely step in the evolution of the tendril pattern.

2 The only instance I know of a cross-head decorated with two vertical and two horizontal bands interlacing is at Barochan, Renfrew, which, however, has two concentric rings, and the bands terminate in plait-work, but they are not carried up from the shaft.-Early Christian Afonamzents, iii., 775. The only other which approaches our Manx crosses in this respect is at Kildaltan, Islay, where the bands turn to right and left and terminate in diagonal rings, interlaced with a circular one.-Carved Stones of Islay, Graham, No. So.

1 The Runic and other Monumental Remains of the Isle of Man, pl. iii., fig. iIb.

2 Elsewhere the nearest approach to this design that I know of is on the Rossmarkie Cross (left side, middle panel), where the interlacing bands split, and are continued in long tendrils to form loops.-Early Christian Monuments in Scotland, iii., p. 67. Since writing the above, I have heard from Mr. Romilly Allen that it occurs on the bell-shrine of St. Mura and in Caedmon's Paraphrase of the Scriptures.

1 Early Sculptured Crosses, etc., in the Diocese of Carlisle, Rev. W. S. Calverley.

2 Ibid., p. 124.

3 Ibid., p. 238. 4 Westwood, Lap. Wal., 84, 1. 5 Langdon, Old Cornish Crosses, p, 356, pl,

1 (1, 3) From Calverley, Crosses, et-, Carlisle; (a, §) Stuart, Sculp. Slones; (4) Du Chaillu; (6) Langdon, Cornish Crosses.

GEOMETRICAL AND ZOOMORPHIC DESIGNS

FIG. 31--GEOMETRICAL DESIGNS ON SCANDINAVIAN PIECES : SPIRALS.
(1) 2, 3, 4, 7) Maughold 72; (5, 6) Bride 97.

A large slab at Michael 89 shows similar workmanship to the last. The head is rounded, the stone tapers to the foot, the spaces between the limbs and the circle are pierced. The cross varies in form from the type I (p. 39), in having limbs and shaft expanding, and the circle connecting the limbs is a true one, not composed of segments of four different circles, as is usually the case. The treatment of the head differs from that of Gaut, though on one face evidently suggested by it. In this the bands at their extremities form outfolding loops, which unite in the centre to make a third middle band; in the centre and in each limb a ring interlaces with the bands. On the other face the bands pass through two central rings to the right and the left arms, where they split to form involved interlacing loops, and pass through the central rings to form a somewhat similar design in the upper limb. The shaft on each face is covered with interlacing, the upper half close and compact, making a plait-of-seven, the lower loose and irregular; on one face two of the bands broadly expand, and terminate in dragons' tails; the spaces in the lower portion where the plait is looser are filled in with pellets. At either side of the shaft is a dragon figure with top-knot interlaced around its body and tail knotted. The circles have plaits, on the one face of four, and on the other of five. All the bands are double-beaded, and the broader ones bordered. The splitting of the bands and the free use of pellets and the form of the dragons with their pear-shaped eyes pointing forwards, their top-knots, and barbed tails, are characteristically Scandinavian.

FIG. 32.-RING DESIGNS.
(I, 5, 6) Maughold 72; (2, 4, 13) Andreas 102; (3, 11) Andreas 103; (7) Michael 74; (8) Braddan 110; (9) Jurby 99 ; (10, 14) Bride 97 ; (12) Michael 104.

A fragment from Mr. Wallace's collection, Michael 90, so closely resembles the last, both in execution and design, that it must almost certainly be by the same hand. It shows the head of a similar dragon on a larger scale, apparently in the same position at the side of the cross; above, as though forming a portion of the circle connecting the limbs, is an elaborate piece of plait-work, the bands, which are double-beaded, running vertically and horizontally, instead of diagonally, which is exceptional.

FIG. 33. TRIQUETRA DESIGNS.
(1) Bride 97 ; (2, 6) Michael 101 ; (3, 4) Michael 105 ; (5) Michael 104 ; (7) Braddan 109 ; !(8) Andreas 103.

Quite different in style and workmanship is the small Maughold piece 91. The cross, type 4, on each face is equal-limbed, the arms connected by a true circle, but a panel is added below as a shaft. On one face this bears ring-chain pattern engraved ; the square ends of the limbs contain irregular plait-work with little spiral flourishes. The circle has plaits-of-four and of three, two ending in the heads of serpents with sharp snouts and round eyes. The other face shows on the shaft, which is bordered with a cable moulding, a plait-of-eight; the limbs bear remains of plait-work, and the circle plain step-pattern.

Very different, again, is the beautiful pillar cross at Braddan 108. Here, for the first time, the stone itself is cruciform. On each face the design upon the head follows Gaut's

FIG. 34,--STEP AND KEY PATTERNS.
(1) Braddan 109; (2) Braddan 108; (3) Maughold 91 ; (4, 7, 8, 11, 12) Maughold 72; (5) Andreas 103 ; (6) Michael 105 ; (9, 10) Bride 97; (13) Ballaugh 77 ; (14) Jurby 78; (15) Braddan 86; (16) Andreas 73.

treatment, the limbs in one case showing interlaced diagonal rings, in the other similar rings continued, with a fold at right angles to form a loop device, the lower limb ending in a triquetra knot with loose tendril at its upper part. The fust is occupied by four pelleted dragons, each showing one fore-leg and hind-leg, the heads lizard-shaped, the top-knots and tails elaborately interlaced. The rather feeble and discontinuous interlacing is quite Scandinavian in character, the loose ends of the bands notched and ornamented with small spirals. A band of step-pattern on one face separates the cross from the fust or pillar. One of the edges is occupied by a single dragon with similar head, but without limbs, which forms, with a double-beaded band, a sort of link-twist. The four corners of the stone are finished with cable moulding-our only example of this treatment.

The broken pillar, Braddan 109, is very similar to the last, and must have been carved by the same hand. The outline of the stone differs only in proportions, the fust expanding more rapidly; as the head is lost, we can only surmise that it was probably something like the last. The general treatment, however, is rather different, a distinctly panel arrange- ment being favoured. The pelleted dragons on one face also show slight differences in shape and proportions. What remains of the other face is occupied by geometrical designs. The panels are separated by bands of step-pattern, scarcely to be distinguished from angular twist; one shows plait-of-seven, the other two double triquetra knots with spiral flourishes. One of the edges has step-pattern between raised flat borders; the dragon face is bordered by flat cable moulding.

GEOMETRICAL PATTERNS, WITH FIGURES

A broken slab, Bride 92, shows on each face the shaft of a cross. One has on the shaft the ring-chain pattern double-beaded. The space to the left, a development from the link twist (see fig. 28, 2), where two bands converging and diverging are connected by double links above and below; the effect is marred by little side.flourishes at intervals. The space to the right had tendril pattern, double-beaded. The other face has on the shaft a plait-of- five; the space to the left a double twist and ring, of which the drawing is irregular, the lower band passing into the ring which is the third from the bottom, where the upper one should. Above this, the pattern breaks into the key-fret. The space to the right shows the feet of some beast; below, an elaborate knot formed from the link-twist (fig. 28) ; below, an animal, crested, with pointed head and straight thick tail, the feet very distinctly two-toed.

The fine slab at Maughold 72 differs in character from any other in the Island. The form of the crosses, with shaft and head of almost equal length and the arms very short, is peculiar. On one face the cross occupies nearly the whole length of the slab. In the centre is a ring of step-pattern, at either side of which is the figure of a cock; on the upper limb, the figure of a man with hands upraised, and a staff by his side; on the shaft, the Virgin and Child, with nimbus, now badly worn, but evidently of the form seen on Michael 100. The recesses between the limbs and circle are occupied by triple divergent spirals, double spirals, and linked rings. At either side of the head is a design of angular key pattern, which in the Island only occurs on one other stone, Bride 97, and, at the sides of the shaft, key, and infolded ring designs. The other face is divided into two panels. The cross in the upper one is plain (fig. 25, 3) ; at the sides, twist, double-beaded step, plait-of-three, and an irregular square loop interlacing; the panels enclosed by a cable moulding border. A lower panel is separated by double spirals and key designs, and contains rude scroll work, key, and irregular plait-of-three, with a triquetra, and figures of stag, hind, hound, and man on horseback. One edge has angular key-fret, diagonal rings, spiral designs, and plait-of-three; the other a curious design of curved lines and cup hollows, suggesting a twist; also a square four-fold ring and plait-of- three. (For fuller account of the figures see Description of the Monuments, Maughold 92.)

A fragment from Jurby 109 shows remains of the head of a cross with circle, both faces displaying Gaut's treatment. On one the circle is plain, on the other it has had a loose plait. Below are remains of some loop-form of interlacing; above the circle figures of a stag and a boar. The other face shows a double triquetra knot, with flourishes and pellets.

FIG. 35.

TWIST-(1) Maughold 72; (2) Andreas 73; (3) Michael 75
DOUBLE Twist-(4) Jurby 99.
TWIST AND RING-(5) Michael 74; (6) Bride 97; (7) Andreas 103; (8) Michael 105.

The large Joalf cross-slab in front of Michael Church gates, 105, is remarkable for the number and variety of animals figured. They do not appear to have any other than a decorative purpose. The cross on each face is of the usual form, but the limbs are connected by a true circle, and the shaft is rather long, in one case being finished off with volutes above and below; in the other, the interlaced design throws out two large volutes at the bottom of the shaft. The whole surface of the cross on the first face is covered with close and regular plait-work, with four breaks on the shaft, which reduces the monotony and lightens the design, giving an appearance of panels, yet avoiding the stiffness of more pronounced divisions. In the execution we notice the use of the drill, which is very conspicuous in the work of the Beckermet cross, Cum- berland'; we meet with a similar use of drill-holes on a fragment from Great Farne and other Anglian pieces (see Canon Greenwell's Catalogue of Sculp- tured and Inscribed Stones in the Cathedral, Durham). It occurs on Greek and Byzantine work. The four corners left between the circle and outline of the stone are occupied with triquetras, which appear again in circular designs of four midway at either side of the shaft, besides forming a base to two decorative crosslets.1 Immediately below the cross are two pelleted dragons, each having one fore-leg and hind-leg, with interlacement, reminding one of the Braddan Crosses 108, 10g. Below is a band of step-pattern. The figures on this face are two stags with hounds at their backs, and a man on horseback followed by a spare horse. Below the band of step-pattern are a stag and hound, much worn.

The other face has the head of the cross with Gaut's design, the shaft with double twist and diamond-shaped ring. Above the head is a stag, and at its back a small bird pursued by a larger one. At the sides are a man on horseback, two riderless horses, and faint trace of the figure of a man; bull, ram, a conventional figure of a bird, and a horse. One edge bears the well-drawn figure of a cock, the space below which is occupied by a single twist and diamond-shaped ring design, with a curious break about two-thirds of the way down, formed by the omission of two of the rings, the bands converging, but not crossing, and connected below by a link. The other edge has, above the inscription, the figure of a warrior armed with spear and shield. The carving is bold and free; the bands and rings are decorated with pellets or double-beaded, and on the edge with small flourishes, notches, and spirals.

A very small fragment from Ballagilley, Maughold 106, shows the fore part of a boar, very well drawn. This must have been one of a series of figures at one side of the shaft of a cross, as on the last. It is in rather higher relief and on a larger scale. The monument must have been a very handsome one, and one wonders how the rest of it can have been so totally lost.

The rectangular Sandulf slab, Andreas 103, is also remarkable for its animal figures. The orm of cross on each face, type 2, is long and narrow (see fig. 25, p. 39), the upper limbs are bare of ornament, and there is no surrounding circle. On one face the shaft bears rather angular plait-of-five, terminated at the top by two of the bands interlacing in a circular ring. The shaft has a narrow border of key-fret and step-pattern. Above, a horizontal band shows plait-of-four, and, below, another band (bordered like the shaft) has twist-and-ring double- beaded. On the other face the shaft bears the ring-chain pattern, terminated at the top by a small Latin cross; above the head, a loose sort of triquetra loop design. Below the cross is a vertical band of tendril pattern. On one face, two rectangular interlaced rings appear at the back of a man on horseback, and on the other a square-folded ring and ring-and-buckle design. The animals appear to be, on one face, a cock and some other bird above the arm of the cross; a stag pursued by hounds, man on horseback armed with a club, boar, and long-horned ox; goat, wolf, hound, hind, and bear. The other face has also above the arms a cock and some other bird with a collar. (See, as to this design of the birds, under Descrip- tion of the Monuments, Maughold 92.) Below are stag, knotted serpent, wolf, calf, wolf and hound, or it may be another wolf; goat, ram, ox, and boar. Below the cross is a robed figure on horseback, and lower down a faintly incised hound.

The broken head of a slab, Michael ioi, shows Gaut's decorative treatment on one face. The bands are pelleted, but variety and artistic effect are given by leaving all within the central ring plain, and by having the loop terminations in the limbs double-beaded; the foliation in the upper limb is more advanced than in any of our other pieces. Above the left quarter of the circle is a pelleted dragon with one fore-paw and top-knot like those on the Joalf cross 105, and the Braddan pillar crosses 108, 109. The corresponding space above the right quarter is occupied by the figure of a man attacked by an eagle-possibly intended t0 suggest a hero fallen in battle, and the great eagle, Hræ-svelgr, carrion swallower. A somewhat similar scene occurs on an early Celtic cross at Aberlemno, Forfarshire, where, among figures of warriors on horse and foot, is one of a man staggering backwards with a bird of prey pecking at his breast.' As, however, ours is a Scandinavian piece, we may expect a motive from the Norse history, folk-lore, or mythology.

The other face also shows Gaut's treatment, still more floreated, and with a certain angularity and want of gracefulness; the bands split to form diagonal rings at the ends which cross and interlace with a circular ring; and owing to the central figure, the bands do not come up from the shaft, and in the upper limb stop short in a rather ugly angle. In the centre is a numbed figure in plaited garment, his arms outstretched in the ancient attitude of prayer or blessing; below his feet are five large pellets, which may be intended t0 suggest clouds, the figure probably representing Christ in ascension. The curious form of nimbus terminating in fringes, and decorated with three crosslets, appears to be peculiar to the Isle of Man (see p. 40).

Above the left arm is the well-drawn figure of a cock, symbol of the Resurrection. The branch on which it rests finishes in a triquetra knot; on the other side, a winged figure of an angel-possibly intended for the Holy Ghost-below which we again see the triquetra.

The fragment from Jurby 98, showing the base of a cross shaft, has on one face the ring-chain pattern, the V-shaped tongue bordered by a fine line. The fore-feet of some animal are seen above a curious design of flat pellets within a border of step-pattern. The other face has on the shaft double twist-and-ring, the inner band of the twist and the rings being pelleted ; on one side are very rudely-drawn figures of a boar and a hart; on the other a robed figure with a trident, and, below, a bird-headed figure with a pole over his shoulders, from the end of which, suspended by a rope, hangs a smaller figure. This would appear to illustrate some scene from the Norse mythology. Thus Odin, the " Eagle-headed," who in the character of Bikke, the evil counsellor, caused Randver's death by hanging; the robed figure above, possibly the aged King Jormanrek, Randver's father. In the animals we may recognise the boar, Saerimner, which affords sport to the champions in Valhalla; and the sacred hart, Eikthyrnir, " that stands o'er Odin's hall, and bites from Laerad's branches."

In the Rumund cross, Michael 100, we have the double twist-and-ring design-the outer bands and rings pelleted ; tendril pattern, double-beaded, and twist-and-ring. On the other face the shaft has the ring-chain design, double-beaded. The spaces at the sides are occupied by figures. Below the circle, on the left, a man, robed, under his arm a book, his hands clasping a crutch-shaped staff ; over his head the peculiar Manx form of nimbus, with fringed ends, decorated with three crosslets. Undoubtedly this is a Christian subject, but the other figures appear to illustrate the heathen idea of the future life; thus the two bird-headed warriors, one head downwards to show that he is dead, may be the champions in Valhalla, who spend their days in hunting and fighting, their nights in feasting. On the other side of the shaft we see Odin with his spear, Gungnir (Tusker) and his two wolves, Geri and Freka. The broken head of a slab, Jurby 99, shows on one face a double twist design; in the upper limb the bands terminate in diagonal oval rings interlaced. This double twist (No. 519 of Mr. Romilly Allen's Analysis) is rare; we find it on a stone at Rothesay, and once at Clonmacnois. The other face has the tendril pattern on the shaft. The terminations in the limbs show a further variation. In the upper one, the two bands connect by an upward loop, then pass on to form a downward loop, the whole presenting the figure of a four-fold ring, with which a diamond-shaped ring is interlaced. The terminations in the arms have no interlaced ring, and though the idea is the same as in the upper limb, the expression of it is more compact, the two loops overlap, and the result is a different design, resolving into a double triquetra knot. both faces have the bands decorated with diagonal scores between fine border lines, giving the appearance of pellets, varied in this by the tendril and the terminations being double-beaded. Some scoring in the loops displays an approach to foliation.

To the left of the shaft is a plain key-fret, the space to the right being occupied by an artistic knot, formed by combining the terminal designs of the two arms, thus making two double triquetras. The other face has the inscription on the left, and on the right a more elaborate knot, which again is a development of the designs in the arms-two sets of diagonal oval rings interlaced with flourishes. Above is the figure of a man in a tunic, blowing an Alpine horn, doubtless intended to represent Heimdall, Warder of the Gods, who at Ragnarok will summon them to the last great battle by a blast on the Giallar horn, which rings through all the nine worlds. Cumming's cast, taken before the top of the stone was broken, shows a flying raven above his head. The long-robed, dog-faced figure on the other face may be the sibyl Hyndla, whose prophecies foretell the fate of Heimdall and the dreadful days of Ragnarok. The fine monument at Bride 97 is truly remarkable for the variety of geometrical designs, as well as for the figure drawing and mythological references. The slab is round-headed, the cross on either face an equal-limbed one, and the spaces between the limbs and surrounding circle-a true one-is pierced (see fig. 25). It shows throughout a panel arrangement suggestive of Celtic models. The crushed angular key-fret which we only have elsewhere on the Maughold slab 72 is freely used. The circle on one face shows twist-and-ring breaking into key-fret, and then into a plait-of-four, with a double spiral and a small triquetra between the beginning and the end; on the other face, twist-and-ring breaks into a plait-of- three; the remaining third is unfortunately chipped off. Three limbs of one face are occupied by the angular key, the fourth by a plait-of-four. The other face has angular key-fret on one limb, and on the other three irregular designs of rings and rods, while the centre is occupied by interlaced diagonal rings. On this face we have, above the circle, a panel of angular key, at either side of which is the figure of a cock and a triquetra. Below is a panel with two double spirals face to face, from which a narrow band of ring-chain pattern runs down the middle. The slab itself has had a double border of step-pattern and of cable work, the former terminating on the right in the head of a serpent. The rest of this face is occupied by figures, by panels of crushed key-fret, and a design of oval interlaced rings. The other face has, above the head, a panel of angular key-fret framed by cable-border, above which is a chevron and arched cable, and at either side the figure of a dwarf and a plain triangle. Below the circle is a panel of angular key-fret bordered by cable-moulding, from which a band of step-pattern runs down the middle. A panel of angular key on the left terminates below in a large volute, with a triquetra to fill up the space on one side ; above are step-pattern and two triquetras ; to the right is a panel with plait-of-nine, double-beaded, and just below the circle a double spiral. Two figure-of-eight knots within a circular ring occupy the spaces between the arms of a human figure.

Of figures we have on one face a bearded man, belted, running with some large object in his hand. Below is a great bird; below, again, a stag with a hound at its back and another under its feet; and below that, separated by a panel of key-fret, two large hounds, back to back. To the left is a bird, a man with spear attacking a stag, another man beneath the stag armed with a knife; at the back of the stag a hound, and another behind it. The other face shows on the right a man with spear, and, below, a panel with a horse, or possibly a stag. As the head is broken off, it is uncertain. This animal is trampling human bodies underfoot.' To the left is a man, belted, attacking a serpent with a knife; below,, a giant and a dwarf, with some curious square figure above the shoulder of the giant. There are also above the circle on one face the two birds referred to above, on the other the two dwarfs.

Though some of these figures are no doubt merely decorative, others appear to illustrate stories from the Norse mythology. The man attacking the serpent is, I think, intended for Thor, who is to meet the dread Midgardsorm at Ragnarok. The corresponding figure with a spear may be Odin. The monstrous figure with arms akimbo and legs outstretched may well represent the giant Rungni, who challenged Thor to combat, and stood upon his shield to meet him; and the little figure at its shoulder, Main, the three-nights-old son of Thor, who released his father from the weight of the fallen giant. Then we have the dwarfs upholding the lift, the two great wolves of the Eclipse, and, I think, other scenes.

The Mal Lumkun cross, Michael 104, is shown by the inscription to be the work of a Swedish, not a Norwegian artist, as most of our Scandinavian pieces are. It is highly decorated with Celtic designs arranged in panels, the decorative treatment and the workman ship being very delicate and artistic. The limbs of the cross, which is of the usual form I (see fig. 25), are connected by a true circle, plain, but with cable borders. The cross also has a border, which on the shaft takes the form of cable-moulding, terminating below in the head and tail of a serpent. In the centre is a ring of cable pattern, surrounded by the ring-chain. Each limb is occupied by a device of four triquetras, double-beaded, continuous, between which and the circle is a design of two diagonal oval rings interlaced with a diamond-shaped one. The shaft has a plait-of-nine, with a horizontal break in the middle; below, a plait-of-seven. The space to the left shows a stag with a hound at its back; below, a robed figure with tau-headed staff. On the right, just below the circle, are the remains of a stag and hound; below, a harper seated and playing on a four-stringed harp, to whom a long-robed female is offering a drinking horn; below, a robed figure corresponding to that on the other side. The Celtic and the Christian character of these figures is evident, except in the case of the harper and the figure offering to him, which must be taken together, and may have a mythological reference; the idea may have been suggested to our artist by carvings in Gotland, Sweden, as on the Tjangvide and the Habbingbo stones, on which robed figures are seen offering mead to Odin. Below the sculpturing is an Ogam alphabet.

The other face of the stone is not sculptured, but bears two inscriptions in runes which differ from those used on our other monuments. In the middle are the worn remains of another inscription scratched in Ogams.

The small Thorwold cross 102 is of the Andreas type (see fig. 25 ). The shaft of each face bears the ring-chain design, terminating above in a small cross with peculiar lyre-shaped horns at either side. The square, looped, or four-fold ring design is seen in the hollow recess above the remaining arm of one face; it is repeated on the arm, but this looks as if added later. Above the arm is another device which at first sight looks like the square four-fold ring, with two other rings crossing and interlaced; it proves on examination, however, to be a ring so folded on itself as to form seven external loops. Of figures, we have on the one face Odin, nude or in tight-fitting garments, with his spear attacking the Fenris wolf, his raven on his shoulder, knotted serpents all around; on the other face is a man with cross and book, knotted serpents and a fish, which in this case I think must be the well-known Christian symbol.

A large slab at Jurby 93 shows variation on Gaut's treatment of the head-the two bands, instead of passing direct to the upper limb, turning to the right and the left arm, and so from the arms to the upper limb. The shaft bears the double twist and diamond-shaped ring design, which displays several irregularities in the drawing. The space by the side of the

cross gives an illustration of Sigurd Fafni's bane. The great dragon is well drawn, decorated with scores which remind one of the wood carvings in Norway. Sigurd, with his sword, is seen in the pit. Below, Sigurd is seen again in conical helmet and long robe, in his right hand a staff, on which, no doubt, was spitted the dragon's heart over flames of fire, now worn away; below, his steed, rather feebly drawn, and a curiously conventional figure of a tree, at one side of which is one of the talking birds. The other face is almost worn away, but shows, above the cross and circle, a badly-drawn figure of a cock. The head of the cross appears to have been decorated in a novel manner with a band of ring-chain design across the arms, possibly crossed by another from the shaft to the upper limb.

The broken Malew slab 94 has on the shaft an effective design of link-twist, the two bands forming the upper portion of which meet about the middle two other bands, which form the lower half, in a figure-of-eight knot. Besides the link form 5 (fig. 27), repeated four times, we meet with a further development, twice repeated, in which the off-shoot springs from the inside of one band to pass downwards, loop round upon itself, and join the inside of the other band. In the lower half the two links are separated by the bands splitting to form two diagonal oval rings interlaced. The spaces at the sides of the shaft are divided into panels. Above, on the left, are the remains of a well-drawn figure of a horse ; on the right, Sigurd in tunic, pointed cap, and with sword by his side, roasting the dragon's heart over the flames of fire; below, Sigurd slaying Fafni, looks like a feeble copy from memory of the same scene on the last.

The other face shows on the shaft the figure of a dragon, with fore-paw and top-knot interlaced. At one side of the shaft is an irregular loop form of plait, characteristically Scandinavian.

FIG. 36.-ZOOMORPHIC INTERLACING AND DRAGON-FIGURES.
(1, 2) Maughold 91; (3, 6) Andreas 95; (4) Bride 97; (5) Michael 100; (7) Michael 89; (8) Michael 90; (9, II) Malew 94; (10) Jurby 93; (22) Michael 101; (23) Michael tos; (24, 26) Braddan 108; (25) Braddan 109.

The broken piece at Andreas 95 shows some affinity to the last in respect of the dragon figures with long top-knots, and in the interlacing, which, however, is purely zoomorphic. Sigurd is shown in profile, in helmet and tunic, roasting the heart of the dragon, and placing his fingers in his mouth to cool. The heart is represented by three rings over three pointed triangles representing flames of fire arising from the hearthstone. His steed and the talking bird appear at his back. Below, Sigurd, helmed, appears again in the act of piercing the dragon with his short sword. The shaft is occupied by three dragons, with top-knots interlaced, and scored in such a manner as to suggest a woodwork prototype. The other face shows Gunnar in the worm-pit in tunic and pointed cap or helmet, his wrists and ankles fettered in gyves; knotted serpents around him, one biting his shoulder. On both faces the bands are double- beaded or bordered. The drawing of the figures, the general treatment, and the execution, show this to be the work of a different artist from the last.

FIG. 37.-BIRDS, FISH, AND SERPENTS.

Birds-On the Cross two birds with ring, representing Chaplet, (1) Maughold 72 ; birds above limbs of cross, (2, 3) Andreas roa ; birds above cross, (4) Bride 97 ; cock, (5, 6) Michael 101, 105 ; large bird chasing small one, (7) Michael 105 ; nondescript bird, (8) Michael 105 ; eagle, (9) Bride 97, (1 1) Michael i01 ; raven, (10) Andreas 102 ; (12) Bride 97; (13) Jurby 99•

Fish-As Christian Symbol; (14) Andreas 102.

Serpents-(15, 17, 20) Andreas 102 ; (16, 19) Andreas 103 ; (18) Andreas 95

The Sigurd fragment found by me in 1902 in the town of Ramsey, Maughold 96, is later work by yet another hand. One face shows a beautiful piece of plait of a broad band in the form of figure-of-eight, with narrow bands closely interlaced, and a free use of pellets. This combination of broad and narrow bands is very exceptional in the Isle of Man; we meet with it in the Scandinavian cross-slab, Michael 89, and to some extent in the Andreas Sigurd piece 95. The interlacing on the other face is loose and irregular in character, and reminds one of the dragon-face of the Malew Sigurd piece 94, while the ring and rod as a decorative design just above the salmon might have been suggested by the figure of the roasted heart, as seen on the Malew and Andreas pieces. The interlacing terminates in a large incised volute or double spiral. There is a figure of a man crouching and throwing stones, evidently Loki; beside him the Otter devouring the Salmon. This, which is the beginning of the story, has not been figured elsewhere. Above stands the steed Grani, with the chest containing the gold hoard on his back.

Finally, we have a Scripture scene in the Temptation of Adam and Eve, Bride 116. This has not been part of a cross, but, I think, an architectural feature of the twelfth century built into the early parish church. It belongs to the series of subjects found in the Catacombs and the early sculptured Sarcophagi. The same scene occurs on about twelve Irish crosses and two Scottish; we find it on a stone at Dacre,l Cumberland, which Mr. Collingwood and the late Mr. Calverley assigned to the tenth century. Mr. Romilly Allen points out2 that in Norman sculpture Adam and Eve are represented as receiving or presenting the apple with one hand, and hiding their nakedness with the other. In our example, Eve holds one hand with the apple to her mouth, and with the other presents an apple to Adam, who stretches out one hand to receive it, hiding his nakedness with the other. A remarkable thing is that in this, instance there is no figure of the Serpent. " The woman tempted me," is perhaps the particular text the sculptor sought to illustrate.

So in the latest as in the earliest of our decorated pieces, whilst we recognise the character of artistic treatment general in the British Isles throughout the period to which many of ou- Manx monuments belong, we see a certain originality, an individuality which distinguishes these from the monuments of the same class in the surrounding lands. This style also obtained much later in the Isle of Man than elsewhere, finally giving place to English art of the thirteenth century.

This exhausts our examples of Celtic art and the Scandinavian development of it in the Isle of Man. If ever we had illuminated manuscripts, or metal-work, such as shrines3 or reliquaries, they have long since been carried away or destroyed; and the total absence of such decorative work in architecture, secular or ecclesiastical, is to be accounted for by the poverty of the land and the unsuitability of material in the Island as well as by the unsettled condition of its people

.1 Crosses, etc., Diocese of Carlisle, p. 38.

1 The triquetra, which appears on sixteen monuments in England (of which seven are in Cornwall), is met with also on nine in Scotland and six in Wales, but on two only in Ireland. In the Isle of Man it occurs singly on four Scandinavian and three of our earlier pieces, in a design of four continuous, on two Scandinavian and one Celtic, and in form of knots on two Celtic and eight Scandinavian pieces-in all on twenty monuments, fourteen of which are Scandinavian sometimes four and even five times repeated.

1 Early Christian Monuments of Stolland, p. 211, fig. 227b.

1 In Vikingerne, p. 227, Dr. Alexander Bugge figures a stone from Ardre Kirke, Gotland, showing two animals con- ventionally treated with very involved interlacings trampling a human form underfoot. He regards the animals as stags, as he does those upon our Bride Stone, to which he refers. I think, however, that in both cases they are meant for horses, and may refer to the death of Swanhild. (See further under the Description of the Monuments, Bride 97.)

1 Crosses, etc., in the Diocese of Carlisle, p. 114-

Christian Symbolism, p. 190.

3 According to the Annals of Ulster, the Shrine of St. Dachonna on Inis Patrick, Peel Island, was broken by the Gentiles in 798.-A. W. Moore, Sodor and Man, p. 25. We read in our Lex Scripta of the "three reliques of Man," but of what they consisted we do not know. See p. 30, footnote.

SUMMARY

We find, then, that the Scandinavian funeral monuments in the Isle of Man, both as regards the character of the memorial and the mode of decorative treatment, are a continuation of the Celtic system introduced with "the Christian religion five centuries previously. Though from Gaut's statement that he made " all in Man," it is evident that he, the first of our Scandi- navian sculptors, was not aware of, and therefore owed nothing, to the earlier Manx artists, it is likely that he had seen some of the carved stones in the North of England, and perhaps in Scotland, whence he may have derived the use of the twist-and-ring, a favourite design in his work, while the ring-chain may have been suggested by the Gosforth cross. Yet his peculiar application of this design, his development of different forms of link-twist and tendril patterns, and in particular his decorative treatment of the head of the cross, exhibit originality and keen artistic feeling. It was evidently to Gaut that his successors owed the ideas of these designs, which became favourite and characteristic of the Manx monuments. The double twist and diamond-ring was probably suggested by Northumbrian work, while the free use of the triquetra as well as other local peculiarities support my view that the geometrical designs on our Scandinavian pieces are for the most part derived directly from early Celtic illuminated MSS. The origin of the figure-drawing is more difficult to determine. It is not from the MSS., though some details, such as the peculiar Manx nimbus on one or two pieces, seem to be so derived. On the Welsh monuments figure drawing is infrequent and very rude. The Irish slabs have none, figures of birds and animals not appearing except on the later high crosses, to the art of which there is no indication that our Manx pieces owe anything. In Scandinavia the drawing is very inferior and rare, while only two or three exceptional pieces are in relief. Nor do I see much resemblance to the figure-drawing on the Viking pieces in the Lake District, which otherwise have much in common with the Manx. Some of the animal forms seem rather to have affinities with the carvings in East Scotland, while many appear to be the original work of local artists derived from direct observation of nature.

As regards execution and workmanship, one notices as a characteristic less regard for accuracy of detail and careful finish, less formality and stiffness, a bolder treatment, and a greater freedom than in our earlier pieces.

The following table, showing the distribution and use of the decorative designs, will be interesting also for comparison with that of the Celtic pieces (p. 35) :-

TABLE OF DESIGNS

...