painted in various colours, and therefore the people of the age must have possessed a damp-resisting vehicle for pigment. We know also that their ornament in MSS. was elaborately coloured, and similar ornament in stone would hardly have seemed to them complete without similar treatment. In Scandinavian museums there are examples of the colour-treatment of interlaced patterns in wood, apparently traditional from early times. This suggests a richness quite in accordance with all we know of the tastes of Anglo-Saxons and Danes before they learnt Renaissance - Classic theories of uncoloured sculpture. This also explains the slight relief and sketchy execution of much work which could never have "carried" or "told" in the open air without colour to discriminate its flat surfaces and shallow ground; and it accounts for the lavish ornament spent on poor material.

TOOLING.

It is not always possible to distinguish the tool-marks; the stones are sometimes too far gone in decay. But for the most part it is fairly easy to see whether the work were good sculptor's carving with the chisel, or merely rude chipping of a pattern with the hammer. In certain cases it is evident that a chisel-dressed stone was finished with hacked or picked work, as though a rough outline were preferred, especially when the surfaces were not rounded and smoothed in the style of the finer carvings. This finishing with the pick is sometimes mistaken for weathering; but no one who has a good personal knowledge of the stones can doubt that the pick was freely used, and to intentional effect. The dotted line of some MS. illuminations (e.g. Bishop More's Prayer-book, dated by Westwood eighth century) is imitated by lines of pick-marks (e.g. Amotherby a), and perhaps may have suggested the trick, which is found on a cross with Anglian head and Anglian runes from St. Ninian's Cave (Wigtownshire), showing that it was known in pre-Danish days. But in classing the stones according to their tool-marks, we shall find that picked or hacked work is generally associated with Viking Age pattern, while fine chiselling is found on monuments which have beautifully designed leaf-scrolls and well drawn figures.

The fine chiselling used to give finish with "surface," not merely to outline a form and sink the ground, occurs in the Brompton "cock" shaft, Easby, Hackness, Hovingham, Kirkby Hill impost, Kirkby Misperton b, Kirkby Moorside impost, Kirkdale "Ethelwald" slab, Kirklevington "bird" shaft, Lastingham dragon head, crossheads a and g, and jambs, Masham crosshead a (the pillar being too weathered to show cutting), Melsonby slabs, Northallerton a, d and g, Stanwick a

and b, West Witton slab, Wensley slabs and c, Wycliffe jamb and shaft b. All of these are of the finer style in design, as well as of good cutting.

The drill has been used with the chisel in Kirkby Hill, Pickhill c, Wath d, and in the Great Edston dial.

Rough or second-rate chiselling, in which no attempt at "surface" is made (the effect being flat), or in which no attempt at finish is made (the effect being rough), is seen in the Bedale pillar, Brompton hogbacks, Crathorne slabs, Forcett stones, Gilling, Great Ayton, Kirklevington "Hart," free-armed head and saint with wands, Northallerton second head and shaft p, North Otterington c, d, g and i, Osmotherley a, g and d, Stainton, Stanwick 7, m and o, Stonegrave 1, Topcliffe head, Wath a, West Witton b, Wensley i and Wycliffe bear. These are not so refined in design; some are evidently of the Danish period, as we shall see later.

The chisel and pick are used together in Birkby, Brompton porch cross, Easington dragon shaft, Kirklevington wheel-cross and shaft v (with Danish loop-pattern), North Otterington a and k, Osmotherley hogback and Stanwick "Hart and Wolf" cross. These are all of a quite different class of design from the finely chiselled stones.

Shaped with the chisel, perhaps, but certainly worked with the pick, are Amotherby cross-heads, Bedale shrine-tomb, Brompton porch hogback, Crathorne dragon shaft, Ellerburn stones, Easington hogbacks and heads (the pick work is very clever on Easington h), Finghall heads, Kirkby Moorside shafts, Kirkdale band, Kirklevington c, s and %, Levisham slab, Middleton a and e, Northallerton wheel-heads, Thornton Steward, and Wycliffe hogback. The ruder character of the design is as evident as the rude cutting, in spite of vigour and occasional decorative effect.

Some stones may have been simply worked with the hammer or pick, being nothing more than rough blocks or slabs with a little shallow ornament hacked on them, and then perhaps painted. Such are Croft e, Hauxwell lintel, Helmsley hogback, Kirkby Hill and ‹, Middleham dragon, Ormesby slabs, Pickhill a and d, Sinnington h, k, n and o, Stanwick k and n, Stonegrave e, f, h and j, Wath b, and Wycliffe ring-plait shaft. Most of these hacked stones bear pattern which is characteristic of the Viking Age.

We have already set up a very definite classification. The finely chiselled stones are also finely designed. The roughly hacked stones rank with Danish motives. It was in the earlier Anglian time that, as we know, stonemasons were brought from abroad to teach English artificers; and the Danes were a people who destroyed churches and

lived in wooden houses, not being stonemasons themselves until they learnt the art from the English. Let us now see what result can be got from a different kind of division, classing the monuments by their types of form.

CROSS-HEADS: (1) FREE ARMS.

Type A. The simplest free-standing head was that in which the form of the whole was contained in a circle, the arms being shaped by cutting out four parabolic spaces. Crathorne h, though the arms are not cut free, may serve as a diagram. The upper cross in Spennithorne is of this type, with all outlines curved. Mr. Romilly Allen notes the form as characteristically Northumbrian; only in the Spennithorne slab the topmost arm is abnormally large, creating the "Hammerhead" variety, common in Cumberland but not in Yorkshire. Finghall a and Kirklevington gh, had they been unmutilated, would have been better examples. Crosses of the Anglian bishopric at Whithorn are of this Type A, and bear Anglian runes, fixing their date.

Type B. The lower cross in Spennithorne has the ends of the lateral arms flattened to fit the frame. This is carried further in Wensley a and b, West Witton a, Middleton g, and Stainton b. This form would be easier to cut than A, and give a rather bigger crosshead from a stone of the same size. Many heads approach this form, having their arm-ends only slightly curved, as Easington b, Gilling a, Kildale a, Kirkdale b, Kirkby Hill f, Northallerton m, and Sinnington d; but Northallerton a is a good example of Type B in perfection. Already there is a tendency to bevel off the acute angles, to facilitate cutting and to prevent chipping, as in Great Ayton cd.

The "Maltese" cross, seen on slabs at Crathorne b and c, was not used for free heads until late, as Welbury d. The rude forms of Forcett a, Kirkby Hill a and Welbury c appear to be unskilled attempts at Type B, though they are seen in Durham and Cumberland also.

Type C. From B, but only in work which shows great care and a high aim, the spatula, created at the ends of the arms by bevelling the acute angles, was lengthened. The curve of the armpits was still kept open and parabolic, and the outline of the arm ends was naturally and rightly tapered, sometimes with a little concavity. This makes the graceful form of Lastingham a and g and Masham a, crosses of the best workmanship. The form must have been so risky and difficult to cut that the arms were shortened and the armpits contracted when the carvers became lazier or less skilful.

Type D. The easiest step would be to shorten the lateral arms, so that the topmost arm remained long; thus a tall head could be got out of a narrower stone-a great advantage to the mason, as Amotherby b, Sinnington and Wath a.

The final form has small circular armpits and broad equal armends, straight at the extremities but sometimes bevelled at the sides, as in Kirkdale a, Osmotherley g, Great Ayton a, Easington c.

Now the Keills, or West Highland type, is this Type D cut from a broad slab, with small circular armpits and big oblong arm-ends; a late development out of these North English crosses. It must not be forgotten that the crosses now seen at Iona are hundreds of years later than the time of St. Columba, or even of the Anglian church.

CROSS-HEADS: (2) WHEEL-HEADS.

By the addition of a ring, wheel-heads were, formed from the above types, excepting C, which did not lend itself to that kind of elaboration. If simpler forms are earlier than complex, then free heads are earlier than wheel-heads, and the Northumbrian type is the parent of all crosses.

Some of the North Riding wheel-heads have the usual piercing of a four-holed cross, but some are not pierced.

Kirklevington a
Middleton e
Kirkby Moorside a
Thornton Steward a
Thornton Steward b

is reduced to lowest

From forms like Brompton e, where the disc terms, may have come Kirklevington and Middleton a, where the quadrants of the circle are replaced by four cylinders; evidently a late invention.

Now comparing these lists with the plates, it will be seen that in nearly every case free arms are associated with more symmetric and correct design and the fine style of carving; whereas the wheel heads nearly always are ornamented with the looser design and ruder cutting which we have seen to characterise the Viking Age. Sometimes the styles overlap, as at Gilling, where there is both a free head and a wheel-head, with the spine and boss, or lorgnette pattern, developed from a superimposed cross, as at Stainton b, Wath a, and the five bosses with rings as seen in Amotherby b. The lorgnette

occurs also at Great Ayton in a free-armed cross with well-cut crucifix, at Northallerton on a finely-cut head, and at Forcett on a rudely designed but neatly chiselled stone. A variant is seen at Stainton. The lorgnette in Cumberland is common, derived from a cross at Carlisle, which in turn is copied from Northallerton a. The pattern is therefore Anglian, surviving into the later period.

The centres of cross-heads usually bear bosses or rings, or both, unless the space is occupied by a crucifix. The raised boss or crucifix may be generally taken to indicate the front, and the flat circle the back, of a cross. Square forms are seen in cross-head centres at Brompton e,j and 7; interlaced bosses at Gilling b, Lastingham b, Masham a; rosettes at Great Ayton a (8 petals), Lastingham (24 petals), Middleton g (16 petals); rings with pellets at Kirkby Hill a, f and h; rings with pellets interlaced and lorgnette at Northallerton a and c. There is a socket to hold a gem or glass bead at Lastingham a, Middleton g, Osmotherley g, Wensley a; and a merely vacant space at Stonegrave a. These details are not without importance in classifying.

As the wheel-heads are associated with the later style of ornament and cutting, it seems that wheel crosses were later in use here than free armed crosses. They may have been introduced from Ireland, either by the natural spread of a picturesque fashion, or by the intercourse of York and Dublin in the tenth century Danish period.

SHAFTS.

Some cross-shafts are square or circular in section; others are thin for their breadth, like the stopes used for gate-posts; and between the two extremes there are various forms. It is obvious that skilled masons and quarrymen would be needed to win sound and thick stones, whereas slabs can often be picked up without much trouble. In fact, slab-shafts characterise the Viking Age crosses of the Isle of Man, Cumberland, and the shores of the Irish Sea generally, in which parts building was rude and unskilled; while thick stones occur in districts where the early civilisation included good stone architecture, or in places like Ruthwell and Bewcastle, where some special cause brought skilled masons to the work.

In the North Riding we find that the finer class of design and cutting is associated with shafts of which the section is quite or nearly square, at Cundall, Hackness, Northallerton d and Wycliffe ; the Masham shaft is cylindrical. Easby and Brompton f are thick stones. Wensley and f, however, are thin for their breadth; but as they have no carving on the reverse, they may possibly be recumbent