Glancing down the columns of a convenient note-book, one observes, as famous examples of deflected buildings, Lichfield cathedral, which leans to the North, and Canterbury, Ely, and York cathedrals, which incline to the South. Other cases of Northern deflection are Brent Pelham, Hertfordshire (“several degrees North of East”); St Michael’s, Coventry; St Mary’s, York; North Curry, Somerset; St Mary’s, Bridlington, and Whitby Abbey[565].

The instances of deflections towards the South are much more numerous in this country. My list includes Bishopstone and Bosham, Sussex; St Mary’s, Coventry; Holy Trinity, Stratford; Priory church, Tynemouth; West Malling, Kent; St Andrew’s, Lammas, Norfolk (15°); Chipstead and Mickleham, Surrey, and many others. Further lists have been compiled from various sources, but, as the direction of the twist is not usually specified, it would serve no purpose to reproduce the whole catalogue.

Examples of unsymmetrical churches abound on the Continent, but the deflection of these is usually to the North. The Southern bend is not unknown, however, for it is found in the cathedral choirs of Geneva and Stuttgart. Some French chancels even, contrary to common belief, lean to the South. Speaking of the deflection of French churches, M. J. K. Huysmans, in his romance, La Cathédrale, says, “This twist in the church is to be seen almost everywhere—in St Jean at Poitiers, at Tours and at Reims[566].” M. de Caumont observed the Northerly deflection in more than a hundred churches of the thirteenth and fourteenth centuries, the examples being widely scattered[567]. The deviation, in the case of Bayonne cathedral, is at once noticed by the visitor; Mr Baring-Gould claims this irregularity as the result of English domination and English architects[568].

Returning to our own churches, we find it necessary to clear the ground somewhat. On the one hand, we have the estimate given by the able editors of Durandus, to the effect that probably about one quarter of our English churches have a deflection, and that it is usually towards the South[569]. The present writer dare not hazard the opinion that the proportion is so high, but he feels convinced that the instances are too numerous to be explainable by chance. But consider an experience of an opposite kind. About a dozen years ago, Mr G. Watson, of Penrith, stated that he had examined the plans of nearly four hundred churches as shown on the Ordnance maps, and found either no deflection at all, or, at most, a trifling variation[570]. Precisely what constitutes a trifling variation we are not told, though it may be freely admitted that the angle is often very slight. Thus, Mr W. J. Maxton and myself found that the Lady Chapel of St Saviour’s, Southwark, inclined but a degree or two East of the nave, yet surely, even this trivial deflection could have been avoided by the architect, had it been desired. It has been asserted, apparently with some reason, that York cathedral, already cited as a “leaning” edifice, has no deflection[571], and to this the reply has been given that the variation would be obvious on a good ground-plan[572]. There is no need to argue about these examples in detail, because there exist sufficient cases of undisputed deflection.

The “skew chancel” is not confined to one epoch. Yet it is submitted that we have no marked examples before A.D. 1200. The feature is most observable in churches of the time of Edward III. It is said to be doubtful whether any instance could be adduced from the Renaissance period of building; the apparent exception in the church St Mary Magdalen, Taunton, being of later date[573]. That the feature has a certain fixity is shown in the case of St Aldate, Oxford, where it exists in spite of the restoration made in 1863[574]. A late example of the “twisted” ground-plan, which is valueless as regards the present discussion, occurs in St Peter’s of the Vatican. Here we have a deflection of some feet, due to an architect’s blunder, when the Greek cross was formed into a Latin one by prolonging the nave[575].

Admitting, as we are compelled by the facts to admit, that there exist many undoubted examples of skew chancels, we next look for an interpretation. Two schools of thought, the symbolist and the rationalist, are soon encountered. The case presented by the rationalist group of authorities may be thus summarized; deflection is the result either of carelessness or of differences in the dates of building the nave and the chancel respectively. We will consider first the explanation involved in the question of dates. It is generally believed that the choir of a church was often consecrated as soon as it was completed, the consecration of the nave or vestibule being held over until that part of the building was, in turn, ready for use. For instance, as Mr Parker pointed out, the Norman choir and transepts of the earlier Westminster Abbey are known to have been consecrated in A.D. 1065, but the nave, if one were ever added, was probably not finished until the twelfth century. Separate alinements may, in such cases, have been set out, and slight errors have been made during the operation. Again, in some of the churches which show inharmonious alinements, the nave and the choir manifestly belong to different epochs, one or other portion having been rebuilt. The architect of the later period unconsciously took an orientation which deviated from the first axis, or, according to the alternative view, he was actually unable to make the axes correspond. If we assume that he employed a magnetic compass, there is a third possibility of error in the variation of that instrument. But all these hypothetical causes of miscalculation are removed if we attach weight to a simple suggestion which has been made by one or two writers. It is submitted that the axis of the nave or the choir, whichever portion was already standing, would probably be hidden from the mason’s view, through the temporary blocking up of the chancel arch. This might render harmonious alinement a task of some difficulty[576]. It has been averred that every case of deflection occurs in a church which has been partially rebuilt; hence the lack of agreement. It is difficult to say whether this statement is absolutely correct, but it seems to be true in many instances. Even so, why are the misalinements practically confined to re-buildings of the thirteenth and fourteenth centuries?

Again, it is asserted that many cases of deflection were discovered only when, in modern times, the rood-screen was, for one reason or another, taken down. In such instances, it is argued, the mason would have his standard line concealed while setting out the new section of the building. But, as we shall shortly see, this fact about the rood-screen may be interpreted in another way. If we accept the Saint’s Day theory, there remains the possibility of error due to the re-dedication, or plural dedication, of churches. The difference might then conceivably originate without any re-building of the fabric. Mr Airy cites the case of Clapham, Bedfordshire. A better known example, possibly, is found in Whitby Abbey. The building, as a whole, is dedicated to St Peter and St Hilda jointly—a fact proved by the Abbey seal. St Peter’s Day is on the 29th of June, and St Hilda’s on the 25th of August[577]; it was therefore tentatively submitted that there have been two dedications for the existing building. At this point we are thrown back on the “two-period” theory, for the Abbey choir dates from the late twelfth century, while the nave, which is considered to be the deflected limb, belongs to the mid-fourteenth century. Yet why should the axis of the nave lie nine feet North of the choir line? Canon Atkinson, who carefully investigated this example, expressed the result thus: the axis of the nave diverges from true East and West by 14°·5, that of the choir by 9°·7. He adds this statement: “That it was planned so requires no elaborate proof[578].” The parish church of Whitby, St Mary’s, which is hard by, and which belongs to the twelfth century, runs exactly parallel to the Abbey choir. Why, again, assuming the theory of separate dedications, should the nave be assigned to one saint, and the choir to another[579]?

This query leads us to consider, for a moment, the subject of dedication to two, or even three patron saints. Miss Frances Arnold-Forster, in her Studies in Church Dedications, suggests several reasons for “twofold ascriptions[580].” The two names may represent that of the founder and that of the patron saint. The Lady chapel or chancel may have been placed under the invocation of one patron, and the rest of the building under another. Or, the founder may have deliberately intended to have a dual dedication. Again, when a church was rebuilt and re-dedicated, a new name may have taken its place alongside the old; such re-naming is thought to have been of frequent occurrence, especially in Cornwall. Lastly, the double dedication may be due to the union and consolidation of two parishes, as is the case with the large majority of our City churches.

Now, with respect to Whitby, we seem to have obtained a clue to the dedication difficulty, at least. The Rev. Canon G. Austen, Rector of Whitby, informs me that the Saxon Abbey was dedicated to St Peter alone. Though somewhat contrary to popular belief, it must be accepted as a proof that the ascription to St Hilda was a later addition. Let us suppose that St Hilda’s name was introduced at one or other of the rebuildings of the Abbey; and let us assume that there was a second formal dedication. This admission does not, perforce, imply our acceptance of the Saint’s Day theory as explaining the discordant alinements. Else we should expect to find many more cases of deflection among the churches with double dedications. Nor, again, does the fact of re-dedication necessarily support the notion that the builder was incapable of setting out a straight line. In short, the solution tendered enforces the very difficulties which it professes to dispel. If it can be clearly shown beyond dispute how the Mediaevalists obtained their axial line, then the theory that the misalinement is the result of re-dedication will have to be faced seriously. Until that time comes, the theory is little more than a fair surmise. As a slight contribution to the inquiry, and as a possible instance of the architect’s incapacity, the case of Leatherhead parish church may be cited. Here the tower is deflected about 3´·6´´ from the axis of the nave, while the nave diverges from the chancel. The church is dedicated to St Mary and St Nicholas, but whether there is further connection between the two facts is uncertain. It was a visit to this church which led Dr J. C. Cox to refer to “symbolic absurdity,” and the “leaning-head” theory as being propounded by “ill-instructed persons.” His own explanation of the divergences is twofold: the “endeavours to obtain the true East at differing periods of the year,” and “the well-known carelessness of Mediaeval builders in following out a true square[581].”

We next consider the arguments of the symbolists. Several solutions are offered by these authorities, though there is no great divergence of fundamental ideas. The favourite explanation, as just hinted, is that Christ, while on the Cross, faced Westwards, and that His head leaned towards the left, or South side. Another school teaches that the inclination was to the North, and, as we have already seen, such Continental churches as have varying axes, are deflected towards that point. In roods, it may be noted, the head is usually made to fall towards the left[582]. The “leaning-head” theory was accepted by Dr Rock. M. Huysmans has evidently paid so much attention to this phase of the question that he may be again quoted. Describing the fine abbey church of Preuilly-sur-Claise, in Touraine, he declares that the builders gave life to the stones. In its serpentine line, in the perspective of its aisles, and the obliquity of its vaulting, the church gives an allegorical presentment of Our Lord on the Cross, and perpetuates “the never-to-be-forgotten moment between the ‘Sitio’ and the ‘Consummatum est’[583].” M. Huysmans mentions, with reserve, the opinion of some writers, that the bent line occasionally represents the body of a saint instead of that of the Saviour; thus the curved axis of St Savin is supposed to be symbolical of the wheel which was the instrument of martyrdom of the saint[584].

There are not lacking authorities who, like Dr Cox, spurn these theories as idle fancies. “Symbolism!” exclaimed Welby Pugin, when viewing a twisted chancel in Leicestershire, “Pack of nonsense: it was because they didn’t know how to build straight[585].” This summary verdict, told by a person who heard it delivered, may, however, be countered by an opinion in the opposite sense, uttered by the same architect, and equally well attested. Asked about the bend in the nave at Whitby Abbey, which has just been mentioned, Pugin declared that it signified that the debt of redemption had been paid; “for, after the Saviour had expired on the Cross, his head would naturally lean or incline to one side[586].” Which of Pugin’s two opinions is the earlier cannot be ascertained, but even if antagonism to the “leaning head” theory were representative of his maturer judgement, he surely could not pretend that the deflection can always be ascribed to a blunder. Mr Parker himself, who advanced the theory that the twist might be due to the consecration of a building before its completion, admitted that many cases of deflection are incapable of a constructional explanation[587]. One may well ask whether the architects who reared our magnificent Gothic cathedrals, edifices whose every part, when untouched by the restorer, exhibits skill in design and workmanship, were really unable to build straight. What of the deflection at St Ouen, in the city of Rouen, a building which Fergusson enthusiastically declares was “beyond comparison the most beautiful and most perfect of the abbey edifices of France[588]?” And, coming to meaner structures, dare we say that a discrepancy of 5° or 10° is likely to be the result of mere carelessness? One might as reasonably argue that the existence of a “low side window,” with its deviations of size, shape, and position, indicates a clumsy arrangement of fenestration.

Rebutting arguments against the “theory of error” may be urged on broad grounds. We have seen that in the greater number of instances, our undeflected buildings bear to the North of the equinoctial point, indicating, if we suppose a calendar alinement at sunrise, that work was begun either a little after the advent of spring, or, less probably, a little before the autumnal equinox. Consider, then, the case of a church of which one member is to be rebuilt. Under similar conditions of starting work, the assumed error in direction might be on either side of the axis. We might perhaps conclude that it would actually tend to be on the Northern side of that axis—a Northern deviation added to a Northern deviation—for, in spite of liturgical rules, a liking for summer orientations (North of East) appears to have been very common. But, not to press this probability, it is at least claimable that, on a balance of instances, the deflections, if due to clumsy workmanship, would range themselves in equal numbers North and South of the earlier alinement. But the fact remains that the Southern variations form the rule, and we are driven to the belief that they are Southern by reason of design.

This leads us to consider an hypothesis which embraces a purpose more aesthetic than symbolical. Put shortly, this hypothesis is, that the bend was designed to produce an artistic illusion—a perspective effect whereby a building appeared to be longer than it actually was. Thus, if one of the side walls of a church with a “weeping chancel” be viewed from the Western entrance, an impression of greater length and of undefined distance is received by the spectator. Again, where a rood screen exists, the bend in the wall, not a pleasing feature, considered separately, would be concealed, but a change in the direction in a lofty roof would still produce an illusion that the church is indefinitely extended. Moreover, as the beholder caught a glimpse of a portion of the sanctuary window, which was often richly ornamented, both in form and in colour, the beauty of the vista was much enhanced. On this hypothesis, the screen, instead of being the immediate cause of the inclined axis, was an accessory to a complete design, a feature of which the effect was foreseen and provided for. To round off this question, it should be added that not many relics of screen-work exist which belong to a period earlier than the fourteenth century. A few specimens of thirteenth century work remain, and doubtless others were swept away at the Reformation, but most of the examples which have come down to us are assignable to the fifteenth century[589]. In short, the middle stage of the development of rood-lofts seems to correspond roughly with the period in which deflection is most commonly observable. This coincidence does not, indeed, solve the riddle, since each of the opposing schools may produce it as testimony.

Seeing that the “aesthetic theory” is unsupported by documentary proof, evidence must be sought in the possible existence of similar contrivances associated with other architectural features. Can such evidence be produced? The reply is in the affirmative. As one enters certain Egyptian temples, he perceives a gradual transition from light to darkness. This effect is brought about by the forced ascent of a few steps, combined with a lowering of the roof; thus the sense of gloom and mystery increases as the worshipper moves forward[590]. Again, we find that, in the columns of the Parthenon and other Greek buildings, there is an entasis, or slight convexity of outline, perchance not more than the length of a finger nail, yet sufficient to prevent the appearance of hollowness. The Parthenon is said to possess scarcely a vertical or a horizontal line throughout the whole design. The columns lean a little inwards, the corners incline diagonally, the entablature is curved and recedes in the centre[591]. Such subtle niceties, trivial when taken singly, prove that, in Classical times at least, the aesthetic phase of architecture was not ignored. The next question is, Can parallel instances of design be observed in our own country? The Norman church of Barfreston, Kent, may be cited. In this building, the axes of the nave and chancel correspond, but the walls of the chancel are so built that this portion appears to incline about 5° North. One can hardly find a rectangle in the plan of the church—“everything is oblique[592],” and apparently oblique by intention. Again, the South arcade of the nave of Scarborough parish church has a distinct curve, as if deliberately thus planned. In the case of Whitby Abbey, before cited, Canon Atkinson discovered another intentional irregularity. The West wall of the North transept projects into the interior at least a foot more than does the East wall. The view, being thus interrupted by the broken line, has a pleasing effect upon the mind of the spectator. Consider, too, the churches, to be met with in many a village, in which the pillars and mouldings of the Northern arcade are simpler than those of the Southern, the Northern windows less rich in tracery, the specimens of painted glass more sombre in motive and execution. It will be understood that a reservation must be made for those churches in which the South arcade is of slightly later date than the North. The symbolism of the Northern aspect of Nature must have been clearly appreciated for these conditions to have arisen (see p. 334 infra). Or, take the feature just alluded to as entasis, by which the optical illusion of hollowness in upright lines is corrected. Many of our spires exhibit this convexity. Mr Francis Bond, a high authority on such matters, states that the entasis probably does not exceed 1 inch for 60 feet in our best examples of Gothic spires. The important point to notice is, that the curvature exists as the result of design. True, it has been mooted whether the bulging is not an incidental consequence of a peculiar method of building, but this idea is not commonly accepted, nor does it seem to satisfy the conditions of the problem. Sometimes, as in the Lincolnshire spires of Caythorpe and Welbourn, and of Glinton (Northants), the limit just given is greatly exceeded, with results not at all pleasing to the eye[593]. Caythorpe spire, once grotesque with its “sugar loaf” protuberance, exhibits the familiar curve even after being shortened and rebuilt[594]. That there appears to be an artistic necessity for the convex treatment, or for some equivalent device, is exemplified once more in the steeple at Louth. There, the impression of concavity is prevented by increasing the projection of the crockets about one-third of the way up the spire[595]. Again, there is the artistic disposition and gradation of ornament, the delicate, enriched carvings being reserved for the parts of the building near the eye, while a “greater effective quantity” is provided in the upper parts, as on pinnacles and parapets. Perhaps the reader will admit that there is no need to press further the plea of aesthetic artifices on the part of the early builders. It is, at least, conceivable that our English masons understood the value and charm of subtle suggestions of spaciousness and mystery. To elaborate their art, and to conceal it, and to keep

may well have been one of the golden rules of the Gothic architect. For the trained eyes of the sympathetic beholder admire “both what they half create, and half perceive,” and they will find privileged pleasures in sublimities of form and colour.

Nevertheless, in the absence of positive testimony that our early builders were actuated by artistic principles such as those just described, there is no justification for putting the case a whit more strongly. Is there, indeed, any evidence at all for the intentional designing of deflected churches? St Charles Borromeo (A.D. 1538-1584) has been cited as recommending deflection towards the South, if deflection be necessary[596]. The reference is probably the result of misunderstanding St Charles’s meaning, and, in that case, the claim is vitiated. What the writer actually required was, that the High Chapel (cappella) should look directly East, equinoctial East. Should dwellings exist which obscured the view, that fact did not relieve the builder of his responsibility. If the axis of the building must be turned, it should be towards the South, and not towards the North. The passage seems to allude to the church as a whole, not to the High Chapel, but I admit that there is some little ambiguity[597].

Attention has been drawn by Mr W. F. Hobson to an early allusion to orientation, involving, as he considers, the principle of deflection also. The passage occurs in the writings of St Paulinus of Nola, c. A.D. 420. Speaking of a church which he had built, St Paulinus remarks, “Prospectus basilicae vero non, ut usitatior mos est, Orientem spectat, sed ad Domini mei B. Felicis basilicam pertinent Memoriam ejus aspiciens[598].” The statement is that the outlook of the church was not directed towards the East, following the more common practice, but towards a certain basilica, containing a particular tomb, that of the martyred presbyter, St Felix. That we have here an early instance of orientation is manifest, but whether “aspiciens” may be interpreted so as to prove deflection is doubtful. It is, however, possible that while the church had its general alinement to the sacred basilica, there was a bend or slant which caused the “chancel”—speaking conventionally—to face the monument. Mr Hobson urges that, from orientation towards a tomb and orientation for a Saint’s Day there is but a slight transition[599].

We will now attempt to sum up the evidence. Skew chancels were, in some cases at least, due to carelessness or unskilfulness. A few deflections seem to have been caused by the architect’s dislike to build on old foundations—a prejudice felt by Wren in re-building St Paul’s. The Saint’s Day theory may hold for some examples, since a few positive instances of observance, if these can be produced, cannot be nullified even by many negative exceptions. As to the symbolism of the leaning head, whimsical though it may at first sight appear, one cannot dismiss it as mere folly. Mediaeval symbolism was no dead art, and it is visible in many guises, though to a less extent than its advocates would insist. Unless, therefore, we are to believe that symbolical explanation is altogether an after-thought—the creation of cultured ecclesiastics—we have to find a reason for its existence in Mediaeval days. And, if the idea be so old as the foundation of our finest churches, we may conjecture that it received an embodiment in architecture. Lastly, the theory of artistic design, though least advocated, seems most in consonance with all the facts. The deflection does cause an agreeable optical illusion; and it is probable that this was intended. While we search in vain for some early document to throw a glimmering of light on the problem, we are bound to consider all tentative solutions unsatisfactory. He would be a bold man who should unhesitatingly affirm that one explanation will meet every case, and hardier still would be the prosaic architect who should dismiss every instance of deflection as the result of pure chance or blundering ignorance.

CHAPTER VI

THE ORIENTATION OF GRAVES

Durandus, who has already been frequently cited, states the rule, and gives as a reason that the Eastward position is properly assumed in prayer[600]. Bede explains the arrangement on the supposition that the Coming of the Great Day and of the “Sun of Righteousness” will be seen from the East, hence the dead should face the sunrise[601]. The Scriptural allusion in Zechariah, too, has often been quoted in support of the practice; the feet of the Messiah, so the verse runs, shall “stand in that day upon the Mount of Olives, which is before Jerusalem on the East[602].” Larousse, to mention but one more writer, refers to the injunction, but adds that the bodies of priests, martyrs, and bishops are laid in a reverse direction—“caput versus altare,” so that they face the West. A curious reason is assigned: churchmen of high or clerical rank are expected to rise first, and to pass onwards with the head looking West[603], a strange contradiction of Bede’s hypothesis.

Since the effective force of a rule is tested by its exceptions, we shall now glance at some breaches of conformity. With such transgressions of custom one might profitably compare the deliberate refusal to orient churches, discussed in the preceding chapter. Thus, in John Aubrey’s Remaines of Gentilisme and Judaisme, we are assured that most of the ancient graves at Middleton Stoney, Oxfordshire, were made to lie North and South. Whether this were due to ignorance or to the spirit of opposition, the author of the statement, Dr White Kennett, was unable to say. In one of the rare and anonymous “Marprelate” tracts (A.D. 1589), now familiar to most students through Professor Arber’s admirable “English Scholar’s Library,” we read that Martin Month “would not be laid East and West (for hee ever went against the haire) but North and South: I think,” adds the tractarian, “because ‘ab aquilone omne malum,’ and the South wind ever brings corruption with it[604].” Martin Month has not lacked followers. Mr Hissey, in his Over Fen and Wold, has recorded a seventeenth century example from Lincolnshire, a stronghold of Puritanism[605]. William Glanville, who died at Wotton, near Dorking, in the year 1750, left documentary injunctions requiring, among other curious details, that he should be buried facing the North. In a case of infringement of the rule, as related by Brand, the motive, strangely enough, was not contumacy, but the desire to exhibit a visible mark of penitence and humiliation[606]. In support of this view, attention may be called to a stone coffin which is built into an exterior angle of two walls at Lindisfarne Priory church. The coffin lies South-East and North-West, and is believed to have held the body of a monk who had broken his vows.

Thomas Hearne, the antiquary, as befitted a philosopher, ordered that his executors should determine the true orientation of his grave by means of a compass (A.D. 1735). This was done, with the result that the monument appeared to be awry, though the actual error, it is affirmed, lay in the incorrect orientation of the surrounding graves[607]. One is inclined to modify this explanation; the divergence was probably caused by the different modes of orientation, the earlier graves having been most probably alined by the sun or the church fabric, while that of the antiquary was set out by the needle, and, it may be, without making allowance for the declination. In a parenthesis, an analogous case, which came under the notice of the present writer, may be given. An oblong kerbing of marble was to be placed around a grave in the churchyard of Chipstead, Surrey, and the mason was told to take his alinement from the wall of the chancel. When the task was finished the kerbing was seen to be out of line with the surrounding graves. Dispute was followed by a careful test. The kerb was true to the chancel wall, but, as was then discovered, the chancel itself was askew, when compared with the nave. Was the twist, then, of a date posterior to that of the earliest grave-mounds, and were these made parallel with the nave? I think the answer to the latter question is probably affirmative, but the puzzle of dates cannot be so readily solved. It is almost certain that none of the existing mounds is co-eval with the rebuilding or partial rebuilding of the chancel[608], but again, the present mounds may have been alined from a succession of earlier ones.

As already suggested, direct infraction of the general rule arrests the attention; were it not for this, familiarity might make us blind. So again, one may easily pass by at least two Shakespearean allusions. “Make her grave straight,” says one sexton to the other, when about to dig Ophelia’s grave[609], and, though certain commentators have considered the words to imply “Make it immediately,” there is good authority against this interpretation. The alternative reading supposes that the command is to dig the grave East and West. This rendering of the words seems very reasonable, seeing that they form part of an answer to the question whether Ophelia shall have Christian burial. “I tell thee she is: and therefore make her grave straight.” The explanation is in full accord with the tradition, held in the South-East of England, that the bodies of suicides were buried in a North and South direction. The instruction given by Guiderius to Cadwal concerning the apparently dead body of Imogen is more specific:

The meaning probably is, “Lay his head to face the East”; although the other interpretation, “Head directed to East, feet to West,” is possible. At any rate the passage clearly shows that the orientation of graves was assumed by the poet to have been observed, even in the solitudes.

A scrap of quaint, but illuminating, evidence comes from the folk-lore of Wales, where the East wind is called “the wind of the dead men’s feet[611].” Nor is the practice of orienting graves confined to Britain. A Scandinavian folk-story makes the grave-diggers, either through stress of bad weather, or “out of mischief,” bury an unpopular person, one Jón Flak, in a grave dug North and South. Every night the dead man haunted the grave-diggers and repeated this verse:

And no rest was obtained until the body was taken up and buried in the proper position[612]. A study of comparative customs will, however, furnish us with departures from the British plan, although haphazard alinements seem to be exceptional. Thus, some Australian tribes allow the head to front the West, others the East. The Samoans, the Fijians, and the North American Indians place the head to the East; this was also the practice of the Peruvians. Some South American tribes and the old Ainus of Yesso made their dead face the sunrise[613]. Plainly, while all these instances involve alinement of the corpse, they illustrate the two opposing ideas mentioned in the last chapter,—that of darkness and death, connected with sunset, and that of resurrection and new life, typified by sunrise. Sir Thomas Browne, in his Hydriotaphia, remarks: “The Persians lay North and South: the Megarians and Phoenicians placed their heads to the East; the Athenians, some think, towards the West, which Christians still retain. And Beda will have it to be the posture of our Saviour[614].”

Various Greek authors attest the practice of orienting graves before the time of Solon (died c. B.C. 558)[615]. Setting this evidence aside, and restricting the inquiry chiefly to our own country, we have good reason to believe that there has been continuity of custom from the Saxon period at least. Mr Romilly Allen asserts that the East-to-West position was usual in the early days of Christianity. Further it was formerly suggested by Mr Reginald A. Smith that, wherever East-to-West graves are met with in Saxon cemeteries, they indicate the burial-places of Christians, in contrast to the tombs of the heathen. But the discoveries at Mitcham, in Surrey, of a large number of skeletons carefully orientated, break the rule, since Mr Smith himself shows that the burials were of pre-Saxon date, while the associated relics point to pagan interment. Again, a cemetery which was discovered at the Roman “level” in Bishopsgate Street, London, contained bodies laid East and West. Since no pagan objects were associated with the skeletons, we may perhaps conjecture that the burials were those of Christians[616]. But the most common position of skeletons in Roman cemeteries, according to Wright, is found to be East and West; usually, though by no means always, the feet are towards the East[617]. This generalization must now be considered too confident. There is the further complication that some of the cemeteries contain burnt, as well as unburnt bodies, suggesting a mingling of pagan and Christian burials.

We next take the pre-Roman period. Two examples from France will serve to illustrate one stage. A cemetery at Charvais, belonging to the earliest Iron Age, contained more than seventy graves, all but two or three of which were so oriented that the head lay at the West end[618]. Out of five graves excavated at Pleurs, in the department of Marne, three exhibited a like arrangement[619]. Examples might be multiplied, so far as the Continent is concerned; the position of the skeleton in English interments of the Iron Age unfortunately seems not to have been much noted by investigators. Even the excellent Guide to the Antiquities of the Early Iron Age does not furnish much information about British orientation, the main cause, doubtless, being the neglect of the excavators to record such details. One remarkable interment, at Kilham, near Driffield, in the East Riding, is worthy of notice. The grave contained the remains of a man, with the head placed at the West end; on each side of the body a goat had been laid, with a similar orientation[620]. Mr J. R. Mortimer, who, in 1897, conducted a series of excavations at Kilham, gave this summary respecting the disposition of the skeletons. Of twenty burials, eleven bodies had the head towards the North, five towards the South, and only four West[621]. At the Late-Celtic cemetery at Harlyn Bay, Cornwall, the majority of the skeletons, which were contracted, lay on their left sides, and faced the East. This evidence, which might be supplemented, plainly shows that the placing of corpses East and West was not uniformly observed during the Early Iron Age.

But we must widen the scope of the inquiry, for it is necessary to remember that the round barrows of the Bronze and Aeneolithic periods, and the long barrows of the Neolithic Age, were nothing but huge graves. Mr Mortimer’s excavations in the round barrows of Yorkshire led him to the conclusion, that, while no rule could be formulated concerning the alinement of the body, “the most prevalent position of the head [was] to the West and the East[622].” Canon Greenwell’s explorations revealed a like absence of regularity in the position of round-barrow skeletons, though the tables which he furnishes indicate a tendency to an East-and-West approximation. One of his results is of great interest. He found that, wherever the head of the skeleton pointed, roughly speaking, towards the West, the body had generally been laid on its right side; where the head was towards the East, the body was usually resting on its left side[623]. The inference has therefore been drawn that the corpse was made to face, not merely the sun, but the position of the sun in the sky at the time of burial. This opinion has been widely held until recent years, but it is now becoming customary to heap ridicule upon it[624]. The scoffers should, however, in fairness, consider collateral testimony.

First, it is well to note an actual record of careful orientations. For example, it is not contested that, in the round barrows of Wiltshire there is a strong tendency for the body to face the South[625]. If the theory just mentioned be correct, the corpse had, in the Wiltshire graves, been placed with the face towards the midday sun, implying a possible, but not demonstrable intention, with some fixity of custom. General Pitt-Rivers, who believed that the Saxons arranged the body with the head towards the rising sun, examined 31 Saxon graves at Winkelbury. He discovered that, while the skeleton was always alined towards an Eastern point, the exact axis varied from E. 19° S. to E. 28° N., but was usually towards the North. Supposing that his theory were valid, he admitted, with great candour, that all but two of the bodies must, from their positions, have been buried in summer. Here, again, we meet a possible, but certainly not provable, adherence to a custom.

Let us now look at Mr Mortimer’s valuable tables, which are found on the page of his book to which reference has just been made. From an examination of 383 interments, Mr Mortimer obtained these results:

The 31 unascertained positions must be ignored. The 71 instances in which the body lay on its back or chest are obviously testimony against the “face-to-sun” hypothesis. Moreover, of the 281 remaining cases, only 61 will be found to have the head pointing North-East, South-East, or in intermediate directions. At first sight, this summary is fatal to the theory. But the question is not easily settled. Other details must be known concerning each individual interment, and these are, to some extent, lacking. It is evident that a corpse may occupy any one of numerous linear positions, and may lie on either its right or left side, and yet face the sun at some time of the day. For the sun constantly changes its position in the course of its daily journey, and also alters its points of rising and setting. The consequence is, that a body lying on its left side might have its head directed to any point from North-East round to North-West approximately, and nevertheless face the sun. If we do not know the season, or the hour, of the interment, we cannot assert that the corpse did not so face the sun. Similarly, if the deceased tribesman were placed on his right side, he might, broadly speaking, occupy any position from East round to West, and yet fulfil the condition. Now the tables show that 245 out of the 281 bodies might conceivably have been laid as the theory supposes. Mr Mortimer gives Greenwell’s list for comparison: it supplies details of 234 bodies. Of these bodies, 112 lay on the right side, and 122 on the left—there is no reference to other positions. An inspection of the columns reveals that about 200 cases might support the “face-to-sun” theory.

It would be highly undesirable to press the facts further than is necessary to appeal for a suspended verdict. The “face-to-sun” hypothesis was never meant to be more than a possible explanation, but it has the virtue of accounting for many instances in which the skeleton lies, not due East and West, or North and South, but obliquely, at some intermediate horizon. And, at least, the alinement of the head, if not the face, towards the sun, is not peculiarly a British feature of prehistoric burials. Among the Tlingits, or Tlinkits, of South Alaska, the corpse was buried with the head towards the sun, so as to allow the soul of the deceased person to return; if the body were laid in the contrary direction, the spirit could never come back. Professor Frazer, who records this custom, asserts that other totemistic peoples inter their dead, according to fixed rules, with the head towards particular points of the compass. Professor E. B. Tylor has also collected abundant testimony corroborative of such customs. In places widely sundered geographically there has existed a strong desire to bury the dead “in the path of the sun[626].” Recently, too, Dr A. W. Howitt has discovered, just in time, at the Australian station of Wotjo, an isolated tribe which still preserves the custom. The arrangement of the head of the corpse is determined by the class and totem of the deceased person, and the direction is fixed by reference to the rising sun. The practice is nearing extinction, for, although Dr Howitt was able to secure sufficient information from the natives to construct a kind of burial compass, they could give him no reason for the varying dispositions. The only explanation proffered was: “Oh, that is what our fathers told us.” Such a man was buried in such a way because he was “nearer to us” than the others[627]. The folk-memory of this tribe will soon become a mere echo of tradition, and the task of reconstructing these primitive burial customs will grow increasingly difficult as each aboriginal patch of the earth’s surface becomes influenced by civilization.

Respecting graves themselves, we readily perceive that a round barrow, from its very nature, cannot be normally oriented. But in the British long barrows, usually accepted as the earlier kind, the long axis generally runs East and West. The broader end of the mound, where the sepulchral deposit is, as a rule, discovered, is directed towards the East[628]. So that, during the Neolithic period, some broad principle of orientation was observed, the constraining motive being probably connected with sun-worship.

Orientation in barrow burials having been considered, one is easily led to discuss an allied subject, in part connected with interments, in part not so strictly related, yet tinged with like primitive ideas. This topic is the orientation of megaliths and ancient earthworks, and it has been so prominently put forward by Sir Norman Lockyer that no archaeologist can pretend to be alert and watchful who has not, to some extent, followed the discussion. Hence a notice, however slight and imperfect, is demanded. Sir Norman Lockyer has made careful triangulations and measurements in connection with the prehistoric monuments of Cornwall, Wiltshire, and other counties, and his general conclusion is, that the axes of these monuments, and their mutual geographical relations, were carefully determined by their constructors. The assumption having been made that the monuments were originally alined to sunrise, the problem has been to calculate when that arrangement could have existed, because now, owing to the alteration of the inclination of the earth’s equator with regard to the ecliptic, the solstitial points have considerably shifted their positions. Stonehenge, which Sir Norman Lockyer considers to have been a solstitial temple, was accurately alined by astronomical means when erected. He believes that the original alinement of the circles was adapted (c. B.C. 2200) to a “farmer’s year,” May to November; but that, about B.C. 1600, a new cult was introduced, probably from Egypt, and the lines of orientation were changed to harmonize with a June-to-December year[629]. The investigation has also been applied to other monuments, and it is claimed that not only can we determine to what star a particular megalith was originally oriented, but, as a consequence of this determination, we can discover the approximate date also of the erection of the stones[630]. To any one who, like the present writer, is not a skilled practical astronomer, the evidence is difficult to handle, for the actual observations and measurements must, of necessity, be implicitly accepted. Yet, on general grounds, one cannot avoid the thought that the assumptions are not self-evidently true, and that even if this were the case, the theory is pushed much further than seems reasonable. These objections have been raised by several writers, and the views of Sir Norman Lockyer have been exposed to severe, and, apparently, disintegrating criticism[631]. It is but fair to add that these rebutting arguments have been met by rejoinders, though of varying value. Do we urge that the theory, with the minute measurements which it involves, presupposes too high a state of culture for those early times, too deep a knowledge of astronomy, too complicated a system of religion or worship? We are met by the enunciation of another hypothesis, that the Druids of Caesar’s time, admittedly men of some learning, were the descendants of astronomer-priests of a more ancient British civilization[632]. Do we ask for proof of re-orientation, as a result of the immigration of a new race of men? We are thrown back on a further series of intricate astronomical data, reinforced by similar details relating to Egyptian temples. It may be that the objector has some misgivings as to the exactitude of the orientation of these very temples—that is a small matter—but he would like to know the reason for placing our rough, unhewn cromlechs and menhirs in the same category. To the general archaeologist, the comparison between our rude “amorpholiths” and an elaborate Eastern temple is not just. Such a person asks questions about the instruments which the builders of the megaliths may have used in their observations, about the possible obstacles which may have blocked out of view portions of the horizon, about the difficulty in ascertaining which standard line was used, where all the monoliths are so crude and irregular. Again, which was considered the time of sunrise—the moment when the sun’s rim peeped above the horizon, or when his centre was clearly visible? Each position has been assumed, the one in England, the other in Egypt. In the ages when not even the Julian calendar had been invented, could primitive folk know, for certain, the exact date of the solstice? In England, the sun rises on the slant: where was the observer’s eye placed in relation to that oblique line? In one Cornish instance, at least, it has been asserted that the stones from which the alinement had been taken cannot be seen from the central group of stones[633]. And of Sidbury Hill, which was deemed so important in the Stonehenge measurements, Dr T. Rice Holmes declares that it cannot be seen from that monument, only the trees on its summit being visible[634].

Somewhat in accord with the conclusions of Sir Norman Lockyer, but by no means in total agreement, are the theories propounded by the Rev. J. W. Hayes, of West Thurrock. Mr Hayes, who has made a close study of stone-circles, especially with regard to their use in Gorseddau and other assemblies, has kindly permitted me to read his unpublished papers, and to glean therefrom such information as is germane to the present question. It is manifestly impossible to give more than the briefest summary of the evidence, or to enumerate the miscellaneous documents, written and printed, which Mr Hayes has consulted. The leading idea is, that the stone-circles were first erected to serve as indicators of seasons. By means of these “dials,” the priestly caste determined the times of fasts and feasts. In spring, there were sacrifices to ensure bountiful crops. Before entering upon a war, there were offerings to propitiate the gods and thus to win success. A comparison is made between the old festivals and our own Easter—itself a pagan feast originally, and still depending on the moon for the exact date of its observance. Analogies are also drawn from the customs of the Jews, the Romans, and the Egyptians. The outstanding menhirs connected with stone-circles are considered to have been intended to provide sight-lines, while the trilithons of Stonehenge, it is suggested, marked out sky-spaces. It is interesting to note, in passing, that Professor W. Gowland has recorded the use of wooden trilithons by modern Japanese sun-worshippers, for determining the position of the sun and of sacred objects.

It will be inferred that Mr Hayes’s theory makes the sacrificial use of the circles subsidiary to a primary astronomical and judicial purpose. Yet he takes note that the old Irish name for a cromlech—the accepted term among modern writers for a “stone-ring”—was siorcal leacht, or “circle-stone-of-death.” In old Irish manuscripts such circles are always associated with pagan priests or druids. The sepulchral aspect of stone-circles, however, seems to be largely ignored by the modern school of investigators.

Mr Hayes lays stress upon a singular fact, the true explanation of which has long been an enigma—the frequent occurrence of a definite number of pillars in the stone-ring. Nine is not an uncommon number in England, though it is recorded from Ireland and Germany. (The stones of the Nine Maidens, near St Columb Major, in Cornwall, form a row, not a circle.) The cromlechs of Boscawen-ûn and the Dawns Mên, in Cornwall, of Whitemoorstone Down, Dartmoor, as well as the inner “horseshoe” at Stonehenge, have nineteen uprights. Thirty, again, is the tale of some circles, as in the outer ring of sarsens at Stonehenge. The group of nine is puzzling, but the arrangement of nineteen is held to represent the metonic, Indian, or lunar cycle of the astronomers. After the lapse of nineteen years, the sun and moon occupy relatively the same positions as at the commencement of the cycle, so that full moon and new moon recur on the same days of the months as before. Whether this astronomical explanation of the number nineteen can be proved or not, one curious feature deserves notice. The circles which show nineteen stones have a gap towards the East, and no explorations have hitherto brought a twentieth stone to light.

It further appears that the old Gorsedd circles (p. 98 supra) consisted of twelve stones, set 30° apart, and said to represent the twelve signs of the Zodiac. The circles reared at the annual Eisteddfod by the neo-Druidic cult are arranged similarly. Besides the twelve pillars, three outstanding menhirs were fixed at the North-East, East, and South-East respectively so that the sun’s rays at the solstices, or the equinoxes, as might be, fell on the central stone, or “chair” within the circle. And, by a strange coincidence—or is it due to imitation?—another plan of the Gorsedd circle, as preserved in ancient manuscripts, gives nineteen stones, with a gap towards the East. In this case, eight outstanding pillars were erected, so as to indicate, by shadows or rays, (1) the summer and winter solstices, (2) the old May-December agricultural year, (3) the equinoxes, and, possibly, (4) other solar and lunar periods.

The weakness in the foregoing theories seems to lie in the assumption, up to the present unproved, that the primary purpose of the cromlechs was to serve as places of assembly. True, the convocations are supposed to have been hedged with much ceremony, to which the stones were accessory. But, dealing with circles, are we not liable to argue in a circle? Until the prime purpose of the cromlechs is ascertained, the student is forced back like the turnspit in Hudibras: