by Prof Carl J.S. Markstander

Norsk Tidsskrift for Sprogvidenskap Bind VIII, Oslo, 1937

[The main body of this paper is in Norwegian - however the important summary is in English and is reproduced below, as access to this Journal in English-speaking Libraries is not easy - note the hyperlinks are, obviously, my addition.]


The Isle of Man is for administrative purposes divided into six sheddings, each of three parishes (apart from Glenfaba that has two only). Each parish contains a varying number of treens and each of these, on an average, four quarterlands. Above this administrative apparatus towers the House of Keys.

It has often been maintained — though not definitely proved— that this administrative system is of Norwegian origin. In my opinion, too, the system originated, in its present form, during the Norwegian period; but at the same time it has its roots deep in the Pre-Norwegian Celtic community in Man, as will be shown below.

The Treen

To start with the smallest administrative unit, the treen may be defined as a district bounded by natural borders such as glens, mires, rivers and streamlets, but varying in size from less than 200 to more than 600 acres. Each treen formed a fiscal unit and paid a fixed annual tax, levied on the four farms or quarters of which it consisted. Apart from the cultivated land, the been also comprised wastes and hill pastures common to all quarterlands, and often also intacks, i. e. land brought under cultivation or utilised otherwise and entered on the Manorial Roll. The been, in consequence, is a combination of farms, a grend, a country district, jointly furnishing a fixed rent.

The Treen-rent

The oldest records of the lord's tenants and the Lord's Rent date for the southern division of the isle from the year 1511, for the northern, from 1515. The average rent amounts to about 17S. 10d. a quarterland or 71S. 4d. a treen (p. 290 ff.).

Age of the treen

The Manorial Rolls of 1511-15 are based on the treen. The term treen occurs in the Ballaugh Register of A. D. 1600 and is common in documents from the 17th to the 19th cent. The earliest reference to the treen is contained in Add. Charter of 1408, where the parochial church of Rushen is mentioned as ecclesia Sanctue Trinitatis inter Prata. Inter Prata is nothing but the Latin rendering of the treen-name Rowany, which Man. Roll 1511 writes as Edremony from Old Manx Etar Da Moini, litt. "between the two moors", cf. Irish names such as Star Da Glass.

But the treen division may be traced still farther back. Even in the 16th century, two-thirds of all the treen-names are still Norwegian, and of the other third several are Norwegian in Gaelic disguise. The conclusion is that the treen division dates back to the Norwegian domination. There can be no question of these Norwegian treen-names having secondarily developed from names of quarterlands; for, corresponding to the social distinction between the Norwegian aristocracy and the Gaelic tenants, nearly all Norwegian farm-names in Man are names of treens, while nearly all names of quarterlands are Gaelic. This conclusion is confirmed by certain legal documents from the 18th cent., produced by Richard Stevenson when defending before the court his ownership of the treen-farm Balladoole in Marown [sic in Arbory] and published by W. Cubbon in the Journal of the Manx Museum 1934. According to these documents, which consist of transcripts of old long lost records of the Lord's Rent, the family of Stevenson had held the farm of Balladoole for 6—7 generations previous to the Manorial Rolls of 1511, and throughout this period had paid the same rent of 56 S. 8 d. as in 1511. This takes one back to the end of the 13th or at all events to the beginning of the 14th cent. The last Norwegian King of Man and the Isles died in the year 1265.

The fact that the treen, as a fiscal unit, originates from the Norwegian administration, does not preclude the possibility that, as a unit of land, it may be older than the Viking Age.

Treen and Church

Church lands and Abbey demesnes were leased in quarterlands to tenants for a rent, which on an average was slightly lower than in the neighbouring treens. But the fact that they themselves were not divided into treens, indicates that the treen division did not originate in the Church or in the Monastery. Several considerations, however, point to a certain connexion between the Church and the treen system. In the first place, no treen and no quarterland conflicts with the parish boundaries, i. e. all quarterlands within one and the same treen are all situated in one and the same parish. Secondly, five treen-names begin with Kirk- "church", and finally, every treen had at least one chapel.

The treen-chapel

Keeills is the standing term for the primitive chapels which once were scattered all over Man to a number of more than 200. The survey (p. 321 ff.) goes to show that about 160 such treen-chapels are known at present in the Isle. Even if no keeill now remains intact and as a rule only the foundations are left, one is able to form an idea of their original appearance. They were built of unhewn stones, slabs, earth and rubble, in ancient times, with no other kind of agglutinate than earth or clay, later, shell mortar or cement. The size varies vary considerably. The Ballachrink-keeill in Marown measures only 10' x 6' inside. Otherwise they may attain 23' x 13' (Keeill Vian, Lonan), even 57' x 18' (St; Patrick's Chapel, Patrick's Isle), and 75' x 24' (St. Trinian's, Marown). The walls vary in thickness from 2' 4" to 4' 8" and are, on the outside, protected by an embankment of earth and stones, in height 2'-5', in depth 4'-10'. The shape is rectangular with no division between nave and chancel. The door which is narrow and tapering towards the top, is usually situated in the western gable. The window — as a rule only one — is built at a height of 2'-3' above the floor. The altar is invariably placed against the eastern wall, attaining a height of about 2'. Usually the keeill was built upon a natural or artificial elevation in the terrain and had attached to it a burial ground, surrounded by a mound, which in height could reach 5'-6'. The whole plan of the keeill recalls, on a small scale, a rath or a liss, upon which it was certainly modelled.

We know but little of the manner in which the keeill was administered. A Norwegian runic inscription from the keeill of St. Mary in Cardle, dating from the12th or 13th cent., mentions a Joan Prest i Kornadal, and as late as this some keeills must have still been in use. They certainly fell into disuse at very different times. We know, however, they were used as burial grounds long after their other functions had ceased. Even in the 19th cent. burials took place in the Lag ny Killey in Patrick.

Keeill types

The keeills may be divided into three groups according to the interrelation of length and breadth:

1. keeills with breadth = 2/3 of the length

2. keeills with breadth = 1/2 of the length

3. keeills with breadth = 1/3 of the length.

The narrower the breadth compared to the length, the larger is the keeill and the more advanced the building technique. Only in group 3 has lime or shell mortar been used as cement.

Keeill survey

On p. 321-333 a list is given of all keeills known in the Isle of Man, with particulars regarding the treen and quarterland to which the keeills belong, their dimensions, the archeological finds made on their sites, their dedications etc.

The list comprises 159 keeills. To this must be added some 15 parochial churches, which in all probability hat e gradually developed from old keeills. Altogether about 174 ancient keeills or keeill-sites are known in the Isle of Man.

Age of the keeill

The age of a keeill is often difficult to decide. Of course, the dedication only gives a terminus ante quem nom Of particular interest are dedications, relating to saints of a relatively recent age, such as Machutus who died in the year 565, Brandan -† 578, Columba - † 597, Comgall -† 600, Cuthhert - † 687 and Adamnan -†703. No keeill can be shown to have been dedicated to any saint later than Adamnan who died three generations before the coming of the Norwegians to the island. No notice is taken of the Church of Olav (sill Amhlaoibh) on Euastad, mentioned in King Gilded Svarte's charter of 1154, which cannot have been a keeill in the usual conception of the word, but a chapel of considerable dimensions.

A far more reliable chronological foundation is furnished by the archeological finds made on the keeill-sites. In 18 keeills and parochial churches Norwegian ornamental cross-slabs and runic inscriptions have been discovered (p. 338).

These keeill consequently date back to the Norwegian period.

But many keeills are considerably older, as is proved by a long series of primitive crosses and cross-fragments of the 6th — possibly even the 5th — cent. from 42 keeills or parochial churches (p. 339).

An examination of the keeill structure lends no support to the supposition that keeills with Norwegian finds belong to a more recent type than those having Pre-Norwegian finds, the dimensions and the technique being the same. As a matter of fact, the three keeills which show Norwegian finds all date from Pre-Norwegian times, as is definitely proved by old cross-slabs (p. 340).

In all keeills which can be shown to be of Pre-Norwegian date, the breadth is at least ½ of the length, even in several cases 2/3 of the length. All keeills of these dimensions may be presumed to be older than the Viking Age.

In fact, the 15 undatable keeills of these dimensions exhibit exactly the same primitive structure as the datable ones. Added to the 42 previously mentioned, they make in all 57 keeills undoubtedly older than the Norwegian period.

In distinct contrast to these primitive keeills, one finds a smaller group of keeills or chapels, distinguished by a manifold increase in size and a reduction in breadth to 1/3 of the length, typical instances being St. Patrick's Church (German) 57' x 18', Maughold parochial church 65' x 18', and St. Trinian's (Marown) 75' x 24'.

In many cases old finds go to prove that these large chapels must have been Christian places of worship centuries before the Viking Age, this holding good e. g. for the parochial church of Maughold and St. Trinian's.

The conclusion necessarily is that: All Manx keeills of the usual small dimensions are, in principle, older than the Norwegian domination. The large keeills, however, with a breadth 1/3 of the length originate from Norwegian times, most of them developing into parochial churches. The majority of the latter being, like many of the large keeills, places of Christian worship long before the Viking Age, the large keeills must be looked upon as reconstructions of older, primitive keeills. In consequence, all keeills in the Isle of Man are in principle of Pre-Norwegian origin, being founded by Celts, not by Norwegians. Nothing could be more natural; for it can scarcely be assumed that a people, able to create such important monuments of art as the ornamental crosses of Gaut and his school, and noted for the magnificently adorned faces in its home country, would ever have built such primitive places of worship as the small keeills.

Treen and keeills

In the ballad Mannanan Beg Mac y Lir, dating probably from the beginning of the 16th cent., it is said that St. German built a small chapel in every treen balley, that the people might come there to pray. By treen balley the author no doubt means the treen, the expression being modelled on kerroo balley "the quarterland farm" (lit. "the quarter of a (treen-)farm"). The tradition, to which the ballad gives expression, can be traced back to the beginning of the 12th cent., the verse of the ballad being only a version of what Joceline, the well-known monk of Furnes, tells of St. Patrick granting to St. German authority over Man, enjoining him to build chapels and confirm the people in the faith.

Is this close connexion between treen and keeill of which the ballad speaks original and historically founded? Or is it due to later conclusions ?

Although the 174 keeills known (parochial churches included) apparently accord with the 175 treens mentioned by Manorial Rolls of 1511-15, this agreement is only coincidental. Of the 174 keeills 15 are parochial churches and 35 are situated on Church lands and Abbey demesnes, the treen-number of which is not known and consequently cannot be taken into account here. The remaining 124 keeills are distributed among 96 treens (with intacks), every treen thus having 11/3 keeill.

In 83 treens no trace of keeills has so far been discovered.

But the fact that a treen has no keeills does not necessarily imply that a keeill has disappeared, for very likely many treens in which no keeill is found, have never possessed one.

Above all, this is certainly the case with the alia-treens, which as shown NTS VI 310 ff.[Norsk Tideskrift for Sprogvidenskap] originated from the division of older, larger treens. In perfect accordance with this theory, the mother-treen and the alia-treen, in no case both possess keeills. In 5 out of 8 cases, the keeill is situated in the mother-been, in two in the alia-treen, while in one (Leodest) no keeill is known to exist in either (p. 348). The keeill of the original, larger treen, being older than the division, consequently served also for the alia-treen.

Other keeill-less treens should be judged in the same way, e. g. Crosby Beg " Small Crosby " in relation to the neighbouring treen Crosby More or "Big Crosby", Cornay Beg to the neighbouring Cornay More and Smeall Beg to the neighbouring treen Smeall. Both Cornays, it is true, possess keeills, but one of them, Keeill Woirrey, being situated on an intack high up the Cardle Valley, may, in olden times, have belonged to the old treen Cornay before being divided.

Similarly the three Baldalls in Braddan only possess one keeill in common. The three Cranstalls in Bride have two, both belonging to the primitive treen Cranstall and thus being, no doubt, older than its division.

At the time when the old treen was divided, the keeill had played its part, being overshadowed by the large keeill or the parochial church developing therefrom, the new treen having no use for a keeill. Tllere can be no doubt that this division of the old treens to a great extent took place in Norwegian times, as indicated by the fact that not only the alia-treens but all treens, the names of which point to a division of an older been, bear Norwegian names: Gresby, Begode, Raby, Colby, Dalby, Gnebe, Sulby, Leodest, Coma, Smeall, Crosby, Cranstall, Baldall.

The parochial churches have grown out of old keeills one-time treen-chapels, in treens the names of which are either known or lost through the treens becoming Church land or Abbey demesne, as is the case with the parochial churches of Malew, Braddan, Maughold, and Lezayre. But in the case of the other parochial churches, the names of the treens can be determined almost with certainty.

The number of old treens having no keeills, is thus reduced from 83 to 57; but further reduction must, no doubt, be made. 'thus several treens, showing a marked deficit of quarterlands, have undoubtedly been segregated from older, larger treens, as for instance the four Renn-treens (Manx rheynn "division, part"), none of which has more than one quarterland.

Small treens bordering on Church land or Abbey demesne may have lost their keeills through parts of the old treens to which they belonged becoming Church or Abbey land. But it has been thought prudent not to deduct such treens from the number of old keeill-less treens, by now reduced from 83 to 53.

The result of our survey so far is: of the 159 keeills known in the Isle of Man, 124 are dispersed over 96 treens, while 35 are situated on Church or Abbey land. In 83 treens no keeills are extant, but at least 30 of these treens have never possessed keeills at all.

Were the treens in olden times provided with keeills according to a definite principle?

Of the 96 treens possessing keeills, no less than 71 or 74 per cent have each one keeill, 21 or 22 per cent two keeills and 3 or 2 per cent, three keeills. If one adds to these 96 the 13 treens the keeills of which have grown into parochial churches, no less than 84 out of 109 treens or 76 per cent have one keeill each, while about 19 per cent have two keeills each. If intacks are disregarded, 80 per cent have one keeill, and 16½ per cent two keeills each.

Our result so far strongly supports the tradition expressed in the ballad, according to which every treen had of old one keeill each.

But what about the remaining 53 treens without keeills?

It is perfectly clear that many of these treens must have once possessed keeills, and also that some of them never have, because of their originating at a time when the decline of the keeill had set in and new keeills were no longer constructed. Now a comparison of the number of quarterlands in treens with and without keeills, shows that the former are 1/4 larger than the latter. It appears that old treens, growing excessively large, gave rise to new, smaller treens which no one, at this late stage, would think of endowing with keeills.

These 53 treens may be divided into three groups:

1. old treens which of old had keeills now lost

2. secondary treens which have never possessed keeills

3. old treens which have never had keeills.

By equal distribution, each of these groups would obtain 17-18 treens each. There can be no doubt whatever about the existence of the two first groups, but, in our opinion, the third group certainly never existed. This small keeill-less group of 17-18 treens contrast too strongly with all the other old treens, numbering about 72 treens + about 30 which at an early date had become Church and Abbey land.

The conclusion necessarily is that all old treens which now possess no keeills, have lost them, and that in consequence all old treens have had at any rate one keeill.

It is hardly to be expected that this state of things would be faithfully preserved in any parish. The decline of the keeill setting in already in the 10th cent., it is almost a wonder that, one millennium later, it is still possible to point out at all events the foundations of 159 keeills in Man. When, nevertheless, several parishes have preserved the ancient tradition intact, no stronger support could be adduced in favour of the old intimate connexion between the treen and the keeill as advanced in the ballad. All old treens in the parishes of Malew, Patrick, and Braddan possess keeills, in those of Maughold, Arbory, Marown, and German all except one, in those of Lonan, Santan, and Ballaugh all except two, while in the remaining parishes the number of treens having no keeills varies from 4 to 9; most of the latter treens are situated in the fertile agricultural districts in the North where many keeills were doubtless ploughed over at an early date.

There can thus be no doubt that, from the first period of the Christian Church in the Isle of Man, every treen possessed at any rate one keeill.

In the 21 cases (24 including the intacks) where a treen has more than one keeill, no distinct difference of age is traceable between keeills within one and the same treen, whether situated on the same quarterland or not. They all give the same impression of great antiquity, they all have burial grounds attached to them, at least everywhere when the site has not been ploughed over or otherwise ruined. There is nothing to show that when a treen possessed two keeills, one has replaced the other, after its destruction or abandonment. The distribution of the keeills in different quarterlands does not support such a supposition. Even in the three cases where the keeills are situated on the same quarterland, they are not close to each other: on Ballachrink the distance between them is about 700 yards, on Ballahowin about 600 yards.

It follows that a treen of old may have had two keeills, in which case the keeills as a rule were situated on different quarterlands.

This conclusion does not, of course, conflict with the tradition of the ballad, the overwhelming majority of treens possessing in fact, in accordance with the words of the ballad, only one keeill each.

Keeill and Manor

What has determined the choice of quarterland when founding a keeill in a treen ? One would expect the keeill to be placed in the leading quarterland within the treen. And as this quarterland must be the same as that chosen later by the Norwegian treen-owner for his residence, it might be assumed that the keeill would preferably be on this very quarterland or on the Manor, as it might be styled here.

A scrutiny of the material concerned gives the surprising result that of 75 known keeills in 62 treens, 27 are situated on the Manor and 48 on other quarterlands.

This result is of great importance for the history of the treen. On the one hand, it shows a somewhat stronger attachment of the keeill to the Manor than to the other quarterlands. On the other, it definitely proves that the keeill is by no means bound to the Manor, and that on the contrary numerous Manors, undoubtedly the great majority, never possessed a keeill at all.

Is the comparatively high percentage of keeills on the Manors due to the Norwegian treen-owner? Do we know at all what his attitude was towards the keeill?

Yes, to a certain extent. So much we know that having given up their pagan fanes, the treen-lords made no small use of the old local keeills, allowing themselves to be buried there throughout the 10th cent., for several generations from the first appearance of Gaut Bjornson down to the beginning of the 11th cent. This is apparent from Norwegian cross-fragments discovered at the following treen-chapels: West Nappin, Ballachonly, Ballagilley, Balure, Rhyne, Ballaqueeney, St. John's. But 26 Norwegian sepulchral crosses from the parochial church-yards of Bride, Andreas, Jurby, Ballaugh, Michael, Patrick, Maughold, Braddan, and Marown prove that, already in Gaut Bjornson's days, it was customary for the Norwegian treen-owners to be buried within the precincts of the parochial churches. Nothing indicates that the Norwegian cross slabs belonging to the small local keeills are older than those from the large keeills, which subsequently grew into parochial churches.

Under these circumstances nothing supports the assumption that the Norwegian treen-owners ever built keeills of their own on the Manor. The seven local keeills where Norwegian treen-members were incontestably buried during the 10th cent., differ in no respect from the other ancient keeills situated on the Manors, all being older than the Viking Age. The keeill, in consequence, never formed an obligatory part of the Norwegian estate in Man. That is why, in the treens of Rhyne and West Nappin, two Norwegian treen-members, some time at the beginning of the 10th cent., were interred in burial grounds of keeills not situated on the Manors, which could not, therefore, have had separate keeills of their own. In agreement with our view, the keeills on the Manors, in no less than 7 of 27 cases, are proved by archaeological finds to be older than the Viking Age.

During the 9th cent. the Norwegian treen-lords probably still had their pagan fanes in Man; but like the big hofs in their home country, one fane probably served many treens. This may explain — if an explanation be required—why the Norwegian treen-owners, after their conversion, did not found keeills on the Manors, but as a matter of course looked upon the large keeill (the later parochial church) as the common Christian hof. It may also explain the important position which these large keeills assumed at such an early date.

The Christian burial custom must have been early adopted by the Norwegians, as so very few pagan finds are extant. The tumulus on Knock y Doonee was erected sometime in the 9th cent., over a pagan Norwegian treen-lord who was buried there in his boat. Mr. B. R. S. Megan[sic Megaw] of the National Manx Museum has kindly placed at our disposal a list of finds of possible Norwegian origin (see p. 384 ff.).

It may thus be taken for granted that the relatively high percentage of keeills situated on the Manors is not due to the Norwegian treen-lords. Nor should it be ascribed to Pre-Norwegian social conditions in the Isle. The explanation is certainly to be found in the size of the Manor, which, on an average, exceeds by more than 1/4 the ordinary quarterlands. The Manor apparently extended its original area, whereby several old keeills, originally belonging to other quarter-lands, automatically became merged in the Manor. This extension of the Manors would be due to the Norwegian treen-lord and many of the old keeills were probably already in ruins when it took place.

Treen, tirung, urisland, davagh

If the keeill is not linked either with the Manor or- a special quarterland, the reason is that the Manor did not originate until the coming of the Norwegians, while the treen and keeill system has its roots deep in a Celtic social order which was much older than the Viking Age. It only to a limited extent admitted of private ownership of land; all land belonged to the tribe or family, the cultivated areas being distributed anew at periodical intervals. It is well known that this runrig or common field system, in a modified form, survived in Man to comparatively recent times. The Norwegians are sure to have encountered it elsewhere also: in the Shetlands and Orkneys, in Scotland, the Hebrides, and Ireland. The main features of society being to such a great degree the same over such extensive parts of the British Isles, it would only be natural if equivalents to the Manx been-system were to be found elsewhere in the isles. In an excellent, concise paper: Leidang in the West, Proc. Ork. Ant. Soc. N11 1934, with the general trend of which I concur, Hugh Marwick draws attention to the striking similarity existing between the Manx treens, the Orkney urislards and the West-Scottisll tirungs. Like the treen, the urisland and the tirung also serve as fiscal units, divided into quarters, Gaelic ceathramh = Manx kerroo, in the Orkneys named skaflands. Tax and rent of an urisland amount in the Orkneys to £4, of a tirung and davagh in the Hebrides and West-Scotland to £3—8—0. This agrees well with the Manx treen which, on an average, paid £3—10—0 each. Like the been, the urisland has one chapel, sometimes — but rarely — two.

The linguistic terms treen and tirung seem identical, the Scotch Gaelic tirung originating from tir unga, the Manx treein (with a palatal n) from tir uinge with a palatal ng well known from MiddleIrish manuscripts; for the development of palatal ng into palatal n, cp. Iuinie (Bishop Phillips), i.e. Ihuinge, gen. of Ihong "ship".

Treen, like tirung, is thus a translation of the Norwegian term eyrisland. It is significant that the Gaelic translation is confined to the Hebrides and Man, whose history is so closely linked together, while in the Orkneys the original Norwegian term urisland survived. This circumstance also bears witness to the predominant position of the Norwegian language in the Orkneys and the increasing influence of Gaelic in the Hebrides and Man

It is hardly a mere coincidence that the term eyrisland (urisland, tirung, treen) has spread over the area of the Orkney Earldom: Man, the Hebrides, Caithness, the Orkneys. It is certainly possible that the term was introduced into Man under Earl Sigurd at the close of the 10th cent., possibly even earlier. That the treens in the Isle of Man were in Norwegian possession as early as 875 may be concluded from the fact that Gudrød Crovan in the year 1075 deprived the treen-owners of their Udal: but in Norway no udal could be acquired before the lapse of six generations.

The term eyrisland does not originate from the leidang (or naval levy), one eyrir (unga) per treen being a far too insignificant assessment. The term must allude to a different tax, probably the veizla, paid for the maintenance of the king and his retainers and apparently identical with the skat termed wattle in the Orkneys and, assessed on each urisland at the rate of 1d. per pennyland. An administrative division founded on the veizla would form a natural basis for the leidang —or vice versa. The leidang may perfectly well be older than the veizla or of the same age.

The urisland, the tirung, and the treen have the appearance of having emanated from one and the same fiscal policy. Probably they also emanated from one and the same political organization.

Outside the domain of the Orkney Earls, the term davagh, in the North of Scotland, is found to correspond to urisland and tirung. Davagh is a Gaelic term (identical with Manx doagh "a vat" though the underlying idea is not apparent) and no doubt older than the Norwegian term eyrisland (treen) by which, to a great extent, it was supplanted.

The Irish conquest of the Isle of Man

The treen-system is, as shown here, centuries older than the Viking Age. Is it also older than the Irish colonisation of Man? In other words: has it its roots in a Celtic or a Pre-Celtic community ?

The Celtization of Man is always associated with the conquest of West-Scotland and the Hebrides by the North-Irish tribe Dál Riada in the 5th and 6th centuries. But several circumstances go to indicate that the island had become Gaelic long before. While Dál Riada was a Christian tribe already in the 5th cent., the two Ogham inscriptions discovered in a prehistoric burial ground, at Ballaqueeney (Rushen), could not possibly be ascribed to Christian Gaels. Of the inscriptions, both of which are ancient, one reads: B I V A I D O N A M A Q I M U C O T C U N A V A [ L I ], i. e. " (this stone was erected over) Bivaidu of the tribe of Conall". Now, the tribe-name, corresponding to the eponym Cunovalos, would in Ogham be Curzovalinion, concurring in all respects with Irish Conailne, Conaille, the name of the famous tribe in Louth, South Ulster, also named Conaille Muirthemne and emanating, as tradition goes, from Conall Cernach, the hero of the Red Branch, who is presumed to have lived at the beginning of our era. The Ogham inscriptions of Ballaqueeney thus pointing to an influx from South Ulster to the Isle of Man, it may not be amiss to call to mind that such an immigration is in fact mentioned in the Annals of Tigernach for the year 254. The Imchad mac Rochado mentioned in the Ogham inscription of knock y Doonee, Andreas, must also in some way be connected with the two brothers Imchad and Rochaid, the sons of Colla Da Crich and nephews of Colla Huais, Monarch of Ireland A.D. 327 to A.D. 331, the progenitor of the Clann Cholla from whom sprang the Lords of the Isles. The Ogham of Knock y Doonee, therefore, cannot in all probability be later than the 4th cent. The inscriptions of Ballaqueeney and Knock y Doonee may be described as the most unique and important historical monuments any people possesses, being original documents, removed in time by only a few generations from the great events that made the Isle of Man a Celtic community, and older by far than Fergus Mor's conquest of Scotland.

The Ballaqueeney inscription raises the Pictish question anew. Conaille Muirthemne belonged, with Antrim, Down, and parts of Meath, to the Pictish domain in Ireland. In the Chronicle of the Picts and Scots all the Conailli of Ireland are said to descend from the Picts of Dál Riada (and thus far the expansion of the Conaille Muirthemne to Man may be looked upon as a precursor of Dál Riada's conquest of West Scotland). From a linguistic point of view, the Pictish question is not a question of either Celtic or non-Celtic language. Even if the Picts originally spoke a non-Indoeuropean tongue, some tribes very likely early adopted the Gaelic language of the Irish tribes with whom they entered into such intimate contact.

One may thus, without fear of being in error, assume that the eyrisland-division in Scotland and Man has its roots either in a Celtic or a Pictish social system. Unfortunately, very little is known of the social organisation of the Picts. But as it does not seem possible to point out a Celtic equivalent to the urisland or davagh, Pictish origin seems a priori more likely; only it must be remembered that as far as the North of Ireland is concerned, Pictish and Irish in olden times would in many cases mean the same thing.

The treen as an ecclesiastical unit

As the Norwegians utilised the treen for their fiscal purposes, the Irish Church had, centuries earlier, turned it to account for their special ends, the treen becoming an ecclesiastical unit with a chapel as its centre. How may we explain that, almost from the first day of Christianity in the Isle, oratories and chapels arose in every treen?

Very likely the treen apart from being a social unit, also formed a strong religious unit with an extremely localised cult. It is no far fetched idea that the Irish missionaries should have found themselves face to face with an ancestor-worship deeply rooted in the life of the tribe; Bivaidu, no doubt, worshipped Conall Cernach, the progenitor of his tribe, as stated by himself on the Oghan stone of Ballaqueeney. The been-chapels served as a counterpoise to the local places of worship and were probably often constructed on the very same sites. As in the Orkneys practically all urisland-chapels have been built over a brooch or other prehistoric monument, so in the Isle of Man no less than 24 keeills [see list p398] can be proved to be built on places of pagan worship, while in numerous cases the keeills are situated in the vicinity of holy wells which in pagan times were the object of worship.

Parish, skibreide, leidang

The tendency of the large keeills (the later parochial churches) to outshadowv the local keeills, manifests itself already prior to the Norwegian period. In this early phase, there is of course no question of parishes, parochiae, but old finds prove that several keeills, on the site of which parochial churches were later built, had a dominating influence long before the 10th cent. as compared with the local keeills in the neighbouring treens. This holds good of the four keeills of Maughold, Conchan, Braddan, and Marown, that later grew into parochial churches. These 4 parishes being situated on a continuous line, there can be no doubt whatever that the districts of the four large keeills mentioned agree with the later parishes.

In the last resort, the parochial division in Man, as so often elsewhere, seems to be based on an older secular, one may say pagan, district division, which the Church utilised for its organisatory purposes.

In support hereof several considerations may be quoted.

In the first place, the finds of the 5th to the 7th cent. preferably group themselves around the keeills which in the 13th cent. appear as parochial churches. This, it is true, does not take us back to pagan times; but it is a reasonable assumption that these keeills were built in places of secular importance.

In the second place, all parishes (except Marown) stretch down to the sea, a feature more easily explained by the requirements of naval defence than of ecclesiastical administration. The peculiar shape of the parishes of Braddan and Lezayre is best explained by the assumption that they were old secular districts which during the Norwegian domination were given access to the sea.

Finally, the composition of the House of Keys also points to the existence of old administrative units, agreeing with the later parishes. From the famous declaration of the 24 to Sir John Stanley of the year 1412, it is evident that the Hebrides and Man, during the Norwegian period and previous to the year 1156, when the political connexion was severed, possessed in common a thing or court of law, in which Man was represented by 16 members, "the Out Isles" by 8. This court may date back to Gudrød Crovan who took possession of the Isle after the battle of Skyhill 1079, or it may be ascribed to Earl Sigurd to whose reign we are inclined to date the term eyrisland. But it can hardly be older; for Harald Harfagre's conquest of Man and the Isles was of too transient a nature to leave lasting traces in the administration.

The sixteen Manx representatives at this Althing of Man and the Isles were certainly chosen from the proceres, principes, or optimates, so frequently mentioned in the Manx Chronicle and referring, not to the ordinary treen-owners whom Gudrød had deprived of their udal, but to men of trust administering certain districts under the king of Man and the Isles. And these administrative districts cannot have differed from those which in the 13th cent. emerge as parishes. As in the Orkneys, the parishes in the Isle of Man were also represented at the thing by one deputy each, like the goðorðs on Iceland.

It follows that one of the 17 Manx parishes must have arisen — probably after 1156 — through the division of a large parish. The fact that Marown, alone among all Manx parishes, does not reach the sea, points to its once having formed one large parish with Santan. These parishes both count among the smallest in the island, Santan in fact being the smallest of all. Together they make 10,833 acres, about the same area as the neighbouring parishes of Braddan, Patrick, and German. While the northern and southern parts of Man were divided into sheadings of three thirds ( þriðjungs) each, the central belt of the island was occupied by the two sheadings, Glenfaba and Middle, divided into two halves each. The reason for this disparity is doubtless to be found in the leidang, Glenfaba and Middle showing the lowest leidang in the Isle, only half that of Rushen and Ayre.

In entire agreement with this conception of the parish as the skibreide, the Watch and Ward was from olden times organized parish by parish. The night-watch stations in all probability originated from the stoðvar or landing-places where the leidang ship was beached, their aim being to protect the leidang ships against such surprise night attacks as that in Peel in the winter 1228, when King Ragnvald set fire to the entire Manx fleet. The harbours of the Norwegian leidang fleet are, in consequence, all known.

The Norwegian term for skibreide was in the Isle of Man skirt, a word borrowed from England where it was used in the same sense. With the establishment of the parish division in the 12-13th cent., the word in Manx and Scotch-Gaelic came to denote the parish.

The parish thus being the skibreide, the quarterland, not the treen, must correspond to the lide, the district responsible for the supply and equipment of one man. No parish comprised at the utmost more than 21 eyrislands, and small leidang ships of 21 men never existed.

With every quarterland supplying one man, it should now be possible to determine the strength of the Manx leidang. The number of treens, which at the beginnng of the 16th cent. is 176, must have been considerably higher in the 11th, as many treens had been lost by becoming Churchland or Abbey demesne. Though the names of the latter treens are long ago lost, the number of the quarterlands they comprised is known, totalling 1294/5, i. e. 32½ treens. That this calculation is approximately correct, is proved by the 41 keeills situated on the Church and Abbey lands and presupposing 34 treens in accordance with the proportion 177 keeills to 147 known secular treens.

Calculating with normal treens of 4 quarterlands, the Isle of Man would thus have provided a leidang of 832 men, say 800 men, as several treens are no doubt later than the Norwegian period. In support of this, reference may be made to the tax of 800 marks, imposed by king Ragnvald on the occasion of his proposed visit to the English Court in the year 1224. This tax seems to have been assessed with one eyed (uncia) per quarterland (100 marks = 800 eyrar).

The strength of the leidang varies considerably from parish to parish. Thus while Malew supplies 84 men, Andreas 64, and Rushen 62, Arbory only provides 36 and Jurby 28 only. It follows that in the Isle of Man, as in Norway, the leidang system presupposes skibreides of varying leidang.

The Manx leidang fleet may have numbered about 20 twenty sessers. But this estimate is of a rather theoretical nature, as the ships were certainly not all of the same size. Jurby would by itself only be able to provide a 13 sesser, the smallest type of leidangship. Probably, therefore, Jurby did not supply a ship alone; either it joined the neighbouring skibreide Ballaugh, or the leidang was assessed for the sheading and levied on the parishes or the treens.

The sheading

While the skeerey (or parish) and the treein continue, as shown above, secular divisions older than the Viking Age, the shearing seems to be a new formation, due to the Norwegian leidang-organization. Every sheading had a thing, or court, attached to it. At this court which met twice a year (a circumstance that may suggest a payment of the leidang in two instalments, spring and autumn, as in Norway), the parties pleaded before a jury of 10 members. After concluding the ordinary lawsuits, the thing constituted itself as a Registry Court, settling all disputes regarding land and tenants.

Comprising three skibreides and dealing with all matters concerning land and leidang, the sheading seems to be the equivalent of the Norwegian district of assessment which in olden times was a permanent institution from Agder to Namdal. It is quite possible that, in Man, the leidang was assessed on the sheading as it was in Norway on the district of assessment; and levied on the treens irrespective of the limits of the skibreides. In support of this one might emphasise the marked variability of the leidang from parish to parish, as compared to its greater stability in the sheadings.

As proved by Gregor IX's edict of 1231, mentioning the Ecclesia Sanctae Trinitatis in Lezayre, the sheading division dates back, at any rate, to the beginning of the 13th cent. But no doubt it is much older. The sheading of Rushen, it is true, cannot have got its name before 1134, the year when the monastery of Rushen was founded by Olav Bitting, for it was the famous monastery which gave its name to the sheading, not the contrary. But it does not necessarily follow that the sheading division is no older than Olav Bitting; for as the name of the sheading, Rushen may have superseded an older name. In all probability the sheading division is just as ancient as the leidang-institution itself in the Isle of Man.

The term sheading, or sheathe, sheetin, has developed from Old Norw. séttungr meaning "a sixth part". It cannot come from Norw. settingr, which would have passed into Manx shetchen (phonetically written ƒet ƒan). But supposing settungr in Manx, under the influence of rheynn "a part", became a feminine, the development of settungr to sheetin, sheaths is quite regular.

Few peoples possess monuments of such a fundamental historical importance as the ornamental Norwegian crosses, the keeills, and the Ogham inscriptions of Ballaqueeney and knock y Doonee.

Thanks to P. M. C. Kermode, the crosses are safe. His excavations of the keeills have brought to light historical materials of the first order. Tynwald should see that his work has not been in vain. All keeills must be fenced and the Manx Museum granted absolute authority over monuments and finds of national importance. Caveant consoles . . .



List of Keeills

Any comments, errors or omissions gratefully received The Editor
HTML Transcription © F.Coakley , 2000