Main Menu - Misc. - Clothing/Textiles - Medieval Wales - Names - Other Medieval - Publications - Harpy Publications

This page last modified April 3, 2006

Snapshot of a Cantref: The Names and Naming Practices in a Mawddwy Court Roll of 1415-16

by Heather Rose Jones
copyright © 2001, all rights reserved


The material in this study comes from "A Mawddwy Court Roll, 1415-16" edited by Keith Williams-Jones in the Bulletin of the Board of Celtic Studies vol. 23 (1970). The original document is a record, in Latin, of the details of various legal actions, and it is the original document that has survived, not a later copy. The nature of the document means that we have a presumably random sampling of the names of the region, but not an exhaustive list. Conversely, we have multiple references to the same individual in many cases, which must be dealt with in some fashion for statistical purposes. The study here makes a useful contrast to that of the Anglesey Submissions of 1406 -- being from the same general region and period. It is unfortunate that the lands of Mawddwy were not included in the 1292 Merioneth Lay Subsidy Roll (Mawddwy was not incorporated into Merioneth until a later period) and so we can't compare names from the same region a century earlier.

This article assumes a familiarity with previous work I have done on medieval Welsh names and naming practices. While all the data from the document is given, in many cases my discussion will focus only on those areas that show some particularly interesting feature, or one that this data-set demonstrates particularly well.

In analyzing statistics, except where I make it clear that I am talking about all the data, I have used only the fullest form of the name given for each person who appears multiple times. I have not attempted to control for individuals who appear multiple times in patronyms (e.g., the father of two brothers who both have entries) as this analysis is much less certain and I have not performed a similar control in previous studies. It is possible that I have counted individuals more than once if different references were not simply subsets of the fullest form, or that I have conflated individuals whose names were identical in their fullest form. When an individual is referred to several times in the same legal case, I can have a high confidence level that I have not erred, however if the individual appears in various cases in the document, matters are less certain.

I have changed my usual format and left discussion of the name patterns for last.

Names of the Ruling Family

The main analysis in this article omits a reference in the document to members of the ruling family of Mawddwy at the time. That is:

"Hugo Bourgh dominus de Mouthoy et Issabella uxor eius, filia Iohannis de Mouthoy ac soror et heres Fulconis filii dicti Iohannis domini de Mouthoy ultimi defuncti"

"Hugo Bourgh, lord of Mawddwy, and Issabella his wife, daughter of Iohannes de Mawddwy, and sister and heir of Fulco son of the said Iohannes, lord of Mawddwy, the last [i.e., Fulco] dead"

One thing that is noteworthy here is the overwhelming Englishness of their names. Hugh Burgh (to use a normalized form) was a first cousin of Owain Glyndwr (via their mothers). Iohannes de Mawddwy was a descendent of the native princes of Powys. And yet no linguistically Welsh element remains in their names at all.

Given Names

For each name heading, I have given the total number of occurrences in each "generation" (as determined solely by the format of the names), first for all entries, then for unique individuals (as defined above). Under each heading, statistics for each spelling are given for all entries. I have not discussed the names themselves except for questionable or unusual examples.

The Bold Face names are modern spellings included for reference only. The written forms that actually occur in the document appear in

Some writen forms represent Latin declensions and are noted as such.


The feminine given names appear only as the names of the cited individuals -- i.e., none appear in metronyms (although see the section on bynames for a probable metronym). Nine women share six names between them, a proportion that is fairly typical for the data-sets I have studied with this small a sample. Four of the names here also appear in the Anglesey data (Angharad, Gwerfyl, Gwenhwyfar, and Lleucu). A slightly different set of four overlap with the Merioneth data from a century previous (Angharad, Dyddgu, Lleucu, and Gwerfyl -- which all fall in the top ten most common names in that data). It is particularly noteworthy that all of these names are of Welsh linguistic origin.

Angharad all: (4); unique (2)

Dillena all: (4); unique (1) -- This is a name I have not previously encountered. The final a, when it occurs, is always an expanded suspension mark, and is presumably a Latin declensional suffix. Since the manuscript occasionally has ll for [l] in Welsh names, we have no certain clue to the quality of the sound meant here. It is possible that this is the same as the feminine given name Dulon which appears in the Old Welsh charters in the Book of Llan Dav, although the vowels would be difficult to explain. It is worth noting that there is a Medieval Welsh word dillyn meaning, as an adjective, "beautiful, fine, neat, chaste", and as a noun, "a thing of beauty or elegance, ornament, precious thing, dear one, darling". Although no feminine form dillen is attested in the GPC, it could easily be produced by analogy to other adjectives and nouns having masculine/feminine pairs ending in -yn and -en respectively. (In the 15th century, there is an example of a feminized dillynes used for "beautiful woman".) This may be the best explanation of this otherwise-unknown name. I can find no Anglo-Norman name that it might be a borrowing of.

Dyddgu all: (2); unique (2)

Gwerfyl all: (1); unique (1)

Gwenhwyfar all: (5); unique (2)

Lleucu all: (1); unique (1) -- Despite the unusual spelling, I am confident of the identification of this name.


A discussion of the overlap of the name pool with previous studies can be found below with the frequency statistics. There are only a few names of non-Welsh linguistic origin: Adam, David, John, Laurence, Philip, and William. Of these, Adam, David, and the vast majority of the forms of John show linguistic evidence of being pre-Norman borrowings. Philip was already quite popular a century before in Merioneth, and William has jumped significantly in popularity since then. Laurence also appears in the contemporary Anglesey data, but, as here, in a single example.

Addaf all: (0,2,2); unique (0,2,1)

Cadwgan all: (0,4,1); unique (0,3,1)

Dafydd all: (32,22,5,1); unique (16,11,14,1)






Dewrwas all: (1); unique (1) --This name has more the "look and feel" of a descriptive byname (it is a compound translating as "brave lad"). The earlier Merioneth data has a number of bynames of the form "Gwas <adjective>", although this particular combination of elements does not occur there. If "Dewrwas" does originate in a byname, it is impossible to determine here whether the byname is being used as a baptismal name, or whether the individual is simply habitually referred to by his nickname rather than his baptismal name. (See the discussion of "Y Mab Lloit" below under bynames for a probable example of the latter.)

Ednyfed all: (3,1,0,1); unique (2,1,0,1)

Einion all: (3,6,8,1); unique (2,5,5,1)

Evadi? all: (3); unique (1) --I don't recognize this as a name I have encountered before. One of the citations is from a list of jurors (which makes it clear the name is masculine) in a clearly nominative form (so we aren't dealing with a Latin genitive).

Griffri all: (0,2,0,1); unique (0,1,0,1)

Gruffudd all: (23,3,3); unique (11,2,3)




Gwrgeneu all: (0,0,3); unique (0,0,1)

Gwilym all: (10,2); unique (4,4) -- In contrast with Ieuan/Iohannes (see below) the "Welsh" spelling Gwill(y)m and the English William or Latin Willelmus are not treated as distinct names, and in two cases both types of spellings are used for the same individual.

Gwyn all: (0,0,3); unique (0,0,1)

Hocnai? all: (1); unique (1) -- From the context, this is clearly a given name. It occurs with no byname or gendered reference, so even assigning it as masculine is a guess.

Hywel all: (3,0,4); unique (2,0,3)

Iorwerth all: (5,4,2); unique (3,2,2)

John/Ieuan all: (33,25,2,1); unique (16,14,2,1) -- The Welsh form Ieuan and the Latin Iohannes are treated as distinct names. One case involves Ieuan ap Griffri Vaughan and Iohann[es] ap Meredith ap Gr. ap Griffri and in the several references to them in the discussion of the case, the language-context of each is consistant. The other man found with the Latin form of the name is "Iohannes Laurence" who may, in fact, not be Welsh.

Ieuan all: (26,25,2,1)

Iocws all: (1)

Iohannes all: (6)

Ithel all: (0,3,1); unique (0,1,1)

Laurence all: (0,1); unique (0,1)

Llywelin all: (10,4,4); unique (6,3,4)

Madog all: (16,18,1,1); unique (7,12,1,1) -- As Morgan & Morgan note, "Badi" is a hypocoristic form of Madog. See the discussion below on patronyms based on bynames.

Badi all: (0,2); unique (0,2)

Meilyr all: (0,0,0,1); unique (0,0,0,1)

Maredudd all: (1,7,2); unique (1,3,2)

Meurig all: (0,0,0,1); unique (0,0,0,1)

Phillip all: (0,1,1); unique (0,1,1)

Rhys all (1,1); unique (1,1)

Rhiryd all (0,0,3); unique (0,0,3)

Tegwared all (6); unique (2)

Tudur all (3,5,1); (1,3,1)

Frequency Statistics for Given Names


There is little to say about the statistics for the feminine names. The following table gives the names in order of frequency (by number of individuals), with the second column giving the total number of citations from the data. Columns have been added showing the frequency ranking from the 1406 Anglesey data and the 1292 Merioneth data for comparison. (Only rankings 1-10 are shown by number, other names that appear in these data-sets are indicated with an "x".)

Name Indiv. Total Angl. Mer.
Gwenhwyfar 2 5 x
Angharad 2 4 3 1
Dyddgu 2 2 4
Dillena 1 4
Gweirful 1 1 1 9
Lleucu 1 1 3 4


The following table is arranged as above except that rankings 1-30 are indicated.

Name Indiv. Total Angl. Mer.
1. John 33 54 1 6
2. Dafydd 32 60 2 2
3. Madog 22 38 3 1
4. Gruffudd 16 29 7 14
5. Einion 13 18 5 3
6. Llywelin 13 18 11 19
7. Gwilym 8 12 17 28
8. Iorwerth 7 11 4 4
9. Maredudd 6 10 13 x
10. Hywel 5 7 6 21
11. Tudur 5 9 9 22
12. Cadwgan 4 5 11
13. Ednyfed 4 5 18 17
14. Addaf 3 4 12 5
15. Rhiryd 3 3 x 13
16. Griffri 2 3 29
17. Ithel 2 4 15 8
18. Phillip 2 2 19 12
19. Rhys 2 2 24 x
20. Tegwared 2 6 14 15
21. Dewrwas 1 1
22. Evadi? 1 3
23. Gwrgeneu 1 3 x 18
24. Gwyn 1 3 28 9
25. Hocnai? 1 1
26. Laurence 1 1 x
27. Meilyr 1 1 30 16
28. Meurig 1 1 21 26

Not surprisingly, there is a large amount of overlap in the name pools of the three samples. The only names in this article that do not occur in either other sample are the three (Dewrwas, Evadi, Hocnai) about which there is some uncertainty. In general, the name pool here shows a greater correlation with that of the contemporary Anglesey data, rather than with the Merioneth data, which is from a closer region geographically, but a century earlier. This suggests that overall chronological patterns in name popularity thoughout Wales may be stronger than patterns in regional preference. Of the eleven most common names from Mawddwy (since there is a tie for tenth place), eight are also in the top eleven from Anglesey, while only five are also in the top eleven from Merioneth. Looking at the top 20 most popular names from Mawddwy, the pattern is quite similar: 16 are also in the top 20 from Anglesey, and 14 in the top 20 from Merioneth -- a slightly higher rate than for the top eleven.


The following list shows the data for bynames (of all types) in a similar format as was used for given names (with respect to the number of total citations versus individuals, and to what "generation" of the name the element occurs in). The bynames that follow immediately after given names occur only with masculine given names. There is one byname appearing alone in a patronym that, if I have interpreted the origin correctly, would appear to refer to a woman -- providing not only the only example of a feminine byname in this data, but the only example of a metronym as well. Examples preceded with an asterisk appear alone (i.e., without a given name) in a patronym. (See the discussion of this topic below.)

Bach "small" all: (1); unique (1)

Bychan "small, junior" all: (3,6); unique (1,5)

Ceisiad "servant of the peace; tax-gatherer" all: (0,0,1); unique (0,0,1)

Cethin "dusky" all: (1,3); unique (1,2)

Coch "red" all: (2,9); unique (1,5)

Cog "cook" all: (0,0,3); unique (0,0,3)

Chwith "left-handed" all: (0,2); unique (0,2)

Crydd? "shoemaker" all: (1); unique (1)

Du "black" all: (1,7); unique (1,3)

Gof "smith" all: (0,1); unique (0,1)

Gwreang "young man, page, squire" all: (0,2); unique (0,1)

Hir "tall" all: (0,2); unique (0,1)

Llwyd "brown, gray" all: (7,4,2); unique (3,2,2)

Mab... "the gray/brown lad of? the wrong-head" -- This is probably the oddest personal identification in the document. No given name is included, simply the phrase as shown below. The Merioneth data includes four people identified solely by a name of the form "Map <adjective>", two of which use color-adjectives as here. Particularly with the inclusion of the definite article here, it would be hard to interpret this as anything but a personal nickname being used in preference to a baptismal name (as occurs somewhat more often in patronyms). The second part of the entry (irbengam) appears to contain the element pengam (stubborn), which occurs as a byname in the Merioneth data. The initial ir- could be the definite article (yr), although one would expect a reduced form (y) here, and the lenition (pengam > bengam) is hard to account for.

Mamaeth? "foster-mother, wet-nurse" all: (0,0,2); unique (0,0,2) --Although this appears in the entries for two different individuals, the person being referenced is almost certainly the same. (Both individuals are sons of Cadogan ap y Vameth.) The most likely origin for this byname is mamaeth ("foster-mother, wet-nurse"), an interpretation that is reinforced by the lenition after the definite article (mamaeth > y famaeth) which is best explained by a feminine noun. Assuming this interpretation is correct, this is the only feminine byname in the data as well as the only metronym.

Moel "bald" all: (3,2); unique (1,1) --The individual indentified as "Moil" in a patronym also appears (in a patronym) as "Moel Gloff" (bald & lame).

Pencilli place name all: (1); unique (1) --This is the only clearly locative byname in the data (other than among the ruling family). Interestingly, if I have interpreted the Latin correctly, he is the only villain mentioned in the document. (This is, of course, insufficient data for any conclusions about naming practices of villains versus free men in this context.)

Person "parson" all: (0,1); unique (0,1)

Porthmon "porter" all: (0,0,1); unique (0,0,1)

Syw "beautiful, wise" all: (0,1); unique (0,1) --I am not entirely confident of my interpretation of this byname, due to its rarity, but another example can be found in the contemporary Anglesey data.

Terfyst? unknown all: (0,1); unique (0,1) --I have not yet found a plausible origin for this byname.

Toppan? unknown all: (0,3); unique (0,3) --the Anglesey data also has an example of "y Toppan", but I still have no good guess as to its origin

Frequency Statistics for Bynames

These are based on the "fullest" record for each distinct individual. As with the given names, this inflates the statistics for some elements. For comparison purposes, the rankings of these elements from the Anglesey and Merioneth data are also given.

Rank Name # Angl. Mer.
1. Bychan 7 6 4
2-3. Coch 6 2 2
Llwyd 6 3 5
4. Du 5 1 1
5. Cethin 3 15 19
6-8. Chwith 2 5 6
Moel (Gloff) 2 7 3
y Famaeth 2
9-21. Bach 1 9 x
Ceisiad 1
Cog 1
Crydd 1 18 7 [Lat. "sutor"]
Gof 1 8 11 [Lat. "faber"]
Gwreang 1
Hir 1 24 x
Pencelli 1
Person 1 16
Porthmon 1
Syw 1 x
Tervist 1
Toppan 1 x
Y Mab Loit... 1

Unsurprisingly, there is a consistency of popularity among the most popular bynames. The eight most popular bynames here overlap each of the eight most popular in the other two data-sets by six items. The four most popular here overlap the other two top four lists by three items each. Cethin is the surprise popular name here, but with a count of three, this may not be significant. The contemporary Anglesey list shows greater overlap overal (particularly if you don't count the correspondence of Welsh occupations here with Latin ones in the earlier document).

Correlation of Name Frequency with Byname Use

One question that I don't believe I've examined directly before is a correlation between the frequency of a given name, and the number of people bearing that name who have a non-patronymic byname. This is similar to the statistical study I did for the Anglesey records comparing the overall name complexity with the given name frequency. Assuming that bynames are distributed relatively randomly with respect to the given name, it is not surprising to find the general correlation that appears: more individuals = more bynames. While it would be difficult to draw strict statistical conclusions, it does appear that bearers of the most popular name (Ieuan) are disproportionately more likely to have a non-patronymic nickname (in addition to whatever patronyms they may bear).

Name # bynames name pop. #B/#G
1. Ieuan 11 1 0.33
2. David 6 2 0.19
3. Madog 6 3 0.27
4-6. Gruffudd 3 4 0.19
Llywelyn 3 5 0.23
Einion 3 6 0.23
7-8. Iorwerth 2 8 0.29
Rhys 2 17 1.00
9. Evadi 1 21 1.00
Total 0.23

The second most popular name, David, appears to have a different method for differentiating: the use of hypocoristic forms of the given name (Dai, Deio, Deicws, Daiwyn). If hypocoristic names are included in the above analysis (including Iocws for Ieuan and Badi for Madog), we get the following. In this comparison, we see much more clearly that, for the most popular given names, a greater proportion are given some additional means of distinguishing one from another. (Rhys and Evadi throw off the general pattern due to the quirks of low sample numbers.)

Name # bynames name pop. #B/#G
1. David 16 2 0.50
2. Ieuan 12 1 0.36
3. Madog 8 3 0.36
4-6. Gruffudd 3 4 0.19
Llywelyn 3 5 0.23
Einion 3 6 0.23
7-8. Iorwerth 2 8 0.29
Rhys 2 17 1.00
9. Evadi 1 21 1.00
Total 0.31

Patronyms Formed from Bynames Alone

This data-set shows a strong pattern of something I have seen only as an option before. Eighteen of the citations (nine different individuals) include a patronym "generation" based on a byname with no given name present. E.g., Mereddith ap y Person (Maredudd son of the parson). There are a number of striking trends in these examples (which do not necessarily hold in other data-sets with similar examples):

Except for moel, there is no overlap between bynames that appear in this type of patronym and those that don't.

Except for moel, all the identifiable meanings of these bynames are occupational in a loose sense of the word. (Toppan has an unknown meaning at this point.) This construction comprises the majority of the occupational bynames used, although a couple occupational bynames also occur in the more usual construction (in combination with a given name).

Except for moel" (there's a pattern here), the definite article is used in all these constructions (i.e., ap y <occupation>).

Patronym units of this type only occur as the last in a string of generations. (Four are the second unit in a two-generation string; five are the third in a three-generation string.)

Very closely related to this construction is one used with a hypocoristic form of a given name. The two examples of ap y Badi represent this type. Note that the definite article is also used in these.

The names containing units of this type (using only one for each individual) in this data-set are:




Name Patterns

In discussing name patterns, I have used only the fullest (i.e., most complex) form of each individual's name. A later section discusses the ways in which names are simplified when multiple references to the same individual are made in the text.

The following patterns are found in the data (including both men and women). As in previous studies, I have used the following coding system:

One "generation"

g 1
gb 5
gl 1
gg 1 (Iohann[es] Laurence -- this might be a patronym or it may be an English-style hereditary surname)
bb 1 (Y Mab Lloit irbengam -- hard to analyze but it appears to have at least two independent byname-like elements)

Two "generations"

gae 2
gag 9
gagb 17
gago 1
gag? 1
gbagb 2
gab 1
gabb 1
gao 1
ga? 1

Three "generations"

gagag 23
gagagb 2
gagbag 1
gbagag 3
gagab 1
gagao 3
gagbao 1

Four "generations"

gagagag 6
gbagagag 1

There are a number of angles from which to look at this data. There is the question of patterns of complexity, of how that complexity expresses itself, and of what types of elements appear where and in what numbers. The last shows some strong trends in this data. The following table shows the patterns arranged by number of generations and presence of various types of elements.

  1 gen 2 gen 3 gen 4 gen total
G only g 1 gae 2 gagag 23 gagagag 6  
gg 1 gag 9          
total   2   11   23   6 42
B present gb 5 gagb 17 gagagb 2 gbagagag 1  
bb 1 gbagb 2 gagbag 1      
    gab 1 gbagag 3      
    gabb 1 gagab 1      
total   6   21   7   1 35
O present     gago 1 gagbao 1      
    gao 1 gagao 1      
total       2   2      
L present gl 1             4
total   1             1
unknown type     gag? 1          
    ga? 1          
total       2         2
all bynames   7   25   9   1 42
Total   9   36   32   7 84


The names are divided exactly evenly between those containing non-patronymic bynames, and those that don't. Only 6% of the whole contain more than one non-patronymic byname and none has more than two. This compares very closely to the Anglesey data, where 48% of the names contain at least one non-patronymic byname, and 5% contain more than one. (0.1% of the Anglesey names contain three non-patronymic bynames, none more than that. If the Mawddwy data were precisely proportional, we would expect to find no names with three bynames -- as is the case.) This is a higher incidence of bynames than in the Merioneth data from a century before. There only 37% of all names (looking only at men's names) have at least one non-patronymic byname, and only 0.4% have two. (None has more than two.)

The distribution of types of bynames, as a percentage of all names with bynames, is shown in the following table for all three studies. (Remember that multiple bynames may cause the total to add to more than 100%.)

Descrip. Occ. Loc.
Mawdd. 86% 10% 2%
Angl. 79% 17% 11%
Mer. 72% 23% 6%

The following table shows what percentage of the names with a particular number of generations contain non-patronymic elements. Note that the use of bynames decreases with the number of generations included. This follows my general thesis that a primary purpose of name structure is differentiation -- names are not made needlessly complex, but rather a number of techniques can be used in various combinations to distinguish between one person and the next.

gen. 1 2 3 4
Mawddwy 78% 69% 28% 14%
Anglesey 100% 72% 17% 15%
Merioneth 90% 4% 7%

There are clear patterns as to which generation, in a multi-generation name, tends to bear non-patronymic bynames. Naturally, in a one-generation name, there is only one generation for the elements to attach to. However when more generations are present, the is a clear trend regarding placement. The following table shows the absolute numbers for which generation the byname appears in when only one byname is present. I have not presented normalized statistics, but it should be noted that, when available, the proportions within each group (by number of generations) are strikingly similar. That is, in two-generation names, the byname is attached to the second (last) given name ten times more often than to the first. When three generations are present, the byname is four times as likely to fall on the last given name than on the first, with the second position even less likely. In the Anglesey data, the pattern continues with all the examples having the byname in the last (fourth) generation.

  2 generations 3 generations 4 generations
  1st 2nd 1st 2nd 3rd 1st 2nd 3rd 4th
Mawddwy 2 19 1   4 1      
Anglesey 59 524 31 8 117       8
Merioneth 4 58     1        

Since the Mawddwy data only includes four names containing more than one byname, it is hard to make significant comparisons, but the following tables show the absolute numbers, and percentage of each sub-group with a particular number of generations of names with two bynames, and the distribution of bynames in names containing two bynames. Note that in neither group did any four-generation name include two bynames.

Proportion of multiple bynames

# of generations 1 2 3 4
Mawddwy 1 (11%) 2 (6%) 1 (3%) 0
Anglesey 49 (23%) 45 (5%) 4 (0.4%) 0
Merioneth 9 (1%) 1 (0.1%) 0 0

Distribution of multiple bynames

  2 generations   3 generations  
Byname appears in 1 & 2 2 & 2 1 & 3 2 & 3
Mawddwy   2   1
Anglesey 25 20 4 2
Merioneth   1    

Note that I haven't included all possible combinations -- only the ones that occur. All names had at least one of the bynames in the last generation. This continues that pattern found above for only one byname: the final generation included is the most likely to include a byname.

While the Mawddwy data isn't numerous enough to do the strict statistical analysis about name complexity that I did for the Anglesey data, a more anecdotal analysis shows similar trends. If the names are split into three roughly similar-sized groups based on the popularity of the first given name, we get a group of the two most popular names (Ieuan and David) with 32 individuals, a middle group (Gruffudd, Madog, and Llywelyn) with 24 individuals, and the remaining, less popular names, with 30 individuals. Using the same "complexity score" system as I did for the Anglesey data (that is, each given name and each byname counts one "point"), the most popular group averages a complexity score of 3.03 (that is, slightly over three elements), the middle group a score of 3.00, and the least popular group a score of 2.67. While the statistical difference (particularly between the top two groups) is not great, the apparent trend follows the admittedly subtle pattern seen in the Anglesey data, for the commonness of the given name to correlate with more complexity in the name structure. (Note that this is a correlation, not an absolute rule. Two names that only appeared once each in the Mawddwy data had complexity scores of four.)


Jones, Heather Rose. "Names and Naming Practices in the Merioneth Lay Subsidy Roll 1292-3" in Proceedings of the Known World Heraldic Symposium (1991). Washington D.C.: SCA College of Arms, 1991.

Jones, Heather Rose. "Names and Naming Practices in the Anglesy Submissions of 1406" in Y Camamseriad 4 (1996).

Williams-Jones, Keith. "A Mawddwy Court Roll, 1415-16" in Bulletin of the Board of Celtic Studies vol. 23 (1970).

This site belongs to Heather Rose Jones. Contact me regarding anything beyond personal, individual use of this material.

Unless otherwise noted, all contents are copyright by Heather Rose Jones, all rights reserved.