Why Bartrum’s dating of the Demetian Arthur is wrong

The nature of the problem

To understand why there is a problem with the birthdate of Arthur ap Pedr given by Bartrum, namely 560, we need to look at a pedigree list, HG 2, that mentions him:

[O]uein m. [H]elen merc [L]oumarc m. Himeyt m. Tancoys[tl] merc Ouein m. Margetiut m. Teudos m. Regin m. Catgocaun m. Cathen m. Cloten m. Nougoy m. Arthur m. Petr m. Cincar m. Guortepir m. Aircol m. Triphun m. Clotri m. Gloitguin m. Nimet m. Dimet m. Maxim Gule[t]ic m. Protec m. Protector m. Ebiud m. Eliud m. Stater m. Pincr Misser m. Constans m. Constantini magni m. Constantii et Helen Lui[t]dauc, que de Brittannia exiuit ad crucem Christi querendam usque ad Ierusalem, et inde attulit secum usque ad Constantinopolin, et est ibi usque in hodiernum diem.

Table 1 shows the above list together with the allocated generation numbers, estimated birthdates and obits sourced from the Annales Cambriae stated by Bartrum in A Welsh Classical Dictionary.

Gen.	Birthdate	HG 2
1	330	Maxim Gule[t]ic (ob. 388)
2	355	Dimet
2	380	Nimet
3	410	Gloitguin
4?	440?	Clotri
4	430	Triphun
5	460	Aircol
6	480	Guortepir
6	510	Cincar
7	535	Petr
8	560	Arthur
9	580	Nougoy
9	600	Cloten
10	625	Cathen
11	650	Catgocaun
11	675	Regin
12	700	Teudos
13		Margetiut (ob. 796)
14?		Ouein (ob. 811)
15?		Tancoys[tl] (ob. 790)
16		Himeyt (ob. 893)
17		[L]oumarc (ob. 904)
18		[H]elen (ob. 929)
18	900	[O]uein

Table 1. Bartrum’s generations and birthdates.

Bartrum’s scheme would have had 17 generations between Maxim Guletic and Margetiut whose obits Bartrum dated as 388 and 796 respectively. This would result in 24 years/generation, contrary to his statement on generation lengths:

This is based on the well known fact that, on the average, three male generations span almost exactly a century … A male generation is the period of time between the birth of father and the birth of son, or daughter. Female generations, birth of mother to birth of son or daughter, are on the average much shorter, and in the period covered are nearer to 20 years, i.e. five to a century.

Bartrum, P.C., 1974, vol. 1, 6.

To achieve a more realistic step-size, he gave a number of parents and offsprings the same generation number, viz. Dimet/Nimet, Guortepir/Cincar, Nougoy/Cloten, Catgocaun/Regin, and possibly Clotri/Triphun.This results in a more satisfactory generation lengh of 34 years/generation, but at the cost of artificially giving parents and offsprings the same generation number. This combining of generations also occurs at the end of the table with Helen/Ouein.
Furthermore, it needs to be pointed out that there is an anomaly in that Bartrum’s tentative birthdate given to Clotri is 10 years after that of Tryffin who Bartrum believed was his nephew.

Bartrum did not continue beyond what he termed generation 1. However, under the entry headed Tryffin, king of Dyfed in A Welsh Classical Dictionary he provided a continuation togehter with some additions derived from other lists with which he attempted to synchronize :

Table 2. The Ancestry of Tryffin.

However, this made the problem worse as now there would have been 23 generations between Maxim Guletic and Margetiut, giving a generation length of 18 years/generation.

The resolution of the problem

The reason why Bartrum’s analysis went astray was because he treated HG 2 as a single line of descent, whereas in reality it consists of three seperate segments as shown in table 3. That the manuscript compilers did not always know when one list ended and the next one began can be seen in, for example, JC 10 to 11 where the ending of the first list is repeated at the start of the following one.

The proposed generation numbers are based on the principle that the generation number of an offspring should be one greater than that of the parent. Arthur’s generation has been numbered gen. 0. For table 3 the birthdates do not give estimates for the individuals listed, as they did in table 1, but rather give the average birthdate for members of each generation. This has been derived from a database of known obits of individuals named in a range of pedigree lists with the assumption that the average life-span was 65 years. Note, the birthdates, on average, increase in a regular fashion by 32 years, consistent with Bartrum’s assertion concerning generation lengths quoted above. The vast majority of generations in my database are male ones.

Gen.	Birthdate	HG 2a	HG 2b	HG 2c
-9	197		Protector
-8	229		Protec
-7	261		Maxim Gule[t]ic	Constantii (ob. 306)	Helen Lui[t]dauc
-6	293		Dimet	Constantini magni
-5	325		Nimet	Constans
-4	357		Gloitguin	[Macsen Wledig] (ob. 388)
-3	389		Clotri	[Custennin]
-2	421		Triphun	Pincr Misser
-1	453	Petr	Aircol	Stater
0	485	Arthur	Guortepir	Eliud
1	517	Nougoy	Cincar	Ebiud
2	549	Cloten
3	581	Cathen
4	613	Catgocaun
5	645	Regin
6	677	Teudos
7	709	Margetiut (ob. 796)
8	741	Ouein (ob. 811)
9	773	Tancoys[tl] (ob. 790)
10	805	Himeyt (ob. 893)
11	838	[L]oumarc (ob. 904)
12	870	[H]elen (ob. 929)
13	902	[O]uein

Table 3. Proposed generations and birthdates.

This analysis gives the optimum date for Arthur’s birth as 485. The Demetian Arthur’s birthdate is in the period that one would expect for the individual around whom the Arthurian cycle was built bearing in mind the dates for Badon and Camlan given in the Annales Cambriae.

One consequence of this proposed analysis is that the Maxim Guletic of table 1 was not a reference to Macsen Wledig but rather to the emperor Constantius Chlorus, as shown in table 3, who died in 306. His wife, St. Helena, was conflated in the first part of the Mabinogion tale, entitled Dream of Macsen Wledig, with Eudaf’s daughter, Elen Luyddog, who was probably St. Helena of Cornwall. Bartrum appears to agree with this identification in table 2 where he synchronizes a number of pedigrees, including JC 13.
The proposed allocation of generations in the relevant segment, JC 13b, of that list is shown in table 4. Note, this list provides the missing names in HG 2c, namely Custennin and Macsen Wledig. The latter individual appears in the correct generation to be identified with the emperor Magnus Maximus who died in 388.

Gen.	JC 13b
-7	Custenint	Elen
-6	Constantinus Maỽr
-5	Maximianus
-4	Maxen Wledic
-3	Custennin
-2	Miser
-1	Ewein
0	Kyngar (m.) Prỽtech
1	Ewein
2	Cyndỽr Bendigeit

Table 4.

Bartrum made use of the dateable event for the birth of Guortepir. His date of 480 cannot be too wrong as Gildas, writing in the 540s in the De Excidio, refers to him as Vortipor saying:

… discolor canescente iam capite … appropinquante sensim uitae limite …

… though thy head is now becoming grey … though the end of life is gradually drawing near …

DE 31. Williams H., 1899.

It may be that Barturm, too, noticed HG 2 needed to be divided into three segments as he wrote:

There seems to be three independent strands of pedigree here, see note to ABT 18a.

Bartrum, P.C., 1966, 126.