From ‘Stemmata Robertson et Durdin’ compiled chiefly from printed authorities by Herbert Robertson. London. 1893-5

The ancestry of this Andrew of Atholia has been the subject of much discussion. It seems clear, however, that Angus of Cowell, the son of Angus Mor was his father. The Book of Clanranald states clearly that Angus of Cowell was son of Angus Mor and ancestor of the Robertsons, and Andrew de Atholia is known also to be one of their ancestors, further his son Duncan is styled MacInnes na Coalich, and universal repute made the Robertsons, including of course Andrew de Atholia, descendants of the Macdonalds until a doubt was thrown on it by Buchanan of Auchmor (out of deference to the family) suggesting that perhaps they claimed to be an independent instead of a branch clan. It being thus established that Angus of Cowell and Andrew were both ancestors of the family, the dates show that they could only have stood towards each other in the relation of father and son, and this theory explains all the difficulties that have arisen.

The history of the difficulty on this point of the pedigree is as follows:- the oldest record, the Book of Clanranald, records that Angus Mor had a son “Aonghus na Conluighe, from whom sprang the Clann Donchaidh and Robertsons”. Until the publication of Skene’s Celtic Scotland, 1880, this record so far as it relates to the Robertson family was quite lost sight of. Duncan, undoubtedly the son of Andrew de Atholia, and style Duncan, MacInnes na Coalich, that is MacAngus of Cowell (Douglas’s baronage, p 405), was treated as son of Angus of Cowell and thus Andrew dropped out of the family. The tradition that Duncan was a descendant of the Macdonalds which always existed is recorded, according to Col J.A. Robertson’s ‘Concise Historical Proofs respecting the Gael’ p437, as early as 1549, by Hugh Munro, Dean of the Isles, in his ‘History of the Macdonalds’ which is not that given in ‘Miscellanea Scotica’ iv, p121, and again by Alexander Skene in 1678 (memorials of the Skene Family,’ by W.F. Skene (New Spalding Club) p6. The MS history of the Macdonalds, temp Charles II, of which a fragment is printed in the ‘Collect de Red Alb’ (Iona Club) p 282, identifies Angus of Cowell, who had thus become father of Duncan, with a certain Angus, called p292 a natural son of Reginald, son of Somerled. The MS evidently confuse Duncan with someone else, as it says he ruled for three years in the Isles. This pedigree would place the birth of Duncan early, instead of late, in the thirteenth century. Crawford’s ‘Scotch Peerage’. 1716, under Macdonald, Lord of the Isles, does not refer to the Robertsons. Sir William Douglas in his ‘Peerage’ (first edition), 1746, under Macdonald, Lord of the Isles, p357, refers to it citing as his authority ‘MS history of the family’, and this is probably the MS temp Charles II already referred to; but Angus of Cowell he identified with Angus Mor, and Duncan himself thus made son of Angus Mor he calls Donald, adding after his name ‘designed Donaldus de Isle, filius Angusii, in Rymer’s Foedera Angliae, and is said to be ancestor of the Robertsons of Strowan’. The change from Angus, natural son of Reginald, to Angus Mor, was evidently made from the old tradition as to the latter being ancestor of the Robertsons, and that from Duncan to Donald was probably due to the erroneous allegation in the MS that Duncan ruled over the Isles, and to the author finding no Duncan de Insulis, but finding Donald des Isles mentioned in Rymer’s ‘Foedera’ (new edition), 1, p761, and apparently misreading the entries so as to make him son instead of father of Angus. In 1798 Douglas published his ‘Baronage’, and so had to make more minute enquiry into the Robertson family (p405). He follows the ‘Peerage’ in identifying Angus of Cowell with Angus Mor, who he says was designated of Cowell, but for the Donald named in the ‘Peerage’ substitutes Duncan de Atholia, thus in this respect going back to the MS. taking the ‘Peerage’ and ‘Baronage’ together Duncan de Atholia for the first time definitely appears as son of Angus Mor, and his actual father and grandfather disappear from the pedigree. The pedigree in the  ‘Baronage’ is copied verbatim into Alexander Robertson’s ‘Struan Family’, published a year or two after. Nisbet’s ‘Heraldry’, 1,p323, published in 1804, simply calls Duncan, Duncan Macdonald. The editor of the second edition of Douglas’s ‘Peerage’, 1813, does not appear to have known of the correction in the ‘Baronage’, as he repeats the entry in the first edition, merely adding a query after the erroneous reference to Rymer. In the meantime, before the appearance of the ‘Baronage’, viz in 1793, Buchanan of Auchmor had published his ‘Account of the Macdonalds’ (Miscellanea Scotica,’ iv, p81). He simply gives as the alleged ancestor of the Robertson family, ‘Duncan Crosda, or cross-grained, a son of the Macdonald, Lord of the Isles, about the reign of William the Lyon.” As William reigned from 1165-1214 this would place Duncan at about the same date assigned to Duncan in the MS, temp Charles II, which he probably had access to. His information evidently came from a Macdonald source, and he appears to think that the Robertsons might object to be treated as a junior branch of another clan instead of as an original clan, for he adds that ‘for divers ages Robertson hath been reputed a distinct surname and had no dependence upon that of Macdonald, so that although this above account be the most generally received in relation to the descent of that surname I am not positive how far it is acquiesced in by those of the same.’ He thus, out of deference to the Robertsons, introduced the first recorded doubt as to their descent from the Macdonalds. The designation Crosda is by Sir Noel Paton (‘Earls of Althol’) attributed to Duncan, son of Robert de Atholia. The unreliability of Douglas’s ‘Peerage’ and ‘Baronage’ and of Buchanan of Auchmor is pointed out in Skene’s ‘Celtic Scotland’.iii B8079. The next work was Skene’s ‘Highlanders’ 1837. Mr Skene appears to have taken considerable interest in the Robertson family of which his own is a branch, and he deals fully with it in vol. ii, p140 et seq. He treats the matter as if the accepted theory was that Duncan was the son of Angus Mor, as given in the ‘Baronage’, and having discovered the charter in which Duncan is styled son of Andrew de Atholia he saw at once that Duncan’s real father had been omitted from the pedigree in error. He, therefore, reinstated this personage by the name of Andrew of Atholia, and set himself to find out who his father was. He does not appear to have known of the entry in the Book of Clanranald, but of course knew of the old traditional descent of the family from the Macdonalds recorded as above. Mr Skene, however, argues that the MSS of 1450, printed in Iona Club ‘Transactions’, affords presumption that this traditional descent is unfounded as it details the branches of the Macdonalds with great minuteness and does not include the Robertsons, forgetting apparently that the entire omission of the Robertsons clan, which is evidently simply an error, prevents any such presumption arising, as the connection with the Macdonalds would only have been given when the clan was dealt with, and he considers the presumption increased by the family possessing such large territories in Athol, by the arms of the two families being different, and by his being unable to discover any connection as a clan between them as previously noted by Buchanan, a circumstance which arose from the families being separated before the first appearance of clans (see supra p5). He then argues that the family were descended from the old earls of Athol on the following grounds:- 1. The title of Andrew and his immediate descendants, viz. de Atholia, as he considers it hardly possible that the possession of a considerable part of Athol could entitle a stranger to the use of such a designation, “as Atholia was a Norman barony, and the names of the baronies were exclusively used by their possessors and their descendants”; but considering the number of persons who bore this name mentioned below this proposition is untenable. 2. Glenerochie (Strowan) belonged to Conan, younger son of Henry, last Celtic Earl of Athol, and passed to his son Ewen and then to Duncan de Atholia; 3. Finoch or Fynoch in Stratherne belonged to Ewen’s wife and passed to Duncan de Atholia. The last three arguments show conclusively that Duncan was a descendant of the old Earls of Athol, and heir to the estates of Ewen, the grandson of the last earl Henry, and this (whether then suggested for the first time or not) has never since been doubted, but they do not prove that he was a descendant of the old earls in the male line. Mr Skene’s suggestions in favour o a male descent are as follows;- 1. Unless Dull was a male fief, inherited by Duncan as heir male of the family, he considers it impossible to see how why the daughters of Earl Henry’s eldest son, who carried the title of Earl to Lowland families, did not inherit it. There is however no reason why it should not have been given By Earl Henry to his son Conan. 2. He states, without any proof, that Glenerochie was a strictly male fief which, if true, would make Duncan male heir to Ewen, and so male descendant of the old earls. His conclusion is that Andrew de Atholia was the son of Angus, son of Ewen, the grandson of Henry, last earl of Athol, hence his son’s title of McInnes, and therefore that the Robertsons of Struan are the male representatives of the old Earls of Athol and of the ancient royal house of Scotland, which was descended from King Duncan, son of Crinan. This magnificent descent, supported by so high an authority, was adopted by the Robertsons of Strowan, and printed the following year 1838 in Burke’’Commoners,’ vol iv. p419, and in Burke’s ‘Landed Gentry’ 1847, Robertson of Struan; and was also followed in James Browne’s ‘History of the Highlanders 1858, vol iv, cap vi, p468 et seq, William Anderson’s ‘Scottish Nation’ 1863.iii. p345, and by Sir Joseph Noel Paton in his ‘Celtic Earls of Atholl’ 1873. Mr Skene appears to have adhered to the correctness of this pedigree. In his ‘Celtic Scotland’ 1880, vol iii, p289, he says, that the family were probably descendants of the old Gaelic earls and at p309 he says they were so descended, and at p330 he says their name (de Atholia) sufficiently indicates that they were the male representatives of these earls, and at p361 he says that by a fraud common to later times they exchanged Andrew de Atholia, for Angus of the Isles. He seems not to have observed the entry in the Book of Clanranald, although it is printed at the end of the volume, p397, and is, according to his statement at p338, one of the oldest Highland pedigrees and worthy of credit. Notwithstanding the weight of Mr Skene’s authority, it seems clear that he has given no sufficient proof of a male descent from the Earls of Athol to displace the original tradition of a descent from the Macdonalds, supported as it is by an entry in the Book of Clanranald, and there is no foundation for such a royal descent as he has traced. It is noticeable that the men of Athol and Argyll had always been closely connected; they fought together, temp Solvathius (See Sir James Dalrymple’s ‘Collections concerning Scottish History’ 1705), also under Somerled in 1164 (Chronicon Hyense in Adamnan’s ‘St Columba’ Bannantyne edition p402) the Earl of Athol had authority in Argyll (Regia in Magistatem, lib i, cap xvii, art 6, cited in Dalrymple’s ‘Collections’, p217). The erroneousness of this pedigree was seen so long ago as 1847 by Major J.A. Robertson, the head of the great family of Lude, and he prepared a new table of descent which (see his ‘Athol;’ p23) he supplied to Mr Burke and which is printed in the ‘Landed Gentry’ of that year, under Robertson of Lude. Subsequently, in 1860, he reprinted the same pedigree in his ‘Comitatus de Atholia’. In this pedigree he very properly added the old traditional descent from the Macdonalds through Angus Mor. He had to connect this Angus Mor with Andrew de Atholia, and evidently did not know of the entry in the Book of Clanranald. he, accordingly, selected as the connecting link, who was to be the son of Angus Mor, and father of Andrew de Atholia, a person whom he called Malcolm de Insulis, and identified with M. de. Asceles (Atheles), mentioned in Robertson’s ‘Index’, pxvii, temp. Alexander III 1249-86, without saying where he got the name Malcolm from. To this person (in order to account for the title de Atholia and for Andrew, and his son Duncan, being styled Earl of Athol in Robertson’s ‘Index’ he gave a wife Lora Countess of Athol, whom he makes, without giving any reason, daughter of Isabella, eldest co-heiress of Henry, last Earl of Athol, by Alan Durward. The only record of any Lora, Countess of Atholl, is an entry in the Chronicle of Melros in the year 1269, recording the death of such a person, without giving her parentage or any comment (probably she was widow of one of the Hastings of Strathbogie families) and the whole of this link, in the pedigree of which Major Robertson was admittedly the inventor, depends upon this entry and the bond mentioned above by “M de Asceles’ probably the MacBeke de Atholia mentioned below), which manifestly afford no foundation for the insertion of either Malcolm or Lora in the pedigree. Major Robertson, as shown below, subsequently though this pedigree erroneous, but it was nevertheless adopted by all branches of the family which appeared in the ‘Landed Gentry’ prior to the 1886 edition. Major Robertson’s ‘Athol’ p23, complains of this being done without acknowledgement. To this supposed Malcolm and Lora Major Robertson at first (see the ‘Landed Gentry’, Lude 1846) attributed an elder son John, Earl of Athol, relying on Andrew and Duncan being styled (erroneously) Earls of Athol, and on a mistake in the ‘Peerage’, by which a charter John Strathbogie, Earl of Athol, who died in 1264, was made to bear date in 1283, instead of 1254 (see Skene’s ‘Celtic Scotland’ iii p289 note) but this error is not repeated in his ‘Athol’. As already stated in 1858 appeared Browne’s ‘Highlanders’ copying Skene’s ‘Highlanders’ and not noticing the pedigree under Lude in the ‘Landed Gentry’ of 1847; and editions of the ‘landed Gentry’ in 1858 and 1863 adopt this pedigree for the Robertsons of Struan and all other branches of the family, and in 1860 Major Robertson’s ‘Athol’ combating Skene’s theory, and repeating the pedigree supplied to Burke’s in 1847. In 1866 Major Robertson published his ‘Concise Historical Proofs respecting the Gael’, and by that time he had evidently come to the conclusion that the supposed Malcolm and Lora could not be the parents of Andrew, for at p437 he suggests that Andreas filius Gilmur Clerauch de Dull, mentioned as living in 1269 in the register of St Andrews, p349 (whom he had referred to in his ‘Athol’) p23), and that this Gilmur was the Gillmur, son of Maoldooch, seneschal or administrator of Dull, who he says is mentioned by Buchanan of Auchmor, and that the latter may have been a son of Reginald of the Isles, the son of Somerled. This view is purely fanciful, and all the evidence is against it. The latest work, Sir Noel Paton’s ‘Celtic Earls of Athol’, 1873, simply copies Mr W.F. Skene’s pedigree as stated above. Burke’s ‘Landed Gentry’ which previously followed the 1863 edition in 1886, omits the earlier pedigree for all families beginning with Duncan de Atholia, who is erroneously styled Earl of Athol, The discovery of the entry in the Book of Clanranald showing that Aonghus na Conluighe was father of Andrew of Athol, furnishes a solution to the difficulty as to the parentage of this Andrew de Atholia.

Major J.A. Robertson’s earlier pedigree of Andrew (in this respect correct) did not account for the ownership of Glenerochie. He, however, in Burke’s ‘Landed Gentry’ 1846, under Robertson of Lude, and in his ‘Athol’ and ‘Concise Historical Proofs’ p437, adopted Mr Skene’s proposition, which seems to be well established, that Duncan de Atholia was a descendant of Ewen of Glenerochie,, whom he made heiress [sic] to her father and her brother Angus, and this view has been adopted by all families recorded in the later editions of ‘Landed Gentry”.