As Sir Henry Howorth has so truly observed, in a presidential address to the members of the Archæological Institute, the transition from the chronicle to the record as a source of mediæval history is one of the most striking and hopeful features in recent historical research. And in no respect, perhaps, has the study of original records modified more profoundly the statements of mediæval chroniclers than in the matter of the figures they contain. Dealing with the introduction of knight-service into England, I was led to give some instances in point,[581] and specially to urge that “sixty thousand” occurs repeatedly as a conventional number ludicrously remote from the truth. It is now, I believe, generally accepted that my estimate of about five thousand for the number of knights’ fees in England[582] is nearer the truth than the “sixty thousand” which, in his History, Mr. Green accepted. But we still read in ‘Social England’ (i. 373) that William I. “is believed to have landed ... with at least 60,000 men”; nor did Mr. Freeman himself reject the statement of Orderic that “sixty thousand” men were gathered on Salisbury Plain for the “Mickle Gemót” of August 1, 1086. We who saw, only last summer, the difficulty of there assembling a force scarcely so large, even with all the modern facilities of transport and organization, can realize, more forcibly than ever, the incredibility of the fact.

“Stephen Segrave,” Dr. Stubbs reminds us, “the minister of Henry III., reckoned 32,000 as the number” of knights’ fees; and even so late as 1371, ministers allowed a parliamentary grant to be calculated on the belief that there were 40,000 parishes in England, when there were, as a fact, less than 9,000.[583] So too, as is well known, Fitz Ralph, archbishop of Armagh, declared at Avignon, that at Oxford, in his early days, there were 30,000 students, although it is probable that they cannot have exceeded 3,000 in number.[584] It is even said that Wycliffe doubled Fitz Ralph’s estimate.

There is nothing, therefore, strange in the fact that two centuries and a half after the Norman Conquest, we still find absurd numbers assigned to armies in the field and accepted with thoughtless readiness, even by modern historians. This, we shall see, has been the case, among many other battles, with that of Bannockburn (1314).

The ultimate “authority” for the numbers engaged at this ever memorable fight is Barbour’s Brus. Of Edward that romancer wrote:

In accordance with this statement we read further of the king, that

Of the Scots we are told that:

On the English side we have a statement in the ‘Vita Edwardi Secundi.’ It is there asserted, of the host marching on Stirling, that

The same authority states that Bruce

Let us now see how modern writers have dealt with the numbers present, remembering that the character and issue of the battle turn largely on the vast numbers assigned to the English host.

In the ‘Dictionary of National Biography’ (1886) Dr. Æneas Mackay adopts the traditional view of the English numbers, following Barbour, indeed, blindly:

A far abler and more cautious writer, Mr. Joseph Bain, F.S.A. Scot., in his ‘Calendar of Documents relating to Scotland’ (1887), reckoned that “the whole English army probably did not exceed 50,000.”[588] Against Hailes on the Scottish side, he supports Hume, who, he writes:

From Mr. Bain I turn to our latest authority, Mr. Oman’s ‘History of the Art of War.’

To the memorable Scottish victory Mr. Oman, as we might expect, devotes special attention (pp. 570–579). He attributes “the most lamentable defeat which an English army ever suffered” to two fatal errors, of which one “was the crowding such a vast army on to a front of no more than two thousand yards” (p. 579). His argument, in detail, is this:

In his first work (1885) Mr. Oman had adopted “100,000 men” as the number of Edward’s host; in 1895 it had become “an army that is rated at nearly 100,000 men by the chronicler.”[589] In 1898 we learn that “the estimate of a hundred thousand men, which the Scottish chroniclers give, is no doubt exaggerated, but that the force was very large is shown by the genuine details which have come down to us” (p. 573). These “genuine details” prove to be the figures in the ‘Foedera,’ on which Mr. Bain relied. Mr. Oman arrives at his figures thus:

To the numbers of Edward’s host he attaches so great an importance that he gives the details, from Rymer, in a note. I make the total, myself, to be 21,540.[590] It is Mr. Oman’s extraordinary delusion that the other English counties were similarly called on for troops, but that the orders have not been “preserved.” On the strength of this illusion alone he adds some 30,000 men to the English host! A glance at Rymer’s list, as given in his own pages, is sufficient to dispel that illusion. As Mr. Bain correctly implies, the counties called on for troops form a compact group, of which Warwick was the southernmost. Moreover, even within that group, the southern counties were evidently called on for much less than the northernmost, Warwick and Leicester only sending 500 men, while Northumberland and Durham were called on to supply 4,000, as was also Yorkshire. We have only to turn to the ‘Rotuli Scotiæ’[591] for 1314 to learn that the writs originally issued (i.e. in March) for the Bannockburn campaign summoned no more than 6,500 men, and these from the counties “beyond Trent” alone.[592] As the peril increased subsequent writs called for a further 6,000 men from these counties, and extended the net so as to obtain 3,000 from Lincolnshire, 500 from Warwickshire and Leicestershire, and 500 from Lancashire (previously omitted); this, with 4,940 men from Wales and its marches, made up the total.

When Edward III. arrayed his host, twenty-five years later, for the French war, he only asked for 500 foot from Northumberland (as against 2,500), and 1,000 from Yorkshire, but from Warwickshire with Leicestershire he exacted 480. These figures speak for themselves. Any one of ordinary intelligence can see that the forces on these two occasions were raised on entirely different principles, Northumberland being called on for five times as many men in 1314 as in 1339, while Warwickshire and Leicester supplied almost as many in the latter as in the former year. And yet Mr. Oman actually makes the comparison himself (p. 593), and prints the numbers in detail for both occasions without any comprehension that this was so. Indeed, he bases on his misapprehension a theory that as, at the later date (1339), the quotas were never more than a third of those demanded for Bannockburn (1314), a comparatively picked force was secured.

We have seen that, on the contrary, in Warwickshire and Leicestershire, the number summoned was almost the same, and that the above theory is, therefore, another delusion. In 1339 the proportion varied from 20 per cent. to 96 per cent. of the numbers summoned in 1314, and did so, as we have seen, on a geographical system. Mr. Oman bases his above assertion on a note in which four lines contain four direct mistakes. It asserts that Yorkshire sent “six thousand,” Lincolnshire “four thousand,” Warwick “five hundred,” and Leicester “five hundred,” in 1314, when the right numbers, as given by himself on page 573 of the same volume, were: Yorkshire four thousand, Lincolnshire three thousand, and Warwick and Leicester together five hundred. The result of this astounding inaccuracy is that he fails to understand the system of these levies in the least.

It is, no doubt, surprising that, after years of study, a writer should produce a work intended to constitute a standard authority on mediæval warfare, in which he has not even grasped so elementary a fact as the raising of English armies, in the 14th century, on geographical principles, and should consequently invent an imaginary host of nearly 30,000 men. Precisely as in 1314, the bulk of the foot for the Scottish expedition were raised from the Northern counties, so in 1345, for the contemplated French expedition, it was from the counties south of the Trent that the infantry (archers) was raised.[594] But it is even more surprising that he should substitute for this system a theory, based on the misquotation of his own figures alone, that, in 1339, we meet with a new system of summoning a comparatively small quota of picked men. It is but a further instance of his grievous lack of accuracy that on page 599 he renders the “homines armati”[595] summoned from the towns as “seventeen hundred archers,” although he prints from Rymer, a few pages earlier, the numbers of the foot summoned in 1339, of whom half are distinguished as archers and half as “armati.”

One would have imagined that the fact of the host being drawn from the northern half of England alone would have been obvious from the dates. The orders from which Mr. Oman takes the numbers demanded were only issued from Newminster on May 27,[596] and ordered a rendezvous of the force at Wark (Northumberland) on June 10. The troops were to be there on that day “armis competentibus bene muniti, ac prompti et parati ad proficiscendum” to the immediate relief of Stirling. The time was desperately short, and haste was enjoined (“exasperes, festines”). Moreover, the English leaders were clearly not such fools as Mr. Oman imagines. The “orders” state that foot are wanted because the Scots

Common sense tells one that 60,000 foot could not be manœuvred in such country, and would only prove an encumbrance. Edward, therefore, only summoned less than 22,000. As to his horse, Mr. Oman writes: if the English “had, as is said, three thousand equites coperti, men-at-arms on barded horses, the whole cavalry was probably ten thousand” (p. 575). But why? At Falkirk, sixteen years before, Edward I., he writes (p. 565), had

If 3,000 “barded horses” implied 4,000 other horsemen in 1298, why should they imply 7,000 in 1314? More especially, why should they do so when, as we have seen, the king, in summoning his foot forces, himself described the scene of the campaign as “ubi equitibus difficilis patebit accessus,” so that he was most unlikely to take a large force of cavalry?[597] Estimating the horse on the Falkirk basis, the English host cannot have amounted to more than 30,000 men instead of the 60,000 or 70,000, horse and foot, at which Mr. Oman reckons it.[598]

And what of the Scotch? Let us compare these passages:

It is true that the Scottish king had a fourth battle in reserve, but, according to Mr. Oman’s plan, it was no larger than the others, if so large. It would only, therefore, add some 8,000 men to the above 25,000. Where then are the 40,000?

From the numbers of the forces I now pass to their disposition on the field. With each of his successive narratives of the battle Mr. Oman has given us a special—and different—ground plan. In all three of these the English ‘battles’ are shown as composed of horse and foot,—the horse in the front of each, the foot behind. But in the earliest of these (1885) three such ‘battles’ are shown, in the second (1895) five, and in the third (1898) ten.[601] Will the number increase indefinitely? Again, as to the famous “pottes,” dug as traps for the English horse. In the earliest narrative these are described as covering the Scottish flank “to the left,” and in the second, as dug by the Scots “on their flanks,” though in both the ground plans they are shown in a cluster on the left flank alone. When we turn, however, to the latest account (1898), we find them shown, no longer on the flanks, but as a single line along the Scottish front, and described as dug by Bruce “in front of his line,” so that they “practically covered the whole assailable front of the Scottish host” (p. 572).

Lastly, on that all-important point, the disposition of the English archers, we are shown in the first ground plan the “English archers considerably in advance of the main body,” and, indeed, almost all on the Scottish side of the burn. In his second they are still in front of the host, but no longer across the burn. In his third there are no “archers” shown, and the English ‘battles’ themselves are depicted as close up to the burn. But to realize the completeness of the contradiction, one must place side by side these two passages:

Poor Edward! He is first made to place his archers in front of his horsemen, and blamed for his folly in doing so; and then he is made to place them behind, and again blamed for his folly.

It is the same with the battle of Creçy (1346). Let any one compare the four narratives given in succession by Mr. Oman,[604] together with the three ground plans, and he will be fairly bewildered. The only thing of which we can be sure is that when Mr. Oman has adopted a view, he will himself afterwards abandon it. It is the same, again, with the numbers also. Mr. Oman, in his second narrative (as apparently in his first), reckons the English host at some 9,300 men (6,000 archers, 2,300 men-at-arms, 1,000 Welsh). In his fourth they exceed 20,000 (11,000 archers, 3,900 men-at-arms, 5,000 or 5,500 Welsh).

Need I pursue further this endless contradiction? It has been my object to warn the reader of Mr. Oman’s works on the Art of War to compare his successive views before adopting a single one of them. Whether on the field of Bannockburn or of Hastings we need a guide who knows, at least, his own mind, and whose “cocksureness” is not proportionate to the mutability of his views.