december, 1899, to march, 1900.

i interpose a chapter here in order to carry events in other parts of the theatre of war up to the date of the capture of bloemfontein.

a sketch will suffice, since specifically mounted operations were few. no body, either of mixed mounted troops or of regular cavalry reckoned separately, comparable in size to that formed by roberts in the main theatre, existed anywhere else. the largest homogeneous mounted force outside this area was brabant’s newly raised colonial division, nominally 3,000 strong, which, in conjunction with gatacre’s troops, had been deputed to push back the invaders of eastern cape colony from dordrecht and stormberg, while clements, succeeding french in the positions opposite colesberg, checked the menace to central cape colony. brabant, however, seems not to have been able to muster an effective strength of more than 2,000 during the period under review. that fine permanent corps, the cape mounted rifles, was a strong, stiffening element in an otherwise raw force of cape colony volunteers. fortunately, the work before them was not severe, for the success of roberts in the north threw the boers into a strictly defensive attitude from the middle of february onwards, and in the early days of march caused 151a general retreat. a successful attack upon labuschagne’s nek, between dordrecht and jamestown, on march 4, gave the recruits confidence.

clements had had a much harder task than gatacre and brabant. stronger forces opposed him, and the boer retreat set in later. early in february all the regular cavalry, save two squadrons of inniskilling dragoons, had been diverted to roberts’s command. there remained, besides these squadrons, 500 australian horsemen, together with infantry and artillery which made up the force to a total strength of about 5,000 men and 14 guns. against clements—if the official estimate is correct—the forces at one time were as great as 11,000. clements, fighting stubbornly, was forced back south of rensburg, and, in the course of the retreat, all his mounted troops, and particularly the australians, did excellent service—fire-tactics, of course, being the universal rule. the danger was soon over. on february 21 clements was reinforced with 900 mounted men and two batteries, and at about the same period the tide of invasion slackened. a week later, on the news of paardeberg, the boers were in full retreat for the north. by the middle of march—two days after the fall of bloemfontein—clements, gatacre, and brabant were all within the free state borders.

we need not enter at any length either into the siege of ladysmith or into the long series of operations which ended in its relief. the numerical facts, broadly speaking, were that white, with 13,000 men and 51 guns, was invested by a force under joubert which originally numbered 23,000 men and 17 guns, but which dwindled gradually by abstractions to the tugela, to cronje, and to colesberg, and finally fell to a strength of about 5,000; while, on the line of the tugela, buller, reinforced in the period following colenso to a strength of 30,000 men and 73 guns, faced louis botha and lukas meyer with a 152strength which varied in round numbers from 7,000 to 9,000 men and about 18 guns.

as in the western theatre and in every other part of the field of war, the rifle, whether in the hands of mounted men or infantry, was the decisive weapon. artillery, as a mere statement of the relative strengths in that arm shows, was comparatively negligible. sword and lance were out of court. every responsible person at the time realized this fact. short as we were of mounted troops, nobody would have dreamed of asking for more troops trained to shock on the ground that shock was either requisite or possible.

the most striking circumstance about the mounted troops in natal—upwards of 5,000 in number—was the fact that rather more than half were locked up in ladysmith during the whole four months of the sieges. four cavalry regiments, besides the natal carbineers, other natal volunteers, and the greater part of the imperial light horse—2,800 men in all—were demobilized in this way. the mistake, no doubt, was serious, and white has been freely blamed for it. at the same time, it is only fair to white to put ourselves in his position, and recognize that the question of retaining or parting with his mounted troops was subsidiary to the much larger problem which originally faced him in deciding what was to be the r?le of the natal army after the battle of ladysmith on october 30, 1899. had he possessed, in his force of professional mounted regiments, troops really capable, in conjunction with the volunteers, of tackling the boer mounted riflemen, it is difficult to believe that, in spite of the moral and material value of ladysmith, he would have accepted investment there as an alternative to the maintenance of his army as an active field-force. but the battle of the 30th, revealing a deficiency in the striking-power of the army as a whole, had revealed a weakness in the cavalry which was in no way attributable to moral 153causes, but simply to armament and training. this circumstance must have influenced him powerfully in resolving to accept investment, a resolve which it is exceedingly difficult to impugn. a retreat to the tugela, harassed by a greatly superior boer force, whose temper was exhilarated by the success at nicholson’s nek, would have been a hazardous operation. it is no reflection on the regular cavalry, but the simple truth, to say that they had not as yet shown the capacity to act as rear-guard for such a retreat.

but what kind of investment was white to accept? here, no doubt, he is open to the charge of compromising between two logical alternatives, the one being to send away instantly the bulk of his mounted troops and field artillery, and with the rest of his force to accept a formal siege, with the purely passive object of detaining as many boers as he could; the other, to keep his force intact, and maintain a defence so active and supple in character as to enable him to cut loose at any moment and co-operate with the relieving force. although something like this latter course was evidently in his mind, as it would naturally be in the mind of any spirited field commander, he did not clearly grasp the determining factors and act accordingly. he did not foresee the initial impotence of buller before the colenso position, also largely attributable to a deficiency in efficient mounted troops. he occupied too small a perimeter to permit of elastic offence, and he forgot that the tactical weakness of his cavalry was an obstacle even more serious to the kind of operations he had in his mind than it was to the larger plan of complete freedom which he had rejected. this weakness again became manifest in the small offensive operations of november 14 and december 7–8. then came buller’s failure at colenso, and henceforth white’s attitude, though courageous and unyielding, was strictly passive. this was all the more 154to be regretted because the boer attitude, save for the one big attack of january 5 on c?sar’s camp and wagon hill, and for the minor attack on november 9, was equally passive, while their numbers sank to a point well below the strength of the garrison.

white’s mounted troops were reduced by degrees to the r?le of foot-soldiers, and in that capacity took their share in the defence. the part played by the regular cavalry, gallant as it was, could not have been, and was not, so important as that played by the irregulars, who were genuine, though improvised riflemen. all alike took part in the great fight of january 5, and by common consent the chief honours belong to the imperial light horse, whose heroic defence of wagon point, the key to the threatened position, at a cost of 25 per cent. of the numbers engaged, was as fine a feat of arms as their final attack at elandslaagte. it was by a detachment of the same regiment, in conjunction with a body of natal mounted volunteers, that the brilliant little sortie of december 7–8 was carried out and the two heavy guns on pepworth hill destroyed.

during the last month of the siege, when forage became scarce, and 75 per cent. of the cavalry horses had to be turned adrift or converted into food, the troopers returned their lances, swords, and carbines to store, received rifles instead, and took regular posts in the defence. that change of weapons once made, it is almost inconceivable that it should not have been adhered to when horses were once more available. why deliberately revert to an inferior firearm? why deliberately resume steel weapons whose futility was manifest? tradition—nothing more: the ineradicable habit of associating together the horse and the steel weapon as complementary elements of the highest mounted efficiency; the same habit which induces general french, in defending the arme blanche, to say that “nothing is gained by ignoring the horse, the 155sword, and the lance,” as though these weapons were inseparable adjuncts of the horse, and as though south african experience were not one long and costly proof of the contrary.

buller’s mounted force, about 2,600 strong during the period following colenso, was composed mainly of south african irregulars,[35] with two and a half cavalry regiments, and a few regular mounted infantry. it played a creditable, though not a distinguished, part in the operations. the battles, from the british point of view, were all pre-eminently infantry battles. in one instance only, so far as i am aware, was a mounted corps employed in conjunction with infantry in a really critical and desperate fight, and that was the detachment of thorneycroft’s mounted infantry at spion kop. for the rest, we find them operating on the wings, seizing advanced positions, and guarding the flanks of the main attack. fire-tactics are the invariable rule, and efficiency in fire-tactics the test of general utility.

there is reason to believe that the mounted troops might have been employed to greater advantage had the higher command of the army been in stronger hands. though they were less than half as numerous as the mounted force at the disposal of lord roberts, they were on the average more than a quarter, and sometimes not far from a third, the strength of the whole boer force opposed to them—a tolerably high proportion, if we reflect that the boers, immensely strong though their position was, had to sustain the attacks of 20,000 infantry, to say nothing of an overwhelming number of guns.

the most hopeful enterprise in which the mounted troops were ever actually engaged was in the opening 156operations of the spion kop campaign (january 18 to 20), when dundonald’s brigade of 1,500 men, including one cavalry regiment, acted as advance-guard to sir charles warren, who, with the greater part of the army, was deputed by buller to turn the boer right, while lyttelton threatened the centre.[36] one of the most disappointing features of a painful story was the waste of a golden opportunity for utilizing mounted strength against an enemy whose high tactical mobility rendered surprise exceedingly difficult. dundonald, a cavalry man, certainly did his utmost, and, as far as he was allowed, did well. unnecessary delays had attended the turning movement from the first, but a considerable measure of surprise was, in fact, obtained. few boers had rallied to the threatened flank; none were entrenched. dundonald, operating boldly in advance, gained on the evening of the 18th a position, overlooking acton homes, which might, under vigorous generalship, have been turned to great strategical advantage. his men were in high fettle owing to the skilful surprise and defeat of a boer detachment which rode out to check them. but warren seems to have regarded his mounted troops wholly in a protective light, and to have resented anything approaching independent action. the chance was thrown away[37] and 157the operations never recovered from the initial sluggishness of movement.

another opportunity for a vigorous use of mounted troops came after the great fight at pieter’s hill (february 27), which led to the relief of ladysmith and to a general retreat of the boer forces both from the beleaguered town and from the tugela heights. if we regard all buller’s previous operations as one long-drawn battle—and in a sense they may so be regarded—now, it would seem, was the time for pursuit. the two leaders of horse were undoubtedly anxious to pursue. men and horses were alike fresh. buller refrained. there is a 158general agreement that he was wrong. whatever the prospects of success, he should unquestionably have tried, for instinctive and habitual mounted energy was the vital need in south africa if a mounted enemy was to be not only defeated, but conquered.

at the same time, a close examination of the facts does not appear to justify the assumption of the times historian that a pursuit would have involved the boers in utter destruction and defeat. the critic lays excessive and indiscriminating stress on the demoralization of the enemy. he forgets that botha’s troops and the investing force combined numbered in all about 13,000 men, as against 2,600 of our mounted troops; that there was not much question of further co-operation by our infantry, who were exhausted by ten days of continuous fighting, and that the encounters which actually did take place between our mounted troops (regulars and irregulars alike) and the boer rear-guard were not of such a character as to warrant a belief that a general pursuit, begun at the earliest possible moment, would have led to the destruction of the boer army.

both the german and british official historians correctly point out that, in order to have been really effective, the intervention of the mounted troops should have begun at, or immediately after, the climax of the great infantry fight on the 27th. here was just the difficulty: the british attack, delivered on a front of about three miles, was threefold—upon railway hill, inniskilling hill and pieter’s hill, the latter representing the extreme boer left, the only quarter at which the mounted troops could possibly have intervened. the two first positions were stormed in magnificent style by the infantry, supported by a tremendous fire of artillery, and were won at about 5 p.m. and 5.30 p.m. respectively—that is, very late in the afternoon. on the left, at pieter’s hill, the boers still stood desperately at bay. it was not till 6.30, in the 159growing dusk, that the southern, or nearest, crest of the hill, held by the standerton and heidelberg commandos, was carried by a final charge of 300 irish fusiliers, who lost a third of their strength engaged and had all their officers killed or wounded. the northern part of the hill was still obstinately held when the battle came to an end, and was evacuated only during the night.

according to the “official history,” the same unyielding attitude was shown by the most valiant among the defenders of the other two hills, who “clung most stubbornly to the broken ground behind these kopjes,” after their trenches had been carried, and it was in view, we are told, of these signs of dangerous resistance that buller abandoned the idea of a mounted pursuit. he was wrong, it must be concluded, even at this late hour, when darkness and the boer rear-guards must have severely limited effective action; but his real fault lay farther back, in retaining the mounted brigades well in the rear and out of sight all day instead of planting them opposite the boer left flank, where they would have acted at least as a passive menace to the enemy, and might have caused a premature retirement during daylight. we may speculate at will on what might have happened. all we can say with confidence is that the boers were never more formidable than on this culminating day of four months’ strenuous resistance, and that only by using their own fire methods with the utmost energy and determination could our troopers have turned a defeat into a rout.

on that night a general boer retreat set in. among the besiegers of ladysmith, who had not fired a shot, something in the nature of a genuine panic reigned, but the great majority of these had a long start in respect both of time and distance. botha’s commandos, too, gained fully twelve hours’ start, for, in spite of a strong 160appeal from barton on pieter’s hill for a prompt advance by a flying column of all arms, buller made no preparation for a swift movement by the mounted troops. on the morning of the 28th they were still behind the tugela. a block on the pontoon-bridge delayed the irregular brigade under dundonald till 8 a.m., and the regular cavalry brigade under burn-murdoch till 9 a.m. their orders were to work north-west and north-east respectively, not to “pursue.” still, limited as their orders were, they experienced considerable difficulty in carrying them out. botha had organized adequate rear-guards to protect his retreat. dundonald was checked twice within two miles of pieter’s station, and, on the second occasion, had to send for the assistance of burn-murdoch, who, by a later order of buller’s, and against his own repeated requests, had been kept inactive in the gorge between pieter’s hill and the station. the combined brigades having eventually driven off this detachment of the enemy, burn-murdoch moved on to the north-east, but in his turn was brought to a complete standstill at the klip river by the rifle and artillery fire of another boer rear-guard, which was covering the withdrawal of guns and waggons from umbulwana mountain. he held his ground till dusk, prevented the destruction of the wooden bridge which spanned the klip at this point, and informed buller that he intended to remain where he was for the night, and to pursue on the morrow. buller, for inadequate reasons, recalled him. dundonald, meanwhile, still meeting with sporadic opposition, pushed on slowly in the late afternoon towards ladysmith, finally sending in two squadrons, whose arrival denoted the definite relief of the town.

buller had now, definitely and finally, set his face against pursuit. yet even on the morning of march 1 the chances of success, which had steadily diminished, were still considerable. although most of the free state 161forces and a substantial part of the transvaal forces were out of danger, the plain east of ladysmith was still thronged with waggons and guns, the last of which did not reach elandslaagte till nightfall. even as near as modder spruit station siege-guns were entrained as late as 11 a.m. despair reigned in the boer army as a whole. a resolute pursuit must, we can fairly surmise, have led to the capture of a considerable quantity of material and many guns. but we are bound equally to affirm that here, as at every previous stage of these operations, and according to our invariable experience through nearly three years of war in south africa, the measure of success would have been the measure of our ability to overcome defensive fire-tactics with yet more vigorous offensive fire-tactics. that botha, who had effectually covered his retreat on the 28th with parties of the same men who had gone through the nerve-shattering experiences of the previous ten days, culminating in the desperate struggle overnight, would have subsequently allowed his transport and guns to be captured without an effort for their defence, is a tempting, but an altogether illusory, hypothesis. analogy points the other way. it was one of the most striking characteristics of the war that, however great the depression of the undisciplined mass, there were always to be found a few indomitable spirits who were prepared to sell their lives dearly to avert disgrace. we saw this at poplar grove, when the opportunity for our mounted troops, if we consider the relative numbers engaged, while making full allowance for the relative condition of the horses, was far better than at ladysmith. botha himself, the ablest of all the boer leaders, had again and again in the last few months proved his power to restore discipline and nerve among his burghers. his rear-guard tactics, whatever the strength he might have managed to raise, would in form have been those of poplar grove and of his own resistance to burn-murdoch and 162dundonald on the 28th. something more effective than french’s action at poplar grove, and more effective than the action of dundonald and burn-murdoch on the 28th, would have been needed to secure results of really supreme importance. as for the arme blanche, we need not regard it seriously as a contingent factor. if it possessed any utility, it had in the course of the war innumerable opportunities of proving the fact—above all, in cases of pursuit against boer rear-guards. we can scarcely draw negative evidence from occasions where the opportunity was denied.[38]

buller has placed on record his reasons for not undertaking a pursuit.[39] the only one that need concern us is, curiously enough, his insistence on this very point—boer skill in rear-guard actions—a skill which he considered it so futile to combat, that, on this occasion, he thought it not even advisable to try. and he bases his view on his own experience in the first boer war, twenty years before. the admission throws much light on his handling of the mounted troops under his command during the south african campaign, and, in particular, on his dispositions during the battle of pieter’s hill. he had calculated rightly on a victory that day, and, departing from the usual practice, deliberately kept his 163mounted men fresh and concentrated in rear of the army, in order to complete the victory by a pursuit. but the kind of victory he hoped for was one which excluded the possibility of rear-guard actions. in other words, he was a prey to that antiquated habit of thought which was an inheritance from the days prior to the magazine rifle, and which took shape in dreams of massed cavalry on fresh mounts, whirling, sabre in hand, at the psychological moment, through hordes of helpless fugitives. even in 1866 this habit of mind was antiquated. it does not seem to have occurred to him, nor does it seem to occur to some of the present advocates of the arme blanche, that skill in rear-guard actions, often sneeringly alluded to as skill in “evasion,” and always spoken of as if it were some miraculous attribute of the boers, was, in reality, the simple exercise, by the use of horse and rifle combined, of one of the most important of the functions of any corps of mounted troops, cavalry included, especially in the case of the numerically weaker side; and that its counterpart—power to pierce a rear-guard, and drive home a victory, a power correspondingly dependent on the use of horse and rifle combined—is a no less crucial test of mounted efficiency. by these tests, among others, cavalry in future wars will be judged.

defensive skill in the boers suggests the allied question: had they, in the course of the long struggle for ladysmith, shown any new development of offensive power? that is a question we must always be asking, as we contrast the merits of the steel weapon and the firearm in war. as i have often before remarked, there can be no sharp distinction between defensive and offensive action: excellence in the one is wrapped up with excellence in the other. the british seizure of spion kop, for example, was an aggressive stroke; the boer counter-attack was a measure of defensive necessity. regarded in this light, botha’s defence of the line of the tugela 164merits the highest praise. make what allowance we will for defects in british generalship, for the ever-present prejudice against incurring heavy loss of life, and for the extraordinary natural strength of the tugela heights, the fact stands out plainly that no class of troops but mounted riflemen, experts in horse, rifle, and spade alike—and first-class men at that—could, with numbers comparatively so small, have held for so long a position whose extent for purposes of defence cannot be estimated at less than thirty miles. neither european cavalry nor european infantry of that date could have held it for a week against a european force of all arms and of the given superiority—the former from lack of spade and rifle power, the latter from lack of mobility. but measuring the boers by their own standard, did they fully develop their own offensive potentialities?

the answer must be, i think, in the negative. but we cannot in this case afford to be too sweeping or positive. we must remember, here as elsewhere, that the dead-weight of numerical superiority, especially in artillery, gives a force of low mobility, like the british force, a defensive power disproportionately greater than its offensive power. still, there were undoubtedly a few occasions when the boers missed opportunities for counter-strokes. by common consent, i think, the best opportunity of all was on february 23 and 24, when the position of buller’s army, huddled together in hart’s hollow and other parts of the colenso basin, after the magnificent but unavailing assaults of the 23rd, was in the highest degree dangerous.[40] a casual outburst of boer fire on the night of the 24th actually caused a partial panic among the troops in hart’s hollow. according to the german historian, who quotes a german officer present with botha at the time, botha’s reason for not ordering a counter-stroke 165on the 24th was that it would “cost too many lives.” if so, it was a costly error, an irreparable error. but there was much excuse for it. moral administrative weaknesses, from which we were free, had sapped their strength from the first, and among these troops on the tugela at this latter end of february, in spite of botha’s untiring efforts, the tension was becoming unbearable. we have only to contrast the same man, leading tried veterans of the same commandos in latter phases of the war, to understand the full aggressive power that mounted riflemen can develop. nevertheless, we must, as far as we can, disentangle technical from moral causes, and it remains true that up to this point the boers had not brought into line the horse and the rifle as the twin factors of aggressive mobility.

the offensive honours rested with the british infantry. i hope by this time that the reader is beginning to realize how indefinable is the border-line between mounted and dismounted attacks, both of which equally draw their power from that master of modern battle-fields, the rifle. look at wagon hill, where soldiers classed as mounted riflemen were engaged against soldiers classed as infantry, mounted riflemen, and cavalry. here is a case where one almost forgets which class had horses and which had not. when we read of the memorable charge of the devons, we care very little whether they were infantry or mounted infantry, recognizing, as we must, that, in the given conditions, such efforts are within the power of both classes alike. our ambition should be to discover how and when the horse may be made to serve as an engine of still more formidable tactics. look, too, at the infantry charges on february 23 and at the battle of pieter’s hill. watch the old problem of mobility versus vulnerability being worked out in terms of foot-soldiers, and, without rushing to the impracticable extreme of demanding that all riflemen should be 166provided with horses, observe how close is the analogy when the same problem is worked out in terms of horse-soldiers. note how the german historian, from whom nothing will force any compromising allusion to shock as a function of cavalry, lest the whole edifice of cavalry theory should tumble about his ears, slips unconsciously into the deprecation of “shock” in infantry, without sufficient fire-preparation.[41] but for those separate mental compartments, would not some glimmering of the analogy have occurred to him? observe, on the other hand, the fundamental differences between the steel weapon of the foot-soldier and the steel weapon of the cavalry, the efficacy of the former being conditional, not only on the vigour and skill of the previous fire-fight, but on being used at the climax of the fire-fight, still in association with the rifle, and still on foot; the efficacy of the latter a minus quantity, and, for the same reason, everywhere and always, because it was not only incompatible with, but by the habits of mind it engendered, and by the nature of equipment it involved, actively prejudicial to the vigorous offensive use of the firearm.

grasp now the nature of the problem which confronted us in this war. our foes were not only riflemen, but mounted riflemen, comparatively few in numbers, but able both to fight stoutly and to retreat safely when overcome in combat. infantry, though they possess the power to overcome and eject mounted riflemen, have not the power to catch and destroy them, simply because infantry move too slowly. the responsibility for securing complete victory lay with our mounted troops acting as mounted riflemen.

widening our horizon to include the whole area of the war at this period, we perceive that the cavalry theory, so far as it was based on the arme blanche, had collapsed. 167the only and not especially remarkable achievement of that weapon is the pursuit at elandslaagte on the second day of hostilities. everywhere else we have seen it directly or indirectly crippling the cavalry, and the greater the numbers employed and the larger the measure of independence permitted, the more unmistakable is the cause. when the cavalry succeed strategically, as in the ride to kimberley and back to paardeberg, they succeed in spite of disabilities traceable to arme blanche doctrine. when they succeed tactically, as in the colesberg operations and in containing cronje’s force on the eve of paardeberg, they succeed through the carbine, in spite of its inferiority as a weapon of precision. in tactical offence, the paramount raison d’être of the arme blanche, and in reconnaissance, they show marked weakness.