countless and inestimable are the chances of war. those who read the story, and still more those who share the dangers, of a campaign feel that every incident is surrounded with a host of possibilities, any one of which, had it become real, would have changed the whole course of events. the influence of fortune is powerfully and continually exerted. in the flickering light of conflict the outlines of solid fact throw on every side the vague shadows of possibility. we live in a world of 'ifs.' 'what happened,' is singular; 'what might have happened,' legion. but to try to gauge the influence of this uncertain force were utterly futile, and it is perhaps wise, and indisputably convenient, to assume that the favourable and adverse chances equate, and then eliminate them both from the calculation.

the 'sirdar's luck' became almost proverbial in the soudan. as the account progresses numerous instances will suggest themselves. it was lucky that the dervishes did not harass the communications, or assail akasha before it was fortified. it was lucky that they fought at firket; that they retired from berber; that mahmud did not advance in january; that he advanced in march; that he did not retire before the battle of the atbara; that the khalifa did not hold the shabluka; that he did not attack on the night before omdurman, and that he did attack at dawn.

but after firket all things were contrary. one unexpected misfortune succeeded another. difficulties were replaced by others as soon as they had been overcome. the autumn of 1896 was marked by delay and disappointment. the state of the nile, the storms, the floods, the cholera, and many minor obstacles, vexed but did not weary the commander. the victory at firket was succeeded by a long pause in the operations. the army had made one spring forward; it must now gather energy for another. the preparations, however, proceeded rapidly. a strong camp was formed at firket. macdonald's brigade occupied suarda two days after the fight, and this place now became the advanced post, just as akasha had been in the first phase of the campaign. the accumuluation of stores at firket and suarda began forthwith. owing to the arrangements which had been made before the engagement it was possible to collect within one week of the action two months' supplies at suarda for the garrison of 2,000 men, and one month's at firket for the 7,000 troops encamped there. thereafter, however, the necessity of hurrying the railway construction and the considerable daily demands of 9,000 men only allowed this margin to be increased very gradually.

the army had now passed beyond the scope of a camel, or other pack-animal, system of supply, except for very short distances, and it was obvious that they could only advance in future along either the railway or a navigable reach of the river, and preferably along both. from the dal cataract near kosheh there is a clear waterway at high nile to merawi. to kosheh, therefore, the railway must be extended before active operations could recommence. a third condition had also to be observed. for the expulsion of the dervishes from kerma and dongola it was desirable that a flotilla of gunboats should co-operate with the land forces. four of these vessels—the tamai, el teb, the metemma, and the abu klea; and three steamers—the kaibar, dal, and akasha, which it was proposed to arm—had, since 1885, patrolled the river from assuan to wady halfa, and assisted in protecting the frontier from dervish raids. all seven were now collected at the foot of the second cataract, and awaited the rise of the river to attempt the passage. to strengthen the flotilla three new and very powerful gunboats had been ordered in england. these were to be brought in sections over the railway to a point above the second cataract, and be fitted together there. it was thus necessary to wait, firstly, for the railway to reach kosheh; secondly, for the nile to rise; thirdly, for the old gunboats to ascend the cataract; fourthly, for the new gunboats to be launched on the clear waterway; and, fifthly, for the accumulation of supplies. with all of these matters the sirdar now busied himself.

the reconstruction of the railway to akasha and its extension beyond this place towards kosheh was pressed forward. by the 26th of june akasha was reached. thenceforward the engineers no longer followed an existing track, but were obliged to survey, and to make the formation for themselves. strong fatigue parties from the egyptian and soudanese battalions were, however, employed on the embankments, and the line grew daily longer. on the 24th of july the first train ran across the battlefield of firket; and on the 4th of august the railway was working to kosheh.

kosheh is six miles south of firket, and consists, like most places in the 'military soudan,' of little more than a name and a few ruined mud-huts which were once a village. on the 5th of july the whole camp was moved thither from the scene of the action. the reasons were clear and apparent. kosheh is a point on the river above the dal cataract whence a clear waterway runs at high nile to beyond dongola. the camp at firket had become foul and insanitary. the bodies of the dead, swelling and decaying in their shallow graves, assailed, as if in revenge, the bodies of the living. the dysentery which had broken out was probably due to the 'green' water of the nile; for during the early period of the flood what is known as 'the false rise' washes the filth and sewage off the foreshore all along the river, and brings down the green and rotting vegetation from the spongy swamps of equatoria. the water is then dangerous and impure. there was nothing else for the army to drink; but it was undesirable to aggravate the evil by keeping the troops in a dirty camp.

the earliest freight which the railway carried to kosheh was the first of the new stern-wheel gunboats. train after train arrived with its load of steel and iron, or with the cumbrous sections of the hull, and a warship in pieces—engines, armaments, fittings and stores—soon lay stacked by the side of the river. an improvised dockyard, equipped with powerful twenty-ton shears and other appliances, was established, and the work—complicated as a chinese puzzle—of fitting and riveting together the hundreds of various parts proceeded swiftly. gradually the strange heaps of parts began to evolve a mighty engine of war. the new gunboats were in every way remarkable. the old vessels had been 90 feet long. these were 140 feet. their breadth was 24 feet. they steamed twelve miles an hour. they had a command of 30 feet. their decks were all protected by steel plates, and prepared by loopholed shields for musketry. their armament was formidable. each carried one twelve-pounder quick-firing gun forward, two six-pounder quick-firing guns in the central battery, and four maxim guns. every modern improvement—such as ammunition hoists, telegraphs, search-lights, and steam-winches—was added. yet with all this they drew only thirty-nine inches of water.

the contract specified that these vessels should be delivered at alexandria by the 5th of september, but, by exertions, the first boat, the zafir, reached egypt on the 23rd of july, having been made in eight weeks, and in time to have assisted in the advance on dongola. the vessels and machinery had been constructed and erected in the works in london; they were then marked, numbered, and taken to pieces, and after being shipped to alexandria and transported to the front were finally put together at kosheh. although in a journey of 4,000 miles they were seven times transhipped, not a single important piece was lost.

the convenience of kosheh on the clear waterway, and the dirty condition of firket, were in themselves sufficient reasons for the change of camp; but another and graver cause lay behind. during the month of june an epidemic of cholera began to creep up the nile from cairo. on the 29th there were some cases at assuan. on the 30th it reached wady halfa. in consequence of this the north staffordshire regiment marched into camp at gemai. their three months' occupation of the town had not improved their health or their spirits. during the sixteen-mile march along the railway track to gemai the first fatal case occurred, and thereafter the sickness clung to the regiment until the middle of august, causing continual deaths.

the cholera spread steadily southward up the river, claiming successive victims in each camp. in the second week of july it reached the new camp at kosheh, whence all possible precautions to exclude it had proved vain. the epidemic was at first of a virulent form. as is usual, when it had expended its destructive energy, the recoveries became more frequent. but of the first thousand cases between assuan and suarda nearly eight hundred proved fatal. nor were the lives thus lost to be altogether measured by the number. [the attacks and deaths from cholera in the dongola expeditionary force were as follow: british troops—24 attacks, 19 deaths; native troops—406 attacks, 260 deaths; followers—788 attacks, 640 deaths.] to all, the time was one of trial, almost of terror. the violence of the battle may be cheaply braved, but the insidious attacks of disease appal the boldest. death moved continually about the ranks of the army—not the death they had been trained to meet unflinchingly, the death in high enthusiasm and the pride of life, with all the world to weep or cheer; but a silent, unnoticed, almost ignominious summons, scarcely less sudden and far more painful than the bullet or the sword-cut. the egyptians, in spite of their fatalistic creed, manifested profound depression. the english soldiers were moody and ill-tempered. even the light-hearted soudanese lost their spirits; their merry grins were seen no longer; their laughter and their drums were stilled. only the british officers preserved a stony cheerfulness, and ceaselessly endeavoured by energy and example to sustain the courage of their men. yet they suffered most of all. their education had developed their imaginations; and imagination, elsewhere a priceless gift, is amid such circumstances a dangerous burden.

it was, indeed, a time of sore trouble. to find the servant dead in the camp kitchen; to catch a hurried glimpse of blanketed shapes hustled quickly to the desert on a stretcher; to hold the lantern over the grave into which a friend or comrade—alive and well six hours before—was hastily lowered, even though it was still night; and through it all to work incessantly at pressure in the solid, roaring heat, with a mind ever on the watch for the earliest of the fatal symptoms and a thirst that could only be quenched by drinking of the deadly and contaminated nile: all these things combined to produce an experience which those who endured are unwilling to remember, but unlikely to forget. one by one some of the best of the field army and the communication staff were stricken down. gallant fenwick, of whom they used to say that he was 'twice a v.c. without a gazette'; polwhele, the railway subaltern, whose strange knowledge of the egyptian soldiers had won their stranger love; trask, an heroic doctor, indifferent alike to pestilence or bullets; mr. vallom, the chief superintendent of engines at halfa; farmer, a young officer already on his fourth campaign; mr. nicholson, the london engineer; long, quaint, kind-hearted 'roddy' owen—all filled graves in halfa cemetery or at the foot of firket mountain. at length the epidemic was stamped out, and by the middle of august it had practically ceased to be a serious danger. but the necessity of enforcing quarantine and other precautions had hampered movement up and down the line of communications, and so delayed the progress of the preparations for an advance.

other unexpected hindrances arose. sir h. kitchener had clearly recognised that the railway, equipped as it then was, would be at the best a doubtful means for the continual supply of a large force many miles ahead of it. he therefore organised an auxiliary boat service and passed gyassas and nuggurs [native sailing craft] freely up the second cataract. during the summer months, in the soudan, a strong north wind prevails, which not only drives the sailing-boats up against the stream—sometimes at the rate of twenty miles a day—but also gratefully cools the air. this year, for forty consecutive days, at the critical period of the campaign, the wind blew hot and adverse from the south. the whole auxiliary boat service was thus practically arrested. but in spite of these aggravating obstacles the preparations for the advance were forced onwards, and it soon became necessary for the gunboats and steamers to be brought on to the upper reach of the river.

the second cataract has a total descent of sixty feet, and is about nine miles long. for this distance the nile flows down a rugged stairway formed by successive ledges of black granite. the flood river deeply submerges these steps, and rushes along above them with tremendous force, but with a smooth though swirling surface. as the nile subsides, the steps begin to show, until the river tumbles violently from ledge to ledge, its whole surface for miles churned to the white foam of broken water, and thickly studded with black rocks. at the second cataract, moreover, the only deep channel of the nile is choked between narrow limits, and the stream struggles furiously between stern walls of rock. these dark gorges present many perils to the navigator. the most formidable, the bab-el-kebir, is only thirty-five feet wide. the river here takes a plunge of ten feet in seventy yards, and drops five feet at a single bound. an extensive pool above, formed by the junction of two arms of the river, increases the volume of the water and the force of the stream, so that the 'gate' constitutes an obstacle of difficulty and danger which might well have been considered insurmountable.

it had been expected that in the beginning of july enough water would be passing down the second cataract to enable the gunboats and steamers waiting below to make the passage. everything depended upon the rise of the river, and in the perversity of circumstances the river this year rose much later and slower than usual. by the middle of august, however, the attempt appeared possible. on the 14th the first gunboat, the metemma, approached the cataract. the north staffordshire regiment from gemai, and the 6th and 7th egyptian battalions from kosheh, marched to the 'gate' to draw the vessel bodily up in spite of the current. the best native pilots had been procured. colonel hunter and the naval officers under commander colville directed the work. the boat had been carefully prepared for the ordeal. to reduce, by raising the free-board, the risk of swamping, the bows were heightened and strengthened, and stout wooden bulwarks were built running from bow to stern. guns and ammunition were then removed, and the vessel lightened by every possible means. a strop of wire rope was passed completely round the hull, and to this strong belt the five cables were fastened—two on each side and one at the bow. so steep was the slope of the water that it was found necessary to draw all the fires, and the steamer was thus dependent entirely upon external force. it was luckily possible to obtain a direct pull, for a crag of black rock rose above the surface of the pool opposite the 'gate.' on this a steel block was fixed, and the hawser was led away at right angles until it reached the east bank, where a smooth stretch of sand afforded a convenient place for the hauling parties. two thousand men were then set to pull at the cables, yet such was the extraordinary force of the current that, although the actual distance in which these great efforts were necessary was scarcely one hundred yards, the passage of each steamer occupied an hour and a half, and required the most strenuous exertions of the soldiers. no accident, however, occurred, and the six other vessels accomplished the ascent on successive days. in a week the whole flotilla steamed safely in the open water of the upper reach.

and now for a moment it seemed that the luck of the expedition had returned. the cholera was practically extinct. the new gunboat zafir was nearly ready at kosheh, and her imposing appearance delighted and impressed the army. on the 23rd of august all the seven steamers which had passed the cataract arrived in a stately procession opposite the camp. almost at the same time the wind changed to the north, and a cool and delicious breeze refreshed the weary men and bore southward to suarda a whole fleet of sailing boats laden with supplies, which had been lying weather-bound during the previous six weeks at the head of the rapids. the preparatory orders for the advance tinkled along the telegraph. the north staffordshire regiment were, to the intense relief of officers and men, warned to hold themselves in readiness for an immediate move. the mounted troops had already returned to the front from the camps in which they had been distributed. at last the miserable delay was over.

from kosheh to kerma, the first dervish position, the distance by river is 127 miles. a study of the map shows that by land marches this can be shortened by nearly forty-one miles; thirty miles being saved by cutting across the great loop of the nile from kosheh to sadin fanti, and eleven miles by avoiding the angle from fereig to abu fatmeh. from kerma to dongola, which latter town was the objective of the expedition, a further distance of thirty-five miles must be traversed, making a total of 120 miles by land or 161 by river. the long desert march from kosheh to sadin fanti was the only natural difficulty by land. although the river from kosheh to kerma is broken by continual rapids, it is, with one interval, freely navigable at half nile. the amara cataract, ten miles beyond kosheh, is easily ascended by sailing boats with a fair wind, and by steamers without assistance. from amara to the kaibar cataract stretches a reach of sixty-five miles of open water. the kaibar cataract is, during the flood, scarcely any hindrance to navigation; but at hannek, about thirty miles further on, the three miles of islands, rocks, rapids, and broken water which are called the third cataract are, except at high nile, a formidable barrier, once this is passed, there is open water for more than 200 miles at all seasons to merawi. the banks of the river, except near sadin fanti, where the hills close in, are flat and low. the eastern bank is lined with a fringe of palm-trees and a thin strip of cultivation, which constitutes what is called 'the fertile province of dongola.' on the other side the desert reaches the water's edge. along the right bank of this part of the river the army was now to move.

the first act of the advance was the occupation of absarat, and on the 23rd of august macdonald's brigade marched thither from suarda, cutting across the desert to sadin fanti, and then following the bank of the nile. the occupation of absarat covered the next movement. on the 26th lewis's brigade was ordered to march across the loop from kosheh to sadin fanti, and reinforce the brigade at absarat. the distance of thirty-seven miles was far too great to be accomplished without a system of watering-places. this the sirdar rapidly organised. water-depots were formed by carrying tanks and water-skins on camels to two points in the desert, and replenishing them by daily convoys. but now a heavy calamity descended on the arrangements of the general and the hopes of the troops.

during the afternoon of the 25th the wind veered suddenly to the south, and thereupon a terrific storm of sand and rain, accompanied by thunder and lightning, burst over the whole of the nubian desert, and swept along the line of communications from suarda to halfa. on the next day a second deluge delayed the march of lewis's brigade. but late on the 27th they started, with disastrous results. before they had reached the first watering-place a third tempest, preceded by its choking sandstorm, overtook them. nearly 300 men fell out during the early part of the night, and crawled and staggered back to kosheh. before the column reached sadin fanti 1,700 more sank exhausted to the ground. out of one battalion 700 strong, only sixty men marched in. nine deaths and eighty serious cases of prostration occurred, and the movement of the brigade from kosheh to absarat was grimly called 'the death march.'

the 'death march' was the least of the misfortunes caused by the storms. the violent rains produced floods such as had not been seen in the soudan for fifty years. the water, pouring down the broad valleys, formed furious torrents in the narrower gorges. more than twelve miles of the railway was washed away. the rails were twisted and bent; the formation entirely destroyed. the telegraph wires were broken. the work of weeks was lost in a few hours. the advance was stopped as soon as it had been begun. at the moment when every military reason demanded speed and suddenness, a hideous delay became inevitable.

in this time of crisis the success of the whole campaign hung in the balance. sir herbert kitchener did not then possess that measure of the confidence and affection of his officers which his military successes have since compelled. public opinion was still undecided on the general question of the war. the initial bad luck had frightened many. all the croakers were ready. 'a jingo government'—'an incapable general'—'another disaster in the soudan'—such were the whispers. a check would be the signal for an outcry. the accounts of 'the death march' had not yet reached england; but the correspondents, irritated at being 'chained to headquarters,' were going to see about that. and, besides all this, there were the army to feed and the dervishes to fight. in this serious emergency, which threatened to wreck his schemes, the sirdar's organising talents shone more brilliantly than at any other moment in this account. travelling swiftly to moghrat, he possessed himself of the telephone, which luckily still worked. he knew the exact position or every soldier, coolie, camel, or donkey at his disposal. in a few hours, in spite of his crippled transport, he concentrated 5,000 men on the damaged sections of the line, and thereafter fed them until the work was finished. in seven days traffic was resumed. the advance had been delayed, but it was not prevented.

on the 5th of september the 1st (lewis) and 2nd (macdonald) brigades moved to dulgo, and at the same time the remainder of the army began to march across the loop from kosheh by sadin fanti to absarat. every available soldier had been collected for the final operation of the campaign.

the expeditionary force was organised as follows:

commander-in-chief: the sirdar

the infantry division: colonel hunter commanding

1st brigade 2nd brigade 3rd brigade 4th brigade

major lewis major macdonald major maxwell major david

3rd egyptians xith soudanese 2nd egyptians 1st egyptians

4th " xiith " 7th " 5th "

ixth soudanese xiiith " 8th " 15th "

xth "

cavalry brigade and mounted forces: major burn-murdoch

cavalry..... 8 squadrons

camel corps.... 6 companies

horse artillery... 1 battery

artillery: major parsons

field artillery... 2 batteries

maxims . ... 1 battery (british)

divisional troops: major currie

north staffordshire regiment.... 1st battalion

the flotilla: commander colville

gunboats... zafir, tamai, abu klea, metemma, el teb

armed steamers... kaibar, dal, akasha

total: 15,000 men, 8 war-vessels, and 36 guns

thus thirteen of the sixteen battalions of the egyptian army were employed at the front. two others, the 6th and xivth, were disposed along the line of communication, holding the various fortified posts. the 16th battalion of reservists remained at suakin. the whole native army was engaged in the war, and the preservation of domestic order in the capital and throughout the khedive's dominions was left entirely to the police and to the british army of occupation. by the 9th all four brigades had reached the rendezvous at dulgo; on the 10th the british regiment, which it was determined to send up in the steamers, was moved to kosheh by rail from sarras and gemai. the sirdar prepared to start with the flotilla on the 12th.

but a culminating disappointment remained. by tremendous exertions the zafir had been finished in time to take part in the operations. throughout the army it was expected that the zafir would be the feature of the campaign. at length the work was finished, and the zafir floated, powerful and majestic, on the waters of the nile. on the afternoon of the 11th of september many officers and men came to witness her trial trip. the bank was lined with spectators. colville took command. the sirdar and his staff embarked. flags were hoisted and amid general cheering the moorings were cast off. but the stern paddle had hardly revolved twice when there was a loud report, like that of a heavy gun, clouds of steam rushed up from the boilers, and the engines stopped. sir h. kitchener and commander colville were on the upper deck. the latter rushed below to learn what had happened, and found that she had burst her low-pressure cylinder, a misfortune impossible to repair until a new one could be obtained from halfa and fitted.

in spite of this, however, the advance was not delayed. on the 13th the 1st, 2nd, and 3rd brigades occupied kaderma. here the flotilla overtook them, and henceforward the boats on the river kept pace with the army on the bank. fareig was reached on the 14th, and as the numerous palms by the water afforded a pleasant shade a halt of two days was ordered. on the 16th the 4th brigade arrived, and the concentration of the force was then complete.

after the annihilation of his strong advanced post at firket, the dervish emir, wad bishara, concentrated his remaining forces in dongola. here during the summer he had awaited, and in the middle of august some small reinforcements under one emir of low rank reached him from omdurman. the khalifa, indeed, promised that many more should follow, but his promises long remained unfulfilled, and the greatest strength that bishara could muster was 900 jehadia, 800 baggara arabs, 2,800 spearmen, 450 camelmen, 650 cavalry—in all 5,600 men, with six small brass cannon and one mitrailleuse gun. to augment in numbers, if not in strength, this small force of regular soldiers, he impressed a large number of the local tribesmen; but as these were, for the most part, anxious to join the government troops at the first opportunity, their effect in the conflict was inconsiderable.

the first sign that the forces were drawing closer was the cutting of the telegraph-wire by a dervish patrol on the 6th of september. on the 10th the sirdar heard that kerma was strongly held. on the 15th of september the egyptian cavalry first established contact with the dervish scouts, and a slight skirmish took place. on the 18th the whole force advanced to sardek, and as bishara still held his position at kerma it looked as if an action was imminent. it was resolved to attack the dervish position at kerma at dawn. although it seemed that only four miles separated the combatants, the night passed quietly. with the first light the army began to move, and when the sun rose the spectacle of the moving masses of men and artillery, with the gunboats on the right, was inspiring. the soldiers braced themselves for the expected action. but no sooner were the village and fort of kerma visible than the report passed along the ranks that they were deserted. rumour was soon merged in certainty, for on reaching kerma it was found that the dervishes had evacuated the place, and only the strong, well-built mud fort attested the recent presence of bishara. whither had he gone? the question was not left unanswered.

half a mile to the southward, on the opposite bank of the river, among the groves of palm-trees ran a long and continuous line of shelter trenches and loopholed walls. the flanks of this new position rested on the deep morasses which extend from the river both on the north and south sides of hafir. a small steamer, a fleet of large gyassas and other sailing vessels moored to the further shore explained what had happened. conscious of his weakness, the prudent emir had adroitly transported himself across the river, and had thus placed that broad flood between his troops and their destruction.

meanwhile the three gunboats—all that now remained of the armed flotilla, for the teb had run on a rock in the hannek cataract—were steaming gradually nearer the enemy, and the army swung to the right, and, forming along the river bank, became spectators of a scene of fascinating interest. at half-past six the horse battery unlimbered at the water's edge, and began to fire obliquely up and across the river. as soon as the first few shells had reached the arab entrenchment the whole line of shelter trenches was edged with smoke, and the dervishes replied with a heavy rifle fire. the distance was, however, too great for their bad rifles and inferior ammunition, and their bullets, although they occasionally struck the ground on which the infantry were drawn up, did not during the day cause any loss to the watching army.

the dervish position was about half a mile in length. as the gunboats approached the northern end they opened fire with their guns, striking the mud entrenchments at every shot, and driving clouds of dust and splinters into the air. the maxim guns began to search the parapets, and two companies of the staffordshire regiment on board the unarmoured steamers dal and akasha fired long-range volleys. now, as on other occasions throughout the war, the dervishes by their military behaviour excited the admiration of their enemies. encouraged by the arrival in the morning of a reinforcement from omdurman of 1,000 black jehadia and 500 spearmen under abdel baki, the dervish gunners stood to their guns and the riflemen to their trenches, and, although suffering severely, maintained a formidable fire.

the gunboats continued to advance, beating up slowly against the strong current. as they came opposite hafir, where the channel narrows to about 600 yards, they were received by a very heavy fire from guns placed in cleverly screened batteries, and from the riflemen sheltered in deep pits by the water's edge or concealed amid the foliage of the tops of the palm-trees. these aerial skirmishers commanded the decks of the vessels, and the shields of the guns were thus rendered of little protection. all the water round the gunboats was torn into foam by the projectiles. the bullets pattered against their sides, and, except where they were protected by steel plates, penetrated. one shell struck the abu klea on the water-line, and entered the magazine. luckily it did not explode, the dervishes having forgotten to set the fuse. three shells struck the metemma. on board the tamai, which was leading, commander colville was severely wounded in the wrist; armourer-sergeant richardson was killed at his maxim gun, and on each boat some casualties occurred. so hot was the fire that it was thought doubtful whether to proceed with the bombardment, and the tamai swung round, and hurried down the river with the current and at full steam to report to the sirdar. the other gunboats remained in action, and continued to shell the dervish defences. the tamai soon returned to the fight, and, steaming again up the river, was immediately hotly re-engaged.

the sight which the army witnessed was thrilling. beyond the flood waters of the river, backed against a sky of staring blue and in the blazing sunlight, the whole of the enemy's position was plainly visible. the long row of shelter trenches was outlined by the white smoke of musketry and dotted with the bright-coloured flags waving defiantly in the wind and with the still brighter flashes of the guns. behind the entrenchments and among the mud houses and enclosures strong bodies of the jibba-clad arabs were arrayed. still further back in the plain a large force of cavalry—conspicuous by the gleams of light reflected from their broad-bladed spears—wheeled and manoeuvred. by the nile all the tops of the palm-trees were crowded with daring riflemen, whose positions were indicated by the smoke-puffs of their rifles, or when some tiny black figure fell, like a shot rook, to the ground. in the foreground the gunboats, panting and puffing up the river, were surrounded on all sides by spouts and spurts of water, thrown up by the shells and bullets. again the flotilla drew near the narrow channel; again the watching army held their breath; and again they saw the leading boat, the metemma, turn and run down stream towards safety, pursued by the wild cheers of the arabs. it was evident that the gunboats were not strong enough to silence the dervish fire. the want of the terrible zafir was acutely felt.

the firing had lasted two hours and a half, and the enemy's resistance was no less vigorous than at the beginning of the action. the sirdar now altered his plans. he saw that his flotilla could not hope to silence the dervishes. he therefore ordered de rougemont—who had assumed the command after colville was wounded—to run past the entrenchments without trying to crush their fire, and steam on to dongola. to support and cover the movement, the three batteries of artillery under major parsons were brought into action from the swampy island of artagasha, which was connected at this season with the right bank by a shoal. at the same time three battalions of infantry were moved along the river until opposite the arab position. at 9 a.m. the eighteen guns on the island opened a tremendous bombardment at 1,200 yards range on the entrenchments, and at the same time the infantry and a rocket detachment concentrated their fire on the tops of the palm-trees. the artillery now succeeded in silencing three of the five dervish guns and in sinking the little dervish steamer tahra, while the infantry by a tremendous long-range fire drove the riflemen out of the palms. profiting by this, the gunboats at ten o'clock moved up the river in line, and, disregarding the fusillade which the arabs still stubbornly maintained, passed by the entrenchment and steamed on towards dongola. after this the firing on both sides became intermittent, and the fight may be said to have ended.

both forces remained during the day facing each other on opposite sides of the river, and the dervishes, who evidently did not admit a defeat, brandished their rifles and waved their flags, and their shouts of loud defiance floated across the water to the troops. but they had suffered very heavily. their brave and skilful leader was severely wounded by the splinters of a shell. the wicked osman azrak had been struck by a bullet, and more than 200 ansar had fallen, including several emirs. moreover, a long train of wounded was seen to start during the afternoon for the south. it is doubtful, however, whether bishara would have retreated, if he had not feared being cut off. he seems to have believed that the sirdar would march along the right bank at once to dongola, and cross there under cover of his gunboats. like all moslem soldiers, he was nervous about his line of retreat. nor, considering the overwhelming force against him, can we wonder. there was, besides this strategic reason for retiring, a more concrete cause. all his supplies of grain were accumulated in the gyassas which lay moored to the west bank. these vessels were under the close and accurate fire of the artillery and maxim guns on artagasha island. several times during the night the hungry dervishes attempted to reach their store; but the moon was bright and the gunners watchful. each time the enemy exposed themselves, a vigorous fire was opened and they were driven back. when morning dawned, it was found that hafir was evacuated, and that the enemy had retreated on dongola.

wad bishara's anxiety about his line of retreat was unnecessary, for the sirdar could not advance on dongola with a strong dervish force on his line of communications: and it was not desirable to divide the army and mask hafir with a covering force. but as soon as the dervishes had left their entrenchments the situation was simplified. at daybreak all the arab boats were brought over to the right bank by the villagers, who reported that bishara and his soldiers had abandoned the defence and were retreating to dongola. thereupon the sirdar, relieved of the necessity of forcing the passage, transported his army peacefully to the other bank. the operation afforded scope to his powers of organisation, and the whole force—complete with cavalry, camels, and guns—was moved across the broad, rushing river in less than thirty-six hours and without any apparent difficulty.

the casualties on the 19th were not numerous, and in a force of nearly 15,000 men they appear insignificant. commander colville was wounded. one british sergeant and one egyptian officer were killed. eleven native soldiers were wounded. the total—fourteen—amounted to less than one per thousand of the troops engaged. nevertheless this picturesque and bloodless affair has been solemnly called the 'battle of hafir.' special despatches were written for it. it is officially counted in records of service as a 'general action.' telegrams of congratulation were received from her majesty and the khedive. a special clasp was struck. of all the instances of cheaply bought glory which the military history of recent years affords, hafir is the most remarkable.

the 20th and part of the 21st were occupied by the passage of the army across the nile. the troops were still crossing when the gunboats returned from dongola. the distance of this place by water from hafir is about thirty-six miles, and the flotilla had arrived opposite the town during the afternoon of the 19th. a few shells expelled the small dervish garrison, and a large number of sailing vessels were captured. the results of the movement of the gunboats to dongola must, however, be looked for at hafir. in consequence of the sirdar's manoeuvre that place was evacuated and the unopposed passage of the river secured.

bishara continued his retreat during the 20th, and, marching all day, reached dongola in the evening. wounded as he was, he re-occupied the town and began forthwith to make preparations for the defence of its considerable fortifications. the knowledge of his employment was not hidden from his enemy, and during the 21st the gunboat abu klea, under lieutenant beatty, r.n., arrived with the design of keeping him occupied. throughout the day a desultory duel was maintained between the entrenchments and the steamer. at daylight on the 22nd, beatty was reinforced by another gunboat, and an unceasing bombardment was made on the town and its defences.

notwithstanding that the army did not finish crossing the river until the afternoon of the 21st, the sirdar determined to continue his advance without delay, and the force accordingly marched twelve miles further south and camped opposite the middle of the large island of argo. at daybreak the troops started again, and before the sun had attained its greatest power reached zowarat. this place was scarcely six miles from dongola, and, as it was expected that an action would be fought the next day, the rest of eighteen hours was welcomed by the weary soldiers. all day long the army remained halted by the palms of the nile bank. looking through their glasses up the river, the officers might watch the gunboats methodically bombarding dongola, and the sound of the guns was clearly heard. at intervals during the day odd parties of dervishes, both horse and foot, approached the outpost line and shots were exchanged.

all these things, together with the consciousness that the culmination of the campaign was now at hand, raised the excitement of the army to a high pitch, and everyone lay down that night warmed by keen anticipations. an atmosphere of unrest hung over the bivouac, and few slept soundly. at three o'clock the troops were aroused, and at half-past four the final advance on dongola had begun.

it was still night. the full moon, shining with tropical brilliancy in a cloudless sky, vaguely revealed the rolling plains of sand and the huge moving mass of the army. as long as it was dark the battalions were closely formed in quarter columns. but presently the warmer, yellower light of dawn began to grow across the river and through the palms, and gradually, as the sun rose and it became daylight, the dense formation of the army was extended to an array more than two miles long. on the left, nearest the river, marched lewis's brigade—three battalions in line and the fourth in column as a reserve. next in order maxwell's three battalions prolonged the line. the artillery were in the centre, supported by the north staffordshire regiment. the gunners of the maxim battery had donned their tunics, so that the lines and columns of yellow and brown were relieved by a vivid flash of british red. macdonald's brigade was on the right. david's brigade followed in rear of the centre as a reserve. the cavalry, the camel corps, and the horse artillery watched the right flank; and on the left the gunboats steamed along the river.

for two hours the army were the only living things visible on the smooth sand, but at seven o'clock a large body of dervish horse appeared on the right flank. the further advance of half a mile discovered the arab forces. their numbers were less than those of the egyptians, but their white uniforms, conspicuous on the sand, and the rows of flags of many colours lent an imposing appearance to their array. their determined aspect, no less than the reputation of bishara, encouraged the belief that they were about to charge.

the disparity of the forces was, however, too great; and as the egyptian army steadily advanced, the dervishes slowly retired. their retreat was cleverly covered by the baggara horse, who, by continually threatening the desert flank, delayed the progress of the troops. bishara did not attempt to re-enter the town, on which the gunboats were now concentrating their fire, but continued to retire in excellent order towards the south and debba.

the egyptian infantry halted in dongola, which when they arrived they found already in the hands of detachments from the flotilla. the red flag with the crescent and star waved once again from the roof of the mudiria. the garrison of 400 black jehadia had capitulated, and were already fraternising with their soudanese captors, whose comrades in arms they were soon to be. while the infantry occupied the town the cavalry and camel corps were despatched in pursuit. the baggara horse, however, maintained a firm attitude, and attempted several charges to cover the retreat of their infantry. in one of these an actual collision occurred, and captain adams's squadron of egyptian cavalry inflicted a loss of six killed on the enemy at a cost to themselves of eight men wounded. the cavalry and camel corps had about twenty casualties in the pursuit. but although the dervishes thus withdrew in an orderly manner from the field, the demoralising influence of retreat soon impaired their discipline and order, and many small parties, becoming detached from the main body, were captured by the pursuers. the line of retreat was strewn with weapons and other effects, and so many babies were abandoned by their parents that an artillery waggon had to be employed to collect and carry them. wad bishara, osman azrak, and the baggara horse, however, made good their flight across the desert to metemma, and, in spite of terrible sufferings from thirst, retained sufficient discipline to detach a force to hold abu klea wells in case the retreat was followed. the dervish infantry made their way along the river to abu hamed, and were much harassed by the gunboats until they reached the fourth cataract, when the pursuit was brought to an end.

the egyptian losses in the capture of dongola and in the subsequent pursuit were: british, nil. native ranks: killed, 1; wounded, 25. total, 26.

the occupation of dongola terminated the campaign of 1896. about 900 prisoners, mostly the black jehadia, all the six brass cannon, large stores of grain, and a great quantity of flags, spears, and swords fell to the victors, and the whole of the province, said to be the most fertile in the soudan, was restored to the egyptian authority. the existence of a perpetual clear waterway from the head of the third cataract to merawi enabled the gunboats at once to steam up the river for more than 200 miles, and in the course of the following month the greater part of the army was established in merawi below the fourth cataract, at debba, or at korti, drawing supplies along the railway, and from railhead by a boat service on the long reach of open water. the position of a strong force at merawi—only 120 miles along the river bank from abu hamed, the northern dervish post—was, as will be seen, convenient to the continuance of the campaign whenever the time should arrive. but a long delay in the advance was now inevitable, and nearly a year was destined to pass without any collision between the forces of the khedive and those of the khalifa.

the success of the operations caused great public satisfaction in england. the first step had been taken. the soudan was re-entered. after ten years of defensive war the dervishes had been attacked, and it was clear that when they were attacked with adequate forces they were not so very terrible after all. the croakers were silent. a general desire was manifested in the country that the operations should continue, and although the government did not yet abandon their tentative policy, or resolve utterly to destroy the khalifa's power, it was decided that, as the road had so far been safe and pleasant, there was at present no need to stop or turn back.

a generous gazette of honours was published. with a single exception, which it would be invidious to specify, all the officers of the egyptian army were mentioned in despatches. sir h. kitchener, colonel hunter, and colonel rundle were promoted major-generals for distinguished service in the field; a special medal—on whose ribbon the blue nile is shown flowing through the yellow desert—was struck; and both the engagement at firket and the affair at hafir were commemorated by clasps. the casualties during the campaign, including the fighting round suakin, were 43 killed and 139 wounded; 130 officers and men died from cholera; and there were 126 deaths from other causes. a large number of british officers were also invalided.