The achievements of the Muslim armies in Central Asia during the reign of Walid I were due in the first place to the complete co-operation between the directive genius of Hajjaj and the military capacity of Qutayba. Qutayba’s strategic abilities have been somewhat overrated, though the Arabic texts are at no pains to conceal the fact that his gifts fell something short of genius. On more than one occasion we are shown in what constant touch the viceroy was kept with the progress of his armies, and how large a part he took in drawing up the plan of campaign, though the credit of carrying it through to a successful issue rightly belongs to Qutayba. Hajjaj seems to have had the fullest confidence in his lieutenant, and if he did not hesitate to utter reproof and warning when occasion required, he was equally quick to express appreciation of Qutayba’s success. The Arabs of all parties soon realised that behind their general lay the authority of Hajjaj, the wholesome respect inspired by whom prevented any open breach during his lifetime. The second factor which materially assisted the conquests was that in their prosecution Qutayba united all parties in Khurasan, Persians and Arabs, Qays and Yemen. It was no small matter to keep their enthusiasm unabated in the face of campaigns so protracted and severe, nor can the enthusiasm be explained only by the attraction of a rich booty. It is by no means improbable that Qutayba’s success was really due more to his talent for administration than to his generalship. He seems to have realised, as no other Arab governor in the east had yet done, that in such a province as Khurasan the safety and security of the Arab government must depend in the long run on the co-operation of the Persian populace, who formed so great a majority in the country. The bitterness of factional strife had shown how unsafe it was to rely on the support of the Arabs alone, especially in the face of such a movement as Yazid had provoked. By his conciliatory attitude, therefore, Qutayba earned the confidence of the Persians and repaid it with confidence; from his constant employment of Persian agents and his growing preference for Persian governors, it would seem even that he came to regard them as forming the “?Ashira” he lacked among the Arabs. Although it earned him the ill-will of the Arabs and played a great part in his fall, it may be that in this he was instrumental in giving the first impulse to the recovery of a national sentiment amongst the Persians of Khurasan.

The situation in Central Asia was also favourable for a renewal of the attempt to annex to the Arab dominions the rich lands of Transoxania, though it is doubtful how much information the Arabs possessed on this point. In 682, while China, weakened internally by the intrigues of the Empress Wu, had her hands tied by the wars with Tibet, the Eastern or Northern Turks had re-asserted their independence. The new Empire never regained its authority over all the western territories of the former Khans, but by constant campaigns had extended its rule over the Ten Tribes of the Ili and Chu, who, we are told, were “almost annihilated.” In 701 the Eastern Turks invaded Sogdiana, but there is no reason to assume, though it has frequently been suggested, that Muhallab’s forces at Kish were affected by this raid. As the necessity of securing hostages for the safety even of the lines of communication shows, the hostility of the local forces is sufficient to explain all the encounters narrated. The devastation and loss that invariably accompanied these raids must have still further weakened the resources of the subject princes, to whom there was small consolation in the appointment of a son of the Khan to command the Ten Tribes. In any case the unceasing warfare which the Eastern Turks had to wage against the TÜrgesh from 699 to 711 effectually prevented them from sending assistance in response to any appeals for support which may have reached them from Sogdiana[41]. Equally if not more impossible was it for the TÜrgesh to intervene in Sogdiana during the same period[42]. By the “Turks,” as we have seen, the Arab historians mean as a general rule the local inhabitants, amongst whom there may quite possibly have been included at that time Turkish elements. Occasional references to the Khaqan (unless they may be taken to refer to local chiefs, which is improbable) are obvious fakhr-developments. The narrative of 98 A.H. on which the theory of TÜrgesh intervention is mainly based, is a pure Bahilite invention. Finally, the experience of the Arabs in later years shows us that, if the resistance of Sogdiana had been backed by large forces of Turks, it would have been impossible for Qutayba to achieve so large a measure of success.

The conquests of Qutayba fall naturally into four periods:

The recovery of Lower Tukharistan.

The first task before Qutayba was to crush the revolt of Lower Tukharistan. In the spring of 86/705 the army was assembled and marched through Merv Rudh and Talaqan on Balkh. According to one of Tabari’s narratives the city was surrendered without a blow. A second account, which, though not explicitly given as Bahilite, may be regarded as such, since it centres on Qutayba’s brother and is intended to establish a Bahilite claim on the Barmakids, speaks of a revolt amongst some of the inhabitants. This may perhaps be the more correct version, since we hear of Balkh being in a ruinous condition four years later (Tab. 1206. 1). The submission of Balkh was followed by that of Tish, king of Chaghanian, who had probably cooperated with Mufaddal in the attack on Tirmidh the year before. His action was, it seems, inspired by a feud with the king of Shuman and Akharun, in the upper valleys of the Surkhan and Penjab rivers, against whom he hoped to use the Arab troops in return for his assistance to them. Mufaddal had actually projected an expedition against Shuman before his recall, and it was now carried out by Qutayba, who was perhaps the more ready to undertake it since it assured the safety of the southern approach to the Iron Gate. After the submission of the King Ghislashtan, who was of Turkish blood, according to Yuan Chwang, Qutayba returned to Merv alone, leaving the army to follow under his brother Salih, who carried out a number of minor raids on the way. It is obvious that, in spite of Baladhuri’s imaginative account, these raids must be located in the districts neighbouring on the Oxus. The readings in Tabari’s narrative are, however, defective[43]. Having thus isolated Nezak in Badghis, the heart of the revolt, Qutayba spent the winter months in negotiating with him through Sulaym “the Counsellor,” an influential Persian whose skill in conducting the most difficult negotiations proved more than once of the utmost value to Qutayba. Nezak was persuaded to surrender and was conducted to Merv, where peace was concluded on condition that Qutayba would not enter Badghis in person. As a precautionary measure however the governor arranged that Nezak should accompany him in all his expeditions. Thus for the moment at least, the danger of an outbreak in Khurasan was averted, in a manner honourable to both parties, and the son of Peroz took his way back to China to await a more favourable opportunity[44].

The Conquest of Bukhara.

In the following year, Qutayba, first making sure of the crossings at Amul and Zamm, opened his campaigns in Bukhara with an attack on Paykand. From the expressions of Narshakhi, on whose history of this period we may place more reliance since his details as a rule fit in with and supplement the other histories, it can be gathered that the principality of Bukhara was weakened by civil war and invasion. During the minority of Tughshada and the regency of Khatun, the ambitious nobles had struggled between themselves for the chief power; most of the territories, including Bukhara itself, had been seized by the prince of Wardana and the remaining districts seem to have been brought under the rule of Khunuk Khudah, a noble who assumed the title of Bukhar Khudah[45]. Paykand was thus more or less isolated and, from Narshakhi’s account, seems to have been left to its fate. The battle with the Sughdians related in Tabari is an obvious anticipation from the events of the following year. After a siege of some two months the city came to terms with Qutayba, who left it under a small garrison and, according to Tabari’s version, began the return march to Merv. An Émeute in Paykand, however, brought him back at once. It seems reasonable to assume that the citizens, imagining Qutayba’s attack to have been no more than an isolated raid, tried to expel the garrison as soon as he retired. The details given in Narshakhi, that on Qutayba’s advance towards Bukhara a certain citizen, enraged by the insulting conduct of the governor, Warqa? b. Nasr al-Bahili, attempted to murder him, are trivial and unconvincing. Whatever the cause of the revolt may have been, however, Qutayba took a terrible revenge. In accordance with mediaeval practice the renegade city was sacked, its fighting men put to death, and its women and children enslaved. The booty taken from this, the first of the great trading cities of Central Asia to be forcibly captured by the Arabs, furnished inexhaustible material for the exaggerated details of later tradition. The most important part of the spoil was an arsenal of weapons and armour, the excellence of which was such that the “forging of Sughd” appears in contemporary verse alongside the traditional “forging of David” for superlative craftsmanship[46]. With the consent of Hajjaj, these weapons were not included in the division of the booty but used to re-equip the army. The statement that there were only 350 suits of armour in the whole army before this is, however, of Bahilite provenance and scarcely worthy of credence. The exemplary punishment thus meted out by Qutayba to Paykand at the beginning of his career was a stern warning to Nezak and the Sogdians. Those who accepted Arab dominion would be humanely treated, but any attempt at rebellion would be inexorably crushed. Nevertheless the sentence on Paykand was somewhat mitigated in the sequel, as Narshakhi adds that the captives were ransomed by the merchants of Paykand on their return from the annual trading expedition to China, and the city, after lying in ruins for many years, was eventually rebuilt.

The disaster at Paykand roused the princes and merchants of Transoxania to the danger of neglecting the invaders. The feud between Wardana and Bukhara was patched up; round Wardan Khudah, the central figure and organiser of the struggle for independence, gathered the forces of all the nearer principalities. Thus when Qutayba, on renewing his expedition in 88/707, had taken the outlying town of Tumushkath (not Numushkath, which was the earlier name of Bukhara) and Ramithana (or Ramtin), he found his communications cut by the troops of Wardana, Bukhara, and Sughd. It is not, perhaps, impossible that the prince of Farghana should have cooperated with the Sughdians, as stated in Mada?ini’s account. On the other hand the Arabic narratives are far from explicit, and the Sughdians here referred to are much more probably those of Kish than of Samarqand, a suspicion which is confirmed by the famous punning order of Hajjaj: “Crush Kish, destroy Nasaf, and drive Wardan back.” Narshakhi and Ya?qubi give an account of the negotiations between Hayyan an-Nabati, representing Qutayba, and Tarkhun king of Sughd, which is certainly to be put, with Tabari, after the conquest of Bukhara two years later. Throughout all these campaigns there is manifest a tendency, common to the early chronicles of all nations, to exaggerate the numbers and composition of the opposing forces. As usual the Bahilite account carries this to the point of absurdity by introducing a TÜrgesh force of no less than 200,000 men, an obvious anachronism, influenced by the later TÜrgesh invasions. The connection is made clear by the mention of Kur Maghanun, whom we find nearly thirty years later (Tab. II. 1602. 2) as “one of the chiefs of the TÜrgesh.” The true account would seem to be that Qutayba did not attempt to fight a pitched battle, but by dilatory tactics wearied out the allies and gave time for their natural inclination towards disunion to operate, then evaded them by a rapid march through the Iron Gate and, except for a rearguard skirmish with the enemy’s cavalry, got his army clear across the river at Tirmidh. The appointment of ?Abdur-Rahman ibn Muslim to command the rearguard gives us the clue, as it was to this brother that Qutayba regularly entrusted all the most difficult commands. In the following year Qutayba was still unable to make headway against the united forces of Wardan Khudah, Kish and Nasaf, and after protracted fighting (in spite of the double victory claimed by the Bahilites) returned to Merv. For this weakness he was severely reprimanded by Hajjaj, who, with the aid of a map, drew up a plan of attack. The invasion of 90/709 seems to have taken Wardan Khudah by surprise, as the Muslim army was able to advance at once to the siege of Bukhara. There is some ground for the conjecture, however, that the death of Wardan Khudah had occurred in the interval and that Qutayba was opposed only by the local forces[47]. This may also explain the hesitation of the forces of Samarqand to intervene. The battle before the walls of Bukhara is described by Tabari in a long Tamimite tradition reminiscent of the ancient “days,” but the actual capture of the city is left to be inferred. This siege is transferred to Wardana by VÁmbÉry (cf. Heart of Asia p. 52) probably on the authority of the Persian Tabari (Zotenberg IV. 165), but Narshakhi, Tabari and all other authorities quite definitely refer to Bukhara. Abu ?Ubayda’s tradition (Bal. 420) of capture by treachery is at best a confusion with the capture of Samarqand. All the details given in Narshakhi relative to Qutayba’s organisation of Bukhara do not refer to this year; most probably the only immediate measures taken were the imposition of a tribute of 200,000 dirhems and the occupation of the citadel by an Arab garrison.

A diplomatic success followed the victory at Bukhara. Tarkhun, king of Samarqand, opened negotiations with Qutayba, who was represented by the commander of his Persian corps, Hayyan an-Nabati, and terms were agreed upon, probably on the basis of the old treaty made by Salm ibn Ziyad. Tarkhun gave hostages for the payment of tribute and Qutayba began the march back to Merv.

Consolidation and Advance.

If the Arabs returned in the autumn of 90/709 elated with their success, they were soon given fresh cause for anxiety. Nezak, finally realising that all hope of recovering independence must be extinguished if Arab rule was strengthened in Khurasan, and perhaps putting down to weakness Qutayba’s willingness to gain his ends if possible by diplomacy, determined on a last effort to overthrow Muslim sovereignty in Lower Tukharistan, at the moment when it was least to be expected. Having obtained permission to revisit his home, he left Qutayba at Amul and made for Balkh, but escaped to Tukharistan in order to avoid re-arrest. From here he corresponded with the rulers of Balkh, Merv Rudh, Talaqan, Faryab, and Juzjan, urging them to undertake a concerted rising in the spring. The king of Chaghanian seems to have refused to countenance the conspiracy, but the weak Jabghu of Tukharistan was induced, possibly by force, to make common cause with Nezak, who hoped doubtless by this means to unite all the subject princes in defence of their suzerain.

Qutayba’s army was already disbanded and the winter was setting in. All that he could do was to despatch the garrison at Merv, some 12,000 men, under ?Abdur-Rahman, with instructions to winter in Balkh, where they could counter any immediate move by Nezak, and advance into Tukharistan in the spring. This resolute action made Qutayba master of the situation and so intimidated the rebels that when, in the early spring, the Arabs marched through the disaffected districts, scarcely a blow was struck and the princes either submitted or fled. The inhabitants were granted a complete amnesty except at Talaqan, concerning which the traditions are hopelessly confused. According to one account, a band of robbers were there executed and crucified, but it is possible that it was selected for special severity because there alone the revolt had openly broken out[48]. There was probably also some reorganization of the administration of Lower Tukharistan, in the direction of conferring fuller powers on the Arab governors installed in each district, though the native princes continued to exercise a nominal authority. From Balkh, Qutayba marched forward and rejoined ?Abdur-Rahman. With the assistance of the lesser princes they pursued and captured Nezak, who was subsequently executed on direct orders from Hajjaj, in violation of Qutayba’s promise of pardon[49]. How little this action was condemned by the prevailing spirit of the age, however, is shown by the contemporary poems quoted by Tabari, lauding the “defender of the precincts of Islam” and comparing his action to the measures formerly adopted against the Jewish tribes of Madina. Yet even at this time we find traces of the new spirit that was to make itself more felt in later years, and hear voices raised, like Thabit Qutna’s, against the “treachery that calls itself resolution.” Tabari inserts at this point the narrative of the putting to death of the hostages of Juzjan, in retaliation for the murder of the Arab hostage in Juzjan, a much more excusable incident. Baladhuri puts it at the beginning of Qutayba’s career, however, as though it belonged to the first pacification of Lower Tukharistan, so that its position in Tabari may possibly be due to its superficial similarity with the case of Nezak. The results of this expedition were of the greatest importance: not only was Nezak’s scheme crushed and Lower Tukharistan henceforth incorporated in the Arab Empire, but also for the first time Arab authority was extended over the Jabghu and his immediate vassals in the Oxus basin. The former, exiled to Damascus, formed a valuable hostage against any attempt to regain independence, and it seems not improbable that the king of Chaghanian was made regent for the young Jabghu (see above, p. 9), ?Abdur-Rahman was appointed governor of Balkh, in order to supervise the administration of the new province.

Qutayba had hardly returned to Merv before he was called to deal with yet another revolt. The king of Shuman, taking advantage of the difficulties of the Arabs, or of their absence in the southern mountains, had re-asserted his independence in spite of the conciliatory offers of Salih ibn Muslim. The full weight of Qutayba’s power was now employed to crush him. His stronghold was attacked with siege artillery, the king himself killed in a sortie and the garrison put to the sword. From this point Shuman and Akharun gradually drop out of the Arabic narratives altogether. Qutayba then resumed his march through the Iron Gate, reduced the districts of Kish and Nasaf, and revisited Bukhara. There seems to have been continual friction between the Arab garrison and the population[50] and it was felt that a drastic re-organisation was necessary. Tughshada, though still a youth, was restored to the position of Bukhar-Khudah, and the leaders of the hostile party (more probably that of Khunuk Khudah than Wardan Khudah) were put to death. By this means, Qutayba no doubt hoped to secure compliance and docility in the native administration. Tughshada had been raised to the throne by the Arabs and it might be expected that he would side with them in consequence. A more solid guarantee for the permanence of the conquest, however, was the establishment of a military colony in Bukhara. Following the precedent set in the colonization of Merv, Arabs were lodged in the houses of the inhabitants, and it is said that the latter were encouraged to attend the Friday prayer and behave as Muslims by the distribution of a small gratuity. The Kushan merchants left their homes and property rather than comply with these orders and founded a new city outside the walls, but it is evident that the Islamization of the city was not yet so thorough as the traditions assert[51]. The building of the Mosque and the organization of the Friday services are dated by Narshakhi in 94 A.H., which points to a further organization of the city after the capture of Samarqand. The organization of the new territories proceeded, in fact, pari passu with the extension and consolidation of the conquests. So long as the Arab authority was insecure in Cisoxania, it was out of the question to establish either military colonies or an elaborate administration beyond the river. Consequently, it was only now that the failure of Nezak’s revolt had definitely secured the Arab dominion in the former Ephthalite lands that it was possible to take the decisive step of settling an Arab garrison in Bukhara. The regularity with which each step followed the last suggests that it was done according to a prearranged plan, or at least that some attention had been devoted to the question of the administration of the occupied territories in the event of the success of the military operations.

Qutayba’s reorganization was not confined to the civil government, however, but extended to the army as well. Hitherto the jealousy of the Arabs for their exclusive rights as a warrior caste had strictly limited the number of Persians in the armies, apart from the clients and camp followers. Thus we are told (Tab. 1290. 20) that the armies of Khurasan at this period were composed as follows: from Basra-Ahl al-?Aliya, 9,000; Bakr, 7,000; Tamim, 10,000; ?Abd al Qays, 4,000; Azd, 10,000: from Kufa, 7,000: and alongside these 47,000 Arabs only 7,000 Mawali, commanded by Hayyan-an-Nabati, who is called variously a Daylamite and a native of Khurasan. Now, however, Qutayba imposed, first on Bukhara, and later on each successive conquest, the obligation of providing an auxiliary corps of local troops, amounting usually to some ten or twenty thousand men, to serve with the Arab armies. It is possible, if the story be true, that this was suggested by the precedent set by Sa?id b. ?Othman in the conquest of Samarqand, but more probable that it represents an entirely new departure in the East, though it had long been a practice in other spheres of the Arab conquests.

We are given no hint of the motives which led to the adoption of the new system, though it would seem that they must have been of some force. Possibly it was no more than a desire to keep the native armies occupied in the service of the Arabs rather than risk a revolt in their rear. Hajjaj and Qutayba perhaps realised too that the Arab forces by themselves, after taking four years to reduce Bukhara alone, were insufficient to ensure success in the greater task of subduing Samarqand. Under the new system—which recalls Pan-ch?ao’s famous aphorism “Use barbarians to attack barbarians”—each conquest in turn made the next more easy. The rapidity of Qutayba’s later conquests in contrast with the early period is thus explained. It is just possible that in this plan Qutayba had an ulterior motive as well: the formation of a Persian army, trained on the same lines as the Arab forces, but more devoted to the person of the governor and able to take his part against the Arabs. How very nearly this plan succeeded, even in Qutayba’s own case, the sequel was to show.

The practice of raising native levies, once started, appears to have become general in Khurasan. We have no information as to when the local forces of Khurasan and Lower Tukharistan were incorporated in the army, nor in what proportions, but we have frequent evidence of their presence and increasing prestige in the wars of the next forty years[52]. On the other hand, though contingents from the towns of Sogdiana were used by later governors if they were available, as in 106 and 112 A.H., in view of the weaker hold of the Arabs on Transoxania Sogdian troops never formed a regular division of the Arab forces up to the end of the Umayyad period. This distinction between the two subject Iranian groups became, as will be seen, of some importance when the ?Abbasid propaganda began to tamper with the loyalty of the armies of Khurasan.

While Qutayba was occupied with the new organization of Bukhara, a detached force, sent under ?Abdur-Rahman from Kish to Samarqand to exact from Tarkhun the tribute agreed upon in the previous year, successfully accomplished its mission. ?Abdur-Rahman, after restoring the hostages to Tarkhun, rejoined his brother at Bukhara, whence they returned to Merv for the winter.

One important vassal of Tukharistan, who had long been a thorn in the side of Hajjaj, still remained unsubdued. This was Rutbil or Zunbil, the Turkish ruler of Zabulistan[53]. In 91, the viceroy united Sijistan to the province of Khurasan, with instructions to Qutayba to undertake a campaign in person against Rutbil. In the following year, therefore, the expeditions into Transoxania were interrupted, and the army again marched southwards. To Qutayba’s great relief (for he disliked to undertake a campaign against this formidable foe who had made Sijistan “an ill-omened frontier”) Rutbil hastened to tender his submission, and at the same time sent an embassy to convey his homage to the Emperor of China[54]. Recognition of Arab suzerainty over Zabulistan involved of course only the payment of a fixed tribute, and no attempt was made at a permanent occupation.

Meanwhile a serious situation had arisen in Sughd. The merchants and nobles of Samarqand had resented the weakness of their king and the payment of tribute: in Qutayba’s absence the party for resistance À outrance gained the upper hand, and Tarkhun, deposed on the ground of incapacity, committed suicide. The choice of the electors fell on Ghurak[55], a prince of whom we would gladly have known more. Under the ever increasing difficulties with which he was confronted during his twenty-seven years of rule, his consummate handling of the most confused situations shows him to have been at once statesman and patriot, and preserved his kingdom from repeated disaster. The action of the Sughdian nobles, however, the Arabic account of which is confirmed by the Chinese records, constituted a challenge to Arab pretensions which Qutayba could not be slow in answering. These considerations clearly disprove the partial tradition of Abu ?Ubayda (Bal. 422), to the effect that Qutayba treacherously attacked Khwarizm and Samarqand in spite of the treaties of Sa?id ibn ?Othman, and the argument based upon it by van Vloten in La Domination Arabe, must also, in consequence, be somewhat modified.

The winter of 93/711, therefore, was spent in preparations for an expedition against Samarqand, but before the opening of the campaigning season, Qutayba received a secret mission from the Khwarizm Shah, who offered to become tributary if the Arabs would rid him of his rebellious brother Khurrazadh. Qutayba agreed, and after publicly announcing his intention of invading Sughd, suddenly appeared at Hazarasp. The followers of the Khwarizm Shah were persuaded to offer no resistance for this year, at least, and accepted the terms, which included, in accordance with the new scheme, the provision of a corps of 10,000 ablebodied men as well as the usual tribute. Qutayba remained at the capital[56] until the army was collected, while ?Abdur-Rahman was employed, according to Tabari, in reducing the king of Khamjird, who from the parallel account in Baladhuri is to be identified either with Khurrazadh, or at least with his party. The Persian Tabari adds a long and doubtless legendary narrative of his surrender. Four thousand prisoners were taken and butchered, probably by order of the Khwarizm Shah.

The later history of Khwarizm under Qutayba’s rule is an unhappy one. His first governor Iyas b. ?Abdullah, proved too weak for his post, and on Qutayba’s withdrawal the Khwarizmians rose in revolt and put to death the king who had betrayed them. Iyas was recalled in disgrace, together with the Persian Hayyan an-Nabati, who had been associated with him, and Qutayba’s brother ?Abdullah (in Baladhuri ?Ubaydullah) was appointed as temporary regent until, after the capture of Samarqand, a strong force under al-Mughira b. ?Abdullah could be sent to effect a reconquest. Qutayba’s retribution on this occasion exceeded even the terror of Paykand and Shuman. We are told by Al-Biruni that the educated classes and more cultured elements in Khwarizm were slaughtered almost to extinction. He refers this by implication to the second expedition of Qutayba (though it does not appear that the governor led the expedition in person), which is borne out by what we know of Qutayba’s methods in similar cases, while there is no instance in his career of such an action on a first conquest. It was in all probability the educated classes (including no doubt the hierarchy) who led the revolt against the traitor king and thus met with the severest punishment. The dynasty, however, was maintained, and it is not improbable that the Arab colony of which we hear shortly afterwards was settled in Khwarizm at the same time[57].

The booty from the first expedition into Khwarizm was enough to satisfy Qutayba’s troops, who demanded to be allowed to return to their homes, but a sudden thrust at Samarqand promised such success that Qutayba and his leaders decided to make the attempt. The Sughdian army had apparently been disbanded, and under cover of a false movement of the advance guard, the Arabs marched directly on Samarqand. The advance guard under ?Abdur-Rahman numbered 20,000 men, while the main body included the new Persian contingents from Khwarizm and Bukhara. The march occupied only a few days and the slight resistance encountered did not prevent the Arabs from proceeding at once to invest the city. Ghurak conducted the defence with vigour, however, and appealed to Shash and Farghana for assistance, reminding them that Samarqand was the bulwark of the Jaxartes valley. A strong force was despatched from Shash with the intention of making a surprise attack on the Arab camp, but was ambushed at night by a picked troop of Arabs and almost annihilated. This reverse, together with the continuous bombardment to which they were subjected, disheartened the Sughdians, but the wall had been breached and an entrance almost effected by the Arabs, stoutly assisted by their new Iranian divisions, before Ghurak sued for peace. Qutayba’s demands were unexpectedly light—an annual tribute, stated in widely varying amounts, and a strong corps of Sughdians, together with a stipulation that the city should be cleared of its fighting men while the Arabs built a mosque and celebrated the ritual prayers. Once within the gates, however, Qutayba refused to restore the city to Ghurak: a strong garrison was established in the citadel, under the command of ?Abdur-Rahman (so Ya?qubi; in Tabari ?Abdullah) and drastic orders were issued excluding all unbelievers except under strict surveillance, doubtless with the intention of avoiding a repetition of the friction that had occurred at Bukhara. Ghurak either could not or would not place himself in the humiliating position of Tughshada, and with his retinue, accompanied possibly by the merchants, withdrew from Samarqand altogether and built a new city, Farankath, some four farsakhs distant in the direction of Ishtikhan[58]. Qutayba’s double-dealing on this occasion, however, tarnished his reputation among both Persians and Arabs, far more than his severity to Paykand and Khwarizm, and left a rankling memory in Sughd. In order to avoid the stigma of treachery attaching to their hero the Bahilite tradition relates this expedition in an entirely different version[59]. Qutayba, we are told, after marching down the right bank of the Oxus and collecting his army at Bukhara, advanced to Rabinjan where he was met by the Sughdians under Ghurak, supported by the troops of Shash and Farghana and the Turks. The enemy retired on Samarqand but engaged in constant rearguard actions, the city being finally entered by force after a decisive battle in the suburbs. Though this account is at first sight borne out to some extent by Ghurak’s own narrative in his letter to the Emperor of China, in which he claims an initial success against the Arabs, but was unable to prevent their advance, both statements must be regarded as exaggerations in opposite interests. At all events it is quite certain that none but Sughdian troops were involved at first.

A further development of the Bahilite tradition has given rise to some controversy. According to this, Ghurak appealed for help not only to Shash but also to the Khaqan, and the squadron sent from Shash appears as a force of Turks, commanded by a son of the Khaqan. This is, of course, an obvious exaggeration on the former narrative. In the Turkish Orkhon inscriptions, however, an expedition under the prince KÜl-tegin into Sogdiana “to organize the Sogdian people” is mentioned, following on a successful campaign against the TÜrgesh in 710/711. Marquart endeavours to prove that this expedition occurred in 712 and is, in fact, corroborated by the Bahilite tradition. Professor Houtsma has raised several objections to this view, the most important being that the chronology of the inscriptions has to be manipulated to allow of this date, as the natural date to assume from the context is at latest 711. These, together with the considerations mentioned above, render Marquart’s hypothesis absolutely untenable.

A second suggestion has been put forward by Professor Barthold, to which, however, Professor Houtsma’s objections would apply with equal force[60]. In the narrative of the historian Ya?qubi (II. 344), there is a brief notice as follows: “Qutayba appointed his brother ?Abdur-Rahman ibn Muslim governor of Samarqand, but the men of Samarqand treacherously revolted against him, and Khaqan, king of the Turks, attacked him also. He wrote to Qutayba, but Qutayba waited until the winter cleared, then marched to join him and routed the army of the Turks.” Professor Barthold takes the view, therefore, that this is the expedition referred to in the inscriptions, and attributes the failure of the Turks to the disastrous effects of a winter campaign in a devastated land, which so severely disabled them that they could not face the formidable army that took the field under Qutayba in the spring. It is questionable, however, how far Ya?qubi’s narrative may be trusted. None of the other historians give the slightest hint of this invasion, nor were the results such as we should expect after a Sughdian revolt. There was no ruthless reconquest, no stamping out of rebellion in blood. Neither does the general tenor of Ya?qubi’s accounts of Qutayba inspire confidence. They are not only confused in detail and chronology—the capture of Samarqand, for instance, is dated 94 A.H.—but in some cases are taken from what we know to be the Bahilite tradition, and in others, such as the narrative under discussion and the account of the conquest of Khwarizm, follow a tradition which seems irreconcilable with our other information. While it cannot be said definitely therefore, that Ya?qubi’s statements in this case contain no truth, it is certainly preferable to regard them as a later development of the narrative, on the lines of the Bahilite tradition.

If the chronological objections raised by Professor Houtsma are sound, there remains still a third possible solution, which, however, as there is no corroborative evidence from either the Arabic or Chinese sources, must remain nothing but a hypothesis. It is surely quite tenable that KÜl-tegin’s “organization of the Sogdian people” had something to do with the deposition of Tarkhun and appointment of Ghurak. With Sogdian trade playing the most important part which we know in the Turkish lands, it would be well worth while to try to prevent the Arabs from obtaining control over it. The very unexpectedness of the description given to this expedition shows clearly that there was some motive for “organization” and it is difficult to see what other motive there could have been. These circumstances would render it quite probable that Ghurak did, in fact, appeal to the Khaqan for assistance against the Arabs, but it seems that the growing power of the TÜrgesh barred the way into Sogdiana against the Northern Khanate for the remainder of its short existence.

By the conquest of Samarqand Qutayba finally established his position in Transoxania. It must not be assumed, however, as many of the Arab historians give the impression of assuming, that the holding of Samarqand meant the conquest of Sughd. All that had been done was to settle an Arab garrison in a country as yet unfriendly. It was the duty of the commanders at Samarqand gradually to extend their authority over the whole district of Sughd by expeditions and razzias[61]. There was thus a radical difference between the conquest of Bukhara and that of Samarqand. The former was the result of a series of campaigns in which the resources of the country had been exhausted and the province annexed piecemeal. The whole population had become subjects of the Arabs and were under constant surveillance: Tughshada himself held his rank on sufferance and was compelled to maintain at least an outward show of loyalty. But Samarqand had been captured in one swift thrust; Sughd as a whole was still unsubdued and only from policy acknowledged the suzerainty of the Arabs for the time being. “Ghurak at Ishtikhan was free to turn either to the Arabs or to the Turks”[58]. Nevertheless in the years that followed there is evidence that friendly relations were formed between the Arab garrison and many of the local leaders and inhabitants[62]. The whole country, however, had suffered terribly in the constant invasions and counter invasions. A contemporary poet gives a vivid picture of its dissipated wealth, its ruined and desolate lands:

“Daily Qutayba gathers spoil, increasing our wealth with new wealth: A Bahilite who has worn the crown till the hair that was black has whitened. Sughd is subdued by his squadrons, its people left sitting in nakedness.... As oft as he lights in a land, his horse leave it furrowed and scarred.”