252 chapter four
rashīd al-Dīn also provides one of the most complete accounts of the
great internal conflicts in the ulus of Jochi. In autumn 1298, toqta mobil-
ised his army to the Dnieper, but could not cross because the river had not
yet frozen for the winter. he made his summer camp on the Don, where
Noghai and his army found him and defeated him. toqta fled beyond the
Volga to Sarai. Noghai was convinced that he had defeated his enemy
and ordered a retreat, entering the crimea where he plundered Solkhat,
then crossed the Dnieper and the Dniester to where he had pitched his
“old tent.” contrary to his expectations, toqta had regrouped his forces
and had begun a counter-offensive. Soon nothing stood between the two
adversaries but the waters of the Dniester, and Noghai tried to infiltrate
troops across the river under the command of his son chaka. toqta’s
scouts discovered the stratagem and the khan ordered his forces across
the Dniester. Both sides suffered considerable losses in the battle, which
toqta won. the aging Noghai was taken captive by a russian soldier and
died as he was being taken to the khan, while his sons sought refuge with
a mere thousand cavalry, fleeing to the hungarians.437
the most surprising account contained in the Mamluk chronicles is
al-Nuwayrī’s, which sets out the two stages of the war as distinct and well-
defined episodes. according to this, in 697 [= 19th october 1297–8th octo-
ber 1298], toqta “declared war against Noghai and his sons, who had more
than two thousand cavalry [.. .] they fought by the river Yaṣī [= pruth],438
which runs between toqta’s lands and Noghai’s lands. toqta and his sol-
diers were defeated, and fled to the Don. Some of them crossed the river
and were saved, some of them drowned. Noghai ordered his troops not
437 tiesenhausen, Sbornik, II, pp. 71–72; the names of the rivers pose few problems,
except for the Dniester, which the editor transliterates as Tarku (with variants). Given that
the arabic-persian script differentiates only minimally between -k- and -l-, it is clear the
two letters have been confused here, so that the root of the word is actually t.r.l.w/u, with
pronunciation determined by the order in which the great steppe rivers are mentioned
in the text; Törlü is the turkish name of the Dniester, while Uzi (strictly: Özü) appears in
the same passage for the Dnieper. for ‘hungarians’ from the term kelar and bashghurd, cf.
ciocîltan, “angaben,” pp. 116 ff. During this period, requests for help reached the Ilkhan
Ghazan from both Jochid commanders; the Ilkhan refused, and offered to mediate (rashīd
al-Dīn/Boyle, p. 102, Soranzo, Papato, p. 392 note 1).
438 for the river names and their identifications, cf. Veselovskiy, Khan, p. 45, Vernadsky,
Mongols, p. 188, ciocîltan, “alanii,” p. 937. polo/Benedetto, pp. 239–444 (= polo/Latham,
pp. 313–318) describes the first clash remarkably precisely, and indeed this account closes
his Il Milione; on p. 315 he says that it took place on the ‘good, wide plain of Nerghi’;
the persian word means a ‘line,ʼ and Vernadsky, Mongols, pp. 187–188, supposed that the
name might refer to the Valley of traian, between the Dniester and the pruth, which is
plausible.