away captive from Mesopotamia. This may or may not be true, though the appellation which he received does not support it. If it is untrue, he was very probably the person whom Muhammad's enemies are said to have accused the Prophet of using as his assistant in the composition of certain parts of the Qur'an; for in Surah XVI., An Nahl, 105, we read: "Truly we know that they say, ‘Verily a human being teacheth him.’ The tongue of him at whom they aim is Persian 1, and this [book] is Arabic, clear." If Salman was not a native of Persia, then the language of the verse suffices to prove that there was some Persian in Muhammad's company who was believed to "teach" him a certain portion of what he was then inserting in the Qur'an. We see then that Persian fables were well enough known 2 in Arabia to be recognized by some at least of the Arabs when incorporated into the supposed Divine Revelation. Nor was Muhammad able to give a satisfactory answer to the charge, for no one supposed that the foreigner was teaching him to improve his Arabic style. The charge affected the matter and not the language of the Qur'an. Moreover, as we have proved that Muhammad borrowed legends from the heathen Arabs and from the Jews, there is no reason why he should not be ready and willing to

1 The word 'Ajami properly means Persian, though capable of being applied to other foreigners.
2 Vide pp. 215, 216, 217.

adopt others from Zoroastrian sources. In fact the instances which we have produced in this chapter prove conclusively that he did so, and that these Persian legends, many of which have been shown to be common to the Persians with other branches of the Aryan family of nations, form another of the original sources of Islam.