THE EVIDENCE OF ISNĀDS 165

an alternative form of the isnād of the first: 'Amr b. Shu'aib— Amr b. Sharīd—Sharīd—Prophet (Tr. I, 50; Ibn Ḥanbal, iv. 389, 390), and with the mixed isnād Qatāda—'Amr b. Shu'aib—Sharīd— Prophet (Ibn Ḥanbal, iv. 388).¹

A significant example of the arbitrary creation of isnāds occurs in Tr. II, 6 (a) and (b). Here we have first three versions of an Iraqian tradition that 'Alī said, or gave orders to say, prayers over the tomb of Sahl b. Ḥunaif. The prayer over the tomb was an Iraqian invention, but did not become prevalent in Iraq (Muw. Shaib. 166 and Shāfi'ī, loc. cit.). Nor did it become prevalent in Medina, although a tradition from the Prophet in its favour found currency there (Muw. ii. 11 and Zurqānī, ad loc.; Muw. Shaib., loc. cit.). The isnād of this tradition uses the son of Sahl b. Ḥunaif: Mālik—Zuhrī—Abū Umāma b. Sahl—the Prophet said prayers over the tomb of a poor woman. This can be dated with certainty in the generation preceding Mālik. It is mursal; the isnād was later completed by inserting Sahl himself and by creating new isnāds through other Companions (Comm. Muw. Shaib., loc. cit.).

The gradual improvement of isnāds goes parallel with, and is partly indistinguishable from, the material growth of traditions which we have discussed in the preceding chapters; the backward growth of isnāds in particular is identical with the projection of doctrines back to higher authorities.² Generally speaking, we can say that the most perfect and complete isnāds are the latest. As is the case with the growth of traditions, the improvement of isnāds extends well into the literary period, as the following examples will show. The Muhammadan scholars chose to take notice of one particular kind of interference with isnāds, the tadlīs;³ we saw that Shāfi'ī disapproved of it, but minimized its occurrence.

Āthār A.Y.: the editor has collected in the Commentary the parallels in the classical and other collections; a comparison shows the extent of the progressive completion, improvement, and backward growth of isnāds.

Muw. iii. 172 and Muw. Shaib. 364: Mālik—Zuhrī—Ibn Musaiyib and Abū Salama—Prophet; this tradition is mursal. Shāfi'ī (Ikh. 258 f.) has the same, but knows it also with the full isnāds Zuhrī—Abū Salama—Jābir—Prophet, and Ibn Juraij—Abul-Zubair—Jābir—Prophet. According to Comm. Muw. Shaib., Ibn Mājashūn,