The following readings would seem to be of some interest (cf. Usener, Rheinisches Museum, xvi. 264 sqq.).
iv. a. 16, §2, ὥσπερ καὶ τὰ ἀίδια τῶν φθαρτῶν: the beginning of the line is lost; the Arabic, however, has
كتقدم الاشياء الازلية للاشياء القابلة للفساد
like the priority of eternal things to things which are corruptible.
iv. a. 20, §3, οὔθ’ ὅλως ἀξιόχρεα φαίνεται παντός:
ولا هی فی الجملة مما جحتاج اليه وينتفع به فی الكل
nor are they altogether such as are needed or can be useful in all or on the whole. This would represent πάντως.
23, αὐτὰ δὲ δι’ αὑτῶν οὐδεμίαν ἔχει φύσιν:
انما هی بمنزلة ما نخترعه فنضعه وضعا واما هي فی انفسها فليست لهل طبيعة قائمة
They are merely likely what we invent and set up ourselves, and as for them in themselves, they have not any abiding nature.
Owing to the loss of the preceding words, it is difficult to say whether ἔχει or ἔχειν was read; nor can we be sure that abiding was a supplement of the translator.
The next words εἰ δὲ μή, οὐχ ὥστε συνάπτει τοῖς τῆς φύσεως ὥστ’ ἐμποιῆσαι καθάπερ ζωὴν καὶ κίνησιν αὐτοῖς are represented by fragments:
وان لم تكن لهل [طبيـ]ـعة قائمة فليس . . . تتصل . . . [طـ]ـبيعة حتى تحدث فيها حياة وحركة مثلا
And if they have no permanent nature; the stroke that remains of the last word seems to me to point to يمكنها ان [they cannot] be connected with the [things appertaining to] nature so as to create in them life and motion as it were.
This would be in favour of the conjecture οἷα τε for ὥστε (Hoffmann). The Arabic continues:
وذلک ان العدد نفسه فضلا عن غيره وهو الذی يعتقد [فيه النا]س انه . . . . طـ . . .
