Manuscripts of the Qur'an
The study of ancient manuscripts is no task for amateurs. Even when reading the descriptions by those best qualified, amateurs can misunderstand and/or draw the wrong conclusions. There is no guarantee that this amateur will get it right, but if you see something wrong, let me know by posting a comment below.
François Déroche is one of the few who could be properly described as an expert in manuscripts of the Qur'an. His article with the same title was published in the Encyclopaedia of the Qur'an in 2003. More recently Déroche published the article "Written Transmission" in a collection edited by Andrew Rippin.*
Muslim tradition ascribes the collection of the Qur'an to the third caliph, 'Uthman. Déroche writes that though there are claims of "'Uthmanic Qur'ans" in existence--six in Istanbul alone--historians are not convinced that any of these are authentic. Rather, the earliest Qur'ans which can be dated or have been dated using reliable evidence "are know to originate from the second quarter of the third/ninth century." (WT, 172)
Déroche helpfully explains the ways in which scholars date Qur'an manuscripts: through codicology, palaeography and philology. He writes that the Arabic script lacked standardization in the 1st/7th century. Certain long vowels were not part of the "consonantal skeleton"; there was no system in place for recording short vowels; and the dots which identify consonants were used with varying frequency by early copyists, sometimes not at all.
"The various deficiencies noted in the hijazi-style manuscripts mean that it was not, in fact, possible to adequately preserve the integrity of the Qur'an through writing" at the time when 'Uthman is reported to have definitively established the text in Muslim tradition. (WT, 173-4)
Déroche gives the details of the changes in the qur'anic codex during the early centuries. The earlier defective script was slowly replaced by a full script, with such marks as vocalization, hamza, sukun and shadda gradually added. "The system as we know it today seems to have been introduced towards the end of the third/ninth century." (WT, 175)
One other point which Déroche includes--perhaps more widely known--is that most Qur'ans today go back to a decision of al-Azhar scholars in 1924 to favour one of the many possible variant readings, that of Hafs 'an 'Asim. Early manuscripts of the Qur'an were not taken into account in the preparation of this now-standard Cairo version, and it received no official sanction except by the shaykhs of al-Azhar. (WT, 184)
*The Blackwell Companion to the Qur'an (Oxford: Blackwell, 2006). At the time of posting, Déroche's article in this collection was accessible online through Google books.