Survival of Latin manuscripts

A very interesting new article in Science applies biological principles to the problem of evaluating how many ancient and medieval works have survived. Treating extant manuscripts as "fossils" of works, the author concludes that most of the technical literature even of Late Antiquity has survived. But, why were many older works from Antiquity lost? The author gives a reasonable suggestion:

But why then have so few actually survived from Antiquity? For instance, only one of seven works by Pliny the Elder and only a small fraction of the approximately 2000 works on which he based his Naturalis Historia (77 A.D.) (17), the foremost scientific encyclopedia of the Middle Ages, have survived. The answer may lie in copyists' changeover from papyrus to parchment during the third and fourth centuries (18). Surviving texts may be mostly those similar to Pliny's (19), which happened to have been in demand during and soon after the transition to the new and more durable medium.

Science, Vol 307, Issue 5713, 1305-1307 , 25 February 2005

How Science Survived: Medieval Manuscripts' "Demography" and Classic Texts' Extinction

John L. Cisne

Determining what fraction of texts and manuscripts have survived from Antiquity and the Middle Ages has been highly problematic. Analyzing the transmission of texts as the "paleodemography" of their manuscripts yields definite and surprisingly high estimates. Parchment copies of the foremost medieval textbooks on arithmetical and calendrical calculation closely fit age distributions expected for populations with logistic growth and manuscripts with exponential survivorship. The estimated half-lives of copies agree with Bischoff's paleographically based suggestion that roughly one in seven manuscripts survive in some form from ninth-century Carolingian workshops. On this basis, many if not most of the leading technical titles circulating in Latin probably survived, even from late Antiquity.

Link