String, or nothing
The first question that always comes up when someone wants to make a medieval paternoster is what kind of beads to use. The second question is almost always, "And what kind of thread do I string them on?"
The answer I usually have to give is that we don't have a lot of evidence to go on. In paintings, usually you can't see that kind of detail, because the painter didn't often paint at such a small scale. To complicate matters, we often have no idea just how realistic the painter was trying to be, as evidenced by the number of painted rosaries whose beads seem to be disproportionately big, or of a number like 28, 16, or 39 that doesn't seem to match any set of plausible prayers.
We also don't have much evidence from surviving rosaries or paternosters. As far as I know, all the surviving pre-1600 beads that have survived have been re-strung — except those very few discovered in the last ten or twenty years, which have been recovered from archaeological contexts and were preserved with their cords. (For a long time, remnants of fibers, strings and cloth were simply tossed on the trash heap by archaeologists — fortunately this rarely happens any more.)
This means that just as we can't completely trust what we see in paintings, we can't completely trust the way in which surviving beads may have been re-strung. We hope that whoever re-strung the beads followed the original pattern carefully, but we usually have no evidence whether they did or not.
All this makes it sound like there's no trustworthy evidence at all: well, we do the best we can through comparing different types of evidence, just as we do for other difficult (I was going to say "knotty"!) questions.
For what it's worth, though, we do have a couple of data points. There's at least one 16th-century set of beads in Köln (Cologne, Germany) that are strung on a bright pink silk cord. There is also a 14th-century English example in the Museum of London: eight amber beads still threaded on a tubular tablet-woven string of silk. (From: Crowfoot, Elisabeth, Frances Pritchard and Kay Staniland. Textiles and Clothing, c.1150-c.1450: Medieval Finds from Excavations in London: 4. London: Boydell Press, 2001.)
My own conclusion is that threading beads on silk twist is the best guess in most cases, given what evidence we have. I would hypothesize that in less wealthy contexts, linen or hemp cord would be plausible. Wool is unlikely because it tends to be considerably less strong by comparison, and doesn't take abrasion well as the beads slide back and forth.
As for knots: a good many of the images of medieval rosaries I've seen are clearly strung on a plain thread, with no knots between beads, so that the beads can be slid along the thread one by one as the prayers are said: for example, in a portrait of the Madonna and Child by Ambrogio Fossano. I have heard rumors of period pearl necklaces with knots between the beads, but have never seen anything resembling evidence for this (it tends to be argued on grounds of "common sense," but we all know that what seems sensible to us didn't necessariy seem so to people in history).
By the way, yes it is more than a little odd to see the Infant Jesus playing — most anachronistically — with a rosary; but there are a lot of medieval paintings showing exactly that!
Posts in this series:
String or Nothing
The thread thread
Threads of silk and gold
Of flexwire and time machines