This needs some more research, but samim has announced some interesting results for fletchery testing on IRCNet. some main points:
-status of fletchery set makes little to no difference
-status of log does - blessed logs give double results
-on quarrel testing Fletchery and archer class power seemed to make no difference to yield quality (perhaps needs looking into more)
-with blessed log you get one of the following number of quarrels: 18, 27, 36, 45, 54, 63, 72
-with uncursed log you get quarrels of: 9, 13, 18, 22, 27, 31
-this means (1d8+1)*9 quarrels, halved for uncursed log
-arrows seem to give more - perhaps (1d8+6)*9 arrows per blessed log
-lower fletchery values gave same results spread, but with a chance of failure that resulted in 0 missiles. also took same length of time.
-10 turns per quarrel made, 7 turns per arrow, 5 turns per tiny quarrel - length of time is halved for level 18 archers.
Link to raw results: http://www.students.tut.fi/~maki36/adom/fletchery.xls