Roman numerals were used, as these could be scribed, chiselled or gouged into the green timbers with ease.
Characters were always added so that IX and XI are both eleven. Many carpenters worked out from the joint so that IX would be paired with XI with the X placed closest to the joint in each case.
There are many hundreds of joints in a house and carpenters
devised a variety of ways to avoid having to use large numbers The most important strategy was for each of the frames to have a different type of mark upon them. If only scribed marks were used then the frames would be differentiated only by tags (see illustration below) but it is more common to find scribed marks, two different size straight chisels, and a curved chisel being used to create four of the sets before tags were necessary.
Tags could be used to differentiate two sets of otherwise similar sets of timbers within the same frame such as the joists on either side of a spine beam. This lowers the highest number to typically seven or eight. An alternative which halves the number of marks on the spine beam is to mark the beam between adjacent pairs of joists and to number these on the side facing the spine number (see illustration).
The carpenter of Daphne Cottage in East Clandon reduced the need for marks by having tagged and untagged marks either side of a spine beam, but only marking I, II on one side of the beam and I' and II' on the other to show how to fit the fully numbered joists.
It would seem likely that each carpenter soon developed his own favourite method of numbering and I had hoped to be able to test this during the dendro-project. However because some houses have secret marking and the survival of others is not comprehensive a larger sample is needed and I anticipate having to visit many more houses before this may be tested.
One generalisation I am prepared to make at this stage relates to the common rafters. These numbers are the least likely to have any form of tag, and I put this down to this being the longest sequence in any group (typically seven rafter pairs per bay, so a four bay building has twenty eight numbers to cut). When using rafter numbers for interpretive purposes remember that the pairs are all the same and were commonly not erected in strict numerical sequence.
The use of different numbering systems had additional
advantages. Firstly at the prefabrication stage it is unlikely that all the frames would have been laid at one time but with this numbering the timbers could be stacked in discrete piles. The unique mark would aid the return of a timber to the correct pile after it had been used to set out an adjoining frame. Secondly the ability to sort out the frames would have been a tremendous help when they were moved from the frame plot to their final site: how otherwise could several hundred pieces of timber be efficiently sorted.