I was digging at the south end of Temple Avenue in weeks 14 and 15. To begin with (c. 60cm depth), there were constant changes in coloration and texture - grey and bright orange squishy clay alternating with iron pan, packed stones and drier earth - and barely any artefacts. About 20cm further down though, these impressive beam slots began to show up! You can clearly see a grey cross shape (room divide?) in the clay natural that would have housed a wooden sleeper with upright posts struck through it; the lack of finds was probably due to the floor level being elevated above these scant foundations. To the left of the picture we later found what seemed to be pit filled with organic material and a scrap of leather. One possible idea is that this area housed cavalry horses in Vindolanda's earliest years. That this is and always was a particularly wet part of the site was borne out by the trench's undulating cross-section and the amount of baling needed each morning after spells of rain: why ever did the Romans build here?!