wrought terrible havoc among ruins which had previously remained more or less untouched. Lower down, however, we succeeded in tracing remains which had been protected by heavy covering masses of debris, and the employment of large numbers of diggers to clear them was easy. After the difficulties to which my previous explorations at desert sites far away from habitations and water had accustomed me, conditions of work at the ruins of Turfan seemed to me, as it were, quite `suburban'. In the end, we recovered at Toyuk a considerable quantity of fine pieces of frescoes and stucco relievos. Fragments of Chinese and Uigur manuscript texts were numerous.
From Toyuk I proceeded by the middle of December to the important site of Bezeklik, below the village of Murtuk. It occupies a conglomerate terrace on the steep west bank of the stream watering the Kara-khoja oasis, where it breaks in a narrow wild gorge through the barren hill chain overlooking the main Turfan depression. There an extensive series of ruined temple cellas, partly cut into the rock, had their walls decorated with paintings in tempera dating from Uigur times and representing scenes of Buddhist legends and worship in considerable variety of style and subject. In richness and artistic merit they surpassed any similar remains in the Turfan region and recalled the pictorial wealth of the `Thousand Buddhas' of Tun-huang. In r go6 Professor Grünwedel, with his intimate knowledge of Buddhist iconography and art, had carefully studied these fine mural paintings, and a considerable selection of fine fresco panels was then removed to Berlin, as one particularly well-preserved set had been previously by the late Professor Von LeCoq.
For centuries the frescoes had been liable to casual injury
IIIF Color HighRes
