1981–1983

3

was needed to determine the range of dates represented by the figurines in the votive deposit. Considerable variation could occur in the handmade features, but on the assumption that a single worker would tend to reproduce the same proportions in his pieces and use repeatedly the same techniques for handling the clay and working his figures, an attempt was made to classify the most diagnostic features: body proportions and poses, and the technique of construction. Further classifications were done for the headdresses worn by both riders and charioteers and for the attributes held in their arms. Around fifty-five variations of headdresses were identified, sixteen human body and nearly thirty horse body types. Each fragment was classified into its appropriate category or series of categories according to how much of the figurine was preserved, and according to the quad in which it had been excavated.

The horses proved to be particularly diagnostic, as little attempt was made to carefully finish the unseen underside, and so individual characteristics in the handling of the clay are clearly discernible. If these characteristics can be taken as identifying features of an individual worker, the date at which he was working can be determined by the style of the latest moulded human face of the riders associated with the horse type. The number and range of earlier face moulds still in use at a given time should also be determinable, as well as the periods in which moulds were no longer available but actual figurines were, to be used as prototypes from which new moulds could be made.

Further work needs to be done. Study of the handmade human and horse heads might show indvudual hands at work, confirming or supplementing the picture formed from the data which has already been compiled. Clay types and painted decoration need to be more carefully documented. Finally, the