Chariots were drawn either by two or three, never by four, horses. They seem to have had but a single pole. Where three horses were used, one must therefore have been attached merely by a rope or thong, like the side horses of the Greeks, and, can scarcely have been of much service for drawing the vehicle. He seems rightly regarded as a supernumerary, intended to take the place of one of the others, should either be disabled by a wound or accident. It is not easy to determine from the sculptures how the two draught horses were attached to the pole. Where chariots are represented without horses, we find indeed that they have always a cross-bar or yoke; but where horses are represented in the act of drawing a chariot, the cross-bar commonly disappears altogether. It would seem that the Assyrian artists, despairing of their ability to represent the yoke properly when it was presented to the eye end-wise, preferred, for the most part, suppressing it wholly to rendering it in an unsatisfactory manner. Probably a yoke did really in every case pass over the shoulders of the two draught horses, and was fastened by straps to the collar which is always seen round their necks. These yokes, or cross-bars, were of various kinds. Sometimes they appear to have consisted of a mere slight circular bar, probably of metal, which passed through the pole; sometimes of a thicker spar, through which the pole itself passed. In this latter case the extremities were occasionally adorned with heads of animals. [PLATE XCI., Fig. 1.] The most common kind of yoke exhibits a double curve, so as to resemble a species of bow unstrung. [PLATE XCI., Fig. 2.] Now and then a specimen is found very curiously complicated, being formed of a bar curved strongly at either end, and exhibiting along its course four other distinct curvatures having opposite to there apertures resembling eyes, with an upper and a lower eyelid. [PLATE XCI., Fig. 3.] It has been suggested that this yoke belonged to a four-horse chariot, and that to each of the four eyes (_a a a a_) there was a steed attached; but, as no representation of a four-horse chariot has been found, this suggestion must be regarded as inadmissible. The probability seems to be that this yoke, like the others, was for two horses, on whose necks it rested at the points marked _b b_, the apertures (_c c c c_) lying thus on either side of the animals’ necks, and furnishing the means whereby the he was fastened to the collar. It is just possible that we have in the sculptures of the later period a representation of the extremities (_d d_) of this kind of yoke, since in them a curious curve appears sometimes on the necks of chariot-horses, just above the upper end of the collar. [Illustration: PLATE 91]