国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

considerably more than natural size. The plant growth, stylised like the hooves,
may be seen in front of the necks and over the backs of the animals, and represents
a forest. Grooved squares form the frame. Although this piece is but little
perforated, the modeling stands out boldly. These two peculiarities are char-
acteristic of a phase that owes much to an observation of nature.

Plate XXVI no. 2 is in strong contrast to such a phase. Here the ibex-heads
are turned full-face, so that the horns are developed on each side. In the centre
of the upper border we find a pair of unconnected horns that serve merely to fill
space. The animals are surrounded by leaves and stand upon a wavy ground.
The back is cut away as much as possible, typical of the Jenissei valley. The
modeling has been replaced by linear drawing in ridges, while the bulk of the
bodies is almost flat. Were there not a horizontal line across the middle, these
bodies would show the same treatment as those on plaques dealing with mytho-
logical scenes (Plate XXI nos. 1 & 2). Here, however, the stylization and geome-
tric forms show progress and warrant a date about the beginning of the second
half of the I millennium.

Still another animal combination is found in the Jenissei as well as in the north
of China, that of two yaks (13). Their heads are turned full-face, the folds of
flesh are comma-shaped, repeated three times. The tuft at the end of each tail
has the same motive. What differentiates this repliqua from its Siberian model
is the emphasis here on linear decoration, which should therefore place it in the
T'ang period. We give the same date to Plate XXVI no. 4, where we find a
single horse that is a repetition of the pair of Plate XXII no.4. The garland
over the contorted form of the animal has also been previously seen on Plate XXVI
no. 6. It is composed of plaited ribbons with animal-heads. There is another
similar ribbon that joins the hind leg to the frame. One type, only found at the
northern frontier and at Minussinsk, and nearly identical in both regions, is that
of Plate XXVII no. 1. It was certainly one of the most popular and most admired
designs of the period. One must not look for this type under our heading in
previous literature on the subject, since it has always been termed an " animal-
combat " (14). But one has only to glance at the hooves, the shape of the heads,
and at the comma-design of the shoulder joints to know that these are two animals
of the same species, two horses. They are playing, one biting at the neck of
the other which in turn bites the foreleg of its companion. In an earlier publication
I called this group " Horses at play " (15). This explanation has been accepted
by Andersson, who goes still farther and calls it " A combat between rutting
stallions " (16). This type is of the usual shape framed in grooved squares, wavy
ground below, and perforated leaves above. In variations of the same scene, the
stylization of the forest usually forms part of the frame, thereby causing less
openwork. In another case, one horse is kneeling and has the forelegs of the
other in his jaws. Since the modeling of this piece is more accentuated than in