Yet, these imageries, more or less reviving the heavy outer-wall above the outer terrace, and consequently standing comparatively high above the lower series we uncovered in 1890 (now covered again) don’t tell us any story or legend, but allude to symbolical ornaments only.
Notwithstanding, they can’t be said to be without sense, though we may not readily understand them.
They represent numberless, but continually modified repetitions of some motives: a man seated near an incense-offering or a flower-vase, and a man standing between two women, nymphs or servants; both scenes every time separated by a single woman’s image provided with a lotus or another symbol. This lotus may refer to female Bodhisattvas, otherwise I should be inclined to think of apsarasas or celestials, because I don’t see any reason for so many Bodhisattvîs. And yet, why not, provided that they are not taken as personal, legendary or historical Bodhisattvîs.
Don’t we also find them in other ruins (tyanḍi Parambanan, and tyanḍi Sévu), and in the Sari and Pĕlahosan cloisters?[22].
And on the top of the heavy cornice covering these imageries, stand—or formerly stood—from distance to distance, just above the sacrificers, small temples of a completely similar form, each of them containing a deep niche, wherein a Buddha image on a lotus-throne provided with the prabha or disc behind his head.
A square spire with screen-shaped stories reminding us of the Siam pagodae or of some tyaityas also represented on the imageries of our temple, crowned each small temple which had been flanked by two wings with similar but lower spires. And between every two small niche-temples stood—or stands—, just above the groups of the three small images, an altar-shaped stone-block, covered by a bell-shaped dagob which has or had been crowned with a conical column.
The front part of each of these dagob-pedestals has been adorned with a sitting man’s or woman’s image with a flower-vase or an incense-offering, or with both of them.
The back parts of these niche- and dagob temples formed—and they partly still form—an (formerly) uninterrupted cornice which carried the small spires and the dagobs, and beneath, a single wall-opening which, following all the re-enterings of the tridodecahedral, was only interrupted by the four doorways which showed us a repetition (on a larger scale) of the small niche-temples.
These stairs were and are still the weak points of the architecture.
Dissimilar as they are in height and depth of the steps, they sometimes occupy the greater part of the floor of the surrounding galleries. Even the doorways once covering them from terrace to terrace, but which now have for the greater part disappeared, were less proportioned to the whole, and therefore not always equally rich in style, and beauty. It still appears from that which has remained that the side-posts of these doorways—just as those of each niche—had been formed by the serpent’s bodies of two nâgas whose tails ended into the mule of a monster-head we saw above the doorway. We already came across this very same motif on our walk round the niches, and on the banisters of tyanḍi Mĕndut and tyanḍi Pawon, and find it back in all the Buddha temples in Java, especially in those of the plain of Parambanan, and in the ruins of the temple group of this name whose buddhistic character will not be easily acknowledged. At the foot of the doorway (or of the niche) these nâga-heads ended into outward turned mythical monster-heads which, at first sight remind us of elephants rather than of snake-like animals, because their upper lips generally (not always) change into a trunk curled up on their foreheads. Wilhelm von Humboldt and after him all European examiners, among whom the Dutch scholar Leemans, therefore took these monstrous figures for elephant’s heads without perceiving however, that they changed into serpent’s bodies when seen on the side-posts of the doorways; they also didn’t see the relation there was between these heads and the monster-head above the doorways and niches.