The building which we had entered was one of great size and elaborateness, and gave us an impressive notion of the architecture of that nameless geologic past. The inner partitions were less massive than the outer walls, but on the lower levels were excellently preserved. Labyrinthine complexity, involving curiously irregular differences in floor levels, characterised the entire arrangement; and we should certainly have been lost at the very outset but for the trail of torn paper left behind us. We decided to explore the more decrepit upper parts first of all, hence climbed aloft in the maze for a distance of some 100 feet, to where the topmost tier of chambers yawned snowily and ruinously open to the polar sky. Ascent was effected over the steep, transversely ribbed stone ramps or inclined planes which everywhere served in lieu of stairs. The rooms we encountered were of all imaginable shapes and proportions, ranging from five-pointed stars to triangles and perfect cubes. It might be safe to say that their general average was about 30 × 30 feet in floor area, and 20 feet in height; though many larger apartments existed. After thoroughly examining the upper regions and the glacial level we descended story by story into the submerged part, where indeed we soon saw we were in a continuous maze of connected chambers and passages probably leading over unlimited areas outside this particular building. The Cyclopean massiveness and giganticism of everything about us became curiously oppressive; and there was something vaguely but deeply unhuman in all the contours, dimensions, proportions, decorations, and constructional nuances of the blasphemously archaic stonework. We soon realised from what the carvings revealed that this monstrous city was many million years old.
We cannot yet explain the engineering principles used in the anomalous balancing and adjustment of the vast rock masses, though the function of the arch was clearly much relied on. The rooms we visited were wholly bare of all portable contents, a circumstance which sustained our belief in the city’s deliberate desertion. The prime decorative feature was the almost universal system of mural sculpture; which tended to run in continuous horizontal bands three feet wide and arranged from floor to ceiling in alternation with bands of equal width given over to geometrical arabesques. There were exceptions to this rule of arrangement, but its preponderance was overwhelming. Often, however, a series of smooth cartouches containing oddly patterned groups of dots would be sunk along one of the arabesque bands.
The technique, we soon saw, was mature, accomplished, and aesthetically evolved to the highest degree of civilised mastery; though utterly alien in every detail to any known art tradition of the human race. In delicacy of execution no sculpture I have ever seen could approach it. The minutest details of elaborate vegetation, or of animal life, were rendered with astonishing vividness despite the bold scale of the carvings; whilst the conventional designs were marvels of skilful intricacy. The arabesques displayed a profound use of mathematical principles, and were made up of obscurely symmetrical curves and angles based on the quantity of five. The pictorial bands followed a highly formalised tradition, and involved a peculiar treatment of perspective; but had an artistic force that moved us profoundly notwithstanding the intervening gulf of vast geologic periods. Their method of design hinged on a singular juxtaposition of the cross-section with the two-dimensional silhouette, and embodied an analytical psychology beyond that of any known race of antiquity. It is useless to try to compare this art with any represented in our museums. Those who see our photographs will probably find its closest analogue in certain grotesque conceptions of the most daring futurists.
The arabesque tracery consisted altogether of depressed lines whose depth on unweathered walls varied from one to two inches. When cartouches with dot-groups appeared — evidently as inscriptions in some unknown and primordial language and alphabet — the depression of the smooth surface was perhaps an inch and a half, and of the dots perhaps a half-inch more. The pictorial bands were in counter-sunk low relief, their background being depressed about two inches from the original wall surface. In some specimens marks of a former colouration could be detected, though for the most part the untold aeons had disintegrated and banished any pigments which may have been applied. The more one studied the marvellous technique the more one admired the things. Beneath their strict conventionalisation one could grasp the minute and accurate observation and graphic skill of the artists; and indeed, the very conventions themselves served to symbolise and accentuate the real essence or vital differentiation of every object delineated. We felt, too, that besides these recognisable excellences there were others lurking beyond the reach of our perceptions. Certain touches here and there gave vague hints of latent symbols and stimuli which another mental and emotional background, and a fuller or different sensory equipment, might have made of profound and poignant significance to us.