As night fell we camped on a deserted island in the middle of the river. We dined in the moonlight sitting on scattered empty turtle shells. How pleasing it was to be safe and together! We imagined how it would be if one man had saved himself alone, wandering these deserted banks, meeting more and more tributaries and unable to swim because of the crocodile and caribe fish. We pictured this sensitive man never knowing what had happened to his companions, more worried about them than himself. If you like surrendering to these sad thoughts it is because escaping from danger makes you feel the need for strong emotions.

Leaving the tableland of Guardia we descended to the Indian village of Santa Cruz. First we reached a steep, extremely slippery slope that the missionaries strangely named Bajada del Purgatorio, or Descent of Purgatory. It consists of eroded slaty sandstone, covered with clay; the slope seems terribly steep. To go down, the mules draw their hind legs to their forelegs, lower their rumps and trust their luck sliding downhill. The rider has nothing to fear as long as he drops the reins and leaves the mule alone. From here to the left we saw the great pyramid of Guàcharo. This calcareous peak looks very picturesque, but we soon lost it to view when we entered the thick jungle known as Montana de Santa María. We spent seven hours crossing it. It is hard to imagine a worse path; a veritable ladder, a kind of gorge where, during the rainy season, torrent water rushes down the rocks step by step. The steps are from 2 to 3 feet high. The hapless animals first have to calculate how to pass their loads between the tree trunks, and then jump from one block to another. Scared of slipping they wait a few moments, as if studying the terrain, and then draw their four legs together like wild goats do. If the mule misses the nearest rock it sinks deep into the soft ochre clay that fills in the gaps between the rocks. When there are no rocks, the rider's feet and the mule's legs are supported by a tangle of enormous tree roots. The creoles have faith in the skill and instinct of their mules and remain in the saddle during the long dangerous descent. We preferred to dismount because we feared fatigue less than they do, and were more prepared to travel slowly as we never stopped collecting plants and examining the rocks.

The Indians of the missions, remote from all civilization, are influenced solely by physical needs, which they satisfy very easily in their favorable climate, and therefore tend to lead dull, monotonous lives, which are reflected in their facial expressions.

A historical narrative covers two quite different aims: whatever happens to the traveler; and the observations he makes during his journey. Unity of composition, which distinguishes good work from bad, can be sought only when the traveler describes what he has seen with his own eyes, and when he has concentrated on the different customs of people, and the great phenomena of nature, rather than on scientific observations. The most accurate picture of customs is one that deals with man's relationships with other men What characterizes savage and civilized life is captured either through the difficulties encountered by a traveler or by the sensations he feels. It is the man himself we wish to see in contact with the objects around him. His narration interests us far more if a local coloring informs the descriptions of the country and its people. This is what excites us in the narrations of the early navigators who were driven more by guts than by scientific curiosity and struggled against the elements as they sought a new world in unknown seas.

On the 15th of July we reached the fundaciòn, or the Villa de Pao, established in 1744 and well situated as a depot between Nueva Barcelona and Angostura. Many geographers have mistaken its position, confusing it with other small towns. Though it was cloudy I was able to determine the latitude and the longitude from the sun. The astronomical fixing of Calabozo and Concepciòn del Pao are very important to the geography of this country where there are so few fixed points in the savannahs. Around about we saw some fruit trees, quite rare in the steppes. We also noticed coconut palms, despite the distance from the sea. I insist on this observation as some have doubted the veracity of travelers describing coconut palms, a coastal plant, in Timbuktu and in the heart of Africa.

Soon after entering the Nueva Barcelona llanos we spent the night in a Frenchman's house. He welcomed us very cordially. He came from Lyon, and had left home when still very young. He seemed quite indifferent to all that was happening across the ocean or, as they scornfully say here, 'del otro lado del charco' ('on the other side of the pond'). He was busy sticking large bits of wood together with a glue called guayca, used by carpenters in Angostura. It is as good as any glue made from animal matter. It is found ready-made between the bark and sap of a creeper of the Combretaceae family (Combretum guayca). It resembles birdlime made from mistletoe berries and the inner bark of the holly. An astonishing amount of this glue pours out from the twining branches of the bejuco de guayca when they are Cut.