Narrative of the surveying voyages of his majesty's ships Adventure and Beagle (vol.3)

Aus;. 1836. pernambuco. 593 '^& at the thought of refusing such a request, or of behaving to a stranger with rudeness. The channel by which we went to, and returned from Ohnda, was bordered on each side by mangroves, which sprang Uke a miniature forest out of the greasy mud-banks. The bright green colour of these bushes always reminded me of the rank grass in a churchyard both are nourished by putrid exhalations; the one speaks of death past, and the other too often of death to come. The most curious thing which I saw in this neighbourhood, was the reef that forms the harbour. It runs for a length of several miles in a perfectly straight line, parallel to, and not far distant from, the shore. It varies in \^idth from thirty to sixty yards ; it is dry at low water, has a level smooth surface, and is composed of obscurely-stratified hard sandstone. Hence, at first sight, it is difficult to credit that it is the work of nature, and not of art. Its utility is great ; close within the inner wall there is a good depth of water, and ships lie moored to old guns, which are fixed in holes on its summit. A lighthouse stands on one extremity, and around it the sea breaks heavily. In enter- ing the harbour, a ship passes within thirty yards of this point, and amidst the foam of the breakers ; close by, on the other hand, are other breakers, which tlius form a narrow gateway. It is ahnost fearful to behold a ship run- ning, as it appears, headlong into such dangers. With respect to the origin of the reef,' I beheve a bar composed of sand and pebbles formerly existed beneath the water (a circumstance no ways improbable) when the low land, on which the town now stands, was occupied by a large bay; and that this bar was first consoHdated, and then elevated. These two processes are of such frequent occur- rence in South America, that there can be no objection to using them in accounting for any remarkable structure in the land. There is another and slightly different explanation which possesses equal probability, namely, that a long spit of sand, like some that now run parallel to parts of the neighbouring coast, had its central part consolidated, and VOL. III. 2 Q