To explain the difference and anomaly of these phenomena we must attend to the different character of these rivers. For while the Bénuwé after having once assumed a westerly direction follows it up with but very little deviation, the great western branch describes three quarters of an immense circle, and having but very little fall in the greater part of its extraordinary winding course, the waters which flow towards it from the more distant quarters require a long time to reach its middle course, so that the rain which falls in the course of September and October in the country of the Wangaráwa, or the South-eastern Mandigoes, will certainly continue to swell the river at Timbúktu till the end of November or even December; for that rain falls in those quarters behind the coast of Sierra Leone and Cape Palmas till the end of September, and perhaps even in October, we may conclude with some degree of certainty from the fact that such is the case on the coast[3]; and this is confirmed as regards Kakóndi and Tímbo by Caillié’s observations.[4] In the mountainous southern provinces of Abessinia, too, whose latitude corresponds exactly with that of the regions from whence the feeders of the Niger take their rise, the most constant fall of rain has been observed in September. Now while the whole region between Jenne and Timbúktu is of a very flat and level character, so that the river, running along at a very slow rate and with a very meandering course, not only fills a very broad stream spreading out over the neighbouring low lands, but forms also a great many backwaters and basins or lakes of which the Débo is only one although perhaps the largest in size, the river lower down beyond Bamba, and especially in the district called Tin-sherífen which we shall visit on our return journey, is shut in and reduced to the width of a few hundred yards, so that the water, having expanded over such an immense tract and not exercising therefore the same pressure which such a volume of water would do under other circumstances if it were kept together in a narrower channel, preserves its level or even still increases in extent and depth, while the surplus produced by the fall of rain in the country higher up has already diminished. This is my mode of accounting for a phenomenon which seems to contradict in so great a measure the whole of the phenomena which have come under our observation with regard to the effects of rain and the rising of the rivers north and south of the equator, and imparts to the upper course of the Niger the same character as the Gabún and other rivers of the equatorial line which reach their highest level in the course of February.