as Coleoptera, and thus render their existence unnecessary. The large proportion of climbing forms of carnivorous beetles is an interesting fact, because it affords another instance of the arboreal character which animal forms tend to assume in equinoctial America, a circumstance which points to the slow adaptation of the Fauna to a forest-clad country throughout an immense lapse of geological time.

The large collections which I made of the animal productions of Pará, especially of insects, enabled me to arrive at some conclusions regarding the relations of the Fauna of the south side of the Amazons Delta to those of neighbouring regions. It is generally allowed that Guiana and Brazil, to the north and south of the Pará district, form two distinct provinces, as regards their animal and vegetable inhabitants. By this it is meant that the two regions have a very large number of forms peculiar to themselves, and which are supposed not to have been derived from other quarters during modern geological times. Each may be considered as a centre of distribution in the latest process of dissemination of species over the surface of tropical America. Pará lies midway between the two centres, each of which has a nucleus of elevated table-land, whilst the intermediate river-valley forms a wide extent of low-lying country. It is, therefore, interesting to ascertain from which the latter received its population, or whether it contains so large a number of endemic species as would warrant the conclusion that it is itself an independent province. To assist in deciding such questions

CHAP. III.

FAUNA OF PARÁ.

109

as these, we must compare closely the species found in the district with those of the other contiguous regions, and endeavour to ascertain whether they are identical, or only slightly modified, or whether they are highly peculiar.

Von Martius, when he visited this part of Brazil forty years ago, coming from the south, was much struck with the dissimilarity of the animal and vegetable productions to those of other parts of Brazil. In fact, the Fauna of Pará, and the lower part of the Amazons, has no close relationship with that of Brazil proper; but it has a very great affinity with that of the coast region of Guiana, from Cayenne to Demerara. If we may judge from the results afforded by the study of certain families of insects, no peculiar Brazilian forms are found in the Pará district; whilst more than one-half the total number are essentially Guiana species, being found nowhere else but in Guiana and Amazonia. Many of them, however, are modified from the Guiana type, and about one-seventh seem to be restricted to Pará. These endemic species are not highly peculiar, and they may be yet found over a great part of Northern Brazil when the country is better explored. They do not warrant us in concluding that the district forms an independent province, although they show that its Fauna is not wholly derivative, and that the land is probably not entirely a new formation. From all these facts, I think we must conclude that the Pará district belongs to the Guiana province, and that, if it is newer land than Guiana, it must have received the great bulk of its animal population from that region. I am informed by Dr. Sclater

that similar results are derivable from the comparison of the birds of these countries.

The interesting problem, how has the Amazons Delta been formed? receives light through this comparison of Faunas. Although the portion of Guiana in question is considerably nearer Pará than are the middle and southern parts of Brazil, yet it is separated from it by two wide expanses of water, which must serve as a barrier to migration in many cases. On the contrary, the land of Brazil proper is quite continuous from Rio Janeiro and Bahia up to Pará; and there are no signs of a barrier ever having existed between these places within recent geological epochs. Some of the species common to Pará and Guiana are not found higher up the river where it is narrower, so they could not have passed round in that direction. The question here arises, has the mouth of the Amazons always existed as a barrier to migration since the present species of the contiguous regions came into existence? It is difficult to decide the question; but the existing evidence goes far to show that it has not. If the mouth of the great river, which, for a long distance, is 170 miles broad, had been originally a wide gulf, and had become gradually filled up by islands formed of sediment brought down by the stream, we should have to decide that an effectual barrier had indeed existed. But the delta of the Amazons is not an alluvial formation like those of the Mississippi and the Nile. The islands in its midst and the margins of both. shores have a foundation of rocks, which lie either bare or very near the surface of the soil. the case towards the sea-coast.

This is especially In ascending the

CHAP. III.

FORMATION OF AMAZONS DELTA.

111

river southward and south-westward, a great extent of country is traversed which seems to have been made up wholly of river deposit, and here the land lies somewhat lower than it does on the sea-coast. The rocky and sandy country of Marajo and other islands of the delta towards the sea, is so similar in its physical configuration to the opposite mainland of Guiana that Von Martius concluded the whole might have been formerly connected, and that the Amazons had forced a way to the Atlantic through what was, perhaps, a close series of islands, or a continuous line of low country.

CHAPTER IV.

THE TOCANTINS AND CAMETÁ.

Preparations for the journey-The bay of Goajará-Grove of fan-leaved palms-The lower Tocantins-Sketch of the river-Vista alegre— Baiao-Rapids-Boat journey to the Guariba falls-Native life on the Tocantins-Second journey to Cametá.

August 26th, 1848.-Mr. Wallace and I started today on the excursion which I have already mentioned as having been planned with Mr. Leavens, up the river Tocantins, whose mouth lies about forty-five miles in a straight line, but eighty miles following the bends of the river channels, to the south-west of Pará. This river, as before stated, has & course of 1600 miles, and stands third in rank amongst the streams which form the Amazons system. The preparations for the journey took a great deal of time and trouble. We had first to hire a proper vessel, a two-masted vigilinga twentyseven feet long, with a flat prow and great breadth of beam and fitted to live in heavy seas; for, although our voyage was only a river trip, there were vast sea-like expanses of water to traverse. It was not decked over, but had two arched awnings formed of strong wickerwork, and thatched with palm leaves. We had then to store it with provisions for three months, the time