>. IV RESIDENCE AT SANTAREM 109 
way through the Igarape Acii into the Tapajoz, 
and encumber the port of Santarem. These float- 
ing Grass-islands are a sure indication of the river 
beginning to rise, and they merit a particular de- 
scription here, from being a remarkable and indeed 
unique characteristic of the Amazon and of its 
tributaries with white or turbid water, but not of 
those with blue or black water, nor indeed of any 
other rivers in the world that I have seen or read 
about. The rafts of driftwood on the Orinoco, 
described by Humboldt, and seen there more lately 
by myself, have their counterparts on the Amazon, 
the Mississippi, etc. ; but the Grass-islands of the 
Amazon are totally different things : they are com- 
pact masses of grass, in a growing state, varying 
from 50 yards in diameter to an extent of several 
acres. What kind of grass they consist of, and 
how they came there, I will now try to show. 
Along low shores of the Amazon, especially in 
deep sheltered bays, there is often a broad belt of 
Caapfm (the Tupi name for grass, in general) ; and 
the same feature, more strongly marked, is seen 
in some of the still parana-min's, and in lakes that 
communicate with the river by a short channel. 
This Caapfm consists chiefly of two species, the 
Canna-rana or Bastard-cane (Echinochloae sp.) and 
the Piri-membeca or Brittle-grass {^Paspahtm pyra- 
midale) — amphibious grasses, for whose production 
white water is essential, as is proved by their ab- 
sence from the Tapajoz and Rio Negro throughout 
their entire course, and from the Trombetas above 
the Euro de Sapuqua. Lakes, it is true, have 
mostly clear water in the dry season, but the lakes 
into which white or turbid water enters during 
