846 A JOURNEY IN BRAZIL. 
brought to market at each locality until the spawning season 
is over. Now, as they cannot breed twice within a few weeks, 
it is evident that the shad which make their appearance suc- 
cessively along the Atlantic coast from February to May are 
not the same. It is the spring which migrates northward, 
calling up the shoals of shad from the deeper sea, as it 
touches in succession different points along the shore. 
Such movements, if thus connected with the advancing 
spring along a whole coast, appear to be migrations from 
south to north, when they are, in fact, only the successive 
rising of the same species from deeper to shallower waters 
at the breeding season. In the same way it is probable that 
the inequality in the seasons of rise and fall, between the 
different tributaries of the Amazons and the various parts 
of its own course, may give a sequence to the appearance of 
the fish in certain localities, which seems like migration with- 
out being so, in fact. 
" Keeping in view all the information I could obtain upon 
this subject, I have attempted, wherever it was possible to 
do so, to make collections simultaneously at different points 
of the Amazons : thus, while I was collecting at Villa Bella 
six months ago, some of my assistants were engaged in the 
same way at Santarem, and higher up on the Tapajoz ; while 
I was working at Teffe, parties were busy in the Hyavary, the 
la, and the Hyutahy ; and during my last stay at Manaos, 
parties have been collecting at Cudajas and at Manacaparu, 
and higher up on the Rio Negro, as well as at some lower 
points on the main river. At some of these stations I have 
been able to repeat my investigations at different seasons, 
though the intervals between the earlier and later collections 
made at the same localities have, of course, not been the 
same. Between the first collections made at Teffe and the 
