NOTES ON FISHING METHODS OF THE SOUTH SEAS. 813 
A mile or so from the shore the beach has the appearance of being 
perfectly smooth, but on nearer approach a fringing reef covered with 
old coral comes into view. Near a narrow cut on the west end of the 
beach is a long wall of upheaved beach rock, and across the mouth of 
the cut are scattering bowlders; immediately on the inside of the cut 
the bottom is quite smooth. We subsequently learned this to be one 
of the native fishing-grounds. The beach above the reef on either 
side of the cut is composed of very white sand, and makes off with 
a gentle slope; the reef also makes off quite level. 
South and east of Mat Island is a chain of islets, with submerged 
reefs between. The islets vary in size and shape, each one encircled by 
a sandy beach. <A few of the reefs, or broken parts of the atoll rim 
lying between the islets, are out of water; at high tide and when the 
sea is running high they are awash. Further to the south the reefs 
are covered with small bowlders and patches of live coral. Two islets 
inside the lagoon afford the only places in this vicinity where collect- 
ing with a seine might be carried on with success. 








UE Smee 
CATH RE 
IT? 
RTT 
NDIA LALA 

Fish-traps, Likieb Atoll. 
There are no seining-beaches near the village, the entire shore being 
covered with fine coral. Ata point some little distance to the north- 
ward is a sandy beach where collecting may be done at high water. 
Just below low-water mark a profusion of coral heads appear. Our 
time being limited, we did no shore collecting. 
In the evening the usual number of hand lines were put over the 
ship’s side, and a crab net and fish-traps were set, but without results. 
A few gar-fish, shrimp, and a number of small fry, which had the 
appearance of young sardines, were attracted by the surface light and 
taken in dip nets. Some 20 feet below the light were some large fish, 
but they would approach no nearer. 
The fish-traps are constructed on the order of afyke net and similar 
to the salmon stream traps used by the Indians of southeast Alaska. 
They are cylinder-shaped, 8 to 12 feet long, and 10 inches in diameter, 
and are much less complicated than those seen by us elsewhere, con- 
sisting merely of a number of wooden hoops set 12 to 14 inches apart, 
having withes and bamboo strips seized on the outside; spaces between 
the strips, 2 inches. Some of the traps have one end larger than the 
other. The lower or bottom end is covered with a piece of webbing, 
