152 REruKT OF COMMISSION EK OF FISH AND FISHERIES. 



tion of the reef-dwelling fishes by means of explosives, but the attempt 

 was attended with but poor success, owing, the natives stated, to the 

 fish having been scared away by pre^^ous operations. Explosives for 

 catching the fishes on the reefs and poisons for taking them in the 

 small tidewater pools, wliere, from their shy and secretive habits, it is 

 difficult to secure them with nets, are perhaps the only feasible means 

 of making extensive ichthyological collections under the conditions 

 prevailing in the South Seas, and the expedition was handicapped 

 by not possessing the means for working along these lines. On the 

 whole, the biological collections on the coral islands were disappoint- 

 ing, and far less than similar effort would have yielded in the waters 

 of the West Indies or on the coast of Japan. 



At the various islands where stops were made a few ethnological 

 specimens, principally fishing and canoe implements and articles of 

 adornment, were gathered, but as a rule the time was too short for col- 

 lecting of any sort. The houses differ somewhat in different islands, 

 but typically consist of rather high cocoanut-thatch roofs supported 

 on blocks of coral rock or posts about 3 feet high. Many of them, 

 but not all, have floors on a level with the eaves, a scuttle or hatch 

 giving access to the compartment above, which is used for sleeping 

 purposes and as a storehouse. On some of the islands where no land- 

 ing was made, e. g., Taput^uea, the corner stones, which are usually 

 about 10 or 12 inches square in cross sections, were seen in places 

 along the beach, sometimes quite in the open, on bare sand flats, the 

 rest of the house having disappeared. 



On all of the islands visited more or less attention is given to the 

 cultivation of a large rank-growing species of taro, which has proba- 

 bly been introduced from some of the volcanic islands of other groups 

 where it is indigenous. The taro patches are artificially constructed 

 trenches dug in the sandy soil and usually for some distance into the 

 underlvina: coral rock and filled with an accumulation of veoretable 

 mold, which lying, as it were, in a more or less impervious basin, is 

 kept constantly moist by the rains. These beds are carefully culti- 

 vated and fertilized by household refuse and other materials, the soil 

 from time to time being loosened up and added to by materials sifted 

 through a sieve of cocoanut fiber. At Apamama a spade made of a 

 pearl shell lashed in a cleft stick is used in agricultural operations. 

 Bread fruit grows sparingly, and in general the fauna is more varied 

 than in the Paumotus. 



The natives are smaller and of slighter build than those of the 

 Ellice Islands, and their color is somewhat darker and the hair gen- 

 erally straighter and coarser. The men wear a pandanus-leaf mat 

 reaching to below the knees, and the women are clothed in skirts of 

 stripped leaves, whicn form a very scant covering. As a rule, they 

 are a wild-eyed people, especially the women, and formerly they were 

 fierce and warlike, completely clothing themselves for battle in armor 

 made of closelv woven cocoanut-fiber sennet. Thev are still under the 



