O'JO 



SCIENCE. 



[N. S. Vol. XXI. No. 540. 



ALBATROSS EXPEDITION TO THE EASTERN 

 PACIFIC * 



III. 



We left the Galapagos (Wreck Bay) for 

 Manga Reva on the tenth of January. On 

 the northern part of this line we did but 

 little work beyond sounding, as we were 

 likely to duplicate our former work to the 

 eastward. The fourth day out, in latitude 

 5° south, we began a series of trawl hauls, 

 surface hauls and intermediate towings to 

 300 fathoms. In the northern part of the 

 line to Manga Reva the hauls were remark- 

 ably rich as long as we remained within 

 the influence of the western extension of 

 the Humboldt Current, and as long as 

 there poured from the surface masses of 

 the radiolarians, diatoms and Globigerinas 

 living in the upper waters. Some of the 

 hauls were remarkable for the number of 

 deep-sea holothurians and siliceous sponges. 

 Among the former I may mention a huge 

 Psychropotes, 55 cm. long. 



As we passed south and gradually drew 

 out of the influence of the western current 

 we entered the same barren region we 

 passed through to the eastward when going 

 to and from Easter Island. By the time 

 we reached latitude 15° south the hauls 

 became quite poor, and this barren bottom 

 district extended to within a short distance 

 of Manga Reva ; and corresponding to it 

 we found a most meager pelagic fauna, 

 both at the surface and down to 300 fath- 

 oms—so poor that it could afford but little 

 food to the few species, if any, living on 

 the bottom in that region. 



We arrived at Manga Reva on the 

 twenty-seventh of January and found our 

 collier awaiting our arrival. 



While at anchor in Port Rikitea we ex- 



* Extract from letter Xo. 3, from Alexander 

 Agassiz to Hon. (Jeor^^e M. Bovvers, U. S. Com- 

 missioner of Fisheries, on the cruise of the Fish- 

 eries Steamer Albatross, in the Eastern Pacific, 

 dated Acapulco. Mexico, Fel)ruary 24, 1005. 



amined Manga Reva, the principal island 

 of the Gambler group, from its central 

 ridge on the pass leading from Rikitea to 

 Kirimiro on the west side of Manga Reva, 

 as well as from the pass leading to Taku. 

 On both these passes we obtained excellent 

 views of the barrier reef to the west, north 

 and east of the Gambler Islands, and we 

 could trace in the panorama before us the 

 western reef extending in a northeasterly 

 direction parallel to the general trend of 

 Manga Reva Island for a distance of about 

 5| miles. 



From the northern horn to nearly op- 

 posite Kirimiro Bay the barrier reef has 

 only three small islets. It is narrow, of 

 uniform width, about one third of a mile, 

 plainly defined, submerged in places, and 

 passing north bounds a large northern 

 bight dotted with numerous interior coral 

 patches from a quarter of a mile to a mile 

 in diameter or length, with from 7 to 11 

 fathoms. The southern part of the western 

 barrier lagoon ofl:' IManga Reva is irreg- 

 ularly dotted with many small patches of 

 reef, with an occasional deep hole of from 

 15 to 20 fathoms near Manga Reva Island. 

 From the islet to the west of Kirimiro 

 there are but few coral patches, indicating 

 a reef which dips gradually in a distance 

 of a mile to a deep channel of from 4 to 6, 

 fathoms, which separates the northern and 

 western reef from the great reef flat lying 

 to the southwest of Tara Vai. This flat 

 has a width of nearly 2 miles, is about 4^ 

 miles long, and is marked at its southwest 

 extremity by a series of low islets arranged 

 in a somewhat circular line, formed by 

 three deep bays and spurs from the outer 

 line of islets, as so frequently occurs on 

 wide reef flats in atolls of the Pacific. 



This part of the reef is called Tokorua. 

 It shelves very gradually from 3^ to 4 

 fathoms on the west face to 7, and con 

 neets with the plateau upon which standr^ 

 Tara Vai and Aga-kanitai. From Tokorua 



