DEPARTMENT OF MARINE BIOLOGY. 195 



Due to local sinkings on a grand scale, an island at one time almost as large 

 as Tau was converted into the existing islands Ofu and Olosega. The south- 

 ern third of the original island foundered beneath the waters of the Pacific, 

 the downward movement taking place along a curved fault, the trace of which 

 is convex to the north. In a similar way the northern third foundered, along 

 a fault-line convex to the south. The two cui'ves are practically tangent at 

 the strait between Ofu and Olosega. The sharp spurs to east and west of the 

 strait present ragged cliffs remarkably similar to the Pali, the famous fault- 

 scarp of Oahu, near Honolulu. The writer suspects that a partial foundering 

 may have cooperated in the formation of Pago Pago Bay in Tutuila. , 



The bearing of the season's work on theories regarding magmatic differentia- 

 tion and the origin of coral-reefs will be discussed in the final report. On the 

 present occasion the writer will merely state two major conclusions: Samoa 

 illustrates once again the high probability that some alkaUne trachytes and 

 phonolites are derivatives of common basalt; and detailed study of the Samoan 

 reefs confirms skepticism as to the validity of the Darwin-Dana theory of 

 coral-reefs as a general explanation. 



Determinations of the Carbon Dioxid in Sea-Water at Tortugas, Florida, 



by Roger C. Wells. 



Having previously found it difficult to decide whether the CO2 content of 

 certain samples of Gulf water should be ascribed to its physical or organic 

 history, 1 the writer welcomed the opportunity to make some further de- 

 terminations at the Marine Laboratory at Tortugas in June 1919. The tests 

 were made on water taken directly from the sea at various points about 

 Loggerhead Key, at a depth of about a foot below the surface, and included 

 determinations of the temperature, chloride content, excess base or alkalinity 

 by titration, and the Ph value, as well as the total CO2. The results are 

 shown in table 1. The chloride was titrated with a solution of silver nitrate, 

 using potassium chromate as indicator, and the density of the water as sampled 

 was calculated by the aid of Knudsen's tables. The titrations were checked 

 with gravimetric determinations and found to be correct. The Ph values were 

 estimated colorimetrically by comparison with a set of standard tubes pre- 

 pared for this purpose and very kindly loaned to the writer by Professor J. F. 

 McClendon. 



The excess base was obtained by titrating 100 c.c. portions with 0.02 normal 

 sulphuric acid, using methyl red as indicator and blowing out the carbon 

 dioxid by means of a stream of pure air for 15 to 20 minutes. Although the 

 method gives good results, it is difficult to interpret them, as the alkaH titrated 

 includes a small amount due to substances other than carbonates and bicar- 

 bonates. Further determinations yielding only the normaUty of the carbonates 

 and bicarbonates in the water are desirable, the latest determinations by the 

 writer indicating a value of about 0.00223 for this quantity. 



The total carbon dioxid was determined by adding an excess of hydro- 

 chloric acid to 500 c.c. portions of the water and boihng about 15 minutes, 

 while a current of pure air was passed through the water, then over calcium 

 chloride, and finally through weighed soda-lime tubes. A counterpoise was 

 used in order to minimize errors likety to be caused by the high humidity. 



No definite relations are apparent between the sUght variations in the 

 temperature and excess base, or between the variations of the total CO2 and 



^ New determinations of carbon dioxid in water of the Gulf of Mexico. U. S. Geol. Survey Prof. 

 Paper 120-A, p. 7, 191S. 



