288 



The National Geographic Magazine 



dantly, and the feldspar may be some- 

 what less calcic. The groundmass is 

 lighter colored, filled with minute clear 

 crystals and grains like feldspar, with 

 few that are colored in a glassy base. 



The only specimen of this type in the 

 collection is No. $b, " Carbet Peak, from 

 a ravine on upper slope of plateau," al- 

 though No. 15, from near the barracks, 

 five miles north of Fort de France, 

 should be mentioned here as closely 

 related. 



Dacite {Quartz Andesite). — One of 

 the most interesting specimens is the 

 dacite No. 4, " Pitons, Mount Carbet, 

 material of peak." It is a light gray, 

 conspicuously porphyritic rock. The 

 light-colored phenocrysts are quartz and 

 feldspar, the dark ones, chiefly horn- 

 blende, attaining in some cases a diame- 

 ter of 8 millimeters. One apparently 

 hexagonal scale suggests biotite. The 

 quartz is much fractured and did not 

 appear in the thin section, but the con- 

 choidal fracture and uniaxial positive 

 character of one of the glassy grains 

 leaves no doubt as to its presence and 

 places the rock among the quartz ande- 

 sites. 



The feldspars are more prominent, and 

 the crystals broader proportionally than 

 in thehypersthene andesites. Twinning 

 bands are broader and sections more 

 common in which these appear, and it 

 is possible that some orthoclase may be 

 present with the plagioclase. Augite 

 and black-bordered deep brown horn- 

 blende are among the phenocrysts. The 

 light gray groundmass is filled with 

 small crystals of feldspar and hypers- 

 thene, with some grains of magnetite in 

 a clear glassy base. 



The dacite is much more closely re- 

 lated to the andesites than the dacites 

 associated with similar andesites about 

 Crater Lake, Oregon. 



From the same locality ( ' ' Pitons Car- 

 bet, decomposed old material") conies 

 specimen 20, which is much altered. It 

 is friable, earthy white, spotted reddish 



brown with oxide of iron. Looked at 

 more carefully, grains of quartz occur 

 and connect it with the dacites. 



Under the microscope the feldspars 

 are found to be entirely replaced by a 

 clear isotropic substance, and the fer- 

 romagnesian silicates are represented by 

 oxide of iron. The groundmass has 

 been converted into an aggregate of 

 minute grains of a light-colored mineral 

 like quartz or feldspar, and stained more 

 or less deeply by oxide of iron. Several 

 of the original quartz-phenocrysts occur 

 in the thin section unaltered. 



PRODUCTS OF THE RECENT ERUP- 

 TIONS FROM MONT PELEE 



Hypersthene-andesite Pumice. — We 

 now come to the material which Mr 

 Hill regards as immediately connected 

 with the great eruption of May 8, 1902, 

 in destroying St Pierre. It is the con- 

 solidated molten material of that out- 

 break. Only two pumiceous fragments 

 (Nos. i6« and i6£) were selected from 

 a large number for examination. They 

 were collected "near Riviere Mare." 

 One of them (160) readily sinks in 

 water, but the other, rounded as if 

 water-worn, floats lightly. They differ 

 only in degree of porosity, due to the 

 difference in number and size of the 

 vesicles. 



The pumice is nearly white, sprinkled 

 with small black spots, which under the 

 microscope are found to be crystals and 

 fragments of crystals of hypersthene 

 and grains of magnetite. With them 

 are clear crystals of plagioclase feldspar, 

 probably labradorite or bytownite, and 

 all are included in a very vesicular, 

 dusty-looking, glassy groundmass. The 

 vesicles range in size from less than .01 

 of a millimeter to several millimeters in 

 diameter, and in shape from spherical 

 to linear. They ma3' be best seen with 

 a small lens in the hand specimen, where 

 the fibrous drawn-out character due to 

 the expansion of the gas in the vesicles 



