PISSUEE DEPOSITS OF QUARTZ. 753 



closely grouped that the separating walls of silica were (juite tliiii or partly 

 wanting. At times they were quite wide apart or in small groups. These 

 tubes are lined with limonite and sometimes nearly filled with it. At times 

 a separate cylinder or open tube of limonite is found free in tlie cavities. 

 The limonite can not be wholly removed from the cavity, but impregnates 

 the silica for a small but definite distance in from the surface of the cavity. 



The most striking circumstance is that the silica in one portion of the 

 mass grades with imperceptible boundary into a mass of distinctlj' banded, 

 siliceous, dove-colored linaestoiie, or ankerite, as it oxidizes into a porous 

 ocher. 



It seems tolerably clear that the general explanation of this must be 

 that a mass of rootlets penetrating a fissure of the trap became coated with 

 limonite and that then a deposit of silica, at first impregnating the limonite 

 there, went' on to fill the whole fissure, while in part of the latter a mixture 

 of calcite and silica completed the work. Where the delicate cjdinders and 

 tubes of limonite rest free in the cavity we may suppose that limonite was 

 deposited 'v\'ithin the bark of the rootlet, replacing or surrounding the 

 shrunken pith. Indeed, a portion of this bark remained in the tube at one 

 jilace and was in part removed and burned. 



MON XXIX i8 



