Yol. 50.] CERTAIN- NOVACtTUTES AND QXJAKTZITES. 387 



Had the rhombohedral crystals been formed in the manner sug- 

 gested by Mr. Griswold it seems doubtful whether they would have 

 been so perfectly developed. Moreover, if the surrounding material 

 consisted of already-formed quartz-grains, there would not have been 

 the selective arrangement of grains around the rhombohedra which 

 is so well shown, at times, in Ouachita stone. Had the rhombohedra 

 been formed last, they would have pushed the surrounding grains 

 aside, or have enveloped them, and of this there is no clear evidence. 

 To my way of thinking, all points to the solidification of silica around 

 already- formed crystals of dolomite. 



If the views advocated in the present paper be correct, we have 

 here, in these Arkansas novaculites, a case of metasomatosis on a 

 large scale. Mr. Griswold states that the novaculites "occur in mas- 

 sive strata, usually presenting plane surfaces and having only thin 

 layers of shale iuterbedded. Five or six hundred feet is the common 

 thickness of the novaculite formation, which generally includes some 

 flinty shales and soft shales or sandstones. The novaculites proper 

 are the prominent members of the formation, however, and occur in 

 massive beds from a few inches to twelve or fifteen feet in thickness. 

 When thinner than about four inches the beds generally lose their 

 novaculite character and are more like flinty shale." 1 



There seems nothing improbable in beds, of the thickness above 

 cited, being wholly replaced by silica. Prof. Hull states that " in 

 the south-western districts of Tipperary, Limerick, Kerry, and Cork, 

 the principal masses of chert occur at the top of the limestone, im- 

 mediately below the shales of the Yoredale series, and sometimes 

 are so abundant as almost completely to replace the limestone itself. 

 .... At the foot of the ridge west of Carlo w, the limestone is com- 

 pletely replaced by masses of greyish chert in thin layers, and over 

 thirty or forty feet in thickness." 2 



If the beds of novaculite be a siliceous replacement of limestone- 

 beds, ic by no means follows that their thickness indicates the former 

 thickness of the limestones, since the removal of limestone may have 

 been greater by far than the amount which has been replaced by 

 silica. 



In connexion with the rhombohedral cavities, it may be pointed 

 out that the component rhombohedra of a dolomite are not all of 

 the same size, as shown in PI. XIX. fig. 6, although there is a 

 general uniformity in this respect. Some of the rhombohedra are 

 occasionally 15 or 20 times as large as others, and these larger crystals 

 would take longer to dissolve and would probably form a fair pro- 

 portion of the residue which ultimately became imprisoned in the 

 silica. That many of the dolomite-crystals were more or less eroded 

 is evident from the irregular forms of some of the cavities in the 

 Ouachita stone. 



1 'Whetstones and the Novaculites of Arkansas,' p. 94. 



2 ' On the Nature and Origin of the Beds of Chert in the Upper Carboniferous 

 Limestone of Ireland,' Sci. Trans. Roy. Dublin Soc. vol. i. (new series) p. 75. 



