EJIvISHA MITCHELL SCIENTIFIC SOCIETY. 33 



of a cementing* material in a clastic rock; and this is 

 true only when thin sections are viewed in ordinary 

 lig-ht. (PI. V, fig-. 4.) For, so long as numerous oli- 

 ivine remnants remain embedded in the serpentine, they 

 are frequently found to exting-uish together over con- 

 siderable areas, showing that the fragments belong- to 

 the same crystalline individual, representing the larger 

 g-rains of the orig-inal rock. Such fragments may be as 

 widely separated by the alteration product as are the 

 Z ^rtions of entirely distinct individuals. E^ven after all 

 li aces of unaltered olivine have disappeared, the orig- 

 inal granular character, and occasionally the outline 

 of a crystal form, are still shown in some of the ser- 

 pentines by the narrow, reticulating bands (* 'mesh- 

 structure") of parallel-polarizing, fibrous serpentine, 

 representing- the first stage of alteration along the bor- 

 ders and fissures of the olivine grains. 



Plate V. is reproduced from photomicrographs of 

 thin sections of typical dunite. A perfectly fresh spec- 

 imen is shown in figure 1. Figure 2 shows a slight 

 beginning- of serpentinization about the borders of the 

 olivine. Figure 3 is taken from a fresh specimen in 

 which the original g-rains, like those in figure 1, have 

 been broken into great numbers of smaller g-rains, with- 

 out perceptibly disturbing their orientation; this sec- 

 tion is seen between crossed nicols. A more advanced 

 stage of alteration than that of figure 2 is shown in 

 fig-ure 4; the olivine grains, the light portions of the 

 field, are reduced to mere remnants. Figure 5 repre- 

 sents the final result of serpentinization, in which no 

 fragment of unaltered olivine remains. With the ex- 

 ception of spots caused by segregation of iron oxides, 

 the rock appears perfectly homogent'ous in ordinary 

 light. When viewed between crossed nicols, however, 



