496 



PROCEEDINGS OF THE NATIONAL MUSEUM. 



vol. 59. 



cracks, like those developed in glass, by the same treatment and show 

 no indication of cleavage. 



The crystals may be attached by any part, but from the develop- 

 ment of those studied it appears that a majority of them are attached 

 by the negative end of the d axis, the front of the crystal being com- 

 pletely developed, while the rear is bounded by fracture or interfer- 

 ence surfaces. This frequently requires that the crystal be set up 

 in two positions, top and bottom, on the 2-circle goniometer in order 

 that all the forms may be measured in the Dana position. In some 

 specimens all of the crystals are more or less parallel in arrangement, 

 while in the majority of cases they may be inclined at every angle. 



Twinning could not be distinctly proven to occur. Many crystals 

 which at first seemed to be twinned proved, when studied, to be 

 merely examples of parallel growth. Whitlock reports a large crystal 



Figs. 22-23.-22, Orthographic projection on c (001) of crystal 65 showing peculiar distortion 

 by elongation parallel to (120). 23, orthographic projection on tt (100) of crystal b10 

 showing the rare forms 9 (140), * (546), and \) (811). 



from Westfield as a penetration twin parallel to a (100), having c 

 for the twinning axis. This is the only reported occurrence of 

 twinning in datolite. 



Parallel growth is very common and characteristic, nearly every 

 crystal being composite and made up of two or more individuals 

 in parallel position. Although made up of more than one crystal 

 these are united in a peculiar manner, some planes being common 

 to both individuals, which merge without any trace of even a suture 

 to show where they are joined, although sutures frequently occur. 

 These parallel growths occur in great variety and are sometimes 

 very striking in appearance and resemble twins. Figure 14 is drawn 

 to show a very typical example, and the "step" growth is well shown 



