1894. 



PROCEEDINGS OF THE NATIONAL MUSEUM. 81 



reference to Pis. iv and v. The individual blades are rarely more than 

 a few inches in length, six and eight inches being the maximuni of the 

 single curved blades such as are shown in PI. iv. (Cat. No. 68142.) 

 In fig. 2, PI. V, the longer blades are 90 mm., by about 24 mm. breadth 

 and 5 mm. thickness. This is in many respects the most remarkable 

 specimen of its kind I have ever seen. The method of growth of tliese 

 forms is plainly by additions to the bottom, or more properly, to the end 

 attached to the wall. They seem to have grown outward precisely as 

 does the hoar frost in loose soil, where the moisture, rising by capil- 

 larity, freezes as soon as a certain level is reached, so that the older and 

 first formed portions are ever pushed upward so long as the supply 

 below is continued. As in the formation of hoar frost, particles of 

 earth are lifted upon the tops of the ice spicules, so here the growing- 

 gypsum having begun forming in a crevice not infrequently forces off 

 pieces of the limestone of considerable size. In fig. 2, PI. v, the force 

 of the growing crystals has even ruptured the stone in three directions. 

 In fig. 1 of the same plate we have proof of two stages of growth. 

 The last formed crystals having pushed the first formed nearly an inch 

 out of place, the line of separation between old and new being indicated 

 by the smaller size of the later formed spicules. As the crystals form 

 and are pushed outward they are in most cases in a condition of strain, 

 which causes them to curl and twist in a remarkable manner, as shown. 

 The individual blades or spicules are but slightly attached to the walls 

 of the cave, and except under very favorable circumstances it is nearly 

 impossible to remove a rosette in a condition at all satisfactory. 

 Proc. N. M. 94 



