OF ARTS AND SCIENCES: FEBRUARY 11, 1873. 513 



expansion are in the course of the winter repeated with every cycle of 

 temperature going above and below freezing, until when exposed to 

 the higher sun of spring, or even the direct sun during the winter, 

 especially in the case of a detached block, the surfaces of the honey- 

 comb structure are separated from each other by the solution of the 

 contiguous walls of ice. 



The production of transverse horizontal columnar structure in the 

 ice of the cracks above referred to is a direct result of the freezing of 

 the water in the crack, first at the surface of contact with ice of a tem- 

 perature below the freezing point. 



A like columnar structure I have observed when water in bottles 

 was frozen, as already alluded to, by exposure out of doors on very 

 cold nights. The ice from the surface in contact with the glass is, in 

 such cases, frequently filamentous, as if slender visible tubes converged 

 in a general way toward the vertical axis of the bottle. Strictly speak- 

 ing, the direction of the filaments is uniformly approaching a perpen- 

 dicular to the surface of the bottle, whatever its conformation. 



These slender white threads I have proved to be primarily, not air- 

 tubes, but empty spaces, though when the ends of the tubes are ex- 

 posed by fracture, they, of course, instantly fill with air. 



As a proof that these are empty spaces, I may mention that I ex- 

 posed a quantity of water in a cylindrical cast-iron vessel with rounded 

 bottom of some half-gallon capacity to a freezing temperature in which 

 the whole water became solid. On examining the transparent block, I 

 noticed near the centre of the upper surface what seemed a large air- 

 bubble shut in below a layer of ice about an eighth of an inch in thick- 

 ness at the thinnest part. Around the margin of this air-bubble I 

 raised a little wall of wax, and filled the space within the wall with 

 colored water. With a needle I then drilled through the film of ice to 

 the space below, when, on withdrawing the needle, the fluid passed 

 with a rush to the before, vacant cavity, filling it completely without 

 escape of bubbles. 



This vacuum had been produced by the shrinking of the main body 

 of the ice, due to the cold after the formation of the exterior crust en- 

 casing the largest volume of the ice, the shrinking of the ice taking 

 place while there was still a portion of water not solidified. The space 

 left by the shrinking, filling with the remaining water, left a vacuum. 



Like cavities I have produced in bottles filled with water' and ex- 

 posed to low temperature. The phenomenon is the same as that ob- 



VOL. VIII. 65 



