58 



typical "frost flowers" had formed. These were smaller than 

 those observed in November, measuring only 3 cm. to 7 cm. in 

 diameter. All had formed close to the ground; in most cases 

 beginning about 1 cm. above the soil line and extending about 

 5 cm. up the stem. Many of them had pushed up under or 

 through a mulch of fallen leaves around the base of the stems. 



Crystals were observed again on these plants on Jan. 29, 

 1935. They were small but typical "frost flowers" and were 

 formed beneath the snow! The snow crust was carefully removed 

 from around the plants and a piece of black mulch paper placed 

 behind and beneath the crystals as a contrasting background. 

 The mulch paper was left in place all day and as the sun shone 

 brightly the snow was entirely melted away from the area within 

 15 cm. of the plants. The crystals melted entirely away from 

 the stems, also. During the night of Jan. 30, typical ice masses 

 2.5 cm. and 3.5 cm. in diameter formed around the stems. These 

 ice masses were quite symmetrical, whereas those formed under 

 the snow were deformed by pressure against the leaves and snow 

 above them. The ice masses on the stems surrounded by paper 

 were not appreciably larger than those formed on other stems 

 nearby. These ice masses had disappeared by 4 :00 P.M. Crystals 

 formed again during the night and, when observed about 9 :30 A.M. 

 the following morning, the rosettes were about 1 .5cm , in diameter. 

 Atmospheric temperature at the time was about 4° C. and the 

 ice was beginning to melt. At 11 :00 o'clock P.M. the night be- 

 fore, the atmospheric temperature (measured about 8 feet above 

 the surface of the ground) was 0° C. 



During the interval between my first and second observa- 

 tions Dr. R. M. Harper sent me a reprint — from Torreya31; 17, 

 Aug., 1931 — of his article recording observations of "frost 

 flowers" in Florida. This reprint also contained an article by Mr. 

 H. M. Jennison suggesting an explanation of the mechanics of 

 this phenomenon which he had studied in the vicinity of Knox- 

 ville, Tenn. My own observations agree with those of Mr. Jen- 

 nison so closely that no detailed description of the ice masses 

 need be recorded here. Attempts to cut off stems and photo- 

 graph the under surface of the ice masses were unsuccessful, the 

 crystals shattering off at the edge of the bark. Immediate ex- 

 amination with a hand lens showed radial plates of ice in the 

 stem. These plates were thicker in the middle than at the edges 



