Observations of ice crystal fragments ejected from a frost surface during a positive growth cycle suggest a multiplication process of glaciations in the atmosphere. Photomicrographic studies reveal that the structures of ice particles are identical to ‘most samples collected in natural convective clouds and the mechanism for their breakage is due to sublimation governed by curvature effect rather than by mechanical fracture.

• A simple mechanism is described which may explain the rapid glaciations which commonly occur in the cold portions of super cooled convective clouds. The process begins with the formation of a few ice particles in the cloud. Since they have no competition for moisture and cloud droplets present, they grow rapidly by crystal growth and the accretion of cloud droplets. As they enlarge through dry growth to form graupel or hail particles, their falling velocity increases so that they encounter warmer super­saturated air. This causes the graupel particles to become coated with a very fragile dendritic frost coating.
• Laboratory studies of this mechanism are very spectacular. A simple demonstration of the basic phenomenon can be observed by placing dry ice in a polyethylene bag in warm air. Within a few seconds a dust-like layer of frost will be seen which rapidly grows until dendritic growths appear. Low power magnification shows a very rapid development of frostiness, accompanied by great mobility of particles. It soon becomes obvious that this growth involves electrification phenomena, the dendritic <TREE> twisting and turning as they respond to the moisture flow, electrical stress and growth environment. At times an entire tree will suddenly shoot off and then disintegrate into myriads of fragments as they literally explode.
• With a less intense temperature gradient, the same effects can be seen but with less rapid changes.
• Figures 1-a, b and c show the moisture and particle relationships observable within a 8-minute period in which there was a temperature difference of 5C degree between simulated graupel particle and its moist environment. 1-a is the initial flow pattern of super-cooled droplets around the particle when it is first placed in a moist environment. 1-b shows the moisture, liquid and ice crystal particle regime which develops in one minute. 1-c depicts some of the ice particles ejected from the frosty environment as the dendritic trees become electrified and are repelled by the particle. It is possible that electrification develops due to the Workman-Reynolds effect.
• Figures 2-a, b and c illustrate the tenuous attachments which bind frost crystals grown from water vapor and accreted frozen cloud droplets. Since these attachments often have diameters of less than 5 microns, they readily evaporate, even when the larger portions are growing, because of the difference in vapor pressuren these slender columnar attachments and the more massive ice structures nearby. In many instances the ice growths forming the dendritic trees appear to be held together by nothing but electrical attraction. Some of the branches of the trees will be seen to rotate with complete freedom. If a charged object is brought nearby, much of the structure responds as in a magnetic field and some of the formations will fly toward the charged object or will be repelled in an opposite direction at high velocity. At times these fragments will be seen to disintegrate into many small particles.
• Careful examination of the ice particles in convective clouds such as produce thunder, in blizzards and in intense lake-effect snow storms, show ice particles very much like the nondescript particles collected from the simulated graupel particle growth regime described here, and will be found illustrated in a number of recent papers published in the scientific literature [1], [2], [3].
• Another simple and spectacular demonstration of this phenomenon can be observed by placing a metal plate, cooled to -10C degree, 1 cm above a water surface and illuminated with daylight or a beam from a flashlight. In less than a minute after placing the cold plate above the water surface, a dense miniature snow storm can be seen in the space between the water and the cold plate.
• Dependent on the temperature profile existing between an icy precipitation particle and its warmer environment, the shape of the frost crystals which form will follow the classical growth pattern outlined by Nakaya [4] or modified by environ­mental gaseous impurities [5] Variations we have observed are shown in Figures 8-a, b, e and d. Notice also in these illustrations, especially in 3-a and 3-d, the very small ice structures which hold the mass together.
• Although Odencrantz has suggested a multiplication process of glaciations which he
• believes is caused by frost growth, the illustrations used by him [6], [7] are very different from the method we are describing here, and we find ourselves in agreement with Smith-Johannsen [8] that the modified crystalline structures he uses to illustrate his papers are due to artifacts produced by the replications processes used.
• The basic experiments illustrated in this paper are not new discoveries. The senior author first encountered the phenomenon in his original studies of frost crystal replication. The concept of frost shedding, causing a sort of chain reaction within a super cooled cloud, was described in his notebook at General Electric on the day he discovered the dry ice effect on July 13, 1946, and a brief reference to this work was published [5] two years later. Electrical aspects of the phenomena were mentioned in several papers [9], [10] and most recently in ICCIN [11].
• The laboratory experiments have been shown to many persons over the years. To our knowledge no one has yet pointed out these interesting relationships.
• It should be emphasized that the phenomena described in this paper are observed only during a strong positive growth cycle and not during an evaporating phase as discussed by Ruskin [12], Cross [13] and others, and is rarely of whisker-like structure but mostly of a nondescript but more < massive > structure than filamentous.
• ACKNOWLEDGEMENTS.
• Although I (Schaefer) have known of this phenomenon for 24 years, described it at a public meeting about 20 years ago, and have referred to it in several papers cited, the present paper resulted directly from an impetus received during the visit last winter to the Island of Hokkaido as part of our Japanese-American Scientific Cooperative Program, NSF Grant GF-280. While stormbound at a ski hostel in the high mountains near Nisekoxx hot Springs, I spent part of the demonstrating to Dr. Choji Magono and two of his associates, the phenomenon described in this paper. Using a chunk of dry ice held in a polyethylene bag, I was able to show him the extremely rapid growth of dendritic frost formations, the high charge which develops on the ice crystals which form, and the manner in which the dendritic trees suddenly leave the cold surface and disintegrate into myriads of tiny ice fragments. Although Dr. Magono is an exceedingly well informed and capable experimentalist, he had never encountered this phenomenon and immediately made plans to make some laboratory studies of the effect which occurs under these conditions.
• This was an unexpected development which flowed from the National Science Foundation grant mentioned, and this support is gratefully acknowledged.
• REFERENCES
• [1] Mossop S.C.--- Bull. Am. Met. Soc., 51. p. 474 (1970).
• [2] Koenig L.R. .J. Atm. Sc.. 25, p. 460 (1968).
• [3] MOSSOP S.C. and ONO A. - J. Atm. Sc. 26, P 130 (1969).
• [4] NAKAYA U. — Snow crystals. Harvard University Press (1954), 504 P.
• [5] SCHAEFER V.J. Chemical Reviews. 44, p. 291 (1949).
• [6] ODENCRANTZ F.K. et a!. ---Sciences, 100, p. 1345 (1968).
• [7] ODENCRANTZ F.K. J. Atm Sc., 27, p. 167 (1970).
• [8] SMITH-JOHANNSEN R.I. ---Science, 163, p. 958 (1969).
• [9] SCHAEFER V.J. — Final Report, Project Cirrus, Contract # DA-36-039-SC-15345. G.E. Res.Lab., Schenectady, N.Y., pp. 52-9 (1953).
• [10] SCHAEFER V.J. — J. Appl. Met., 7, p. 452 (1968).
• [11] SCHAEFER V.J. — Proc. of 7th Int. Conf. on Condensation and Ice Nuclei, Prague. p. 167 (1969)
• [12] Ruskin R.E. — Science, 166, p. 906 (1969).
• [13] Cross J.D. —- Science. 164. p. 174 (1969).

• JOURNAL de RECHERES ATMOSPHERIQUES (1971) "The Production of Ice Crystal Fragments by Sublimation and Electrification" Vol V. No I.-1971 SCHAEFER and CHENG