While wandering over the Lower Grindelwald Glacier in 1938 I happened to find a peculiar formation of ice crystal which, I believe, has never been described. I wrote a short account of these crystals in Die Alpen.Footnote * I found them again several times afterwards so that it is not a question of a single occurrence.

In glacier tongues we often find crevasses or holes filled with water which does not freeze (see Fig. 4, p. 177). On the occasion in question I threw a stone into one of these holes and ice crystals of most unusual shapes and great beauty floated to the surface. I repeated this twenty or thirty times, and each time new crystals were dislodged from below. Naturally there was an appreciable interval before they reached the surface, as the holes are often very deep. The crystals were all flat-shaped and had the form of greatly magnified snowflakes (see Fig. 4). One surface was always quite flat; the other appeared to be carved or crenulated. Occasionally they had a kind of twinned structure in which two of the crenulated surfaces were in contact with the two flat surfaces pointing outwards.

Photographs by A. Srhiseider

Fig. 4. Submerged Ice Crystals. Top left a small water-filled hole about 1 m. across and several metres deep

In 1938 and 1942 I found these crystals without exception in very deep water-filled holes and came to the conclusion that depth played a part in their development, but my most recent investigations (1947) have made me alter this view. I also believed that the crystal masses grew in a horizontal direction. I was strengthened in this on learning that crystals growing in partly frozen water tanks developed in this manner.

My investigations were made easier in 1947 by the discovery of a hole which contained no water but showed remnants of these crystals. They were found at much smaller depths than before, namely 1–3 m., and I was therefore able for the first time to inspect them in their place of growth. In this case they grew nearly vertically downwards on projections in the narrow crevasse. I could not find any crystals growing horizontally.

In another crevasse, this time filled with water and inclined at an angle of 60°, I found the same conditions as in the dry hole, except that the crystals were much more abundant and more fully developed. They were growing downwards and were to be found at depths of only 1 or 2 m. These crystals were always better formed than in the empty crevasses, in which they showed clear signs of melting. In addition to nearly structureless plates, I found many beautiful crenulated designs. The former were irregular and appeared to have developed at random, but the majorit always seemed to branch at the customary angle of 60°, as is usual with most ice structures. The smallest was about the size of a child’s hand, but I also found some as long as 30 to 40 cm.

At this stage it is far from clear whether we have here crystals which have really grown beneath the surface of the water, or whether they originated as crystals of the crevasse hoar type which had sublimed in empty crevasses and had subsequently become “drowned” by infiltrating melt water. In about a hundred water-filled crevasses which I have explored I have only found one or two cases of crystals which appeared to be of the latter type.

In many cases it is not easy to remove the crystals since their downward direction of growth renders it easy for stones to miss them. It is also not easy to dislodge them with long rods, as they usually lie so deep as to be quite out of sight, even in the clearest water; casual prodding seldom finds them.

It is only possible at this stage to guess at the origin of these beautiful designs, but I hope in the near future to make further investigations which may throw more light on the problem.