I. Alpine Research on Glacier Ice
A concluding summary of some of the earlier workFootnote † of the Jungfraujoch Research Party of 1937–38, led by G. Seligman and with M. F. Perutz as crystallographer, has recently become available (SeligmanReference Seligman 16 ). Other inferences regarding the mechanism of glacier movement had been published previously (Seligman, p. 307, 312Reference Seligman 14 ).
As far as metamorphic petrology is concerned, Seligman’s main conclusions of 1941Reference Seligman 14 (first item below) and of 1949Reference Seligman 16 (other items: Seligman, p. 262–65Reference Seligman 16 ), based on the study of the more superficial layers of the Aletsch Glacier and on observations made on other alpine glaciers, are as follows:
-
Movement in glacier ice involves (a) slip of layers of ice one upon another (laminar motion), and (b) plastic deformation (gliding) in individual crystals, accompanied by crystal growth. In the firn or nèvè region, laminar motion and plastic deformation are absent; movement simply involves relative slip or rotation of crystal units or clusters.
-
Ice crystals near the surface of an actixe glacier are smallest on the lines of fastest flow,Footnote * that is to say normally in the centre of the ice stream; towards the margins they increase gradually in size.
-
The steeper the glacier, the smaller the crystals.
-
Crystal size increases from bergschrund to snout.
-
The longer the glacier, the larger are the crystals at its end.
-
Crystal size is, to a greater or lesser extent, dependent on time.
-
Glacier movement may cause crystal growth by the operation of local shearing stresses and by local pressure variations.
-
Crystals in dead ice grow to Iarger sizes than in moving ice; hence movement is not essential to crystal growth.
The following exception to generalization (4) is of importance. In the deeper layers of glacier ice (investigated only in a tunnel in the snout of the Upper Grindelwald Glacier) Seligman found assemblages of abnormally small crystals. After discussing various possibilities, he has suggested tentatively that they may be due to recrystallization under abnormally great stresses caused by an ice fall not far “up-stream” (Seligman 1949, p. 264Reference Seligman 16 ; 1950, p. 379–80Reference Seligman 17 ). In this suggestion he was influenced by Deeley and Fletcher who, some fifty years ago, attributed to shearing (and in part to fracture) the formation of many small elongated ice crystals observed in tunnels in the Upper Grindelwald and other glaciers (Deeley and Fletcher, p. 155–57Reference Deeley and Fletcher. 5 ). Seligman has also many times found very small ice crystals in shear-planes in glacier ice (Seligman, p. 264Reference Seligman 16 ).
II. Petrological Significance of the Results of Research on the Crystal Structure of Ice
Seligman (p. 254Reference Seligman 16 ) quotes J. D. Bernal as saying “A glacier may be considered as a model… with a fairly rapid rate of transformation, of … sedimentary rock undergoing dynamic metamorphism. The relation of crystallisation to thrust and fault planes, and the size of the crystals. may throw light on the crystallisation that occurs in these rocks.” This idea, which has also occurred to others (e.g. Niggli, p. ix, xReference Niggli 9 ), appears to be particularly apposite when one compares the effect of shearing stress on ice and on quartz, for they are, respectively, the only mineral of glaciers and the most important constituent mineral of the great majority of dynamically metamorphosed sediments; and both are hexagonal. Petrologists will be grateful to Mr. Seligman for envisaging the study of glacier-ice crystals under a polarizing microscope, and to Dr. Perutz for carrying out this work, with its difficult technique (Perutz and Seligman, p. 340−41Reference Perutz and Seligman 11 Seligman, p. 295, 298, Fig. 2 and Plate 19Reference Seligman 14 ; Bader, p. 49Reference Bader 3 ).
1. Ice and Quartz Mosaics in Regions of Minimum and Maximum Shearing Stress
Within recent years the writer has been studying Scottish psammitic granulites (metamorphosed sandstones) of the Moine Series, which H. H. ReadReference Read 13 had found to exhibit a “normal” quartz-crystallization in parts of east-central Sutherland, and to become progressively “abnormal” westwards, as they approach the outcrop of the Moine thrust-plane, along which the Moines have been thrust westwards over other rocks for at least ten miles. The “abnormality” was thus produced by conditions of marked shearing stress in the vicinity of Scotland’s major thrust-plane, which has a gentle inclination eastwards.
In “normal” Moines the quartz is tessellate, that is to say it formsFootnote * a mosaic of more or Iess equidimensional crystals with relatively smooth mutual junctions. In “abnormal” Moines the quartz appears as a highly sutured, interlocking, inequigranular mosaic of crystals that show pronounced optical anomalies due to strain.Footnote † The average grain-size of the inequigranular sutured and strain-shadowed mosaic gets smaller on close approach to the Moine thrust; that is to say, grain-size decreases with increasing shearing stresses.
To the writer there seem to be striking resemblances between (a) these two contrasted styles of quartz-crystallization known to be due to differences in shearing stress (Figs. 1 and 3, p. 567), and (b) two contrasted styles of ice-crystallization illustrated in glacier publications of Seligman and Perutz, and of others, viz. the tessellate (equidimensional and unsutured) mosaics of firn and of the upper layers of ice in the névé region; and the sutured inequigranular ice crystals of the main body of the glacier (Figs. 2 and 4, p. 567; see also Ahlmann and Droessler, Fig. 2, p. 270Reference Ahlmann, Droessler and . 1 ).Footnote ‡
Fig. 1 Tessellate quartz-mosaic of psammitic granulite, east of Beinnnan Losgann, Ardnamurchan. Mosaic largely quartz, with sporadic alkali feldspar and mica. Nicols crossed. × 32. Compare Fig. 2
Fig. 2 Tessellate mosaic of new ice (6 years old) at 23 metres, in a crevasse in the Mönchfirn, near the head of the Aletsch Glacier. Dark blebs with high relief represent enclosed air spaces. Nicols crossed. × 4.1. Compare Fig. 1
Fig. 3 Inequigranular, sutured and strain-shadowed quartz-mosaic of psammitic granulite, east of Mallaigmore, North Morar. Mosaic largely quartz, with subordinate alkali feldspar, mica and epidote. Nicols crossed. × 32. Compare Fig. 4
Fig. 4 Outlines of insquigranular sutured ice crystals in tunnel in snout of Upper Grindelwald Glacier. Reduced (scale on photograph). Compare Fig. 3
Tessellate ice characterizes a region at the head of a glacier where shearing stress is at a minimum, and crystallization goes on under conditions of compression due to the weight of overlying fim and snow (Perutz and Seligman, p. 335, 340 and Plate 17Reference Perutz and Seligman 11 ; Seligman, Plates 24, 25Reference Seligman 14 ).
Highly inequigranular sutured ice crystals characterize the ice of the glacier tongue (the main body of the glacier); and abnormally small sutured crystals have been observed along shear-planes in glaciers, and in the Grindelwald ice tunnel situated not far downstream from an ice fall. Seligman has consequently made the tentative suggestion that abnormally small ice crystals (which his illustrations show to be sutured) are formed as the result of the operation of abnormally great shearing stresses.
Published statements by Seligman and Perutz regarding the genetic significance of the abnormally small ice crystals are at first sight incompatible. To clarify Perutz’s difference of approach to the problem it is necessary to summarize, in the first instance, his explanation of the general increase in size of ice crystals, near the surface of an active glacier, from bergschrund to snout (Perutz, p. 133–34Reference Perutz 10 ). He suggests as a working hypothesis that deformation and growth in glacier ice are interconnected in the following manner: “Crystals having the right orientation for yielding to stresses by glide along their basal planes would have a higher energy than others which cannot yield; the former would therefore have a tendency to grow at the expense of the Iatter by molecular exchange across the crystal boundaries. In glaciers the change in crystal orientation brought about by intra-crystalline gliding and the alterations in stress to which the ice is subjected in the course of its flow to the valley would continuously give rise to fresh energy-differences between neighbouring crystals, and to consequent crystal growth.” Perutz states also that (a) the ice crystals of the glacier tongue (i.e. of the main body of the glacier) are often of very complex shape and always closely interlocked; (b) it is not certain whether growth follows or accompanies deformation; and (c) it appears certain, as a result of glacier observations, and of laboratory experiments by Tammann and Dreyer, that increase in average crystal size is invariably to be observed after strain has taken place.
This hypothesis, in which the time-element (length of period of crystallization) seems to be important (cf. Seligman, p. 262Reference Seligman 16 ), appears to have no obvious direct applications to the problem of quartz-crystallization in metamorphic rocks. To illustrate this statement, let us suppose that shearing stress (up to a certain limiting value) acting on a psammitic sediment, has resulted in increase of the average size of quartz grains; such an effect would, in general, be impossible to detect because the original size of the quartz grains would be unknown.
Recession of the Glacier du Valtournanche
Upper photograph, August 1942
Lower photograph, August 1949
Photographs by Manfredo Vanm
Seligman (p. 264Reference Seligman 16 ) has however suggested that a similar process, combined with recrystallization due to abnormally great shearing stresses (connected with the presence of an ice fall) may account for abnormally small ice crystals that he found in relatively deep layers of ice in a tunnel in the snout of the Upper Grindelwald Glacier; these crystals are shown by his illustrations to be inequigranular and sutured (see Fig. 4, p. 567; also Seligman 1948, Fig. on p. 485Reference Seligman 15 ; and 1949, Fig. 3, p. 267Reference Seligman 16 ) Seligman’s tentative hypothesis is thus apparently inconsistent with Perutz’s view that increase of average crystal size is the invariable result of strain. The two views may perhaps be reconciled by taking into consideration the increase in crystal size that has been observed by Seligman on the upper surfaces of ice-tables carved out in tunnels in the Upper and Lower Grindelwald Glaciers; Seligman attributes the increase in grain-size to relief of pressure (Seligman, p. 380Reference Seligman 17 ). We have seen that the grain-size of quartz close to the Moine Thrust was reduced as the result of increasing shearing stress. In this case, of course, relief of pressure took place millions of years later, under conditions that made renewed growth of quartz crystals impossible. It would thus appear that Seligman’s ice hypothesis gains support from petrographic evidence.
Seligman thinks that actual rupture of ice crystals is unlikely to accompany his postulated recrystallization; even if he is right, it would seem that an effect similar to that of mechanical granulation accompanied by recrystallization, might well be produced by a combination of unusually pronounced translation-gliding and the development of shear-planes (cf. Turner, p. 252Reference Turner 18 : explanation of mechanical granulation and concomitant recrystallization of quartz).
It seems probable to the writer that, in psammitic granulites, sutured margins of quartz crystals that show marked undulose extinction due to strain may be accounted for by a somewhat similar mechanism. Although there is as yet no certainty regarding the number and orientation of glide-planes in quartz, undulose extinction has been definitely correlated by petrologists with translation-gliding (Turner, p. 227, 255–66Reference Turner 18 ). It seems reasonable to infer that all suturing of strain-shadowed quartz is due to variable localized pressures and shearing stresses connected with internal differential movement along crystal glide-planes; these physical controls, operating in conjunction with a pore-fluid, may be expected to lead to marginal irregular, highly localized and migratory, solution and recrudescence of crystal growth, involving perhaps, in addition, molecular transfer across crystal boundaries (cf. Erdmannsdörffer, p. 283–84Reference Erdinannsdörffer 6 ; Griggs, p.1003, 1009Reference Griggs 7 ; Seligman, p. 263Reference Seligman 16 ).
Such transient recrystallization at constantly changing positions on mutual crystal boundaries is just what one would expect to produce the sutured, irregularly interlocking and inequigranalar quartz-mosaic of psammitic granulites (cf. Turner, p. 252Reference Turner 18 : on mylonites).
The fact that the existence of similar glide-planes in ice and quartz has not been established, does not appear to affect the general analogy; the important point is that in each case translation-gliding has been independently inferred. The writer is not aware whether undulose extinction characterizes the inequigranular sutured ice of the Upper Grindelwald tunnel; but even if it is not developed, this may well be due to the fact that shearing stress produces translation-gliding much more readily in ice than in quartz. Again, a general analogy between the behaviour of ice and quartz under shearing stress would not appear to be vitiated by the fact that the scale of differential flow is often much greater in a glacier than it is in rocks undergoing dynamic metamorphism. (See also addendum, p. 571).
2. Effects of Shearing Stress on the Orientation of the Principal Crystallographic Axes of Crystals of Ice and Quartz
Observations that appear to have petrological significance have been made on the orientation of the principal crystallographic axes of ice crystals in ice subjected to shearing stress. This evidence suggests that extended and more detailed work on ice may go far to settle the present geological controversy regarding the relationship between the direction of tectonic thrust and the orientation of quartz crystals in psammitic granulites (Anderson, p. 122Reference Anderson 2 ; Phillips, p. 286Reference Phillips 12 ). In this connection (Perutz and Seligman, p. 350–55Reference Perutz and Seligman 11 Footnote *) it has been found that in glacier tongues there exists a general tendency for ice crystals to be arranged with their principal axes perpendicular to planes of shear. Moreover, a study has been made of the orientation of crystals of natural ice, in relation to direction of movement, in an ice-apron on the north wall of the Sphinx ridge, near the Jungfraujoch (Seligman, Figs. 2 and 7, p. 299, 314Reference Seligman 14 ). Here, in an ice-grotto pillar distorted by creep, most of the principal axes were found to be perpendicular to the plane of shear, and all principal axes were distributed in a plane normal to the plane of shear and normal to the direction of flow. These findings have so far, however, not been reconciled with the results of Bader’s study of crystal orientation in a block of artificial ice after it had been subjected experimentally to shearing. Prior to the experiment, the crystals of ice had a random orientation; after shearing, the grain-size had increased but suturing had not developed. According to Perutz and Seligman,Footnote † who have summarized Bader’s results, most of the principal axes were found, after the experiment, to have taken up positions perpendicular to the plane of shear, and all principal axes were distributed in a plane normal to the plane of shear and parallel to the direction of flow. If we draw an analogy between the behaviour of crystals of ice and quartz, the ice-grotto results appear to support E. M. Anderson’sReference Anderson 2 view that the planes of “quartz girdles” in Scottish psammitic granulites are normal to the direction of movement (and to the lineation). On the other hand, the Bader-Haefeli experiment apparently supports F. C. Phillips’sReference Phillips 12 view that the Scottish “quartz-girdle” planes, although normal to the lineation, are parallel to the direction of movement. A satisfactory explanation of the apparently contradictory results of Perutz and Bader may thus be of fundamental importance to tectonic research and to petrofabric studies.
It may be mentioned in passing that a study of glacier motion has recently been recommended to petrofabric workers as a guide to the definition and interpretation of lineations in rocks (Cloos, p. 22–25Reference Cloos 4 ).
In conclusion the writer would record, with gratitude, his indebtedness to Mr. Seligman for the active interest he has taken in this paper, and for his assistance in illustrating it, and to Sir Edward Bailey and Dr. E. M. Anderson for encouragement and helpful criticism.
In a subsequent communication dated 22 April 1951 Dr. MacGregor writes:
Addendum. Since the above paper went to press, the writer has come across striking confirmation of some of the inferences of Section II.1. Thirty years ago a Japanese Professor, while at Chicago University, (a) twisted cylinders of ice, and (b) bent rectangular ice bars formed of subparallel crystals. The behaviour of the ice bars, and their microscopic appearance before and after bending, were described and illustrated in his paper (Figs. 3, 12, 13 and p. 613–15, 624–31).Footnote * Ice aggregate with the relatively smooth mutual crystal boundaries of the present writer’s “tessellate” type of mosaic was converted, as a result of bending, into recrystallized ice forming a highly sutured and more inequigranular mosaic, just like that of the “abnormal” Moine granulites. Faint sub-parallel lines were developed, representing planes said to be parallel to optic axes. These lines had a different orientation in different ice crystals; they started from the angular points of the zigzag (i.e. sutured) mutual crystal boundaries. Sometimes two sets of these lines, nearly at right angles, were seen in a single crystal unit of the new mosaic. Uniform extinction (between crossed nicols) was generally observed throughout each individual new crystal; but in some crystals portions divided by the straight lines showed slight differences in extinction. From all his experimental results Matsuyama inferred that gliding planes parallel to the base of each crystal are not the controlling factor in the deformation of ice and are probably not even an important factor.
