A glance at Fig. 40 will render all that I have said intelligible. The three circles are first stamped on the mud in the same transverse line; but after they have moved downwards they will be in the same straight line no longer. The central one will be the foremost; while the lateral ones have their forms changed from circles to ovals. In a glacier of the shape of this canal exactly similar effects are produced. Now the shorter axis _m n_ of each oval is a line of squeezing or pressure; the longer axis is a line of strain or tension; and the associated glacier-phenomena are as follows:--Across the line _m n_, or perpendicular to the pressure, we have the _veined structure_ developed, while across the line of tension the glacier usually breaks and forms _marginal crevasses_. Mr. Hopkins has shown that the lines of greatest pressure and of greatest strain are at right angles to each other, and that in valleys of a uniform width they enclose an angle of forty-five degrees with the side of the glacier. To the structure thus formed I have applied the term _marginal structure_. Here, then, we see that there are mechanical agencies at work near the side of such a glacier which are absent from the centre, and we have effects developed--I believe _by the pressure_--in the lateral ice, which are not produced in the central. I have used the term "uniform inclination" in connexion with the marginal structure, and my reason for doing so will now appear. In many glaciers the structure, instead of being confined to the margins, sweeps quite across them. This is the case, for example, on the Glacier du Geant, the structure of which is prolonged into the Mer de Glace. In passing the strait at Trelaporte, however, the curves are squeezed and their apices bruised, so that the structure is thrown into a state of confusion; and thus upon the Mer de Glace we encounter difficulty in tracing it fairly from side to side. Now the key to this transverse structure I believe to be the following: Where the inclination of the glacier suddenly changes from a steep slope to a gentler, as at the bases of the "cascades,"--the ice to a certain depth must be thrown into a state of violent longitudinal compression; and along with this we have the resistance which the gentler slope throws athwart the ice descending from the steep one. At such places a structure is developed transverse to the axis of the glacier, and likewise transverse to the pressure. The quicker flow of the centre causes this structure to bend more and more, and after a time it sweeps in vast curves across the entire glacier. [Sidenote: STRUCTURE OF GRINDELWALD GLACIER.]