Frosting of quartz grains is thought to be due in minor degree to mechanical action (1) by wind and (2) by water, but mainly to chemical action (3) by corrosive solutions and (4) by alternate solution and deposition of matter, especially in desert areas. Defrosting under water is likewise a physico-chemical process involving incipient crystal growth or solution. These physico-chemical actions, in contrast with mechanical attack, affect the whole surface of the grain equally, even in deep indentures. The typical circular markings of 3-10 μ diameter described by Cailleux as percussion cones are shown to be diffraction images of the light source. They can be observed also on the etched surface (e.g., by hydrofluoric acid) of crushed quartz, while electron micrographs of frosted grains fail to show any circular pattern of a few microns in diameter. Because of the diffraction effect, the resolution of the microscope is insufficient to distinguish chemical and mechanical action readily. True percussion cones occur on a minor proportion of well-rounded large quartz grains. Roughly speaking, the diameter is one-fifteenth or less of the long axis of the grain. On pebbles and cobbles this ratio is one-tenth. Experiments on wind abrasion and defrosting confirm these conclusions. On the one hand, mechanical aeolian abrasion is limited to corners and edges. It results in a coarser finish than natural chemical frost. On the other hand experimental fluviatile transport does not defrost but tends to frost polished grains on their protruding parts. Dry rolling in a bottle produces an ideal polish on protruding parts. The action of dew is invoked as the cause of desert frost by alternate solution and deposition. The chemical frost (noted by Sorby, Tricart, and others) is evidently a result of solution, possibly combined with deposition. The mechanical frost of aeolian action is somewhat suppressed in the desert by the healing over by dew action. Mechanical frost, both aqueous and aeolian, and chemical frost can be removed under water by solution or deposition according to the nature of the water involved. Experimental defrosting of sandpapered alum crystals in a saturated solution strongly supports the latter deduction. Experimental solution can lead to a polished finish and to rounding. The important solvent action of sea water will be treated in a later paper. A distinction is emphasized between reaction solution, in which case regrowth by evaporation is excluded, and non-reaction solution, in which case regrowth is possible. It would appear that the latter process produces smoothly rounded, polished surfaces, whereas the former tends to etch and cause frosted finish. In a short note (Kuenen and Perdok, 1961) the main points of the present paper were already recorded.