Josef Halda has built and/or supervised the building of hundreds of rock gardens around the world.

The rocks we used are slightly metamorphosed weathered pieces of columnar basaltic andesite. Those of you who visited some of the hiking sites in the Western Oregon Cascades during the 1998 NARGS Annual meeting have seen this type of rock in natural outcrop. Long jointed columns are produced by a unique set of thermal and geochemical circumstances. The dominant set of joints is very elongate and a secondary set promotes natural breakage into large prismatic pieces of rock. Loren located the quarry and one of the nicest characteristic of these rocks is that the all of those that Loren obtained are naturally shed pieces. This means that breakage is old and all surfaces have a nice soft alteration to them that causes them to look especially mellow. Sometimes in quarry rock, the excess of freshly broken surfaces can be ugly and jarring.

Josef Halda used this long axis and a unique spacing in his design to create the effect of dipping bedded rock more characteristic of sediments. Instead of placing the rock prisms upright as we typically find them, he laid them on their side. Sometimes this type of formation is seen "sideways" or even end-grain. The famous Giant's Causeway is where you can best see the pattern of polygons that the end-grain vantage point produces. Halda quickly studied the geometry of each rock that he selected not only to create a consistent bedding aspect or expression, but also for stability from a construction perspective.

A good foundation is the first requisite to stability. Ponder some of the basic causes of landslides: a slip surface or plane of weakness, an excess load, or poorly consolidated material --and always the big G acting on any or all of the above. One thing that especially will cause a landslide is too steep a slope for the load of material. Josef Halda chose to begin the garden in a slightly depressed area of the hillside that acts as a natural cradle to hold the rock load. We are lucky foundation-wise in that the hillside has dense mudstone not far below the soil. The soil here helps with stability because it has good cohesion. How we sometimes curse clay in soil, but it can act as a terrific glue that holds rock. This is needed only at the very base of the rocks, so the clay is kept out of the upper portions of the crevices. Halda also used some of the blockier large rocks with some amount of "rock root" protruding into the hillside like the roots of teeth.

He was careful to place the rocks so that they interlocked thus adding to their stability. This interlocking is very important whether you are working with large boulders or sand. Angularity in pieces of any size enhances their ability to lock together. And speaking of angularity...This promotes drainage because the points of contact between pieces are irregular and sharp, especially when envisioned in three dimensions. For this reason, we chose crushed rock and builders or concrete sand. The crushed rock also helps the large rocks to lock together, whereas rounded pieces would slide past each other. Halda used chock-stones here and there to add to the stability, to define the crevices, and to slow down possible water movement in them. The sand is "washed" or "clean" sand so that the air and drainage space is not filled in with finer material.

His basic approach once he had a well-rooted large rock in place was to lay courses of rock with the long axis parallel. To keep this from becoming too monotonous a design, he interspersed some larger blockier pieces among the columns. A playful touch of dissonance was created when he suddenly used a single recessed and extra-flat piece. Not only did this further break what could have become a linear monotony, it also added a sense of depth. This area will be an especially interesting focal point for a plant or two that benefit from extra protection. This could be repeated a time or two, but not overdone. To the geologically-trained eye, he caused basaltic andesite to look like steeply-dipping sedimentary beds. And yet the total effect is remarkably natural-looking and has its own geologic and structural fidelity.