Abstract The failure of rubber insulation brought about by soil microörganisms can best be detected by a decrease of several decades in the insulation resistance. To study the mechanism of attack and the physical structure that makes rubber insulation vulnerable to such attack, it is necessary to have means of locating and marking the failed spots. Several methods of doing this have been described. The electrodeposition of copper in the micropores created by microbes affords the best insight into the physical structure of the faults. Two distinct modes of attack have been shown. The simpler one, which is characteristic of natural rubber, is a progressive eating away of the bulk insulation. The second, a more obscure type, creates micropores through the wall, but without visible surface erosion. This mode is typical of the synthetics, in which the base hydrocarbon or polymer, if freed of extraneous microbe food, does not support microbe growth. The imperfections consist of streaks of microbe food scattered in a random manner through otherwise inert material. Their distribution is affected by physical processing. These streaks are not hygroscopic and do not lead to failure in prolonged immersion in either hot or cold water. They do, however, tend to lower the d-c breakdown of samples after several days soaking in hot water. Thus with a given susceptible compound it is possible to predict which processing will produce the greatest stability in soil by breakdown tests on samples not exposed to soil. The breakdown tests relate to physical structure. To test the inherent resistance of a material to soil exposure, the simple visual fungus test is very convenient. The validity of this test as a criterion of stability has so far been confirmed by the routine resistance tests after accelerated laboratory soil exposure. It is not safe to predict the stability of a complete compound from the results of visual tests on the base material. The vulcanized insulation must be tested. Insulating compounds shown to be promising by the visual test should be given a thorough soil exposure test. The testing of preventives is a field in itself. It is not simple. Materials which are strongly antiseptic in water may not be so when mixed in a rubber compound. Many of them would be completely unsuitable in rubber insulation. Some preventives are effective in a compound before it is vulcanized but not afterwards. Some are completely inactivated by certain other necessary or desirable ingredients. Others may leach out in time and cease to protect the material. Oxidation inactivates some preventives. A nicely balanced compound is required to produce insulation with sufficient stability in soil, proper physical characteristics, and high electrical quality.