(Gulbarga District, Karnataka) reported that one government programme insisted they use earthen bunds, even though soil was very scarce, because they were found to be optimal under research station conditions. A more flexible programme would offer more sensible, cost- effective designs. A second implication concerns places where lack of materials is the major constraint to soil conservation investment. Government programmes might find it cost effective to transport stones from places where they are too abundant to places where they are too scarce, but could be used to construct bunds or waste weirs. In Shirapur (Sholapur District, Maharashtra), for example, some of the land is littered with surplus stones excavated during construction of a canal. Farmers indicate that stone bunds are cheap in Shirapur as a result, and also that some land is uncultivable because of the stones. Transporting these stones to nearby regions may not only promote soil conservation in those places, but also clear land for cultivation along the canal. Quality: Farmers indicate that they invest more in soil conservation on land with higher potential productivity than on land with lower potential productivity, given equal levels of erosion. For example, farmers in Aurepalle say that their black soils receive the most soil conservation investment, followed by less productive shallow red soils. On one pocket of saline soil in Aurepalle, old soil conservation structures are not maintained, and new ones are not built because yields on that land are so meager that the soil is not considered worth conserving. It also might be expected that farmers would be more concerned about losing scarce soil on shallow fields than on good land where soil is deep and abundant. However, farmers are more concerned about erosion on good land than erosion on bad land, mainly because good land generally receives greater applications of fertilizer and farm yard manure. As farmers say that removal of nutrients is often the most serious implication of erosion, top quality land is likely to receive the greatest soil conservation investment. Irrigated land receives the most soil conservation investment. The primary objective of this investment is water management, with soil conservation as a by-product. This further strengthens the notion that soil conservation programmes should look for complementarities between invest- ments with short-term and long-term payoffs. Farmers who own and operate both irrigated and unirrigated land appear to invest little in soil conservation on their unirrigated plots. Their irrigated land provides opportunities for productive investments that cannot be matched by SWC measures on dryland. Caring for unirrigated land becomes a low priority for those farmers. Preliminary evidence suggests that farmers without irrigation take better care of their dryland plots than do farmers who also own irrigated land. This is another disturbing prospect, since irrigation development is an important component of agricultural development strategies, and farmers tend to try to gain access to some irrigated land to protect against weather-related risk. As irrigation spreads, it is likely that dryland plots will be relatively neglected. GATEKEEPER SERIES NO. SA34