Diagnosis of herbicide injury can be complex because of the large number and interaction of factors leading to herbicide injury. Computer-based expert systems have great potential to assist users, particularly nonexperts, in accurate diagnosis of herbicide injury. Rule-based and case-based reasoning are the most widely used forms of expert systems, and each system has strengths and limitations. Approaches that integrate rule-based and case-based reasoning may augment the positive aspects of the two reasoning methods and simultaneously minimize their negative aspects. The Herbicide Injury Diagnostic Expert System (HIDES) integrates rule-based and case-based reasoning and uses field-specific information, injury symptoms, herbicide use history, and herbicide information to diagnose crop injury from herbicides. The HIDES program uses a set of rules to identify suspect herbicide(s) that is the candidate for causing the observed injury and possible sources of the suspect herbicide(s). Case-based reasoning is used to propose a probable cause of injury by making an analogy to previously solved cases. A four-step procedure is followed when using HIDES: information collection, suspect herbicide identification, suspect herbicide source determination, injury reason suggestion, and knowledge accumulation.

A survey was conducted in 2000 across 38 counties in Mississippi on 192 randomly selected soybean fields to assess the most common occurring weeds. Statewide, prickly sida, which was present in 40% of the fields sampled, was the most common. Pitted and entireleaf morningglory were present in 34 and 29% of the soybean fields, respectively. Broadleaf signalgrass and barnyardgrass were the most common annual grasses, and yellow nutsedge was the most common sedge observed. Trumpetcreeper and redvine were the most common perennial vines. In the Mississippi Delta region of Mississippi, prickly sida was present in 45% of the fields sampled. The trend of occurrence of other species in the Delta mirrored statewide results. In eastern Mississippi, prickly sida and broadleaf signalgrass were found in 43% of soybean fields. Sicklepod, common cocklebur, and balloonvine were more prevalent in eastern Mississippi, when compared with the Mississippi Delta. Since 1982, there has been a sevenfold decline in the occurrence of common cocklebur and a fourfold decline in the occurrence of johnsongrass in Mississippi soybean. Also, the occurrences of redroot pigweed, common ragweed, and fall panicum have declined. Conversely, the occurrences of yellow nutsedge and broadleaf signalgrass have increased. The occurrences of barnyardgrass, prickly sida, redvine and trumpetcreeper have been relatively static over the past two decades.