Human gnathostomiasis caused by Gnathostoma spinigerum has been reported to be a prevalent nematode infection in many regions of Thailand. The diagnosis of this parasitic infection is only persumptive, based on clinical features, for instances, intermittent migratory swelling, itching,pain and history of consuming half-cooked meat by individuals in endemic areas. Attempts have been made to diagnose this infection by conventional laboratory methos, such as, complete blood count which showed a high percentage of eosinophil in the circulation. A number of immunodiagnostic methods for gnathostomiasis have been developed to confirm the presumptive clinical diagnosis. However, the development of specific and sensitive immunodiagnostic method is essential to obitin a reliable diagnosis. In this study, the various third-stage larval G. spinigerum (L(,3)G) antigen preparations including somatic extract, excretory-secretory product (ES) and surface estract were characto-rized, employing a number of physicochemical and immunological methods in order to define specific component (s) with potential for immunodiagnosis of gnathostomiasis. An additional study was also undertaken to detect these parasitic antigens in cerebrospinal fluid (CSF) of patients with central nervous system (CNS) involvements in order to monitor the active infection. Detection of specific anti-bodies in both serum and CSF of these cerebral gnathostomiasis patients had been undertaken in this study with a main objectiver of looking for the posibility of using CSF as a specimen for immuno-diagnosis. The SDS-PAGE pattern of somatic antigen was highly complex, consisting of proteins and glycoproteing with a molecular weight ranging from more than 116 to 13 KD. On the other hand, the ES antigen and surface extract consisted of components with more narrow molecular weight range of 98 to 12 KD and 70 to 16 KD respectively. The predominant somatic counterpart with a molecular weight of 38 KD was the only major glycoprotein detected in the somatic extract as demonstrated by concanavalin-A. On the contrary, a majority of the ES antigen, particularly those with molecular weight range of 55 to 40 KD, were glycoproteins. However, both somatic and ES antigens showed strong reactions with sera from infected humans, mice and rabbits immunized with various L(3)G antigen preparations as demon-strated by immunoblotting technique. Specificity of these two antigens were analyzed with angiostrongyliasis sernm obtained from patients suspected of having been infected with Angiostrongylus cantonensis, a common nematode found within CNS. The immunoblot pattern showed that the low molecular weight components of the somatic antigen (26, 21 and 19 KD) reacted specifically with gnatho-stomiasis sera. On the other hand, those with high molecular weighe components (more the 38 KD) reacted strongly with the angiostrongyliasis serum. On the contrary, the ES antigen failed to react with angiostrongyliasis serum. In this study, the sensitive and specific biothin - strepta - vidin ELISA (B-SA ELISA) was also undertaken to be used for antigen detectionin CSF of this group of patients. Although this B-SA ELISA could detect the presence of antigen to be level of 2 ng protein/ml, only one of the twently-eight patients showed a positive antigen in his CSF specimen. It should be noted that no antibody could detect in his CSF. On the other hand, the other CSF specimens had high antibody levels in their CSF, therefore , any ES antigen may be in a form of immune complexes. In fact, one out of these CSF specimens, immune complexes were detected by complement consumption test. An alternative approach with regarding to analyze both serum and CSF specimens of these patients for the presence of anti-body reactive with the somatic antigen was also undertaken. Furthermore, by comparing the specific antibodies obtained in both serum and CSF of the individual patients and analyzed in conjunction with other immunological parameters, for examples, albumin ratio (serum albumin/CSF albumin). IgG-albumin index (CSF IgG/serum IgG ratio) / (CSF albumin/serum albumin ratio) and specific antibody activity (specific IgG titer/total IgG) all pointed to the fact that in addition to serum antibody which may be present in various degrees in CSF, local antibody production does occur within the CNS in two cases out of thirty-two patients. An additional data clearly demonstrated that a large majority of the patients with CNS involvements gave high serum antibodies and the specific antibody could be readily detected in the CSF specimens. From the overall study, it can conclude that the detection of antigen in CSF of cerebral gnathostomiasis patients is not suitabel for immunodiagnosis. An alternative approach for detection of specific antibody is probably more effective, especially when the more refined ES component is available. The latter failed to react with angiostrongyliasis serum in immunoblotting analysis and therefore using the ES component as a diagnostic antigen would provide one with a reliable diagnosis for differentiation of gnathostomiasis and angiostrongyliasis. Either serum or CSF specimen can be used for antibody detection in cerebral gnathostomiasis patients.