Summary Reasons for performing study: There is much evidence to suggest that group III Clostridium botulinum (types C and D) are involved in the aetiology of equine grass sickness (EGS). Antibodies have been detected previously in the blood and high levels associated with resistance to disease. Specific mucosal antibodies in the gastrointestinal (GI) tract are likely to be important in protection, and this study was performed to ascertain if such antibodies could be detected and if their levels were related to disease state. Objectives: To develop a method for quantifying IgA antibodies to C. botulinum types C and D in the GI tract of horses and to relate antibody levels to disease status. Methods: Samples of tissue (n = 25: 6 duodenum, 7 jejunum and 12 ileum) were taken from acute grass sickness (AGS) cases and from control horses (n = 12; 4 samples from each site) at post mortem. They were extracted with the detergent saponin in the presence of protease inhibitors and assayed for total IgA, for specific IgA against botulinum neurotoxins types C and D (BoNT/C or BoNT/D), and against surface antigens of a BoNT/C negative strain of C. botulinum type C (SA) and of Clostridium tetani (TetSA), as a control. Specific IgA was expressed as percentage total IgA. Results: Compared to controls, significantly higher levels of specific IgA against BoNT/C were detected in the jejunum (P = 0.04) and ileum (P = 0.02) of AGS cases. Similarly, higher specific levels against BoNT/D were demonstrated in duodenum (P = 0.01) and jejunum (P = 0.02). Significantly higher levels of IgA against SA were demonstrated only in duodenal samples (P = 0.01). Conclusions: Levels of IgA antibody to BoNTs in control horses were at near undetectable levels, suggesting no recent exposure to toxins. In AGS cases, significantly higher levels of specific IgA were detected predominantly in jejunum and ileum. Potential relevance: If specific IgA is protective then any successful vaccine for EGS should induce a mucosal response.