The aim of the study was to develope and validate a serological test system for extended epidemiological investigations to which extend dogs were exposed to Borrelia burgdorferi. For this purpose, 121 samples of dogs which were suspect of an infection and submitted to the laboratory for serological testing, were investigated in an immunofluorescence test (IFT) and to different enzyme-linked immunosorbent assays (ELISA). Valuation of the ELISA systems was assessed in relation to the IFT. Sensitivity, specificity and predictive values were calculated for negative, positive and also borderline values. In the screening test all samples with a positive titre in IFT were judged positive. Samples negative in IFT showed negative results in the screening in only 86 %. All samples positive in screening or of borderline value in IFT were again tested using a confirmatory ELISA. By this procedure a specificity of 100 % and a sensitivity of 85 % was calculated for samples with positive or negative IFT titres, respectively. Samples with IFT borderline titres (1:64 or 1:128) were judged negative in this ELISA in 80 % and borderline in 16 %. Checking of selected samples in an immunodot test cnfirmed the ELISA results. Sera being negative in ELISA showed also no specific reaction in this test. When considering the possibility of false positive reactions in IFT, rather high percentages of sensitivity and specificity could be found. Because all real positive samples could be detected using the confirmatory ELISA as a single test, there is no further need in epidemiological investigations to use the screening test. For specific problems the Western blot can be used. Thus, for serological testing of Borreliosis a low-budget and safe test system is available, and the investigation of random samples in an extended area can be started to estimate the geographical distribution of Borrelia burgdorferi.