The objective of this work was to develop electrochemical sensors for the rapid screening of residues of the anabolically active androgens Boldenone, Methylboldenone, Stanozolol, Testosterone, Methyltestosterone and 19-Nortestosterone; used by some producers to boost animal growth during the production cycle. Such use of hormonal substances for growth promotion is prohibited by EU Directive 88/146/EEC, as it is considered to pose a health risk for consumers. The immunosensors are designed to meet an MRPL (Minimum Required Performance Level) of 1 ppb for detection of hormone residues in bovine urine. Any neat urine matrix effects affecting signal to noise ratios of the system are reduced with the 20-fold dilution of each sample with buffer. Detailed metabolism profiles have been obtained through treatment of animals with each of the six androgens and collection of urine at regular time points. Background testosterone levels are measured in five different animal types: cows, heifers, steers, bulls and veal calves. It is necessary to establish a threshold value for each animal when developing the prototype instrument. Results are very promising; limits of detection (LOD) for all androgens are below 0.1ppb. A multi-analyte sensor has also been developed using three working electrodes on a single strip to detect the six analytes of interest. In this project a small pen sized device (identical in size to the glucose diabetes pen), with a insertable strip containing 5 screen printed electrodes, 3 for detection, one reference, one counter, has been developed for detection of 6 illegal androgens:testosterone, methyl testosterone, 19-nortestosterone, boldenone, methyl boldenone and stanozolol.Following reduction of signal for naturally occuring androgens in 5 types of cows the device will register 3 sigma levels of comounds present in the urine. We can conclude that our sensors provide a rapid, simple, inexpensive, highly sensitive and specific assay for determining androgenic concentrations in bovine urine. Problems developed only in the software: methyl testosterone and methyl boldenone were detected in channel 1, unexpected originally but justifiably because of cross reactivity with the antibody used. Most samples show �+� which actually should be negative in working electrode 2 by UCC, UCD and UL; all samples show �-� in working electrodes 3 by UCC, UCD and UL. This ia a problem in the algorithm and critical level values that were programmed in the instrument software. But these problems can be easily corrected when a further prototype(final instrument) is designed and sold using new software. We have successfully done this manually with a correction program, showing the first prototype could detect all 6 androgens at MPLs as required in the contract. Both polyclonal and monoclonal antibodies were developed at CSIC and TUM and were successfully used in the project. In addition other molecular recognition agents were developed at UCC (aptamers),at CSIC and CBC(MIPs and synthetic receptors)..Some of these are available for commercialisation. RESULTS The multi- analyte sensors set-up in the prototype instrument were: Working electrode 1 � detecting testosterone, boldenone and 19-nortestosterone Working electrode 2 � detecting methylboldenone and methyltestosterone Working electrode 3 - detecting stanozolol Conclusion To validate the prototype, 9 known samples (6 positive and 3 negative) and 100 unknown (blind) samples from UCD were analysed by UCC, UCD and UL.Electrode 1 seems to work satisfactorily except for some cross reactivity with methyltestosterone and methyl bold none. This turned out as a bonus since the prototype should detect all six androgens in one of the 3 channels. Most samples show �+� which actually should be negative in working electrode 2 by UCC, UCD and UL; all samples show �-� in working electrodes 3 by UCC, UCD and UL. This ia a problem in the algorithm and critical level values that were programmed in the instrument software, which was discussed in the previous part. But these problems can be easily corrected when a further prototype (final instrument) is designed and sold using new software. We have successfully done this manually with a correction program, showing the first prototype could detect all 6 androgens at MPLs as required in the contract. The prototype instrument developed by Senslab appears to function very well in all its hardware aspects. The software needs serious modification and when corrected (not a difficult matter as we have shown)should perform well. Unfortunately the company, Senslab, has undergone restructuring and sale to another German company, so a redoing of software and development of the final prototype by them is not feasible. Hopefully the TIP will entice others to take over this instrument redesign and sales to the EU.