A nationwide electronic system for the identification of all pigs is a means to achieve a tighter control of livestock and meat in the Netherlands. In order to examine the use of electronic identification transponders, two field trails were performed. Transponders supplied by three separate companies were tested on pigs on commercial farms. In phase 1, each device was examined on separate farms and in phase 2, the three devices were tested on each farm. A total of 3,436 and 5,947 transponders from the different suppliers were injected in the base of the ear at weaning in phase 1 and 2 on seven and five farms, respectively. The following aspects were examined: technical labour for injection and reading, readability of the transponders, impact on tissues at the injection site, and retrieval of the transponder after slaughter. After instruction the farmer was well able to inject a transponder in a restrained piglet. The results show that in phases 1 and 2 1.6% to 7.3% of the transponders were unreadable at retrieval in the slaughter line, which is significantly (p < 0.05) higher than the required maximum loss of 1%. The 1.6% failure rate in phase 1 involved transponders from a single supplier. Loss of identification was associated with rejection after injection, expulsion during inflammation and technical failure. Three weeks after injection on average 0.6% of the piglets had an observable inflammation and at the time of retrieval pus was found around, on average, 1.2% of the transponders. An average of between 37% and 88% of the transponders were retrieved in the slaughter line from the base of the ear in phases 1 and 2. The other transponders were retrieved medial or caudal to this position. This positional variation meant that it was not consistently possible to remove the transponder from the carcass within the required 4 second time period. It was concluded that the systems should be improved before recommending their introduction on a large scale, because the variation in readability and location is too high.