One of these was that the efficiency of the measurements was too low (known as the detection loophole). Although the data obtained violated Bell's inequality test, it may not be a representative sample of a complete set due because some photons in the experiment couldn't be detected. Another loophole was that the measurements were too slow (the locality loophole). If the measurement devices were able to communicate via some unknown, slower-than-light channel they could share information and influence the outcome of the impending measurement.