Transmitters for dense wavelength division multiplexing (DWDM) consist of banks of lasers, with one laser for each channel. It would be nice if a single laser could produce a comb of frequencies exactly aligned with the channels. Researchers at Bell Labs in Hilversum, the Netherlands, and at National Technical University of Athens in Greece developed a device that promises to do just that.The design incorporates a semiconductor laser, external mirrors and an intracavity Lyot filter. Developed in the early 20th century by the French scientist Bernard Lyot, a simple Lyot filter consists of a birefringent crystal between two polarizers placed at 45° to the optical axis of the birefringent material, so the entering light is divided evenly between ordinary and extraordinary polarizations. The polarizations propagate with different phase velocities, so that, for a given length of birefringent crystal, only one set of wavelengths is perfectly aligned to pass the exit polarizer.A Lyot filter, therefore, produces a comb of frequencies, and the separation between the "teeth" of the comb depends on the length of the birefringent crystal. Normally, the filter employs multiple birefringent crystals of different lengths, yielding a number of combs of different spacings. Because very few frequencies align with all the teeth, the result is a filter with a few widely separated and extremely narrow passbands.In this case, the researchers constructed the Lyot filter by substituting a 5.77-m length of polarization-maintaining fiber for the birefringent crystal. By placing the Lyot filter inside the resonator with an InGaAsP/InP laser, they were able to produce 23 output wavelengths, each separated from its neighbors by 100 GHz, the typical spacing for telecom channels. The power across the 23 channels was constant to within 3 dB, and the extinction between the channels was 17 dB. By scanning the individual lines with a fiber Fabry-Perot filter, they determined that each channel had a bandwidth of 12.5 GHz.The total output from the laser in all channels was approximately 42 µW. The scientists suggest that the device may find a place in laboratory DWDM transmission evaluation experiments to reduce the cost of the transmitter, but they do not indicate that it would be suitable for field deployment.