2004A&A...425..255J

We show that all long nulls in PSR B0818-13 are trains of rapidly alternating nulls and pulses (each shorter than one pulse period). Sometimes only the nulls coincide with our pulse window, resulting in one of the apparently long nulls seen occasionally. We show these are seen as often as expected if during such a train the probability for nulls is 1.2 times less than for pulses. During nulls, the subpulse drift-speed appears to increase. We assume that the carousel of sparks that possibly underlies the subpulses actually slows down, as it does in similar pulsars like PSR B0809+74, and conclude that the subpulse-drift in this pulsar must be aliased. The carousel must then rotate in 30s or less, making it the fastest found to date.