The measurement series behind the three glitches
PSR J1637-4642 was followed with the Murriyang telescope at Parkes Observatory from February 2009 to October 2024, a span of 15.5 years. That series holds three glitches, and their sizes spread across three orders of magnitude: 2.7 parts per million, 2.2 parts per billion and 28 parts per billion. Only the first shows an exponential recovery, with a decay time of about 100 days and a recovery fraction of around 0.015.[1]
The data underneath that series is thin. Because the pulsar is radio-faint, only 166 reliable times of arrival came out of the 15.5 years, from observations roughly every 2-4 weeks with integrations of 2-15 minutes. The glitch epochs are correspondingly loose, quoted with margins of 30, 39 and 16 days. None of the three glitches has a known day, only a known window.[1]
What the word quiet was describing
PSR J1637-4642 went about a decade after its discovery without a detected glitch, even though it sat inside the Parkes young pulsar timing programme. The two steps found later are far smaller than the first, at 2.2 parts per billion and 28 parts per billion. Holding those two numbers together suggests the silence reported on the smallest step a fortnightly cadence and 166 arrival times could resolve, as much as on the star's behaviour. That is not the only reading: the star may genuinely have gone 9.5 years without a large unpinning event, with the small steps beginning only afterwards.[1]
A seismology study published the same week lights up the same distinction from the other side. A deep-learning system pulled 174,929 PKP precursor signals out of more than 2 million seismograms recorded between 1990 and 2024, roughly 10 times the combined total of every earlier global compilation. Earth's core-mantle boundary did not change over that period; what changed was the instrument that could count the signal. In both studies the number in the headline describes the reach of the apparatus doing the measuring as much as the world being measured. The difference is that the seismology team could lower its threshold deliberately and show what that bought, while in pulsar timing the threshold is set by telescope time and by how faint the source is.[1], [2]
The measurement that would settle it
The authors say that measuring the true waiting-time distribution requires timing to continue, and that leaves a testable claim. If Parkes keeps the same fortnightly cadence, I expect a timing update published by the end of 2027 to report at least one further step at or below 28 parts per billion for PSR J1637-4642. An update that reports none at that level would count against reading the earlier silence as a sensitivity limit.[1]