Eighty-four hypersoft X-ray sources surfaced in Chandra channels astronomers had skipped

The selection carries its own control. A detector artifact or a cosmic-ray hit has no reason to respect a spectrum, so a point source that is present in the softest band and gone from the harder image of the same field ^[4] is not easily written off as junk, and the combination it implies, very soft X-rays alongside bright ultraviolet, is not one that known accreting binaries have shown ^[21]. Eighteen years of Chandra pointings went into the search ^[18].

The denominator deserves a note. Eighty-four objects across six galaxies works out to about 14 apiece ^[17], but that average is fragile: the targets are not a matched set ^[6], and the exposures are whatever the archive happened to hold ^[1]. What is published is a count, not a rate.

Then there is the word "class". Rosanne Di Stefano's reading is that these must be stellar remnants, a black hole, neutron star or white dwarf, ripping material off a companion that emits X-rays as it falls ^[8]. Jimmy Irwin puts the looseness on the record: "They're probably a heterogeneous mix of different types of objects," and "a hypersoft source isn't necessarily a type of object but a representation of what kind of energy is coming out" ^[9]. All that is known about the members so far is roughly how bright and how hot they are ^[10]. That makes the finding a spectral category which may contain three unrelated kinds of stellar corpse, a weaker claim than NASA's description of objects behaving unlike any seen before ^[11], and a more interesting one, because a category defined by output rather than by mechanism is exactly what you get when a detection window opens before the physics does.

Both advertised payoffs sit on the far side of that ambiguity. Accreting white dwarfs are candidate Type Ia progenitors, and those progenitors have been sought for years without success ^[12]; the extreme ultraviolet from these sources might account for ionized gas in galaxies that have no active nucleus to do the ionizing ^[13]. Each requires sorting the 84 into physical types first.

The thing this doesn't tell you is how many there are. The hydrogen and helium between the stars absorbs the energetic ultraviolet that makes these sources interesting ^[14], so the ones detected are the ones sitting where the light escapes. Di Stefano says each source seen stands for many more in the same galaxy that cannot be, the tip of an iceberg ^[15]. An ionizing budget built on hypersoft sources therefore runs through a hidden-to-visible ratio that nobody has measured.

The part of this that travels beyond one paper is the least glamorous. According to Irwin, the very lowest-energy channels were passed over for years because they were thought unimportant ^[16], and a doctoral candidate and two colleagues reading them anyway ^[3] produced 84 detections without an hour of new telescope time ^[1].