Question: When is a supernova not a supernova?
Answer: Now that's an interesting story...
It all started on Christmas night 2005, when astronomers using the Katzman Automatic Imaging Telescope (KAIT) in California discovered an apparent supernova not far from the center of the elliptical galaxy NGC 2274. There was nothing there on an image they had taken two weeks prior. Twelve hours later, Astronomers at the National Astronomical Observatory of China confirmed the 18th magnitude object was real. It was named SN2005md and the discovery was announced in CBET #332 on December 26.
A spectrogram taken on December 28 showed it to be most probably a "young Type-II supernova". This was announced in an IAU Circular (8650) on the 29th of December. Subsequent KAIT images showed that SN2005md faded rather quickly and it was fainter than magnitude 19.8 by January 2006.
Normally that would be the end of the story, but this time it wasn't.
First, it is generally accepted that the progenitors of core-collapse supernovae are massive young stars. These massive young stars are almost always found in spiral or irregular galaxies dominated by young stellar populations.
NGC 2274 is a strangely shaped early irregular galaxy (an E-type galaxy), so SN2005md was unusual. In fact, it was only one of 22 examples found in an extensive literature search of all early irregular galaxies containing core-collapse supernovae in history.
A paper published in 2008 by Hakobyan et al. (2008, A&A, 488, 523) examined all these cases and found that 19 of the galaxies had been mis-classified, and were actually spiral (17), irregular (1) or ring (1) galaxies. Of the 3 remaining galaxies with early type classification, one (NGC 2768) is a suspected merger remnant, another (NGC 4589) is definitely a merger, and the third (our NGC 2274) is in close interaction with another galaxy. This seemed to explain the contradiction of core-collapse stars residing in old non-star-forming irregular galaxies, since some amount of young stellar population in these interacting galaxies is expected.
Well then, all was right in the Universe once more...or was it.
In February 2008, while Hakobyan and company were putting the final touches to their paper for submission, an electronic telegram (CBET 1265) was issued announcing that either a new supernova in NGC 2274, very close to the position of SN2005md has erupted at magnitude 18.5, or that SN2005md itself had suddenly re-brightened!
The difference between the previously reported position of SN2005md and the "new" object was on 0.1 arc seconds in R.A. and 0.4 arc seconds in declination, but at the distance of NGC 2274 (estimated to be 70 mega parsecs) that could mean they were unrelated objects 120 parsecs apart. Measuring the exact positions of anything that faint close to a galaxy is tricky business and the likelihood they were the same object seemed greater than the probability they were two SN in the same galaxy that close together.
The fact that the previously reported spectrum only showed a featureless blue continuum, with no obvious broad supernova features, and that the object faded so quickly added to the suspicion that SN2005md wasn't a supernova at all.
The telegram went on to explain if the new object was indeed a re-brightening of 2005md, possible explanations were that it was the super-outbursts of a luminous blue variable (LBV), or multiple flares of the LBV as part of an extended eruption. Other possible explanations included a Galactic variable star or a background AGN/blazar.
Needless to say, SN2005md was a mystery. Further observations were encouraged.
Flash forward to July 2010. Astronomers Telegram (ATEL) #2750 finally sorts it all out for us. A fully reduced spectrum taken with the LRISp on the Keck I 10 meter telescope on December 31, 2005 shows that the object originally classified as a young Type IIb supernova is in fact a galactic cataclysmic variable. That's right, it's in our own Milky Way galaxy. NGC 2274 just happens to lie in the background very close to its position on the sky. The CVs spectrum shows features typical of a dwarf nova in outburst. The Balmer emission lines were the clincher. They have an average redshift of about 300km/second, which is far to little to be part of a galaxy estimated to be receding from us at 5000+km/second.
This also explains the re-brightening in 2008, since CVs are prone to outburst over and over on various timescales from weeks to years. It also resolves the conundrum of having a core-collapse supernova in an E-type galaxy with few signs of active star formation.
And once again, order has been restored to the Universe.
The only mystery that remains is why it took them so long to figure this out. The spectrum that resolved this issue was obtained New Years Eve, 2005!