A by-eye search of VST images multiplies the known dwarf galaxies around three hosts twelvefold

Forty-three of the 47 candidates had not been catalogued before ^[18]. For the same fields, the new list is nearly twelve times the old one ^[19], and it averages close to 16 candidates per host ^[20].

A deeper image only adds objects. That makes an incomplete census a clean explanation for a shortfall of satellites and a poor one for an excess. The single route by which this list gets shorter runs through spectroscopy. The team classified the candidates on colour and structural properties consistent with other dwarf galaxies, and says distance and velocity can come only from future spectroscopic observations ^[12]. Every candidate placed at another distance comes off the host's tally.

Automated detection was not used, because algorithms would have easily missed objects this faint. Three of the authors inspected the images independently ^[10], after masking the light of foreground stars and other artifacts so the faint light could be measured accurately ^[11]. Independent passes cross-check which candidates are real, but on their own they do not give a detection efficiency: the fraction of dwarfs of a given brightness and size the search would have recovered. The phys.org account does not say whether the team measured one by injecting artificial dwarfs into its images.

The hosts sit at very different distances: NGC 5068 at about 17 million light-years, the lenticular NGC 5084 at 80, the elliptical NGC 5087 at 100 ^[6]. By inverse square, a dwarf of the same intrinsic luminosity at NGC 5087's distance arrives about 35 times fainter than the same object at NGC 5068's ^[21]. One survey with one exposure depth therefore reaches a different intrinsic faintness limit around each of the three.

"Studying these islands of nearly invisible stars is not just astronomical stamp collecting," said Crescenzo Tortora, the INAF researcher who is principal investigator of the VST-SMASH survey and lead author of the paper ^[15]. He also turned the search into a browser game: "To give people a taste of the thrill of being a dwarf galaxy hunter, I created a small online game that reproduces a process very similar to the one we used," he said ^[16].

The plane reported around NGC 5084 rests on the same unconfirmed membership as the count, so the two wait on the same measurements. Whether dwarfs generally lie in a plane perpendicular to their host's disk, as the Milky Way's satellites do, is still a matter of debate ^[5].