Andrew Cooper
Oct 6, 2020    Waikoloa, HI (map)
20cm f/6 Newtonian, Cave Astrola @ 76x
Seeing: 5 Transparency: 6 Moon: 0%

Nothing notable appears visually at the listed position, particularly anything that matches William Herschel's description "A cl of sc stars filling the field". There are certianly more suggestive clumps of stars in the region, with the obvious NGC 6645 just 20' minutes north.

The original William Herschel coordinates precess to 18h32.7' -17°17 (J2000). Archinal and Hynes pick a clump around BD-17 5224 at 18h31.5' -17°21' with a modest dark nebula running along the southwestern side in agreement with SIMBAD. Corwin picks a clump a bit north and west at 18h33 -17°14 near BD-17 5231 on the other side of the dark structure.

Harold Corwin

NGC 6647 = H VIII 14. William Herschel found this on 18 June 1784; he called it "A Cl of sc pL sts" in his PT paper. The sweep has it as "A cl of sc stars filling the field; not rich and the stars less than those in the foregoing [M 25 = IC 4725, which see]". John Herschel also claims to have seen it; his position -- somewhat east of his father's and marked "+-" in RA -- is adopted in GC and NGC. Curiously, he has it as "A very loose parcel of v small stars, hardly noticeable as a cluster," not in very good agreement with his father's description. Further, in the GC, John Herschel has "Cl, L, Ri, lC, sts vS." How he got that out of his and his father's observations is a mystery to me. Perhaps he penned the description in haste, or mistakenly copied it from another object.

There is nothing obvious at either position. William Herschel's original position is about 8 arcmin west-northwest of his son's. There is no evidence in the sweep that William Herschel's position is incorrect; reducing it with respect to another star (HD 170680) in the sweep gives the same position as the star that William Herschel uses in his published list (43 Sgr). The position for M 25 that William Herschel finds is about 25 seconds of time west of the modern positions, but that cluster is much larger than his field, so he would have had some trouble estimating the center.

Just four arcmin northeast of William Herschel's position is a group, about four arcmin across, of a couple of dozen stars. The brightest is around 12th magnitude. These may be some the stars that William Herschel took to be a cluster, so I had earlier listed it as a possibility for the NGC object.

However, having finally (in early February 2016) seen the sweep with it's fuller description, it's clear that William Herschel's cluster is much larger, on the order of the field size, 15 arcminutes. The only DSS image that offers a fairly clean view of the field is the DSS1 blue image. Mentally filtering out the myriads of background stars leaves a sprinkling of brighter stars that more or less match William Herschel's description. I make this apparent "cluster" approximately 13 by 10 arcminutes, and put the center almost exactly at William Herschel's position. But seeing it in the DSS image leaves me unimpressed; I'd be surprised to find a real cluster here.

Curious to know if this larger grouping persists at the eyepiece, I checked Steve Gottlieb's visual observation from July 2013 with a 24-inch telescope at 125X; his field was 50 arcminutes across. Steve sees only half a dozen 10th magnitude stars centered a few arcminutes northeast of William Herschel's position. These may be the same clump that I picked out earlier, and are included in the larger "cluster" that I think I see on the DSS. Like me -- and John Herschel -- Steve is underwhelmed.

Digging into the literature, I found a reference to the cluster in one of Walter Scott Houston's Sky and Telescope columns (in the January 1976 issue), asking for information on this cluster and other of the RNGC "nonexistent" clusters.

This led me to Archinal and Hynes who suggested the star cloud 1.2 minutes west and five arcminutes south of William Herschel's position. Tracing this back to Brent's Webb Society Monograph, I see that he credits one of my early lists of (R)NGC corrections based on my SEGC sweeps for galaxies -- I was indeed checking every NGC number in the south-equatorial zone. I now think this star cloud is very unlikely as there no hint of an error in William Herschel's observation that would push his position back this far to the west. And William Herschel describes a scattering of brighter stars, not the fainter ones comprising the cloud.

The professional catalogues are no help, either. The Alter et al (1970) catalog has only a token entry for the cluster at the same position as Archinal and Hynes give (which is probably from Alter et al), though does give an integrated magnitude of 8.0. Lynga (1987) repeats this, but makes the declination just an arcminute to the north, perhaps a typo.

Finally, working with the Archinal and Hynes position as input, Kharchenko et al. (A&A 558, 53, 2013) used PPMXL and 2MASS to find a "remnant cluster" here. Their algorithms put the remnant at 5.25 kpc away with a barely significant proper motion, and give it 136 stars over a diameter of thirteen arcminutes, with 43 in the central 7.2 arcminutes, and just three in the 1.5 arcminute core. They also note "Poorly seen in the rich background." My feeling is that they have applied their computer pipeline to a random patch of the Milky Way. This, by the way, is cluster number 2939 in their catalog. My overall assessment of this "remnant cluster" -- "Underwhelmed."

Whatever the case, the clump of stars that I believe to be William Herschel's object does not match the NGC description, though it does more or less fit what William Herschel himself recorded. So, I am going to take it, with some trepidation, as NGC 6647.

― NGC Notes by Harold Corwin

Drawings, descriptions, and CCD photos are copyright Andrew Cooper unless otherwise noted, no usage without permission.

A complete list of credits and sources can be found on the about page