Article MT307
Note: place cursor on red asterisks for footnotes.  Sound clips are shown by the name of the song being in underlined bold italic red text.  Click the name and your installed MP3 player will start.

The Ballad of George Collins - versions or variants?

A preliminary survey

Many scholars, including Tristram Coffin, A L Lloyd and Bertrand Bronson, agree that when Professor Francis James Child was working on his monumental work The English and Scottish Popular Ballads he included two ballads, Clerk Colvill (Child 42) and Lady Alice (Child 85), which were actually two parts of the same story.  The bulk of the story may be found in Clerk Colvill, who, ignoring his mother's or sweetheart's advice, goes to the water's edge where he is seduced by a mermaid (or possibly some form of water-spirit).  Clerk Colvill complains that his head is aching and he is told that this is an omen of his impending death.  In Lady Alice, a much shorter ballad, we find that Clerk Colvill here called Giles Collins, has died and Lady Alice is mourning for her loss, thus rounding off the story.  According to A L Lloyd, 'Either these are two separate songs which have been combined to form George Collins or (which seems more likely) or they are two fragments of the complete ballad.'  Versions of Clerk Colvill have been found scattered throughout Western Europe, the earliest known one being dated to a German manuscript poem of c.1310, whereas versions of Lady Alice appear much later in printed form.  A version printed by David Herd in his Ancient and Modern Songs, Heroic Ballads, &c. can be dated to 1769, for example.

What I want to do now is to consider the second ballad, Lady Alice, on its own, because it is this ballad which has entered oral tradition in North America and where several versions were recorded commercially in the 1920s and '30s.  Why, I wonder, did this ballad become so popular and what exactly was it that made it so popular in the first place?  I will also be using the title George Collins for these versions of Lady Alice.  I hope that this will not be too confusing, but most, if not all, the American singers mentioned here used the title of George Collins rather than Lady Alice and so I am following their example.

But firstly, though, let us consider Professor Child's three texts to Lady Alice.  Version 'A' comes from Robert Bell's Ancient Poems, Ballads and Songs of England (1857).  Version 'B' is from the 1810 edition of Gammer Gurton's Garland, while Version 'C' comes from Miss M H Mason's Nursery Rhymes and Country Songs, printed in 1877.  Gammer Gurton's Garland is subtitled 'The Nursery Parnassus: A Choice Collection of Pretty Songs and Verses for the amusement of all Little Good Children who can neither read nor run' and I find it of interest to see that both Child's 'B' and 'C' texts come from publications which contain songs from, or suitable for, children, especially when we consider that today this ballad would hardly be considered to be a story that was suitable for young children.

According to Bell 'This old ballad is regularly published by the stall printers.  The termination resembles that of Lord Lovel and other ballads.'  Today, we might say 'broadside printers', rather than 'stall printers'.  Child also notes that a version of Bell's text was 'heard sung forty years before 1856 (in) Philadelphia', in other words sometime c.1816.  Bell's text is as follows:

*A quoif, or coif, is a close fitting cap.  It covers the top, back and sides of the head and was in use in Britain from the 13th to the 17th century, although it did survive longer in some rural areas, where it was worn by women and young children.

**Mention of 'a garland of marjoram, and of lemon-thyme, and rue' is interesting, in that such listings of herbs occur as a refrain in other ballads, such as versions of The Elfin Knight (Child 2), which have a supernatural element.  The idea was that by singing the names of herbs which had magical qualities, the singer was protecting himself or herself from the Devil or some other form of evil.  Is this, therefore, a lingering echo of the mermaid/water spirit who seduces Clerk Colvill in the earlier part of the ballad?

Professor Child's 'B' and 'C' texts contain a number of stanzas which occur in these two versions, but which do not occur in his 'A' text.  For example, in both texts there is mention of Giles Collin's mother making 'water-gruel' for her son.  In the 'C' text we also find Lady Alice's mother making 'plum-gruel' for her daughter.  In fact, this verse does sound slightly odd to my ears:

It is the line about the doctor who 'ate up the rest', which rather seems to be out of place here.  Could it be that, at one time, the song was sung on the stage, rather in the manner that the Music Hall singer Sam Cowell would perform traditional songs in his act? Bell's comment about the similarity to the ballad of Lord Lovel (Child 75) is clearly shown in the final two stanzas of Version 'C'.  The stanzas also occur in many versions of Barbara Allen (Child 84).

Let us now return to Version 'A' - Bell's text.  If we remove the final two stanzas we are left with just four stanzas.  In stanza one Lady Alice is sitting in her 'bower-window' mending, presumably by sewing, her quoif.  She sees a corpse being carried past the window.  The body is identified as Giles Collins in stanza two.  In stanza three Lady Alice orders that the body be laid upon the grass and says that she will also be dead 'when the sun goes down'.  Then, in stanza four, Lady Alice asks to be buried 'all for my love so true' and requests that a garland be made of marjoram, lemon-thyme and rue, three herbs which had magical qualities.  If we compare this text with the versions which were commercially recorded in America in the 1920s and '30s, then we find that only two stanzas, stanzas one and two, correspond with stanzas found in the recorded texts.  And we may also say that only parts of versions 'B' and 'C' occur in the recorded versions.  So it would seem that the recorded versions are not directly descended from these three British texts.  But, we do know that at least one version of George Collins was being sung in America c.1816 and I would suggest that it is possible that texts were by then being printed on American broadsides and chapbooks.

I will now consider the American recordings.

I am not going to examine just why these American recordings were made because I have already done so in two previous Musical Traditions articles - When Cecil Left the Mountains and When Cecil Left the Mountains Part 21 - and readers wanting to know more on this subject should consult these articles.  I would, however, simply say that the recordings were made by singers who could possibly have learnt some of their repertoire from oral sources and so I will also be later considering versions of the ballad which had been collected from Appalachian singers prior to when these recordings were made. 

In the case of George Collins, the title most often used on the recordings, we know that at least twelve different versions of the ballad were recorded prior to the Second World War, though four of these versions were never issued.  One singer, Henry Whitter, had two different recording issued, so that means that we have eight recordings, from seven performers, available to study.  The recordings which I wish to consider are:
   SingerTitleRecording dateLabel
    Kelly Harrell
Henry Whitter
Henry Whitter
Henry Whitter
Roy Harvey & North Carolina Ramblers
Dillard Smith
Dick Justice
Jack Hicks & Charlie Dykes
Emry Arthur & Della Hatfield
Jess Johnston
Riley Puckett
Dixon Brothers
Dying Hobo
George Collins
George Collins
George Collins
George Collins
George Collins
One Cold December Day
George Collins
George Collins
George Collins
George Collins
The Story of George Collins
June 9th 1926
c. August 1926
September 7th 1926
October 18th 1926
February 16th 1928
May 3rd 1929
May 20th 1929
November 8th 1929
c. Oct/Nov 1929
November 14th 1931
March 29th 1934
September 25th 1938
Vi 20527.
Bwy 8024, Her 75536.
OK unissued.
OK 45081.
Br 250
Ge rejected.
Br 367.
Ge rejected.
Pm 3222.
Ge unissued.
BB B-5818.  MW M-4551.
MW M-7580.
BB = Bluebird.  Br = Brunswick.  Bwy = Broadway.  Ge = Gennett.  MW = Montgomery Ward.  OK = Okeh.  Pm = Paramount.  Vi = Victor.

It seems clear that, with so many recordings, George Collins was once a popular song.  Recently, on an on-line web-page issued by an American record collector, we find this comment, 'George Collins was a common tune from the area where Henry (Whitter) lived.  George Collins may have been a person from that area.'  I mention this, not to criticize the record collector's lack of knowledge about the origins of the ballad, but rather to show how someone could assume that the ballad was based on a real person who may have lived in the Appalachian region, where Henry Whitter lived.  And, if a person listening to one of the George Collins recordings could think this way, then is it possible that Appalachian singers also considered George Collins to have been a local man, one of their own?  I presume that the collector's comment, that it was a 'common tune', came from the fact that it had been recorded by so many performers.  And this is an interesting point.  Why should so many recordings have been made?  George Collins is not the best-known ballad in Professor Child's collection, when compared with, say, Barbara Allen (Child 84).  At the time that the above recordings of George Collins were being made we know that fourteen recordings were made of versions of Barbara Allen.  However, on closer examination, we find that one performer, the prolific Vernon Dalhart, recorded the song on six different occasions for different record companies, and that two performers, Frank Luther and Bradley Kincaid, each made two recordings for different companies.  So, to be accurate, we may say that only seven performers actually recorded Barbara Allen.  In other words, there were fewer versions of Barbara Allen recorded than there were of George Collins.

If we begin with the earliest recording, by Kelly Harrell - a Virginian singer who came from Fieldale, a musically rich area of the State close to the small towns of Galax, Hillsville, Fries, Independence and Mt Airy - we find the following text:

If we analyse this text we find four stanzas, which are normally associated with the ballad, sandwiched between an opening stanza from another song, The Dying Hobo, and a so-called 'floating' verse, one which turns up in most of the versions considered here.2  We might consider this 'rogue' opening stanza something of an oddity, were it not for the fact that another recorded version, this time by Dick Justice, a West Virginian, also includes a similar opening stanza.3 While it is clear that both the Kelly Harrell and Dick Justice recordings have a similar opening stanza, from another song, the Justice version only contains two stanzas from George Collins and these are followed by two and a half 'floating' stanzas, including the one about the lonesome dove.  As Justice repeats two of the 'floating' stanza (actually one and a half, to be accurate) I would suggest that he was not really familiar with the ballad of George Collins and that the final stanzas are sung rather as 'fillers'.  So where did Dick Justice get his version of the ballad from? 

We can, however, suggest that another singer, Ollie Gilbert from Mountain View in Arkansas, learnt her version of George Collins directly, or indirectly, from a recording.

Ms Gilbert used the same six stanzas as did Kelly Harrell, though the final two stanzas are reversed in Ms Gilbert's version.  We can see that the lines are not word for word identical, but there is, I think, sufficient similarity to suggest a connection between both of these versions. 

The next recording was made by Henry Whitter, from Fries in Virginia, a small town close to Kelly Harrell's hometown of Fieldale.  Whitter recorded the ballad three times in 1926.  The first version was recorded c. August, 1926 and issued on both Broadway 8024 and Herwin 75536.  A second version, recorded on September 7th 1926 for the Okeh company was unissued, but he recorded it again for Okeh on October 18th 1926 and this recording was issued on Okeh 45081.  This is the text from the Broadway recording. Another version, this time by Roy Harvey & The North Carolina Ramblers, was recorded in February, 1928, over a year before the Justice recording was made, and this version, which deals mainly with George Collins, does contain two 'floaters', both similar to the ones that Dick Justice employed. Roy Harvey, like Dick Justice, was also originally from West Virginia.  It may be that the version recorded by Emry Arthur and Della Hatfield some eighteen months later was influenced by the Harvey recording.  Interestingly, both the Harvey and the Arthur/Hatfield recordings are also backed with recordings of the same song, The Bluefield Murder, on the B side of their respective records.  An examination of the two Bluefield Murder texts shows that both texts are, to all intent, identical.4  If, as looks possible, Arthur is copying the Harvey recording of The Bluefield Murder then there is a possibility that he had heard the Harvey recording.  An analysis, (see below), of both the Harvey and Arthur recordings of George Collins also suggests that both versions are related.  They may not be identical in every word, but there are certainly sufficient similarities to suggest a relationship. There then followed an almost five year hiatus before the Georgia singer Riley Puckett laid down his version of the ballad.  Puckett was both a solo singer/guitarist as well as being a member of Gid Tanner's Skillet Lickers and recorded a vast number of songs and tunes. The final commercial recording was that made four years later by the Dixon Brothers (Howard & Dorsey) who were originally from South Carolina, though they later moved to North Carolina to find work in the Cotton Mills.  This version begins with a more or less standard two stanza opening, before we find a stanza where Mamie enters the 'chambry of death' to view George's body.  The ballad ends with its own stanza, beginning with the line 'The golden Sun sinking in the west', before ending. Having considered these examples of commercially recorded versions of George Collins I now wish to consider the individual elements that make up the recorded examples.  I have chosen eleven separate points which occur within these versions.  Not all of these points occur within any single version, though several points do occur within separate versions.  The numbers indicate in which stanza the elements occur.

 Dying hoboRode homeSewingWept/mournedWhy weep?Open Coffin
Kelly Harrell1 2345
Dick Justice1 2  3
Henry Whitter 1235 4
Roy Harvey 12 4 3
Emry Arthur 12354
Riley Puckett 123 4
Dixon Brothers 12364
 Stay with body
all nigh
Golden sun sinking
in the West
Kelly Harrell 6   
Dick Justice 645 
Henry Whitter 6   
Roy Harvey 6 5 
Emry Arthur 7 6 
Riley Puckett 5   
Dixon Brothers57  8

We can see that the opening Dying Hobo stanza occurs in two versions (Kelly Harrell and Dick Justice) and that this element replaces the opening stanza - where George Collins rode home - which occurs in all of the other versions.  Three other elements also occur in all of the recorded versions.  These are where (Lady Alice - actually his girl, a girl, this maid, sweet Nell (x2), Mary and Mamie) is sewing, where she orders the Coffin to be Opened so that she may Kiss the Body, and the floating verse about the Lonesome Dove.  Assuming that the Lonesome Dove stanza is indeed a floating verse, one that possibly became attached to the ballad once it was in America, then we may indeed confirm that the recorded versions of the ballad only share two stanzas with the three versions printed by Professor Child.  Two other elements - Why weep?, which occurs in five of the recordings, and Stay with the body all night also seem to belong to early versions of George Collins, whereas the other elements, Lonesome train, Lonesome Road and Golden sun sinking in the west, are probably other American floating verses.  If, as I suggest, the Why weep? and the Stay with the body all night elements are from other (Old World?) versions of George Collins then these must be versions that were unknown to Professor Child.

It will be seen that all of the recorded versions are quite short, each one containing only a few stanzas from the actual ballad, together with one or more 'floating' verses.  Field recordings from England show that the ballad could actually be much longer.  For example, a version collected in Hampshire contains ten stanzas, while a set collected in Gloucestershire contains eleven stanzas.5  It may be felt that the recorded versions of the ballad are of similar length because of the limited time available on a 78rpm disc - namely about 3 minutes - and that these versions had been shortened to fit the recording times.  But, if we look at a selection of collected versions of the ballad we find that many of these versions are of a similar length.  This set is from the singer Frank Proffitt (1913 - 1965).  Proffitt, who was from Watauga County, NC, said that he had known the ballad since he was a little boy and that his mother, and his mother's sisters, used to sing it to him.  Frank's mother was Rebecca Alice Proffitt, née Creed, and she had been born on November 11th, 1878, in Surry County, NC.  If, as Frank Proffitt says, he heard the song 'since he was a little boy', then we may assume that his mother, and her sisters, learnt it possibly before 1920 - 25, in other words before it had been commercially recorded. 

It is also possible to examine American versions of George Collins which predate the recorded versions.  Cecil Sharp, for example, printed five versions in his Appalachian book.6  Three versions are printed with full text, while the two remaining sets only show the tune and the first stanza.
 Rode homeSewingWept/mournedWhy weep?Open Coffin
Frank Proffit1234 
Mrs Dora Shelton12346/7
Mrs Hester House11/2235/6
Miss Mary McKinney1234/56
 Stay with body
all nigh
Golden sun sinking
Frank Proffit 5   
Mrs Dora Shelton 5   
Mrs Hester House 4   
Miss Mary McKinney 7   

Again, in these for texts we see that the core of the ballad is to be found in the Rode home/Sewing/Wept or mourned/ Why weep? motifs, which occur in all four versions.  The open coffin motif occurs in all three of Sharp's version, though not in the Frank Proffit version, while the Lonesome Dove motif again occurs in all four versions.  In other words, the similarities found in these four texts suggest that, at one time or other, they may have been derived from a common source.

The ballad of George Collins was clearly once popular in parts of the Appalachian Mountains of North America.  Collectors, such as Cecil Sharp, found many versions in the early part of the 20th century and several singers recorded the ballad in the 1920s and '30s.  We know that at least one version of the ballad was being sung in Philadelphia c.1816 and that it must have travelled with settlers into the Mountains.  But why should this ballad have survived on the lips of so many singers?  Many of the versions collected before the ballad began to be recorded commercially show a number of similarities, which indicates that they may all stem from a common source either priinted or oral.  And the same may be said for most of the commercially recorded versions.  Most of these versions are clearly related to the versions which were collected by people like Cecil Sharp prior to the commercially recorded versions being made.  There are anomalies, of course, such as the opening stanzas recorded by both Kelly Harrell and Dick Justice (the stanza about the dying hobo) which would seem to have crept in from another song and, once recorded by Kelly Harrell, may have been heard by Dick Justice who based his recording on that by Harrell.  And the 1938 Dixon Brothers recording ends with a unique stanza, one which may have been learnt in South Carolina, before the brothers moved to North Carolina in search of work.  If this is the case, then their version would have been learnt some good distance from where the other singers lived.  I sometimes wonder if we assume too much when it comes to songs found in specific areas.

Cecil Sharp covered a huge area of the Appalachians in some six States - and I can imagine people thinking that there would be an almost unlimited number of songs being sung there.  But, after about 50 weeks spent collecting during the period 1916 - 1918, Sharp came to realise that he was no longer finding 'new' songs.  He was actually hearing versions of songs and ballads which he had previously collected there and that is why he ended his Appalachian journey.  He agreed that there could still be some songs that he had not heard, but he definitely believed that he had the bulk of the material down in his notebooks.

If we examine Sharp's Appalachian collection we do find that many songs and ballads were sung to similar tunes, as is the case with recordings of George Collins.  Why should this be?  I believe that if singers and singing communities are isolated from other singers and communities, then there is every likelihood that tunes and texts will develop separately and become distinct.  We might expect this to have happened in Appalachia, where communities were once separated by deeply wooded valleys and mountain chains.  But singers were singing similar songs and ballads, and this suggests that there must have been considerable movement between the early settlers, who were possibly not as isolated as we tend to imagine.  Many singers around Beech Mountain NC have recalled that they learnt songs from a wandering minstrel called 'Lie-hue', who would travel among the communities - turning up once every seven years, while other mountaineers mention similar song-carriers passing through their regions.  The ballad of George Collins was one piece that minstrels such as 'Lie-hue' may have carried with then.  Cecil Sharp also confirmed that he was told of singers who had moved away from their singing communities.  When he later went to visit these singers he discovered that they had often taken their songs with them.

I previously mentioned another well-known ballad, Barbara Allen, and perhaps there is a parallel between George Collins and the ballad Barbara Allen.  Professor Child clearly saw a connection, because he placed them next to one another in his collection.  Barbara Allen was given number 84 in his collection, and Lady Alice number 85.  Both ballads carry the same, simple story, namely the death of two lovers, one dying upon hearing the news of the death of the other.  It is a Romeo and Juliet story in miniature.  And then there is the rose and briar ending.  When the Appalachian singer Dan Tate gave me his version of Barbara Allen he added that 'those last two verses just cain't be beat'.  There is a strange beauty in the symbolism of the two plants intertwining from the two graves, a beauty that has lasted for generations Perhaps we will never fully know just why the folk chose to sing George Collins, rather than some of the other ballads which occur in Child's collection.  But honesty, truth and beauty, must surely feature somewhere in that riddle.


Article MT307

Top Home Page MT Records Articles Reviews News Editorial Map

Site designed and maintained by Musical Traditions Web Services   Updated: 22.3.17