A mx: 78256 (take 3); recorded New York, NY, January 17, 1919.
A mx: 78256 (take 4); recorded New York, NY, February 5, 1919.
B mx: 78255 (takes 2 & 3): recorded New York, NY, January 17, 1919.
Label printing date codes:
BX (= February 1919)
DX (= April 1919)
The Talking Machine World, May 15, 1919, page 164: Advance Record Bulletins for June, 1919 (i.e., released in late May 1919).
Images
Number:3553116 THUMBNAIL Uploaded By:han enderman Edited By:scrough Description: Columbia 2707 A side label (DX with 2 lines of patent text = later April 1919 label pressing)
Number:3553113 Uploaded By:han enderman Edited By:scrough Description: Columbia 2707 B side label (BX with 2 lines of patent text = February 1919 label pressing)
Number:3553109 Uploaded By:han enderman Edited By:scrough Description: Columbia 2707 A side label (BX with 2 lines of patent text = February 1919 label pressing)
Number:3553117 Uploaded By:han enderman Edited By:scrough Description: Columbia 2707 B side label (DX with 2 lines of patent text = later April 1919 label pressing)
Number:3553106 Uploaded By:han enderman Edited By:scrough Description: Columbia 2707 A side label (DX with 1 line of patent text = early April 1919 label pressing)
Number:3553108 Uploaded By:han enderman Edited By:scrough Description: Columbia 2707 B side label (DX with 1 line of patent text = early April 1919 label pressing)
Number:1034013 HIDDEN Uploaded By:GumboStu Description: Columbia 2707 A side label (DX with 2 lines of patent text = later April 1919 label pressing)
Number:1034012 HIDDEN Uploaded By:GumboStu Description: Columbia 2707 B side label (DX with 2 lines of patent text = later April 1919 label pressing)
Thanks again, scrough, for taking on this editing job! I've reordered the labels chronologically. I don't know if the A and B sides always belong together, of course. The labels were certainly not uploaded in this order.
The older label design of the two is the one with the single line of patent text at the bottom (sub-variety ii). It was replaced by the design with two lines of unequal length (the lower one longer) (sub-variety iii) as early as January 1919 (AX).
That said, we can see here that older label designs were sometimes used up for months. Thus, the two images {3553116} and {3553113} with the earliest date codes (BX = Febr. 1919) are, in fact, on the newer label design (sub-variety iii, which, as I said, was introduced as early as Jan. 1919).
Images {3553109} and {3553117}, _if_ they are from the same record (but the different lighting speaks against it), would show a transitional phase: both have the two rows of patent text, but one was printed in February 1919 (BX), the back side (DX) in April 1919.
Images {3553106} and {3553108} are late uses of label sub-variety ii with the single line of patent text, both with DX (April 1919) date code. This design was introduced in June 1918 (FY). The latest use I have seen has an EX date code (= May 1919), see Columbia E4236 on Discogs.
The lower quality images that I have now hidden, {1034013} and {1034012}, both have the more recent label variant (subvariety iii) with DX date code (April 1919).
It's fascinating to me that there is not a single actual repress among these 8 pictures, despite their variety of date codes and designs. All were printed before the record's official release in May 1919.
Source, as always, M. Sherman & K. Nauck, "Note the Notes," page 28 (they date the subvariety ii to "ca. 1918)."
Xiph: Another label ordering please? All images now edited. I thnk I understand the printing codes, but here the older label designs have later dates than the newer.