Golden Gate Orchestra (a.k.a. California Ramblers)
A: Fox Trot. Vocal chorus by Marvin Young (a.k.a. Irving Kaufman).
B: Fox Trot. Vocal chorus by Arthur Fields.
A mx: 147672 (label) 147672 3-G-2 (i.e., take 3) (runout); recorded New York, NY, November (?) 28, 1928.
B mx: 147702 (label) 147702 1-G-3 (i.e., take 1) (runout); recorded New York, NY, December 6, 1928.
It appears that Rust had a hand in all of this data, including that which is on the DAHR website. Some of what is there is certainly of more recent vintage than the American Dance Band Discography, and possibly also than Jazz Records. As you note, there is a mention there of take 3 of mx 147672, and it is shown as recorded on the same day as takes -1 and -2. The DAHR is also the only place where Rust listed Ben Selvin's mx 147680.
I wouldn't worry overly much about the Hawaiian mxs. They are clearly anomalous and out of sequence. The seven month gap exists between the Hawaiian matrixes and both what comes before and after. The gap between the last pre-Hawaiian block mx and the first post-Hawaiian block mx is exactly one day. Business as usual.
The date for mx 147680 is still suspicious for being one month out of step with its surroundings. And of course, the dates associated with mx.147702 are still out of step with everything. I would note, however, that Rust adamantly stated in the introductory essays to Jazz Records that if he gave a specific date, it was because the available company records documented it. The difference, if there was one, between the 6th Edition and his previous works was due to the completeness of the records he had gained access to in the intervening time. On the other hand, I have not noticed that the editorial exactness of Rust's works changed markedly over the years. Proofreading was not something his editors spent much time or money on. If Rust made a boo-boo, it was unlikely that anyone except his readers would catch it.
Sadly, Rust is gone now and I hear he left little behind to document his processes or to support his conclusions beyond what was incorporated into his books. We would do well to take a lesson from that and try to be more transparent, publishing not just conclusions, but the basis for them whenever possible.
Thanks for the additional info on how Rust over the years shifted the dates he assigned to these matrix numbers by an entire month. Again, either he had at first misread and incorrectly copied down what he found in his sources, or his dates were simply estimates that he changed in the light of new, conflicting evidence.
The Hawaiian block of matrix numbers had me puzzled too. Here's why:
So even within the tiny Hawaiian block, we can observe a jumping back and forth between dates that makes no sense. And when we leave it with mx 147679, we jump ahead by seven months, only to jump back exactly one month with mx 147680.
Truly, it's a mess.
What can be said, though, is that take 3 of "Glad Rag Doll" (147672) is ascribed to the same day as the first two takes, and that seems very likely since recording up to 3 takes in one session appears to have been the general practice.
It will get muddier before it gets clearer, if in fact it ever does. In the American Dance Band Discography, Rust had the session that produced mxs 147702-1-2 occurring on November 6, 1928. He changed that to Dec. 6 by the sixth edition of Jazz Records, thus reducing the disparity to only 22 days. We should not overlook the fact that the take of mx.147672 in question is the third take, which may or may not have been recorded at the first session when that mx number was used. A gap of only thirty mx numbers between two recordings makes the likelihood of their being essentially contemporaneous events fairly high. A gap of three weeks between an initial session and a re-make session would not be unheard of.
What needs explaining to me is the mx numbers attributed to Honolulu, Hawaii being used in May and June. Compared to many mx numbers used for other California Ramblers recordings made in October, November and December, 1928, those numbers are way ahead of the curve. The most probable explanation, given the unusual location, is that a small block of numbers were allocated for use on the field trip well ahead of the actual date of the field trip, thereby resulting in their being used out of sequence. (Alternatively, the Hawaii recordings could have been cut using somebody else's facilities (and masters), brought back to NY and assigned Columbia matrices when they got there, five-six months later.)
The dates that Rust and DAHR give both for Velvet Tone 1816-V and for Harmony 816-H are patently wrong. Rust claims that the lower matrix number (147672) was recorded 22 days (!) later than the higher matrix number, 147702. If anything, it should be the other way around. December 17, though, was a mistake on my side that I can't explain myself.
Columbia matrix 147701 and 147703 are both dated New York, NY, December 6, 1928. If that is correct, this would confirm a Dec. 6 date for matrix 147702.
Interestingly, though, the Columbia matrix 147673 (Harmony, as you know, was a budget label of Columbia's that used the same matrix numbers) is dated to June 6, 1928, recorded in Honolulu, Hawaii. The following matrix number, 147674, is supposed to have been recorded also in Hawaii on May 20, 1928, so again several days earlier! Columbia matrix 147672, as you said, is dated by Rust to New York, December 28, 1928. Columbia matrix 147680, however, is supposed to have been recorded in New York on November 28, 1928. I find all this very confusing.
The truth, hidden in all this confusion, might be:
Matrix 147672, recorded November 28, 1928. [Maybe Rust wrongly copied 11/28/28 as 12/28/28?]
Matrix 147702, recorded December 6, 1928.
That's what I have now entered in the Notes above.
In the 6th edition of Jazz Records, Brian Rust gives a recording date of Dec. 6, 1928 for the first and second takes of mx 147702 ("My Troubles Are Over). He gives a recording date of Dec. 28, 1928 for the third take of mx 147672-3 ("Glad Rag Doll"). These are the same dates cited by the DAHR and the Online Discography Project (78rpmdiscography). I'm sure I'm not alone in wondering what the basis for your new dates is.