A couple of references from Google Books:
https://books.google.it/books?id=xjRBBAAAQBAJ&pg=PA55&dq=leading+point+s...
https://books.google.it/books?id=IXHCKKhswdgC&pg=PA20&dq=leading+point+s...

I don’t understand why you would want an authoritative reference for that. It’s pure mathematical logic and therefore undeniable. It doesn’t require any “metal type guy” saying so in a “metal type book”.

The first baseline is at “top edge” (0) + x.
The top edge of the second line is at 0 + point size (y).
The second line’s baseline is at top edge of the second line (=point size) + x.
The difference between each consecutive baselines is therefore y + x - x = y (=point size).

I say: it is true.

I’m a letterpress printer. Hence an authority on the matter (if that’s sufficient).

;)

Well, it certainly is true from mathematical logic when the type in both lines belongs to the same font.

It can be false, though, if the first line is Century Expanded (set on the Standard Line) and the second line is Cloister Lightface (set on the Art Line or Script Line). In the case of metal type, the position of the baseline within the metal type slug is variable depending on the typeface. (I see, though, you specifically said "in a given font".)

In the late 19th century, for the convenience of printers, the major typefaces made their types lining by having only three standard baselines - one for classic typefaces like Caslon or Jenson, used also for script faces, one for the workhorse faces of the day with fairly short descenders, and one for titling faces.

The differences between these baselines were made to be in whole point increments as far as possible - and 1/2 or 1/4 point if not.

Before that, the baseline was an individual matter, depending on the typeface.

Sounds like Mr. Phinney is having to deal with the kind of blockhead that denies anything that isn’t in a printed book.

While this is almost always true, interestingly, Bringhurst chooses to point out a case where this isn't true, in his only example of a metal type sort in the newly added spread on setting lead type.

For a reference, see Ralph W. Polk's The Practice of Printing, or even better, Printing for School and Shop by Frank S. Henry — both high school printing textbooks from the first half of the twentieth century.

Sounds like Mr. Phinney is having to deal with the kind of blockhead that denies anything that isn’t in a printed book.

That's entirely possible, in which case my reply would not have helped him. One possibility is that he was looking for a reference for Wikipedia.

Since the question is one of terminology, I presume the Oxford English Dictionary should be a reliable reference. Here is the corresponding entry:

c. Typogr. Having no leads between the lines; unleaded; (see also quot. 1888).

and here is quot. 1888

1888   C. T. Jacobi Printers' Vocab. 128   Solid matter, type composed without leads; also applied to type with but few quadrats in.

Here is something else from the Treasury Board of Canada Secretariat http://www.tbs-sct.gc.ca/fip-pcim/glo-eng.asp

Blank line (ligne de blanc)
A blank space equal to the type size used. For example, a blank line for 10-point type, set solid, is specified as 10 points, but measures 20 points from base line to base line.

Not completely true. Monotype would cast type on a smaller measure on request, eg 12 on 11, etc. I have seen references to this in catalogues.

Bert: True, but that would be specially requested, and then the size of the slug itself would be affected.

Michel: Those citations, although all clear, don’t directly address the combination of factors in question.

Yes, I should have used the caveat that it is only certain with the same font on both lines (although many typefaces use a standard/common alignment). I thought of that, and said as much in my report.

Agreed with all that it seems a simple case of logic and understanding, not needing a reference. I didn't use one, in the end.