Last year, Ronan posted a bit called “Making Adjustments” at By Common Consent (here, with useful comments all the way to the end) that hashed out some of the issues with and hermeneutical strategies for bringing together revealed and scholarly understandings. (See also: Joel’s post from Friday.) The Gold Plates’ putative chemical composition provides an example of revealed-subsequently canonized-language “adjusting.”  Joseph Smith-History 1:34, quotes Moroni, an angel, as saying “there was a book deposited, written upon gold plates, giving an account of the former inhabitants of this continent.” Does “gold” mean “100% pure, elemental gold,” a gold-based alloy, or a color?  How much could such plates plausibly weigh?
Joseph Smith used the word “gold” from very early, but I’ll begin with the Eight Witnesses, who attested only “the appearance of gold.” The Book of Mormon itself doesn’t require that either Nephi’s Small Plates or Mormon’s plates be pure gold, or, for that matter, gold at all. It’s also vague about how, when, and to where the plates were moved, but seems to imply that Moroni could carry them by himself. Smith described transporting and hiding the plates in various circumstances, including an at least two-mile foot chase. Reports (of varying reliability) indicate fifteen other individuals hefting the plates, including five women. 
The earliest two weight estimates—“something more than 20 lbs” (1831) and “[Smith]…put it [the Gold Plates] under his arm and run all the way home, a distance of about two miles. He said he should think it would weigh sixty pounds, and was sure it would weigh forty” (1833)—come from antagonistic sources, neither of which complain about the weight.  La Roy Sunderland did complain in 1838: “…admitting there were as many plates as there are pages in the book; and that each plate weighed not less than one pound each; these plates must have weighed not less than five hundred and fifty pounds.” A few months later, Parley Pratt rejoindered that “there is only half as many leaves in a book as there are pages, for one leaf makes two pages…. Besides, a thin gold plate, about 7 by 8 inches, and about the thickness of tin, would not weigh a pound: and you should know that the ‘Egyptian’ is a much shorter language than the English. 
Joseph Smith – History 1, cited above, was written in 1839 under Smith’s close supervision. The following year, Orson Pratt published “An Interesting Account…,” whose description of the plates became a portion of the 1842 Wentworth Letter after Smith tweaked the wording and reduced the dimensions from “not far from seven” x8″x6″ to 6″x8″x6″—though he left in “appearance of gold.”  Smith’s editing and subsequent use make this statement the non-canonical gold standard (Ha!) of his considered opinion about the physical Plates. In 1853, Martin Harris reported a stack height of “about four inches” and then, in 1857, John Hyde brought the chemistry and the size estimates together, using 7″x8″x6″ dimensions, data about a box of plate tin, and a series of ratios to estimate 220 pounds for the Gold Plates. Since the plates “were not so compactly pressed as boxed tin,” he reduced the estimate to “nearly 200 lbs.” In contrast to Sunderland, for whom the plates, as described, were too small, Hyde’s objection was that they were too heavy to carry while beating off “two ruffians,” and running two miles; “Statements must be probable, and, therefore, these ought to be rejected.” The 10% weight reduction Hyde allows for stacking inefficiency amounts to about 8% of the stack height. 
Over the subsequent few decades, the original observers made various statements, most of which matched earlier valuations and interspersed “gold” and “appearance of gold” language. In 1880 JS-H 1 was canonized and in 1884 William Smith became the first to explicitly suggest an alloy: “a mixture of gold and copper.”  Martin T. Lamb joined the fray in 1887 but, unlike Hyde, focused on information density: “there were only plates enough to furnish from one-third to one-eighth of the contents of the Book of Mormon.”  He asserted the “well-known fact” that Egyptian characters “occupy a great deal of space,” which—together with Harris’ 4″ stack with 2/3 sealed, 50 plates per inch, and a one-to-one page/plate correspondence—yielded 200 total plates with 67 available for translation, which fell far short of the 563 pages in the Book of Mormon. Lamb acknowledged Orson Pratt’s objections—6″ versus 4″ stack, double-sided plates, and space-efficient Egyptian—but rejected them out of hand.
Moving to 1908: one B.F. Cummings responded in the Liahona: The Elders’ Journal to a piece in the Christian Standard.  He began with a swipe: “This argument is old and has been revamped we don’t know how many times. Few of those who use it are honest enough to credit it to the first man who published it, Hyde.” He then parries the issue of God’s intervention:
Stale, flat and dishonest as it is, this argument bears as close a resemblance to a scientific objection to the Book of Mormon as the present writer remembers ever to have seen. ¶ And as a scientific objection let us examine it. The reader will please remember that we are now in the domain of physical science, and are dealing with its demonstrated truths and laws, with which our opinions and conclusions must square or be cast aside. While we believe God to be a God of miracles we are to say nothing about them.
Cummings’ emphasized inefficient stacking, estimating that each plate occupied two to four times its thickness. He noted that engraving removed some metal and concluded that the plates weighed not more than 65 pounds and plausibly less than 50. Chemistry anchored his argument: “The plates…were very thin…and being of pure gold were easily wrinkled and indented.” So, while Hyde allowed an 8% inefficiency, Cummings posited 100 to 200%.
Janne M. Sjödahl responded to Lamb in 1927.  With Mormon 9:32-33 establishing Hebrew as an upper epigraphic limit, Sjödahl calculated the surface area required for Hebrew and Phoenician Books of Mormon, including the 116 lost manuscript pages, and thence estimated 45 plates—which fit easily under Lamb’s 67-plate limit. On the weight, Sjödahl articulated a composite explanation of concise text, copper-alloyed gold, and inefficient stacking due partially to hammered rather than cast plates and to engraving. Following Lamb’s 4″ stack and 50 plates per inch, Sjödahl predicted a raw weight of 123 pounds from which the inefficiencies would be subtracted, concluding that “the volume must have weighed considerably less than a hundred pounds.” Finally, he reversed the information density problem: since he estimated 45 plates for the translated third, he estimated 135 plates for the whole set. The raw estimate for 200 plates was 123 pounds, so 135 similar plates would weigh about 83 pounds, which, with inefficiencies subtracted, yielded a final weight of no more than fifty pounds. Interestingly, though Sjödahl’s analysis is more sophisticated than its predecessors, he also wiggles, emphasizing that the dimensions are approximate, that Smith might have only carried the unsealed portion, that Smith “was an unusually strong man, physically as well as mentally,” and that critics also challenge the weight of the ark the covenant.  Additionally, he fudges, using the 4″ instead of 6″ stack height in his calculations. 
Reed H. Putnam advanced Gold Plate Chemistry studies in a 1966 Improvement Era article that analyzed some of the implications of gold-alloy plates and suggested an archeological basis for the alloying technology.  His analysis calls for plates made from between 8- and 12-carat gold to avoid the electrolysis and brittleness of an extremely low-gold alloy and the excessive weight and malleability of a high-gold alloy. He also addresses the electrolytic issues of a copper-gold mix and how those might have been avoided, as well as specifying at least a 125% stacking inefficiency. More recent apologetics have done likewise.  In 1986 Ron Doxey recapitulated the composite explanation but additionally emphasized the number of individuals who reported hefting the plates; in 2001 John Gee verified the plausibility of Sjödahl’s text-size estimates. 
Apologists point out that neither 1830s nor contemporary English require that “gold” mean “24-carat gold” and that it is not uncommon to refer to a ring made of 10-carat gold as a “gold ring.” On the other hand, Sunday Schoolers still speculate about how much plates of pure gold would be worth, Primary children sing enthusiastically and without complicating details of “golden plates,” and artistic renderings show perfectly stacked plates. Further, the weight/composition objection remains viable in some circles.
Significantly, almost all of the apologists herein cited avoid interventionist-God arguments. The closest are Cummings and Sjödahl, who make naturalist cases but hedge with Deus ex machina fallbacks. The problem of overly-heavy plates is easily solved by a well-attested miracle: make the hefters temporarily stronger. Such a miracle accords with the Book of Mormon (Nephi bursting bands, Ammon dis-arming sheep rustlers), the Bible (Samson), subsequent Mormon history (the invisible-angels-pushing-handcarts episode), and any number of non-Mormon-specific faith-promoting rumors about lifting cars to save babies, and so on. More emphatically, Smith’s story of the Gold Plates begins, ends, and runs through with angels, some of whom moved the Plates. Nevertheless, on the question of Gold Plate Chemistry, apologists have tended to pursue naturalism-in church-sponsored periodicals-and imagine it as helping religion advance rather than retreat in the face of secularism.
Sorry, purists, but I left out many of the footnotes. Also, if tense disagreement or inconsistency bothers you, well, I’m sorry about that, too.
 I will use capitalized “Gold Plates” or “Plates” to refer to the bound set of plates from which Joseph Smith translated the Book of Mormon, thus distinguishing them from the various other individual and collected plates.
 Many reliable sources put the word “gold” in Smith’s or the angel’s mouth, so the issue would face us regardless, but the “Standard Work” status ups the ante.
 The Eight Witnesses; Martin, Lucy, and daughter Harris; Lucy, Emma, William, and Catherine Smith.
 Abner Cole, “Gold Bible, No. 6.” Palmyra Reflector, 19 March 1831, Vol. 2, Series 1, No. 16. Howe, Eber Dudley. Mormonism Unvailed. Telegraph Press, Painesville Ohio, 1834, p. 246 (quoting Willard Chase in an 1833 affidavit).
 Pratt’s plates were thin enough to stack more than 40 per inch. Sunderland’s “Mormonism Exposed and Refuted” was a newspaper series and then a pamphlet (Zion’s Watchman, New York, New York, 13 January to 3 March 1838). I have only examined the portions quoted in Parley P. Pratt “Mormonism unveiled: Zion’s Watchman unmasked, and its editor, Mr. L. R. Sunderland, exposed: truth vindicated: the Devil mad, and priestcraft in danger!” (New York, 1838, p. 36.
 Orson Pratt, “An Interesting Account of Several Remarkable Visions, and of the Late Discovery of Ancient American Records,” Ballantyne and Hughes, Edinburgh, Scotland, May 1840. Joseph Smith, Jr., “Church History,” Nauvoo Times and Seasons, 1 March 1842, Nauvoo, Illinois. 3:707 (706-710); also HC 4:535-541.
 Martin Harris: David B. Dille, letter, 15 Sep 1853, in Dan Vogel, Early Mormon Documents 2:298-99. Hyde, John. Mormonism: Its Leaders and Designs. W.P. Fetridge, New York, 1857, p. 243-4. I will report stacking inefficiencies as the percentage of the actual stack height relative to the optimal stack height. Thus, a solid block would have an inefficiency of 0%; a stack where each plate occupied twice its height would have an inefficiency of 100%.
 William Smith interview, The Saints’ Herald, 4 October 1884, p. 644.
 Martin Thomas Lamb, The Golden Bible, or The Book of Mormon, Is It from God? (New York: Ward & Drummond, 1887), p. 245-50.
 B. F. Cummings, “Weight of the Plates,” Liahona: The Elders’ Journal, 18 July 1908, pp. 108-10.
 The book chapter expanded a 1923 article. Janne M. Sjodahl, “The Book of Mormon Plates,” Improvement Era, April 1923, as reprinted in JBMS 10(1, 2001):22-24. Janne M. Sjodahl, An Introduction to the Study of the Book of Mormon (Salt Lake City: Deseret News Press, 1927), p. 35-46.
 He questions whether Smith “had charge of the sealed part of the volume, as well as of the part that was not sealed.” He quotes Orson Pratt: “…about two-thirds were sealed up, and Joseph was commanded not to break the seal; that part of the record was hid up,” (emphasis in Sjödahl) interpreting “hid up” to possibly mean that “the sealed part was hidden somewhere when the translation of the other part was in progress,” which would obviate the weight concern. I find his evidence weak and the counter-evidence strong.
 He quoted Smith’s 6″ estimate in the chapter. Since he was responding to Lamb, he needed to use the same dimensions as him, but Lamb did not then turn about and make an argument about the plates’ weight. What was a “fair” assumption for the information density question becomes a major-favorable-under-estimation for the weight question. Using the six-inch height, the “raw” weight becomes 188 pounds, which becomes 127 pounds for 135 plates. Whereas originally he only had to assume a 40% reduction due to stacking inefficiency, engraving, and alloying, with the six-inch numbers he would have to assume a 61% reduction to arrive at the same 50-pound final weight.
 Reed H. Putnam, “Were the Golden Plates Made of Tumbaga?” Improvement Era 69, no. 9 (Sep 1966):788-9, 828-831. John A. Widtsoe and Franklin S. Harris, Jr. addressed topic in 1937 and Sidney B. Sperry did so in 1964, but I have not examined their arguments. John A. Widtsoe and Franklin S. Harris, Jr., Seven Claims of the Book of Mormon (Independence, MO: Zion’s Printing and Publishing Company, 1937), p. 36-38. Sidney B. Sperry, Problems of the Book of Mormon (Salt Lake City: Bookcraft, 1964).
 Ash (1998) rehashes Putnam’s analysis bolstered by a 1984 paper by Heather Lechtman, “Pre-Columbian Surface Metallurgy,” and an analysis by Robert Smith (“The ‘Golden’ Plates,” 1992). See also Henrichsen (2001).
 Roy W. Doxey, “I Have a Question,” Ensign, Dec. 1986, 65. John Gee, “Epigraphic Considerations on Janne Sjodahl’s Experiment with Nephite Writing,” JBMS, 10(1, 2001):25.