Walter Breen
Prior to 1828, all American coins - including proofs -were struck in screw presses, from radially dentilated or sawtooth-bordered dies wider than the coins, to accommodate flan expansion on impact. Planchets were cut out from strips rolled to approximately the thickness of finished coins, the cutting being done in a press by a device analogous to a cookie cutter. Each planchet was weighed, too heavy ones "adjusted" by being given one or several strokes with a file, lightweight ones remelted. Planchets intended for proof coins were cut from strips which had been thoroughly cleaned and burnished. Planchets intended for gold and silver coins (and copper ones through 1795) were run through the Castaing machine, earlier described, coming out with mathematical roundness and reeded, ornamented or lettered edges. After various additional inspections, weighings, cleanings and dryings, they went to the appropriate striking presses. Regular planchets intended for ordinary circulating coins were given one impression from the dies, then while still hot knocked into a hopper or basket, as likely as not jostling against their neighbors and receiving mint abrasions. Proofs, however, received other treatment. Dies intended for them were specially polished on the fields, the (intaglio) devices left untouched. In some instances they were made up with special care; in others, apparently, they were regular dies removed from the press and polished up for the occasion. In many instances, especially before 1834, they were later used On regular business strikes. Proofs were given two to four blows from these polished dies, the coin being carefully refitted into the lower die between each two successive impressions. Occasional early proofs will therefore show traces of double, triple or quadruple impressions here and there, especially on obverse. These made before 1828 will not show knife-rims or "wire edges", because this phenomenon requires the close collar not then in use. They will normally show very sharp striking-up evenon borders, stars (most stars will show centers), and upper edges of letters. Details of hair, feathers, and drapery will be much clearer than on normal business strikes. Occasional proofs will have only the obverse die polished but the striking quality will be excellent on both sides. A few copper proofs in the late 1820's and early 1830's may be bronzed.
Earlier press operators, through 1827 on all denominations and through some years in the 1830's on some, used "open collars" - thin, resilient plates with openings larger than the finished coin, and placed just around and above the striking surface of the lower die -to aid in centering planchets at striking. The point of using open collars was precisely not to constrict the edges, as applying pressure on edges would squash the edge reeding, lettering or ornamentation, and impede ejection from dies. With the advantage of having edges protected against clipping by such ornamentation came the disadvantage, especially on larger coins, that it was often difficult to bring up the designs fully or to prevent very indistinct striking of details at and near borders. Some half-dollars in the 1817-34 period are unattributable precisely because dies dies differed almost wholly in peripheral details - the very same details often not struck up properly. It is thought that heavier presses were used to make proofs in that period precisely because they were necessary to force stars, letters and borders to come up properly on the finished coin, sometimes even requiring multiple impressions.
Beginning in 1828, Mint engraver William Kneass began a long-range plan of improving appearance and physical qualities of U.S. coins. This involved adopting Boulton's device of a close collar (called a "collar die" in the Newman-Bressett book). Coins so struck would have not only a raised protective rim but also equality of diameters, greater thickness, improved stacking qualities, improved striking-up of details near borders. Beginning with 1828 for dimes, 1829 for half-dimes, quarter-eagles and half-eagles, 1831 for quarters and half-cents, 1834 for some cents, 1836 for dollars and half-dollars, Kneass made dies with deeply cut plain rims surrounding beaded borders, and these dies were closely fitted to the intended diameters of coins to be struck from them. In addition, they were intended for use in heavier presses with "close collars." A close collar fitted closely over the neck of the lower die, and was much thicker and heavier than former open collars; planchets would fit closely into it before striking, and extremely tightly after striking. The newly struck coins would therefore come out only by a special ejection process: the lower die would rise through the collar, pushing out the coin and allowing it to be brushed into the basket by metal fingers. This immediately rendered obsolete the Castaing machine. Lettered edges were no longer possible: raised letters would either be sheared off or squashed to invisibility, incused letters (as on dollars and half-dollars prior to 1836) would be squashed, reeding would be squashed. Edge reeding could be imparted with a grooved close collar, but it would have to be vertical only.
Proofs so made could be struck with fewer blows, most likely two (as at present) or at most three. They continued to be struck in screw presses long after steam coinage (1836). Many, though not all, come with "squared" borders and "wire edges," more properly called "knife-rims or knife-edges" (in British publications). This phenomenon occurs when metal is forced minutely into the extremely harrow crevice between die and close collar, and this could happen if the collar or die gave way even slightly, or if unusual amounts of pressure were exerted. Knife-rims are unusual in this period on coins for circulation, but they do occasionally occur; most often from giving way of collar or die. Proofs in this period are found more often without them than with them.
At this juncture, Newcomb's criteria for intentional brilliant proofs need to be re-examined. Newcomb was writing about U.S. large cents from 1816 to 1857, and was troubled - as have been many numismatists since - by coins claimed to be proofs by those with commercial motives for exaggerating the importance of any coin with shiny surfaces. Newcomb insisted that those and only those coins are proofs which satisfy the following criteria: perfection of strike, without weak areas anywhere on the coin; perfect centering; brilliant mirrorlike fields and dull or frosty devices; dies sharp and new, perfect or at worst with very faint cracks; all stars show their centers; border beading complete, uniformly strong. In particular, coins with any frosty areas in fields are not proofs.
There are two tests of these criteria: (1) Do all coins in or from original proof sets of the year conform to them? (2) Do coins struck from proof-only dies all conform to them? American proof sets dated prior to 1858 are very rare, those dated prior to 1850 hardly seen once in a generation. Nevertheless, they do exist. Photographs of the King of Siam's 1834 proof set - the one with the 1804 dollar and plain-t 1804 eagle - provide additional evidence to that given by the early proofs in the Smithsonian and several other collections. More important, from 1834 on certain die varieties are known to exist only in proof, and comparing various examples of these coins with each other and with other alleged proofs of the period has proved very instructive. Newcomb's criteria are unequivocally too high. Aside from restrikes, early proofs (those prior to introduction of closed collars) have sharpness on all parts of devices superior to that of uncirculated coins of the same date and type; mirrorlike surface on fields is uninterrupted; borders are strong, though centering is not always perfect; stars are sharp, though not every star will always show its center even on the best proofs in some dies; occasional coins as late as 1836 may have proof obverse and frosty reverse. The mirrorlike surface generally possesses more brilliance and 'depth' than will that of a regular business strike made from new dies retaining polish; in general, a proof will clearly and accurately mirror a finger or pencil point held above the field, while a first strike will seldom do so.