In presenting these cuts of the shifted transfers, no attempt was made to illustrate any of the recutting, nor was an attempt made to reproduce the shifted parts of the tasselated portion of the border or of the doubling of the rosettes. Stamps from a few positions on the plate show a slip of the engraver's tool from the lines he was attempting to recut, and such slips resulted in a double line. Such slips, however, rarely occurred, and generally show up best on copies from very early impressions from the plate. Probably the best example of one of these slips is noticable in the doubling of the inner right frame line on 89LI. (See Figures 27 and 27A.)

Earliest Known Dates of Use

Fig. 27A

The earliest dated copy of an Imperforate from Plate 1 of which we have a record is February 7, 1852 (Chase), although we understand a Detroit collector has seen a cover dated October 18, 1851. As very few copies, comparatively speaking, were used for domestic rates, it is not strange that an earlier dated cover is unknown to us.

Will all those interested in the study of early U. S. look up their covers bearing copies of the 12c imperforate, and advise either of us of any covers used in 1851, or prior to February 7, 1852? The earliest record we have of the use of a 12c perforate from Plate "1" is a cover mailed from Charleston, S. C., August 25, 1857. Can you show an earlier date?

The following is a list of various cancellations noted by Col. Tracy on singles which were well marked:

Paper, Gum and Shades

The paper used for the Imperforates was a hard, white wove paper, and varied in thickness from very thin to medium, and thick.

The gum was rather thick and was spread very liberally, and varied in color from almost white to an amber shade.

The shades of the stamps ran from a gray black to a very deep black.

For the Perforates, the paper used in 1857 and during the early part of 1858 was thin, hard, and brittle, and generally very white, though most copies show a yellowish brown tint from the gum.

After 1858, a much softer paper was used. The shades vary from a gray black to full black, and late impressions are apt to show a smudgy black appearance. The gum used varied in color from light to dark brown. Plate Three

As stated above, this plate was either made late in 1859 or early in 1860. No doubt it was made just prior to the time the 30c and 90c Plates were made. Stamps from it were probably placed on sale in the spring of 1860, as the earliest date of use of which we have a record is June 1, 1860.

The outstanding characteristic of this plate is the very noticeable uneven or broken outside Frame Lines. These breaks occur in the top and bottom lines as well as in both side lines, and in our opinion result from the fact that the Frame Lines were very lightly engraved on the original die and when transferred from the relief to the surface of the steel plate, breaks in these fine lines occurred because of the uneven surface of the plate. These breaks were corrected on Plate "r" by recutting, whereas, no attempt was made to do so on Plate 3. The Plate was probably rocked in with a one Relief Transfer Roll, and the partial reconstruction of the right pane by Col. Tracy seems to confirm this theory.

When Mr. Perry was compiling his notes on the 12 Cent Plate "3," Col. Tracy furnished him with the information he used regarding the right pane.

We are reproducing herewith that part of the Perry Article (Mekeel's Weekly Stamp News, page 319, June 12, 1920):

"Strong indication that Plate No. 3 was rocked in from bottom to top with a single relief transfer is found in the 8th vertical row of the right pane. Stamp No. 98 has a noticeable shift at the bottom; stamps Nos. 88, 78 and 68 are normal; stamps Nos. 58, 48 and 38 have a blur in the "S" of U. S.; stamp No. 28 shows a very noticeable shift; stamp No. 18 is the only stamp on the pane which is much out of vertical alignment, being closer to 28 than are any other two stamps on the plate to each other. The spacing between No. 18 and No. 8 is normal and this brings No. 8 too low compared to the others in the top row. The obvious explanation is that after No. 68 was rocked in, something got on the roller and was not noticed until it had produced the blur in the "S" found on Nos. 58, 48 and 38; that the workman then raised the roller and cleaned it, but did not replace it in exactly the right position for No. 28, and a shift resulted, or else he commenced to rock in No. 28, and happening to notice the imperfection on the subjects below it, which had just been rocked in, he then lifted the roller to clean it and on replacing it, did not get it in exactly the correct position; hence the shift.

"Perhaps he got a bit rattled then and didn't space No. 18 far enough away from No. 28, and followed up that error by spacing No. 8 the correct distance from No. 18 instead of moving it further upward to the line horizontally with No. 7 and No. 9. The writer cannot conceive how the varieties in the 8th vertical row could have been produced from a roller on which there was more than one relief."

Supplemting the above, are some additional facts. 28R3 is one of the best shifts on the plate.

The spacing between 18R3 and 28R3 is about 4 MM whereas, the spacing horizontally always measures at least 7% MM.

The blur on 58R3, 48R3 and 38R3, in the "S" of "U. S." is located in exactly the same place on each of these stamps, and therefore was surely the result of some foreign substance that adhered to the relief, and this same relief rocked three adjoining positions in a vertical row.

No. 8R3 is the regulation distance of % MM above 18R3, but still it measures about 2 MM below 7R3 and 9R3.

Why Toppan Carpenter used a one relief roll in making Plate "I" of the twelve cent in 1851, when they were using three relief rolls to make other plates, and why this firm nine years later used a one relief roll to make a similar 12 cent plate, when they were using multiple relief rolls in making other plates, cannot be explained by us at this time. It is logical to suppose that the same One Relief Roll used in 1851 was again used in 1860.

From Plate "3" come the varieties of the 12c stamps, known as "double side lines." It has been incorrectly stated in articles in the past that these double side line stamps come from both the right and left panes of Plate 3, but the partial reconstruction of this plate by Col. Tracy virtually proves that all such varieties originate only from the left pane. These extra side line varieties are the result of vertical lines ruled from the top to the bottom of the plate, and while not always continuous, they generally are found extending the full length of the stamp.

It is believed that each position in the third vertical row had an extra line to the right of the right outside frame line, and each position in the fourth vertical row had an extra line to the left of the left outside frame line.

Fig. 28

Thus two extra vertical lines were ruled between the 3rd and 4th vertical rows. The distance of these extra lines varies from what appears to be on some copies as a doubled appearance of the frame line to 3% MM. (See Fig. 28.)

Both of these extra lines are of about the same thickness as the outside frame lines.

On some of the positions in the fourth vertical row the extra line is so close to the outside left frame line, that it has the appearance in single copies of an attempt to recut the frame line; however, such was not the case. (See Fig. 29 for typical example.)

Between the 7th and 8th vertical rows of the left pane is a single line, which we believe runs from the top to the bottom of the plate. This line measures from 2 to 34 MM to the right of the right frame line of stamps in the 7th row, and about 12 MM from stamps in the eighth row, and is heavier than the other extra lines, and therefore heavier than the side lines of the stamps.

Fig. 29

Fig. 30

Figure 30 shows part of a block of 4 which includes 87L3, 88L3, 97L3, and 98L3. The extra line can be distinguished in the cut just to the right of the perforations. This extra line is very similar to the center pane line, as the distance between the right and left panes measures only 134 to 12 MM, which is less than the distance between certain vertical rows on the plate.

In reference to the Center Line of the plate, this line is heavily drawn, especially so at the bottom, and measures 5% to 34 MM from stamps in the right pane, and 8 to 1 MM from stamps in the left pane. (Fig. 31.)

There is a slight break in the Center Line at the top, opposite the top of the stamps in the top row.

Stamps showing the Center Line generally show a straight edge at right or left along the Center Line, whereas, stamps from the 7th and 8th rows are always perforated. Thus it will be noted that out of every 200 stamps printed from Plate "3," there were 30 copies which had extra side lines exclusive of those showing the regular pane or Center Line.

We do not know at this time why these extra lines were drawn on the plate.

While the reconstruction of the plate is not complete, the distribution of the guide dots seems to be as follows:

Dots are found in or near the lower

right hand corner of nearly all posi-
tions, except the 10th vertical rows of
cach panc.

This confirms our theory that Plate
"I" was rocked from top to bottom and
Plate "3" from bottom to top, as dots
in the top row of Plate "I" are above
the upper right and left corners of the
designs, with no dots below the posi-
tions in the bottom row, whereas, on
Plate "3," dots are found below the
lower right corner of the design of the
bottom row and also in the top row,
but no dots are found above the de-
signs in the latter row.

Fig. 31

The dots on Plate "3" are not consistent in size, shape, or position, but vary greatly.

Another point confirming our theory regarding how the two plates were rocked is as follows: Plate One has the greatest number of shifts in the top row, whereas Plate Three has the greatest number in the bottom row.

The principal shifts on Plate "3" are 8R3, 28R3 (Fig. 32), 38R3, 92R3 (Fig. 33), 95R3, 98R3 (Fig. 34), and 92L3. Shift 92R3 is the most pronounced on the plate Other minor shifted transfers have been noted, but their positions are unknown at this time, due to the fact that the reconstruction of the plate has not been completed.