From Clive Barrett's site:
http://www.mythographica.fsnet.co.uk/t/rw/print.htm
.......................................................................................
Essentially there are three versions of the original artwork produced by Pamela Colman Smith - A, B and C.
Why this should be so may be hinted at in a letter from Pamela Colman Smith, quoted by Kaplan (vol. III). In it she appears to have had her doubts about the quality of the printing, prior to seeing the finished cards.
The 1909 (A) printing did prove to be inferior and was improved for the second printing. From inspection of the 1910 (B) deck, it would appear that this was not as good as it might have been, so it is possible that a decision was made not to use the black plate from B for the Pictorial Key (C).
There is some evidence that this may have been the case as the Pictorial Key (C) illustrations are superior to those of the 1910 (B), and the errors of the copyist interpretation have been corrected.
The cards were printed with line blocks (solid colour) rather than halftone blocks (graduated colour). The colours used were black, red, blue, yellow and grey.
To quote from Modern Printing, J. Southward, London 1913 -
" A line block cannot be made from a photograph or wash drawing without first making a pen and ink drawing (extra cost)."
If one assumes that Pamela Colman Smith's original illustrations were in colour, (line and wash), then of necessary tracings would have been made - for both the cards and the Pictorial Key illustrations.
The method as follows:-
First a tracing of the outline is made, from this tracing a negative is produced from which a zinc plate is etched. Using the same negative bichromate prints are made on zinc for each colour, after a slight etching they are painted with lithographic ink where colour is desired, (tints may be added when and where desired).
Errors occur on several of the 1910 (B) cards which were corrected for the Pictorial Key (C). Some are minor such as on the Sun, one of the sunflowers does not have a stork, others may be more significant. For example:-
Fool - The eagle which some see on the Fool's bag cannot be seen.
Magician - Loses snake belt.
Hierophant - only one of the circles on the floor has a cross, instead of all four.
Chariot - The symbols on his cloths are confused and indecipherable.
Death - The fallen kings crown is distorted and barely recognizable as such. The child's flowers are unclear.
King of Cups - The fish pendent is not recognizable.
That the two are copied from the same original artwork may be seen by close inspection. The split pen stroke caused by increased hand pressure, which occurs in other of Pamela Colman Smith's illustrations is to be found in the C illustrations but not in B deck. However in many examples of Pamela Colman Smith's work a characteristic line ending (the line is terminated by an elongated comma), may be seen. This may be found in B but is absent from C.
Generally the artistic quality of the B is poorer than C, but strangely the reverse is the case with the Empress, the detail, especially the decoration of the cushions, is much clearer on the B than C. A printer might well have more than one person capable of copying in his employ, which may account for this.
So, it is assumed that both C and B were traced from the missing artwork. This being so, then the copyists errors in B would suggest that the C version was a truer representation of Pamela Colman Smith's designs.
The printing plates for the Rider Waite Tarot Deck were 'lost' during World War Two, (this is a euphemism for melted down for the scrap metal). So when the Pictorial Key to the Tarot was reprinted in 1972, it was a photo-lithographic reprint of the 1910 edition. At the same time a deck was produced by adding colour to the Pictorial Key's illustrations. The colouring being done with reference to the earlier edition of the deck. This also was printed by offset lithography, in five colours - black, cyan, magenta, yellow and grey.
A further point, the printed cards both B and 1972 (D) have flat solid colours (with the occasional added tint), Pamela Colman Smith's illustrations would have had a more varied and natural colouring. Unless the original artwork is discovered the true appearance of the cards can only be imagined.
I have not had a chance to examine a set of 1909 (A) cards and unfortunately due to the small size of Dummett's reproductions little can be said about them, however, the line of the right hand border has been moved inwards making the cards narrower than either B or C.
From the above observations it would appear that of all the editions of the Rider Waite Deck, the 1972 (D) deck (and susequent reprints) is the closest we have to the artists original intention.
The above is based on the works mentioned and also Gilbert's A.E.Waite: A Bibliography, Aquarian Press 1983.
....................................................................................................
Storks on the Sun card? What storks? I don't see no steenkin storks. Maybe he meant stalks.
1909 deck from Holly Volley's collection:
http://www.sacred-texts.com/tarot/pkt/img/ar19.jpg
Comparisons of the Sun 1910 1931 1937 and unknown. Be sure to click on the images to see them BIG!
http://home.comcast.net/~vilex/SunComparison/main-Sun.html
As for all them Fools, it looks like the "new" RWS of US Games System not only has changed the font, but has had a Tangerine Attack! Look at the color change.
http://home.comcast.net/~vilex/ShipofFools.html