In current practice, an artist prints an edition consisting of a specified number of copies of an image from the block (wood engravings and woodcuts), the plate (etchings, drypoints, mezzotints and the like), the stone (lithographs), or the screen (serigraphs). The edition is thereby capped - or limited - by the stated number of copies. By custom, the artist may also print an additional number of proofs not exceeding ten per cent of the edition size without violating the limitation of the edition. When the maximum allowable number of prints has been reached, the original work - or matrix - will be destroyed or, more probably, cancelled either by defacing the image or by altering it in some distinctive way.
The artist may print the edition personally, in which case the edition printing can take years to complete because the artist need only produce as many copies as will satisfy immediate demand. This has obvious advantages because it not only minimizes the outlay for costly printing papers, but also eliminates the burden of unsold inventory while avoiding the risk of damage or degradation to unsold copies of the edition after the matrix has been permanently altered. However, it is often the case that far fewer than the stated number of prints have been made owing to lack of demand. It is often impossible to tell if this is the case, since many artists fail to keep accurate records, though this is not always so.
The artist may instead rely upon the services of a skilled fine arts printer as many famous artists including Rockwell Kent, John Sloan, Käthe Kollwitz and many others have done. The professional typically prints a series of trial proofs that aim to fulfill the artist's described intent, carefully adjusting his press and the ink as deemed necessary, and then submits these proofs to the artist for review. The artist may request additional proofs if it is deemed necessary, but ultimately selects the impression which best conforms to his original intent, marking it "BAT" ("Bon à tirer", or good to print). This becomes the reference master used by the printer to verify the consistency of prints in the edition. Over time, the collaboration between printer and artist often evolves to a level of mutual trust and understanding, and trial proofs may no longer be called for.
The size of the edition depends on both the demand for the artist's work and the durability of the matrix. The relationship between demand and edition size is directly proportional. Rockwell Kent, for example, could comfortably print lithographs in editions of 150 without fear of unsold inventory while others are more likely to print much smaller editions.
It is also current practice for the artist to sign and number each print in pencil outside of the image area, most frequently in the lower margin. Occasionally, the artist will also add his or her signature to the matrix at the time that it is created. In earlier days, prints signed by hand were far less commonplace; Whistler, for example, signed his work in the plate. Happily, artists are creatures of habit like the rest of us, so if you collect the work of a particular artist the likelihood is very good that most, if not all of his prints will all be signed in the same way.
A work of art on paper often evolves over time. The artist's initial delight with his work may give way to dissatisfaction as minor deficiencies become ever more annoying and ultimately rise to the level of gross offenses to his or her aesthetic sensibilities. Thus the artist will return to the matrix and re-work the offending parts of the image - adding or softening tone here, strengthening or weakening lines there, until the result is again satisfactory. In his catalog of the works of Käthe Kollwitz, August Klipstein notes fifteen different states for the woodcut titled "Hunger" (1925), catalog number 207. This is an extreme example, but serves to underscore Kollwitz's relentless quest for complete personal satisfaction with her work. This is one of the prices to be paid when the artist is her own harshest critic.
If a catalog raisonné has been published for the work of an artist whom you are collecting, you will have need to refer to it more frequently than you may imagine. It is the one reliable place to sort out the different states of a given print, and is often replete with useful contextual background for a particular work. So it is important to either own a personal copy or to have access to the catalog at your local public library (a more feasible option in larger cities or regions with an inter-library loan program). These books can be expensive because they are published in rather small editions aimed at universities, research libraries, museums, and galleries.
If you are looking for such a book, I heartily recommend that you begin and end your search on line at www.vialibri.net. viaLibri aggregates almost all of the major antiquarian bookseller collaborative organizations worldwide, like AbeBooks, Alibris, and Biblio. It's fast and free, but you can also subscribe to an array of research features more appropriate to the professional antiquarian book trade if you so choose.
A restrike is the term used to describe a print made using the original matrix (block, plate, screen, etc.) subsequent to, and unrelated to, the original publishing venture. This is typically a practice sanctioned by the artist or the artist's estate. Such subsequent prints are produced with the same care and craftsmanship as their predecessors. Our inventory includes a number of restrikes made from the original plates created by Käthe Kollwitz by Emil Richter Verlag, her first publisher, and by Alexander von der Becke & Son, which succeeded Richter as her publisher. Kollwitz etchings were printed into the late 1960s, and logically were commissioned to answer an unsatisfied demand for the artist's work. They have traditionally been issued with restraint in order to avoid undermining the market for her work. A signed Kollwitz print has more value than an unsigned Kollwitz print published during her lifetime, which in turn has more value than a print produced subsequent to her death. It is important to note, however, that there is often little qualitative difference in the prints produced across this spectrum save for the inevitable gradual degradation of the matrix caused by the effects of the printing process.