P/x coding on early P decor

It is not clear when precisely the production of P décor objects at PZH started. Rozenburg had shown their new range of eggshell bone china objects with finely hatched floral decors before the range was presented at the 1900 Paris World Exhibition. But it is most likely it was the great acclaim and publicity the range had in Paris that made Egbert Estië decide to copy the décor, but on heavier PZH earthenware. Based on artists hired from Rozenburg to paint the new P (for porcelain) décor for PZH main production likely started in early 1901.

The early P decor pieces do in their majority have a P/x marking next to the Lazarus gate logo and the monogram of the decorator, and carried the ‘Made in Zuid-Holland’ branding. They never carried the letter code found on the earliest PZH objects that were long believed to be a year code. At some point, probably around 1906/1907, the branding of P décor changed back to ‘Zuid-Holland, Gouda’. By this point, the P/x code was not in use anymore and all production was marked with a model number/décor number code. This did not change with the introduction of N(ew) P décor in 1907/1908. Year marks were only re-introduced in 1918, when production of P, NP and décor 304 had probably as good as stopped. 

The P/x marks on P decor pieces from the early period were long a mystery until collector W. Spijker stated in his 2001 book ‘Nederlandse Keramiek van de Jugendstil tot 2000’ that they were ‘with very high likelihood’ year marks, starting with P/a for 1901 until P/e for 1905. This theory was taken over in later publications.

When this author started to build up a database of pre-1920 PZH objects to learn more about his collection of P décor, the amount of early P décor objects per year mark he encountered seemed very uneven. Was there an explanation for this?

The following research is based on a dataset of  390 P decor objects, representing items in his collection and images shared by and with other collectors, museums, dealers and auctions and all from the early P décor period from c.1901 until c.1907. Of the  390 objects, 275 had legible P/x codes and these were distributed as follows:

                 27 % P/a for 1901

                 66 % P/b for 1902

                   7 % P/c for 1903

It is extremely unlikely that the production of P décor peaked in 1902 and dropped rapidly to zero in the years 1904-1905, as it still continued for many more years after.

A likely explanation could be that the discipline of adding the year code did rapidly disintegrated after 1902. If we evenly distribute the 115 objects without useable P/x code over the years 1903-1905 the distribution becomes:

                 19 % in 1901

                 46 % in 1902

                 15 % in 1903

                 10 % in 1904

                 10 % in 1905

This looks more realistic for a product that after its successful introduction continued to sell in decent numbers. While NP décor replaced much of the P décor sales after it was introduced in 1907/1908, P décor continued to be produced. The 1909 PZH price list for The Netherlands (the earliest surviving document of this kind) shows that of the 285 models in the PZH range 193 could be ordered with a P decor (although in these years with a limited range of decors, some from the early years and some new decors) and 223 with an NP decor.

While this might be an acceptable explanation for the distribution of the year codes over the years 1901-1905, another anomaly presented itself after grouping images per year. All P/a coded objects appeared to be exclusively painted with hatching in one particular style, namely linear lines in multiple colours. And a relationship between year letters and décor was also present for the next year letters.  The hatching on P/b coded objects was almost predominantly executed with curved line hatching and P/c coded objects had a “hatching” pattern that filled the space between other decoration elements with parallel contour lines in a fairly random patron. The photos below illustrate the three styles:

For more examples and a description of these styles see here.

When we compare the correlation for the 275 P/x coded objects between coding and decoration style it is 100% for P/a and linear hatching and 100 %  for P/c and contour hatching (although 2 of the decors had some curved hatching as well).  For P/b coding and curved hatching, 6 objects have bird decors with what resembles contour (P/c) hatching but P/b coded decors One non-bird décor is also coded P/b for what appears a P/c décor. These decors might have been seen by PZH as a special subset within the P/b décor family. Yet, the correlation between P/b coding and curved hatching in 96 %

For all  390 objects, the distribution was 25 % in linear, 68 % in curved and 7 % in contour style hatching, a similar distribution to the P/x distribution of marked pieces. The dataset has only one object that at first inspection seemed to be coded P/d. It appears it was a  mistake by the artist that marked a P/b décor with ‘P/a’ and extended the stalk of his typical ‘a’ somewhat. Objects coded P/e and P/f mentioned in the literature might have been misinterpreted  P/c resp. P/b marks.

The above makes the use of P/x as a year mark for the author less likely: it would mean that after a year of only decors with linear hatching all decors were changed to different decors with curved hatching in the second year and for the third year to yet another style. From objects with P décor made after 1906, we know they still had decors with curved hatching. Only one P/a decor with linear hatching is known from after 1907 and non with P/c decors with contour hatching. The distribution of pre-1906 P/a, P/b etc objects mentioned above does probably say something about their popularity at the time, and hence the stronger survival of curved hatching.

So, was the P/x code used by PZH to indicate different décor families? The author could see only one possible alternative: that the P/x code was a year code used only in the first year of production of a décor family, with the un-coded pieces produced in subsequent years. While this leads to a similar distribution over the production years as on the bottom of the previous page it seems a farfetched assumption as it would serve no practical purpose. And it still meant that completely new décor families were introduced in each of the first three years of production on P décor.

In conclusion:  the author has reasonable doubt that the P/x code was a year coding system and thinks it is more likely that it does represent a code to group décor families executed in a certain style. At present this is presented as a theory; any arguments pro or con will be greatly appreciated.


With thanks to Dennis van den Hoek for suggestions and support.

Scroll to Top