The Great Inkwell Experiment—Conclusion

Last Updated on

The great inkwell experiment is complete! While none of the results are particularly surprising, it has definitely given us some good information and insight as to what to look for in a working inkwell.

The final documentation is available here. It is a stand-alone paper, in PDF format, and includes all of the data, including actual tenth-of-a-gram measurements, observations, and discussion from this forum, and conclusions, etc. Feel free to copy it, distribute it, give it away, use the information in any way that will benefit the pen community. I have digitally signed it to prevent changes, and would prefer that no one sell it and make millions, but those are the only restrictions.

Much of the information in the PDF documentation duplicates what has appeared in earlier posts here on our blog. In order to round out the information here, the remaining excerpts follow:


Our complete data, including specific measures, are recorded in an MS Excel spreadsheet. Copies are included in the PDF documentation. (Realistically, I could spend several hours transferring them to html tables. But I shan’t.)


The trends we see in this experiment are established quickly, and follow through the entire data collection period. Especially when viewed in graph form (see page seven), there are three main groupings: low, medium, and high evaporation rates. When it happened, evaporation happened steadily and decisively.

It is obvious from observation that the design of an inkwell has a strong impact on it’s performance as a means of efficient storage. Those containers with soft gaskets were the top performers, those with a close-fit lid did less well, and those with plenty of air space and loose-fitting (or absent) caps did the worst.

The most efficient group, as far as concerns liquid retention, were the modern ink bottle, the pottery bottle, and the travelling inkwell. The top two performers had almost no measurable loss, although there is a miniscule amount of loss in the raw data not entered on the spreadsheet (which I can make available upon request.) The middle group includes the two looser-fitting-cap inkwells, and, perhaps most surprisingly, the open control. The underdog of the group was the brass-cased glass insert. This one was, as they say, an epic fail which was evident from the first few days. It should be noted that I took a good look at the physical integrity of the insert, as well as for evidence of leakage into the case, but it was sound.

We also observed that what minor fluctuations in humidity and temperature there were had minimal impact on our measurements.


One interesting conclusion is that the role of an ink container is to slow down evaporation as much as possible. Even the best materials and design—even modern ones— will still lose solvent , given a long enough period of time. Given that ink is meant to be used regularly, miniscule loss is acceptable. There are, however, some obvious conclusions at which we can arrive:

  1. Inkwells with gaskets or tightly threaded caps are superior in terms of long-term ink storage.
  2. The age of an inkwell is irrelevant as long as the materials are sound, and the original design is superior. With these caveats, an antique inkwell is a viable choice for ink storage, either long- or short-term.
  3. Badly-designed inkwells might be candidates for retrofitting with a gasket or cork, depending on design.
  4. Certain inkwells are best reserved for aesthetic purposes, or for temporary use. Any inkwell with a loosely-fitting lid needs to have its contents either removed daily, or watered-down every few days, so as to avoid either ink loss, or a change in the quality of the ink due to evaporation.

The numbers show that, in a low-performing container, a ten dollar bottle of ink will lose at least 18¢ of ink per day. This means that the entire amount of ink will have disappeared in just over a month’s time! On a more positive note, a well-made and well–engineered inkwell can keep ink sound and in good condition almost indefinitely, with no attention from it’s owner.

As an interesting side note, the many antique inkwells on the market today are the loosely-fit, hinged type. This put us in mind of the old axiom about suits of armour in museums: their small size is not because the people in medieval times were physically smaller than we are today. They are smaller because the surviving suits were ornamentally made, largely for the young nobility. The truly representative armour was always used to pieces, literally, on the battlefield; too valuable to store away where it would be preserved for posterity and scholars. Perhaps the really efficient inkwells were used so much that there aren’t many left?

What design factors are good when it comes to ink storage? A tight-fitting lid, preferably one which is threaded on, is also very effective. Obviously, many inkwells designed for travelling will have such effective design features. Any soft material seems to do well in preventing evaporation; interestingly, natural materials, such as cork and wood, proved nearly as effective as modern materials. It reminded us of the Japanese chests, made of paulonia wood (the same wood as many traditionally made Japanese kitchen knife handles). One significant characteristic of paulonia is that it absorbs huge amounts of water, very quickly. The reports are that, when caught in a flood, the drawers of these chests swell shut almost instantly, perfectly preserving the clothing inside them. It made us wonder if the wood of inkwell D is simply swelling shut enough to limit air exchange, and thus evaporation.

Finally, it is obvious that there is more work to be done in analyzing the efficacy of older inkwells. Why is it that C failed in such a spectacular way? Would similar designs likewise perform poorly? How do early plastics affect efficient ink storage? What retrofits are possible in order to increase the efficiency of poorly-designed inkwells? Could such retrofits be done in a such way that the aesthetics of an antique inkwell could be preserved? Further experimentation may well shed light on the answers to these, and other, questions.


Various discussions and brainstormings were done on a thread on the Fountain Pen Network; partial transcripts are included in the documentation, but the full thread can be read here. A number of good suggestions were made, and the discussions certainly helped to refine the entire process.

No comments yet.

Leave a Reply