This (G. Tammann, Z. Anorg. Allg. Chem. 111 (1920), p. 78) is a paper on the rate of corrosion of metal. I cited it in a paper, as it is supposed to be the earliest mention of the length ∝ square root of time relationship. The copy I found was in a volume at our library that was literally falling apart, so I'm archiving the photocopy I made here.
I am not sure of the copyright status of this paper. It was published in pre-war Germany. I believe that the relevant law is "Directive 2001/29/EC of the European Parliament and of the Council of 22 May 2001 on the harmonisation of certain aspects of copyright and related rights in the information society", but I am not sure of this. In any case, I would probably need to know when G. Tamman died. If it was more than 70 years ago (e.g. 2009 − 70 = 1939), it is probably out of copyright. Maybe it even requires 80 years. But it doesn't really matter because I have no good idea when Tamman died, or even what his first name is. Life expectancy was around 55 years at the time, so I'm putting it down as 50/50.
Update: This paper is also, in a more official-looking capacity, scanned and on the web at Wiley InterScience. They claim "© 1943 Verlag GmbH & Co. KGaA, Weinheim". I don't know where 1943 comes from for an article dated 1920. Also, their scan is much lower quality than mine! So I'm leaving mine here until/unless I receive a (credible) cease and desist from the putative copyright holder.
Due to the equipment available to me in a practical way, this is a scan of a photocopy of a photocopy of the journal. The second photocopy step was used to center the pages for scanner input. The scanner JPEG-damages everything it sees without asking if that's what you want. Nevertheless, it's perfectly readable.
Here is the paper in three formats:
