People who wish to analyze nature without using mathematics must settle for a reduced understanding.
Richard P. Feynman
Rudolf Virchow's original 1858 paper in German describing the discovery of neural glial cells. pdf
Bernhard Riemann's original 1859 hand written paper in German describing the Riemann hypothesis. pdf
Alan Hodgkin's and Andrew Huxley's original 1952 paper in the Journal of Physiology describing their model of the neuronal action potential. They received the Nobel Prize in Medicine or Physiology in 1963 for their work. pdf
Roger Tsien's 1998 review on green fluorescent protein in Annual Review of Biochemistry, published exactly ten years before his Nobel Prize in chemistry for his discovery of green fluorescent protein (shared with Osamu Shimomura and Martin Chalfie). pdf
Grigori Perelman's three 2006 papers published to the ArXiv archives that prove the Poincaré conjecture. Of note, Perelman famously declined both the Field's Medal and the million dollar Clay Institute prize (http://www.claymath.org/millennium-problems) for his achievement. In fact, he does not mention the Poincaré conjecture in his papers at all. Instead proving Thurston's geometrization conjecture which implied the Poincaré conjecture. pdf pdf pdf
Andrew Wiles' 1995 proof of Fermat's last theorem published in the Annals of Mathematics. pdf
Re-print of Claude Shannon's 1948 paper in the Bell Systems Technical Journal seminal introduction and development of what would become information theory. pdf
Shannon's development of information theory was motivated to a significant degree by the original work of Ralph Hartley in 1927 on the "Transmission of information" (pdf) and Harry Nyqvist's 1924 work on the speed of information transmission in the telegraph (pdf). Both originals are reproduced here.
Although Shannon's information theory is better known, one of the most beautiful and eloquent descriptions of information is Kolmogorov complexity, also referred to as algorithmic complexity or algorithmic information theory. It was developed independently and in somewhat different flavors by Andrei Kolmogorov (http://www.kolmogorov.com), Ray Solomonoff (http://world.std.com/~rjs/), and Gregory Chaitin (https://ufrj.academia.edu/GregoryChaitin). An almost comprehensive repository of on-going research and publications related to algorithmic information theory can be found here. Here we reproduce Kolmogorov's original 1968 paper. pdf
John Wheeler's original 1989 essay on the information-based 'it from bit' nature of the universe. pdf
The papers and commentaries by Albert Einstein reproduced here represent only a cursory sampling of his entire written legacy. Einstein was a prolific writer, with about 80,000 manuscripts and pieces of correspondence (by him or too him) known. There are two projects that are superb scholarly resources that reproduce in digitized high resolution most (all) of Einstein's writings, including digital reproductions of his hand written manuscripts. These projects are the The Einstein Papers Project at CalTech (http://www.einstein.caltech.edu/ , and see http://einsteinpapers.press.princeton.edu/ for their collection of digitized manuscripts), and The Einstein Archives Online (http://www.alberteinstein.info/). These two sources are the authoritative scholarly online repositories of Einstein's collected works.
Einstein's 1945 paper published in the Annals of Mathematics introducing general relativity. pdf
Einstein's Nobel Prize lecture. pdf
His 1936 paper from The Annas of Mathematics. pdf
"On the method of theoretical physics", Herbert Spencer lecture, Oxford University, June 10, 1943. pdf
Commentary in Science (1920). pdf
Commentary in Science (1936). pdf
Commentary in Science (1940). pdf
Commentar in Science (1951). pdf
Collection of correspondence between Einstein and Barrie Stavis, 1953. pdf