If a filtered bacteria-free phage lysate of a broth culture of coli bacteria is shaken with aluminum oxide C g, the phosphate-containing high-molecular phage protein is adsorbed preferentially. It can be eluted again from the centrifuged and washed adsorbant by means of a phosphate buffer (m/10 to m/30, pH 7). This adsorption procedure is repeated several consecutive times on the lysate [and] the eluates are pooled and dialyzed until the removal of phosphate ions. The adsorption procedure can be repeated on the purified phage solution. Freshly precipitated aluminum oxide is also suitable for adsorption; it adsorbs better, but is apparently less selective. More details on the technique will be communicated soon. The success of the adsorption technique is illustrated by the following example. In a phage assay (0.5 ml phage solution + 5.0 ml coli broth), a lysate pre-purified by aluminum oxide C g was still lytic after dilution to a protein content of 2 × 10−12 g/ml (visibly positive = +). The pure phage solution prepared from this lysate, with a protein content of 3 × 10−15 g/ml, was strongly lytic (+++) in the phage assay. The smallest lytic protein quantity is probably 1–3 × 10−16 g. Starting from the assumption that bacteriophages, like phytopathogenic viruses, are high-molecular protein bodies, the maximal molecular weight of a phage may be calculated as about 100 millions. The diameter of the[se] supposedly spherical molecules is thus approximately 60 mm. The specific clouding of the protein increases with growing purity. The above-mentioned lysate had a specific opacity of 0.8 and the phage solution here had an opacity of 22 (values in other cases were 10–22). If such phage solutions are examined in the Siemens-hypermicroscope of E. Ruska and B. v. Borries according to the technique developed for other viruses, one obtains images as shown in Fig. 1 and and2.2. Phages appear in pure preparations as very small rounded corpuscles (Fig. 1), whose dimensions seem to depend on the phage protein concentration in the solution, so that higher concentrations lead to larger aggregates. The phage solution seems to become more homodisperse in higher dilutions. Isolated corpuscles show then diameters of 40–80 mm. These values are of the same order of magnitude as the particle diameter calculated above from (biological) activity. Presently, it cannot be decided whether we are in presence of phage protein molecules or still aggregates. It was observed with certainty that phages can be destroyed by long and intensive electron irradiation, leaving round, hollow structures behind. In addition, we refer to the article of H. Ruska, in this issue, on the visualization of bacteriophage lysis. Fig. 1. III 73/39. Highly purified suspension of coliphages, strongly lytic till 10 g/ml. El. opt. : 25 000:1. Fig. 2. III 71/39. As Fig. 1 after intensive and prolonged electron irradiation. El. opt. : 22 000:1. Berlin-Dahlem, Biologische Reichsanstalt fur Land- und Forstwirtschaft, and Berlin-Siemensstadt, Electron Optics Laboratory of Siemens & Halske A.G.