But Roscoe’s most important contribution to inorganic chemistry was unquestionably his research upon vanadium and its compounds, which occupied him for the greater part of five years. About 1865 his attention was drawn to the occurrence of vanadium in some of the copper-bearing beds of the Lower Keuper Sandstone of the Trias which were then being worked at Alderley Edge and Mottram St. Andrews in Cheshire. He obtained possession of a large quantity of a lime precipitate, which was found to contain about 2 per cent. of vanadic acid. It was a most unpromising material, but eventually a method was worked out by which the vanadium was extracted as an ammonium vanadate: this on heating yielded vanadic acid. Great difficulty was met with in freeing the vanadic acid from accompanying phosphoric acid. Even small quantities of phosphoric acid cause the vanadic acid after fusion to solidify as a pitch-like amorphous mass. It was the writer’s privilege to assist in the early stages of this investigation, and it fell to his lot to carry out the various experiments which eventually served to establish the composition of the oxides of vanadium, the true nature of its volatile chloride, the existence of hitherto unknown oxychlorides, and of the mononitride which Berzelius had regarded as the metal, and lastly to fix its real atomic weight and to show that it was approximately 16 below that assumed by Berzelius. It was only very gradually that the true chemical relationships of vanadium revealed themselves. For a time the indications were contradictory and perplexing. The first clue was given by Rammelsberg’s observation that vanadinite is isomorphous with pyromorphite and mimetesite—analogously constituted minerals containing phosphorus and arsenic. The next significant fact to be discovered was that by the action of a reducing agent it was possible to obtain a solution of a vanadium oxide which on reoxidation to vanadic acid appeared to require as much oxygen as Berzelius’s vanadium, regarded as metal, would have needed. When it was discovered that the volatile chloride which Berzelius had considered was a trichloride and free from oxygen, in reality contained oxygen, and was analogous in constitution to phosphoryl chloride, the whole matter was rapidly cleared up, and the chemical affinities of vanadium to phosphorus, arsenic, and the other members of the trivalent group were established. This, of course, necessitated altering the formulæ of the vanadium compounds hitherto described. At the close of the College Session 1866-1867, Roscoe took away with him the laboratory journals containing the results of the inquiry as far as it had progressed, and worked at them at Roddam Hall, near Alnwick, which he had taken for the Long Vacation. The following letters have reference to this matter: RODDAM, NEAR ALNWICK, _August 26, 1867_.