The great essential for success in water culture work is _strict attention to detail_. Cleanliness of apparatus and purity of reagents are absolutely indispensable, as the failure of a set of cultures can often be traced to a slight irregularity in one of these two directions. Purity of distilled water is perhaps the greatest essential of all. Plant roots are extraordinarily sensitive to the presence of small traces of deleterious matter in the distilled water, especially when they are grown in the absence of food salts. Ordinary commercial distilled water is generally useless as the steam frequently passes through tubes and chambers which get incrusted with various impurities, metallic and otherwise, of which slight traces get into the distilled water. Loew (1891) showed that water which contained slight traces of copper, lead or zinc derived from distilling apparatus exercised a toxic influence which was not evident in glass distilled water. This poisonous effect was removed by filtering through carbon dust or flowers of sulphur. Apparently only about the first 25 litres of distilled water were toxic, in the later distillate the deleterious substance was not evident. The best water to use is that distilled in a jena glass still, the steam being passed through a jena glass condenser. For work on a large scale, however, it is impossible to get a sufficient supply of such water, while the danger of breakage is very great. Experiments at Rothamsted were made to find a metallic still that would supply pure water. While silver salts are very injurious to plant growth it was found that water that had been in contact with _pure_ metallic silver had no harmful action. Consequently a still was constructed in which the cooling dome and the gutters were made of pure silver without any alloy, so placed that the steam impinged upon the silver dome, condensed into the silver gutter and was carried off by a glass tube into the receptacle. Such water proved perfectly satisfactory so long as any necessary repairs to the still were made with pure silver, but a toxic action set in directly ordinary solder was employed. More recently a new tinned copper still has been employed with good results, but this is somewhat dangerous for general purposes, as in the event of the tin wearing off in any place, copper poisoning sets in at once. The water is always filtered through a good layer of charcoal as a final precaution against impurity.