7th. A great excess of strength over any legitimate strain, the boiler being so constructed as to be free from strains due to unequal expansion, and, if possible, to avoid joints exposed to the direct action of the fire. 8th. A combustion chamber so arranged that the combustion of the gases started in the furnace may be completed before the gases escape to the chimney. 9th. The heating surface as nearly as possible at right angles to the currents of heated gases, so as to break up the currents and extract the entire available heat from the gases. 10th. All parts readily accessible for cleaning and repairs. This is a point of the greatest importance as regards safety and economy. 11th. Proportioned for the work to be done, and capable of working to its full rated capacity with the highest economy. 12th. Equipped with the very best gauges, safety valves and other fixtures. The exhaustive study made of each one of these requirements is shown by the following extract from a lecture delivered by Mr. Geo. H. Babcock at Cornell University in 1890 upon the subject: THE CIRCULATION OF WATER IN STEAM BOILERS You have all noticed a kettle of water boiling over the fire, the fluid rising somewhat tumultuously around the edges of the vessel, and tumbling toward the center, where it descends. Similar currents are in action while the water is simply being heated, but they are not perceptible unless there are floating particles in the liquid. These currents are caused by the joint action of the added temperature and two or more qualities which the water possesses. 1st. Water, in common with most other substances, expands when heated; a statement, however, strictly true only when referred to a temperature above 39 degrees F. or 4 degrees C., but as in the making of steam we rarely have to do with temperatures so low as that, we may, for our present purposes, ignore that exception. 2nd. Water is practically a non-conductor of heat, though not entirely so. If ice-cold water was kept boiling at the surface the heat would not penetrate sufficiently to begin melting ice at a depth of 3 inches in less than about two hours. As, therefore, the heated water cannot impart its heat to its neighboring particles, it remains expanded and rises by its levity, while colder portions come to be heated in turn, thus setting up currents in the fluid.