The Rowe top was invented by a young man of that name at a plant where they had suffered greatly from bad distribution. The extreme simplicity of his design enabled it to be applied to the furnace with only a short shut-down, and this application resulted in an immediate and decided improvement in the practice. Unfortunately no drawing of this top is available, but it may be described without much difficulty.

A vertical plate was set up in the hopper immediately under the dumping position of each skip and in its central plane; this served to divide the skip load into two precisely equal halves. To this vertical plate were fastened inclined plates which ran down into the throat of the hopper, one of them sloping from the top corner next the skip-way to the opposite bottom corner, and its companion plate on the other side sloping from the top corner farthest from the skipway to the diagonally opposite corner of the vertical plates, so that the planes of the two plates made a letter X, but one line of the X was on one side of the vertical center plate, the other on the opposite side. The bottom ends of the inclined plates were located approximately at the quarter points around the circumference of the throat leading to the bell. Consequently as each skip load came up it was divided into two halves, and those two halves placed on two adjacent quarter points around the throat; similarly when the other skip came up its load was halved and delivered to the other two quarter points.

It is obvious that this action would be but little affected by a change in the velocity of the skip in approaching its dumping position, or by changes in the size, stickiness, etc., of the materials, or at least the changes so produced would be very much smaller than they would if the skips dumped directly and delivered all their material on one side or the other of the throat in accordance with variations in the conditions.

One of the earliest designs was that of M. A. Neeland, at the Ohio works of the National Steel Company, where the throat of the furnace was made quite high, and had set inside of it sloping vanes, roughly like a screw thread, which caused the material to whirl around as it descended through the throat, and as a result of this whirling a reasonably uniform distribution of the stock was obtained.

Very much may be done to secure regularity by correct proportioning of the main parts of the apparatus, the location of the skips in their dumping position, the size and shape of the receiving hopper which delivers the material from both skips alike to the throat, and many other details. Each has its effect on the general result, but while many devices such as those mentioned are available, it is far better not to trust to any of these alone but to eliminate the principal irregularities as far as possible by correct design in the first place; even then there will be plenty for the secondary devices to do to secure good distribution.

These irregularities were in several cases within my own knowledge very great. It did not require accurate measurement of the heaps and gaging of the size of the materials in each portion to show this; it was obvious to the most casual observer. The really remarkable feature of the whole situation was that many furnacemen did not recognize the source of their trouble or in some cases did recognize it, but were not allowed to correct it, because to do so would reflect on a design which had been selected by their superior officer.

At one plant using lump ore which I visited about ten years ago I was taken up to the top of the furnace by the stock-house foreman. Of the skip load which lay on the upper bell about two-thirds lay on the side next the skipway and consisted of all the finest portion of the load, while the one-third on the opposite side contained all the coarse! The stock-house foreman said he thought that gave a very good distribution, and when I casually asked him where the "hot spot" was of which he had told me he pointed down directly in the line of descent of the coarse material.