the figuring-die into intimate contact with the material of the glass over the entire cross-section of the ribbed portion."

"Our improved apparatus," say the inventors, "is distinct from all prior machines in the use of a forming or molding die which does not operate merely to cut into the surface, but which acts to mold the upper body of the previously-rolled mass of glass, so as to destroy what would otherwise be a level surface and impart to the body of the material different transverse sectional form."

This fairly represents the existing state of the prism glass art at the time the inventions in suit were projected into it, and the advance upon it which was thereby made. According to the Heidt (1899) patent for instance, for the molding of prism tiles, the glass in a plastic state was pressed by a plunger upon a die or mold, by which prism ridges were formed upon its surface. The prisms as made were of fairly sharp outline. But the operation being single, and the glass having not only to be displaced vertically into the angles of the die, in order to make the prisms, but also having to spread laterally, in order to reach the edges of the mold, the tiles so formed were not only of necessarily limited size, but the glass, of which they were made, was subjected to such internal molecular strain that it was very much weakened and could not be successfully cut or scored. A number of tiles, also, having to be assembled together and set in a metal frame in order to make up the required area for a window, not only did this add to the cost, but the rays of light, none too many, were materially obstructed, and a place furthermore afforded for the lodgment of dirt and dust, which, particularly in a place like Pittsburg, detracted seriously from the commercial acceptance and efficiency of the article.

The same objections do not obtain with regard to the rolled prism plate, but there were others which were encountered. By the method of making this, found in the Cummings patent (British 1898)—with which the Walsh (1869) the Stevens (1870) and the Boughton (1880), also all British, are to be classed-the glass in plastic state was poured onto a table having longitudinal V-shaped grooves, and a roller was passed over it, or the table might be flat and the roller grooved. The glass in this way could be spread to any desired extent, without structural distortion, but the difficulty was in securing sharpness and accuracy of outline, as well as depth or prominence of the prisms. If the material, for example, was too hot, it adhered to the grooves of the table or roller, leaving roughnesses where the prisms were detached; or, on the other hand, as the formation of the prisms depended on the plasticity of the glass rather than the pressure of the roller, which was but momentary, if the glass was not hot enough, becoming chilled by contact with the table, it was liable not to flow down into the angles of the grooves as it should, allowing only the production of shallow prisms at the best, and these often rounded at the point, instead of being sharp and well defined. Rolled prism plate had also to be made so thin that it could not be polished, and had thus to be put upon the market in a comparatively imperfect state.

The two processes which have so been referred to were the only ones prior to that of the complainants', which were strictly devoted to the production of prism glass. Another, however, in which it is

claimed to be involved is the Miller (British) patent (1895), also taken out shortly afterwards in the United States in a somewhat modified form. Like Cummings and the others mentioned, it had to do with the making of plate, if not prism plate, glass; and the particular part of the prolix, not to say ambiguous specifications, upon which stress is laid, is the provision by which, after the glass is rolled into a sheet, it is subjected to pressure between an upper descending table or platen and the table upon which it rests moved upwards by hydraulic force; it being further provided that the reversible double sections of the plates, which make up the face of the lower table, may have one side figured (the other being plain) so as to impart a design to the sheet, if desired, either depressed or raised, and that the upper table may have similar plates. But the impression upon the glass, which is so arranged for, is made in the initial operation, when the glass is first spread out and rolled over the figured face of the table, the same practically as in the Walsh, Stevens, Boughton, and Cummings patents. It is not effected by the upper table which has no part in it; the only purpose in subjecting the glass to pressure between the two tables, as declared in the patent, being to render the plate "compact, smooth, and clear." The idea that the upper table may be also figured is a clear misapprehension; the provision that it shall have face plates similar to those of the under table evidently meaning of similar mechanical construction-that is to say, in double sections, spaced apart, so as to allow of the introduction of heat retaining material. If similarity in every part was intended, why was it necessary to specially declare, as is done, that the plates should be provided with air vents, the same as those of the lower table? This was the view first taken by the defendants' expert; and his subsequent retraction is not convincing. It is confirmed, if that be necessary, by a reference to the specifications in the American patent to the same inventor, where nothing is said about any configurations upon the reverse side of the sectional plates of the lower table, and yet there is the same provision for similarity in those of the upper. Moreover, the patent being a foreign one, if there is any ambiguity, there is no occasion to be particularly concerned with it. Hanifen v. Godshalk, 84 Fed. 649, 28 C. C. A. 507; Hanifen v. Armitage (C. C.) 117 Fed. 845.

The second Sievert (British) patent (1892) is also referred to. The object of this, however, was not to produce plate glass, prism faced or otherwise, but to cut or stamp out flat glass objects, such as letters, numbers, lenses, frames, crosses, buttons, and rings, all of which are specified. For this purpose, a plate or sheet of glass is first rolled out in the ordinary way, and then, while still in a plastic state, a second roller, armed with sharp edged fillets or rims, designed to produce the contours or outlines of the figures desired, is passed over the sheet, stamping or marking it, without cutting through; or, in place of a roller, the figures may be stamped by a descending die. Still in sheet form, the glass is then annealed, and, after having been cooled, the figures are broken out of it; or the sheet may be ground on one or both sides, and made smooth, bright, or obscure, according to the particular effect which is to be produced; and then be broken up. The forming of prisms, among other things, is spoken of; but only, like the rest, as

detached articles, for decorative purposes, and not as constituent parts of a continuous plate or sheet.

The only other reference of which mention is specially made is the Wise (British) patent (1895), which is concerned with the production of figured plate, cut or impressed with lettering or designs. With this object in view, two or more sheets of glass are rolled one upon the other, of possibly different colors, where that effect is sought to be produced, and upon the composite mass, while still in a plastic state, a platen or die, the lower surface or face of which is suitably incised, is made to descend, and cut out of, or impress upon the glass the configuration or lettering which is desired. Where two or more colors are to be shown, as, for instance, a white decoration upon a red or blue ground, the die is made to cut deep enough to expose the under sheet or sheets. It is also provided that the design may be inscribed upon the face of the table, or may be impressed on the glass by the flatteningout roller, or by a second roller following it. And it may also be arranged to have the figures cut clear through the plate, from face to face, so as to produce open work ornamentation.

It is not contended, as I understand it, that either of the patents which have been so described is a direct anticipation of those in suit. The point made is that, having regard to the prior art as thus exemplified, there was no invention in the advance made, which at the most amounted to mere mechanical adaptation, for which the case of Daylight Mfg. Co. v. American Prismatic Light Co., 142 Fed. 454, 73 C. C. A. 570, is relied on. The Cummings patent, for rolling prism plate, was the one which was there considered; the method pursued by it, as noted above, being by spreading the molten glass on a grooved table and passing a roller over it, or, as an alternative, by having the table smooth and the roller grooved circumferentially. But this, as is well pointed out, was substantially the same operation as was shown in the preceding patents to Walsh, to Stevens, and to Boughton; the only difference being in the form of the corrugations or ribs, which, instead of being V-shaped, to make prisms, were semicircular, waving, or ratchet-toothed, a purely mechanical variation, which involved no invention to devise or make, and the patent was therefore properly declared void. The discovery of Cummings was simply that rolled prism plate could be successfully annealed and cut, and thus a commercial article of that character be produced, where, from the experience with molded prism tiles, the contrary was supposed to be the But this correction of a wrong impression entertained by the manufacturing trade had nothing inventive in it. It was the mere development of a structural quality, which could hardly impart invention to a machine to manufacture the article, in its essential mechanism not new. No such situation, however, is presented here. As the above review of the art has shown, both apparatus and process differ materially from any which had gone before. There is nothing, for instance, in the Heidt or the Cummings patents on which to base a comparison, each being distinctly a single-step process-the glass in the one being both spread and pressed by the die or plunger (there being also no attempt to make anything larger than a tile); and, in the other, both the spreading and the pressing being done, at one and the same time, by

the roller. In the Miller also, so far as the rolling and molding are concerned, the operation is single; the subsequent pressure of the glass between the upper and the lower table being simply to compact and smooth it, although it may possibly also have the effect (unsuggested) of setting or bringing out the figure. And while in the Sievert and the Wise there is the two step, or divided process, the plate being first rolled and then the desired configurations impressed upon it, yet in each of these the second is distinctly and intendedly a cutting and not a molding operation, nor is it, as in the patents in suit, accompanied with compression, with which, indeed, it does not comport.

These considerations serve also to suggest the distinctions upon which the inventive novelty of the present patents may be properly claimed. No doubt, as in many another case, old devices and methods have been made use of, but in a new and hitherto unrelated way, and with new and most important and useful results. So that, while it may not have been old to flatten out a plate by means of a roller or descending press, and there may not seem to be anything especially novel or inventive in molding prisms, any more than any other configurations, on the surface of the glass, this does not, after all, state the whole case. It is to be remembered that, in the manufacture of prism plate, two things have to be contended with and provided for upon which success depends: first, that perfect prisms in transverse parallel ribs or ridges shall be formed, which are of sharp and welldefined outline, and of the desired prominence or depth; and, second, that this shall be accomplished over an extended surface, without lateral or structural strain, so as to produce a plate that can be annealed and cut. This had been regarded as a most difficult, if not an impossible, problem which had been hitherto only partially solved. Cummings did much by establishing that, contrary to the accepted idea, rolled prism plate could be successfully annealed. But it remained to Ripley and Wadsworth, in the patents under discussion, to present a complete solution, by combining and adapting the rolling and the molding processes, first producing the sheet or plate, and then impressing it with prismatic ridges, under hydraulic force, so that every part of the die, in the words of the patent, should make "actual pressing contact with the glass beneath." For reasons already given, this pressing is clearly to be distinguished from a cutting or piercing of the glass, such as appears in other methods noticed above; and, accompanied as it is by a forcible vertical displacement of the material, it constitutes something more than the ordinary molding, having really the compressing effect of a die, essential to the formation of sharp and welldefined prism angles, but without spread or strain. Combined with the preceding rolling out of the plate, by which the completed article is turned out in a single, two-step operation, it has made possible a most valuable commercial product, which is in some respects new of its kind. The method of making it, as well as the apparatus by which this is carried out, must, therefore, be regarded as both novel and disclosing invention; and the patents which cover them must be sustained. Not so, however, with regard to the third patent, for the manufactured article or resultant product. Notwithstanding what has just been said, Ripley and Wadsworth were not the first to make prism

plate. It is Cummings, if any one, who is entitled to credit for that, which is not affected by the fact that his patent was not good. It is true that his plate had defects; the prisms not being so sharp and clean cut as they should be, and there being a limit to their prominence or depth, and neither, as it seems, could the plate be polished or ground. All of this detracted from its value and is remedied in the complainants' glass, which tells, no doubt, in the market, both with regard to its acceptability and price. But we are not to be carried away by this result. The difference is one of quality or degree, and not of kind. In structure, as in intended use, the plate is the same, whether made by the one process or the other. In fact, the only distinction between them is that this is the case, and the superiority which thereby ensues. A new article, in other words, was not created by the process in question, whatever be the favorable view of the trade, but merely an old one perfected and brought out, which is not enough. Kahn v. Starrells (C. C.) 131 Fed. 464; Id., 135 Fed. 532, 68 C. C. A. 82. the product patent therefore, the bill must be dismissed.

As to

The question of infringement remains, and as to the machine patent, at least, is not difficult. In the mechanism employed by the defendants the glass in plastic state is first spread by a roller, and immediately following it is acted upon by a succession of prism-faced sectional dies, which are forced downwards, against the action of retracting springs, by a traveling shoe or cam passing over their upper ends. Dies enough are provided to cover the whole sheet, and, as they descend in rotation, parallel prism ribs are impressed upon its surface, until the whole, in sectional areas, has been progressively operated upon. This, to a certain extent, is a modification of the Shuman (1901) patent, the idea of which was to spread the glass and form the prisms, at one and the same time, the sectional dies, without anything more being utilized to both spread and impress the sheet. This method was devised by Shuman after his attention had been called to the prism plate manufactured by the Ripley and Wadsworth process, and with knowledge of the patents which they had taken out. It was an attempt, as he says, to simplify it; but it did not work, and had to be abandoned. A roller to spread the glass was then put ahead of the dies, the rest of the mechanism being retained, and the process divided up, as in the patents in suit, with corresponding success. The first claim of the machine patent is the one that is charged to be infringed (the fourth, which was at first relied on, having been withdrawn); that is to say:

"1. Apparatus for making prism-glass comprising a spreading-table, on which the glass is rolled, a roll, and an upper pressing-die having prism-form shaping cavities, said cavities being provided with air-vents, and means arranged to force said spreading-table and die towards one another said die being shaped to make pressing-contact of the surface of the said cavities with the entire surface of the glass beneath them; substantially as described."

The defendants claim to escape infringement because their dies, as it is said, do not make pressing contact with the entire surface of the glass beneath, within the meaning of these terms. The contact which is contemplated and called for, according to this, is one which extends over the entire surface of the sheet, and not a contact by sections, as in the case of the defendants' dies; or, otherwise put, the die must be